Ovarian stimulation protocols: don’t immediately change a losing team

“Never change a winning team” is an often-used saying in football, attributed to Sir Alfred Ernest “Alf” Ramsey, who guided England to their only FIFA World Cup victory in 1966. In assisted reproductive technology (ART), based on the results of the study by Wald et al. (1

Wald K, Hariton E, Morris JR, Chi EA, Jaswa EG, Cedars MI, et al. Changing stimulation protocol on repeat conventional ovarian stimulation cycles does not lead to improved laboratory outcomes. Fertil Steril. In press.

), this may be modified to, “Don’t immediately change a losing team.” In the results-driven world of sports, there is quick pressure from fans to change the team or sack the manager after short periods without success. In the results-driven world of ART, patients may demand to “change the team” after an unsuccessful in vitro fertilization (IVF) cycle, and change means hope. Patients may have an easier time regaining an optimistic view of their reproductive future if they are told that their next cycle includes a “magic” new strategy in the form of a different stimulation protocol that will change everything. Does the evidence support this approach?In their single-center retrospective cohort study, Wald et al. (1

Wald K, Hariton E, Morris JR, Chi EA, Jaswa EG, Cedars MI, et al. Changing stimulation protocol on repeat conventional ovarian stimulation cycles does not lead to improved laboratory outcomes. Fertil Steril. In press.

) analyzed laboratory outcome parameters from 4,458 patients undergoing >1 IVF cycle in 1 year. Approximately half of the patients (49%) completed another cycle with the same protocol, whereas the other half (51%) underwent the second stimulation with a different protocol. Overall, the investigators found a minor, but statistically significant improvement in certain laboratory outcomes in the group where the same approach was used a second time, including a statistically significant increase in the number of usable embryos (1

Wald K, Hariton E, Morris JR, Chi EA, Jaswa EG, Cedars MI, et al. Changing stimulation protocol on repeat conventional ovarian stimulation cycles does not lead to improved laboratory outcomes. Fertil Steril. In press.

).

The limitations of the study include the lack of data regarding pregnancy outcomes and its retrospective single-center nature. Potential confounding factors include other changes from the first to the second stimulation, such as changes in the stimulation medication dosages and the use of adjuvant medications, although these changes may have occurred to a similar extent in both groups. The arbitrary nonrandomized nature of the decision on whether to change the protocol or stay on the course is another source of bias. The group of patients where a protocol change was instituted may have fundamentally different characteristics from the group where the same protocol was repeated. However, the study provides “real-world” information on a large number of patients and cycles, and provides an incentive to question our clinical decision-making with regards to controlled ovarian hyperstimulation protocols.

When Edwards et al. (2Edwards R.G. Steptoe P.C. Purdy J.M. Establishing full-term human pregnancies using cleaving embryos grown in vitro.) published their original case series on the establishment of human pregnancies by the use of fertilization in vitro, they noticed a cluster of “three abortive pregnancies in patients given gonadotrophins to stimulate the maturation of oocytes.” Given the “comparative failure of attempts to establish pregnancy following the use of human menopausal gonadotropin,” they forecasted that “this method may have been abandoned permanently in favor of monitoring the natural menstrual cycle.” Given their enormous contributions to the field of reproductive medicine, these pioneers may be forgiven for their incorrect prophecy; it was later discovered that the correction of stimulation-induced alterations of the luteal phase could remedy the observed phenomenon.Controlled ovarian hyperstimulation is now an integral part of a contemporary IVF cycle, along with some form of preceding follicular synchronization, prevention of premature ovulation during stimulation, induction of oocyte maturation, egg retrieval, and luteal support if a fresh transfer is planned. Over the last four decades, various ovarian stimulation protocols have been introduced, differing primarily in the methods for ovarian downregulation and the prevention of premature ovulation. Both are achieved via various forms of manipulation of the hypothalamic-pituitary-ovarian axis, taking advantage of the discovery of gonadotropin-releasing hormone (GnRH) in the 1970s that set the stage for the development of GnRH agonists and antagonists in the 1980s (3Pharmacology of medications used for ovarian stimulation.). The most commonly used protocols include GnRH agonist-based protocols (long, short, or flare) and GnRH antagonist-based protocols with various options for follicular synchronization before stimulation, including estrogen priming.Given the diverse population of patients undergoing IVF, the complex multifactorial nature of assisted reproduction cycles, and the fine but significant differences in the execution of each protocol between centers, comparing stimulation protocols has proven to be a challenge. There has been a trend toward increased use of GnRH antagonist-based protocols over the last decade, given that equivalent live birth outcomes compared with agonist-based protocols can be achieved with significantly lower rates of ovarian hyperstimulation syndrome (4Al-Inany H.G. Youssef M.A. Aboulghar M. Broekmans F. Sterrenburg M. Smit J. et al.GnRH antagonists are safer than agonists: an update of a Cochrane review.). The latter finding is principally because of the ability to use the shorter-acting GnRH agonist trigger for final oocyte maturation in the setting of a GnRH-naïve pituitary gland.In daily clinical practice, fertility providers encounter patients across the whole spectrum of ovarian reserve characteristics, with vastly different responses to stimulation, requiring the individualization of controlled ovarian hyperstimulation protocols. The cohort studied by Wald et al. (1

