FEDERAL COURT OF AUSTRALIA
Gill v Ethicon Sàrl (No 5)  FCA 1905
NSD 1590 of 2012
JOHNSON & JOHNSON MEDICAL PTY LTD (ACN 000 160 403)
DATE OF ORDER:
21 november 2019
THE COURT ORDERS THAT:
1. By 20 December 2019, each applicant is to notify the respondents and the Court of her election as to whether she will accept an award of damages under the relevant provisions of the Trade Practices Act 1974 (Cth) or at common law, as modified by the statutory scheme, if any, that operates in the State in which she lives.
2. By 14 February 2020, the parties bring in short minutes giving effect to these reasons.
3. In the event that the parties are unable to agree on the form of orders or all of the orders, including the form of the common questions and the answers to them for the purposes of s 33ZB of the Federal Court of Australia Act 1976 (Cth), a timetable for the filing and exchange of submissions be forwarded to chambers by 17 January 2020, with a view to a hearing in the week commencing 10 February 2020.
4. Liberty to apply be granted on two (2) days’ notice.
Note: Entry of orders is dealt with in Rule 39.32 of the Federal Court Rules 2011.
REASONS FOR JUDGMENT
Table of contents
1 This is a representative action concerning nine medical devices, all made using knitted polypropylene, a thermoplastic polymer. The devices were designed to be surgically implanted in women to alleviate either stress urinary incontinence or pelvic organ prolapse. The action calls into question the safety of those devices. Its outcome will affect thousands of women who were implanted with any of the devices in Australia and who suffered any complications caused by those devices at any time before 4 July 2017. A number of the final orders that will be made in the proceedings will bind all such women except for those who have exercised their rights to opt out of the action.
2 The medical devices used for the treatment of stress urinary incontinence which are the subject of this action are known by the trade names Gynecare Tension-free Vaginal Tape System (TVT), Gynecare TVT Obturator System (TVT-O), Gynecare TVT Secur System (TVT Secur), Gynecare TVT Exact Continence System (TVT Exact) and Gynecare TVT Abbrevo Continence System (TVT Abbrevo). When I refer to them collectively in this judgment, I use the shorthand expression “the SUI devices”.
3 The medical devices used for the treatment of pelvic organ prolapse which are the subject of this action are known by the trade names Gynecare Gynemesh Prolene Soft (Gynemesh PS), Gynecare Prolift Pelvic Floor Repair System (Prolift), Gynecare Prolift+M Pelvic Floor Repair System (Prolift+M) and Gynecare Prosima Pelvic Floor Repair System (Prosima). I refer to them collectively as “the POP devices”.
4 When I refer to the SUI devices as well as the POP devices, I call them “the Ethicon devices”.
5 In a nutshell, the applicants’ case is that all the Ethicon devices cause a number of potentially serious complications but the respondents were so driven by commercial interests that they neglected to undertake the investigations reasonably necessary to inform themselves and the community of the extent, if not also the nature, of the risks posed by the devices and to take appropriate or sufficient remedial action. To the extent that they were aware, they failed to make adequate disclosure. As the applicants put it in their opening submissions:
[A]t all times from the development of the first TVT products, to the commencement of proceedings, the Respondents were motivated by a sales-driven culture, with an emphasis on the perceived urgent need to get products to doctors before competitors. There was a focus on profit and maintaining market share, without adequate focus on safety, efficacy and potential complications. In the race to sell as many products as possible, the Respondents failed to undertake proper or adequate evaluations or [disclose] known complications and risks. 1
6 Since the commencement of this proceeding, all the POP devices have been removed from the market but, with the exception of TVT Secur, all the SUI devices are still available for sale.2
7 The devices in question were made by two foreign companies: the first respondent, Ethicon Sàrl (Société Anonyme à Responsabilité Limitée or Limited Liability Company) and; the second respondent, Ethicon Inc., both members of the Johnson & Johnson group of companies.
8 Ethicon Sàrl is a Swiss corporation and the manufacturer of a variety of polypropylene implantable medical devices, including all but one of the devices the subject of this litigation. Ethicon Inc. is an American corporation and the manufacturer of the remaining one, Gynemesh PS.3 It was also involved in the marketing of all the relevant devices through its Gynecare division. Ethicon Inc. is a subsidiary of Johnson & Johnson, an American public company, listed on the New York Stock Exchange. Ethicon Sàrl is a subsidiary of another Swiss company but the ultimate parent company is Johnson & Johnson. In these circumstances, unless it is necessary for a particular reason to distinguish between them, I will generally refer to both Ethicon companies simply as “Ethicon”.
9 Ethicon Sàrl and Ethicon Inc. supplied the Ethicon devices to the third respondent, a related Australian company, Johnson & Johnson Medical Pty Limited (JJM). JJM promoted and supplied the Ethicon devices to Australian hospitals and doctors.
10 The respondents were represented by the same lawyers and, through their senior counsel, made it clear that no point would be taken that the knowledge of one of the companies in the group was not shared by another.4
11 According to statements made on the Australian Register of Therapeutic Goods (ARTG), the purpose of each of the SUI devices is to, by way of a sling, treat stress urinary incontinence and female urinary incontinence resulting from urethral hypermobility and/or intrinsic sphincter deficiency.5 The purpose of the POP devices was to provide “tissue reinforcement and long-lasting stabilisation of fascial structures of the pelvic floor in vaginal wall prolapses”.6
12 The devices were promoted as effective in restoring normal anatomy and sexual function, as not subject to degradation or weakening by tissue enzymes, and as having high patient satisfaction. Instructions for use (IFU) supplied with the devices warned of some possible adverse reactions and contraindications. Yet, a number of potential complications, known to the respondents from the time the devices were first supplied, were not disclosed and those that were disclosed were often minimised. Some of the potential complications could occur with any form of pelvic surgery. Others were unique to implantable synthetic mesh devices.
13 This action is brought under Pt IVA of the Federal Court of Australia Act 1976 (Cth) (FCA Act) by three applicants, both on their own behalf and on behalf of other women who claim to have suffered complications from the implantation of one or other of the Ethicon devices during the relevant period. I was informed at the beginning of the trial that there are some 700 registered group members but as the class is an open one, and more than 90,000 Ethicon devices have been supplied in Australia,7 it is highly likely that there are many more.
14 The three applicants are Kathryn Gill, Diane Dawson, and Ann Sanders. Mrs Gill had pelvic organ prolapse and was implanted with a Prolift device on 12 January 2007. Mrs Dawson also had pelvic organ prolapse and was implanted with Gynemesh PS on 8 May 2009. Mrs Sanders suffered from stress urinary incontinence and was implanted with TVT on 12 March 2001.8
15 The applicants allege that the respondents contravened various provisions of the Trade Practices Act 1974 (Cth) (TPA) and the Competition and Consumer Act 2010 (Cth) (CCA). The statutory claims are based on:
s 74B of the Trade Practices Act, because it is alleged that the Ethicon devices were not reasonably fit for the particular purpose for which they were acquired by the group members and not fit for the disclosed purpose under s 55 of Sch 2 to the Competition and Consumer Act — the Australian Consumer Law (ACL);
s 74D of the Trade Practices Act, because it is alleged that the Ethicon devices were not of merchantable quality and did not comply with the guarantee given by s 54 of the ACL in that they were not of acceptable quality;
s 75AD of the Trade Practices Act in that the Ethicon devices were allegedly supplied with a “defect” and s 138 of the ACL because they were allegedly supplied with “a safety defect”; and
s 52 of the Trade Practices Act and s 18 of the ACL in that the information the respondents released in connection with the Ethicon devices, including the IFUs accompanying the products, and the way in which they were marketed and promoted was misleading or deceptive or likely to mislead or deceive consumers.
16 The applicants also allege that the respondents are liable to the applicants in negligence for:
failing to undertake any, or any adequate, clinical or other evaluation of the Ethicon devices before releasing them in Australia;
failing to conduct any, or any adequate, evaluation of safety and effectiveness of the Ethicon devices after their release in Australia; and
failing to inform them, their treating doctors, and/or the hospitals in which the treatments were administered, of the inadequate evaluations about, and the risks of, or susceptibilities to, complications of the kinds from which they suffered.
17 The respondents deny liability and, in the cases of Mrs Gill and Mrs Sanders, they contend that, in any event, their actions are statute-barred.
18 An Originating Application and Statement of Claim were filed on 15 October 2012 naming a single applicant, Mrs Julie Davis. On 6 April 2016, by orders of the Court, Mrs Gill, Mrs Dawson, and Mrs Sanders were substituted for Mrs Davis, and two sub-groups were created, one referred to in the pleading as the “Mesh Sub-Group”, mesh being a reference to the POP devices, the other as the “Tape Sub-Group”, tape being a reference to the SUI devices.
19 The trial commenced at the beginning of July 2017 and did not conclude until the end of February 2018. Evidence was adduced from 48 witnesses, 35 of whom gave oral evidence. Of the 48 witnesses, 37 were experts hailing from nine different disciplines. Each witness gave evidence individually, as my proposal that experts in the same discipline prepare joint reports and give concurrent evidence was opposed by both parties. Extensive written and oral submissions were made.
20 More than 5,500 documents were tendered, running to over 164,000 pages. Mercifully, the trial was conducted electronically, which, to some extent, eased the burden on the Court.
21 Each of the applicants gave evidence on affidavit and was cross-examined. Affidavits were also read from their husbands, only one of whom, Steven Gill, was required for cross-examination.
22 The applicants adduced evidence from four urogynaecologists. They were Andrew Korda, Wael Agur, Jerry Blaivas, and Michael Thomas Margolis. Professor Korda is an Australian. Dr Agur practises in Scotland and was described by one of the respondents’ experts as an opinion leader in Great Britain.9 Professor Blaivas and Assistant Professor Margolis hail from the United States. Urogynaecology, I should explain, is a surgical subspecialty of urology and gynaecology, which involves the diagnosis and treatment of female pelvic floor disorders.10
23 After the respondents had indicated they would not be calling him, the applicants tendered a report prepared for the respondents by Malcolm Frazer, another urogynaecologist.
24 The applicants also tendered reports and read affidavits from a number of other expert witnesses in a range of disciplines. They included the following people who, unless otherwise indicated, are based in Australia: Bilal Chughtai, a urologist, from the United States; Uwe Klinge, a general surgeon and biomaterials researcher who lives in Belgium; Bernd Klosterhalfen from Germany and Vladimir Iakovlev from Canada, both pathologists; two biomechanical engineers, Russell Dunn and Scott Guelcher from the United States; four regulatory experts, Derrick Beech, Bryan Allman from the United Kingdom, Peggy Pence and Anne Holland from the United States; three epidemiologists, Howard Hu from Canada, Cara Krulewitch from the United States, and Mark Woodward; Ian Gordon, a biostatistician; two colorectal surgeons, Alan Meagher and Anthony Eyers; Patricia Jungfer, a psychiatrist; Joseph Slesenger, a specialist in occupational medicine; and Lindy Williams and Timothy Walsh, both occupational therapists.
25 Of these witnesses, only Dr Beech, Professor Woodward, Ms Williams, and Mr Walsh were not required for cross-examination.
26 Affidavits were also read from Robyn Leake, an obstetrician and gynaecologist who treated Mrs Gill for a number of years, James Swan, an obstetrician and gynaecologist who treated Mrs Dawson, and Sandra McNeill, an obstetrician and gynaecologist who assisted in the operation in which Mrs Sanders was implanted with TVT.
27 The respondents adduced evidence from six urogynaecologists: Piet Hinoul from the United States, Pierre Collinet from France, Jan Deprest from Belgium, Alan Lam, Jan-Paul Roovers from the Netherlands, and Anna Rosamilia; an engineer, Steven McLean from the United States; Paul Santerre, a professor of biomaterials from Canada; Thomas Wright, a pathologist from the United States; three psychiatrists, Lisa Brown, Anthony Samuels, and Rosalie Wilcox; and an occupational therapist, Susan Borthwick. All of these witnesses, except Dr Wilcox, were required for cross-examination.
28 Dr Hinoul was the only witness from any of the three respondents to give evidence. Dr Hinoul is a urogynaecologist born in Belgium and educated in the United States, Belgium, and the Netherlands. He holds a PhD in bio-medical sciences from the University of Amsterdam. Since June 2014 he has held the position of Vice President–Medical Affairs at Ethicon Inc., based in Somerville, New Jersey, USA.11 He gave evidence about the Ethicon devices and their development and Ethicon’s conduct. Some of that evidence was based on company records. Not all of it concerned matters of fact.
29 Dr Hinoul joined Ethicon in 2008, when he was appointed Director of Medical Affairs – Europe, Middle East and Africa (Women’s Health and Urology), based in Paris, France.12 According to Dr Hinoul, the Director of Medical Affairs is responsible for “generating evidence for new medical devices”, “for ensuring that there is sufficient evidence to enable products to be released onto the market” and for products that are already on the market, and for “monitoring the literature and the various studies involving the products”.13 The Director of Medical Affairs is also responsible for assessing their risks and benefits. As Director of Medical Affairs, Dr Hinoul said that he was responsible for Ethicon’s pelvic floor repair and incontinence repair products. His duties included undertaking risk/benefit analyses of devices, drafting clinical evaluation reports, assessing clinical literature, presenting at panel meetings, and reviewing adverse events. He also contributed to various Ethicon documents which required “medical input”. They included IFUs supplied with the devices, patient brochures, marketing, design verification, and professional education materials.14
30 In December 2010 Dr Hinoul became the Worldwide Director of Medical Affairs (Women’s Health and Urology) for Ethicon Inc., based in Somerville, New Jersey. In June 2012 his responsibilities increased to cover other aspects of the Ethicon business and from April 2013, until his promotion to Vice-President–Medical Affairs, he held the position of Worldwide Director of Medical Affairs (Ethicon Endo-Surgery (Energy Franchise)).15
31 Dr Hinoul presented as a company spokesman. His affidavit was lengthy (363 pages) but not full and frank. It cast Ethicon’s conduct in the most favourable light. In cross-examination, Dr Hinoul was inclined not to give responsive answers to potentially uncomfortable questions and tended to be evasive where direct answers would not suit the respondents’ interests. At times he steadfastly defended the indefensible.
32 Of the Medical Affairs Directors who preceded him, only his immediate predecessor, Dr Axel Arnaud,16 remains with Ethicon.17 Dr Arnaud was a key figure in the development and post-market surveillance of a number, if not all, of the Ethicon devices. In particular, he investigated the TVT procedure in 1996 and TVT-O in 2002 and was a moving force in the development of the POP devices. No explanation was provided to the Court for his absence nor for the absence of the other Medical Affairs Directors. Presumably the respondents were content to rely solely on their reports, despite the criticisms made of them by the applicants’ experts and others.
33 The rest of the evidence consisted of documents. One legacy of the parties’ approach to tendering documents is that documents referenced by multiple witnesses or submissions were tendered multiple times. Sometimes the same document was tendered by both parties. Sometimes the same document was tendered multiple times by the same party. The result was that one document might have two or more document identification (ID) numbers. Rather than refer to multiple document ID numbers for the same documents, I have elected to use one only, and where a document is mentioned more than once in these reasons, I have endeavoured to refer to the document by the same ID number throughout. No significance should be attached to the choice of ID tender number. Moreover, a source cited in a footnote should not be regarded as the only source upon which I relied to form the view expressed in the relevant sentence or paragraph. The ID numbers have been included in the footnotes to assist the parties to identify source material. They are otherwise of no consequence.
34 In a case of this magnitude, it is impossible to refer to all the evidence and it is unnecessary to do so. My failure to refer to any particular item of evidence, however, ought not to be taken as an indication that I have not considered it. Having regard to the length of the judgment and the connection between the subject-matter of its various parts, a degree of repetition is unavoidable. Indeed, it is often necessary. I have tried to keep repetition to a minimum, although some readers may consider that I did not try hard enough.
35 At this point I also wish to make some remarks about the conduct of the trial. An extraordinary amount of work goes into a trial of this size and nature and the stakes are high. No doubt these pressures take their toll on lawyers and their clients. Despite these pressures, and notwithstanding the zealous prosecution of their clients’ interests, counsel conducted the trial with a good deal of equanimity and courtesy. This made the herculean task that confronted me far less daunting and much more tolerable than it might otherwise have been. For that, I am truly grateful.
36 On the other hand, the case could certainly have been conducted with greater efficiency.
37 It is questionable, for example, whether it was necessary for the applicants to plead so many causes of action. The applicants’ pleadings were convoluted and, owing to cumbersome cross-referencing, often difficult to follow. The parties quarrelled about points upon which they should have been able to reach agreement. Insufficient judgment was exercised about the number of experts that should be retained in each discipline, the way in which the evidence should be elicited, and the documents that should be tendered. Sources for some submissions were not always provided, which meant that I had to trawl through the evidence to find them for myself. I appreciate that the litigation was a substantial logistical exercise and that it is easy to be wise after the event. I do hope, however, that when the dust settles, both sides reflect on what could have been done better, as I have done.
38 Despite some sharp and often irreconcilable differences of opinion, there was a good deal of common ground. Unless otherwise indicated, my account of the facts is based on agreed facts, and evidence which was not contradicted or challenged.
39 Prolene is a polypropylene resin first made by Ethicon in the 1960s.18 Subject to certain conditions, it was approved by the Food and Drug Administration (FDA) in the United States in April 1969 for use in non-absorbable surgical sutures.19 Ethicon Inc. had sought approval for its product as a new “drug”. Strict conditions had to be satisfied before approval could be given. Ethicon was required to submit detailed reports of preclinical investigations, including studies on laboratory animals. In order for an animal study to be considered appropriate, proper attention had to be given to the conditions of use recommended in the proposed labelling, including whether the “drug” was intended for short or long-term administration.20 Ethicon conducted studies of tissue reactions to the sutures in rats, rabbits, and dogs. It conducted studies of the tensile strength of the sutures when implanted in rats. Ethicon was also required to submit reports of all clinical testing and all information relating to the evaluation of the safety and effectiveness of the sutures.21 After approval had been granted, Ethicon continued to test its sutures and submitted quarterly reports to the FDA.22
40 Prolene sutures continue to be widely used, including in pelvic reconstructive surgery. Dr Hinoul described them as “functional, safe and effective”.23 The applicants did not suggest otherwise.
41 Synthetic mesh was first used in the repair of hernias in the abdominal wall. Francis Usher was the pioneer. In 1958 he published on the use of synthetic mesh in six dogs to reinforce the abdominal wall and close the abdominal wall and thoracic tissue defects. Dr Usher initially used a woven material made of polyester but rapidly changed to a knitted fabric made of a high-density polyethylene known as Marlex.24 Marlex became increasingly stiff after implantation and at times caused considerable local wound complications, seroma formation, infection, and “stiff belly”.25 At some later point in time, another polymer, polypropylene, was used in the manufacture of Marlex.
42 In the early 1970s, Ethicon developed Prolene sutures into a knitted flat mesh26 and, in 1997, into a three-dimensional form known as the “Prolene Hernia System”.27 Ethicon conducted a good deal of research on Prolene sutures, both before and after it had obtained regulatory approval for their use. But the Prolene Hernia System was cleared for sale on the back of the approval for Prolene sutures, based on their supposed “substantial equivalence”. The first of the Ethicon devices was cleared for sale on the back of the regulatory approval of Prolene sutures and, in part, because of its supposed “substantial equivalence” to the Prolene Hernia System, despite the differences in design, use, anatomy, and site-specific considerations.
43 Stress urinary incontinence is the involuntary leakage of urine during activities such as coughing, sneezing, lifting, laughing or exercising.28 It is to be contrasted with urge urinary incontinence which is involuntary leakage of urine accompanied by a compelling urge or need to urinate.29 For completeness, overactive bladder is a symptom complex characterised by urgency (the compelling need to urinate), usually with urinary frequency and nocturia (rising at night to urinate) and, sometimes, with urge incontinence.30 Mixed urinary incontinence is a combination of stress urinary incontinence and urge urinary incontinence.31
44 The only symptom of stress urinary incontinence is leakage of urine during activities accompanied by increased abdominal pressure like coughing, sneezing, lifting, running, and various other forms of physical exercise. It does not affect any aspect of the anatomy. It affects the quality of the sufferer’s life. It is undoubtedly distressing. But it is never life-threatening.
45 The urethra (the duct or tube extending from the bladder down to the wall of the upper vagina through which urine is carried out of the body) and the bladder are supported by pelvic floor muscles, which contract during coughing, sneezing and exercise to prevent leakage of urine.
46 Whereas urge incontinence is a “bladder problem”, caused by involuntary contraction of the detrusor muscle), stress urinary incontinence is a “sphincter problem”.32 The urinary sphincter is an arrangement of muscles situated closest to the bladder. Ordinarily, voluntary urination causes the sphincter to relax and the detrusor muscle (the smooth muscle in the wall of the bladder) to squeeze or contract, resulting in the expulsion of urine from the bladder down the urethra and out of the body. Weakness in the muscles or damage to the bladder neck support can result in leakage.
47 There are two types of stress urinary incontinence: urethral hypermobility and intrinsic sphincter deficiency. Urethral hypermobility describes the situation in which the urethra has moved outside the pelvis and activities such as coughing or sneezing (known as Valsalva activities) put added pressure on the bladder, causing leakage. Intrinsic sphincter deficiency simply refers to weak urethral sphincter muscles or poor urethral closure function.33
48 Risk factors include pregnancy and vaginal birth and other conditions which cause an increase in abdominal pressure, such as obesity, chronic cough, chronic heavy lifting and constipation.
49 Treatment is always elective.
50 There are surgical and non-surgical options.
51 Non-surgical (or conservative) treatments include general lifestyle changes, pelvic floor exercises, and the use of continence devices such as a pessary. Conservative treatments are not always successful. Some women experience alleviation of their symptoms; some do not and elect to have surgery.
52 Traditional surgical treatment options include:
(1) Burch colposuspension, named after Dr John Burch who first described it in 1961,34 which is used to correct urodynamic stress incontinence.
(2) needle suspension procedures;
(3) sling procedures using either the patient’s own connective tissue (fascia) (known as autologous slings) or foreign graft material; and
(4) use of urethral bulking agents, involving injection of a variety of different substances around the bladder neck and into the urethral sphincter, to thicken the urethral wall so as to provide greater urethral resistance during increases in abdominal pressure.35
53 In Australia, at least until the late 1990s, the Burch colposuspension was described as “the gold standard” surgical treatment for stress urinary incontinence. Associate Professor Rosamilia said that it had an 85% cure rate at five years, although she did not cite a source.36
54 The Burch colposuspension is performed either through abdominal incision (open abdominal surgery) or laparoscopically (keyhole surgery) in which the retropubic space (the space between the pubic bone and the bladder) is dissected and the neck of the bladder is elevated (or suspended) by sutures passed through the Cooper’s ligament, which borders the femoral ring, and attached to the pubic bone.37 The effect on the anatomy is to prevent or minimise descent of the proximal urethra during increases in abdominal pressure and to create a sling of tissue against which the urethra is compressed;38
55 The object of autologous slings is to cause the urethra to hit against something during a Valsalva activity so that the urethra momentarily shuts off. A Valsalva manoeuvre, named after an Italian anatomist, Antonio Valsalva, is the action of attempting to exhale with a closed mouth and nostrils. It increases pressure in the middle ear and chest and is used as a means of equalising pressure in the ears.39 The slings are placed under the neck of the bladder and are commonly referred to as “pubovaginal slings”.
56 The autologous fascial sling procedure was first described in 1907.40 Although it was a mode of treatment favoured by Professor Blaivas, it never achieved widespread popularity, because of its high complication rates, particularly in inexperienced hands.41 The complications with which it was associated included increased rates of urinary tract infection, urge incontinence, voiding dysfunction, erosion, and, occasionally, the need for surgical revision to improve voiding.42 It is not in dispute that the complications include continued incontinence, voiding dysfunction, urinary retention, pain, and dyspareunia (pain with sexual intercourse).43 Historically, the procedure was reserved for cases of intrinsic sphincter deficiency or prior surgical failure but there was evidence that it was an effective treatment for all types of stress incontinence with acceptable “long-term” efficacy.44 In 1998, Chaikin et al published the results of a prospective and retrospective study of 251 women with all types of stress incontinence who underwent pubovaginal fascial sling surgery by a single surgeon, and were followed up after a median period of three years (with a range of one to 15 years). Chaikin et al (1998) claimed that they had “demonstrated that the procedure [could] be performed in a reproducible fashion with minimal morbidity”.45 They said that postoperative urinary retention should be minimal if the sling is not tied with excessive tension. They acknowledged, however, that persistent and de novo urge incontinence remained “a vexing problem”. In cross-examination, Professor Blaivas, who was a co-author of the Chaikin article, suggested that in 1998 a five year follow-up would have been considered “long term”, although he would not consider it to be long term now.46
57 The SUI devices are midurethral slings. Midurethral slings may be surgically implanted in a number of ways. They may pass under the midurethra, run behind the pubic bone through the retropubic space (the cavity between the urethra and the muscles above it, the rectus muscles and the peritoneum) and exit above the pubic bone. This type of sling is known as a retropubic sling. Then there are those which are also placed under the midurethra but, unlike the retropubic slings, they pass through the obturator space and exit in the groin. They are called transobturator slings.
58 There are also single-incision slings, sometimes called “mini-slings”, which also pass under the midurethra, but are typically shorter in length than the multi-incision midurethral slings, and can be placed via either the retropubic or the transobturator route.
59 Despite the variety of approaches and materials used, it is widely accepted that all these slings should be placed under “minimal tension” to prevent the development of additional voiding dysfunction, such as obstruction with incomplete voiding.47
60 Every one of these surgical treatment options is attended by risks, although the nature and extent of the risks vary from procedure to procedure. The complications associated with the Burch colposuspension, for example, include urinary tract infection, urinary retention, urethral obstruction, de novo overactive bladder, haematuria, neurologic symptoms, pelvic pain, dyspareunia, and death. The most common complications associated with the other traditional forms of surgery are voiding dysfunction and urethral and urinary retention. They are managed conservatively with either intermittent self-catheterisation or an indwelling catheter. Refractory pain, that is pain that is resistant to treatment, is exceedingly rare after any of these procedures.48 The complications of treatment with the SUI devices are discussed below.
61 Regardless of the nature of the surgery, clinical outcomes appear to be worse for patients who have had previous surgery for stress urinary incontinence.49
62 It is common ground that, before deciding on the most appropriate course or method of treatment, a treating surgeon would consult with the patient, obtain her medical and surgical history, and assess her clinical needs.
63 As I have noted, the SUI devices are midurethral slings. The object of midurethral and pubovaginal slings is similar, but, unlike pubovaginal slings, midurethral slings are placed under the midurethra. Presumably for this reason, they are sometimes referred to as “suburethral slings”. The SUI devices are designed to be inserted without tension.
64 Tension-free vaginal tape was the brainchild of an Australian pelvic floor surgeon, Peter Petros. Its story begins in 1986 with two unrelated observations, which I gather were made by Dr Petros. The first was that a haemostat (an instrument for preventing blood flow by compression of a blood vessel) applied immediately behind the pubic symphysis (the immovable joint between the pubic bones in the centre of the pelvis), at the level of the midurethra, controlled urine loss on coughing. The second was that, on implantation, polytetrafluoroethylene (PTFE, such as Teflon) tape created a collagenous tissue reaction around it. These observations generated two hypotheses: that a loose pubourethral ligament causes stress urinary incontinence and that a tape implanted in the exact position of the damaged pubourethral ligament would create a collagenous “neoligament” to reinforce it, restoring function and continence.50
65 Dr Petros began animal studies at the Royal Perth Hospital in 1987 using a tape made from Mersilene, a polyester material produced by Ethicon.51 The tape was inserted in the vaginal cavities of 13 dogs for periods from six to 12 weeks. Apart from a sticky yellow vaginal discharge, there were apparently no “ill effects”.
66 Human studies began over the following two years and continued into the 1990s. Between 1998 and 1999, 30 women underwent a “combined intravaginal sling and tuck” procedure, which involved creating an artificial pubourethral ligament and tightening the suburethral vagina. At the end of six weeks, all patients remained continent.52 Within two weeks of the tape being removed, however, 50% of the patients reported recurrent incontinence.53
67 In the early 1990s, doctors from the Department of Obstetrics and Gynaecology at the University Hospital in Uppsala, Sweden, developed a technique for treating female urinary incontinence under local anaesthetic. Their objective was to restore the pubourethral ligament and the suburethral vaginal hammock. The technique was described in a paper by Professor Ulmsten and Dr Petros entitled “Intravaginal Slingplasty (IVS): an ambulatory surgical procedure for treatment of female urinary incontinence” published in 1995. 54 It involved placing under the urethra a tape-like strip of mesh in a U-shape sling formation using a specially designed instrument referred to as a “tunneller” introduced through paraurethrally dissected canals. The technique was used on 50 patients and drew on previous experimental and clinical studies. Thirty-eight of the 50 patients suffered from genuine stress incontinence, objectively verified, and 12 had symptoms and signs of both urge and stress incontinence. The mesh used was Mersilene (in 37 patients), Goretex (in five), Teflon (in six), and Lyodura (in two). Patients were evaluated post-operatively at intervals of one month, six months, and one to two years.55
68 The work of Ulmsten and Petros appears to have been driven by the prevalence of urinary leakage in post-menopausal women and the fact that the surgical treatment was generally extensive and required general anaesthesia. Ulmsten and Petros considered that there was “a strong need for more precise and simple ambulatory surgical methods to treat female urinary incontinence”. They specified the following criteria for these procedures:
(1) the type of defect causing the patient’s leakage must be defined and described in the work-up, as should the manner in which surgery can restore the dysfunction;
(2) the procedure should allow the patient to return to work shortly after the operation;
(3) the operation should be carried out under local anaesthesia, enabling the patient’s cooperation during the procedure and reducing costs and the number of services otherwise required; and
(4) both subjective and objective evaluation of the patient should be carried out, “recognizing her symptoms and the quality of life situation both before and after the operation”.56
69 Early results appeared promising. The authors wrote that 39 patients (78%) were completely cured of their stress incontinence and another six (12%) reported a considerable improvement of their urinary incontinence, leaking only occasionally. They also wrote that there were no intra or post-operative complications. They acknowledged, however, that, since the technique was new, no long-term results were available, although they speculated that the long-term results of their prosthetic sling would be similar to those involving “conventional” sling procedures. Ulmsten and Petros concluded that if their results were to be reproduced in ongoing follow-up studies including more patients (how many more they did not say) their method “should be considered as a promising new ambulatory procedure for treatment of female urinary incontinence”.57
70 It transpired, however, that, upon surgical insertion, the Goretex and Mersilene tapes were rejected in eight to 10 percent of patients. Further work followed, and in 1996 the results of another study by Ulmsten and others were published.58 This time the tapes were fashioned from Prolene and the technique was said to have been improved. I return to this study later in these reasons. It is sufficient at this point to note that the work of Professor Ulmsten and his team culminated in the production of the Ethicon midurethral sling known as TVT.
71 As noted above, the SUI devices the subject of the current litigation are TVT, TVT-O, TVT Secur, TVT Exact and TVT Abbrevo.
72 Each of the SUI devices consists of a sterile, single use device comprising polypropylene mesh in the form of a tape and a set of instruments to facilitate mesh implant placement. Each was designed for the treatment of stress urinary incontinence resulting from urethral hypermobility or intrinsic sphincter deficiency. TVT and TVT Exact were designed for use as pubourethral slings. TVT-O, TVT Secur and TVT Abbrevo were designed for use as suburethral slings.
73 The mesh used in the SUI devices is made from Prolene, which, as I mentioned earlier, was developed by Ethicon in the 1960s as a suture material. Prolene consists of a non-absorbable polypropylene base resin and a number of additives. Polypropylene is a polymer made from the monomer propylene, a component of natural gas made from carbon and hydrogen.59 A polymer, I interpolate, is a material composed of long chains of chemical building blocks, called monomers, bonded together. Propylene is polymerised into polypropylene in a chemical reaction in which propylene molecules (monomers) are combined together in a step-wise fashion to ultimately form linear, chain-like macromolecules60 which turn into flakes, chips or pellets.61 In the manufacturing process, efforts are made to control the lengths of the individual chains (referred to as polymer molecular weight) and the variation in lengths between the different chains (referred to as polydispersity).62 The mechanical properties of polypropylene are directly influenced by its molecular weight and/or molecular weight distribution, which is the name given to the range of polymer chain lengths.63
74 Prolene fibres are manufactured by melting polypropylene flakes or pellets in a heated extruder which exit through a spinneret (a metal end cap on the extruder with holes of various shapes) forming strands or filaments.64 During this process, various additives are added to the polypropylene, including:
(1) two antioxidants: dilauryl thiodipropionate (or DLTDP), to improve long-term storage of the resin and fibre and to reduce the potential oxidative reaction with UV light, and Santonox R, to promote stability during compounding and extrusion;
(2) two lubricants: calcium stearate and Procol LA-10 (previously Luberol), to help reduce tissue drag and promote tissue passage; and
(3) a colourant: copper phthalocyanate (or CPC), to enhance visibility when it is desired to have a blue dyed monofilament fibre rather than a clear monofilament fibre.65
75 The additive package in use in 2003 was the same as that used in the original 1991 formulation save that the Santonox levels were reduced slightly (by 0.05%).66 There has been no change since. Neither side contended that the reduction in Santonox was of significance.
76 The single strands of Prolene (referred to as monofilaments) were then knitted into a specific pattern to form the mesh and the mesh was then scoured and annealed.67
77 Finally, the mesh was cut into various sizes. The mesh was originally mechanically cut and sealed using a hot knife.68 From June 2006, laser cutting was introduced after Ethicon was inundated with complaints from surgeons about particle loss. 69 In a PowerPoint presentation from August 2006,70 no doubt designed to showcase the benefit of laser-cut over mechanically cut mesh, samples of mesh that had been laser-cut were compared with samples that had been mechanically cut after both sets had been pulled to 50% elongation and then relaxed.71 The mechanically cut (MCM) samples were described as follows:
The MCM samples show the degradation of the structure of the mesh in certain areas where, because of particle loss, the knit has opened and a portion of the construction has been lost. The area may also be stretched and narrowed resulting in roping due to this occurrence.
