The Fetal Pillow Deflates—Lessons for All

Abstract

After the recent retractions of three publications reporting clinical studies of the fetal pillow (Cooper-Surgical), the National Institute for Health and Care Excellence (NICE) has reversed its recommendation of the device, and the Royal College of Obstetrics and Gynaecology has retracted a Scientific Impact Paper [1]. Ongoing use of the device is based on a very fragile evidence base. This commentary summarises the history of the fetal pillow and highlights lessons for medical device regulation, journal processes for manuscript review, physicians' enthusiasm for new technology, evidence synthesis groups' rigour in evaluating evidence, and publishers' and journals' assessments of concerns about publication integrity. Impacted fetal head at the time of caesarean delivery is an obstetrical emergency that may affect 1.5% of births [2]. Such caesareans can be difficult if the fetal head is deeply impacted within the maternal pelvis. The condition is associated with maternal uterine and bladder injuries, and fetal skull and brain injuries. Optimal management is uncertain. It was in this context that the fetal pillow was invented in 2006 as a ‘foetal disimpacting system’, later presented to regulators as a ‘foetal membrane manipulator’ and finally to the medical world as the more gentle ‘fetal pillow’. It is intended to assist obstetricians performing emergency caesarean section during the second stage, or late in the first stage of labour. The idea is that the pillow elevates the head out of the maternal pelvis before the operation starts. The inventor of the fetal pillow established a company, Safe Obstetrics Systems, of which he was the Chief Executive Officer. The fetal pillow was launched first in the United Kingdom, in July 2011, after approval by the European Union regulators [3]. At that time, regulation of medical devices in Europe was delegated to private companies (‘notified bodies’), which were charged with evaluating safety and efficacy but whose activities were opaque, inconsistent and subject to commercial conflicts of interest [4]. This unsatisfactory situation still exists today in the United Kingdom, with minor differences. Promotional material for the fetal pillow refers to ‘years of design, development, mechanical and clinical testing before the device could be used in “real clinical settings”’ [3]. Yet, the only published evidence at the time of its UK launch was a 2008 paper, which described a ‘pilot study’ in 30 women in advanced labour with a deeply engaged fetal head [5]. That study was reportedly conducted at the Institute of Obstetrics and Gynaecology and Government Hospital for Women and Children in Chennai, India, but no staff members of either institution were coauthors or acknowledged in the publication. Details of ethical oversight were not included. The paper reported that the inflated device elevated the fetal head by about 3 cm, but neither fetal nor maternal outcomes were reported. Although the authors stated that a ‘prospective randomised trial of its use in caesarean section with a deeply engaged head is presently underway’, clinical trial evidence, now retracted [6], did not emerge until 8 years later. Approval to market the fetal pillow was granted by the Therapeutic Goods Administration in Australia in 2014 (Australian Register of Therapeutic Goods ID 225818). Details of the application are not publicly available, but it was supported by an obstetric unit, which had experienced two fetal deaths from impacted fetal head, after a staff member heard a 2012 conference presentation about the device. Thus, when the fetal pillow entered clinical use in Europe, there was one publication, with integrity concerns, that reported a surrogate end point in 30 participants. When approval to market in Australia was obtained, the only additional data were from conference presentations reporting a nonrandomised study that compared outcomes in 50 deliveries using the fetal pillow with historical data from 124 deliveries that did not use it [7], and a non-prespecified interim analysis of a randomised clinical trial [8]. The publications that resulted from each conference presentation were recently retracted because of concerns about the integrity of data and results [6, 9]. Approval from the Food and Drug Administration (FDA) in the United States was obtained in July 2017 [10]. Supporting clinical evidence included a retrospective cohort study conducted in 2013–2015 in Australia by researchers at the unit whose staff had supported the regulatory submission in that country, which reported improvements in the surrogate outcomes of estimated procedural blood loss and umbilical cord pH [11]. However, the pivotal supporting evidence was from a 240-participant randomised controlled trial, now retracted [6], which reported fewer extensions of uterine incisions, shorter operative time, fewer maternal blood transfusions and shorter hospital stay in the fetal pillow group. With regulatory approvals in hand, Safe Obstetrics Systems successfully marketed the fetal pillow, such that by 2020, 12 500 devices were sold annually, generating revenue of GBP 3.5 million that year. The company, and its sole product the fetal pillow, was acquired by Cooper-Surgical in March 2021 for US$50 million. In 2024, this single-use device is marketed at 350GBP. By the time Safe Obstetrics Systems sold the fetal pillow to Cooper-Surgical, widespread concerns had been raised about the integrity of two of the studies that underpinned regulatory approval and clinical use of the device. The concerns, which spanned eight of the 11 domains of the REAPPRAISED publication integrity checklist [12], were notified to both journals in a joint email in March 2020 and posted on PubPeer in October 2020. What followed was unfortunate, and unfortunately not unusual. The affected journals did not coordinate efforts to address the concerns. J Obstet Gynaecol and its publisher acknowledged receipt of the concerns but did not otherwise respond for > 3 