Future ovarian stimulation strategies should avoid maximizing oocyte yield, but aim at generating a sufficient number of chromosomally normal embryos by reduced interference with ovarian physiology.

No abstract

Replacing GnRH agonist cotreatment for the prevention of a premature rise in LH during ovarian stimulation for in vitro fertilization (IVF) by the late follicular phase administration of GnRH antagonist may render supplementation of the luteal phase redundant, because of the known rapid recovery of pituitary function after antagonist cessation. This randomized two-center study was performed to compare nonsupplemented luteal phase characteristics after three different strategies for inducing final oocyte maturation. Forty patients underwent ovarian stimulation using recombinant (r-)FSH (150 IU/d, fixed) combined with a GnRH antagonist (antide; 1 mg/d) during the late follicular phase. When at least one follicle above 18 mm was observed, patients were randomized to induce oocyte maturation by a single injection of either r-human (h)CG (250 microg) (n = 11), r-LH (1 mg) (n = 13), or GnRH agonist (triptorelin; 0.2 mg) (n = 15). Retrieved oocytes were fertilized by either IVF or intracytoplasmatic sperm injection, depending on sperm quality. Embryo transfer was performed 3-4 d after oocyte retrieval. No luteal support was provided. Serum concentrations of FSH, LH, estradiol (E(2)), progesterone (P), and hCG were assessed at fixed intervals during the follicular and luteal phase. The median duration of the luteal phase was 13, 10, and 9 d for the r-hCG, the r-LH, and the GnRH agonist group, respectively (P = 0.005). The median area under the curve per day (from 4 d post randomization until the onset of menses) for LH was 0.50, 2.34, and 1.07 for the r-hCG, the r-LH, and the GnRH agonist group, respectively (P = 0.001). The median area under the curve per day for P was 269 vs. 41 and 16 for the r-hCG, the r-LH, and the GnRH agonist group, respectively (P < 0.001). Low pregnancy rates (overall, 7.5%; range, 0-18% per started cycle) were observed in all groups. In conclusion, the nonsupplemented luteal phase was insufficient in all three groups. In the patients receiving r-hCG, the luteal phase was less disturbed, compared with both other groups, presumably because of prolonged clearance of hCG from the circulation and the resulting extended support of the corpus luteum. Despite high P and E(2) concentrations during the early luteal phase in all three groups, luteolysis started prematurely, presumably because of excessive negative steroid feedback resulting in suppressed pituitary LH release. Hence, support of corpus luteum function remains mandatory after ovarian stimulation for IVF with GnRH antagonist cotreatment.