78 Accompanying photographs illustrated the description and also showed fraying of the mesh, a phenomenon described by Carol Holloway, Product Complaint Analyst, Worldwide Customer Quality, for Gynecare, in a letter dated 12 October 2005 as “inherent in the product based on the mesh construction”.72
79 In contrast, the laser-cut samples showed no degradation of the structure of the mesh, because no, or nearly no, particles were lost. The knit construction stayed intact and there was no roping. Although the mesh could be stretched and narrowed it was “generally less than the [mechanical cut mesh]”.73
80 The TVT system (sometimes referred to as “TVT Classic” or “TVT Retropubic”) is a sterile, single patient use device made up of one piece of undyed or blue Prolene, in the form of a tape, covered by a plastic (polyethylene) sheath (cut and overlapping in the middle) and held between two curved stainless steel needles (trocars) that are bonded to the mesh and sheath with plastic collars. The trocars (also called introducers) have two parts: a handle and an inserted threaded metal shaft designed to facilitate the passage of the tape from the vagina to the skin of the abdomen.74
81 The tape is 1.1cm wide and 45cm long.75 According to the description in the TVT technical file, maintained by Ethicon, it is about 0.7mm thick.76 It is inserted at the midurethral region, to create a sling on which the urethra can rest when there is a sudden increase in abdominal pressure.77
82 When the device was first supplied, however, it appears that the tape was shorter. It was 40cm, not 45cm, in length.78
83 The images below show the tape (left) and the tape together with the trocars (right):
84 TVT requires the surgeon to use a retropubic (bottom to top) transvaginal surgical approach. The sling is inserted by means of the trocar through a small vaginal incision under the midurethra, passes vertically behind the pubic bone through the retropubic space, and exits through the skin in the lower abdomen. This is known as a “U” shaped sling orientation. The surgeon is unable to see where the trocar goes. It is not in dispute that the blind passage of the trocar creates a risk of perforation of the bladder and urethra and of damage to blood vessels and nerves in the retropubic space.79 Serious complications due to perforation of the large vessels and intestinal viscera have also been reported.80
85 TVT gained regulatory approval in Europe in 199781 and was cleared by the FDA on 28 January 1998.82 On 21 July 1998 it was approved by the Therapeutic Goods Administration (TGA) as a class IIB device for supply in Australia. It was first sold here in October 1999.83
86 Dr Hinoul deposed that TVT was cleared by the FDA on the basis that it was substantially equivalent to the ProteGen sling manufactured by Boston Scientific, one of Ethicon’s competitors, which was already on the market.84 Both TVT and ProteGen were intended for use as pubourethral slings for the treatment of stress urinary incontinence but, as Dr Bryan Allman, a regulatory expert who worked for Boston Scientific at the time, observed, the two devices were very different.85 For a start, they were made of different materials. TVT was made of knitted filaments of extended polypropylene (Prolene). ProteGen was made of woven polyester impregnated with bovine collagen. TVT was entirely non-absorbable. ProteGen had an absorbable collagen coating. The tools were also different. TVT was supplied with a reusable introducer and rigid catheter guide. ProteGen was supplied with disposable single use instruments: a bone locator, a suture anchor system, a suture passer, and a suture spacer. Both were intended for anterior fixation, but TVT was to be fixed to abdominal skin by friction at first and later by tissue ingrowth, whereas ProteGen was fixed by a suture anchored to bone.86
87 Certain modifications were made over time. I have already mentioned the change in the cutting technique. In addition, in 1999 the diameter of the needles used with the device was reduced from 6mm to 5mm.87 A blue version of the TVT tape was developed in October 2001, by incorporating a pigmented blue fibre into the knitted mesh. The stated purpose of this development was to enhance visibility in the surgical field.88
88 Fundamentally, however, the tape did not change. It has always been made from Prolene and the composition of Prolene has not altered in any material respect.
89 TVT-O is also a sterile, single patient use device, which consists of a piece of undyed or blue Prolene mesh tape covered by a plastic sheath overlapping in the middle. Plastic tube receptacles are attached at each end. TVT-O is made from the same Prolene mesh used in TVT and its dimensions are identical. It differs from TVT in only two respects: the tools or instruments supplied with the device and the method of attachment. The tools or instruments consist of two helical passers, shown in the photograph below taken from Dr Hinoul’s affidavit, and a winged guide, which facilitates the passage of the helical passers through the obturator membrane.
90 The helical passers are inserted through the vagina but then pass through the obturator foramen, a large opening between the ischium (the curved bone forming the base of each half of the pelvis and one of the three bones that fuse to form the hip) and the pubic bones, rather than the retropubic space as with TVT and, unlike TVT, exit through the upper legs, rather than the abdomen.89
91 Like TVT, TVT-O was developed by a surgeon. As we shall see, Ethicon’s interest in such a device was largely, if not entirely, motivated by its concern about increasing competition in the marketplace that followed the early success of TVT.
93 TVT Secur was indicated for use in women as a suburethral sling for the treatment of stress urinary incontinence resulting from urethral hypermobility and/or intrinsic sphincter deficiency.92 But it was significantly different from its predecessors in composition, size, and method of fixation.
94 The “Gynecare TVT Secur System” consisted of the implantable device together with a set of instruments to facilitate the placement of the device in the pelvis.
95 The device was a piece of blue Prolene mesh in the form of a tape (8cm long and 1.1cm wide) sandwiched between layers of absorbable fleece (2cm long) made from Vicryl (polyglactin 910) and PDS (poly-p-dioxanone) undyed yarn. The instruments were two metal inserters, a finger pad, a protective cover, and a release wire.93
96 This is what it looked like, with the instruments:
97 The tape was inserted via a single, small vaginal incision. It could be implanted in either a “U” position (which is comparable to a retropubic position) or a “hammock” position (which is comparable to the transobturator sling).94 In the former case, the mesh is implanted towards the retropubic space but not through it. In the latter case, the tape is implanted towards the obturator opening but does not exit the opening or the externus muscle. In neither case is there an exit site.95 Unlike TVT and TVT-O, the mesh in TVT Secur was not covered with a plastic sheath96 and the mesh used to make the tape was always laser-cut rather than mechanically cut.97
98 TVT Secur was the only one of the SUI devices that required a single incision.
99 A single-incision sling is defined as a sling that does not involve either a retropubic or transobturator passage of the tape or trocar and involves only a single vaginal incision so there are no exit wounds in the groin or lower abdomen.98 The background to the development of single-incision slings was described by Nambiar et al (2014):
Historically many types of surgery have been performed to treat women with stress urinary incontinence. Over the past 10 years, the accepted standard technique has been the mid-urethral sling operation, whereby an artificial tape or mesh is placed directly beneath the urethra and is anchored to the tissues in adjacent parts of the groin or just above the pubic bone. Examples of such slings that are commonly used are tension-free vaginal tape (TVT™) and transobturator tape (TOT). These operations are usually quite successful, with success rates approaching 80%or 90%. However, they have been shown to result in significant side effects, which can be bothersome and sometimes even dangerous, such as damage to the bladder caused by tape insertion, erosion of the tape into the urethra during the healing period or chronic thigh/groin pain.
In an effort to maintain efficacy while eliminating some of the side effects, a new generation of slings has been developed, called ’single incision slings’ or ’mini-slings’ (sic); these slings are the subject of this review. They are designed to be shorter (in length) than standard mid-urethral slings and do not penetrate the tissues as deeply as standard slings. It was therefore thought that they would cause fewer side effects while being no less effective…99
100 TVT Secur was cleared for sale in the United States on 28 November 2005 and in Europe on 4 May 2006.100 It was launched in Australia in April 2007,101 but sold for less than a year. Sales were halted in March 2008 and its registration was cancelled by the TGA in June 2012.102
101 According to a Clinical Expert Report signed by two of Ethicon’s medical directors on 28 February 2006,103 TVT Secur was intended to address the complication of bladder perforation (with the retropubic approach) and thigh pain (with the obturator approach) using a less invasive procedure.
102 TVT Exact is a retropubic sling made of Prolene mesh, just like TVT, and, despite the respondents’ submission to the contrary,104 at 1.1 x 45cm, the dimensions of its sling are identical to the TVT sling.105 Unlike TVT, however, all the instruments supplied with the device are fully disposable. The mesh is covered by a clear plastic implant sheath and held between two white trocars which are bonded to the implant and the implant sheath. The trocar is slightly thinner (at 4.2mm) than the TVT trocar (at 5.0mm).106
103 TVT Exact was released to the Australian market in July 2010.107
104 TVT Abbrevo gained regulatory clearance in the United States on 1 July 2010, and in Europe in September 2010. It was first supplied in Australia in October 2010.108
105 It is a transobturator sling like TVT-O.109 It is made up of a single piece of laser-cut blue Prolene mesh tape covered by clear polyethylene sheaths and was supplied with a set of instruments to facilitate placement of the device: a placement loop, a pair of helical passers, and a winged guide.110 The mesh used in the system, however, is considerably shorter than TVT and TVT-O (12cm as against 45cm111) and has sometimes been called a mini TVT-O. Unlike TVT-O, the mesh does not need to be trimmed on either end once it has been implanted.112
106 According to a clinical evaluation report dated 17 August 2010 and signed by Dr Hinoul, TVT Abbrevo was designed to reduce the amount of mesh left behind in the body, to reduce pain possibly caused by the presence of the tape in the adductor muscles, and to improve the ergonomics of the TVT-O procedure. Dr Hinoul claimed that this modified TVT-O device “leads to reduction of tissue trauma, and reduction in the total length of mesh left behind in the body to a total of 12cm”.113
107 Pelvic organ prolapse is the downward displacement of a pelvic organ, which, in the case of a woman means the uterus, the different vaginal compartments or neighbouring organs such as the bladder, rectum or bowel.114 As Professor Deprest stressed, it is an anatomical finding, that is, a change from the normal anatomy, and it may not be associated with any symptoms.115
108 Prolapse occurs when the muscles, ligaments and fascia (the network of supporting tissues that hold the organs in their correct positions) fall or slip out of place.
109 Pelvic organ prolapse is defined as an anatomical change in which there is downward displacement of a pelvic organ.116 There are three different types of pelvic organ prolapse: uterine prolapse in which the uterus descends, cervical prolapse involving the descent of the cervix (the neck of the uterus) and vaginal prolapse involving the descent of one or more of the compartments of the vagina.117 In a prolapse of the anterior compartment of the vagina, either the bladder or uterus (or, in the absence of a uterus the vaginal vault) bulges into the front wall of the vagina.118 This is also referred to as a cystocoele (from the Greek kustis meaning bladder and kele meaning “tumour”) or urethrocoele. In an apical compartment prolapse (sometimes called a vault, uterine, or middle compartment prolapse), the uterus descends or herniates into or outside the vagina or, if the uterus has been removed, the vaginal vault descends.119 In a posterior compartment prolapse, the rectum (the lower part of the large bowel) or part of the small intestine bulges into the upper part of the back wall of the vagina. The former is known as a rectocoele and the latter as an enterocoele.120
110 Diagnosis is based on a combination of information provided by a patient to her doctor about her medical history and symptoms, and a medical examination. The major cause is vaginal childbirth, which requires the levator ani muscle to substantially distend.121 A levator muscle is a muscle whose contraction causes a part of the body to be lifted. The levator ani muscle supports the pelvic organs and helps to prevent urinary incontinence. Professor Korda explained that, when a woman gives birth vaginally, the levator ani muscle in the average woman stretches by 150%. It is known that if that muscle is stretched in a laboratory setting beyond 150% it will never be able to return to normal tensile strength or length. This overstretching causes the natural “hammock” support of the pelvic floor to disappear and the downward descent of the pelvic organs to occur.122
111 Prolapse is described in stages, indicating the extent of the descent. Different staging or grading systems have been devised, including the Baden-Walker classification and the Pelvic Organ Prolapse Quantification (POP-Q) classification.
112 Risk factors for pelvic organ prolapse include pregnancy and childbirth; ageing and menopause; and conditions that cause excessive pressure on the pelvic floor, such as obesity, chronic cough, and chronic constipation.
113 Other factors which have been implicated in pelvic organ prolapse are asthma, genetic predisposition, and metabolic disorders. Age is universally cited as an established risk factor and uterine prolapse appears to be age-related. But Professor Korda’s evidence was that the general proposition was not supported by unexpectedly high rates of mild or moderate prolapse in young or asymptomatic women and increased tissue stiffness after menopause. He said that anterior and central compartment prolapse appear to deteriorate up until the age of 55 but thereafter improvement occurs.123
114 Symptoms of pelvic organ prolapse in women can include a heavy dragging feeling in the vagina or lower back; the feeling of a lump or bulge in or outside the vagina; urinary symptoms such as slow urinary stream, a feeling of incomplete bladder emptying, urinary frequency (needing to pass urine eight or more times a day), 124 or urgent desire to pass urine; 125 bowel symptoms such as difficulty moving the bowel or a feeling of incomplete bowel evacuation, or needing to press on the vaginal wall to empty the bowel; and pain, including lower back pain.126
115 Lower urinary tract symptoms, such as incontinence, are sometimes associated with prolapse but are not caused by it. Recurrent urinary tract infections can be caused by prolapse due to incomplete bladder emptying.
116 Prolapse can affect sexual function and cause dyspareunia, inability to penetrate the vagina due to obstruction, vaginal laxity, and loss of libido.
117 Depending on its severity or extent, while pelvic organ prolapse is not a life-threatening illness, as Associate Professor Rosamilia observed it can have a drastic effect on a woman’s quality of life.127
118 As with stress urinary incontinence, there are both non-surgical and surgical treatment options.
119 Non-surgical treatments include lifestyle interventions, pessaries and pelvic physiotherapy. Symptoms may be improved with the use of vaginal oestrogen to improve the epithelium in post-menopausal women, avoidance of constipation and chronic cough, and training the pelvic floor muscles.128
120 Surgical treatment options include reconstructive surgery and vaginal closure or removal surgery (also known as obliterative procedures). The objective of all these treatments is to correct the prolapse while maintaining or improving vaginal sexual function and remedy lower urinary tract symptoms and disorders. Although it is not always feasible, their broad purpose is to restore the so-called normal pelvic anatomy.129
121 Reconstructive surgery may be accomplished using vaginal or abdominal approaches.
122 One of the forms of vaginal closure is colporrhaphy (literally, suturing the vagina in place). This is a procedure used to repair an anterior or posterior vaginal wall prolapse. It involves an incision into the relevant compartment of the vagina and plication (folding) of the pubocervical fascia (connective tissue forming layers between the vagina and bladder) with sutures. The objective is to repair the midline fascial defects by using and tightening up the stretched-out fascia.130
123 Apical prolapse can be repaired vaginally using sacrospinous ligament fixation (also called colpopexy), which involves suspension of the apex of the vagina to the sacrospinous ligament by means of (usually absorbable) sutures or uterosacral ligament suspension, which involves suturing the vagina to the uterosacral ligaments.131
124 A culdoplasty is performed to correct an apical prolapse after hysterectomy. Sutures are used to suspend the vaginal vault at the origin of the uterosacral ligaments, which support the vagina, and to close off the pouch of Douglas (a small area between the uterus and the rectum).132
125 The abdominal procedures include sacrocolpopexy (also called “sacral colpopexy”), which can be performed laparoscopically, robotically or by open surgery via an abdominal approach in the lower abdomen. This procedure is used to correct apical prolapse. The apex of the vagina is stitched or fixed to the sacrum by means of sutures and a small amount of surgical mesh.133 The procedure is also used to correct a combination of apical, anterior and posterior prolapse. It is only available for women who have undergone a hysterectomy. Often the two procedures are performed in the same operation. Another option is an abdominal hysteropexy, which is also used to correct apical prolapse but in which the uterus is preserved. As with a sacrocolpopexy, the apex of the vagina is stitched or fixated to the sacrum by means of a small quantity of surgical mesh, and the surgery can be performed either laparoscopically or by open surgery via an abdominal approach in the lower abdomen.134
126 Obliterative procedures include vaginal closure surgery (colpocleisis), which involves stitching the vaginal walls together to create a barrier in order to prevent the prolapse re-occurring,135 or total colpectomy, which involves the total excision of the vagina in a woman with no uterus and vaginal eversion (that is, where the vagina is turned outwards or inside out).
127 Surgery which uses the patient’s own tissue is commonly referred to as “native tissue repair”.
128 Prosthetic material used during surgery may be made from absorbable material such as animal tissue, non-absorbable synthetic material, or a combination of absorbable and non-absorbable material.
129 There is no single best approach for all patients. Treatment is individualised according to each patient’s symptoms.
130 All surgical treatment options are associated with risks, although the evidence indicates that complications from native tissue repair are generally short-lived and treatable.136 Professor Korda listed them as failure; injury to adjacent and contiguous organs, such as the bladder, urethra, ureters, bowel, major blood vessels, and nerves; the development of lower urinary tract symptoms such as urinary incontinence, narrowing of the vagina, and dyspareunia.137
131 The medical devices used for the treatment of prolapse which are the subject of the current litigation are Gynemesh PS, Prolift, Prolift+M and Prosima.
132 Prolift, Prolift+M, and Prosima were intended for tissue reinforcement and long-lasting stabilisation of fascial structures of the pelvic floor in vaginal wall prolapse where surgical treatment was intended, either as mechanical support or bridging material for the fascial defect. Gynemesh PS had the same indication until March 2013, when its indication for use was changed, as discussed below at .
133 Before any of these devices was available, however, it appears that surgeons concerned about the failure rate with native tissue repairs experimented with synthetic hernia mesh to treat pelvic organ prolapse. Dr Hinoul was one of these surgeons. He said he began to use a polypropylene hernia mesh to treat pelvic organ prolapse in the late 1990s or early 2000s. He used a transvaginal approach, inserting the mesh through the obturator foramen using needle drivers and anchoring it in the muscles of the groin using a technique similar to that later used to implant the Prolift device.138
134 The meshes used in all these devices were made from non-absorbable polypropylene (Ethicon’s Prolene Soft, except for Prolift+M, which is made from Gynemesh M.
135 Prolene Soft mesh was developed by Ethicon in the late 1990s for use in hernia repair and approved by the FDA for that purpose. Like Prolene it was and is made from knitted filaments of polypropylene.139 Prolene Soft was described in a memorandum from Dr Ning Wang of Corporate Product Characterization at Ethicon Inc. as “a revised construction [Prolene] mesh, fabricated of knitted filaments of natural color and blue pigmented polypropylene identical in composition to that used in [Prolene]”.140 It differs from Prolene, however, in two respects. It has a smaller filament diameter (at 3.5mm as against 6mm), which is said to give it a softer feel,141and it has a different knit design.142
136 Gynemesh M is a partially absorbable mesh consisting of approximately equal parts of Prolene Soft and Monocryl, an absorbable poliglecaprone-25 monofilament fibre.143
137 Gynemesh PS received regulatory clearance in the United States on 8 January 2002, in Europe on 20 March 2003, and in Australia on 26 May 2003. It was first supplied to the Australian market in July 2003.144
138 Gynemesh PS is made from Prolene Soft mesh.145 The “PS” in its name is an acronym for Prolene Soft. It consists of rectangular sheets of mesh that may be cut as desired by the surgeon. This is what it looks like.
139 According to Dr Hinoul, the only difference between Prolene Soft and Gynemesh PS was that Gynemesh PS was cleared for the specific indication for use in “tissue reinforcement and long-lasting stablization of the pelvic floor in vaginal prolapse”, whereas Prolene Soft was cleared for the general indication for use in “repair of hernia or other fascial defects that require the addition of a reinforcing or bridging material”.146 Ethicon treated the change of indication as a matter of no consequence. Its position was that the change was not a new intended use, which might have led to additional regulatory obligations, but a subset of the previous one, which enabled Ethicon to rely on earlier regulatory approval or clearance. In her 2005 Clinical Expert Report on Prolift, Dr Charlotte Owens, then Medical Director for Gynecare, claimed that this approach arose from the “[r]ealization that pelvic floor disorders are physiologic and functional fascial hernias”.147
140 Between March 2003 and March 2013, Gynemesh PS was indicated for tissue reinforcement and long-lasting stabilisation of fascial structures of the pelvic floor in vaginal wall prolapse where surgical treatment was intended, either as mechanical support or bridging material for the fascial defect. On 16 March 2013, however, the indication for use of Gynemesh PS was narrowed to “a bridging material for apical vaginal and uterine prolapse where surgical treatment (laparotomy or laparoscopic approach) is warranted”.148 In other words, it was no longer indicated for transvaginal use but only for prolapse repair using an abdominal approach. The reasons for the change of indication are discussed later in these reasons.
141 On 17 August 2017, JJM notified the TGA that it would be discontinuing Gynemesh effective immediately and on 22 August 2017 the TGA cancelled its entry on the ARTG.149
142 Prolift was cleared for sale in Europe on 2 March 2005, around the same time in the United States, and in Australia on 30 March 2005. It was supplied in this country in and from June 2005 until 15 August 2012.150 Registration was cancelled on 21 April 2015.151
143 The Prolift pelvic floor repair system consisted of a pre-cut Gynemesh PS mesh implant (featuring a central mesh body and mesh arms) and included surgical tools used to facilitate insertion of the mesh fabric through the vaginal area (the Prolift Guide, the Prolift cannula and the Prolift retrieval device).
144 There were three different Prolift systems: anterior, posterior and apical, and total vaginal repairs (the “Anterior”, “Posterior” and “Total” pelvic floor repair systems respectively). Each system was sold separately and included the same guides, cannulas, and retrieval devices but the numbers of accessories varied. The mesh fabric was pre-cut into different shapes to accomplish the different types of repair, but a surgeon could choose to trim the mesh to suit the patient’s anatomy.
145 “Prolift Anterior” consisted of a piece of Prolene Soft with four arms secured through a transobturator approach.152 Self-evidently, it was designed for the repair of anterior vaginal defects. “Prolift Posterior” consisted of a piece of Prolene Soft with two arms secured using a transgluteal approach.153 It was designed to repair posterior and/or apical vaginal vault defects. “Prolift Total” was a combination of a Prolift Anterior and a Prolift Posterior.154 It was designed for a total vaginal repair. This is how the mesh implants were depicted in the IFU:
146 All three devices were developed for the surgical management of pelvic organ prolapse via a vaginal, as opposed to an abdominal, approach.
In the mid 2000, at a time Gynecare [a division of Ethicon Inc.] was enjoying a great monopolistic position on the market for slings, I thought that the next opportunity for the company could be the development of meshes for pelvic organ prolapse repair. Unfortunately, no one such as Ulmsten had come to me with a great solution and I had to be proactive.155
It was quite obvious that no advance could be made in this area as long as a standardized procedure would not be described.
Thus, I took the initiative of setting up a group of experts in order to work on the development of a standardized procedure which would make sense and would be reproducible by the average gynecologist.
148 Dr Arnaud said that he visited Professor Bernard Jacquetin, whom he described “an indisputable [key opinion leader] in the area of prolapse repair” and invited him to become “the ‘Ulmsten’ of POP repair”. Professor Jacquetin took up the invitation and the two of them chose a number of other French experts to work on the project. Their objectives, according to Dr Arnaud, were twofold: to work on developing a standardised technique for the surgical management of pelvic organ prolapse with mesh via a vaginal approach and to try to better understand the mechanism of vaginal erosion which was associated with the use of synthetic materials in order to reduce its occurrence. The group of experts, which became known as the “TVM Group”, met for the first time in June 2000. Gynecare France was placed in charge of its logistic and material coordination.156
149 Before going any further, I should explain what is meant by “erosion” in this context. Like a good deal of the nomenclature in this area, different terms are sometimes used to mean the same thing and there is a lack of precision in the terminology. Some of the professional associations have tried to do something about it, but the evidence in the present case suggests that their efforts have not had any significant effect.
150 The medical definition of an erosion is the “state of being worn away, as by friction or pressure”.157 In 2011 the Working Group on Complications Terminology, established by the International Urogynecological Association (IUGA) and the International Continence Society (ICS), considered that this definition was not necessarily apt to describe the clinical process and recommended that the term be abandoned and replaced with two new terms:
Exposure: A condition of displaying, revealing, exhibiting or making accessible … (e.g. vaginal mesh visualized through separated vaginal epithelium …)
Extrusion: Passage gradually out of a body structure or tissue … (e.g. a loop of tape protruding into the vaginal cavity …)158
151 Sometimes, the evidence indicates, the term “erosion” is used to refer to the migration of mesh into an organ, like the bladder or the urethra, which is an extension of the process of extrusion described in the IUGA/ICS lexicon and “exposure” to refer to anything short of that. Often, “erosion” is used in the literature, and also by some of the expert witnesses in this case, as a synonym for exposure and/or extrusion. Other commentators used the term “protrusion” in lieu of extrusion.159 Generally speaking, in this judgment I use the terms used by the witnesses when I refer to their evidence and the terms used in the scientific literature when I refer to that evidence. Otherwise, I may use the term “erosion” to capture any or all of these events.
152 I now return to the narrative.
153 Once he was convinced of the viability of the TVM project, Dr Arnaud set out to persuade others within Ethicon. In his memorandum on the development of Prolift to which I referred above, Dr Arnaud confessed to having had a hard time persuading some of “the marketing people” of the wisdom of the project. He said that they “wanted to kill [the] project as they found the procedure was too difficult and could only be a procedure for some happy few”. By early 2005, however, Dr Arnaud had prevailed160and, once Ethicon agreed to proceed, the evidence indicates that the marketing people enthusiastically embraced the project. It was only a matter of a few months before regulatory approval was obtained and the Prolift device was released to the market.
154 I hasten to add that the fact that this initiative was taken to exploit a marketing opportunity does not mean that advances in surgical treatment for pelvic organ prolapse were not desirable. There was a common perception that the rate of recurrent prolapse after native tissue repair was too high. The TVM Group put it at 20 to 30%. Surgeons had experimented with mesh as a means of reducing the rate, but the Group noted that “[a]lthough meshes … have been used to repair type 3 or 4 prolapse recurrence, authors still do not recommend their use for primary repair because of reported complications”.161 The desire to improve the status quo through innovation and to commercialise the resulting technology was not in itself problematic. It is, however, part of the context in which to consider the more critical question, which is whether the clinical data available to Ethicon were sufficient to support the conclusions expressed in its pre-market clinical evaluation report for Prolift and the decision to take the device to market, a question to which I will come in due course.
155 It will be recalled that one of Dr Arnaud’s two objectives in developing meshes for the repair of pelvic organ prolapse was to come up with a standardised procedure which the average gynaecologist would understand and could reproduce. Ethicon described the objective of the Prolift procedure as the achievement of “a complete anatomic repair of pelvic floor defects in a standardized way”.162
156 The technique for inserting the Prolift prostheses is described in a 30-page illustrated guide entitled “Gynecare Prolift – Total Pelvic Floor Repair, Anterior Pelvic Floor Repair, Posterior Pelvic Floor Repair – Surgical Technique” (the Guide).163 The metadata in the electronic court book indicates that this document was completed in early January 2005. There is no evidence as to its distribution but Dr Hinoul deposed that the Prolift IFU advised surgeons to review the Guide because it instructed them to refer to “the recommended surgical technique” for “further information” on the Prolift procedures.164 There was no relevant mention of a guide or guidebook in the Prolift IFU. It was mentioned in the Prolift+M IFU, which instructs surgeons to contact their “company sales representative” to obtain the Guide.165
157 A summary of the surgical technique, explained in greater length in the Guide, appears in the first publication on the technique. That was an article by Fatton et al (2007) entitled “Transvaginal repair of genital prolapse: preliminary results of a new tension-free vaginal mesh (Prolift™ technique) – a case series multicentric study”.166 The authors explained that the technique had been refined over a five-year period through more than 600 surgical interventions by nine French skilled vaginal surgeons. They described the technique in this way:
The synthetic material is a pre-cut non-absorbable monofilament soft prolene mesh. The mesh has three distinct parts … The anterior part is inserted between the bladder and the vagina and secured bilaterally by two arms through each obturator foramen. The posterior part is placed between the rectum and the vagina and is secured bilaterally by one arm passing through each ischiorectal fossa and sacrospinous ligament. The intermediate section corresponding to the vaginal apex separates the anterior and posterior parts and can be cut if needed. Instruments were designed to facilitate proper implant placement … Cannula-equipped guides were used to optimize the passage through the tissues, preventing muscle trauma and disruption of the arcus tendinous fascia pelvis (ATFP) and to allow the placement of the retrieval device. Retrieval devices were provided to easily catch and smoothly pass each prosthesis strap through the pelvis …
This procedure may be performed under spinal or general anesthesia. All patients were placed in the lithotomy position with thighs flexed at approximately 90°. After cleaning the entire surgical area with antiseptic, a urine culture is performed and an in-dwelling catheter is placed. All patients had an intravenous perioperative antibiotic prophylaxis.
158 More detail is provided in the article. It is unnecessary to repeat it here. It is sufficient to make the following observations.
159 First, is common ground that the technique is difficult and that it involved complex pelvic surgery.167
160 Associate Professor Rosamilia described Prolift surgery as “technically challenging”. She believed it required “considerable skill with sacrospinous colpopexy and deeper dissection”.168 She said that surgery with anchored mesh kits (like Prolift or Prolift+M) is “more complex” than native tissue repair.169
161 Second, although Fatton et al stated that “[p]rospective evaluation of the functional outcome [was] necessary to support the widespread use of this technique and to recommend it to young women with weakened tissues”,170 no such evaluation took place before Prolift was released to the market.
162 Prolift+M obtained regulatory clearance in Europe on 18 March 2008, in the United States on 15 May 2008, and was first supplied in Australia in December 2009.171
163 As I mentioned earlier, Prolift+M was made from Gynemesh M.
164 Like Prolift, there were three different Prolift+M systems: for anterior, posterior and apical, and total vaginal repairs (Prolift+M Anterior, Posterior and Total). Each system was sold separately and included the same guides, cannulas and retrieval devices supplied with Prolift. Like Prolift, the mesh was pre-cut into different shapes according to the different types of repair. The surgical technique was the same. A surgeon could also choose to trim the mesh to the patient’s anatomy.
165 This is what it looked like:
166 A is Total, B Anterior, and C Posterior.
167 Prolift+M was designed to address problems with Prolift. Yet, when Prolift+M was launched, Ethicon did not withdraw Prolift from sale and continued to promote it. In fact, all the devices remained in the Australian marketplace until all the POP devices except for Gynemesh PS were “decommercialised” on 15 August 2012, a little over three months after they were “decommercialised” in the United States. Registration was cancelled on 21 April 2015. 172
168 Prosima was first supplied in Australia in April 2010. It was not supplied after 15 August 2012 and, along with the other mesh kits, the TGA cancelled its registration on 21 April 2015.173
169 The Prosima system was the brainchild of an Australian surgeon, Marcus Carey, with whom Ethicon entered into a consultancy agreement in September 2004.174 The agreement offered Dr Carey incentives in the form of milestone payments which included: payment of USD 100,000 to be paid after publication of clinical study results in an internationally recognised, peer reviewed journal; and payment of USD 400,000 after the device became available for sale in the United States or the European Union in accordance with the terms of the agreement. Dr Carey was also entitled to royalties on the net sales of the product. This incentive arrangement meant that Dr Carey’s research, discussed below, was potentially tainted by bias or, at least, the perception of bias.
170 The development of Prosima was known internally at Ethicon as “Project Mint”. Like Prolift+M, Prosima was conceived and designed to improve perceived deficiencies in predicate devices.
171 “Mint” was regarded as an opportunity to expand the customer base. A PowerPoint presentation dated May 2006 entitled “Project Mint: Volume Estimate Summary”175 by Ethicon Women’s Health & Urology stated that “Mint [had] potential to be positioned: … [f]or patients whose prolapse is not severe enough to warrant Prolift” and was seen as an “[o]pportunity to expand [the] patient base” to include women with stage 2 and 3 prolapses.
172 Notwithstanding the safety concerns in the medical community related to the use of Prolift, the document identified a “[n]eed to find opportunities for Mint that do not cannibalize Prolift”. As always, Ethicon did not contemplate replacing a device because of safety issues, only introducing an additional one which might attract new customers or customers who were turning away from the old one.
173 Like the other mesh kits, there were three different Prosima systems: for anterior, posterior and apical, and combined vaginal repairs (Prosima Anterior, Posterior, and Total). Like Prolift, Prosima was made from Gynemesh PS. But the device was a different size and shape from Prolift and it could be placed either in the anterior or posterior position or, using the combined system, in both, as required. This is what the mesh looked like:
174 Two different inserter tools to facilitate placement anteriorly and posteriorly were included with the Prosima Anterior, Posterior, and Total systems as appropriate.
175 Unlike the Prolift and Prolift+M systems, the Prosima system provided maintenance of the vaginal canal during the period of healing following surgical repair of vaginal wall prolapse with a silicone vaginal support device (VSD) and a balloon. A surgeon could choose to trim the mesh or the VSD to suit the patient’s anatomy. Dr Hinoul explained that the VSD was designed to maintain the mesh in place as tissue ingrowth occurs.176
176 The VSD and balloon were inserted in the vaginal canal by the surgeon following implantation of the mesh fabric. The balloon was then inflated using a syringe which was supplied with the VSD. This replaced the need for gauze packing of the surgical area and provided support to the patient’s vaginal tissues post-operatively. The balloon was designed to be deflated and removed after 24 hours. The VSD could be left in the patient for up to four weeks post-operation in order to provide further post-operative support to the patient’s vaginal tissues before being removed by the patient’s surgeon.
177 In the case of Prosima, for implanting the anterior prosthesis, an incision was first made similar to that which was made to implant Prolift. Next, a full thickness dissection was undertaken with two channels created on either side of the bladder to the pelvic sidewall in the direction of the ischial spine.177 The arms were then introduced by means of an instrument called an introducer, so that the distal part of the arm was flush with the parietal fascia of the obturator internus muscle. The body of the device had a proximal tab which allowed for the mesh to be tacked to the apex of the vagina during placement of the arms.178
178 For the posterior prosthesis, a full thickness vaginal wall dissection was again carried out. The dissection was then extended on either side of the rectum to the levator ani muscle at the level of the ischial spine. After this extension, the distal parts of the straps or arms were placed up to and abutting the sacrospinous ligaments using an instrument called an inserter. Then the VSD was inserted to provide support for the vaginal tissues in the post-operative period.179
179 Professor Collinet maintained that the technique was standardised and reproducible.180 But the applicants argued that standardisation was never likely to succeed because of the known variations in surgeon technical skills as well as patient tissue and anatomy, the tendency for surgeons to modify their technique, the desire of surgeons to be able to individualise the technique to individual patients, and the unpredictable nature of the tissue response, including contraction.
180 The applicants’ argument was supported by Professor Korda, whose evidence I accept. Despite Professor Collinet’s position, it was not seriously challenged. What is more, it accords with common sense. Professor Korda stated:
Given the variability of tissue thickness, compliance and vaginal anatomy in individual patients, the difference in surgical training, skill and ability of individual surgeons and the level of precision required to place the mesh at a depth so it does not erode, it is almost impossible to get the mesh placement in a perfect position in the hands of ordinary skilled gynaecologists, urogynaecologists and urologists.