years. At that point, they stated that they had contacted the authors and institution in March 2020, but there was no ‘clear picture’ of what had transpired. Nearly another year passed before the paper was retracted because ‘the authors did not respond to our (the journal's) queries and despite several attempts to obtain a response from the authors and their institution we have not received any original data or necessary supporting information’. Presumably, the initial contact in March 2020 went unanswered. No expression of concern (EoC) was published during the > 4 years of investigation. Seven months after receiving the concerns about its publication, Int J Gynaecol Obstet said it had conducted a thorough investigation and concluded that ‘the major concerns had been addressed’ after ‘satisfactory explanations’ were provided by the authors. It declined to share the authors' responses to the detailed concerns or to state what information was considered during its assessment. When the concerns were reiterated to the publisher, its integrity team became involved to ‘support the journal in their (sic) enquiries’. By June 2021, a ‘more fulsome’ response had been obtained from the authors, but no response had been forthcoming from several attempts to contact the institution. Nothing happened for the next 18 months, until an EoC was published in January 2023. The EoC did not reflect the nature or scope of the concerns raised. In June 2023, the publication was retracted—the notice stated that no patient data were available. Thus, it took > 3 years to resolve the integrity concerns about each publication, and almost 3 years before publication of an EoC for either. During that time, clinicians and patients continued to use the device, unaware that the pivotal evidence supporting its use was unreliable. During that time, the 2016 trial publication featured prominently in guidance published in November 2022 by NICE, which had presumably not considered the integrity concerns posted on PubPeer in 2020. NICE concluded that the fetal pillow was safe and effective in managing the impacted fetal head. NICE also misconstrued another fetal pillow publication as a randomised trial [13], and failed during its deliberations to detect a slew of integrity concerns about that paper, which was recently retracted [13]. As the dust settles, how do things stand? It seems unlikely that a responsible regulatory body or clinical governance group would now endorse the use of the fetal pillow if it were to consider the remaining evidence. A meta-analysis [14] of four unretracted observational studies (428 fetal pillow users) [11, 15-17], and a small randomised trial (30 users) [18] concluded that there was no evidence of benefit or harm with the device but that the certainty of evidence was very low. A subsequent observational study of 375 fetal pillow users also reported no benefit [19]. It remains possible that the fetal pillow is an effective intervention for impacted fetal head, but rigorous evidence is needed. A double-blind randomised clinical trial of the device, which is underway, with primary outcome maternal uterine incision extension, will be helpful. Current Cooper-Surgical promotional material cherry-picks the remaining evidence to claim both maternal and fetal benefits from use of the fetal pillow. NICE now refers readers to its 2015 advice, which is that a balloon device ‘should only be used with special arrangements for clinical governance and audit or research’. That seems generous to the device manufacturers. We think that NICE should recommend that the fetal pillow should not be used outside a controlled clinical trial. Diab and colleagues wrote in response to the first description of the fetal pillow that ‘anecdotal evidence and “tips” about surgical technique have always passed between generations of clinicians but it is imperative that new techniques and devices are rigorously developed and evaluated prior to introduction in clinical practice’ [20]. For the fetal pillow, those wise words were ignored. A combination of clinical need, enthusiasm for new technology, deficiencies in journal processes at manuscript submission, lax regulatory oversight, successful marketing, poor evidence synthesis processes and inefficient publisher responses to concerns about publication integrity resulted in the fetal pillow becoming embedded in clinical practice without reliable supporting evidence. Consequently, despite the expenditure of millions of dollars, clinicians still do not know whether the money has been wasted on an ineffective device, patients have been harmed, or a useful invention is being underused. How could things be improved? Clinicians should require better quality evidence than was ever available for the fetal pillow before changing practice. This should apply to any device that claims to favourably and safely improve clinically important outcomes. Publishers and journals should require and review raw study data and key study documents at manuscript submission to improve the quality of their published product. Regulators should do the same when considering whether to approve a device. Evidence synthesis groups should screen influential publications such as randomised controlled trials for compromised integrity during their deliberations. Checklists to assist assessment of publication integrity already exist [12], and a tool to assist authors of Cochrane systematic reviews is in development [21]. Publishers and journals should conduct assessments of publication integrity concerns in a systematic, robust and efficient manner, and publicly notify readers early in the process. Achieving even some of these goals would help to prevent recurrences of debacles such as that which occurred for the fetal pillow. All authors contributed to the conception and writing of the maunscript. All authors approved the final version. The authors declare no conflicts of interest. Data sharing is not applicable to this article as no new data were created or analyzed in this study.

Topics & Concepts