Volume 126, Issue 2 p. 138-141
Commentary
Free Access

Reproductive medicine: still more ART than science?

J Wilkinson

Corresponding Author

J Wilkinson

Centre for Biostatistics, University of Manchester, Manchester, UK

Correspondence: J Wilkinson, Centre for Biostatistics, Division of Population Health, Health Services Research and Primary Care, School of Health Sciences, University of Manchester, Rm 1.307 Jean McFarlane Building, University Place, Oxford Rd, Manchester M13 9PL, UK. Email: [email protected]Search for more papers by this author
S Bhattacharya

S Bhattacharya

College of Biomedical and Life Sciences, Cardiff University School of Medicine, Cardiff, UK

Search for more papers by this author
JMN Duffy

JMN Duffy

Primary Care Health Sciences, University of Oxford, Oxford, UK

Balliol College, University of Oxford, Oxford, UK

Search for more papers by this author
MS Kamath

MS Kamath

Reproductive Medicine Unit, Christian Medical College, Vellore, India

Search for more papers by this author
J Marjoribanks

J Marjoribanks

Cochrane Gynecology and Fertility Group, University of Auckland, Auckland, New Zealand

Search for more papers by this author
S Repping

S Repping

Centre for Reproductive Medicine, Academic Medical Centre, University of Amsterdam, Amsterdam, the Netherlands

Search for more papers by this author
A Vail

A Vail

Centre for Biostatistics, University of Manchester, Manchester, UK

Search for more papers by this author
M van Wely

M van Wely

Centre for Reproductive Medicine, Academic Medical Centre, University of Amsterdam, Amsterdam, the Netherlands

Search for more papers by this author
CM Farquhar

CM Farquhar

Cochrane Gynecology and Fertility Group, University of Auckland, Auckland, New Zealand

Search for more papers by this author
First published: 16 July 2018
Citations: 44

This article includes Author Insights, a video abstract available at https://vimeo.com/rcog/authorinsights15409

Abstract

image This article includes Author Insights, a video abstract available at https://vimeo.com/rcog/authorinsights15409

Author-Provided Video

Reproductive medicine: still more ART than science?

by Wilkinson et al.

The history of obstetrics and gynaecology is not a tale of evidence-based practice. Tradition, expert opinion, and the lure of new technology have frequently superseded evidence as the primary driver for clinical decision making.1 The proof can be found in a litany of dubious interventions that have gained widespread popularity despite an absence of high-quality data attesting to their effectiveness and, in some cases, ample credible evidence demonstrating harm. As a specialty, we have relied on investigations including X-ray pelvimetry and antenatal stress tests, subjected innumerable women to stilboestrol and thalidomide, and have performed routine episiotomy in all primigravid women.1 It is no surprise that, in 1979, Archie Cochrane famously awarded obstetrics the ‘wooden spoon’ for being the least evidence-based specialty.

Reproductive medicine has come a long way since then, and there have been dramatic advances in the field, such that biological parenthood is a reality for many couples for whom adoption was the only available route in the past. The gold standard in reproductive medicine is now ‘to provide childless couples with the best possible management of their fertility problems, while at the same time ensuring that they are not exposed to unnecessary risks or ineffective treatments’.2 In order to achieve this standard, we need reliable evidence that can be used to develop clinical practice guidelines. Unfortunately, even where we have good evidence, its uptake is far from universal.1 This is particularly true in the field of assisted reproduction, which has become increasingly commercialized, with a pronounced predilection for overdiagnosis, overuse, and overtreatment.

Too much treatment?

One in seven couples will struggle to conceive within a year, and will be labelled as infertile. Yet, population-based data show that half of this group will go on to conceive in the next 12 months.3 Meanwhile, success rates for fertility treatments remain modest, with live birth rates for in vitro fertilisation (IVF) estimated to be in the region of 27% per treatment cycle (www.hfea.gov.uk/media/2563/hfea-fertility-trends-andfigures-2017-v2.pdf). Despite this fact, media-fuelled unrealistic expectations on the part of couples combined with skilful marketing by commercial fertility clinics has caused a global explosion in the use of assisted reproductive technology. As few governments fund infertility treatment, the private sector is buoyant. Contrary to guidelines, assisted reproductive technology is frequently initiated early for couples who still have a reasonable chance of conception, promoting overtreatment, increasing avoidable harm, and increasing healthcare costs.4 A plethora of essentially unproven add-on interventions are offered to couples in a technological arms race between fertility clinics.5 As financial considerations often influence treatment policies, the novelty and apparent sophistication of these new technologies take precedence over proof of effectiveness and safety. For many couples, the additional expense of add-on therapies is itself interpreted as a sign of quality. It is the perfect storm for exploitation.