It is my view that given the inconsistency of the skill level of ordinary skilled surgeons, the variation of the vaginal anatomy in individual patients and the level of precision required for mesh placement, mesh implant surgery is not within the skill base of the average surgeon and as a result a standardised technique cannot realistically be achieved. I do not believe that it is ever possible to achieve a “reliably safe result” as asserted by Professor Collinet.181
181 Dr Agur expressed a similar opinion. He acknowledged that the TVM studies attempted to standardise the surgical techniques but said it was inevitable that variations would occur in the hands of ordinary surgeons. He referred in particular to variations in the nature of dissection, in the precise placement of the mesh, in the amount of tension applied, and in the use of transfixing sutures. He also said that expanding the use of mesh to general gynaecologists and urologists with variable technical and patient selection skills inevitably introduced significant variations. He said that the large surface area of the mesh resulted in the presence of mesh in in the lateral pelvic wall close to nerves and muscles, where it was not needed, () and that the use of trocars to guide placement was inevitably associated with the increased risk of organ damage. His evidence was that the so-called standardised technique was “not reliably safe” and its long-term efficacy was not yet established, not even in the hands of sub-specialist urogynaecologists and experienced pelvic floor surgeons.182 He noted that, when a group of experienced surgeons embarked on testing the reproducibility of the results of the TVM study in their own hands, within a research context and independent of the manufacturer, they stopped the trial prematurely due to the high incidence of adverse events.183 This was a reference to the Iglesia study published in 2010,184 discussed below.
182 Professor Korda’s view was that, before transvaginal mesh was released to the market, studies should have been published demonstrating the reproducibility of the technique with a cohort of competent ordinary skilled surgeons who would ultimately use it. As we shall see, no such studies were published.
183 In this Part of the judgment, I identify and describe the adverse events the applicants alleged could be caused by the Ethicon devices. In Part VI, I discuss the evidence about the prevalence of the complications and their clinical significance. Later, in Part XI, I look at the warnings and other information supplied by the respondents about these matters.
184 The applicants pleaded that the SUI devices could cause the following complications, referred to in the pleadings as “the Tape Complications”:
a chronic inflammatory reaction of the tissues surrounding or attached to the implants;
extrusion or erosion of the mesh into surrounding organs, including the vaginal wall, bladder or urethra;
dyspareunia and/or apareunia (avoidance of sexual intercourse);
offensive vaginal discharge;
de novo or recurrent urinary incontinence;
damage to surrounding organs, nerves, ligaments, tissue and/or blood vessels;
reoperation or revision surgery associated with complications; and
185 The applicants alleged that the POP devices could cause the same complications with the addition of difficulty defecating and recurrence of prolapse. The chronic inflammatory reaction was described slightly differently, as “a chronic inflammatory reaction, or an inflammatory reaction which is not minimal or transient in nature, of the tissues in which the mesh implants were implanted, attached and in addition, or alternatively, the surrounding tissues”.186 I do not take the applicants, however, to have intended to make any distinction. In other words, this was a distinction without a difference. It is apparent that the applicants’ intention was to refer to the foreign body response which inevitably occurs when a medical device is implanted and which is described in detail in Part V of these reasons.
186 In relation to the POP devices, they also listed as potential complications “protrusion” and “exposure”, but in the end no distinction was made between the SUI and POP devices in this regard. Likewise, the applicants pleaded that one of the complications attending the POP devices was the need for surgery to remove the mesh implant as a consequence of those particular complications, but this would have been required for the SUI devices too and, in any event, was covered by other alleged complications, such as “reoperation or revision surgery associated with complications”.
187 The applicants further alleged that both the SUI and the POP devices were difficult, if not impossible, to remove safely from patients suffering from one or more of the pleaded complications, that one or more surgical procedures might be required, and that removal carried the risk of new complications or of aggravating existing complications. In the pleadings, these matters were referred to as the “Tape Removal Complications” or the “Mesh Removal Complications”, as the case may be.
188 In their defence, the respondents asserted that “all surgical procedures present risks” but otherwise denied these allegations. The basis for the respondents’ denial was obscure. It did not emerge in evidence and early in the cross-examination of Dr Hinoul it quickly became clear that the denial was unsustainable.
189 In cross-examination, Dr Hinoul acknowledged that, from the time each of the devices was first supplied anywhere in the world, Ethicon knew of its potential to cause each of the pleaded complications. He conceded that, from the time of first supply, Ethicon was aware that a foreign body reaction to surrounding tissue would create a scar, that the mesh could be subjected to a contracting force applied by surrounding scar tissue, that the response of the host tissue was variable, and that any significant degree of contraction could lead to pain as could the scarring itself.187 Furthermore, he admitted that from that time Ethicon knew that there was a risk of mesh exposure and extrusion into the vaginal canal or another organ, that mesh exposure or extrusion could be difficult to treat, and that it could cause pain or discomfort.188 He also admitted that at the date of first supply Ethicon knew that both mesh erosion or extrusion and pain could occur many years after any of the devices had been implanted.189 Moreover, he admitted that Ethicon knew at that time that implantation carried a lifelong risk of erosion and pain,190 as well as risks of: dyspareunia and, as a consequence, apareunia; difficulty voiding; difficulty defecating; offensive discharge; leg weakness; and damage to surrounding organs, ligaments, tissues, and blood vessels.191
190 Dr Hinoul also conceded that Ethicon knew at that time that both acute and chronic pain could be caused by each of the devices, that chronic pain could be very damaging and debilitating, indeed “life altering”, and that multiple operations might be necessary to attempt to alleviate the pain. He agreed that the mesh could be difficult, if not impossible, to remove safely or without complications and that, in the case of Prolift, it could be disastrous. He said that Prolift could be removed safely but admitted that there was always a risk in so doing of causing damage to surrounding structures.192 He also admitted that, at the time each of the Ethicon devices was launched, Ethicon knew that, in the event of complications, the original condition (stress urinary incontinence or pelvic organ prolapse) could recur.193
191 Ultimately, then, there was no dispute that all of the complications could be caused by implantation of the Ethicon devices. What is more, lead counsel for the respondents, Mr Finch SC, told the Court in closing argument that the respondents accepted that each of the pleaded complications was clinically significant.194
192 These concessions were consistent with what appeared in Ethicon’s internal documents, which makes the respondents’ pleading all the more puzzling. In a Health Hazard Evaluation, for example, completed on 10 January 2014, Dr Hinoul acknowledged that “[n]o new adverse events [had] been identified since the initial launch of Ethicon Inc’s mesh based solutions…”.195
193 During cross-examination, the only qualifications Dr Hinoul made were that many of the pleaded complications could occur with any kind of surgery or pelvic floor surgery and/or that, at least some of them, were “small”, “very”, “highly” or “extremely” unlikely, “rare” or “extremely rare”.196 Dr Hinoul said that he was relying on reports in the literature and his personal experience.197 Both were unreliable foundations for an opinion as to the likelihood of complications and particularly at the time the devices were launched, having regard to the absence of data from long-term studies and properly conducted randomised controlled trials.
194 Be that as it may, in the Health Hazard Evaluation to which I referred above, Dr Hinoul said that:
Ethicon Inc. does acknowledge that in a small, yet significant subgroup of patients, these complications are to be considered serious.
195 Still, Dr Hinoul’s evidence exposed the real dispute. There was no issue that the devices could cause the pleaded complications. The dispute largely turned on the magnitude of the risks, the gravity of those risks, and the extent to which the respondents were obliged to disclose them.
196 After the trial concluded, the applicants applied for, and were granted, leave to amend their pleadings so as to add an allegation that the respondents failed to give any or any sufficient information or warning that the chronic inflammatory response to the implants could be affected by conditions which affect the immune response and healing, including autoimmune and connective tissue disorders. My reasons for doing so are set out in Gill v Ethicon SÀRL  FCA 470 at – (Gill (No 1)). For convenience, I will refer to these conditions as “autoimmune disorders”.
197 From now on I will refer to all the risks discussed above, including the heightened risk of a chronic inflammatory response for women with autoimmune disorders, as “the pleaded complications”.
198 Notwithstanding the concessions made by Dr Hinoul, it remains necessary to refer to other evidence about the risks. There is a large body of such evidence, a good deal of which was doubtless generated by the respondents’ defence.
199 The respondents emphasised that several of the complications could occur regardless of the use of mesh. But the weight of the evidence indicates that the use of mesh causes complications of a kind, degree or rate different from or greater than those associated with traditional methods of pelvic floor repair. These complications cannot wholly be explained by insufficient surgical training or experience. Other factors, including product design, mesh porosity, the quantity of mesh used, the route and methods of implantation and patient-specific factors cause or contribute to the development of adverse reactions following the implantation of synthetic mesh, including the various Ethicon devices.
200 In this section of the judgment, I discuss the nature of these complications.
201 The evidence on this question was given largely by the urogynaecologists, gynaecologists, and urologists.
202 It is convenient at this point to introduce them and to provide a snapshot of their qualifications and experience.
203 Professor Korda is an Australian clinician and academic. The bulk of his career was spent at the King George V and Royal Prince Alfred Hospitals in Sydney either on staff or as a Visiting Medical Officer. From 1991 to 1997 he was the Head of the Department of General Gynaecology and Urogynaecology at those hospitals and from 1996 to 2013 he was the Chairman of the Pelvic Floor Unit at Royal Prince Alfred and Concord Hospitals. He was also a Visiting Gynaecologist at the Mater Misericordiae Hospital in Sydney from 1986 to 2001. He was a clinical lecturer in obstetrics and gynaecology at the University of Sydney from 1975 to 2013. In 2008 he was appointed Foundation Professor of Obstetrics and Gynaecology at Western Sydney University and in 2015 Conjoint Professor.
204 Dr Agur is a consultant gynaecologist and obstetrician and lead urogynaecologist at the University Hospital Crosshouse in Kilmarnock, an Honorary Clinical Senior Lecturer at the University of Glasgow and a Career Fellow with NHS Research in Scotland. He is considerably younger than Professor Korda. He did not become a member of the Royal College of Obstetricians and Gynaecologists until May 2003 and did not attain subspecialty accreditation in urogynaecology until 2009.
205 Professor Blaivas is Adjunct Professor of Urology at SUNY Downstate Medical School in Brooklyn New York, a position he has held since 2008, and has practised as a urologist since the late 1970s. He has also published widely in his areas of expertise.
206 Dr Margolis is Assistant Clinical Professor in the Department of Obstetrics & Gynaecology at the David Geffen School of Medicine at the University of California, Los Angeles and also works in private practice. He has held a variety of academic positions over the years at a number of prestigious American universities in the fields of obstetrics and gynaecology, including urogynaecology and pelvic surgery.
207 Professor Collinet is a professor of medicine, consultant gynaecologist, and head of the gynaecologic surgical unit of the Jeanne de Flandre Hospital in Lille, France. He obtained his qualifications in gynaecology and obstetrics in 2003. After the launch of Prolift, he was engaged on several occasions by Ethicon to take part in and lead training sessions around the world in surgical techniques and, in particular, the implantation of Prolift. His work in urogynaecology, both clinical and research, was mostly confined to a period of seven years, ceasing in 2012. During that period he undertook hundreds of prolapse operations using vaginal meshes and treated what he referred to in his report as “specific complications” identified as prosthetic erosions and retraction.198 Since then he has conducted about 20 surgical prolapse procedures a year.199
208 Professor Deprest is a consultant urogynaecologist, a professor of obstetrics and gynaecology at the Catholic University in Leuven, Belgium and an honorary professor at University College, London. He is also the co-director of the multidisciplinary pelvic floor unit at Universitair Ziekenhuis (UZ) Leuven, the largest university hospital in Europe.200 He has a research interest in pelvic floor dysfunction, including aetiology, secondary prevention and treatment. Since 2002 he has been involved in translational research in urogynaecologic surgery and has studied the host response to biological graft materials and synthetic mesh implants. His clinical research has mainly focussed on the outcomes of specific operations including laparoscopic sacrocolpopexy, native tissue repair of anterior vaginal wall prolapse, vaginal mesh surgery and tape insertion. He is the author of over 470 peer-reviewed articles published in the international literature and over 60 book chapters. At the time he prepared his report, he was a member of the board of the Group of Biomedical Sciences and he has held various positions with a number of professional organisations including the IUGA, the European Uroyanecological Association (EUGA), the European Society of Gynaecological Endoscopy, and the Flemish Society of Obstetrics and Gynaecology. He was also involved in the IUGA working party on meshes in prolapse and incontinence.201
209 Professor Roovers is a Dutch urogynaecologist. He completed his training as a gynaecologist in 2006 and the same year undertook an eight-month fellowship in urogynaecology at the UZ in Leuven. Since he has been head of the gynaecology department at the Academical Medical Centre in Amsterdam and was appointed the chair of gynaecology at the University of Amsterdam in 2014. Since 2016 he has also been an honorary professor at the University of Cape Town in South Africa. He was a supervisor of Dr Hinoul’s PhD thesis.202 He is a former chair of the education committee of the IUGA and is currently the chair of the scientific committee of the EUGA.203
210 Associate Professor Lam and Associate Professor Rosamilia are both Australian practitioners.
211 Associate Professor Lam is the director of the Centre for Advanced Reproductive Endosurgery in Sydney. He currently holds appointments as clinical associate professor at Sydney Medical School, gynaecological endoscopic surgeon at Royal North Shore Hospital, The Mater Hospital, and St George Private Hospital. He is a board member of the World Endometriosis Society and the International Society for Gynecologic Endoscopy and a former president of the Australasian Gynaecological Endoscopy and Surgical Society.204
212 Adjunct Associate Professor Rosamilia is a urogynaecologist. Since 2008 she has been the head of the pelvic floor clinic at Monash Medical Centre, one of three tertiary obstetrics and gynaecology teaching hospitals in Melbourne.205 She has been heavily involved in urogynaecology education and has held a number of leadership roles in various professional societies, including the IUGA and the Australasian Gynaecological Endoscopy and Surgical Society. Since 2010 she served as a urogynaecology member of the Urogenital Prostheses Clinical Advisory Group to the Department of Health.
213 Professor Deprest described a number of complications directly associated with the use of mesh (including the slings made from the mesh), sometimes called “graft associated complications” or “graft related complications”.206 They included erosion; contraction or shrinkage; infection; pain; and reintervention.
214 Mesh can extrude, or “erode”, into the bladder and/or the urethra. Professor Chughtai described mesh erosion as “a debilitating complication of transvaginal mesh placement, which can require a myriad of surgeries to repair” but which may have no remedial effect.207 Professor Deprest said that long-term extrusions into the bowel are rare with stress urinary incontinence devices but more common in mesh repair for prolapse, particularly if the device is placed between the rectum and vagina.208
215 Erosion (commonly referred to as exposure) refers to the exposure of the foreign or graft material to the vaginal lumen (canal). Frequent symptoms of exposure are vaginal bleeding, pain, and dyspareunia.209 Recurrent urinary tract infections, haematuria (blood in the urine), and irritative bladder symptoms can also be symptoms of erosion. Exposure of the mesh graft through the vaginal epithelium can also be associated with vaginal discharge, pain and recurrent infections.210
216 Mesh exposure can occur immediately or soon after surgery. But it can also occur much later, including years after surgery, and away from the incision line.
217 It appears that the mechanism by which mesh erodes (extrudes or becomes exposed) is not known.211 Professor Deprest said that several mechanisms have been suggested including thinning of the epithelium on account of aging or menopause, for example, and an active process occurring underneath the epithelium, such as folding of the mesh or blood collection.212 He said that erosion which occurs sometime after surgery is likely to be due to either migration of the mesh or secondary loss of tissue integrity over the area of the implant. Consequently, he attributed late onset exposure to erosion of the covering epithelium (the membranous lining).213 Although Dr Hinoul was quite insistent that the mesh does not migrate or move,214 the weight of the evidence is against him.
218 Professor Korda considered that mesh erosion occurs because the mesh is a rigid substance and during movement, such as sitting or standing or sexual activity, the vagina stretches, but the mesh cannot stretch at the same rate or time. The result, he said, is that the vaginal tissue abrades against the mesh, which can cause the mesh to erode.215 While he deferred to the biomedical scientists and pathologists in this respect, his evidence was supported by Professor Klosterhalfen. Of particular relevance to this matter is the evidence of Professor Klosterhalfen about the “mechanical mismatch” between polypropylene mesh and vaginal tissue, discussed in Part V below.
219 Professor Klinge said that erosion can be caused by contraction of the mesh (strictly contraction of the mesh/tissue scar complex).216 Professor Klosterhalfen said that it could be caused by folds and curls in the mesh due to bridging of the scar tissue (bridging fibrosis).217 Both these matters are discussed at length in Part V.
220 Professor Blaivas said that erosion can be caused by surgical error where the mesh is inadvertently passed through the lower urinary tract or vagina at the time of surgery. But he discounted this as a factor in “the modern era” since almost all patients undergo cystoscopy and vaginal inspection at the time of implantation. He also said that erosion can be caused during surgery if the mesh is placed too close to the vaginal epithelium or too close to the lower urinary tract or in its wall. Finally, he said that erosion may occur because of the acute and chronic inflammatory process, sometimes associated with bacterial infection, much in the same way that a seroma (a collection of fluid under the surface of the skin) or subcutaneous abscess (a collection of pus under the skin) can burst through the skin.218
221 Notes from a “mesh exposure meeting” produced by Ethicon, identified a number of potential causes of mesh exposure, including mesh folding, mechanical irritation, infection, surface biocompatibility, the use of the transvaginal route, as well as patient factors such as a thin vaginal wall and poor collagen in scar tissue.219
222 It was common ground that patient factors, such as obesity, poorly-controlled diabetes (associated with more infections), vaginal atrophy and smoking, may all impede wound healing220 and that delayed or inadequate healing has been identified as a cause of mesh exposure.221
223 Notwithstanding the attempt by Dr Hinoul to minimise its significance, exposure may have serious consequences. Professor Korda explained that, since the vagina is never free of organisms (bacteria), once the mesh is exposed or erodes, the vagina “inevitably” becomes infected and the nidus (focus) of the infection is virtually impossible to treat because antibiotics have great difficulty penetrating through the mesh.222 He explained that antibiotics could not get to the bacteria sitting on the implant as there are no blood vessels in the mesh fibres. Professor Korda also said that chronic inflammation is the result of exposed mesh and exposure may also aggravate chronic inflammation, particularly when bacteria colonise the exposed mesh.223
224 Professor Deprest made the same point. He reported that when the mesh is exposed to the vaginal lumen, “it will almost certainly become infected” as resident bacteria will attach to the implant. Infection can also occur, he added, as “a primary process” and when it does it can cause mesh exposure.224
225 Professor Korda pointed out that the relationship between inflammation and pain has been known since the first century AD. Along with redness, heat and swelling, he said that pain is one of the classic signs of inflammation.225
226 None of this evidence was apparently controvercial. Professor Korda was not cross-examined on these points. It was logical and persuasive. There is no good reason why I should not accept it and I do. I will return to the question of infection a little later.
227 The next graft-related complication to which Professor Deprest referred was pain, which can occur with or without mesh exposure. Professor Deprest said that pain can be felt along the entire trajectory of the implant, which in the case of retropubic tapes can be around the urethra and in the case of transobturator tapes through the muscles. Sometimes it is difficult for patients to localise their pain. Professor Deprest explained:
It is believed that this process starts locally, yet when the problem of pain is long standing (more than 3 to 6 months), the pain may become “centralized”. This means the process becomes less dependent on what happens locally, yet with the unfortunate side effect that even local removal of the causative factor may not solve the problem. However, this is phenomenon is seen in the absence of implants, and is sometimes referred to, without any criticism of the patient, as ghost [phantom] pain.226
228 The evidence indicates that pain can be both more severe and more enduring after repair with procedures involving the use of mesh than after procedures which do not. Professor Korda said that dyspareunia is more severe after mesh repair than after traditional prolapse surgery.
229 Professor Korda said that, while prolapse surgery with or without mesh can cause damage to the urethra, the bladder and the rectum, there is a greater incidence of damage to pelvic nerves and major vessels with mesh surgery.227 He also said that, for anterior compartment repair, the use of mesh is associated with longer operating time, greater blood loss; increased incidence of bladder damage; more severe dyspareunia;228 de novo urinary incontinence;229 chronic and severe pain; and repeat surgery as a result of mesh erosion or contraction.230 He added that mesh surgery in the posterior compartment is associated with a higher incidence of rectal damage and a significant mesh erosion rate (up to 10%).231
230 Professor Chughtai, whose evidence was confined to the POP devices, said that the most commonly reported complications amongst women with transvaginal mesh used to treat pelvic organ prolapse are erosion into the vagina and mesh contraction or shrinkage, both of which can cause severe side effects including pain in the pelvis and dyspareunia.232 He said that even if the mesh is removed, patients may likely have residual pelvic pain and dyspareunia.233
231 Professor Deprest also gave evidence that there is a greater problem with pain in patients who have been implanted with mesh than patients who have native tissue repairs.234
232 Moreover, in contrast with native tissue repair, pain after mesh repair can arise well after surgery, sometimes years later.
233 Professor Korda said that in his long period of clinical practice in urogynaecology (around 30 years) during which he had performed thousands of native tissue repairs, the severe, chronic and intractable pain that can occur after mesh surgery was “virtually non-existent” after native tissue repair.235 He was adamant that, while pain and dyspareunia can occur after native tissue repair, they are never of the same magnitude, duration or gravity as the pain and dyspareunia that occur after mesh repair.236 He said that the incidence of pain after native tissue repair is “extremely low” and is usually related to vaginal incisional pain which resolves within four to six weeks of the operation.237 He elaborated on this evidence in his oral testimony:
If you do a simple vaginal operation, and you don’t – with native tissue repair and you don’t actually incorporate nerves inadvertently into the repair, then usually there is pain immediately after the operation and perhaps for two to three weeks or maybe four weeks until the sutures dissolve. And then the pain disappears, and, you know, I’ve been operating on people for a long time, and it’s virtually never seen as an ongoing, prolonged chronic problem … [W]hat we are seeing in these cases in this case, in this matter, is that some of these ladies have suffered chronic intractable pain.238
234 The probable explanation for this chronic intractable pain, he added, is that it is “due to shrinkage of vaginal tissue around the tape or around the mesh which sets up an inflammatory reaction …”.
235 In an article published in 2010, Dr Donald Ostergard attributed the enduring nature of the pain to both mesh shrinkage or contraction and tension on the arms of the mesh with neuroma formation. As he explained it, since mesh is anchored in tissue, its shrinkage will increase the tension on the anchoring tissue with resulting pain. He said that “[n]o mesh seems to be immune from this process”.239
236 The main difference between mesh repair and native tissue repair, according to Professor Korda, was that mesh surgery carried “new and unique” risks likely to be permanently incapacitating, such as chronic and severe pain, mesh contracture, and mesh erosion.240 He said that complications such as mesh erosion, mesh contraction and bunching, severe chronic pelvic pain, severe dyspareunia and pain on movement, sitting, and standing were not encountered in pelvic floor surgery before the introduction of synthetic mesh.241 He also said that, with the exception of sacrospinous ligament suspension, nerve injuries are virtually unknown after native tissue repairs but are “reasonably frequent” after repair of prolapse using mesh kits. The reason for this, he explained, was that the fixation arms of the kits are often hooked to the coccygeus sacrospinous ligament complex which is very close to the sciatic nerve and the pudendal nerve and its branches, and are inserted through the obturator foramen and can damage the obturator nerves.242
237 Further, mesh implant surgery can cause hispareunia (severe pain in the male partner during sexual intercourse) usually due to contact with the sharp fibres of eroded mesh poking through the vaginal mucosa.243
238 Dr Blaivas, whose evidence was largely confined to the SUI devices, said that there were two complications that were virtually unique to mesh implants: erosion into the lower urinary tract, vagina, or bowel; and refractory pain (pain that does not yield to treatment). He said that before the introduction of mesh surgeries, they were “almost never” mentioned in the peer-reviewed literature.244 There is evidence, however, that intractable suprapubic pain has been described after colposuspension.245
239 In a position statement on midurethral slings issued in May 2017, the Royal Australian and New Zealand College of Obstetricians and Gynaecologists (RANZCOG) observed that “sling insertion can cause pain and dyspareunia”, that transobturator slings can cause groin as well as pelvic pain, that the pain can persist and in some cases become intractable, and that removal surgery may be required, but that complete removal can be difficult, can cause complications of its own, and, in any case, may not completely resolve chronic pain or other adverse symptoms.246
240 Professor Korda expressed the view that the complications of mesh surgery for prolapse are generally long-term and sometimes incurable. He listed these as:
reoperation after vaginal mesh surgery at statistically significant higher rates of 8.5% to 22% compared to native tissue repair (3.2% to 9.7%);
reoperation after laparoscopic sacrocolpopexy (5.9%, 2.9% being for specific complications and 1.7% for prolapse recurrence);
vaginal mesh exposure (mean rate of exposure of 13.1% based on the literature), much lower at 3% and 2.5% after open and laparoscopic sacrocolpopexy respectively;
late visceral (bladder and rectum) mesh exposure after laparoscopic sacralcolpopexy (0.5% to 1%) and also vaginal surgery (rate not given);
infection (the rate of which is unknown, with some series reporting 80% bacterial mesh colonisation), abscess, cellulitis, spondylodiscitis;
painful mesh contraction (involving vaginal pain on movement, dyspareunia and localised tenderness over prominent and tense portions of the mesh), with an incidence rate of between 4 and 11%; and
chronic pelvic pain, which has been poorly studied, but occurs in an estimated 0% to 30% of patients following transvaginal mesh placement.247
241 Professor Korda went on to say that there were few procedures available that could permanently reverse or treat the complications of mesh implant surgery.248
242 Evidence to similar effect was given by Professor Chughtai.
243 Professor Chughtai noted that women who underwent mesh repair for pelvic organ prolapse tended to have a higher retreatment rate within two years of surgery, and a higher reoperation rate, in comparison to those undergoing native tissue repairs. He attributed the difference to “major complications” of mesh repair surgery, such as dyspareunia and chronic pelvic pain caused by mesh shrinkage or extrusion into the vagina.249 Professor Collinet described Professor Chughtai as an opinion leader because of his participation in many clinical trials and his scientific output (in excess of 100 scientific papers on the care of vaginal prolapse and stress urinary incontinence).250
244 Removal of the mesh (which may be required in cases of mesh exposure, extrusion, erosion, chronic pain, or recurrent prolapse) presents additional problems. Mesh exposure through the vaginal wall can often be treated with topical oestrogen or local antibiotics, especially if the exposure is small. But if conservative treatment is unsuccessful, the exposure is not small, or there is extensive pain in the operation area, then surgical excision will be required and it may be necessary to remove the mesh.251
245 In an article published in the International Urogynecology Journal in 2006, members of the TVM Group noted that when the mesh becomes exposed, “the complete and rapid removal of the material” is warranted “in order to prevent serious surgical sequelae”.252
246 Yet the evidence on both sides was that the mesh is difficult, if not impossible, to remove entirely as it is or becomes integrated in the connective tissue, and that removal surgery may not relieve pain. 253
247 In her evidence in chief, Associate Professor Rosamilia’s said that most patients with mesh exposure require one operation only to manage the problem.254 She acknowledged, however, that more than one procedure can often be required to manage chronic pain.255
248 Professor Korda accepted that small areas of the mesh can be excised in many patients and the vagina “appropriately” closed afterwards. But he said that “in a large percentage of patients mesh exposure and erosion creates intractable pain and offensive continuous blood stained vaginal discharge, dyspareunia, difficulty sitting or moving and in such patients removing the exposed part of the mesh and closing the epithelium does not resolve their problem”.256
249 The process of excising mesh is, as Professor Korda put it, “technically difficult and fraught with danger”257. Professor Korda explained that the mesh becomes intimately incorporated into the vaginal epithelium, the underlying fascia, and the adjacent and contiguous organs (such as the urethra, bladder, and small and large bowel) and dissection can cause inadvertent injury to those structures. It can also cause the formation of fistulas (abnormal communications between two hollow organs or between a hollow organ and the exterior).258
250 At a meeting with, amongst others, Laura Angelini and Brigitte Hellhammer from Ethicon, held on 20 December 2000, Dr (later Professor) Klinge, who was experienced in explanting meshes after abdominal wall and inguinal hernia repair, described it as “a demanding task only to be performed by an experienced surgeon”.259 Associate Professor Lam testified that this reflected his own view in relation to explanting Prolift or Gynemesh.260
251 Both Associate Professor Lam and Associate Professor Rosamilia insisted that it is not usually necessary to remove all the mesh.261 Professor Korda accepted that in many patients in whom mesh exposure occurs small areas of the mesh can be excised and the vagina can be closed appropriately afterwards. He observed, however, that inability to remove the mesh entirely often leaves patients with unresolved and continuing symptoms.262
252 Professor Chughtai,263 Dr Agur,264 and Professor Deprest gave evidence to the same effect.265 Associate Professor Lam said that removal of mesh via the transvaginal route may be impossible if surgical access is poor due to extensive vaginal scarring either from the mesh itself or from previous pelvic floor operations.266
253 Professor Korda said that, in a large percentage of patients, mesh exposure and erosion create intractable pain and offensive continuous blood-stained vaginal discharge, dyspareunia, and difficulty sitting or moving. In those patients, these problems are not solved by removing the excised portion of the mesh and closing the epithelium because “the host response to the mesh is one of an inflammatory nature and smoulders on”. Further, if the mesh becomes infected while exposed, the inflammatory process and infection spread along the fibres of the mesh underneath the vaginal skin, causing inflammation and a severe host response leading to fibrosis.267
254 Dr Agur said that the retropubic devices were generally easier to remove in their entirety, regardless of the time of implantation. This was because the anatomy of the route they traverse is well understood by pelvic surgeons and they pass through fewer muscles. On the other hand, he said that the transobturator tape traverses more muscles (in the groin and thigh) to which it becomes densely adherent and the route is unfamiliar to pelvic surgeons. In general, surgery aiming at complete mesh removal is required for patients suffering from implant-related chronic pain and surgery of this kind is complex and highly specialised. He said it poses a unique technical challenge and is associated with a relatively high risk of injury to nearby pelvic organs, surrounding nerves and blood vessels. He said that complete surgical removal is often impossible. It requires extensive surgery with local damage to tissues and pelvic organs causing scar tissue which is usually permanent and commonly chronic pain persists even if all the implanted material is removed.268
255 Complete removal is technically very difficult or impossible for most of the POP devices due to the large surface area, dense adherence to native tissues and proximity to pelvic organs. Dr Agur said that safe removal is impossible in the majority of cases unless the device was implanted within a period of two to four weeks before removal is attempted.269 The same is true of the transobturator SUI devices.270 Speaking of the POP implants, Professor Korda stated:
It is my view that the mesh implant usually becomes attached and densely adherent to underlying structures.
In the case of the anterior compartment the mesh is attached to the urethra and the bladder, in case of the apex of the vagina the mesh is adherent to the vagina, sometimes the small bowel and the ligamentous structures such as the sacrospinous ligament and in the case of the posterior compartment the mesh is densely adherent not only to the vagina but to the rectum.
Removal of the mesh requires the dissection of the mesh from both the overlying vaginal epithelium, the fascia below the epithelium and the underlying structures mentioned above. The adherence of the mesh to these structures is such that during even the most careful dissection it is possible to damage the urethra, the bladder, the ureters, the pudendal nerve, the pudendal artery and the rectum.
It is my view that it is often impossible to remove the mesh in its entirety without causing life threatening damage.271
256 Professor Korda added that it was particularly difficult to excise and remove mesh in its entirety when the fixation arms went through the obturator canal because of the prospect of bleeding and damage to the obturator nerve.272
257 Associate Professor Rosamilia agreed that the mesh is difficult to remove completely as it becomes integrated in the connective tissue.273
258 Although limited vaginal dissection is required for insertion of the SUI devices compared with other surgical procedures for stress urinary incontinence, the venous plexus (a complex network of veins) in the retropubic space is vulnerable to damage during the blind passage of the introducer involved in the implantation of the retropubic devices. Striking pelvic vessels with a needle will cause bleeding from the retropubic space, which may present as increased intra-operative blood loss or later as a retropubic haematoma. Major vascular injury can occur and is potentially fatal. 274 Haemorrhage can be “dramatic” and is difficult to manage.275 The symptoms of retropubic haematoma include pain, voiding difficulty, pelvic mass and a drop in haemoglobin.276
259 Bladder perforation is a well-recognised risk of the retropubic procedures277 because of the blind passage of the needle in the retropubic space. Professor Deprest said that this occurs in up to 5% of cases, more frequently during “the learning curve” and where there has been previous retropubic surgery.278
260 Obstructive voiding symptoms (including, rarely, retention) can also occur, possibly in association with a “higher than desirable tension”.279
261 The introduction of the transobturator tape (including TVT-O) dramatically lowered the rate of bladder perforations but gave rise to different risks. Professor Deprest explained the position well. He wrote:
TOT essentially also provides mid-urethral support, yet the tape is not directed upwards between the bladder and the pubis, however lateral to the obturator membrane. The tape trajectory is continued through a number of muscles going to the medial side of the thigh.
On that trajectory one can be close to or on small nerves which run in unpredictable locations. The piercing or vicinity of permanent tape material may cause specific pain at these locations, unique to this approach.280
262 Injuries to the bladder, urethra, rectum and pelvic nerves may arise from the technique used to implant the mesh kits (Prolift, Prolift+M, and Prosima).281 Professor Korda explained that:
Injury to the bladder, urethra and rectum can occur during dissection and are usually inadvertent.
With the exception of sacrospinous ligament suspension, nerve injuries are virtually unknown after native tissue repairs but are reasonably frequent after mesh kits as the fixation arms of the mesh kits are often hooked to the coccygeus sacrospinous ligament complex which is in very close anatomical proximity to the sciatic nerve and pudendal nerve and its branches[.]
263 It is common ground that all meshes used in the Ethicon devices must be inserted without tension. Professor Korda said that if the mesh tension is too tight “all complications are more likely to occur”.282 I did not understand this evidence to be controversial. He was certainly not cross-examined on it.
