Corifollitropin alfa followed by highly purified HMG versus recombinant FSH in young poor ovarian responders: a multicentre randomized controlled clinical trial

STUDY QUESTION: Does administration of corifollitropin alfa followed by highly purified (hp) HMG result in higher ongoing pregnancy rates compared with daily recombinant FSH (rFSH) in young poor responders? SUMMARY ANSWER: Corifollitropin alfa followed by hp-HMG does not increase ongoing pregnancy rates compared with rFSH in young poor responders, although more supernumerary cryopreserved embryos were obtained with corifollitropin alfa and hp-HMG. WHAT IS KNOWN ALREADY: Poor ovarian response remains one of the main therapeutic challenges in women undergoing ovarian stimulation, given that very low live birth rates of 6% have been reported in this particular group of infertile patients. Nevertheless, concerns have been raised that a degree of heterogeneity remains, as the prognostic effect of individual factors is still unclear, particularly for the young poor responder group. The rationale for conducting the current randomized trial was based on the results of a previous pilot study demonstrating promising results with the administration of hp-HMG following corifollitropin alpha in women younger than 40 years of age, fulfilling the 'Bologna' criteria. STUDY DESIGN, SIZE, DURATION: A multicenter, phase III, superiority, randomized trial was conducted using a parallel two-arm design. The study included 152 patients younger than 40 years old and fulfilling the 'Bologna' criteria for poor ovarian response, from one tertiary referral centre in Europe and one tertiary referral centre in Asia. Enrolment was performed from March 2013 to May 2016. PARTICIPANTS/MATERIALS, SETTING, METHODS: Eligible patients were randomized to either administration of 150 mu g corifollitropin alfa followed by 300 IU hp-HMG (Group A) or to 300 IU of daily recombinant FSH (Group B) in a fixed GnRH antagonist protocol. The randomization sequence was created using a computer generated randomization list stratified by centre, using 1: 1 allocation. The primary outcome was ongoing pregnancy rate (defined as the presence of an intrauterine gestational sac with an embryonic pole demonstrating cardiac activity at 9-10 weeks of gestation). Secondary outcomes included embryo cryopreservation rates, clinical and biochemical pregnancy rates and number of oocytes retrieved. MAIN RESULTS AND THE ROLE OF CHANCE: Overall, 152 poor ovarian responders defined by the 'Bologna' criteria were included in the study. Using an intention-to treat analysis, the ongoing pregnancy rates did not differ significantly between Group A 11/77 (14.3%) and Group B 11/70 (15.7%), absolute difference: -0.4 (-11.5 to 10.8), OR = 0.9 (0.4-2.4). Biochemical and clinical pregnancy rates, live birth rates and the number of oocytes retrieved were also comparable between the two groups. Nevertheless, more patients in the corifollitropin alfa group had cryopreserved embryos compared to the rFSH group [22 ( 28.6%) versus 10 ( 14.3%), OR = 2.4 (1.01-5.5)]. Incidentally, Asian patients had significantly lower cancellation rates compared to European poor responders [2/64 (3.1%) versus 17/ 83 (20.4%), OR = 0.12 ( 0.03-0.5)]. This discrepancy could be explained by the fact that Asian women were better prognosis patients than European patients, with significantly lower FSH [9.8 (5.3) versus 11.5 (5.4), P = 0.017] and significantly higher AMH [1.1 (0.9) versus 0.4 (0.3), P-value < 0.001] levels. LIMITATIONS, REASONS FOR CAUTION: Ongoing pregnancy rates close to 14% for both treatment groups differ significantly from the hypothesized primary outcome rates used in the power calculation. Therefore, our randomized trial might have been underpowered to detect smaller differences. The use of multiple secondary outcomes and multiple comparisons could have increased a Type 1 error. Finally, although the chance of selection biases remains low given the nature of the infertile population, the open-label design could have been a limitation. WIDER IMPLICATIONS OF THE FINDINGS: Poor ovarian response represents a challenge and although a specific protocol may have increased the number of cryopreserved embryos, no difference was observed in ongoing pregnancy rates. Our study, being one of the largest RCTs in 'Bologna' criteria poor responders, highlights that baseline characteristics may play a crucial role in clinical prognosis of this population. Given that ovarian stimulation using novel protocols does not seem to significantly increase pregnancy rates even in young women, we suggest that future clinical research should focus on increasing the number of recruitable follicles and on oocyte quality rather than evaluating different stimulation protocols.