Protocol change improves live birth and recurrent cycle cancellation rates after a previous IVF cycle cancellation: an analysis of 13 000 autologous cycles reported to SART CORS

Abstract STUDY QUESTION After an IVF cycle cancellation, does changing the stimulation protocol affect the odds of live birth and recurrent cancellation in the subsequent cycle? SUMMARY ANSWER After IVF cycle cancellation, compared to those who repeated the same stimulation protocol, those who changed their protocol had higher odds of live birth and lower odds of recurrent cycle cancellation. WHAT IS KNOWN ALREADY There is limited data addressing the effect of changing the stimulation protocol after an IVF cycle is cancelled during initial stimulation. The odds of live birth outcomes are not known so far in studies addressing the effect of changing the protocol. STUDY DESIGN, SIZE, DURATION Retrospective Cohort Study using the 2014–2017 Society for Assisted Reproductive Technology Clinic Outcome Reporting System (SART CORS) database PARTICIPANTS/MATERIALS, SETTING, METHODS The data included 13 135 patients with a first autologous IVF cycle that resulted in a cycle cancellation and was followed by a second autologous cycle within the study period. We excluded fertility preservation cycles, supernumerary cycle attempts after the second IVF cycle attempt, and cycles with more than one stimulation protocol documented per cycle start. Patients who received the same protocol for both cycles (n = 6434) were compared to those who changed their protocol in the second cycle (n = 6701). Multivariable logistic regression analyses were performed to estimate the adjusted odds of live birth and recurrent cancellation. MAIN RESULTS AND THE ROLE OF CHANCE Changing the protocol in the second cycle resulted 14% lower odds of recurrent cycle cancellation (P = 0.01) and 17% higher odds of live birth after fresh transfers (P = 0.04). When stratifying the data by specific combinations of protocol change (agonist flare, agonist suppression, antagonist), there was an increase in live birth when switching from antagonist to agonist suppression (odds ratio (OR) = 1.36, P = 0.03) and from agonist suppression to antagonist (OR = 1.73, P = 0.01) compared to those who repeated their same stimulation protocol. Specifically in poor responders, outcomes were worse when using the agonist flare protocol and significantly improved with the agonist suppression protocol. LIMITATIONS, REASONS FOR CAUTION Comparison of response to stimulation between first and second cycles cannot be made in this study because the index IVF cycle was cancelled during ovarian stimulation, and thus there is no reportable outcome data for that cycle. Additionally, SART only tracks the three stimulation protocols addressed in this study and does not have data on more contemporary protocols that are used in poor responders thus limiting the generalizability of our findings WIDER IMPLICATIONS OF THE FINDINGS Using the SART CORS database, which includes >90% of all reported IVF cycles in the USA, provides generalizability to the demographically diverse IVF populations found here. In agreement with prior studies assessing change in IVF protocols, the agonist flare protocol seems to result in worse IVF outcomes, and based on our results, we believe that there is no role for the agonist flare protocol in patients with a prior poor response to stimulation. STUDY FUNDING/COMPETING INTEREST(S) None declared. TRIAL REGISTRATION NUMBER N/A.