264 Yet achieving the requisite amount of tension is not easy. Experience helps, but even the most experienced surgeons can occasionally fail. Professor Korda pointed out, for example, that there is a significantly greater incidence of voiding problems after retropubic TVT surgery because of difficulty in getting the right amount of tension.283 He also pointed out that the introduction of trocars into the various anatomical spaces to retrieve the arms of the mesh kits (Prolift, Prolift+M and Prosima) is “essentially a blind procedure”, conducted by touch, and requires a thorough knowledge of the bony structures of the pelvis. The tension is achieved by pulling on the arms of the devices. Since some of the anatomical spaces through which the arms are placed are very narrow, Professor Korda said that it was often difficult to be accurate or consistent in achieving the right amount of tension.284 He said that the Prolift procedure (also used to implant Prolift+M) was not reproducible by the average, reasonably competent gynaecologist in order to achieve reliably safe results.285
265 It was common ground that the use of both the TVM technique and the use of Prolift could cause hardening of the vagina and vaginal, perineal and anal pain. It was also common ground that insertion of a vaginal prosthesis can have a deleterious effect on sexuality and sexual function and cause de novo dyspareunia. Professor Collinet conceded as much.286
266 Professor Collinet noted that, irrespective of the device used, procedures involving the use of vaginal mesh required “special surgical training, and therefore new surgical expertise” due to the need for particular dissections not previously called for in pelvic floor surgery. He observed and said that all surgical innovations require special training and involve a learning curve and he attributed the “sizable numbers” of mesh complications, to the widespread adoption of the mesh kits and the lack of adequate initial training.287 Evidence to like effect was given by Associate Professor Rosamilia and Professor Roovers.288
267 Associate Professor Rosamilia relied on two articles in support of the notion that surgeon experience is a significant factor in mesh exposure rates.289 One, by Dwyer and O’Reilly, published in August 2004, a retrospective study of 97 women, related to the use of Atrium polypropylene mesh in the repair of large or recurrent anterior and posterior prolapse.290 The study began in February 1999 and finished in May 2002. The women were followed up at six weeks, six months, 12 months, and two years. Erosion was the most common complication. Over the study period the rate of mesh erosion decreased from 19% in 1999 to 13% in 2000 and 4% in 2001/2002. The authors concluded that the incidence of this complication was related to surgical experience. Associate Professor Rosamilia emphasised this point in her report. She omitted to point out, however, that the differences were not statistically significant.
268 The second article upon which Associate Professor Rosamilia relied was by Achtari et al, published online in January 2005. This was a retrospective review of the records of 198 women who had undergone transvaginal mesh repair surgery for pelvic organ prolapse using either Atrium or Vypro II. Vypro II was a composite mesh made from polypropylene and polyglactin 910 mesh. The authors detected a correlation in the differences in erosion rates according to experience but in cross-examination Associate Professor Rosamilia conceded that they were not statistically significant. 291
269 In both cases, Associate Professor Rosamilia described the omission to refer to these qualifications as an oversight. 292
270 Notwithstanding the successful assault on Associate Professor Rosamilia’s evidence, I accept that inexperience and lack of adequate training in the surgical techniques for implanting the various devices contributed to the incidence of complications. Professor Korda’s evidence on this question was not materially different.293 Moreover, it stands to reason that complications would be higher in less experienced hands. As Dr Agur observed, however, there is no good evidence that the learning curve accounts to any significant extent for “the inherent mesh complications”.294 Even in the best of hands, using “great surgical skill”, mesh erosion can occur and has been reported.295
271 A report on the outcomes of 917 consecutive transvaginal mesh procedures conducted at the Centre for Advanced Reproductive Endosurgery in Sydney where Associate Professor Lam works, and of which he was a co-author, showed an erosion rate for Prolift of 46 out of 232 cases (20.4%).296 When his attention was drawn to these figures during cross-examination,297 he tried to distance himself from the evidence, claimed it was an error, and that the real figure was “around 10 percent”.298 Although he had referenced the report in his written evidence, he later insisted it was a draft but was unable to produce the final version.299 Whether or not the document was a draft, his attempts to resile from the higher erosion rate were quite unconvincing. What is more, regardless of the true figure, his evidence supports the applicants’ case that mesh erosion can occur even in the best of hands.
272 Dr Leake, a gynaecologist who operated on Mrs Gill for one of her mesh erosions, deposed that transvaginal mesh kits like Prolift were “notoriously difficult for a surgeon to get right every time” even for a surgeon skilled in prolapse surgery.300 According to Dr Leake, tension was the critical factor in the success of Prolift and if implants were put in too tightly patients often experienced pain and dyspareunia.
273 Dr Leake said that mesh may also become too tight because of variations in the patient’s response to it. She explained that “[s]ome patients scar a lot more vigorously than others” and she had found that some ligaments could not withstand the trauma associated with insertion or fixation of mesh to them, which also affected the outcome from mesh surgery. For some patients, the deepest anchor point for the anterior part of the device (the anchor point closest to the patient’s head) was not adequate to hold the prolapse up and the mesh sometimes rolled forwards, predisposing the patient to erosion and the feeling of recurrent prolapse. Further, if the anchor points sagged, this led to inversion of the top of the vagina, which could lead to excessive tension elsewhere, particularly where the mesh rolled over on itself, and cause lower back pain, chronic pelvic pain, and dyspareunia.301
274 If a patient developed a haematoma from the surgery, a not uncommon occurrence in vaginal surgery with or without mesh, Dr Leake found, through experience, that healing was often impaired and erosions could occur. She said that the development of a haematoma after mesh implantation can become a far greater problem than a haematoma after native tissue repair because of healing difficulties in the presence of mesh.
275 Dr Leake was of the opinion that the “pulley system” used for fixing the upper ends or arms of the Prolift contributed to the problem of excessive tensioning. The Prolift technique involved “a blind approach”, which made it difficult to determine whether or not there would be sufficient or excessive tension after the operation was complete.302
276 Like the other witnesses, Dr Leake also testified that excising mesh from the body is very difficult. She said that, although it is possible to excise those parts of the mesh which have eroded through the vagina, if there is tension in the mesh arms the vaginal skin can get scarred and, once the erosion has been repaired, it can be hard to stretch again.303
277 Dr Leake was not required for cross-examination and I accept her evidence.
278 In Gill (No 1), I gave the respondents leave to make further submissions on the matters raised by the applicants’ allegation that the chronic inflammatory response to the implants could be affected by conditions which affect the immune response and healing, including autoimmune disorders. In response to this invitation, the respondents filed supplementary submissions on 14 June 2018. They argued that there were three evidentiary gaps in the applicants’ case. The first was that the allegation was not put to Dr Hinoul so that, to the extent that the applicants claim that the proposition was within Ethicon’s knowledge, they needed to prove that independently. The second was that the applicants had not adduced evidence from any of the surgeons with whom the discussions about implant surgery had taken place. The third was that the applicants had not demonstrated the clinical significance of autoimmune disorders and the chronic inflammatory response to the Ethicon devices.
279 At this point, I will deal only with the first and third arguments. I will deal with the second argument in Parts XIII and XIV when I discuss the applicants’ case concerning the failure of the respondents to provide adequate warnings and other information about the risks posed by the Ethicon devices.
280 The applicants relied on the concessions from Dr Hinoul that when the devices were supplied to the market, the respondents knew that the foreign body response to the mesh was chronic, that mesh exposure was a risk, that there was a lifelong risk of erosion, that mesh exposure or extrusion could cause pain, and that there was a lifetime risk of pain, including life-altering chronic pain. The applicants also submitted that the respondents had known for “many years” that Prolene produced a chronic inflammatory response. And the applicants referred to the evidence that inflammation can be associated with pain.
281 Dr Hinoul’s concessions and this evidence are beside the point. The point the respondents were making was that the substance of the allegation about the potential effect of the chronic inflammatory response on women with these conditions was not put to Dr Hinoul. Of that there can be no doubt. Nevertheless, there was a wealth of other evidence, discussed below, to prove that immunocompromised patients were at particular risk of harm from the chronic inflammatory response and that the respondents either knew or ought to have known that was so at least from the time the very first Ethicon device was launched. The proposition that this was not a matter of clinical significance must be rejected.
282 The immune response to any foreign material has been described as “complex, dynamic, and patient-specific”.304 It was uncontroversial that the individual host response is unpredictable.305 The evidence established that a heightened chronic inflammatory response to implantation with non-absorbable polypropylene may occur in patients who have autoimmune disorders or have been using immune-suppressants for a long time and that Ethicon knew this at the time the first of the Ethicon devices was released to the market.
283 On 12 October 1990 the FDA notified Ethicon that it had reclassified non-absorbable polypropylene sutures from class III to class II and enclosed a letter dated 5 July 1990 to the Vice-President of the United States Surgical Corporation (USSC), another manufacturer of non-absorbable polypropylene surgical sutures.306 USSC had apparently petitioned the FDA to reclassify the non-absorbable polypropylene suture, the FDA’s decision to reclassify the suture followed a review of its petition, and this letter advised them of the outcome of that review and the reasons for the FDA’s decision. That letter, which was tendered by the respondents, included the following passage:
The patient’s health and response to the suture material may affect wound healing (Refs. 11, 17, 18, 48, 66, 74, 80, 81, 87, 107, 108, 109 and 121). Patients whose health has been compromised or weakened by poor nutrition, advanced age, obesity, uncontrolled diabetes, infection, anemia, or with certain forms of cancer, may exhibit delayed wound healing (Refs. 11, 17, 18, 48, 66, 74, 80, 81, 87, 107, 108, 109 and 121) which may increase the likelihood of suture failure.307
284 The letter also stated:
Also, many of the above-identified performance parameters and risks can be adequately controlled by labeling disclosures which may be incorporated into a class II standard or required by the class I misbranding controls, which include, among other things, the requirement of adequate directions for use.
285 This letter put Ethicon on notice that patients whose health had been compromised or weakened by a variety of factors could exhibit delayed wound healing in response to non-absorbable polypropylene and that the risk to their health could be adequately controlled by labelling disclosures, including adequate directions for use. This letter did not directly address the applicants’ allegation. Nevertheless, it was sufficient to alert the respondents that implanting non-absorbable polypropylene, a key material used in the manufacture of all the Ethicon devices, could pose a greater risk of harm to women who were immunocompromised — at least a greater risk of delayed wound healing. It also advised them of action that could be taken to address the risk.
286 Professor Santerre, who, it will be recalled, was a biomaterials expert called by the respondents, stated in his first report that the response to the implantation of a foreign body (like the Ethicon devices) will vary from patient to patient but that polypropylene mesh “would likely be the wrong choice of a material by the surgeon for a patient with any compromised immune system which could delay tissue integration with the mesh”.308 Earlier in his report he explained:
Immune response will depend on a patient’s genetics (inherently has an active or compromised, or moderate immune reactivity), disease state (e.g. diabetes, connective tissue disorders, blood disorders), age (young people typically have much more active and accommodating wound healing than older patients), gender, extent of injury (e.g. more damage to local tissues from the surgical process will generate a greater inflammatory reaction and more challenging host response), environmental conditions (smoking, diet, stress, etc.) and sensitivity to biomaterial chemistry (non-common patient allergies) … If new tissue cannot integrate fast through a porous biomaterial then the material can possibly dislocate, and in some cases ultimately migrate and protrude from the unresolved surgical insertion site.
Hence, it follows on to say that not all patients will achieve a successful integration of biomaterials with new tissue because it is well known that the above parameters compromise wound closure and healing for some patients. If wound healing is different from one human to another, then it also follows that wound healing differs between species. Hence, animal models are not absolute substitutes for human clinical trials and pre-market evaluation.309
287 As I observed in Gill (No 1), several sources were given for the statement in the first sentence. The earliest was a 1997 publication at which time none of the Ethicon devices was available for sale in Australia. None of the sources Professor Santerre cited was a publication one might reasonably expect gynaecologists or urogynaecologists to read, at least not routinely. But as manufacturers of biomaterials, even if it was not actually aware that “wound closure and healing” would be compromised for patients with immunosuppressed conditions, it should have been. Moreover, they should also have been aware that these patients were also at a higher risk than others of migration, extrusion, and erosion of their implants.
288 Professor Santerre discussed the relationship between immune response and infection. He stated:
In addition, if the immune response for a specific patient is not compromised then macrophage activity will be very elevated early on and may contribute to controlling bacteria contamination at the site until healing is finished and new-blood vessel formation is complete (note macrophages have an inherent role of digesting bacteria). If the thickness of the fibrous capsule around the implant is thin and macrophages can easily penetrate the mesh structure from the new blood vessels, then there will remain a relatively low level (e.g. an active basal level) of macrophage activity contributing in part to keeping bacteria contamination in check within the mesh’s integrated tissues.
However, in patients with a compromised immune response (for any of the reasons discussed below in section 4.3), then infection will be a challenge, no matter what synthetic biomaterial will be used in this application. 310
289 Both paragraphs in the above extract were supported by references to a 2001 article on the in vitro interaction of bacteria with polypropylene and polyester prostheses published in the journal Biomaterials in 2001.311
290 In cross-examination, Professor Santerre said that “high responders” were likely to have an adverse outcome from a mesh implant. He agreed, in effect, that the high responders included patients who had autoimmune disorders or were undergoing long-term immunosuppressant therapy.312
291 Professor Santerre considered that the early pre-market studies conducted by Ethicon between 1991 and 1997 “were not immune to the existence of patient to patient variability”, noting that it was observed in early mesh testing and could reasonably be anticipated.313
292 The effect of Professor Santerre’s evidence was that patients who are immunocompromised for one reason or another are likely to experience a more intense inflammatory reaction than patients who are not. It follows that they are and have always been at higher risk of developing complications, such as erosion/exposure, infection, and chronic pain.
293 In the chapter on mesh infection and migration in the book Meshes: benefits and risks, edited by Professor Volker Schumpelick and published in 2004, it was noted that “[i]mmune-suppressed patients are very susceptible to infections …”.314 The book contained the papers presented to the third Suvretta conference in Switzerland in January 2003, which was sponsored by Ethicon315 and attended by a number of Ethicon personnel, including Brigitte Hellhammer and Boris Batke.316
294 In their article on the conclusions of the IUGA Roundtable published online on 6 May 2006, Davila et al noted that the occurrence of healing abnormalities after graft implantation, such as erosion, rejection, and infection, is becoming increasingly recognised as a potentially serious problem.317 They compiled a list of “[c]urrently accepted relative contraindications to the use of biomaterials”. One of those contradictions was “[i]mmunosuppressed patients”. The full list was:
1. History of previous pelvic radiation
2. Severe urogenital atrophy
3. Immunosuppressed patient
4. Presence of active pelvic or vaginal infection
5. Patient currently on systemic steroids
6. Host factors including:
a. Poorly controlled diabetes
b. Morbid obesity
c. Heavy smokers
295 Davila et al (2006) wrote that contributing factors to healing abnormalities like erosion, rejection, and infection were likely to include immune reaction to the graft and individual host factors including patients with sensitive skin and/or multiple allergies. They emphasised the importance of integration of a graft into the host tissue:
Integration of a graft into host tissue is important. The implanted graft should allow for prompt collagen in-growth and neovascularization. This should occur without the occurrence of infection or significant inflammatory reaction. A limited amount of inflammatory reaction is necessary to promote neovascularization and collagen in-growth. Grafts that are poorly integrated include microporous synthetic grafts and biologic grafts treated with chemical cross-linking. These grafts may become encapsulated, leading to hardening, shrinkage, and other graft changes, which may subsequently lead to dyspareunia, alteration of normal anatomy, and increased risk of failure.
296 In 2012 the results were published of a study by the Nordic Transvaginal Mesh Group aimed at identifying risk factors for mesh exposures after anterior pelvic organ prolapse repair using Prolift.318 Elmér et al (2012) wrote:
Somatic inflammatory disease showed a significant, albeit weak, association with mesh exposures. We recognize that the term “somatic inflammatory disease” entails several conditions with a more or less heterogeneous pathoetiology. However, inflammatory diseases such as rheumoid (sic) arthritis, Sjögreńs disease, and lupus erythematosus, also share characteristics such as multiple organ involvement, afflicted connective tissues, and underlying autoimmune mechanisms. These disease traits may by themselves influence wound healing but treatment is often accompanied by immunosuppressant medications which may prolong wound healing. Despite the overall large number of patients in the study, the number of cases with somatic inflammatory disease was few (indicated by the wide confidence intervals) and the association is in need of corroboration in further studies.
297 They concluded that, along with smoking and multiparity (two or more births), somatic inflammatory diseases were possible risk factors for mesh exposures after implantation with Prolift.
298 Ethicon and Gynecare Scandinavia were among the sponsors of the study. Two of the authors, Drs Altman and Falconer, declared that they were advisers to Ethicon and/or Gynecare Scandinavia.
299 It was common ground that the biomechanical properties of the mesh and mesh stiffness in particular are amongst the main reasons for the post-surgery complications of mesh implantation.319 Witnesses on both sides testified that exposure of polypropylene mesh, including the mesh used in the various Ethicon devices, can cause infection. As Professor Deprest put it in remarks expressly approved by Professor Korda,320 “when a material is exposed to the vaginal lumen it will almost certainly become infected: resident bacteria will attach to the implant”.321 Infection may also be a primary process, of course. In such a case it was also common ground that it can cause exposure of the mesh.
300 A good deal of evidence was led by the applicants to explain the mechanisms by which the admitted complications arise or might arise. The evidence was very interesting and some of it was controversial. As a matter of law, however, having regard to Dr Hinoul’s concessions that the devices can cause the pleaded complications, it was unnecessary to prove the mechanism(s) by which they may occur. The respondents accepted as much during closing argument.322 For this reason, my discussion will not reflect all the evidence or all the submissions made in relation to it. Above all, I do not propose to decide disputes simply to satisfy someone’s intellectual curiosity. There is enough to be done without that.
301 The evidence about these issues was given for the applicants by Professors Klosterhalfen and Klinge, Associate Professor Guelcher, and Dr Dunn, and for the respondents by Professors Deprest, Wright and Santerre, and Dr MacLean.
302 I was impressed by the evidence of Professors Klosterhalfen and Klinge. They both struck me as honest and trustworthy. Their knowledge of, and experience with, the use of polypropylene implants, in general, and the POP devices, in particular, was far superior to that of any other witness. Their evidence was informed by their extensive and profound experience.
303 On the other hand I was not impressed with a number of the respondents’ witnesses.
304 I begin with Professor Wright.
305 In contrast to the generally favourable impression I formed of Professors Klinge and Klosterhalfen, I found Professor Wright to be generally unimpressive. He appeared too eager to please the respondents. In his first report to the Court, for example, he wrote that the inflammatory response to the implants was transitory (resolving within to two to three weeks), just as the respondents had written in all the instructions for use (IFUs) supplied with the Ethicon devices.323 In cross-examination, however, he conceded that the inflammatory response was chronic.324 The report cited passages in references in support of his conclusions without referring to passages in the same references that detracted from them. Indeed, his review of the literature appeared to be selective. He was coy about referring to scar tissue. He defined fibrosis as the “[f]ormation of excess fibrous connective tissue in a damaged organ or tissue”.325 But Professor Klinge gave unchallenged evidence that only in certain rare primary diseases does fibrosis occur without any previous tissue injury. Otherwise, he deposed, “fibrosis means scar”.326 Professor Wright also stated that scar contraction typically occurs in burn patients or on the palms, soles, and thorax, and is not normally seen at other anatomical sites.327 But Professor Klinge pointed out that it has been known for over a hundred years that scar contraction occurs everywhere in the body.328 Not only was Professor Klinge’s evidence in this regard not challenged, but there was no dispute between the parties that scars contract around meshes.
306 Professor Wright’s experience and expertise in the area of polypropylene meshes is inconsequential in comparison with that of Professors Klosterhalfen and Klinge. In contrast to the extensive research undertaken by them, Professor Wright admitted to not having conducted independent research on the host response to polypropylene meshes or the biocompatibility of polypropylene meshes for the treatment of pelvic organ prolapse. His experience in the area is recent and confined to the litigation context. It involved looking at “over 50 explants”, about 75% of which, the cross-examination revealed, were only tissue biopsies which included partially excised mesh from SUI or POP devices.329
307 In cross-examination Professor Wright refused to answer certain questions claiming that he did not have expertise in biomaterials or urogynaecology, despite having opined on matters falling within these fields in his two reports. On a number of occasions, he drew distinctions that were unconvincing or unclear (such as the point at which fibrous connective tissue becomes dense enough to be characterised as scar tissue). He ridiculed the concept of bridging fibrosis, discussed below, but when asked whether he could identify an article that was critical of it, he said he could not.330 Some of his positions proved to be inconsistent.
308 Overall, he struck me as arrogant and partisan.
309 For all these reasons, where their opinions conflict, I prefer the evidence given by Professors Klosterhalfen and Klinge to that of Professor Wright.
310 Professor Santerre’s reports read like an unqualified endorsement of Ethicon’s products. He cited a number of papers to support his favourable opinions of Ethicon’s conduct without referring to important parts of those papers that were critical of them. That was most evident in his selective use of chapter 13 of Host Response to Biomaterials, which was written by Barone et al (2015) and titled “Host Response to Biomaterials for Pelvic Floor Reconstruction”.331 Professor Santerre’s reports were also peppered with selective and incomplete citations of complication rates and statements from clinical studies.332
311 In a number of instances Professor Santerre’s evidence was internally inconsistent. For example, it was put to Professor Santerre in cross-examination that “agitation of vaginal tissue which contains an implant could have an effect on the tissue reaction to that implant”. He replied: “The answer is yes, if there is poor integration of that implant around”. He later changed this to: “particularly if there is poor integration of that implant around”. It was then put to him, “So even after full integration it’s wise not to engage in too much physical activity, at the risk of a negative reaction to the implant?” He responded, “Everybody – everybody will be different, and I will leave that to a clinician’s call”.333
312 During cross-examination, Professor Santerre often failed to respond directly to questions.
313 When pressed to answer questions directly, however, he made a number of important concessions. For example, he accepted that, for as long as a foreign body (such as a mesh implant) is present in the human body, there is likely to be an ongoing or chronic inflammatory response.334 In addition, he accepted that it was likely the antioxidants in the Prolene meshes would be depleted and oxidation would occur at the surface of the implants, which could account for the appearance of cracks and flaking observed microscopically by Professor Iakovlev and others.335 In fact, as the applicants submitted, he ultimately agreed with each of the central contentions of the applicants’ biomaterials case.336
314 At first, Professor Deprest appeared to be a model witness. Unlike Professor Santerre and some of the other witnesses, he gave direct and responsive answers to questions put to him. I found him generally impressive, although I became troubled by some aspects of his report. For one thing, I was very concerned about his unqualified statement that “the recent SCENIHR [Scientific Committee on Emerging and Newly Identified Health Risks] (2015) report recommends that PP [polypropylene] should be the only product used in vaginal implants”.337 As far as I can tell, the report itself makes no such recommendation, at least not expressly. Be that as it may, it appeared from his evidence in chief that Professor Deprest endorsed such a recommendation. The cross-examination revealed that nothing could be further from the truth. Professor Deprest was a co-author, with Drs Slack and Davilla, on behalf of the board of the IUGA, of a submission made in response to a draft of the SCENIHR report made available for public comment. The conclusion of the IUGA submission was completely at odds with such a recommendation. Indeed, the submission described the recommendation as indefensible. It reads:
In conclusion we reject the recommendation that Polypropylene is safe for vaginal use… as this is not based on any translational or clinical evidence, and therefore cannot be defended.338
315 Professor Deprest agreed in cross-examination that this was his personal opinion at the time and that he remained of this opinion.339
316 When asked whether it occurred to him that the Court might like to know that his opinion differed from the SCENIHR recommendation, he said that it had.340 His attempt to justify the omission from his report of any reference to that opinion was incomprehensible. At first, he said he was quoting from “the official [SCENIR] report” when he was not. Even if he were, that was no reason not to venture his own opinion. After all, one of the critical matters with which his report was concerned was his opinion on the safety of Ethicon’s urogynaecological polypropylene meshes. Citing the recommendation of the SCENIHR report without mentioning his own opinion was apt to mislead the Court. It amounted to a representation that he shared the view of the Committee or, at least, that he did not disagree with it. Why he put himself in this position is a mystery. It leads me to believe that he was not as impartial as I first thought.
317 Other matters of significance emerged in cross-examination.
318 When asked why he did not include a reference to a paper he had co-authored,341 which contained statements apparently favourable to the applicants’ case, Professor Deprest replied: “I didn’t add it to my evidence because we usually don’t cite ourselves”.342 I must say I found this reply difficult to believe, especially since he had cited in his report some 19 articles upon which his name appeared as an author.
319 A party may only call an expert to give expert evidence at a trial if the party has served a report from the expert which, amongst other things, contains an acknowledgment that the expert has read, understood and complied with the practice note on expert evidence: Federal Court Rules 2011 (Cth) (Rules), r 23.11, read with r 23.13.
320 Professor Deprest claimed to have read and complied with the Expert Evidence Practice Note (GPN-EXT). Moreover, he agreed to be bound by it. Clause 5.2 of the Practice Note requires that the expert attach or exhibit to his or her report documents that record any instructions given to him or her. Professor Deprest admitted to receiving a letter from Clayton Utz, the respondents’ solicitors, but did not attach that letter to his report. He also said that he had had numerous teleconferences with the lawyers. It is inconceivable that no notes were taken at those teleconferences. Later, however, in response to a call for documents recording instructions, including notes of conversations, the respondents’ lead counsel, Mr Finch SC informed the Court that there were no such documents. 343 The absence of any such document is troubling. While it may not be a breach of the letter of the Practice Note, it constitutes a failure to adhere to its spirit.
321 In BrisConnections Finance Pty Limited (Receivers and Managers Appointed) v Arup Pty Limited (2017) 252 FCR 450, Lee J considered the importance of a letter of instruction in providing clarity to the Court about the questions the expert has been asked and the assumptions he or she was asked to make. His Honour’s remarks arose in the context of a letter of instruction that was apparently provided (almost) contemporaneously with the finalisation of the expert report. His Honour stated at :
The point of a letter of instruction being annexed to a report is not to act out a stylised ritual, but to provide to (scil) the Court with a transparent indication of what has been provided to the expert and the questions that the expert was actually asked to address. It should be able to be read literally without being silly. As is (at the very least) implicit in FCR 23, the work of the expert is to attend to the questions “the expert was asked to address”, not to invert the process by using the expert’s specialised knowledge in order to identify the questions that should have been asked and the assumptions that should have been given. … The integrity of the expert evidence process and the independence of experts is best facilitated by transparency in what is being asked of experts prior to, or at the time, they are forming their opinions and, if the questions need to change because they are misdirected, a record being made by way of supplementary instructions as to what has changed.
(Original emphasis in bold, italicised emphasis added)
322 I respectfully agree. The lack of transparency means that the Court may never know the extent to which the witness’s opinion has been influenced by the party retaining the witness.
323 No doubt for good reason, Professor Deprest was not asked in cross-examination what oral communications he had had with the respondents’ lawyers. For a start, the cross-examiner would not know the answer he might receive to such a question. In any case, he might well not have remembered. Even if he purported to have a memory of them, it would in all likelihood be unreliable. Finally, any answer he might have given could not be tested.
324 Dr MacLean, for the most part, responded to questions with what appeared to be well-reasoned answers. But his failure to fully disclose the relationship with Ethicon of his company (Exponent) and his colleagues (Drs White and Ong) raised doubts about his independence and the weight to be given to one of the studies on which he relied (the Thames study).344 This was particularly so as he was given an opportunity during examination-in-chief to disclose any past relationships with Ethicon.
325 The absence of any letter of instructions to Dr MacLean (from either Clayton Utz, which represent the respondents in the present proceeding or the US firm, Butler Snow, which represented Ethicon in US litigation) is also of note. When asked to describe his interactions with Chris Hutchison of Butler Snow, Dr MacLean was not precise in his responses, giving me cause for concern around the content of his instructions.
326 There is no dispute that “the ideal mesh” is yet to be found and that the optimal biomechanical properties for urogynaecological meshes remain unknown.345 This was a point the respondents made repeatedly as if it were a mitigating, if not exculpatory, factor.
327 Moreover, the respondents accepted that pore size is a critical component of the biocompatibility of any mesh used in vivo and that pores must be of sufficient size to allow the entry of macrophages “and other processes” to ensure that bacteria are not housed in the pores of the mesh.346 Critically, they also accepted that the pore sizes change in vivo and that the mesh contracts.347
328 There were seven broad issues in contention:
(1) whether the changes to pore sizes in vivo are significant;
(2) whether the classification system upon which many of the respondents’ witnesses relied is still applicable;
(3) whether data derived from the pelvis data pool compiled by Professor Klosterhalfen and others at the University of Aachen is unreliable and, if so, the extent to which that affects his and Professor Klinge’s opinions;
(4) whether bridging fibrosis can occur with any of the Ethicon devices and, if so, whether it is of clinical significance;
(5) whether mesh contraction is of clinical significance;
(6) whether chronic pain is attributable to nerve entrapment in the contracted mesh or scar tissue; and
(7) whether the polypropylene used in the Ethicon devices degrades in vivo and, if so, whether the degradation is clinically significant.
329 Before dealing with these issues, however, it is necessary to provide some context.
330 All the Ethicon devices are made from biomaterials. A biomaterial is a material used in a medical device which is intended to interact with biological systems.348 Professor Wright said that a biomaterial is regarded as biocompatible if it has the ability to perform with an appropriate host response in a specific situation. This definition of biocompatibility was drawn from the writings of Professor David Williams, a noted expert in the field, and has been current since 1987.349 It is not materially different from the definition given by the Second Consensus Conference on Definitions in Biomaterials in Chester, UK, in 1991 (“the ability of the material to perform with an appropriate host response to the application”) to which Associate Professor Guelcher, one of the applicants’ biomaterials experts, referred.350 It drew criticism, however, from Professor Klosterhalfen:
In my opinion, the “Williams definition” is unclear: I do not understand the term “appropriate host response in a specific situation”. Who defines what is appropriate? Further, this definition only relates to the host response (in particular, tissue response), but not to the function of the material or device.
Moreover, the definition is only appropriate for short-term applications, such as preclinical animal studies. For permanent implanted devices (like surgical meshes) the definition should be modified as follows: the biocompatibility of a permanent implantable medical device refers to the ability of the device to perform its intended function, with the desired degree of incorporation in the host, without eliciting any undesirable local or systemic effects in that host.351
331 I doubt, however, that there is much difference between the views of Professors Klosterhalfen and Williams on this point. Indeed, Professor Klinge cited Professor Williams in support of the definition given by Professor Kosterhalfen.352 In another article, published in the journal Biomaterials in 2008, Professor Williams wrote:
This deﬁnition, which clearly places the word in the category of a concept rather than a practical descriptor of a process, is based on the three tenets that a material has to perform and not simply exist in the tissues, that the response which it evokes has to be appropriate for the application, and that the nature of the response to a speciﬁc material and its appropriateness may vary from one situation to another. 353
332 In an earlier article, published in the Journal of Applied Polymer Science in 1994, of which he was a co-author, Professor Williams made it clear that to be biocompatible a permanent implant should not have any harmful effect on human tissue.354
333 As I mentioned earlier, Professor Williams defined biocompatibility as “the ability to perform with an appropriate host response in a specific situation”. In a chapter entitled “General Concepts of Biocompatibility” published in Black J and Hastings G (eds), Handbook of Biomaterial Properties (Springer Science+Business Media, 1998) he also made the following pertinent observation:
To recognize the very effective performance of a material under one set of conditions but then to assume that the same material can perform equally well under entirely different circumstances is inherently dangerous since it takes into account neither the variations one might expect to see in the host response from site to site nor the fact that what is appropriate for one situation may not be appropriate for another.355
334 Yet, in 2003, six years after TVT was launched and many years after Prolene had first been used for hernia repair, Professor Klosterhalfen and others wrote in an article on “Biological Response to Mesh” published in the journal, European Surgery, that “[o]ur knowledge about the long-term biocompatibility and tissue response of mesh in humans is still poor, although a few reports exist”.356 In September of the same year, Drs Arnaud and Robinson, in a joint presentation for Ethicon, observed that “the scientific knowledge about the use of meshes in surgery is still in its infancy, at least for pelvic floor applications”.357 They emphasised the necessity for the mesh to resist infection and the high risk of infection for pelvic meshes, observing that “[t]he vaginal approach is a rather unique situation in surgery as a prosthetic material is placed through a septic cavity” and that before TVT, apart from the mouth and ENT surgery, “meshes [were] never used in such a condition”.
335 In the chapter to which I referred, Professor Williams went on to point out that for a material to be biocompatible “the interaction between the material and the tissues is one which leads to an acceptable balance between inflammation and repair”. This is presumably what he meant by “an appropriate host response”. Professor Williams emphasised that “host variables are as important as material variables in the determination of biocompatibility”. Not all tissues, he added, not even all tissues of the same variety, will be able to respond in the same way and host variables like age and overall health status will have a major effect.358 At the same time, he noted that there are significant regional and tissue-specific variations in the local host response.359
336 One of the other points Professor Williams made in this chapter was that it was necessary to recognise that, if there is an interfacial reaction (that is, a reaction at the place where the foreign body is in contact with the tissue), there is no reason why the products of that reaction and their effects have to be confined to the locality of the interface. He said that the presence of a benign local response does not necessarily mean there will be no systemic or remote side effects.360
337 In an article published in 1982 in the Journal of Materials Science, Professor Williams also wrote:
Although the specific materials requirements will differ according to the nature of the application, it is a fundamental requirement in each and every case that the polymer should display adequate biocompatibility. This implies that, for permanent implant applications, the material should not degrade within the physiological environment, nor should it have any adverse effect on the tissue …361
338 To similar effect, Professor Klinge’s evidence was that the biocompatibility of a long-term implantable medical device can be defined as the ability of the device to perform its intended function, with the desired degree of incorporation in the host, without eliciting any undesirable local or systematic effects in that host.362
339 In an article tendered by the respondents, published in the International Urogynecology Journal in July 2003, Dr Michel Cosson and his fellow authors reviewed the literature on all types of synthetic implants used in both prolapse repair and the treatment of stress urinary incontinence and analysed the mechanical properties of, and tolerance to, the various devices. They concluded:
This review of existing prosthetic products demonstrates that no perfect product currently exists. Two categories of product seem to us to have promising properties with regard to their use in transvaginal surgery for restoring pelvic function. On the one hand are synthetic implants with mechanical properties of strength and elasticity, essentially made of polypropylene. Their strength is unchallenged, but it remains to be established whether they are well tolerated when introduced by the vaginal route.363
340 In these circumstances, it is difficult to understand how the Ethicon devices could have been considered biocompatible or suitable for release to the market without first testing their performance in vaginal tissue in a wide selection of patients and over a sufficient period of time.
341 When a biomaterial such as polypropylene is implanted in the body, it provokes an inflammatory response in the host tissue. This inflammatory response is known as the “foreign body reaction” or “foreign body response”. It causes a layer of scar tissue to form around the implant or the pores of the implant, which is weaker and more rigid than normal healthy tissue.364 Indeed, synthetic non-degradable meshes are believed to work by inducing this foreign body reaction and the consequential fibrosis which enables the foreign material to adhere to the host tissue to provide support to weakened support structures.365
342 The scarring is apparent after about two to three weeks and certainly after about six weeks.366 The extent of the foreign body response (and therefore the extent of the inflammation) is influenced by a number of factors to which I will come in due course.