STUDY QUESTION Is the presence or absence of certain vaginal bacteria associated with failure or success to become pregnant after an in vitro fertilization (IVF) or IVF with intracytoplasmic sperm injection (IVF-ICSI) treatment? SUMMARY ANSWER Microbiome profiling with the use of interspace profiling (IS-pro) technique enables stratification of the chance of becoming pregnant prior to the start of an IVF or IVF-ICSI treatment. WHAT IS KNOWN ALREADY Live-birth rates for an IVF or IVF-ICSI treatment vary between 25 and 35% per cycle and it is difficult to predict who will or will not get pregnant after embryo transfer (ET). Recently, it was suggested that the composition of the vaginal microbiota prior to treatment might predict pregnancy outcome. Analysis of the vaginal microbiome prior to treatment might, therefore, offer an opportunity to improve the success rate of IVF or IVF-ICSI. STUDY DESIGN, SIZE, DURATION In a prospective cohort study, 303 women (age, 20–42 years) undergoing IVF or IVF-ICSI treatment in the Netherlands were included between June 2015 and March 2016. PARTICIPANTS/MATERIALS, SETTING, METHODS Study subjects provided a vaginal sample before the start of the IVF or IVF-ICSI procedure. The vaginal microbiota composition was determined using the IS-pro technique. IS-pro is a eubacterial technique based on the detection and categorization of the length of the 16S–23S rRNA gene interspace region. Microbiome profiles were assigned to community state types based on the dominant bacterial species. The predictive accuracy of the microbiome profiles for IVF and IVF-ICSI outcome of fresh ET was evaluated by a combined prediction model based on a small number of bacterial species. From this cohort, a model was built to predict outcome of fertility treatment. This model was externally validated in a cohort of 50 women who were undergoing IVF or IVF-ICSI treatment between March 2018 and May 2018 in the Dutch division of the MVZ VivaNeo Kinderwunschzentrum Düsseldorf, Germany. MAIN RESULTS AND THE ROLE OF CHANCE In total, the vaginal microbiota of 192 women who underwent a fresh ET could be analysed. Women with a low percentage of Lactobacillus in their vaginal sample were less likely to have a successful embryo implantation. The prediction model identified a subgroup of women (17.7%, n = 34) who had a low chance to become pregnant following fresh ET. This failure was correctly predicted in 32 out of 34 women based on the vaginal microbiota composition, resulting in a predictive accuracy of 94% (sensitivity, 26%; specificity, 97%). Additionally, the degree of dominance of Lactobacillus crispatus was an important factor in predicting pregnancy. Women who had a favourable profile as well as <60% L. crispatus had a high chance of pregnancy: more than half of these women (50 out of 95) became pregnant. In the external validation cohort, none of the women who had a negative prediction (low chance of pregnancy) became pregnant. LIMITATIONS, REASONS FOR CAUTION Because our study uses a well-defined study population, the results will be limited to the IVF or IVF-ICSI population. Whether these results can be extrapolated to the general population trying to achieve pregnancy without ART cannot be determined from these data. WIDER IMPLICATIONS OF THE FINDINGS Our results indicate that vaginal microbiome profiling using the IS-pro technique enables stratification of the chance of becoming pregnant prior to the start of an IVF or IVF-ICSI treatment. Knowledge of their vaginal microbiota may enable couples to make a more balanced decision regarding timing and continuation of their IVF or IVF-ICSI treatment cycles. STUDY FUNDING/COMPETING INTEREST(S) This study was financed by NGI Pre-Seed 2014–2016, RedMedTech Discovery Fund 2014–2017, STW Valorisation grant 1 2014–2015, STW Take-off early phase trajectory 2015–2016 and Eurostars VALBIOME grant (reference number: 8884). The employer of W.J.S.S.C. has in collaboration with ARTPred acquired a MIND subsidy to cover part of the costs of this collaboration project. The following grants are received but not used to finance this study: grants from Innovatie Prestatie Contract, MIT Haalbaarheid, other from Dutch R&D tax credit WBSO, RedMedTech Discovery Fund, (J.D.d.J.). Grants from Ferring (J.S.E.L., K.F., C.B.L. and J.M.J.S.S.), Merck Serono (K.F. and C.B.L.), Dutch Heart Foundation (J.S.E.L.), Metagenics Inc. (J.S.E.L.), GoodLife (K.F.), Guerbet (C.B.L.). R.K. is employed by ARTPred B.V. during her PhD at Erasmus Medical Centre (MC). S.A.M. has a 100% University appointment. I.S.P.H.M.S., S.A.M. and A.E.B. are co-owners of IS-Diagnostics Ltd. J.D.d.J. is co-owner of ARTPred B.V., from which he reports personal fees. P.H.M.S. reports non-financial support from ARTPred B.V. P.H.M.S., J.D.d.J. and A.E.B. have obtained patents `Microbial population analysis’ (9506109) and `Microbial population analysis’ (20170159108), both licenced to ARTPred B.V. J.D.d.J. and A.E.B. report patent applications `Method and kit for predicting the outcome of an assisted reproductive technology procedure’ (392EPP0) and patent `Method and kit for altering the outcome of an assisted reproductive technology procedure’ by ARTPred. W.J.S.S.C. received personal consultancy and educational fees from Goodlife Fertility B.V. J.S.E.L. reports personal consultancy fees from ARTPred B.V., Titus Health B.V., Danone, Euroscreen and Roche during the conduct of the study. J.S.E.L. and N.G.M.B. are co-applicants on an Erasmus MC patent (New method and kit for prediction success of in vitro fertilization) licenced to ARTPred B.V. F.J.M.B. reports personal fees from Advisory Board Ferring, Advisory Board Merck Serono, Advisory Board Gedeon Richter and personal fees from Educational activities for Ferring, outside the submitted work. K.F. reports personal fees from Ferring (commercial sponsor) and personal fees from GoodLife (commercial sponsor). C.B.L. received speakers’ fee from Ferring. J.M.J.S.S. reports personal fees and other from Merck Serono and personal fees from Ferring, unrelated to the submitted paper. The other authors declare that they have no competing interests. TRIAL REGISTRATION NUMBER ISRCTN83157250. Registered 17 August 2018. Retrospectively registered.

In a first feasibility study, the efficacy and safety of a single dose of recombinant long-acting FSH (FSH-CTP) were investigated in in vitro fertilization (IVF) patients undergoing controlled ovarian stimulation with a flexible GnRH antagonist protocol. Eligible subjects were randomized to receive a single dose of 120 micro g (n = 25), 180 microg (n = 24), or 240 microg (n = 25) corifollitropin alfa (FSH-CTP) or to start daily fixed doses of 150 IU recombinant FSH (rFSH) (n = 24, reference). Subjects who received a single dose of FSH-CTP continued 1 wk after injection (treatment d 8) with fixed daily doses of 150 IU rFSH (Puregon/Follistim) until the day of triggering final oocyte maturation. The terminal half-life of FSH-CTP was, on average, 65 h and dose independent. Cycle cancellation before human chorionic gonadotropin (hCG) administration occurred in only three subjects treated with FSH-CTP. The median duration of stimulation was 10.0 d in each FSH-CTP group and 9.0 d in the daily rFSH group. The total number of follicles at least 11 mm at stimulation d 8 and at the day of hCG administration tended to increase with dose of FSH-CTP, although a significant dose-response relationship was revealed only for the number of follicles at least 15 mm on the day of hCG (P = 0.03). Serum estradiol levels and inhibin-B levels were not significantly different between the four groups on d 8 and on the day of hCG. In total, 12 subjects (17.6%) in the FSH-CTP groups and two subjects (8.3%) in the rFSH group experienced a premature LH rise (defined as LH >or= 10 IU/liter) before the start of the GnRH antagonist (P value not significant between groups). This relatively high incidence of women demonstrating an early LH rise in the FSH-CTP groups may be related to the higher initial rises of serum estradiol and the use of a flexible GnRH antagonist protocol. The mean number of oocytes recovered per started cycle was higher in FSH-CTP-treated subjects compared with rFSH-treated subjects (significant at P = 0.03 for the 240- microg FSH-CTP group), but no difference could be noted between the number of good quality embryos (range of means, 3.8-4.8 per attempt), and equal numbers of embryos were available for embryo transfer. In summary, FSH-CTP appeared to be a potent inducer of multiple follicular growth; additional research will be needed to select the optimal FSH-CTP dose and treatment time interval.