O-113 Effectiveness and treatment cost of assisted reproduction technology for women stimulated by gonadotropin in France: A cohort study using the National Health Database

Study question How effective is Assisted Reproduction Technology (ART) in terms of cumulative live birth rate (CLBR) in France, depending on the gonadotropin used? Summary answer Among 214,539 stimulations, originator follitropin-alfa was associated with significantly higher CLBR when compared to Highly Purified-Human Menopausal Gonadotropin (HP-HMG) and biosimilars. What is known already Deciding which type of gonadotropin to prescribe for a woman undergoing controlled ovarian stimulation (COS) remains difficult. The effectiveness of different gonadotropins is one factor to consider. However, studies comparing r-hFSH-alfa, its biosimilars and HP-HMG are scarce and are mostly based on a single ART treatment cycle and fresh embryo transfers. Some clinical trials have shown similar pregnancy, pregnancy loss, and live birth rates after fresh embryo transfer (ET) between HP-HMG and r-hFSH. However, because more oocytes are retrieved with r-hFSH when compared to HP-HMG, it is logical to hypothesize that the CLBR is higher with r-hFSH. Study design, size, duration A non-interventional study based on the French National Health System (SNDS) database was designed to assess the CLBR and treatment costs from the national payer perspective of four gonadotropin groups (originator follitropin-alfa (r-hFSH-alfa), its biosimilars, HP-HMG and r-hFSH-beta) used for COS cycles leading to oocyte pick-up (OPU) between 01/01/2013 and 31/12/2017 with a follow-up period up to 31/12/2018. The study compared CLBR, with originator r-hFSH-alfa as the reference. Participants/materials, setting, methods Women with COS cycles resulting in OPU with one of the specified gonadotropins were included. Data were extracted from billing and reimbursement records of outpatient healthcare consumption and national hospital discharge databases using a unique, anonymized patient number. CLBR was estimated using an Andersen–Gill model, adjusted for clinical baseline, stimulation and ET variables. Costs were reported as secondary outcomes. Main results and the role of chance 135,752 women (mean age 34.1), underwent 214,539 stimulations leading to OPU and contributed one (61.5%), two (24.8%), three (9.4%) or four (3.2%) COS cycles. COS cycles were stimulated with either Originator r-hFSH-alfa (46%), HP-HMG (29%), r-hFSH-beta (21%) or r-hFSH-alfa biosimilars (4%). Over the study period, CLBR reached 20.5%; 21.9% with originator r-hFSH-alfa, 17.9% with HP-HMG, 21.3% with r-hFSH-beta and 18.4% with r-hFSH-alfa biosimilars. After adjusting for age, pre-treatment, GnRH analog, ovulation triggering, luteal phase support, previous COS, fresh or frozen ET and type of center, as possible cofounding variables, the adjusted hazard ratio (HR) for CLBR (delivery [originator r-hFSH-alfa as reference]) was 0.88 (95% CI 0.86 to 0.95, p < 0.0001) with HP-HMG; 0.98 (95% CI 0.95 to 1.00, p = 0.1020) with r-hFSH-beta, and 0.84 (95% CI 0.79 to 0.90, p < 0.0001) with r-hFSH-alfa biosimilars. Although the mean acquisition cost of r-hFSH-alfa during the study was 33% higher than HP-HMG and 20% higher than r-hFSH-alfa biosimilars, the global ART management costs were only 4% higher than HP-HMG, 3% higher than r-hFSH-beta, and similar to r-hFSH-alfa biosimilars. Limitations, reasons for caution Patients were included only from oocyte pick-up, due to missing data in the SNDS database, meaning that it was not possible to estimate the proportion of cancelled cycles. Furthermore, as r-hFSH-alfa biosimilars were only available since 2015, results for biosimilars should be interpreted with caution. Wider implications of the findings This population-wide French study confirms other Real-World and meta-analysis evidence that CLBR is higher with originator r-hFSH-alfa than with HP-HMG or r-hFSH-alfa biosimilars, respectively, and are relevant for healthcare professionals to support gonadotropin treatment decision making. To further support this, the cost analysis should be completed by a cost-effectiveness analysis. Trial registration number Not applicable