343 The inflammatory response takes place in stages. First proteins are adsorbed. In the early stages, mononuclear cells (monocytes) migrate to the surface of the biomaterial, where they can adhere and participate in the foreign body reaction.367 Monocytes, produced by the foreign body reaction, can differentiate into macrophages and foreign body giant cells, which are a collection of used macrophages.368 Macrophages (literally “large eaters” or “feeders”) are specialised phagocytes (literally “feeder cells”). They remove dying or dead cells and cellular debris and clear bacteria.369 Associate Professor Guelcher described the space between the surface of the biomaterial and the adhered macrophage as a “privileged microenvironment” where these types of cells release reactive molecular species called “reactive oxygen intermediates” or “reactive oxygen species”, as well as acids and enzymes. 370 Reactive oxygen species are chemical compounds containing oxygen (such as hydroxyl radicals and hydrogen peroxide) that are secreted by the inflammatory cells while they are adhering to the foreign body.371 I will return to the subject of the reactive oxygen species later in these reasons when discussing the applicants’ case that the Ethicon devices are subjected to oxidative degradation after implantation.
344 Macrophages then fuse and become multinucleated foreign body giant cells, which adhere to the surface of the mesh. They attract and activate fibroblasts (the main cells that form scar tissue) and angiogenesis (the formation of new blood cells). A major effect of the foreign body reaction is that the implant is covered by, or embedded in, a granuloma (a mass of granular tissue) with an inner layer of macrophages and foreign body giant cells and an outer layer of fibrotic (scar) tissue.372
345 The foreign body reaction to an implant can continue over an extended period of time. Unlike normal wound healing, where the number of inflammatory cells decreases over time leaving only the scar tissue, in a foreign body reaction the inflammatory cells may persist for years, even decades, becoming “a chronic inflammatory response”.373 For a permanent implant, like each of the Ethicon devices, the response endures for the duration of the time the implant remains in the body. As Professor Deprest put it in his first report:
Between the openings (pores) present in the implant, the body “grows” scar tissue, so that the implant eventually integrates or becomes part of the body, and ideally exerts its supportive function … When an implant is not degradable the remodelling and inflammation process is permanent …374
346 The persistence of the foreign body reaction is important, especially in younger patients in whom the mesh will remain for several decades. As Professor Klinge observed, it means that in many clinical studies with short periods of follow-up, the morbidity rates are likely to be underestimated.375
347 The foreign body reaction can also vary in intensity, and the intensity of the foreign body reaction, more particularly the extent of the inflammatory response, correlates with the extent of the fibrosis.376 Professor Klinge said that histopathological investigations have shown that the amount of fibrosis is directly related to the amount of the inflammatory, cellular foreign body reaction induced at the interface of the biomaterial and the host tissues. The Aachen Group directed its early mesh research to reducing the amount of foreign material used, its surface area, as well as the density of the mesh primarily in order to minimise the degree of inflammation.377
348 Professor Klosterhalfen’s evidence was that the intensity of the foreign body reaction to polypropylene can be explained by the failure of macrophages to resolve the inflammation in the initial period following implantation,378 the characteristics of the surface area of the mesh in contact with the host tissues,379 differences in the surrounding tissues and the mechanical loads and stresses to which the implant is subjected,380 and the genetic background of the particular individual.381
349 The greater the surface area of the device, both in terms of the amount of fibre length and the density and diameter of the fibres, the greater the foreign body reaction and the inflammatory response.382 Studies have demonstrate that the extent of the reaction is dose-dependent.383
350 Furthermore, the respondents recognised that decreasing the biomaterial content of polypropylene meshes reduces the inflammatory response.384
351 Professor Klosterhalfen also deposed that mechanically cut mesh was “particularly problematic” because it frayed and pieces of polypropylene could break off during implantation. These characteristics, he said, created a greater inflammatory response leading to “greater fibrosis and scarring”.385 He also said that mechanically cut mesh also creates sharp edges that can cut tissues, which can cause pain and increase the likelihood of erosion, and cutting mesh mechanically can cause the mesh to rope.386 He was not cross-examined about any of this evidence and I did not understand it to be controversial.
352 As I indicated earlier, studies demonstrated that the foreign body reaction in response to the implantation of polypropylene mesh continues until the mesh is removed. Despite that, until relatively recently, in the instructions for use issued for all the Ethicon devices the foreign body reaction was described as “transitory”. Notwithstanding what they said in the instructions for use, Ethicon was well aware of that.
353 Ultimately, despite what the respondents were telling consumers, it was common ground that as long as the implant remains in the body, the foreign body or inflammatory response to polypropylene implants, including the devices in question, is not “transient” or “transitory” but chronic and permanent.387 Indeed, this was an intended outcome, as some fibrosis (scarring) is required to enable the device to adhere to the tissue and remain in place. The problem is when the inflammatory response is greater than necessary to create the desired level of fibrosis. Professor Klinge said in his affidavit that “incorporation of the mesh into tissue is not the problem; it is the incorporation of mesh into scar tissue that is the issue”.388 Scar tissue is non-functional tissue and the amount of scar tissue around the mesh is responsible for the stiffness and “functional outcome of the implanted tissue area”.389
354 The respondents submitted that, given the consensus about the duration of the foreign body reaction, the question for determination is whether the applicants have proved that the foreign body reaction has a clinically significant effect and causes the asserted complications.390 The answer to the question is that they did prove that the foreign body reaction is clinically significant and can cause many, if not all, of the pleaded complications.
355 Cobb et al (2005), for example, wrote that the intensity and extent of the foreign body reaction was the most important factor influencing the biocompatibility or tolerance of a subject to the foreign body.391
356 Deprest et al (2006) said that local complications in synthetic meshes, including low-weight, large pore, monofilament materials, seem to be related to an increased foreign body reaction.392
357 Moreover, as the applicants pointed out, the study by Elmer et al on the histological inflammatory response to Prolift, published in the Journal of Urology in 2009, and relied upon by both Professors Santerre and Wright, suggested a relationship between the inflammatory response to polypropylene mesh and complications, such as erosions.393 It found that eight out of 10 patients (80%) had mild granular formation and two erosions one year after surgery. The authors emphasised that a larger study population of patients who had undergone transvaginal mesh surgery “would undoubtedly generate an increased number of mesh complications”. They added that clinicians and patients should be aware of the possibility of late onset “mesh related inflammatory reactions” when using a large polypropylene mesh (like Prolift).394
358 In cross-examination, Dr Hinoul acknowledged that “the mesh tissue interaction leads to scar contraction”.395 Professor Roovers also testified that mesh shrinkage is caused by the fibrotic reaction and that this can cause the scar to contract and the mesh.396 For the reasons given below, scar contraction can cause a number of complications.
359 Professor Klosterhalfen said that an intense foreign body reaction with bridging fibrosis (discussed below) can cause chronic pain.397 He noted that, while fibrosis is useful for the integration of the mesh, “excessive fibrosis with bridging is fatal for mesh function and the long-term outcome”.398
360 Dr Daniela Ulrich and others (including Associate Professor Rosamilia) wrote in an article published in 2012 that, depending on the nature and extent of the response, this inflammatory response is primarily responsible for the significant complications arising from the use of synthetic meshes, including mesh exposure/erosion, mesh contraction, infection, and pain.399
361 Ethicon told the FDA in February 2008 that the benefit of Prolift+M over Prolift was its ability to reduce the amount of foreign body remaining in the patient after implantation.400 If the foreign body response was not clinically significant, there would be no need to reduce the amount of the foreign body.
362 In March 2009, Dr Hinoul and his collaborators from Ethicon wrote that the foreign body reaction induced by the mesh used in Prolift+M (both acute and chronic) can lead to both exposure and shrinkage.401
363 A good deal of evidence on the biocompatibility of the Ethicon devices and the complications that can be attributed to the use of polypropylene mesh derived from the work of a multi-disciplinary group of doctors and scientists based in Aachen in northwest Germany. Professors Klosterhalfen and Klinge were members of that group.
364 Professor Klosterhalfen is a surgical pathologist who has worked in the field of anatomical and clinical pathology for over 25 years. He has devoted much of his career to the study of the body’s response to implanted devices, including polypropylene meshes, and how the design of those devices influences biocompatibility.402 He claims to be the first pathologist to describe in detail the foreign body reaction to polypropylene mesh.403
365 As a resident in the Department of Surgery at the Medical University in Aachen, Professor Klosterhalfen met Dr Uwe Klinge (now Professor Klinge), an abdominal surgeon who was then specialising in general surgery.
366 Professor Klosterhalfen became a consultant in general pathology in 1998, was appointed an associate professor of surgical pathology the following year and in 2003, only five years after he had completed his training in pathology, was appointed to his current chair.404 In 2003 he established the Institute of Pathology in Düren, Germany, which he still heads. It is one of the top three institutes of pathology in Germany. It processes over 150,000 tissue specimens a year, including about 20,000 tissue samples from the female pelvic floor and genital tract. Professor Klosterhalfen is personally involved in the study of about 50,000 samples each year and, since the Institute was established, he estimated, he has had responsibility for the analysis of over 600,000 tissue specimens.405
367 Both he and Professor Klinge are prolific researchers and have published widely on the effect of biomaterials used in both abdominal and pelvic surgery.
368 Professor Klosterhalfen is a highly respected pathologist. He described himself as a world leader in understanding the mechanical properties of polypropylene mesh and the response of human tissue to it.406 The respondents did not quarrel with this description. His opinions on these matters are regularly sought by scientists, physicians, and industry.407 Those people who have had professional dealings with him have the utmost respect for his opinions and the quality of his work. That was apparent from a host of Ethicon documents. Amongst others, Professor Deprest sent explants to him. In cross-examination Professor Deprest said that he did not question Professor Klosterhalfen’s analyses of the explants and did not doubt his knowledge.408 Dr Hinoul described him in an email to colleagues at Ethicon as “the god of surgical pathology on the subject of textile implants in this solar system”.
369 JJM or another company in the Johnson & Johnson group (the evidence does not permit me to say) engaged an external consulting firm, PA Consulting Group, to investigate mesh erosion in pelvic floor repair. In particular, PA Consulting was asked to review the literature and conduct “a broad analysis of the problem”, which was to include an informal meta-analysis of the literature, interviews of both internal and external experts, and consideration of potential animal models. The Group reported on 18 May 2011.409 Professor Klosterhalfen was the only pathologist who was interviewed.
370 While working as a general surgeon Professor Klinge performed thousands of abdominal operations, many of which involved the use of synthetic surgical mesh. He became a specialist in abdominal or visceral surgery in 2004. In 2005 he was appointed professor of surgery at Aachen University.410 He stopped operating in 2006 to concentrate on research and for the last decade he has been the principal scientific investigator in the university’s surgical department.411 Although he has never performed surgery for the repair of stress urinary incontinence or pelvic organ prolapse, Professor Klinge has used meshes manufactured by Ethicon in the course of hernia repairs.412
371 In the early 1990s, as a result of observing and performing revision operations on patients who had been implanted with surgical mesh for the treatment of hernias, amongst other things, Professor Klinge became aware of complications related to the use of mesh including the build-up of wound fluid around the mesh, pain, infection, and fistula formation into the bowels. Keen to learn how to avoid those complications, he started research in the area of biomaterials science.413
372 At that time Professor Klinge invited Professor Klosterhalfen to become part of an interdisciplinary working group at the University of Aachen. The Interdisciplinary Centre for Biomaterials Research or IZKF BIOMAT was thereby established.414 Its object was to bring together specialists from multiple disciplines, including chemistry, medicine, engineering, physics, and biology for the purpose of researching and developing biomaterials, including surgical meshes.415 The sub-group examining surgical meshes (the Aachen Group as it was sometimes known) aimed to find a textile construction that was best adapted to the physiological requirements of the abdominal wall or groin so as to improve the biocompatibility of mesh used in hernia repair surgery.416 Before long, however, the work of the group expanded to include pelvic mesh. It was in the early 2000s that Professor Klosterhalfen began examining tissue samples from female pelvises which had been implanted with synthetic mesh.417
373 The pelvis data pool was created following an approach in 2003 by Dr Peter Meier on behalf of Ethicon. Professor Klosterhalfen understood that Dr Meier wanted him to assess tissue reaction (including the foreign body reaction) to pelvic meshes in the same way he had done with hernia meshes. The first 172 explants were sent to him by Ethicon.418 Later, clinicians from around the world started sending him mesh explants for analysis directly.
374 In the course of this work, Professor Klosterhalfen has examined thousands of tissue samples containing mesh and has built two data pools, one relating to hernia mesh (the hernia data pool) and another relating to pelvic mesh (the pelvis data pool).419 At the time he swore his first affidavit in early October 2016, the combined data pool, which he referred to as his data pool, contained more than 15,000 explanted polypropylene mesh tissue samples. Professor Klosterhalfen believes the data pool to be the largest such data pool in the world. Of the 15,000 explanted polypropylene mesh tissue samples, approximately 5,000 are human explants. The rest are preclinical animal explants. Of the 5,000 human explants, approximately 4,300 are hernia mesh explants and 700 are pelvic mesh explants.420 Of the 700 pelvic mesh explants, approximately 490 are sling explants and the remaining 210 are prolapse mesh explants. The data pool contains explants from each of the nine Ethicon devices, the subject of the proceeding.421 Some 1,000 of the 4,300 hernia mesh explants are full explants; that is to say that in each case the entire implant and tissue was removed for analysis.422
375 The year 1998 saw the publication in the European Journal of Surgery of an article by Professors Klinge, Klosterhalfen and others, including Dr Oblenski of the Institute for the Research and Development of Textile Materials in Aachen, entitled “Modified Mesh for Hernia Repair that is Adapted to the Physiology of the Abdominal Wall”.423 The article reported on the results of a rat study, the objective of which was the development of a new mesh for hernia repairs. The authors determined that the tensile strengths of meshes used for hernia surgery were far beyond what was necessary for fascial closure. Adjusting the meshes to the physiologically-required forces allowed for a considerable reduction in the amount of material. The authors postulated that reducing the material might reduce the rate of local wound complications and the degree of restriction of mobility of the abdominal wall caused by the then current meshes. The textiles they analysed were Mersilene (a polyester mesh made by Ethicon), Prolene, and the new combination of polypropylene with polyglactin 910, which they called Soft Hernia Mesh, which became Prolene Soft, and which was used to make UltraPro hernia mesh and Prolift+M. The effect of the addition of polyglactin was to reduce the non-absorbable component and therefore the amount of mesh after implantation.
376 This was one of the first papers to describe how the structure and mechanics of mesh affect its biocompatibility. Until then, according to Professor Klosterhalfen, the general view was that any complications with mesh were due to the surgery and not the product.424
377 Professor Klosterhalfen also demonstrated that a reduction in the amount of material reduced the foreign body response.425 That same year, Professor Klosterhalfen published the results of a dog study which showed a reduction in shrinkage of mesh with the reduction of polypropylene content.426 By 2000, if not earlier, Ethicon accepted that the amount of implanted material affects the extent of “mesh-induced inflammation”.427
378 Professor Klinge had a long-standing relationship with Ethicon. In December 1993, while he was preparing for a talk about mesh at a conference on inguinal hernia supported by various mesh manufacturers, he met one of Ethicon’s engineers, Dr Boris Obolenski.428 In February the following year Professor Klinge attended a meeting at the Suvretta House Hotel in St Moritz, Switzerland, to discuss hernia repair surgery and, in particular, the use of mesh. In his presentation he stressed that there was insufficient data to relate the properties of mesh to the biochemical demands and argued that it was necessary to look for textile constructions designed specifically for the purpose or reinforcing the abdominal wall.429 This was the first of five meetings which became known as the Suvretta meetings or conferences. Professor Klinge attended them all. So did Professor Klosterhalfen.430 Ethicon was one of a number of sponsors of the 1994 meeting but the sole sponsor of all subsequent meetings.431
379 After the 1994 Suvretta meeting, Professor Klinge prepared a draft protocol for a project the main objective of which was to find a textile construction better adapted to the physiological requirements of the abdominal wall or groin and so improve the biocompatibility of hernia meshes. Ethicon agreed to join the project and provide the filaments for the necessary tests and the currently available mesh constructs for comparison.432
380 Between 1994 and 2000 the Aachen Group met regularly with Ethicon scientists and other personnel. Aachen University entered into a contract with Ethicon in 1995 and the project continued for 10 years, supported by research grants.433 During this time, Professor Klinge said that the Group learned a great deal about the textiles, defined some standard biomechanical characterisation for better comparison, established models for testing the tissue response in animals, looked for parameters that reflected the inflammatory and fibrotic activity of the foreign body reaction, developed a technique to quantify the biomechanical impact of the meshes on the stiffness of the abdominal wall, and measured the biomechanical properties of tissues.434
381 In 1999, the year after the second Suvretta meeting, a textbook entitled Incisional Hernia was published.435 Professors Klinge and Klosterhalfen contributed chapters on experimental and histological aspects of biomaterials. There, they emphasised the importance of biocompatibility and called for an examination of whether meshes were necessary at all and whether they were “specifically suitable”. They pointed out that, “in contrast to normal suture material, it is necessary to choose the material very carefully”.436 Amongst “the important points” of their histological studies on surgical meshes, they noted that surgical meshes are not inert, even years after implantation; that they lead to chronic irritation of the host tissues; that the tissue reaction is dose-dependent (the greater the weight and surface area in contact with the mesh the greater the irritation of the tissues); and that the use of “heavy-weight meshes such as Prolene” should be avoided.
382 In May 2000, Professors Klinge and Klosterhalfen attended a meeting in Hamburg with various Ethicon personnel including Dr Holste. There, amongst other things, they discussed reducing mesh material by between 5% and 10%.437
383 In November 2001, Ethicon engaged Professor Klosterhalfen as a consultant.438 Professor Klosterhalfen also participated in studies of mesh implants, including those sent by Ethicon for histopathological examination and analysis, and reported periodically to Ethicon on his observations.439 For the next ten years he shared the results of his studies and research with Ethicon. He provided advice to Ethicon about how it could improve the design of its mesh so as to achieve better tissue integration and reduce mesh complications, including the foreign body reaction, infection, erosion, and contraction.440 In August 2002, for example, he sent Ethicon a copy of a paper he had written on the biological responses to mesh which pointed out that, while modern biomaterials including polymers are physically and chemically inert, they are not biologically inert. To the contrary, he wrote, they trigger a wide variety of adverse responses in vivo including inflammation, fibrosis, calcification, thrombosis, and infection.441
384 He also discussed his research results during regular visits to Aachen by members of Ethicon’s Research and Development team, especially Dr Engel and Dr Holste, and he would attend Ethicon’s Hamburg premises where he would discuss his research and give presentations.442
385 The work of the Aachen Group led to the launch by Ethicon in 1999 of Prolene Soft mesh (used in Gynemesh PS, Prolift, and Prosima) and of a number of hybrid meshes, made of a combination of polypropylene and an absorbable polymer to improve handling and reduce the amount of polypropylene. One of those hybrid meshes was UltraPro, launched in 2002, which was the mesh used in Prolift+M.443 All the tests the Group conducted on these meshes were premised on their use in the abdominal wall or the groin (the area of the hip on each side of the body located in the folds where the abdomen joins the upper thigh, also known as the inguinal area).
386 In July 2007 Professor Klosterhalfen met with Drs Meier and Holste. At that meeting Professor Klosterhalfen discussed the need to design a monofilament, very large pore mesh that is elastic and adapts to the biomechanics of the vagina.444 At the same meeting, Professor Klosterhalfen suggested that Ethicon contact Professor Prescher in Aachen about the morphology of the pelvic floor.445 I was not taken to any evidence to indicate that this suggestion was taken up.
387 In October 2007, after examining about 20 pelvic floor meshes received from Professor Deprest, Professor Klosterhalfen concluded that nearly all exhibited (signs of) erosions and demonstrated the same tissue reaction as hernia meshes but with greater incidence of infection.446
388 By early April 2008 Professor Klosterhalfen had evaluated about 100 explanted pelvic mesh samples. He deposed that his analysis showed that mesh erosion was typically accompanied by mesh infection and that mesh shrinkage and contraction was more apparent than in hernia explants.447 In June 2009 he prepared an intermediate report for Ethicon based on his analysis of 172 pelvic mesh explants. He found that the tissue reaction to pelvic and hernia mesh was the same but that folding occurred more frequently in the pelvic mesh explants which he attributed to “the mechanical mismatch” (discussed below) between the implant and the pelvic tissue.448
389 Professor Klosterhalfen’s consultancy agreement with Ethicon, signed on 26 November 2001,449 was renewed on several occasions. In December 2011, however, Professor Klosterhalfen advised Dr Holste that he would not be renewing his contract.450 In his affidavit he stated that his relationship had become strained during 2009 when he felt that Ethicon was no longer interested in collaborating with him but wanted him as an opinion leader, to become more involved in marketing and promoting its products. Professor Klosterhalfen went on to explain:
Ethicon wanted me to say that Prolift + M and UltraPro solved all of the problems which I had identified with mesh. However, I did not believe that. I considered that the fundamental problem with all polypropylene mesh was that it did not have the mechanical characteristics suitable for the female pelvis.451
This evidence was not challenged in cross-examination.
390 Professors Klosterhalfen and Klinge expressed strong views about the unsuitability of the Ethicon devices for the purpose for which they were supplied and the deficiencies in the IFUs accompanying them.
391 Based on their research, testing conducted by others, including Ethicon’s own scientists, and their examination of explanted meshes, Professors Klosterhalfen and Klinge concluded that pore size and geometry are the most important influences on the outcome of tissue repair using a synthetic material. They concluded that pore sizes change in vivo due to the biomechanical forces to which the meshes are exposed so that, contrary to what had earlier been assumed, a “large-pore” mesh, as the respondents contended both Prolene and Prolene Soft were, can deform under load. If that occurs, Professors Klosterhalfen and Klinge said, there is an increased risk of injury by scarring.452 What matters, they argued, is not just porosity, but “effective porosity”.
392 Professor Klinge explained that, except for some cells in the heart and the brain, cells in the body are permanently subjected to remodelling; they die and are renewed or replaced by other cells. Cells in soft tissues die within weeks or months and are replaced by local stem cells. Scars, on the other hand, are always replaced by scar tissue, fat tissue by fat tissue, and fascial tissue by fascial tissue. One effect of this permanent cellular turnover is that any strain imposed on a foreign body (including a mesh implant) will cause the foreign body to migrate to escape the stress and an area of “enhanced remodelling” will be preserved at the interface as the inflammatory cells will be replaced after a shorter period than scar or collagen.453 The relevant forces include those exerted during ordinary activities such as straining on the toilet, exercise, sexual intercourse, even just standing up. The degree of risk increases with the degree of strain.454 These matters are discussed at greater length below.
393 Professor Wright said that it was difficult to envisage a mechanism that would allow mesh to migrate once fibrous connective tissue has grown into the pores and anchored the mesh in the adjacent vaginal tissue.455 This evidence reflected his lack of relevant clinical expertise.
394 Ethicon considered that insufficient integration of fascial tissue caused migration of mesh.456 Members of the TVM Group described migration of mesh prostheses as a complication of the use of mesh in the surgical treatment of vaginal prolapse.457
395 Professor Klosterhalfen said that all the studies undertaken by the Aachen Group indicate that even decades after implantation surgical meshes show a tissue and cell turnover at the mesh interface. Consequently, there is never complete integration or an endpoint to integration of the mesh.458
396 Professor Klinge pointed out that migration of foreign bodies is an everyday experience in orthopedic surgery and also in general surgery for the treatment of anal fistula. He went on to say that:
Remodelling of the soft tissues is a fact, and the more inflammation, the more proliferation, the more remodelling, and the higher the risk for migration. The presence of granulomas always means chronic inflammation, which always favours migration in case of mechanical strain (as in the pelvic floor area).
397 Professors Klosterhalfen and Klinge referred to the different forces at work in the abdomen and the pelvic floor and the unique nature of the vaginal environment.459 In its pre-market notification to the FDA for Prolene Soft, Ethicon claimed that the process by which the mesh is knitted, interlinking each fibre junction, provides for elasticity in both directions and this “bi-directional elastic property allows adaption to various stresses encountered in the body”.460 The representation of elasticity in both directions was incorporated into the instructions for use for all the POP devices. Professor Klinge considered that, while this constituted an acknowledgment that elastic elongation occurs as a result of mechanical stress and that the devices are subjected to strain in vivo, the concept of bi-directional elasticity does not account for the anisotropic behaviour of pelvic tissues. “Anisotropic” derives from the Greek words anisos meaning unequal and tropos meaning turn. Anisotropic objects or substances have a physical property which has a different value when measured in different directions. Bilateral elasticity is not enough to allow the mesh to adapt to the various stresses encountered in the body.461 Moreover, the reference to elasticity is misleading. An elastic material is one that spontaneously resumes its normal bulk or shape after having been contracted, dilated, or distorted by external force, whether after a short or a long interval.462 Both professors pointed out that polypropylene is not an elastic polymer since, once stretched, it does not return to its original state. In cross-examination, Professor Klinge explained that deformation of the pores of the mesh stretches the small loops connecting the fibres and, over time, causes both to elongate with use.463
398 Professor Klinge said that the Prolene mesh in the SUI devices contracts or shrinks by 30-50% after implantation, citing studies of the mesh after implantation in dogs conducted by the Aachen Group in 1998, before the launch of TVT.464 He also said:
A knitted surgical mesh device like the TVT that is permanently implanted in human tissue must be designed in such a manner that the pores of the mesh do not collapse and deform upon the expected forces of implantation as well as the expected in vivo forces. Under minimal strain, the TVT mesh pores deform and collapse thereby increasing the risk of injury to patients in which it is implanted and is a less safe design than products that better withstand these conditions and do not display these poor outcomes. It is my opinion that permanent deformation and pore collapse of the TVT mesh leads to fibrotic bridging, scar plate formation, excessive scarring through and around the mesh and a host of tissue complications that can lead to chronic pain, recurrence, erosions, dyspareunia and need for reoperation, to name a few, making it unnecessarily unsafe for its intended purpose of being permanently implanted in a woman’s pelvic tissue.
He added that shrinkage of the mesh after implantation, caused by fibrosis, also leads to such complications.465
399 Professor Klinge deposed that when tension is placed on the mesh, it curls and ropes causing increased scarring between the fibres and its frayed, un-bordered edges shed particles of polypropylene before, during, and after surgery. He said that the release of particles into the surrounding tissue leads to an increased inflammatory response, erosion, chronic pelvic pain, implant failure, chronic sexual dysfunction and dyspareunia, organ damage, urinary dysfunction, inability to remove the device and the need for surgical intervention.466
400 He said that the Prolene mesh used in the SUI devices was over-engineered in that it was many times stronger than necessary for its intended purpose and that it leaves far more polymer material behind than is necessary. He expressed the view that “any pelvic mesh designed with this much excess surface area and weight unreasonably and unnecessarily increases the risk of injury to the patient …”.467
401 He regarded as unreasonable Ethicon’s failure to make critical design changes to each of its SUI devices before launch. He considered that the use of Prolene in the SUI devices unnecessarily compromised patient safety, leading to such complications as chronic inflammation, excessive scarring through and around the mesh, nerve entrapment, chronic pain, dyspareunia, erosion, recurrence, and the need for corrective surgery.468
402 Professor Klinge expressed similar opinions with respect to Prolene Soft (Gynemesh PS) used in the various POP devices. He said that “[i]t is impossible to establish reliable parameters for the design of a device for use in POP repair surgery” without understanding the biomechanics of the pelvic floor. Yet he said that Ethicon had “a very poor understanding” and apparently still does. In this respect, he was supported by Professor Deprest who had written that the mechanical requirements for implants in pelvic surgery had not been defined by 2006 and the mechanisms causing local complications in urogynaecology were poorly understood, then and still.469
403 Professor Klinge deposed that the POP devices were not specifically designed to function in the pelvic floor; they were over-engineered (that is, they were much stronger than necessary); they will create “an intensified and chronic [foreign body reaction]”; they will have pores that are too small to resist fibrotic bridging and the formation of scar plates; and they will curl, rope and fray causing particle loss and sharp edges. All these “design failures”, he added, “cause an unnecessary risk of patient complications and injuries”, including (amongst other things) chronic pain, dyspareunia, erosion, exposure, infection, and the recurrence of the condition which the implantation of the device was designed to treat.470 He said that it was unreasonable of Ethicon not “to properly study and/or make the necessary design changes” to avoid the “known and serious complication” of mesh shrinkage and the other safety hazards already mentioned. This opinion, he said, applied to all the POP devices that used Prolene Soft mesh.471
404 Professor Klosterhalfen’s opinions were to the same effect.
405 On the face of things, having regard to the extensive work undertaken in this area by both Professors Klinge and Klosterhalfen and their close collaboration with Ethicon over a substantial period of time, their opinions were compelling. On the other hand, the logical consequence of the opinions they expressed, at least on one construction, was that every woman in whom these devices was implanted would suffer from the pleaded complications at least if she lived long enough. While it is possible that this is so, the evidence does not establish that every woman with an Ethicon device will more probably than not experience all or even any of those complications. Nonetheless, the evidence does establish on the balance of probabilities that she could.
406 Since the evidence of Professors Klosterhalfen and Klinge was so powerful, before going any further it is necessary to address the respondents’ attempt to undercut its force.
407 It is common ground that, until the hearing in this case, the respondents never questioned the reliability of the information derived from the pelvis data pool or the information provided by Professor Klosterhalfen in his reports to Ethicon about his findings.472 Yet, drawing upon answers derived from their cross-examination of the Professor about data recorded in some spreadsheets relating to the data pool, the respondents submitted that it was “plain” that there was “substantial doubt as to the accuracy and reliability of the theories developed by Professors Klosterhalfen and Klinge”. The reason the respondents gave was that “they are reliant on the data from the Pelvis Data Pool” and, to the extent that the data could be verified, “significant errors (acknowledged by Professor Klosterhalfen) have been shown”.473 They made this submission, I should add, while at the same time acknowledging that the “theories” and work of the two men draw on both the hernia and the pelvis data pool.474 In particular, the respondents submitted that the Court should be cautious about coming to conclusions solely on the basis of the data pool, particularly where those conclusions conflict with other evidence. They urged the Court to be mindful of its “inherent limitations”.475
408 I will come to the matters upon which the respondents rely shortly but one point should be made at the outset. That is this. No conclusion I reach could ever be based solely on the data pool. If the respondents mean that I should be cautious about accepting opinions reached solely on the basis of the data pool, that will depend on the extent to which the particular shortcomings raised by the respondents affect the opinions of the witnesses that were derived from the data pool.
409 I now turn to the respondents’ complaints.
410 First, the respondents pointed to errors in an Excel spreadsheet relating to the pelvis data pool.476 The spreadsheet, which was cited in Professor Klosterhalfen’s second affidavit, contained four worksheets, entitled “Raw data”, “Final I”, “Final II statistics” and “Final PP (J&J)”.477 The respondents submitted that there were errors and inconsistencies in the first three worksheets. They also put weight on the fact that these worksheets were prepared by a student of Professor Klosterhalfen and not the Professor himself. It should be noted, that the student who collected the “raw data” was a medical graduate, although not yet a pathologist.478
411 In cross-examination Professor Klosterhalfen explained that tissue samples were sent by clinicians to his centre together with background information. It was then analysed by the student.479 The student’s analysis, together with the additional information from the clinician, were recorded in the first worksheet (“Raw Data”). For each sample, information was recorded under the following headings: infection, FBR (foreign body reaction), fibrosis, polymer, mono/multi, other (which recorded information about observable adverse events such as erosion and shrinkage), age (of the patient), and mesh class (usually by reference to the weight and size of the pores). Additional information was recorded without any headings, most notably about the identity of the devices (such as TVT-O or Prolift), the condition for which the device was implanted, and adverse events reported by the clinician (such as pain).
412 Professor Klosterhalfen described the compilation of the information in the first worksheet as follows:
She [the student] was looking for – or responsible for this data pool. She made the first analysis. She made the macroscopy and she collected all the slides and tried to get out – to collect the clinical information, and that’s why the spreadsheet is called raw data. These are uncontrolled data. It’s her view what she was collecting there.480
413 The second worksheet (“Final I”) replicated some of the headings and added new ones. Importantly, there was a category entitled “mesh class”, which categorised the device as either “1” (for large pore mesh) or “2” (for small pore mesh).481 The characterisation was made by the student.482 Professor Klosterhalfen accepted that any errors in the first worksheet would “infect” the second worksheet, and because of the errors present, one would not rely on the mesh classification in the second sheet.483 The same was the case for the third worksheet.484
414 Professor Klosterhalfen was cross-examined about some of these errors.485 He was taken to a couple of entries, for example, in which Vypro II was described as a heavyweight small pore mesh, when it is actually a lightweight large pore mesh (before implantation at least).486 Professor Klosterhalfen said that the error could have been in the identification of the mesh rather than the description of its characteristics or vice versa.487 The respondents claimed that there were numerous occasions where an explant was incorrectly described although they did not take the Court to those occasions. But these were the student’s errors, not the Professor’s.
