Surgical mesh and patient safetyBMJ 2018; 363 doi: https://doi.org/10.1136/bmj.k4231 (Published 10 October 2018) Cite this as: BMJ 2018;363:k4231
- Correspondence to: C Heneghan
The use of surgical mesh for vaginal prolapse and stress urinary incontinence increased rapidly for 20 years and then fell from grace, mainly because of a failure to take patient safety seriously. The BMJ investigation by Jonathan Gornall123 itemises the many failings that made this rise and fall inevitable: the lack of postmarketing studies; the failure to establish device registries; the influence of financial conflicts of interest; and, in the UK, the ineffectual role of the National Institute for Health and Care Excellence. Between them, these factors distorted the research evidence and adversely affected the care that women received, causing many of them unnecessary and irreparable harm. “Nobody involved with the mesh revolution emerges covered in glory,” says Gornall.1
The first duty of healthcare systems should be to avoid unintended or unexpected harm during the provision of healthcare. The mesh story shows how healthcare systems around the world have failed in this duty. Systemic problems exist with how medical devices are approved, and, once they are on the market, how they are assessed as safe and effective.
Regulatory approval is not a marker of either effectiveness or safety, and this is unlikely to change with the updated EU medical device directives.4 The problem is, as Gornall explains, that it is not feasible to study absolute long term safety and performance of any implant before market launch.1 This can be achieved only after approval, through the diligent collection and independent monitoring of data on patients’ outcomes.
In 2003, NICE recommended that data on the effectiveness and safety of mesh should be collected “over a period of 10 years or more,” but this didn’t happen.5 Other UK regulatory bodies, including the MHRA, discussed the need for a national registry, but none was forthcoming. In 2018, NHS Digital audited Hospital Episodes Statistics,6 reporting “experimental statistics” that have several limitations that affect interpretation (eg, reporting of the primary diagnosis in outpatient episodes is not mandatory).
Globally, more than 100 000 women are suing manufacturers—over 1000 in the UK alone.7 A 2016 Cochrane review of transvaginal mesh for vaginal prolapse reported 10% of women required reoperation in the mesh group.8 A retrospective review of NHS vaginal mesh procedures for stress urinary incontinence reported complication rates in women who had concomitant procedures of 9.8% within five years.9
In December 2017, too late to prevent substantial harm, NICE issued guidance, effectively banning mesh for use in pelvic organ prolapse because the evidence of long term effectiveness was “inadequate in quality and quantity.”10 In July 2018, it ordered a full suspension of the use of vaginal mesh in England to avoid further risk of “life-changing injuries” to women. NICE’s latest draft guidance (published 9 October 2018) states non-surgical options for stress urinary incontinence or pelvic organ prolapse must be offered before any operation, the procedures should be done only by specialist surgeons, and “surgeons must also record any intervention using surgical mesh/tape in a national database.”11
National registries of interventional procedures can improve outcomes and patient safety. In Sweden, government administered registries are complemented by over 100 Swedish healthcare quality registries designed to improve care.12 For example, the Swedish national cataract registry has data on 96% of all cataract removals since 1992 and is used to drive down postoperative complications.13 In South Korea, national health insurance data contain information on treatments, procedures, and diagnoses for almost 50 million beneficiaries and inform the ongoing monitoring and prevention of adverse events.14 The Australian Orthopaedic Association National Joint Replacement Registry, funded by the Commonwealth Department of Health, was instrumental in uncovering the problems with metal-on-metal hip prostheses15; analysis on mortality risk and revision rates has been recently informed by 1 335 938 joint replacement procedures.16 In the UK, the National Joint Registry—set up by the Department of Health in 2002—collects information on all of the major joint replacement operations to monitor performance, effectiveness, and improve clinical standards.
The sustainability, coverage, and completeness of registries are primarily influenced by national policy. Registries must be adequately funded: Sweden spent $47m (£36m; €41m) in 2016 on its registries.17 Legislation is needed to mandate what is measured and in whom, since voluntary entry of data has been shown to fail, and to ensure reporting requirements are adhered to and completeness of information is maintained. High quality anonymised data should also be made available for clinical research. Finally, registries should collect data on the whole chain of care from start to finish, not just on isolated incidents.
The postmarketing assessment of vaginal mesh has been a shameful episode in the history of implantable devices. Surgeons, manufacturers, regulators, and governments have all played their part in this failing. Mandated national registries are needed for all implantable devices to protect patients, improve outcomes, reduce costs, and identify best practice. Another mesh tragedy is otherwise inevitable.
Competing interests: We have read and understood BMJ policy on declaration of interests and declare that CH is a clinical adviser to the All Party Parliamentary Group (APPG) on surgical mesh has received expenses and fees for his media work, and grant funding from the NIHR, the NIHR School of Primary Care Research and the NIHR BRC Oxford. He has received financial remuneration from an asbestos case and given free legal advice on mesh cases. He is an NHS GP working in urgent care, and editor in chief of BMJ Evidence-Based Medicine.
Provenance and peer review: Commissioned; not externally peer reviewed.