P-238 Non-interventional study comparing the effectiveness and safety of r-hFSH:r-hLH versus r-hFSH in women aged 35-42 years with normal ovarian reserve undergoing GnRH antagonist ART treatment

Abstract

Abstract Study question Is ovarian stimulation (OS) with r-hFSH:r-hLH more effective than OS with r-hFSH alone in women aged 35-42 with normal ovarian reserve treated with ART? Summary answer Compared to r-hFSH alone, OS with r-hFSH:r-hLH for ART treatment was associated with a significant increase in clinical pregnancy and live birth per initiated cycle. What is known already Meta-analyses and secondary analyses of clinical data provide evidence that OS during ART treatment with r-hFSH:r-hLH is associated with improved reproductive outcomes when compared to r-hFSH alone in various patient groups with: a) advanced maternal age , b) low ovarian reserve or , c) hypo response to OS. Other subgroups of patients might also benefit of the use of r-hFSH:r-hLH during OS for ART treatment. We aimed to determine if, in a real-world setting, OS with r-hFSH:r-hLH is associated with improved clinical benefit when compared to r-hFSH alone in patients aged 35-42 years with normal ovarian reserve. Study design, size, duration A non-interventional study including women aged 35-42 years with normal ovarian reserve biomarkers (menstrual cycle day 2 or 3 FSH<12IU/L or 9<AFC<20 with a 2<diameter<10 mm, or AMH ≥1.2ng/ml) undergoing OS for ART treatment (fresh transfer only; GnRH antagonist protocol) in 12 French centers between 1/1/2008 and 31/12/2016, with a follow-up period up to 31/12/2017. The analysis included two cohorts: a) 4,323 OS cycles with r-hFSH alone, and b) 1,070 OS cycles with r-hFSH:r-hLH. Participants/materials, setting, methods Anonymized data were extracted from the Retrospective Multicenter Statistical database, including 12 French centers. A generalized linear model was used to address lack of randomization. The fixed effect part included age, an ovarian reserve estimator and stimulation; the random effect included age, AMH, AFC, FSH, cause of infertility and number of attempts, center and patient effect. Clusters (matched subsets) were built based on the propensity score by using a two-stage algorithm. Main results and the role of chance We included in our analysis 4,472 women receiving ART treatment after OS with r-hFSH (mean age: 37.8 years; 4,323 cycles) or with r-hFSH:r-hLH (mean age: 38.1 years; 1,070 cycles). In the r-hFSH cohort the mean AMH level was 2.81 ng/ml and mean AFC was 12.0, in the r-hFSH:r-hLH cohort they were 2.14 ng/ml and 10.0, respectively. The mean total r-hFSH dose were 2,275 IU (r-hFSH cohort) and 3,002 IU (r-hFSH:r-hLH cohort). Mean ± SD number of oocytes were 8.1 ± 5.7 (r-hFSH cohort) and 6.1 ± 4.3 (r-hFSH:r-hLH cohort); mean ± SD number of embryos were 4.0 ± 3.4 (r-hFSH cohort) and 3.3 ± 3.2 (r-hFSH:r-hLH cohort). The unadjusted ongoing pregnancy rate reached 17.1% in the r-hFSH cohort and 20.4% in the r-hFSH:r-hLH cohort, whilst the unadjusted live birth rate reached 17.0% and 20.2%, respectively. Unadjusted pregnancy failure rate was 6.6% and 3.7%, respectively. After adjusting for confounders, when compared to OS with r-hFSH, OS with rhFSH:rhLH was associated with a higher clinical pregnancy rate (Risk ratio,RR 1.232; 95% CI 1.058 to 1.436; p-value = 0.007) and a higher live birth rate (RR 1.452; 95% CI 1.227-1.1719, p-value <0.001) per initiated cycle. Limitations, reasons for caution Since this was a real-world data study, no random assignment of stimulation was performed, and some residual confounding is possible due to unmeasured confounders. Live births resulting from further transfers of frozen embryos were not included in the analysis. Wider implications of the findings This study demonstrated that OS with r-hFSH:r-hLH improves live birth rate in women > 35 years old with a normal ovarian reserve, compared to r-hFSH only. Sound RWE studies can be combined with RCTs to guide clinical and payer’s decision making, paving the way to precision medicine. Trial registration number Not applicable