415 Importantly, Professor Klosterhalfen testified that he worked with the student on the fourth and final worksheet (“Final PP (J&J)”) and double-checked the data.488 As I understood the evidence, this was the data upon which Professor Klosterhalfen relied, not the data in the earlier worksheets.489 He explained that the data pools are “a scientific work tool”. It was only after the data were checked with the clinicians and the slides were checked, that he accessed the data pool.490
416 Nevertheless, because the original analysis was undertaken by the student, the respondents submitted that there was room for error in both the identification of the type of mesh explanted as well as its characterisation (as a large or small pore mesh). Yet, as to errors in identifying the type of mesh, Professor Klosterhalfen explained that clinical data including the type of implant was checked two or three times with the treating clinician before being included in the data pool and any mesh that could not be confirmed with the treating clinician was not included in the pool.491 As to any errors in the original characterisation by the student while collecting the raw data, in cross-examination Professor Klosterhalfen readily acknowledged that errors would arise at that point: “You will find a lot of that, most probably” and “You can go through all 500; you will find mistakes”.492 While a couple of column headings in the final worksheet were misleading or inapt (“PP” instead of “small or large pore” and “J&J” when not all the explants were Ethicon explants), Professor Klosterhalfen did not accept that errors in earlier worksheets made their way into the final worksheets.493 Indeed, he insisted that the data in those spreadsheets were accurate:
MR FINCH: …You will accept this part of the spreadsheet, that the student one is full of errors of allocation, both as to – possibly as to the nomination of what product it was and certainly as to the equation of the products with the size pores?---Yes. But the next step is that you control that and that you have to get in contact with the doctors, that you have to get the history of the patient, and the - - -
Okay?--- - - - how do you say, the protocols of the operation. And then finally you will find these correct data in the last spreadsheets.494
417 Still, the respondents argued that, because the doctor who implanted the mesh was very often not the doctor who explanted it, there was room for error on that account in the identification of the mesh. While Professor Klosterhalfen acknowledged, in effect, that there was room for error on the part of the clinician who provided the explant, the evidence indicates that any such error would not find its way into the final spreadsheets. He explained:
See, basically what you have to know, you see: you get these examples and the clinical colleague is writing what he believes what have been implanted. And it’s very often that doctor who has implanted this mesh is not the doctor who explants the mesh finally, especially after eight years. And all this clinical data, you have to check it twice or third.495
You see, what is relevant here, if you ask me what is the definite data, the last spreadsheet, for instance. So – and basically, as I told you before, for me these data pools are a scientific work tool. And when all these data are checked with the clinicians and the slides are checked, I go into this data pool and basically if you look at the last spreadsheet you see only one and two, one and two, yes, no, yes no. So a statistical analysis is possible. That’s what we – what is interesting for me.496
418 Professor Klosterhalfen said that the complication rates he mentioned in his presentations, including his presentation in Istanbul on the “Pathology of Meshes in the Pelvic Floor: Actual Data” given on 6 September 2015 (the Istanbul presentation),497 were drawn from the last of the four worksheets (or a version of it).498 He stated that he had personally checked the data he used for the presentation. Where, for example, there was a discrepancy between what the student had recorded and the information provided by the clinician, and he was able to do so, he double-checked the information including, it appears, by examining the sample microscopically.499 I accept that he did check the data. His credit was never impugned and it was not suggested that this evidence was unreliable.
419 Second, although the respondents accepted that the fourth and final worksheet was double-checked by Professor Klosterhalfen, they complained that they had “no way to test its contents” because they had not had access to the underlying data used to prepare the spreadsheet. In a footnote to their closing submissions they asserted that a notice to produce “the documents or things which Professor Klosterhalfen created, considered or relied upon in relation to his examination of the Pelvis Data Pool” (to which the applicants had taken objection) was served on the applicants on 3 August 2017.500
420 The respondents argued that the applicants’ failure to produce the data was important in a context where there were differences in the conclusions reached by Professor Klosterhalfen and the clinicians who provided the original data. They pointed to instances where Professor Klosterhalfen said he could not identify an erosion from an explant or slide although the clinician who provided it had recorded one. They submitted that this necessarily casts doubt upon the reliability of the data provided by the clinicians (or, arguably, on Professor Klosterhalfen) which in turn casts doubt upon the reliability of the data pool, particularly when the record of the reasons for explanation, such as chronic pain, dyspareunia, discomfort/pain, recurrence, migration of mesh, and fistula/abscesses, was derived from calls to the clinicians, calls made by the student and not verified by the Professor through a second call.501
421 Quite apart from the inappropriateness of serving the notice to produce on the applicants when the material to which it referred would obviously not have been in their possession or control, the notice to produce was not in evidence and was never called upon. In those circumstances, the complaint about the failure to produce the underlying data goes nowhere. In the first place, in the absence of a call on the notice there was no failure to produce the documents. In any case, the evidence was that not all notes were kept and it was not suggested that there was anything untoward about that. Although the slides were retained, Professor Klosterhalfen observed that they were the patients’ property and could only be released with their permission.502 This observation was not challenged.
422 Finally, to the extent that the respondents claimed that they were deprived of an opportunity to verify the information, that complaint is difficult to take seriously when they had a longstanding contractual relationship with Professor Klosterhalfen. It will be recalled that it was Ethicon (through Dr Meier) suggested the type of information to be collected and understood the way it would be collected and could at any time have asked to review the Professor’s data and methods, but never did so.503
423 The applicants rightly took issue with the characterisation of Professor Klosterhalfen’s concessions as an acknowledgement of “significant errors” in the pelvis data pool. As they argued, the respondents’ submission was not accompanied by a transcript reference and no errors were acknowledged or disclosed in the final worksheet, which was the one upon which the Professor relied.504
424 As to the admitted errors, I fail to see how they matter — even if they can be said to be “significant”. There will always be errors in working documents. It is the final document that counts.
425 Third, the respondents argued that there were inherent limitations in the conclusions that can be reached on the basis of the pelvic data pool because the explants necessarily represent cases with “a symptomatic complication”, which is why they were explanted, and so the complication rates drawn from the data pool represent a percentage of that cohort (with a symptomatic complication) rather than a broader, representative population of patients including those with no symptoms. The respondents submitted that the data pool does not provide evidence of “the pathological processes associated with a successful (or asymptomatic) outcome”. They argued that “the Data Pools comprise solely of explants, which represent bad cases (or they would not have been explanted)”.505
426 It is not strictly true to say that the data pools are comprised solely of explants representing “bad cases”. This was an assumption. The same assumption was made by Professor Deprest.506 It was an understandable assumption, but it was wrong. Professor Klosterhalfen pointed out in his second affidavit that the data pool does have some internal controls. 507 He explained:
It is true that the structure of a database sometimes lacks appropriate controls. Therefore human mesh samples or biopsies of meshes without any complication have to be included. In my Hernia Data Pool, mesh samples explanted in the context of other surgical procedures (e.g. large bowel cancer, prostate cancer or bladder cancer) serve as an internal control. In total, my Hernia Data Pool includes 112 mesh biopsies and 33 full explants, serving as controls.
427 Professor Klosterhalfen’s opinion about the significance of nerve entrapment in explants was influenced by the comparison with the controls. He said that in the hernia data pool it was rare to find entrapped nerves in mesh explanted other than for pain but that they could often be seen in cases where mesh was explanted for chronic pain.508
428 Nevertheless, I accept that the complication rates mentioned in the Istanbul presentation are percentages of complications revealed in the explants provided to Professor Klosterhalfen and are not representative of the complication rates in all patients implanted with polypropylene mesh. I did not understand Professor Klosterhalfen to have represented that they were.
429 Fourth, the respondents contended that the results Professor Klosterhalfen recorded in the Istanbul presentation cast doubt on his own thesis. That is because slide 44 records an erosion rate of 64.5% for large porous meshes (on average 24.7 months after implantation) and 40.8% for small porous meshes (on average 39.4 months after implantation).509 The respondents argued that the data is either inconsistent with the Professor’s “theory” or the data is wrong because of error in the collection or recording of it.
430 I reject the argument. I accept the applicants’ submission that the data is not wrong and the figures are consistent with Professor Klosterhalfen’s theory of effective porosity or, at least, not inconsistent with it. “Effective porosity”, discussed in more detail below, is the percentage of the area of mesh that is filled only by sufficiently large (effective) pores (to allow for the ingrowth of healthy tissue)510 and is concerned with the size of the pores after implantation and under load.511 The concept was first referred to publicly in a 2007 article by the Aachen Group: “New objective measurement to characterize the porosity of textile implants” (Mühl et al (2007)512. As a result of the studies carried out by the Aachen Group, it was concluded that to allow for the ingrowth of healthy tissues, the pores in the mesh need to be greater than 1000 microns (micrometres or μm and the equivalent of 1mm) in all directions.513 In his testimony in this case, Professor Klinge explained that, if you apply tension to a large pore mesh (UItraPro was the example he gave), small pores can be created, removing the advantages of the large pore mesh and, in effect, transforming it into a small pore mesh.514
431 Fifth, the respondents complained that only some of the complications were independently checked by Professor Klosterhalfen and, when they were checked, there were differences between the Professor’s views and those of the clinicians who provided the initial data.515
432 This was at best a simplistic representation of the effect of the evidence.
433 The evidence was that, if the student had recorded an erosion when the clinician had not, Professor Klosterhalfen, himself, checked the slide or slides.516 He said that you can see an erosion in the microscope. He added, however, that sometimes the clinician did not record an erosion when it was clearly visible in the slides. He said that the same was true for infection. As he explained later in his evidence, certain cells are markers of infection. They are the polymorphonuclear granulocytes (PMNs) and he had seen those cells in explants.517 On the other hand, he was reliant on the clinicians for information as to other complications like pain, migration, and recurrence. It is inconceivable that Ethicon would not have been aware of this at the time Dr Meier asked Professor Klosterhalfen to collect the information.
434 The mere fact that a student collected and checked the information from the clinicians does not make the information unreliable. Given the size of the data pools, I am not persuaded that any such errors are likely to have had any significant effect or that they detract from the force of the opinions expressed by Professors Klosterhalfen and Klinge, even if, contrary to my understanding of the evidence, one or more of them did creep into the final iterations of the spreadsheets.
435 Sixth, the respondents submitted that Professor Klosterhalfen’s evidence should not be accepted because the “theories” he and Professor Klinge propounded were inconsistent with Professor Wright’s experience and “the rigorous testing [by Ethicon] of the meshes used in the Implants”.518
436 This submission must be rejected.
437 For the reasons given above, Professor Wright’s experience is not comparable to the experience of Professors Klinge and Klosterhalfen. In the circumstances, any difference of opinion between Professor Wright on the one hand and Professors Klosterhalfen and Klinge on the other cannot be taken as a valid criticism of the data pool.
438 Moreover, the evidence does not establish that Ethicon subjected the meshes used in the Ethicon devices, particularly the pelvic meshes, to “rigorous testing” — far from it in fact.
439 In any case, the conclusions Professor Klosterhalfen reached and the opinions he formed, like those of Professor Klinge, were not based solely on the data in the data pool, let alone the pelvis data pool. Much of their thinking was informed by their animal studies, their observations of the tissue in which the mesh was embedded or encased, their tests, their extensive reading and research, and the studies of hernia explants.
440 The concept of bridging fibrosis, for example, was first described in reports and articles published before the pelvis data pool was set up.519 It was described, for example, in the report of a study of the foreign body reaction to meshes used for the repair of abdominal wall hernias by Professors Klinge and Klosterhalfen and others, published in the European Journal of Surgery in 1999: Klinge et al (1999).520
441 It was in the early 2000s that they recognised the connection between bridging fibrosis and pore size. On 6 March 2002 the results were published from the Aachen Group’s study of the impact of pore sizes in rats.521 That study compared Marlex (made by the UK firm Bard) and Vypro (made by Ethicon). In the body of the article the authors, who included Professors Klinge and Klosterhalfen, wrote:
[T]he monofilament mesh with its smaller pores almost exclusively is embedded into granulomas and scar tissue, which bridges the whole pore diameter of less than 1 mm. In the case of small pores the extent of the activated foreign body reaction does not permit the ingrowth or persistence of local tissue. A large distance between the single filaments not only guarantees the maintenance of elasticity but will hamper the bridging of the inflammation across the pore.
442 Six months later, in September 2002, the results of another rat study by the group comparing Prolene hernia explants to explants of another material, Vypro II (made of polypropylene and polyglactin multifilaments), found that the Prolene mesh was entirely encapsulated in a scar plate. The authors, who again included Professors Klinge and Klosterhalfen, wrote that the data confirmed that a reduction in the total amount of implanted material and an increased pore size have a favourable impact on integration of the mesh with fewer undesirable side effects. They concluded on the basis of the results, that for proper incorporation of healthy tissue pore sizes should be “at least 1000 µm”, in contrast to the then prevailing wisdom that 75 to 100µm was sufficient.522
443 An article published the following month on mesh implants in hernia repair reported on a rat study comparing the inflammatory cell response to Prolene (according to the abstract, Marlex according to the body of the article), Mersilene, and Vypro after 7 and 90 days of implantation. The authors, all from Aachen University and who included Professors Klinge and Klosterhalfen, wrote that the findings underlined “a more pronounced inflammatory reaction and cell turnover around heavy-weight and small-pored [polypropylene] mesh implants”, consistent with other investigations describing “an accented foreign body reaction and fibrosis for heavy-weight polypropylene meshes”.523
444 Professor Klosterhalfen stated at various points in his evidence that he had consistently observed in histopathology of explanted meshes “contraction of tissue” and “shrinkage”, folding and curling of mesh along with dense scarification, chronic inflammatory responses, and nerve entrapment.524 He also noted that those “in vivo conditions” had been widely reported in the published literature.525 He said that the opinion he expressed that those pathological findings were consistent with, and directly related to, the pain and contraction experienced by women who had received the implants was based on his examination of over 600 samples of histopathology as well as his own testing, research, review of medical and scientific literature and work he had done with experts across the world.526 He was not challenged about this in cross-examination.
445 A number of the opinions Professor Klosterhalfen said were informed by the pelvic data pool were not seriously disputed. They include his opinions about the foreign body reaction to polypropylene mesh implants, which I consider next, the extent to which the mesh shrinks or contracts in vivo, the occurrence of infection with erosion, and the possibility of late infections.
446 More particularly, the respondents did not point to any specific opinion given by either Professors Klosterhalfen or Klinge that was founded on, or undermined or weakened by, the errors in the worksheets. Indeed, the respondents made no attempt to explain the connection, if there was one, between the errors and the witnesses’ opinions. Professor Klosterhalfen’s evidence was that he relied on the final spreadsheet and there was no evidence of error in that document.
447 Furthermore, it was not put to Professor Klosterhalfen that the errors in the worksheets undermined his opinions and Professor Klinge was not even cross-examined on the subject. Since it was not put to either witness, the submission that errors made by the student cast doubt on the accuracy or reliability of their opinions ought not to have been made: Browne v Dunn (1893) 6 R 67 (HL).
448 Finally, it will be recalled that the two data pools included over 15,000 explants. The vast majority of human explants were hernia explants. In other words the hernia data pool was much larger than the pelvic data pool. No attack was made on the hernia data pool. No questions were asked of either Professor Klosterhalfen or Professor Klinge about that data pool. It was not suggested that the conclusions derived from the data collected from hernia explants or explants from animals were unreliable. Yet it is abundantly clear that, to the extent that the opinions of the professors were based on what they had learned from their studies of explants, those opinions were derived from the entirety of the information in both data pools.
449 For all these reasons, I reject the submission that there is substantial doubt about the reliability of the theories developed by Professors Klosterhalfen and Klinge because of “significant errors” in three of the four worksheets.
450 The pore of a textile is the area or space between filaments.527 As the applicants submitted, none of the respondents’ witnesses seriously questioned the importance of pore size to the biocompatibility of the devices and the relationship between pore size and the inflammatory response to polypropylene mesh.528 The respondents accepted that pore size is a critical component of the biocompatibility of any mesh used in the body.529
451 By 1998, when Professor Klosterhalfen first published on the subject, it was known that there were higher complication rates when small pore mesh was used.530 It was well accepted, for example, that large pore meshes had significantly lower infection rates than small pore meshes.531
452 The debate in this case turned on the appropriate pore size, including the classification of pore sizes, and the concept of bridging fibrosis.
453 The specific dispute concerns whether the classification of biomaterials by Professor Parviz Amid in his seminal article, “Classification of biomaterials and their related complications in abdominal wall hernia surgery”, published in 1997 in the journal Hernia, has any continuing relevance, particularly in relation to pelvic meshes.532 The respondents maintained that it has and that it should be applied, doubtless because of what Professor Amid said about Prolene. The answer to the question is that the Amid classification is of limited utility, certainly in relation to meshes in the pelvic floor, although many of the observations made by Professor Amid in that article remain important.
454 In his 1997 article Professor Amid classified the hernia meshes then on the market according to pore size. He also discussed the importance of pore size, molecular permeation, and shrinkage of biomaterials, the role of biomaterials in increasing the risk of infection, the formation of seromas and fistulas, and the failure of repair due to “shrinkage” of the mesh. Professor Amid said that “[p]revention of biomaterial-related complications requires in-depth knowledge and understanding of the physical properties of prostheses, of which, the porosity and the pore size of the materials are of paramount importance”. None of this was controversial.
455 Professor Amid divided meshes into four types. Type 1 included meshes with pores greater than 75μm (0.075mm). He referred to them as “totally macroporous” prostheses. Type II included meshes with pores less than 10μm (0.01mm). He described type III as “[m]acroporous prosthesis with multifilamentous or microporous components”. Type IV were films with submicronic pores of less than 1μm. The difficulty with Professor Amid’s classification, at least for the purpose of classifying pelvic meshes, is that it assumes that the pore size does not change after implantation. The evidence before the Court shows that this assumption was incorrect.
456 Professor Amid classified both Marlex and Prolene as type I or macroporous (large pore) meshes. He wrote:
These prostheses contain pores larger than 75 microns, which is the required pore size for admission of macrophages, fibroblasts (fibroplasia), blood vessels (angiogenesis) and collagen fibers into the pores ...533
457 Professor Amid explained that surgical infection, promoted by implantation of biomaterials, is caused by infiltration and proliferation of bacteria into and within the pores and interstices of these synthetic materials. When interstices or pores are less than 10μm in each of their three dimensions, he wrote, bacteria averaging 1μm cannot be eliminated by macrophages and neutrophilic granulocytes, which are too big to enter the space. Consequently, small pore prostheses harbour bacteria and can promote their growth resulting in “biomaterials-related infection”.
458 The TVM Group emphasised the problem about the interstices. According to the English translation, they wrote:534
The concept of interstice must be clearly understood of the risk of infection associated with endovaginal surgery (sic).
Porosity is also a major mesh characteristic. Porosity will promote colonization of the prosthetic tissue “like a lattice by ivy”.
Fibroblasts, collagen fibers, and neo-vessels will be incorporated if they can circulate easily between the stitches, provided the pores are large enough.
459 Until the Aachen Group’s research, however, it seems that little, if any, thought was given to the possibility that the size of the pores would be affected by the forces operating in the body after implantation.
460 Professor Klosterhalfen explained what should have been obvious from the outset:
[M]eshes are knitted and therefore have numerous large and small pores. It is not enough to simply look at the largest pores. Rather, it is necessary to look at the distribution of pores (porosity), the effective pore size under load and after tissue integration, and the pore geometry (the smallest diameter of the pore determines the risk of bridging fibrosis). All these factors need to be taken into account when looking at the pore size of the [SUI devices] and the [POP devices]. Experimental studies have shown that the effective pore size of [polypropylene] meshes is less than 1mm …535
461 While Professor Amid’s classification is still used by many hernia surgeons to assess the risk of infection, Professors Klosterhalfen and Klinge considered it was not appropriate for the assessment of tissue integration in mesh.536 Professor Klosterhalfen pointed out that it only takes into account meshes produced and used for hernia repair before 1998 and it fails to take into account what happens to the mesh after implantation and under load. 537 Moreover, as Professor Klinge observed, Professor Amid’s focus was on the risk of bacterial infection and not on biocompatibility, the foreign body reaction, or the risk for excessive scar formation.538 These matters were the focus of subsequent research, including but by no means confined to that conducted by the Aachen Group. 539
462 In 2011, based on an analysis of 1,000 explanted meshes, Professors Klinge and Klosterhalfen modified the classification of surgical meshes for hernia repair. They came up with six different classes. Under this modified system, described in an article in Hernia published in May 2012, they placed UltraPro in the large pore class (characterised by a textile porosity of greater than 60% or an effective porosity greater than 0%) and Prolene and Marlex as small pore meshes (characterised by a textile porosity of less than 60% and without any effective porosity).540
463 This classification covers particular mesh modifications, including mesh for laparoscopic and intraperitoneal incisional hernia repair, which were not covered by Professor Amid’s classification. As Professor Klosterhalfen explained in evidence, it was based on increased knowledge about the significance of pore size, pore geometry, and pore stability under mechanical load.541
464 Despite the increase in knowledge over the last two decades and their acknowledgment that pore sizes change in vivo, the respondents maintained that the Amid classification “remains applicable for classifying meshes”, including meshes used in the treatment of stress urinary incontinence and pelvic organ prolapse. They also contended that a minimum pore size of 75μmis sufficient to avoid the formation of a scar plate. They offered four reasons: first, there are examples of scientific literature that positively refer to the Amid classification; second, the majority of the scientific literature refers to the classification without comment; third, there are significant limitations to the reliability of the data from Professor Klosterhalfen’s data pools; and fourth, the effective porosity “theory” of Professors Klosterhalfen and Klinge is inconsistent with Professor Wright’s experience.542
465 I have already addressed the question of the reliability of Professor Klosterhalfen’s data. I set out below my response to the other three reasons.
466 Dealing with the last point first, it is difficult to understand what the respondents meant by Professor Wright’s “experience” in this context. While he disagreed with a statement in an Ethicon document that fibrotic bridging “is usually observed in all modifications with pore sizes of less than 1 millimetre”, he conceded he had undertaken no research of his own. He said he disagreed because “the limit” was based on an extensive study of the literature, a study conducted for the purpose of US and Australian litigation. He acknowledged that (in contrast) Professors Klosterhalfen and Klinge had spent most of their professional lives conducting independent research on this very topic.543
467 In any case, as I have already said, Professor Wright’s experience is inconsequential in comparison with the experience of Professors Klosterhalfen and Klinge.
468 Second, the fact that a number of scientific papers have uncritically adopted the Amid classification does not prove its continuing utility, particularly in the area of pelvic floor meshes. To enable any weight to be placed on the reliance by the authors of the papers on the use of the classification, it would be necessary to know whether they were aware of the research of the Aachen Group and had considered the matters that Professors Klosterhalfen and Klinge raised in their evidence in this case, including the important point they make that pore sizes change in vivo. I was taken to no such evidence.
469 Third, Professor Klosterhalfen gave unchallenged evidence that the validity of the Amid classification has never been proven in a clinical study.544
470 Fourth, the Amid classification was devised for synthetic mesh used in abdominal wall hernia surgery. Consequently, it did not take into account the different environment of, or the forces at work in, the female pelvic floor.
471 Fifth, Barone et al (2015) observed that the Amid classification did not consider the spaces created due to the methods of construction of the mesh. They noted that knit construction, for example, “creates small voids around the location at which filaments are joined”, which “provide microporous elements even in large pore meshes”. 545
472 Sixth, since the Amid classification was devised, there has been an increase in knowledge about pore geometry and the effect of mechanical loading on the pores. There is nothing to indicate, for example, that Professor Amid was cognisant of the fact that pore sizes change in vivo or, if he was, that he took it into account.
473 The research conducted by Barone et al demonstrated what happens to the pores under mechanical load. They cited with evident approval the work of Professors Klinge and Klosterhalfen, but not just their work. Of particular relevance in the present context is the following passage:
Notably, it has been shown that effective tissue in-growth with polypropylene mesh, characterized by the quality of the tissue which forms around mesh fibers, only occurs in mesh pores with a diameter of 1000μm or greater (Klinge and Klosterhalfen, 2011). Pore sizes less than 1000μm have greatly enhanced inflammatory and fibrotic responses (Weyhe et al., 2006; Bellon et al., 2002). While a pore size of 1000μm appears to be the threshold for polypropylene, the thickness of fibrous encapsulation is expected to vary depending on the polymer used as that (sic) the degree of fibrous connective tissue deposition is believed to be dependent on protein interaction with the fiber surface and related to hydrophobicity of the polymer (Klinge et al., 2002d). It is important to note that these findings on pore size were all determined using an abdominal wall model. Though the general foreign body response to mesh should consist of similar mechanisms in the vagina and adjacent supportive tissues, there are distinct differences in the biology of these sites, which likely impacts the host response to SUI, and POP meshes. Therefore, the critical pore diameter for urogynecological meshes to minimize scar plate formation may be distinct from that found in abdominal wall studies.
474 Seventh, much earlier the following observations were made at the IUGA Grafts Roundtable and published online in an article written by Professor Deprest and others on 6 May 2006 (Davila, 2006): 546
Currently existing graft classification systems do not apply to the use of grafts in the pelvis. The Amid classification published in 1997 is frequently quoted as the accepted classification system for synthetic grafts (see Dwyer and Deprest contributions). As only Type l mesh is recommended for use in the pelvis, this classification does not, in general, apply to the use of synthetic grafts in the pelvis. There are currently multiple subtypes of Type l mesh, with markedly variable physical characteristics, including softness, weave, elasticity, and pore size, among others. There are, thus, great and clinically significant differences among Type l synthetic mesh materials. To improve clarity when referring to Type l mesh, other physical characteristics of the mesh should be described. Clear examples of these characteristics include directionality and distortion with the stretch, which can play a significant role in the outcomes of a suburethral sling procedure.
475 Eighth, by 2008, if not earlier, Ethicon itself recognised the shortcomings of the Amid classification.
476 Terri Dunn of Ethicon Women’s Health & Urology, for example, wrote in an email dated 24 June 2008:
The AMID chart is truly out of date in relationship (sic) to what pelvic floor mesh needs to be in pore size.547
477 In October 2008 Dr Trezwik and Dr Meier, both Ethicon employees, submitted to Johnson & Johnson Medical GmbH an “invention disclosure” for a vaginal implant for pelvic floor reconstruction adapted to the biomechanical environment of the pelvic floor. In their description of the benefits of the invention. They wrote:
A reduced mesh pore size (< 1mm) is identified as a major cause of “bridging fibrosis” causing reduced tissue compliance in the area of the mesh implants. The temporary stress shielding of the mesh during the period of implantation circumvent to problem of collapsing mesh pores under load during implantation.548
478 I discuss “bridging fibrosis” below. “Stress shielding” has been well described by the Moalli group. It occurs when a stiffer material (here the mesh) bears the bulk of the load and buffers or shields surrounding tissues from these forces: Barone et al (2015).549 The less stiff material (the soft tissue), now shielded from the load it normally carries, undergoes “a maladaptive remodelling response characterized by degeneration and atrophy (loss of collagen, elastin, and skeletal/smooth muscle)”.550
479 A 2010 PowerPoint presentation by Dr Boris Batke, then Ethicon’s Assistant Director of Research and Development, entitled “Mesh Innovation for Pelvic Floor Repair”, classified a small pore size polypropylene mesh as one which was less than 1mm.551 Dr Batke contrasted mesh of this kind with partially absorbable lightweight meshes. He treated only the latter as macroporous.552 Incidentally, he stated that “[c]linical data have shown more chronic pain, testicular pain and foreign body sensation” with microporous meshes than with partially absorbable macroporous meshes. The note below this slide (slide 10) reads:
Today we have clinical evidence that the mesh construction, the amount of foreign material, the pore size and filament thickness do play a role in post operative pain, and plenty of studies on animal models, explanted meshes etc can explain this by looking at how the body’s immune system reacts to the mesh implant.
480 Three references were given. None was to the work of the Aachen Group. That is not to say that Dr Batke was not influenced by the work of the group. Elsewhere in his presentation he referred to the article by Klosterhalfen et al “The lightweight and large porous mesh concept for hernia repair”.553 But it suggests that his conclusions were informed by other research and independent thought.
481 Dr Batke also asserted that there was no place for the microporous (which he referred to as “[h]eavyweight”) mesh in modern pelvic organ prolapse repair — in other words, there was no place for mesh with pores of less than 1mm.554
482 On 27 June 2011, in a response to a request from a urogynaecologist for some images — “[e]ven the old slides that have the old AMID (sic) classification” — for a presentation he was giving, Brian Luscombe, Ethicon’s US Product Director, Female Pelvic Medicine & Reconstructive Surgery, observed:
As I am sure you know, the world you knew 7-10 years ago on AMID classification is no longer even relevant to some extent. What people think they know about mesh is changing. … In the next year or two it is going to change further with more discussion on topics like partially absorbable mesh, bridging fibrosis, and stress shielding.555
483 Although he had steadfastly refused to accept that it was out of date, when confronted with the Luscombe email in cross-examination, Professor Wright agreed that the Amid classification was no longer relevant “to some extent” “in terms of biocompatibility and biomaterials science”.556
484 Four months earlier, in February 2011, a confidential memorandum under the joint letterheads of Ethicon Women’s Health & Urology and JJM contained a graphic illustration of how pore sizes are affected by load. It also stated:
For mesh strips, even a mild loading of 1N [one newton] has a dramatic impact on the resulting pores size as far as soft meshes like Ultrapro are concerned. Under this loading the initial pore size of between 3–4 mm decreases to values down to 0,3mm.557
485 I interpolate that it was common ground that 1N is the amount of force necessary to keep an apple weighing about 100g from falling.558 One acquires a good idea of just how mild that load is when one sees from a letter Ethicon wrote to the FDA in February 2008 that the force required to pull an arm of a Prolift+M implant through a cannula is 0.73lbf (pounds of force), which equates to 3.24N, and that 12lbf (53.38N) is the figure given for the forces applied to Prolift during insertion.559
486 The February 2011 confidential memorandum went on to say that “[t]he phenomena (sic) of pore size variation as a function of the applied load is related to kinematical effects based on textile mesh design” and “is closely related to the angle geometry of the pore forming filaments”.560
487 The memorandum also contained the following candid admission:
The development of knowledge to understand the mechanics of pelvic floor disorders is imperative; yet, we are only just beginning to determine the necessary criteria on which to base design for pelvic floor implants.
488 The author of this document is not certain. Professor Klinge said that it was written by Jürgen Trzewik, an Ethicon senior scientist, and Christophe Vaihe, principal scientist. This may well have been an assumption, albeit a reasonable one. But no-one suggested otherwise. The document was unsigned but their names appeared on the first page under the heading “Approvals”. Although the document was unsigned, there is no reason why it should not be given weight. Certainly the respondents did not proffer one.
489 Similarly, the effect of applying a mechanical load to polypropylene mesh was graphically illustrated in a slide featuring in a PowerPoint presentation on the importance of mesh properties prepared by Dr Meier, then Ethicon’s Principal Scientist R&D:561
490 The text appearing under the images reads:
Figure 1 illustrate[s] the pore size dependency of Ultrapro for two load cases. Microscopic view on individual pores at zero tension state 0 N (left) and under uni-axle [(scil) Axial] load of 1 N (middle). The right image shows the macroscopic view [of] an Ultrapro strip elongated at a uniaxial (scil) load of 1N.
491 Similarly, in another PowerPoint presentation, Dr Holste, then Ethicon’s Senior Research Fellow, deviated from the Amin classifications. He referred to a mesh of less than 800μm (0.8mm) in diameter (that is, a mesh about 10 times larger than the 75+ μm in Amid’s Type 1 classification) as a small pore mesh, stating that fibrosis is more pronounced in meshes with pores of these dimensions “forming scar plates”.562
492 The unchallenged and uncontradicted evidence of Professor Klosterhalfen is that, after the publication of the Aachen Group’s findings on bridging fibrosis in pores less than 1mm, which followed numerous experiments with experimental polypropylene mesh modifications with different and defined pore sizes, “all companies modified their [polypropylene] mesh portfolio adding [polypropylene] meshes with pores greater than 1mm” (original emphasis).563
493 There was a dispute in this case about the size of the pores in the Ethicon devices
494 Even before implantation the pore sizes of the mesh used in each of the relevant Ethicon devices are not uniform. According to a table in the Gynemesh PS technical file, Prolene mesh (used in all the SUI tapes) has a maximum pore size of 0.37mm2 and a minimum of 0.24mm2. Its average porosity is said to be 53.1%. Prolene Soft mesh, the mesh used in Gynemesh PS and all the other POP devices, is said to have a maximum pore size of 2.38mm2 and a minimum pore size of 0.29mm2. Its average porosity is recorded as 65.6%.564 Pore sizes for Prolift+M (based on figures for UltraPro) range from 3–5mm.565
495 The figures given in the Gynemesh PS technical file or figures like it appear in other Ethicon internal documents as well, including documents prepared by members of its research and development team. For example, in his 2010 PowerPoint presentation Dr Batke said that the pore size of Gynemesh PS was 2.44mm2 and the pore size of UltraPro was 4mm.566 Presumably these were references to the largest pore sizes. Dr Batke did not specifically refer to the other Ethicon meshes.
496 The applicants relied on the figures in the Gynemesh PS technical file. The respondents, on the other hand, relied on measurements in a report by an organisation called LNE, dated 28 November 2008, whose figures were related by Dr Hinoul in his affidavit. LNE had been asked by Ethicon to evaluate the properties of TVT and Prolift. LNE said that it randomly defined 10 zones of fabric and in each zone measured the height and width of five pores randomly selected.567 It reported a mean width pore size of 1.06mm (with a standard deviation of 0.05mm) and a mean height pore size of 1.55mm for one SUI tape and, for one POP device, a mean width pore size of 2.34mm (with a standard deviation of 0.07mm) and a mean height of 1.39mm (and a standard deviation of 0.08mm).568
497 Professor Klosterhalfen challenged these figures. He said that TVT has a number of different size pores and these figures measure the length and width of the largest pores. Professor Klinge said, with a mean pore size of 1.06mm, a figure he also disputed, at least 50% of the pores would be smaller than 1mm.569
498 The respondents did not account for the differences between the LNE figures and those they had recorded themselves. In any case, the decision to provide mean figures obscures the full extent of the differences between sizes of the pores and masks the problem of the interstices and the small voids around the location at which filaments are joined.
499 That is apparent from an Excel spreadsheet dated 24 October 2013 prepared by Daniel Burkley, Ethicon’s principal scientist, which documented pore sizes for Prolene Soft mesh based on two image files taken on 26 March 2006 and 10 July 2010. Three sets of measurements were taken from each image and also averaged. Pore sizes ranged from 0.04mm² to 3.14mm². In total, 54 measurements were taken. As Professor Klosterhalfen pointed out, the vast majority of the pores were well under 1mm. Eighteen were below 0.28mm, the minimum figure given in the Gynemesh PS file. Dr Burkley indicated that some variation in pore size from lot to lot was expected.570
500 As I have already observed, regardless of the true range of pore sizes in the pristine product, it was common ground that pore sizes will reduce after implantation. Professor Klosterhalfen’s evidence was that the pores can collapse altogether. This was how he explained it.