Wald K, Hariton E, Morris JR, Chi EA, Jaswa EG, Cedars MI, et al. Changing stimulation protocol on repeat conventional ovarian stimulation cycles does not lead to improved laboratory outcomes. Fertil Steril. In press.

) was an “all-comer” sample, precluding conclusions regarding the value of protocol changes in particular subpopulations of patients.

One plausible interpretation of their overall findings is that we are simply overestimating the impact of the ovarian stimulation protocol; what matters is giving patients another opportunity to get pregnant, more so than the particular protocol. If the same protocol is chosen, the experience effect of doing another cycle on the side of the patient, and the knowledge of prior response to this particular protocol on the clinician side may have a larger beneficial effect than a protocol change would have.

The well-known tendency to obtain measurements closer to the mean after an outlier observation, known as regression to the mean, may be the most important statistical concept to know for reproductive endocrinology and infertility specialists. Countless studies in our field have been published where a cohort of unsuccessful patients with a live birth rate of zero received a novel intervention, leading to a live birth rate unequal to zero in the next treatment attempt. Regression to the mean renders this type of study meaningless. Outside the research setting, the consumer nature of our field turns this statistical phenomenon into a threat to evidence-based practice: the perpetual unindicated use of vaginal progesterone in the setting of unexplained recurrent pregnancy loss is a prime example.

However, counseling a patient not to take a certain medicine or not to change a protocol takes more time than the opposite action. Also, helping our patients make excellent decisions may be the most rewarding, but also the most challenging aspect of our field. Every day, we are confronted with questions, such as, “How many more intrauterine insemination attempts do you recommend before doing IVF?” or “After two unsuccessful retrieval cycles, should I try one more or use an egg donor?” or “After two failed euploid embryo transfers, should I use a gestational carrier?” The answers to these questions are usually not black or white, but come in shades of gray, and require a lot of patience and counseling. Whereas a surgeon admitting a patient with appendicitis can more or less tell the patient “sign here” and proceed with the surgery, our field is filled with shared patient-physician judgment calls. Ultimately, odds and probabilities only have a limited ability to assist with decision-making. Brian Zikmund-Fisher, Professor of Health Behavior and Health Education, experienced a major health crisis as a graduate student for behavioral decision theory (5What the odds fail to capture when a health crisis hits. National Public Radio 2014.). He was faced with the decision of whether to undergo a bone marrow transplant for myelodysplastic syndrome, which would cure him of the disease and increase his life expectancy tremendously, but is also associated with a 25%–30% mortality. He concluded from this experience and from his research that probabilities, while useful, are quite limited in their ability to predict what will happen to any one person.

However, research data remain the best tool to counsel our patients based on available evidence. Sometimes the data tell us to try treating our patients again with the same approach, hoping that it will become a “winning team.”

References

Wald K, Hariton E, Morris JR, Chi EA, Jaswa EG, Cedars MI, et al. Changing stimulation protocol on repeat conventional ovarian stimulation cycles does not lead to improved laboratory outcomes. Fertil Steril. In press.

Edwards R.G. Steptoe P.C. Purdy J.M.

Establishing full-term human pregnancies using cleaving embryos grown in vitro.

Br J Obstet Gynaecol. 87: 737-756

Pharmacology of medications used for ovarian stimulation.

Best Pract Res Clin Endocrinol Metab. 33: 21-33Al-Inany H.G. Youssef M.A. Aboulghar M. Broekmans F. Sterrenburg M. Smit J. et al.

GnRH antagonists are safer than agonists: an update of a Cochrane review.

Hum Reprod Update. 17: 435

What the odds fail to capture when a health crisis hits. National Public Radio 2014.

()Article InfoPublication History

Published online: July 11, 2021

Publication stageIn Press Corrected ProofFootnotes

You can discuss this article with its authors and other readers at https://www.fertstertdialog.com/posts/33335