The phenomenon of explaining the cause of poor cycle outcomes: looking under the streetlight, or the role of cognitive bias in Reproductive Endocrinology and Infertility practices. In response to: “Unexpected poor oocyte retrieval: the phenomenon of the borderline response to the Gonadotropin-Releasing Hormone (GnRH) agonist trigger”

We read with great interest the editorial about borderline response (1), defined in part as poor oocyte yield and/or low oocyte maturity in in vitro fertilization cycles triggered with gonadotropin-releasing hormone (GnRH) agonist. Understandably, such cycle outcomes are upsetting to patients and caregivers alike and lead to an immediate search for a culprit. However, the GnRH agonist trigger may be inferior to the human chorionic gonadotropin (hCG) trigger in eliciting full nuclear and cytoplasmic maturation of the oocyte and further follicular disintegration, thus causing faulty final follicular maturation or oocyte retrieval failure. There are many other possibilities that ought to be examined before making a conclusion regarding the cause of poor cycle outcomes. A rushed diagnosis risks clinicians with potential diagnostic errors resulting from cognitive bias. We suspect that this may be the cause of growing “GnRH agonist trigger-phobia.”

There is a large body of literature on decision-making driven by cognitive biases in the medical field. The Society for Maternal-Fetal Medicine has even issued a special statement on the role of cognitive bias and medical error in obstetrics in 2022 (2). Publications exploring cognitive biases specific to the field of reproductive endocrinology and infertility (REI) are lacking. We find that assigning poor oocyte yield and low maturity rates to the type of trigger agent is a good example of cognitive bias at play in REI decision-making.

To determine whether common cognitive errors were at play in this diagnosis, we suggest exploring the following:

• •Anchoring bias: could this assignment of causality be due to the excessive weighting of initial information and the inability to adjust the initial diagnostic hypothesis when further information becomes available? In other words, do we rush to a conclusion only on the basis of the information available at the time of the disappointing oocyte retrieval? Do we neglect the fact that GnRH agonist trigger, more often than not, is used in patients who are high responders with polycystic ovary syndrome. This population has a higher incidence of asynchronous follicular growth, which, in turn, results in a low yield of mature oocytes despite multiple growing follicles. Has the hypothesis of low maturity rate as the result of patient and cycle characteristics rather than of a specific type of trigger been adequately considered? Do we know the incidence of unexpected low yield or low maturity in GnRH agonist triggered cycle compared with other trigger types in this population?
• •Confirmation bias: is the assignment of causality a result of our nature to favor, interpret, or recall information in a way that confirms our pre-existing beliefs? As REI physicians, we pride ourselves on knowing and applying physiology to explain and overcome reproductive challenges. Explaining poor oocyte yield as “GnRH agonist trigger failure” fits well with our belief system and reinforces our confidence in fully understanding reproductive physiology. However, we must acknowledge our knowledge limitations and be open to a reassessment of dogmas as reproductive science develops.
• •Availability bias: could this assignment of causality be due to relying on what immediately comes to mind? Certainly, the only things between the trigger administration and oocyte number are the triggering agent and the technical aspects of the retrieval procedure. However, have we accounted for the base rate of the poor oocyte yield? Although the literature is sparse on low yield, there is a body of literature on empty follicle syndrome (EFS), possibly a related phenomenon. Stevenson and Lashen (3), in their 2008 review, reported that 67% of EFS cases occurred as a result of human error. Notably, this was in the era before the routine use of GnRH agonists for final follicular maturation triggering, suggesting a significant baseline rate of this condition unrelated to utilization of GnRH agonist. Furthermore, Castillo et al. (4) reported a similar incidence of EFS after GnRH agonist and hCG triggers. These reports further cast doubt on the proposed linkage between the type of trigger and borderline response.

Decision-making under uncertainty is an important concept in behavioral sciences that can be applied to clinical practice in reproductive endocrinology and specifically in decisions related to triggering final follicular maturation. A good example of this concept is the practice of measuring the serum hormone levels 12 hours after trigger injection, which is highly controversial and depends on individual clinic practices and provider comfort level with uncertainty and risk of rare events such as incidence of EFS. Our survey of practices regarding posttrigger serum hormone testing in Canada revealed that only half of the clinics follow up on the serum luteinizing hormone levels, whereas almost none follow the hCG levels (5).

We agree that the clinical observation of unexpectedly low oocyte yield and low maturity in in vitro fertilization cycles warrants further investigations, yet we believe that it is premature to declare a causative effect between the type of trigger and this phenomenon. We find that the editorial by Paulson (1) and the results of our Canadian practice survey both highlight the mistrust of our REI community in the effectiveness of the GnRH agonist trigger and suggest further exploration into the roles of cognitive biases in REI practice.

CRediT Authorship Contribution Statement

Jennia Michaeli: Writing – original draft, Conceptualization, Writing – review & editing. Heather Shapiro: Writing – review & editing, Writing – original draft, Conceptualization.

Declaration of Interests

J.M. has nothing to disclose. H.S. has nothing to disclose.