501 Once mesh is implanted in the body, it moves and will deform with bodily movement. Mechanical stresses placed on the mesh affect pore size. In the pelvic floor there are numerous sources of mechanical stress. Even defaecation and urination, for example, place tension on the mesh.571 If the textile structure is not stable, mechanical load can cause pores to collapse. After implantation, mesh fibres become embodied in the granuloma and this process significantly reduces the pore size of the pristine mesh. Consequently, the size of the pores before implantation is not the size of the pores after implantation. Whether the mesh is sufficiently porous depends on the effective size of the pores. Effective porosity is what you get after implantation and growth of fibrotic tissue.572 Effective pores and high porosity limit fibrotic bridging, encapsulation and contraction of mesh. To be effective, it is necessary that the pore size be stable and maintained consistently under load or stress.573 The collapse of pores under mechanical load results in pore sizes of less than 1mm (1,000µm) with subsequent fibrotic bridging and mesh encapsulation.574
502 The problem of collapsing pores under load was accepted by Ethicon. It was one of the reasons behind Project Thunder, discussed below. It was identified by Drs Trezwik and Meier in their invention disclosure.575
503 Moalli et al (2008) noted that one of the primary problems of using TVT is that the mesh easily deforms when tensioning under the urethra. They conducted mechanical testing on six commonly used midurethral polypropylene slings, including two made from Prolene: TVT and TVT-O.576 Moalli et al developed a cyclical testing protocol to measure permanent elongation resulting from repetitive loading within a range of forces relevant for in vivo loading: cough, sneeze, straining, and moving from a sitting to a standing position. Cyclical loading induced permanent elongation in all six slings. After 10 cycles of between 0.5 and 5N (about 0.1 and 1.1lbs), the authors reported that the “Gynecare samples had permanently elongated by 17.5±4.2%, indicating that although very little force applied, there is irreversible deformation of the TVT™”. After 10 cycles between 0.5 and 15N (about 0.1 and 3.4lbs), Gynecare samples permanently elongated 42.3±3.2%. “Importantly”, the authors observed, “elongation after cyclical loading was not reversible, indicating that once a sling deforms, it does not revert to its initial form”. They noted that their results were similar to those obtained for the Gynecare in a previous study published in 2003.577
504 The following images of TVT taken from the article, clearly illustrate the collapse of pores under load. The image marked “a” shows the Ethicon sling before loading. The other images show the sling at different stages of loading. The authors commented that, in image “d”, “the sling has elongated more than 70% of its original length and has lost its structural integrity such that it resembles a string”.
505 Moalli et al (2008) identified their “most important finding” in the following paragraph:
The most important finding of the paper is that Gynecare TVT™ mesh has a unique tensile behavior which is characterized by an initial region of very low stiffness in which the mesh easily elongates in response to small changes in force. This is followed by a transition period (inflection point) and an area of high stiffness. As a result of this behavior, after cyclical loading at low loads (0.5 to 5 N or 0.1 to 1.1 lbs), Gynecare mesh permanently elongated by more than 10% of its initial length, confirming the easy permanent deformability of this mesh that is observed clinically during placement. The mesh with mechanical behavior most similar to Gynecare was that supplied by AMS [Amercian Medical Systems].
506 They indicated that the behavior of these slings in vivo and after incorporation into host tissue could be inferred from their study findings but that rigorous in-vivo studies were the next logical next step for determining the relationship between textile and tensile properties of polypropylene slings, tissue behavior, efficacy, patient morbidity, and patient satisfaction.
507 In his evidence in the present case, Professor Klosterhalfen confirmed that Prolene in general (by which he meant both Prolene and Prolene Soft) has no effective pores because, under mechanical load, its pores effectively disappear.578 The collapse of the pores, he added, is likely to occur during or shortly after implantation.579
508 Professor Klinge gave evidence to similar effect. He pointed to tests of Prolene and Prolene Soft under load that he had conducted in 2014 with Professor Thomas Mühl in the context of US litigation. Those tests showed that under 4.9N/cm of strain (which Professor Klinge described as “minimal force”) the textile porosity of Prolene decreased to 50% if the strain was applied perpendicular to the warp course and to 57% if the strain was applied in line with the warp fibres and that any large pores completely disappeared if the structure was stressed at right angles to the warp fibres. The figures were different for the three Prolift arms. Textile porosity of arms 1 and 2 at rest was 58 to 58.9% and decreased to 10% for arm 3. Effective porosity disappeared, however, at a strain of 4.9N/cm in arms 1 and 2 and at a lower strain of 2.5N/cm in arm 3.580
509 While the textile porosity of Prolene Soft was 63% in comparison to 51% for Prolene in the Klinge/Mühl study, the effective porosity of Prolene Soft was 26% and, in the arms, decreased rapidly under strain, dropping to 0% at a strain of 4.9N/cm.581
510 Those tests also showed that under minimal strain the Prolift arms curled (or roped) under minimal load (5.2538N/cm) and did not return to their original shape. Professor Klinge said that, as strips of mesh begin to curl, the pores reduce in size as the fibres “become situated too close together”, enhancing the chronic inflammatory response, and leading to bridging fibrosis.582
511 Professor Deprest expressed similar opinions. He wrote that after implantation, all meshes are “loaded by physiologic forces”, which may cause “tremendous geometrical changes” at the pore level. He added that, if the implant is deformed by these “loading forces” and does not return to its initial state, “the integration process into the body may be compromised as well”.583
512 Furthermore, as Professor Klosterhalfen observed, Professor Deprest’s opinion was supported by numerous studies, which he cited in his report. He went on to refer in particular to the conclusions of Barone et al in their article, “Textile properties of synthetic prolapse mesh in response to uniaxial loading”, published in the American Journal of Obstetrics & Gynecology in September 2016, which Professor Deprest also cited.
513 This was a report on a study conducted by Dr William Barone and his fellow authors from the University of Pittsburgh, Professor Pamela Moalli, and Dr Steven Abramowitch, who had participated in the study of the slings reported in Moalli et al (2008). This later study was of four synthetic meshes with distinct pore geometries to determine the effect of tensile loading and pore orientation on mesh porosity and pore dimensions. One of the meshes was Gynemesh PS. Barone et al (2016) wrote:
In this study, it was shown that the pore size of current prolapse meshes dramatically decreases in response to mechanical loading. These findings suggest that prolapse meshes, which are more likely to experience tensile forces in vivo relative to hernia repair meshes, have pores that are unfavorable for tissue integration after surgical tensioning and/or loading in urogynecologic surgeries. Such decreases in pore geometry support the hypothesis that regional increases in the concentration of mesh leads to an enhanced local foreign body response. Although pore deformation in transvaginal meshes requires further characterization, the findings presented here provide a mechanical understanding that can be used to recognize potential areas of concern for complex mesh geometries. Understanding mesh mechanics in response to surgical and in vivo loading conditions may provide improved design criteria for mesh and a refinement of surgical techniques, ultimately leading to better patient outcomes.584
515 A major finding of the study was that the porosity of nearly all tested products approached 0% in response to just 10N of applied force, which is within the expected physiologic range and consistent with the forces they said they would have expected meshes to experience before tissue incorporation during implantation and in vivo.
516 Professor Wright agreed that the pore size of a mesh before implantation might not be the pore size of the mesh after implantation. He assigned two possible causes for the reduction in pore size. One was the possibility of stretching during implantation. The other was that shortly after implantation “chronic inflammation will surround the mesh fibers”, causing the mesh to shrink somewhat. Despite this, however, Professor Wright said that the pore sizes in vivo were large enough in all Ethicon meshes used in gynaecological surgery to allow good incorporation of the mesh into the host tissue. He said that, once fibrous ingrowth occurs, he could not envisage a potential mechanism by which the various loads and stresses placed on the mesh would affect its pore size without also tearing the fibrous connective tissue around the mesh fibres.585 But as Professor Klinge pointed out, Professor Wright did not consider the physiological forces placed on the mesh after implantation, nor did he consider the remodelling of soft tissue as a mechanism for pore deformation that also occurs after implantation.586Professor Klosterhalfen responded to these opinions of Professor Wright at – of his second affidavit. He wrote:
120 Dr. Wright has not understood the concept of “effective pores”. The concept of effective porosity includes that no fibrotic tissue is formed by the FBR and that the granuloma does not fill out the pores, ideally the whole mesh indicates only a perifilamental fibrosis. If the fibrotic reaction leaves no space in the pore and the mesh is encapsulated by a fibrotic layer on the top, the reaction is called bridging fibrosis …
121 Prolene in general has no effective pores, the minimum size required for effective pores in PP meshes is greater than 1mm. PS [Prolene Soft] has effective pores (26%), however under mechanical load the number drops to 0% ... Therefore, Dr. Wright’s statement that “the size of the pore in vivo remains large enough with all of the Ethicon meshes used in gynaecological surgery to allow good incorporation of the mesh into the host tissue” is incorrect.
122 The collapse of the pores of PS is likely to occur during or shortly after the implantation due to the mechanical load. Later, after mesh integration, mechanical loads lead to an increased cell turnover in the mesh interface to compensate for mechanical stresses. Cell turnover due to mechanical loads or stress again is the basis of tissue remodelling; for example, the correction of tooth position by braces. The same is true for meshes under mechanical load. The phenomenon of a chronic inflammatory reaction with consecutive and persisting cell turnover and tissue remodelling explains changes of the mesh configuration or shape after implantation and explains late complications like erosion or chronic pain.
123 Contrary to Dr. Wright’s opinion, studies indicate that tissue integration after mesh implantation has no “endpoint”, as tissue remodelling can be observed even years after implantation.587
517 Once implanted, Professor Klosterhalfen said, to avoid fibrotic bridging and permit appropriate ingrowth of fatty tissue or loose connective tissue (rather than dense scar tissue) pore sizes need to be greater than 1mm between two polypropylene filaments.588 Professor Klinge’s evidence was to the same effect.589
518 Since Professor Amid did not consider the impact of mechanical load on mesh after implantation or the risk of excessive scar formation, let alone what pore sizes might be necessary in mesh used in the repair of the female pelvic floor, I am persuaded that his classification is not useful for present purposes. I am also persuaded that under mechanical load the pores of polypropylene mesh, including Prolene and Prolene Soft, will reduce in size and can, indeed, collapse. It is unnecessary to decide, as the respondents invited me to do, whether a pore size of more than 1mm is required in order to minimise complications. Having said that, given that I accept the evidence of Professors Klosterhalfen and Klinge. It follows that I accept their evidence about the problems of meshes with smaller pore sizes.
519 Professor Klinge’s evidence was that the failure to account for the phenomenon of pore size variations as a function of the applied load was an unacceptable design failure of Prolift.590 He argued that a knitted surgical mesh device that is permanently implanted in human tissue must be designed in such a way that its pores do not collapse and deform when subjected to the expected forces both during implantation and after implantation. Yet, the tests he and Professor Mühl performed on TVT, the results of which were replicated by Professor Moalli’s group, showed that under minimal strain its pores deform and collapse. “Permanent deformation and pore collapse”, he continued, “leads to fibrotic bridging, scar plate formation, excessive scarring through and around the mesh and a host of tissue complications that can lead to chronic pain, recurrence, erosions, dyspareunia and need for reoperation, to name a few, making it unsafe for its intended purpose of being permanently implanted in a woman’s pelvic tissue”.591
520 Professor Klosterhalfen said that, if the implant is integrated into the body it is covered by fibrotic (scar) tissue, which may encapsulate the mesh in part or in whole. Whether it is completely encapsulated depends on the structure of the mesh. Foils and microporous meshes, he said, are always encapsulated in that the fibrotic reaction forms a complete capsule around the mesh. Professor Klosterhalfen said that this phenomenon is called “bridging fibrosis”.592 Yet, Professor Klinge made it clear that there is a difference between encapsulation and fibrotic bridging. He clarified the position at  of his affidavit:
The encapsulation of a mesh implant is not the same as fibrotic bridging. Pure encapsulation without ingrowth into the mesh refers to the formation of a capsule that surrounds the entire mesh. In the case of encapsulation, the entire mesh can usually be surgically removed from the capsule without any problems as the mesh is not incorporated into the tissue. This type of encapsulation occurs with some meshes, for example, GoreTex mesh. Conversely, “bridging” means that the pores are filled by scar tissue, and that the scar bridges the distance between two filaments, thereby surrounding and encapsulating the entire mesh. This type of encapsulation is different from pure encapsulation as the formation of scar within the mesh makes it very difficult to extract the mesh from the tissue.593
521 Later in his affidavit at , Professor Klinge said that fibrotic bridging occurs when the fibrotic capsule of the foreign body granulomas side by side, crosses the pore, filling out the entire distance between fibres by scar tissue forming a rigid “scar/mesh compound” or “scar plate”, enclosing the mesh and leaving no room for further tissue ingrowth. He said that the formation of a scar plate leads to a number of complications including loss of elasticity and pain associated with the rigidity, shrinkage or contraction of the mesh, mesh erosion, nerve entrapment, bacterial encasement, chronic pain and dyspareunia (in the case of pelvic mesh).
522 These opinions were based on the numerous studies conducted by members of the Aachen Group. As the applicants submitted, they were also supported by the findings of Dr William Cobb and others published in 2005.594
523 Professor Moalli’s group have also reported on this phenomenon. They wrote that “[f]ibrous encapsulation and its potential contraction by resident myofibroblasts may induce pain after mesh implantation”: Barone et al (2016).595
524 In Mühl et al (2008) stated that bridging will occur with polypropylene filaments of less than 600–800 microns in diameter and will form a dense scar plate around the entire mesh.596
525 At least two of the respondents’ witnesses were familiar with the concept of bridging fibrosis.
526 Professor Deprest described bridging fibrosis in his report. He said he had observed that smaller pore materials with closer filaments experience bridging fibrosis “where the fibrous tissue around on (sic) filament comes in contact with the fibrous tissue around the next one”. He illustrated his point with copies of histologic slides from his own research.597
527 Although he did not discuss the matter in his evidence in chief, in cross-examination Professor Santerre agreed that a permanent mesh implant is surrounded with a fibrotic capsule and (for hernia meshes at least) if the fibres are too close, the fibrotic response to neighbouring fibres overlaps or bridges, resulting in bridging fibrosis or encapsulation of the mesh.598 He also agreed that bridging fibrosis leads to the formation of a continuous scar plate and prevents tissue from growing into the mesh structure.599
528 On the other hand, Professor Wright ridiculed the concept of bridging fibrosis. In cross-examination he referred to it as a “silly term”. He insinuated that the concept was recognised by only a small minority of people.600 He claimed that no pathologist in the United States would diagnose bridging fibrosis, although he never identified a reason or basis for that claim.601 Nevertheless, he said he knew of no article which criticised the term.602 Moreover, in cross-examination he appeared to accept the existence of the concept. He agreed that chronic inflammation adjacent to the mesh filaments can lead to increased fibrosis adjacent to the filaments and that, depending on the size of the pores, the fibrosis can fill the whole of the pores or the spaces between the individual filaments. He agreed that it can surround the mesh and can change its surface area. He also agreed that some meshes can become encapsulated.603
529 Despite Professor Wright’s contempt for the term, the respondents did not deny that bridging fibrosis can occur.604 Rather, for the reasons they gave in support of the continuing relevance of the Amid classification, the respondents argued that a minimum pore size of 75µm was enough to avoid it.
530 I have already discussed the problems with the respondents’ argument. It is sufficient to record in this context that Professor Amid did not deal with the question of the sufficiency of the pore size to avoid bridging fibrosis. Indeed, he did not address bridging fibrosis. Nor did many of the articles upon which the respondents relied in defence of the 75µm threshold. To the extent that they commented at all on the rationale, they were concerned with the pore sizes necessary to prevent the growth of bacteria and the risk of infection. While the concepts are related, they are not the same.
531 I also reject any suggestion that bridging fibrosis has not been documented to occur with the Ethicon devices. Professor Klinge said that clinical studies of explanted meshes as well as all animal experiments showed that Prolene mesh bridges, regardless of the site of implantation, and that Prolene Soft (Gynemesh PS) and UltraPro (used in Prolift+M) show bridging when explanted from areas of tension as in the pelvic floor.605 In a report of a study of local tissue reactions to a number of non-absorbable meshes used in abdominal wall hernia repairs (including Marlex, Prolene and Vypro) published in 1999, to which Ethicon provided financial support,606 Professor Klinge and his fellow researchers reported that both the polyester and polypropylene meshes showed pronounced chronic inflammation and “a strong interlinking formation of connective tissue [scar tissue] through the mesh-pores”. The authors went on to say that the embedded connective tissue forms a rigid scar plate, which is responsible for substantial mesh shrinkage,607 a subject to which I will come shortly. They wrote:
All the polypropylene meshes [Marlex, Prolene, and Atrium] (implantation time 3-24 months) had similar histological patterns after explantation.
Commonly, there was a predominant foreign body reaction with typical foreign body granulomas including epithelioid cells and giant cells. However, and contrary to ePTFE, polyester, and the reduced polypropylene mesh there was persistent acute inflammation with varying amounts of CD15-positive PMN and focal fibrinoid necrosis in most cases. The inflammatory process was accompanied by pronounced perifilamentous fibrosis with an extensive amount of deposited collagen fibres… Adjacent to the mesh the fibres were mainly orientated parallel to the polypropylene threads. In the periphery, connective tissue with numerous collagen fibres formed a thick capsule in which the whole mesh was integrated. These mesh modifications were characterised by complete penetration of connective tissue into the pores.
532 Further, in the 2007 Ethicon swine study all three meshes in the subcutaneous model (Prolene, Prolene Soft, and UltraPro) were reported as exhibiting “thicker connective tissue that ‘bridged’ between mesh nodes”. The report also noted that “[c]onnective tissue bridging between nodes has often been observed in meshes experiencing contraction”.608
533 I find that bridging fibrosis has been documented to occur with Prolene, Prolene Soft, and UltraPro., which means that it can occur following implantation of all the Ethicon devices.
534 As to whether bridging fibrosis is of clinical significance, the respondents’ position is inconsistent with the position taken by its own scientists. For example, in a 2001 Ethicon report of a rat study of the tissue reaction over 21 days to various polypropylene meshes, including Prolene, Prolene Soft, and various versions of Vypro, the authors (Dr Richard Hutchinson, Principal Scientist, Corporate Product Characterization and Dr Thomas Barbolt, a pathologist and Senior Research Fellow, Corporate Product Characterization) wrote that:
Integration of the mesh by surrounding connective tissue plays a major role in the ability of the mesh to confer supporting strength to the clinical repair, i.e., the weak link is the interface between the device and living tissue. Thus, it is critical that the mesh allows for this tissue ingrowth and of sufficient magnitude to provide this important function. However, it has been recognized that excessive formation of scar tissue (fibrosis), and subsequent contraction as it matures, can result in discomfort for the patient.609
535 In a white paper on early clinical experience with Gynemesh PS produced in May 2004 (said to reflect the clinical experience of a number of surgeons, including Dr Lucente and Dr Kirkemo),610 Ethicon stated that the formation of rigid scar plates around the mesh could cause deformation of the repair site, bunching of the mesh, and decrease mobility of the repaired tissue.611
536 In a PowerPoint presentation prepared in 2007 by Kerstin Spychaj, a senior scientist in Ethicon’s Research and Development team,612 which reviewed both the literature and internal studies, small porous meshes (defined as less than 1mm) were said to lead to fibrotic bridging which, in turn, resulted in increased shrinkage.613
537 I accept this evidence. I find that bridging fibrosis is, indeed, of clinical significance.
538 There is a wealth of evidence that contraction, often called shrinkage and “retraction” by the French, is a complication of the use of polypropylene meshes. Almost invariably this phenomenon is referred to in the literature as either “mesh contraction” or “mesh shrinkage”. As both Professor Klinge for the applicants and Professor Deprest for the respondents explained, however, the mesh does not contract or shrink as a pullover might in the washing machine. What happens is that there is a reduction in the surface area originally covered by the mesh brought about by the retraction of the fibrotic tissues around the mesh. Contraction of scar tissue is a physiological reaction of maturing scar characterised by a constant water loss.614 The greater the area of scarring, the greater the degree of contraction.615
539 Professor Klinge said that, as a hernia surgeon he had removed “a lot of contracted [polypropylene] meshes from patients”, which exhibited “a huge amount of scar tissue”. He has published over 30 peer-reviewed publications on the subject of mesh contraction including the contraction of polypropylene meshes.616
540 The applicants’ case is that contraction leads to folding, stiffening, and loss of tissue function,617 extrusion, erosion, and infection, and can cause other complications such as chronic pain, dyspareunia, and a recurrence of the disorder for which the device was implanted.618 Professor Chughtai said that mesh contraction can lead to vaginal pain, vaginal shortening and/or tightening which, in turn, can cause severe side-effects, most commonly pelvic pain and dyspareunia.619 Professor Klinge explained that:
Contraction not only leads to poor coverage leading to recurrence, but will also increase the amount of [foreign body reaction] due to pore collapse. This phenomenon then leads to additional complications, including pain, dyspareunia, nerve entrapment, increased inflammation, urinary and faecal incontinence, urinary retention, blood vessel injury, chronic pelvic pain, erosions, an inability to remove the device and the need for painful and, at times, dangerous revision surgery.
541 Dr Donald Ostergard wrote in 2010 that:
Because the vagina is a tubular structure, a decreased calibre and shortening are the inevitable results [of mesh shrinkage]. Dyspareunia can be explained by such mesh shrinkage, as well as by tension on mesh arms with neuroma formation. Because the mesh is anchored in tissue, its shrinkage will put more and more tension on the anchoring tissue, with resulting pain. No mesh seems to be immune from this process … Unfortunately, this fibrous tissue will continue to contract regardless of what the surgeon trying to remove the mesh is able to do. The more surgery, the more scar tissue that will form.620
542 I referred above to the study findings of the Aachen Group on mesh contraction or shrinkage. Fatton et al (2007) also noted shrinkage of mesh in 18 out of 110 patients (17%) three months after implantation of Prolift in a retrospective multicentric study.621
543 In his affidavit Dr Hinoul asserted that the applicants’ experts were incorrect to say that mesh contraction was a cause of dyspareunia and expressed the view that mesh does not contract but that during wound healing tissues may contract, regardless of whether mesh is present. He added that, since studies indicate no difference in total vaginal length or caliber with Gynemesh PS and Prolift in comparison to native tissue repair, if contraction does occur, it is no different than that which is seen with native tissue.622
544 These statements were not only at odds with the overwhelming weight of the evidence, they were also disingenuous. I shall deal first with the issue of whether tissue contraction has an effect on the size of the mesh.
545 In 1997 Professor Amid reported findings of mesh shrinkage of 20% in hernia explants 10 months after implantation.623
546 In 1998 Professor Klinge and his colleagues reported on the shrinkage of polypropylene mesh in vivo in an experimental animal study comparing Marlex to Vypro. They found that polypropylene meshes (Marlex) shrank to about 30 to 50% of their original size within two to six months of implantation.624
547 In a PowerPoint presentation on Gynecare’s pelvic floor strategy produced in November 2004, Zenobia Walji noted:
We have experienced 20–40% of contraction using mesh. This requires a larger piece to cover the prolapse and using a “tension-free approach” – Dr. Mauro Cervigni (Italy), TVM Group (France)625
548 The TVM Group reported 80 cases of mesh shrinkage in its study of 684 patients (a rate of 11.7%) at seven centres in France during the period November 2002 to December 2004. Eighteen of 110 Prolift patients (16.9%) studied over a shorter period (March to October 2005) were found to have mesh shrinkage. 626
549 A prospective study of 125 Prolift patients operated on at the University Hospital in Clermont-Ferrand in France between March 2005 and August 2006 revealed a rate of “[p]ainful mesh shrinkage” of 19.6% at 18 months follow-up.627
550 In 2005 Cobb et al reported that all meshes, regardless of their composition undergo a 20 to 50 percent reduction in their initial size.628 The following year saw the publication of their findings from an animal study on shrinkage in three different meshes which showed no significant difference in the degree of contraction in the three meshes tested: UltraPro, Prolene Soft and Marlex. UltraPro shrank by 29%, Prolene Soft by 33% and Marlex by 28%.629
551 A 2007 study of mesh contraction and tissue integration in swine models over a 13 week period conducted by Ethicon Inc.,630 which Dr Hinoul neglected to mention, showed that Prolene Soft (the mesh used in Gynemesh PS, Prolift, and Prosima) contracted by 27 to 35% and UltraPro (the mesh used in Prolift+M) by about 14%. Furthermore, gross observations of the mesh showed folding, rippling, and distortion of the meshes, greater at the subcutaneous site than in the pre-peritoneum site, with Prolene Soft and UltraPro exhibiting the greater incidence of folding, rippling, and distortion in comparison with Prolene.631 Importantly, the Ethicon swine study also revealed that all meshes were stiffer than native tissue and that Prolene mesh was stiffer than the other meshes.632
552 In a paper on the subject of “mesh shrinkage” prepared for the 2008 IUGA annual meeting in Taipei, Professor Bernard Jacquetin and Dr Luka Velemir noted that mesh retraction occurs during the scarring and remodelling process and is related to the extent of the tissue inflammation around the mesh after implantation which induces the wound contraction.633
553 On 18 January 2011 Professor Klosterhalfen told consultants engaged by Ethicon and Michel Richter of Ethicon at a three-hour meeting that:634
At the high level, there are two classes of “shrinkage” observed with mesh implants (Note: the term “shrinkage” is a misnomer. Tissue reaction over time encapsulates the mesh with connective tissue and effectively ‘crushes’ the mesh into a ball (like crushing a sheet of paper); the mesh does not truly shrink):
– The first is in the immediate short term following implant; the implant is observed to lift and may ‘roll up’ from its position. This occurs as a result of poor positioning, placement and/or suturing of the implant by the clinician.
– The second class of shrinkage is the formation of scar tissue; observed in the longer term (months) following implantation. This scar tissue can reduce and compact, causing the mesh to crumple up.635
554 The minutes of that meeting also record a note “that all implanted meshes will show surface area reduction with time and this was up to 50% with older, small pore meshes”.
555 Professor Klinge deposed that Professor Klosterhalfen merely restated what was widely known in the scientific community and the manufacturing industry in the 1990s, namely, that older, heavyweight small pore meshes like Prolene used in Ethicon’s SUI devices increase the risk of shrinkage or contraction of the mesh by up to 50% of its area.636
556 Professor Wright’s evidence was that it was impossible to observe contraction or folding.637 I reject this evidence. Professor Klinge’s experience, which is far greater than Professor Wright’s, proves otherwise. Professor Klinge deposed that shrinkage is apparent when you measure the distance between two filaments of a pristine mesh and compare them to the distance between filaments after implantation, shrinkage. He said that folding (where there are two or more doubled layers of mesh) is also clearly seen in explants and is not affected by the orientation in vivo.638 Furthermore, in 2008 Hinoul et al reported that shrinkage of the mesh had been observed in a patient six months after surgery.639 In cross-examination, Dr Hinoul said that when the mesh-tissue construct shrinks, the mesh folds.640 He said that when he carries his procedures he always tries to avoid getting folds in the mesh by flattening it out.641
557 In cross-examination of the applicants’ experts, it was not suggested that shrinkage did not occur. Rather, it was put to Professor Klinge that more intense shrinkage with heavyweight small poor mesh was not established by 2005. Without hesitation, he repudiated the notion.642
558 Professor Wright admitted in cross-examination that the meshes used in all the Ethicon devices could “show shrinkage”, that shrinkage can reduce the size of the pores, and that the number of pores within a mesh that are greater than 1mm might decrease as a result of shrinkage and fill with “fibrous connective [scar] tissue”.643 As a matter of logic, it seems to me, this would increase the chances of infection and pain.
559 Professor Deprest said that it was initially thought that contraction was an active process resulting from contracting fibrotic scar. I interpolate that in 2007 Klinge et al had written that “[w]ound contraction of the fibrous integration leads to shrinkage of the implant area …”.644 More recently, however, Professor Deprest said that it has been shown to occur very early after implantation. He explained that what is thought to be happening is that, due to load on the implant when a woman awakes or starts to move around, the implant undergoes deformation and does not return to its original state. His opinion was that both processes are likely to be in play.645 That opinion is supported by the work of Professor Moalli’s group which identified two potential mechanisms: mechanical loading and fibroblast-induced contraction as part of the foreign body response. Barone et al wrote that both these mechanisms are likely to be related to the geometry of the mesh, the loading environment in which the mesh is placed, and the characteristics of the mesh such as pore size.646
560 Professor Klosterhalfen said that mesh shrinkage is closely related to “bridging” or “bridging fibrosis”. He explained that in all mesh implants with a granuloma size around each mesh fibre of more than half the pore size of the mesh the granuloma of one fibre starts to blend with the granulomas formed on neighbouring fibres increasing the area of granuloma. He continued:
As a result, the outer fibrotic layer of each granuloma around each polypropylene fibre form an outer fibrotic capsule, forming a fibrotic tissue layer in parallel to the surface to the mesh. Accordingly, the outer fibrotic tissue layer covers not only single fibres of the mesh, but larger areas of whole mesh. Finally, if this fibrotic capsule starts to retract or shrink, the mesh has to follow and adapt to the reduced surface area. As a result the mesh wrinkles and buckles, forming folds. Ordinarily, bridging occurs in small, porous meshes.647
561 Dr Hinoul’s statement that, if contraction occurs, it is no different for native tissue, makes no sense. As Professor Klinge explained:
In general contraction is an active process, the cause of which may be the changed configuration of the collagen molecules by dehydration. While the wound area gets smaller by growing cells and regeneration from the border, an area may be “reduced” but will not actively shrink. Correspondingly, tissue may grow over a mesh if it is visible in the subcutaneous space. In contrast, a wound made of collagen will develop shrinkage, as collagen contracts, and therefore it will push the mesh together with folds, and this happens in all parts of the body. A reduction in the area is a positive sign of healing if it is done by regeneration, whereas scar contraction often causes functional problems by increased stiffness, reduced stretchability, forming edges with local accumulation of material with its subsequent foreign body inflammation …648
562 Professor Klinge said that the greater the scarring the greater the degree of contraction. Furthermore, a mature scar without a foreign body usually does not show chronic inflammation.649
563 Notwithstanding the evidence given in chief by Dr Hinoul that suggested otherwise, as I have already indicated there was ultimately no issue that contraction occurs with the mesh used in all the Ethicon devices. The only dispute concerned its clinical significance. Despite conceding that there was a potential link between pain (including dyspareunia) and mesh contraction,650 Dr Hinoul maintained that mesh contraction had no clinical significance651 and the respondents argued that the applicants have not been able to establish that mesh deforms in the human body, with clinically significant effects.652
564 I reject the respondents’ argument. The evidence establishes that contraction does have clinical significance.
565 The respondents relied on an article about a study using a rabbit model published in 2009 which tested four types of polypropylene prostheses, all of which contracted.653 The severity of the shrinkage was unpredictable for any of them. For this reason, the authors wrote, “this could be a problem when prostheses are inserted into the vagina; [i]t is possible that the retraction produced pain and dyspareunia”. The respondents submitted that the authors’ failure to express a concluded view was “consistent with the uncertain clinical significance of contraction”.654
566 Since this was a study of hernia mesh and it involved a rabbit, the inconclusive nature of the authors’ remarks is neither here nor there. For a start, pain and dyspareunia are subjective. It is difficult to know how the rabbit would communicate her feelings to the researchers. Besides, none of the meshes were made by Ethicon.
567 The respondents also relied to a study of Prolift mesh by Svabik et al (2010)655 in which the authors found that, despite shrinkage of approximately 15%, the data did not allow for “any conclusions as to whether intraoperative folding or postoperative mesh shrinkage is of clinical significance”.656
568 This was a small study involving only 36 patients, comparing pre-implantation of the mesh with ultrasound appearances four days and three to five months after surgery. It was not designed to determine whether intraoperative folding or postoperative mesh shrinkage was of clinical significance. As the authors themselves stated, it is “only plausible to use ultrasound imaging for follow-up after mesh implantation to gain information about placement and extent of the mesh and, ultimately, to correlate such findings with clinical success and failure”.
569 The respondents also relied on evidence from Associate Professor Rosamilia that she did not see any evidence that contraction is a clinical problem for slings for incontinence procedures657 and Professor Wright’s evidence that he had not seen any pathological consequence of contracted meshes.658
570 This evidence was at odds, not only with the applicants’ evidence, but also with evidence adduced given by other witnesses called by the respondents.
571 Professor Roovers confirmed in cross-examination that mesh contraction can cause pain and dyspareunia.659
572 Professor Collinet, another of the respondents’ urogynaecological experts, agreed in cross-examination that mesh contraction can cause dyspareunia.660
573 Associate Professor Lam accepted that mesh contraction may lead to “painful scars”.661
574 Professor Deprest’s evidence on this subject was also instructive, although the respondents neglected to refer to it. It also supported the applicants’ case.
575 In his first report, Professor Deprest deposed that “[a] physically deformed implant may cause pain or other discomfort, both at rest as well as on effort or during intercourse”, even without exposure, and that the pain or other discomfort can be felt along the entire trajectory of the implant.662
576 Moreover, the respondents’ own records show that they accepted that contraction was clinically significant.
577 In the minutes of a meeting held on 4 April 2007 Dr Meier, from Ethicon R&D noted that the main customer unmet needs included “[l]ess mesh shrinkage and tissue contraction (to reduce vaginal stiffness and pain)”.663 There is no suggestion that anyone at the meeting questioned the connection between shrinkage and tissue contraction on the one hand and vaginal stiffness and pain on the other.
578 A document produced on 6 September 2007 by Ethicon R&D on the clinical strategy for Prolift+M described mesh shrinkage as a more serious complication than mesh exposure:
Mesh retraction (“shrinkage”) is less common but it is considered more serious. It can cause vaginal anatomic distortion, which may eventually have a negative impact on sexual function. Its treatment is difficult. Additionally, the scar plate that forms with in-growth of tissue into the mesh can cause stiffness of the vagina that further impacts sexual function in a negative manner.664
579 Indeed, the stated rationale for developing Prolift+M was to minimise these complications.665
580 In an article published in 2011, of which Dr Hinoul was the second-named author, describing the rationale for the development of Prolift+M, the following statements appear:
Among the most prevalent complications are mesh exposure and shrinkage of tissue around the mesh. These may result in pelvic pain and dyspareunia.
One of the key rationales for adopting a new, lighter-weight mesh with improved directional elastic properties was to minimize tissue shrinkage, which may lead to dyspareunia.666
581 These statements that shrinkage or contraction may result in pelvic pain and dyspareunia were not intended to signify that the connection was theoretical, only that shrinkage or contraction did not invariably have these consequences.
582 The respondents also argued that the first published study to report evidence of a correlation between contraction and complications was one by Rogowski et al published in 2013 (discussed below) and later studies were inconclusive.667
583 I reject this argument, too. There were studies reporting a correlation well before 2013 and later studies were not inconclusive.
584 In as early as 1981, mesh contraction was cited as one of two factors responsible for mesh extrusion in a paper by Randle Voyles et al (“Emergency Abdominal Wall Reconstruction with Polypropylene Mesh: Short-term Benefits Versus Long-term Complications”).668 That paper reported on surgery using Marlex.
585 An article by Clarke et al in the Journal of Surgical Research published in 1996, and referred to in Ethicon’s technical file on Prolene, reported on a study of 12 dogs into whose abdominal walls both Prolene mesh and xenogeneic (porcine) small intestine submucosa (SIS) was implanted. The Prolene mesh, in contrast to the control, was found to be wrinkled in all dogs. The authors thought that the wrinkling of polypropylene mesh after implantation could be caused by contraction of the original defect during wound healing and/or muscle relaxation from anaesthesia. They also considered that the wrinkling could contribute to complications such as seroma visceral erosion, skin graft erosion, and infection.669
586 In October 2010 an article on vaginal mesh contraction by two Australian surgeons, Drs Benjamin Feiner and Christopher Maher, was published in the journal Obstetrics and Gynecology.670 The article referred to “a number of patients who developed substantial morbidity related to mesh contraction”.