The gap between clinical practice and evidence

Although the absence of evidence in many areas of reproductive medicine remains a problem, blatant disregard for evidence that is available is an inexcusable but all too common phenomenon. One of the most notable and regrettable examples of ignoring evidence in assisted reproductive technology is the persistent practice of transferring two or more embryos at a time. This option appears attractive to many couples, who see it as a means to maximise their chances while obviating the need for further treatment to have a second child. Some IVF clinics continue to support this practice on grounds of patient autonomy, while downplaying the considerable maternal, fetal, neonatal, and childhood risks associated with multiple pregnancy. Although multiple pregnancy is almost completely preventable by the practice of single embryo transfer, many healthcare providers across the UK, Europe, and North America continue to transfer multiple embryos. One explanation is that IVF is often self-funded, whereas maternity and neonatal care are often covered by the state or by insurance companies. If the same funder paid not only for IVF but also for maternity and neonatal care, then there would be a strong incentive to reduce multiple pregnancies in order to avoid increases in fetal, neonatal, and childhood morbidity, and the attendant costs. In 2005, New Zealand increased funding for IVF from one cycle to two, provided that the strategy of single embryo transfer was used. As a result, New Zealand now has the lowest rate of IVF-related multiple pregnancies (4%) in the world (https://npesu.unsw.edu.au/surveillance/assisted-reproductive-technology-australia-and-new-zealand-2015).

Another example where common practice defies the evidence is the expansion of intracytoplasmic sperm injection (ICSI), originally designed for men with very poor semen quality, to include couples where the semen parameters are normal. Indeed, it has now become the dominant form of IVF in many countries. The Cochrane review of IVF versus ICSI for non-male infertility includes a single randomised trial involving 415 couples that reported 70/213 continuing pregnancies following IVF, compared with 51/202 in the ICSI group (odds ratio 1.44, 95% confidence interval 0.95–2.21).6 Data from a subsequent Canadian study that randomised 60 women with unexplained infertility to IVF or ICSI failed to show a statistically significant difference in live birth rate.7 In this context, it is the focus on a surrogate outcome (fertilisation rate) rather than the clinical outcome (live birth) that has presumably influenced clinical decision making; trials of randomised oocytes suggest that a strategy of ICSI in non-male infertility could reduce the number of cases of total fertilisation failure, but an improvement in live births has not been demonstrated. Similarly, pre-implantation genetic testing and time-lapse systems for embryo incubation are becoming routine practice in many clinics, despite a lack of robust evidence of clinical benefit.8, 9

Marketing trumps informed patient choice

The primary criterion used by many people with infertility to select an IVF clinic is its success rate. In order to be competitive, fertility clinics must convince couples that the path to parenthood leads through their doors. But there is a clear tension between direct-to-consumer advertising of assisted reproductive technology and informed patient choice. Clinics frequently wield emotive language and creative statistics to persuade rather than to inform.10, 11 Websites speak of ‘dreams’ and ‘miracles’,10 and bespoke success rates can be constructed from a large array of statistics in order to cast the performance of a clinic in a favourable light.11 With the lack of binding standards on assisted reproductive technology outcome reporting, couples may be misled about the likelihood of success and may compare clinics on the basis of incomparable figures. Self-regulation does not appear to furnish couples with impartial information regarding the effectiveness of different treatment interventions.

Challenges for evidence-based assisted reproductive technology

Before being implemented as routine, any assisted reproductive technology intervention should undergo preclinical and clinical trials, followed by randomised controlled trials (RCTs) to prove efficacy and safety, with long-term follow-up to evaluate the continuing effectiveness and safety.5 In assisted reproductive technology, however, an eagerness to identify attractive new interventions has led to a quantity over quality approach to RCTs. Consequently, the clinical effectiveness of interventions is often not robustly assessed. A pivotal limitation is the size of the trials undertaken. In order to achieve 80% power to detect an improvement in birth rate from 27 to 32% at a 5% significance threshold, a trial size of 2610 women is required. This far exceeds the numbers of women we actually see in the vast majority of IVF trials. Routinely collected data and large electronic databases have been touted as a solution, but the absence of random allocation introduces doubt regarding the accuracy of any conclusions drawn. These are essentially large-scale observational studies, and do not offer a ‘free pass’ to the evaluation of assisted reproductive technologies.