587 Feiner and Maher reported on a case series of 17 patients who had been implanted with transvaginal mesh: six with Prolift Total; four with Prolift Anterior; four with Perigee, a device made by American Medical Systems Inc. (AMS); the remaining three with Apogee-Perigee (both AMS products, analogues of Prolift Anterior and Posterior).671 The focus of the study was “the clinical expression of mesh contraction”. All 17 of the patients had vaginal mesh contraction and severe vaginal pain, aggravated by movement; all of those in the cohort who were sexually active (14) experienced severe dyspareunia. The authors reported that:
On a thorough vaginal examination localized areas of prominent tense mesh were noticed under the vaginal epithelium in all the patients. Palpation of the contracted mesh reproduced the pain these women experienced with movement and sexual intercourse. After primary surgical intervention to release the tension caused by the contracted mesh, 88% of patients had resolution or substantial reduction of the vaginal pain. All women had resolution of the pain if including the three who underwent further excision of the entire accessible mesh.
588 They concluded that vaginal mesh contraction was a serious complication after pelvic organ prolapse using polypropylene mesh, characterised by severe vaginal pain and dyspareunia and on vaginal examination focal tenderness over contracted portions of the mesh. They defined vaginal mesh contraction as an adverse outcome from prolapse repair with armed polypropylene mesh in women who experience vaginal pain with movement and dyspareunia and on examination have localised areas of prominent tense and tender mesh under the vaginal epithelium.
589 In cross-examination, Associate Professor Lam said that he had made similar findings in patients with painful mesh contraction and agreed with the authors’ definition of vaginal mesh contraction.672
590 Withagen et al wrote in 2011 that both “[m]esh exposure and shrinkage of fibrous tissues around the mesh may result in pelvic pain and dyspareunia”.673
591 In his 2012 review article, Dr Daniel Elliott of the Department of Urology, Female Urology and Voiding Dysfunction at the prestigious Mayo Clinic, wrote that “[w]hen vaginal mesh contracts, it causes vaginal fibrosis, infection, chronic vaginal pain, chronic pelvic pain, vaginal shortening, vaginal narrowing, vaginal extrusion, adjacent organ erosion, and dyspareunia”.674
592 In its 2015 opinion on the safety of urogynaecological meshes, SCENIHR noted that “excessive fibrosis may lead to mesh contraction resulting in increased pull on adjacent tissues leading to complications such as voiding dysfunction, pain and painful intercourse”.675
593 The respondents’ submissions were also inconsistent with their own documents, the opinions of the TVM Group, and the position of the FDA.
594 The minutes of a meeting of a scientific advisory panel on pelvic floor repair held in Chicago on 22 June 2001, apparently constituted by Ethicon and attended by a number of Ethicon employees, including Dr Martin Weisberg, Dr Brigitte Hellhammer, and Zeonbia Walji, record that contraction is a contributing factor to erosion and it is necessary to avoid contraction of the tissue in order to avoid erosion.676
595 The “main concern” of Professor Jacquetin and Dr Cosson, two members of the TVM Group, in 2004, according to an email sent on 10 May 2004 by Ophélie Berthier, Gynecare Marketing Manager France, to Zenobia Walji of Ethicon US, was “the shrinkage of the mesh which may lead to pain [and] dyspareunia”.677
596 In their November 2004 article, the TVM Group described mesh contraction and its after-effects as “much more worrying” than erosion and referred to dyspareunia as “the symptomatic manifestation” of contraction.678
597 In their 2008 paper on mesh shrinkage, Professor Jacquetin and Dr Velemir wrote that over time it appeared to the TVM Group that “mesh retraction or shrinkage (reduction of the mesh area and loss of compliance after tissue incorporation) was probably the most contributing factor to recurrences, postoperative pain and dyspareunia” (original emphasis).679Jacquetin and Velemir found shrinkage of between 15 and 25% of patients operated on between March 2005 and August 2006 and said that the retraction (contraction) was associated with mesh thickening on ultrasound.680 They provided advice on how to manage the complications. They emphasised that:
When a mesh procedure seems indicated, it is important to remember that severe mesh retraction may result in severe complications including dyspareunia, pain and recurrence; unfortunately, risk factors for these cannot always be identified.
They said that better understanding, assessment and prevention of mesh retraction at the time of augmented reconstructive pelvic surgery remained their “principal challenge” for the years to come.
598 This document was apparently discovered by Ethicon and cited by Professor Klinge in his affidavit, but none of the respondents witnesses referred to it.
599 Contrary to the respondents’ submission, largely based on Professor Wright’s evidence, that contraction occurs (only) within the first few weeks of implantation,681 the TVM Group described it as a “frequent late complication”.682
600 On 13 April 2005 Gene Kammerer sent an email to his colleague Dr Barbolt at Ethicon US, copied to numerous others in Ethicon’s R&D department, on the subject of “ULTRAPRO vs PROLENE Soft Mesh” in which he said (without alteration):
Regarding which attributes to investigate to show a difference between materials, I have this input. The issue which I am trying to investigate/solve is one of scar contracture around the mesh. In pelvic floor repair even with the PSM [presumably Prolene Soft Mesh], we have seen some scar contracture which translates into procedural complications. I don’t want to state % here because the situation which produces the complication is in itself complicated and specific to each patient. Also, most of the data comes from VOC [voice of the customer, that is, feedback from surgeons] and not our documented studies. However, it is important to know that the surgeons who are our consultants on the ProLift product are asking for a mesh which is better than PSM in this area.
The complications which are identified in the market are 1) recurrence of the prolapse 2) pain 3) stiffness 4) erosion and 5) discomfort during sex. The surgeons attribute these conditions to scar contracture. If we could find a way to reduce the scar formation by some % and subsequently the contracture it would give us a significant advantage over the competition as well as make the procedure better for the patient …683
601 Dr Linda Cardozo, a British urogynaecologist under whom Dr Hinoul trained and whom he told the Court he holds in high regard, informed Ethicon in June 2005 in an interview relating to mesh in prolapse repair that mesh folding would increase complications.684
602 Amongst the “highlights” noted from the discussion at an Ethicon expert meeting on 2 June 2006 attended by (amongst others) Professors Cosson, Klosterhalfen, Deprest, Jacquetin, Lucente, and, from Ethicon, Ms Berthier, Ms Hellhammer, Dr Meier and Dr Holste, were that “shrinkage is not controlled by ‘softness’ of mesh” and that the top priorities included the elimination of “shrinkage”.685 The position had not changed by 23 February 2007 when a similar meeting took place.686 It is inconceivable that elimination of shrinkage would be a top priority unless shrinkage was clinically significant.
603 The Prolift Surgeon’s Resource Monograph issued in 2007 contained the observation that:
Contraction of the mesh and/or reduction in the vaginal epithelial dimension is the primary exam finding in a subset of patients with dyspareunia.687
604 In its clinical strategy for Prolift+M, dated 6 September 2007, Ethicon stated that mesh contraction “can cause vaginal anatomic distortion, which may eventually have a negative impact on sexual function”, and is difficult to treat.688
605 The confidential memorandum bearing the joint letterheads of Ethicon Women’s Health & Urology and JJM, dated 16 February 2011, which I mentioned earlier, includes the following statement:
[T]here is evidence that meshes shrink in vivo leading to increased stiffness, pain, and poor restoration of the normal properties of the vagina compliance (Jones et al., 2009).689
The reference was to an article by Keisha Jones, Andrew Feola, Leslie Meyn, Steven Abramowitch, and Pamela Moalli entitled “Tensile properties of commonly used prolapse meshes” (2009) 20(7) Int Urogynecol J Pelvic Floor Dysfunct 847–853.
606 Christophe Vailhé, Mathew Krever, and Suzanne Landgrebe of Ethicon wrote in December 2011 that “[a] good graft design will require resistance to folding, or “balling” which could be due, amongst other things, to “excessive tissue contraction”, since it is believed to be a “significant factor in promoting exposure of the implant”.690
607 In 2008 the TVM Group reported that in their retrospective study of 684 patients there was “a significant link” between “prosthetic retractions” (contractions) and vaginal “expositions” (exposure) and granulomas. The Group also reported that retractions and relapse of prolapse were “significantly linked”.691 They postulated that retraction “favored” relapse because it creates non-reinforced zones692. In addition the Group wrote that prosthetic retractions can cause pelvic pain, dyspareunia, and dyschaesia (painful defaecation). Professor Collinet was a co-author of this article.
608 A Safety Communication issued by the FDA in 2011 relevantly stated:693
Reports in the literature associate mesh contraction with vaginal shortening, vaginal tightening and vaginal pain.
Both mesh erosion and mesh contraction may lead to severe pelvic pain, painful sexual intercourse or an inability to engage in sexual intercourse…
609 In 2013 a group of Polish researchers published the results of a study on possible correlations between mesh retraction (contraction) after anterior vaginal mesh repair with Prolift and de novo stress urinary incontinence, overactive bladder (OAB), and vaginal pain symptoms: Rogowski et al (2013).694 They found that mesh retraction was “significantly larger” in a subgroup of patients who reported postoperative vaginal pain compared with those who did not and that “[a] significant correlation was found between mesh retraction and the severity of vaginal pain”. They also found that mesh retraction was “significantly greater” in the subgroup reporting de novo vaginal pain at follow-up compared to those without de novo vaginal pain. And they said that mesh retraction was also “significantly larger” in patients with OAB than in patients who did not present with this complication at follow-up. Furthermore, the authors reported that the proportion of patients with postoperative vaginal pain and the proportion with postoperative OAB symptoms were in each case significantly higher in the high mesh retraction group.
610 Taking into account all the evidence on the subject, I am not only satisfied that polypropylene mesh, including those used in the Ethicon devices, contracts in the way Professors Klinge, Klosterhalfen and Deprest explained, but I am also satisfied that contraction of the mesh has clinical significance. It can increase the risk of recurrence of the condition which it was designed to arrest and it can cause a number of complications such as mesh exposure/erosion and chronic pain, including at rest and with sexual intercourse.
611 Professor Klosterhalfen’s evidence was that polypropylene mesh products are incompatible with the female pelvis because, while the meshes are flexible, they are not elastic. He said that flexible structures like polypropylene meshes are only able to elongate in one direction but an elastic structure like the vagina can stretch in all directions.695 He explained that the mesh acts as an inelastic material in an area needing elasticity and motion.696 He referred to this as a “mechanical mismatch”:697 He said that the mechanical mismatch is responsible for the main complications of polypropylene meshes in the pelvic floor. He explained:
From a mechanical point-of-view, a harder material in contact with a softer one can induce erosion in the latter [and] from a biological point-of-view, mechanical mismatch induces inflammation and tissue remodelling”.
612 Professor Klosterhalfen said that the mesh is “overstretched” and remains elongated, forming folds and wrinkles. 698
613 A PowerPoint presentation prepared by Ethicon Women’s Health & Urology for an Ethicon meeting in May 2008 about the development of a new device for pelvic floor repair, noted that “there is still NO evidence of a Device created specifically for the female pelvis”, “no patient-centric PF material”, pelvic floor materials were “still over-engineered”, less foreign body material was required, materials should “correlate to measured female pelvic physiological characteristics”, and the elasticity of the implant had to “differ according to direction”.699
614 An Ethicon PowerPoint presentation from March 2009, which compared the elasticity of the vagina with the elasticity of Prolift and Prolift+M before and after Monocryl absorption, showed that neither mesh matched the elasticity of the vagina and that Prolift was the least elastic.700 Professor Klosterhalfen said that this was particularly important because Prolift+M has Monocryl present for at least 84 days and that is “the most crucial implantation time period for a mesh with respect to tissue ingrowth”.
615 Professor Klosterhalfen agreed with the findings reported in this presentation. They accorded with his own experience, including his examination of explanted meshes. His opinion was that the mesh used in the Ethicon devices is “too stiff, inelastic, has too much material and inadequate pore size”. He also expressed the view that these characteristics cause excessive inflammatory response and excessive contraction which, in turn results in pain in the pelvis including dyspareunia. He was adamant that none of the Ethicon meshes used in the devices has the elasticity and “physiologic biomechanics” that are specifically adapted for surrounding tissues in the female pelvis. He reminded us that the meshes were all created for the purpose of hernia repair without regard to the biomechanics of the female pelvis.701
616 Similarly, Professor Deprest said that the main reasons for post-surgery complications and with complications associated with erosion included the biomechanical properties of the mesh, especially “mesh stiffness”.702
617 This concept of mechanical mismatch may well have been drawn from Dietz et al (2003). 703 In their paper, published in the International Urogynecology Journal in August 2003, 17 months before Prolift was launched, Professor Hans Peter Dietz and others from the Royal Hospital for Women in Sydney observed that, although tissue reaction is clearly an important factor, “common sense would suggest that any differential between the biomechanical properties of implant and surrounding tissue is likely to influence the likelihood of erosion”. They added:
This introduces the concept of the tissue-implant interface, which plays an important role in all biomaterial interactions. A mismatch between implant and tissue properties does not allow for the appropriate transmission of loads at the interface and may attribute to the poor clinical results observed with some of these materials.
618 They pointed out that in vivo clinical performance will depend on a range of factors, including surgical technique (especially the extent of the dissection), the mechanical properties of the host tissue, “biological ingrowth/ongrowth of the implants, and not least on in-vivo loading characteristics”.
619 Professor Wright’s opinion, however, was that, because Prolene Soft and TVT use the same synthetic polypropylene as Prolene used in sutures and hernia mesh, the host tissue responses to the individual fibres are expected to be the same.704 He then proceeded to cite three tissue studies which he said “clearly document that Prolene Soft usually evokes an initial minimal to mild acute inflammatory response, followed by a minimal to mild chronic inflammation that is accompanied by a minimal to mild fibrosis”.
620 There are several problems with this opinion.
621 First, two of the three tissue studies were animal studies, involving rat and rabbit models, over relatively short periods of time. These were small scale studies with short periods of follow-up. Professor Wright neglected to mention the sites of implantation, the size of the studies, or the length of follow-up. One, by Ethicon researchers (2001), was a 91-day rat study in which the meshes were implanted under the skin over the sacral vertebral column.705 The other, by Boulanger et al (2005), was a study of five mesh types in 12 sows over 72 days and only 10 prostheses were available for study.706 The meshes were implanted in the peritoneum (the lining of the abdomen).
622 Second, while the Boulanger study showed that tissue integration was best with polypropylene meshes and that Prolene and Prolene Soft induced the least severe inflammatory reactions, Professor Wright failed to refer to the cautionary note in the last sentence of the article in which the authors stated:
The role of atrophy, bacteriological environment and devascularization of the vagina in complications like erosion or infection must be explored, and the long-term efficiency and tolerance evaluated before a wide diffusion and promotion of these surgical techniques.
623 Third, Professor Wright did not take heed of the observations made in one of the articles he cited (by Elmer et al, 2009)707 that “rodent immunology may differ substantially from that in humans and the appropriateness of transferring data on biocompatibility between species is questionable”.708
624 Fourth, Professor Wright also cited Elmer et al (2009) in support of the proposition that Prolene Soft usually evokes an initial minimal to mild acute inflammatory response, followed by a minimal to mild chronic inflammation accompanied by a minimal to mild fibrosis.709 But he did not refer to the qualification made by the authors themselves that:
[I]t is important to recognize that a larger study population with transvaginal mesh surgery would undoubtedly generate an increased number of mesh related complications. Clinicians and patients should be aware of the possibility of late mesh related inflammatory reactions when using a large polypropylene mesh.710
625 Professor Wright went on to say that studies that used other forms of polypropylene mesh have shown that with longer implantation times the host tissue response includes moderately dense fibrous connective tissue.
626 One was a study by Pierce et al (2009)711 of 45 rabbits, 22 of whom were implanted with polypropylene mesh in the abdomen and most also had mesh placed in the posterior vagina. The control group received biologic graft materials. The period of follow-up was nine months. Notably, there were significant differences in results for the abdominal and vaginal implants. Erosion occurred in 6/22 (27%) of rabbits with vaginal polypropylene implants and those with longer grafts (2.9cm as opposed to 2.0cm) suffered vaginal erosions at a much higher rate (5/12 or 42%) compared with 1/10 or 10% in those who received the shorter pieces. No erosions occurred at the abdominal sites.
627 On the face of things, this study did not support Professor Wright’s assertion that the host tissue response should be the same for hernia meshes and pelvic meshes. The authors themselves, in a passage to which Professor Wright did not refer, stated:
This study importantly provides data to substantiate prior speculation that implantation of graft materials in the vagina elicits a different host tissue response than abdominal implantation.712
The authors also emphasised the need for additional research “to evaluate the long-term safety and efficacy of specific graft materials used in pelvic reconstructive surgery”. This was a decade after the release in Australia of TVT and six years after the release of Gynemesh PS.
628 The second of those studies was a 2002 study of 76 hernia explants conducted by Professors Klosterhalfen and Klinge. 713 At this point neither of them had begun to examine pelvic explants. The pelvis data pool was not even in contemplation. It should have been obvious to Professor Wright that the views of Professors Klosterhalfen and Klinge have evolved since then.
629 Professor Klosterhalfen said that it was an over-simplification to say that the host tissue responses to the individual fibres in Prolene Soft and TVT are expected to be the same as seen with Prolene sutures and hernia mesh. While the quality and nature of the tissue reaction is generally the same at the level of each individual fibre, it is necessary to consider the integration of the mesh as a whole. Differences are apparent when the integration of the entire mesh is considered. He said that Professor Wright neglected to take into account the importance of textile structure to the biocompatibility and function of the mesh. Professor Klosterhalfen explained that the intensity of the foreign body reaction depends on the surface area of polypropylene in contact with the host tissues; and the size of this area mainly depends on the porosity of the mesh or distribution of the pores. If the pores are too small, granulomas (small masses of inflamed granular tissue) have to fuse together and the fibrosis (scarring) can encapsulate the entire mesh causing the mesh to become stiff and lose function. Since the meshes are knitted, there are numerous large and small pores and it is not enough to just look at the large pores.714 That is obvious in an image of Prolene Soft reproduced in Professor Klosterhalfen’s second affidavit:
630 Professor Klosterhalfen explained that it is necessary to look at the distribution of the pores, the effective pore size under load and after tissue integration, and the pore geometry. The smallest diameter of the pore determines the risk of bridging fibrosis. He said that Prolene as used in the TVT devices has a porosity of 50.1% and an effective porosity of 0%.715 Similarly, Prolene Soft has a porosity of 62% which drops to 10% under load and an effective porosity of 26%, which falls to 0% under load. Measurements of pore sizes confirmed that the majority of pores are much smaller than 1mm.716
631 For all these reasons, as Professor Klosterhalfen deposed, it is inappropriate to transfer results of experimental studies on polypropylene fibres or sutures to polypropylene meshes.
632 Furthermore, almost all the following propositions are uncontroversial. Many of them are self-evident.
633 First, a suture is a single fibre with only one compartment (loop).717 All the meshes in question consist of a considerably larger amount of foreign material and multiple compartments (pores).718 Consequently, the nature and extent of the reaction to a mesh implant will be different from the nature and extent of the reaction to a suture or sutures. Professor Deprest and his fellow authors explained in a 2006 article that “[t]he nature and extent of the inflammatory reaction [generated by synthetic mesh implants] is regulated by the chemical and physical structure of the implant, the amount of material, and surface of the contact-area with the host”.719
634 Second, hernia mesh is used as a flat mesh in largely tension-free conditions whereas the pelvic floor is an area subject to a great deal of stress and strain. Professor Klinge explained the position this way in re-examination:
The – the introduction of meshes for hernia repair, it was related to the term “tension free”. That means that you have a – a huge flat mesh that reinforces the tissue. And it was – it was considered to be in contrast to the suture repair which causes some – some tension to the tissue, with all the problems of cutting through the tissue or creating pain. So this tension free is a general principle and the major advantage of using meshes for hernia repair, and so later on it – it became evident that if you’re using these hernia meshes in – in an area where you applied some tension to it, that this will create some problems. We have it in the abdominal cavity as well, when we are placing or when we tried to place these meshes close to the oesophagus in hiatal area, you have movement of the diaphragm, you have movement of the muscles, you applied some tension to the meshes, and then we have complications such – of erosion, migration, serious complications in this area. So the application of – of meshes that are designed for mesh – for a tension free application, in an area where you applied some tension, is not a good idea, or is – makes it dangerous to the patient and, from my point of view, a similar thing happens in the pelvic floor where you have to consider forces that are changing the appearance of the meshes in a completely different way than in the tension free procedures of – of the abdominal wall.720
635 Third, the clinical studies conducted by the Aachen Group confirmed that the more foreign body implanted or surface area covered, the greater the degree of inflammation.721
636 Fourth, the effect of polypropylene mesh on tissue differs according to the environment in which it is implanted.
637 As Professor Iakovlev pointed out, the mesh cannot expand or contract and, in order to function properly, several organs in the pelvis need to expand and contract. The mobility and function of the anterior abdominal wall, on the other hand, are much more limited.722 Further, the Ethicon devices are designed to be implanted superficially under sensitive mucosa and can easily erode through the mucosa. In contrast, erosions of hernia meshes are rare.
638 Notwithstanding the approach taken in the present case and in the development of the Ethicon devices, there is some evidence that the respondents recognised this.
639 In one of the slides in a PowerPoint presentation bearing his name dated 17 November 2004, Dr David Robinson, then an Ethicon medical director, emphasised that “[t]he Vagina is NOT the Abdomen (nor similar to any other surgical environment)”.723 Slide 17 identified the following “[u]nique issues to the vagina in the placement of large meshes”:
- Cannot be sterilized
- Relatively thin overlay with no real fascial layer
- Attachment sites are difficult to access
- The three dimensional architecture and various vector forces are complex
- Subject to great forces with little or no bony (and often pelvic floor muscle) reinforcement
- Must remain pliable for filling and emptying of pelvic organs and maintain the capacity for sexual function.
640 A “confidential” PowerPoint presentation prepared by Ethicon’s Women’s Health & Urology Research and Development team acknowledged that:
Today’s vaginal implants do not consider the patients’ biomechanical needs.724
It recognised that “unmet biomechanics” lead to “misfunction, pain & shrinkage”, which in turn, leads to a “handicapped patient”. The document is undated but, since it refers to a 2006 article, was obviously created after the launch of Prolift.
641 At some stage during the first decade of this century, Ethicon began work on designing a mesh that was intended to be adapted to the pelvic floor and that would resist pore collapse. The project was originally referred to as Project T-Pro and later Project Thunder. Nothing eventually came of it, however, and it appears that it was abandoned in about 2011 following budgetary cuts.725
642 In their Invention Disclosure, Drs Trezwik and Meier identified the problem the proposed mesh was designed to solve in the following way:
Each implant introduced into the human body has to mimic the mechanical properties of the tissue it should support, no matter if it is a total hip replacement, vascular graft or mesh for soft tissue reinforcement. In any of these areas the tissue properties are not uniform (isotropic) but change continuously within different areas of the respective anatomical region. Often these variations correlate with the specific mechanical load that occurs in these tissues anisotropy). Furthermore different kinds of involved biologic structures (e.g. Fascia, ligaments, muscle, tendons etc.) have different biomechanical properties. The ideal implant must take these variations into consideration and minic the different mechanical tissue properties …
Mesh based implants which are currently used in pelvic floor reconstruction are based on mesh constructions originally designed for the treatment of hernias in the abdominal wall region. It is important to understand that the biomechanical properties of the abdominal wall and the pelvic floor differ especially in regard of elasticity and anisotropic material behaviour. To fulfil the desired biomechanical compatibility of mesh based implants for pelvic floor reconstruction, it is important to take the biomechanical properties of the implantation site into consideration.726
643 At one point, and despite their reliance on Professor Wright, the respondents submitted that it was apparent that the forces in the female pelvis are not the same as those in the abdominal wall. They contended that the forces are unlikely to be unidirectional and more likely to be multidirectional, and in the case of POP mesh, there are differences between the forces applied to the arms and the body of the mesh.727 These submissions should be accepted. They reflect the evidence given by Professors Klosterhalfen and Klinge.
644 The opinions of Professor Klosterhalfen and Klinge were also supported by the work of Professor Moalli and her colleagues.
645 Barone et al (2015) (Professor Moalli included) noted that literature and marketing pamphlets distributed by manufacturers of urogynaecological mesh demonstrate (or perhaps more accurately seek to demonstrate) biocompatibility of SUI and POP products via implantation studies in the abdominal wall. They accepted that these studies were necessary “to initially demonstrate the ability for mesh designs to be implanted in a host withut overt rejection”. At the same time, however, they emphasized that “the abdominal wall and the pelvic floor are quite distinct in terms of both biological factors and the structural functions that a mesh device is intended to withstand”. Consequently, they wrote that “current urogynaecological mesh designs are more similar to a prototype solution rather than an optimal one”.728
646 Barone et al explained that the environment of the pelvic floor is much more complicated than the abdominal wall from both a biological and mechanical perspective.729
647 For a start, they pointed out, hernia repair meshes are only in direct contact with the abdominal fascia. Pelvic meshes, on the other hand, are placed in an environment with a wide range of soft tissues, including smooth and striated muscle, various kinds of connective tissue, and specialised organs.
648 What is more, the vagina is regarded as a “clean-contaminated” field. When polypropylene is implanted transvaginally, it is seeded with bacteria which contribute to infection and inflammation in the tissues.
649 Professor Klosterhalfen said that even a low-grade infection can enhance the foreign body reaction, contraction, and encapsulation, and that infection occurs with erosion of polypropylene mesh. In 90% of the cases of erosion seen in Professor Kosterhalfen’s pelvis data pool, there were histological signs of infection.730
650 Professor Klinge deposed that the implantation of mesh can result in a biofilm which will make it difficult for the host cells to kill the infection and the development of a biofilm will protect the harmful bacteria the host cells set out to kill.731 A biofilm is an assembly of colonies of bacteria fixed upon a support and locked into an encapsulating matrix which is resistant to both stress and antimicrobials.732
651 Professor Klinge said that one of the major causes of mesh-related infections in patients who have been implanted with pelvic floor meshes transvaginally is the formation of a biofilm.733 I accept this evidence. In a 2007 article relating to the use of synthetic mesh in the treatment of stress urinary incontinence, Klinge et al wrote that infections are usually caused by an intraoperative colonisation of biofilm-forming bacteria that may reactivate in the patient’s later life and present as late abscesses or fistula. They said that sterile handling of the implant and the use of intravenous antibiotics during surgery significantly reduces infection rates, but pointed out that late-onset infections are mainly dependent on the patients’ individual immune system.734
The TVM Group were alive to this. Berrocal et al (2004) said that acute infection was “very rare”, “undoubtedly decapitated by prophylactic antimicrobial therapy” presumably during implantation, but “the real problem associated with the placement of [mesh prostheses]”, including those made of Prolene and Prolene Soft, was chronic infection. 735
652 The respondents take issue with the applicants’ case that infections may occur because of bridging fibrosis or mesh deformation causing a reduction in pore size effected by either phenomenon.736 To the extent that the reduction in pore size prevents the macrophages and neutrophils from doing their work, the evidence is persuasive that, if the pore size is reduced at least to 75µm or below (the figure given in 1997 by Professor Amid), then anything that causes the pore sizes to diminish or, worse still, collapse can increase the risk of infection. It is unnecessary to resolve the question, however, because there is no dispute and no doubt on the evidence that the mesh can both cause an infection and exacerbate an existing infection.
653 In numerous other respects the environment of the pelvic floor is very different from the wall of the abdomen or, for that matter, the groin.
654 Barone et al (2015) pointed out that, “[i]n general, the soft tissues of the pelvic floor are metabolically active, with compositions that have been shown to change dramatically with normal aging and in response to hormone-driven events such as pregnancy, menstrual cycle, and menopause”. They referred to a number of animal studies in which “markedly different host responses” had been demonstrated to abdominal hernia mesh and POP mesh.737
655 Then there is the mechanical environment. Barone et al noted that the structures making up the pelvic floor, like most biological tissues, respond to the presence or absence of mechanical loads.738 They regarded the most significant consideration for pelvic meshes to be the fact that they are subjected to “predominantly uniaxial tensile loading conditions”, an environment, they wrote, which is “quite different” from the loading conditions to which abdominal hernia mesh is subjected. They explained:
During a hernia repair, the mesh is placed within the abdominal wall, fixed along its entire perimeter. Since the abdominal cavity can be thought of as a pressurized vessel, the mesh graft must function as the wall of a pressurized vessel. Under the pressure exerted within the abdominal cavity, the abdomen expands or is “inflated,” and the abdominal wall resists and limits this expansion. Given this loading condition, hernia mesh is placed in tension along all axes simultaneously, much like the surface of a balloon upon inflation. This loading environment helps the mesh maintain its original geometry and pore sizes … However, incontinence meshes are used as suspension structures for slings in the retropubic or transobturator space. Similarly prolapse meshes act as suspension cables, attaching the vagina to the sacrum or pelvic sidewall. These configurations load the mesh, particularly the fixation arms of these devices, primarily in a tensile uniaxial fashion. Even while the name TVT suggests that the device does not experience tension, its intended function is to have the urethra compressed along the sling, which would place the mesh in tension. The tension-free component is thus more representative of the surgical technique rather than the long-term in vivo function of the device.
656 Importantly, Barone et al referred to a number of studies, including one of their own published in 2008, which showed a dramatic reduction in maximum pore size and porosity of “nearly all synthetic meshes” under uni-axial load. In many cases, they said, all the pores in the tensioned mesh were smaller than 1mm in diameter, “diminishing the potential for tissue ingrowth and promoting bridging of fibrous encapsulations such that long-term tissue incorporation is compromised” (citations omitted). In these studies forces of 5 and 10N were shown to cause the dramatic reductions in pore size. It will be recalled that Ethicon’s own research showed that a force of merely 1N could have this effect. It will also be recalled that Professor Moalli’s team established in the study published a year later (Barone et al (2016), that all current prolapse meshes were likely to experience a complete collapse of pores “at some load”.
657 In cross-examination Professor Wright revealed that he had not come across the Barone et al (2015) chapter (chapter 13).739 He said that it would not show up in standard literature reviews.740 Yet, the book from which it was extracted (Host Response to Biomaterials) was described by Professor Santerre as “a pretty comprehensive encyclopaedia (scil) in the international biomaterials world”.741 Professor Santerre volunteered that chapter 13 was “an important chapter to read because it was a compendium review from a key opinion leader in the field”.
658 In cross-examination Professor Wright agreed that uni-axial loads could behave as Barone et al described but, despite the opinions he had expressed in his report about the suitability of the Ethicon devices, admitted he knew nothing about “the forces which are involved with mesh in the vagina”.742
659 Barone et al (2015) also observed that there was “significant evidence” to show that the stiffness of an implantable device can change tissue remodelling and response. They wrote that, in general, “implant stiffness is believed to induce a maladaptive remodelling response through a phenomenon known as stress shielding”. They regarded the concepts of “force distribution” and “stress shielding” as “extremely important” considerations since it has been shown that the vagina responds to mechanical stimuli.743
660 Furthermore, the female genital area has a much higher nerve density in comparison with the anterior abdominal wall and the groin. Professor Iakovlev said that the scar inhabiting and surrounding transvaginal mesh has the highest nerve density of all explanted surgical meshes he has examined as a pathologist.744 Consequently, placement of mesh implants in the female pelvis carries a higher risk of chronic pain than the placement of mesh for hernia repair, whether ventral or in the groin.745 Professor Wright quibbled with aspects of this evidence. There was also a dispute about whether the nerves in question are sensory (rather than motor) and whether, even if they are (or include) sensory nerves, this is the cause of chronic pain. Still, in cross-examination Professor Wright accepted that there was a vastly higher rate of mesh exposure/erosion with vaginal implants compared to abdominal wall implants. In cross-examination he said that he “did not expect abdominally-implanted mesh to have [an] erosion rate of any significance”.746 In his view this was explicable by anatomic differences in the thickness of the overlying tissue. 747
661 There are other differences of significance that are likely to account for the higher rate of complications in the pelvic floor with the use of polypropylene mesh in general, and the Ethicon meshes in particular.
662 The differences between the anterior abdominal wall and the female pelvis were well summarised by Professor Iakovlev in his second report:
a. There are no sharply separated parallel anatomical planes, but rather several round mobile organs in a narrow pelvic space. There is also no real fascia and no subcutaneous fat.
b. There are several organs in the pelvis that need to expand and contract for their function. The mesh cannot contract or expand; moreover, together with the encapsulating scar it provides non-physiological connections to the mobile organs. In comparison, the mobility and the function of the anterior abdominal wall are much more limited.
c. The devices are designed to be implanted superficially under sensitive mucosa and can easily become eroded through the mucosa. Erosions of hernia meshes are rare.
d. The devices are designed to correct the anatomical position of the organs and therefore press against the organs. This pressure can force the mesh to migrate into the organs. In contrast, a hernia mesh would have a much larger area of pressure distribution than a Tape Implant.
e. Vaginal mesh devices cross paths of normal innervation and vascular supply to the mucosa, the bladder and the rectum. After implantation, the innervation and vascular supply need to be restored either through or around the mesh. In comparison, hernia mesh is placed parallel to the anatomical layers and parallel to the nerve and vessel direction.
f. The female genital area is one of the most sensitive areas in the female body. It is much more sensitive than the anterior abdominal wall.
g. The arms of the [POP devices] and the ends of the [SUI devices] cross the muscles and innervation network of the soft tissues and skin. Hernia meshes are placed parallel to the muscular layers, nerves and vessels.748
663 Dr Iakovlev explained the effect of these differences on the incidence and severity of complications as follows.749
664 First, erosion through the vaginal mucosa is one of the most common complications of implantation with polypropylene mesh but it is rare to encounter mesh erosion through abdominal skin or into internal organs with hernia mesh.
665 Second, chronic pain is a complication of both hernia and vaginal mesh. The differences are mainly in the distribution and pattern of radiation. Since the arms of the POP devices and the ends of the SUI devices cross many structures in the pelvis, “pain distribution can involve areas from suprapubic to vaginal, introital, deep pelvic/vaginal, obturator/hip/groin, and into the buttock”. Pain can also radiate into the medial thigh.
666 Third, with hernia mesh, pain on intercourse can occur in cases of mesh migration into the spermatic cord, but dyspareunia is more prevalent as a complication of vaginal mesh.
667 Fourth, urinary obstruction is a complication “almost unique to vaginal implants”.