In principle, the machinery of meta-analysis offers a means to overcome the limitations of small individual trials, but this is predicated on sufficient cumulative sample sizes after pooling studies. The trend to ‘test’ many interventions in small trials does not guarantee this will hold. Moreover, if trials do not report outcomes in a consistent fashion, meta-analysis is precluded. Recent reviews of reporting standards in this field suggest that they do not.12 Moreover, many trials measure outcomes in such a way that the benefits of randomisation are lost, and many fail to report cumulative birth rates per couple, which is probably the most relevant outcome for patients. The need to establish statistically valid, patient-centred outcome measures is the motivation behind the Core Outcome Measures for Infertility Trials (COMMIT) project, which will establish core outcomes to be reported in all assisted reproductive technology trials (www.phc.ox.ac.uk/research/hypertension/pregnancy/commit). Fewer, larger RCTs developed through priority-sharing partnerships and enabled by clinical trial networks committed to the use of common reporting standards will ensure high-quality evidence with maximum clinical utility. A new priority-setting partnership has been established for infertility aiming to bring together professionals, researchers, and people with infertility to prioritise unanswered research questions (www.phc.ox.ac.uk/infertility). In Table 1, we present our top ten research gaps in reproductive medicine.

Table 1. Ten research gaps in reproductive medicine
Population Interventions
In women with unexplained subfertility What is the effect of IVF versus expectant management
In men with azoospermia and severe oligospermia What is the evidence for available screening tests (e.g. genetic testing, ultrasound, screening for cancer, others)
In women with repeated IVF failure What is the effect of different treatment adjuncts, such as endometrial scratch, hysteroscopy, immunoglobulin, G-CSFa
In women having IVF who are undergoing frozen embryo cycles What is the optimal regimen (e.g. endometrial preparation, triggering, luteal phase support)
In women having IVF who are undergoing final oocyte maturation What is the effect of Gn RhA trigger versus HcG triggera
In men with reduced semen quality What is the most optimal treatment modality in relation to the semen quality (expectant management, IUI, IVF or ICSI)a
In women having IVF What is the effect of cleavage stage versus blastocyst transfer
In women having IVF What is the optimal culture environment
In women having IVF and their children What is the long-term effect (over >20 years)
In women having IVF What is the optimal embryo selection method [morphology, genetic testing (PGS), time-lapse algorithm]?a
  • a HcG, human chorionic gonadoptropin; G-CSF, granulocyte colony stimulating factor; Gn RhA, gonadotropin-releasing hormone agonist; ICSI, intracytoplasmic sperm injection; IUI, intrauterine insemination; IVF, in vitro fertilisation; PGS, pre-implantation genetic screening.

Conclusions

How to introduce new technologies in reproductive medicine responsibly?

With respect to new assisted reproductive technologies, our responsibility extends beyond the couples we treat to the wellbeing of future generations. Under the status quo, however, robust evidence of effectiveness and safety is not a prerequisite for the adoption of novel interventions.5 One solution to the trend for treatment to be delivered in defiance of evidence would be stronger regulatory standards governing the introduction of new therapies in assisted reproductive technology. A study on the acceptability of regulating such techniques found broad support from clinicians, patients, and the general public for the regulation of new reproductive technologies by a national bioethics committee, which would consult with advisors from various backgrounds.13 We are not sympathetic to the argument that a requirement for rigorous prospective treatment evaluation would impede progress in the field, as it is not at all clear that a ‘right to try’ philosophy, where treatments are sold to vulnerable people on a speculative basis, benefits anyone other than the people making the sale, and may well cause harm. Until the time comes when we insist that the generation of evidence precedes the implementation of new technologies, rather than follows it, reproductive medical research will be largely preoccupied with reversing perverse practice. We believe that this continuing need for medical reversal in reproductive medicine is the real enemy of progress.

Disclosure of interests

JW declares that publishing in peer reviewed journals benefits his career. SB reports grants from the Chief Scientist Office, Scotland, during the period of this study. JMND reports grants from The Royal Society of New Zealand, during the conduct of the study. MSK, JM, SR, and MVW have nothing to disclose. AV has received grants from the Human Fertilisation and Embryology Authority, outside the submitted work. CMF declares that she is the coordinating editor of Cochrane Gynaecology and Fertility. Full disclosure of interests available to view online as supporting information.

Contribution to authorship

All authors devised and wrote the commentary and gave approval for submission.

Details of ethics approval

No ethical approval was sought in writing this commentary.

Funding

No specific funding supported the writing of this commentary.