Predictors of oocyte yield in controlled ovarian hyperstimulation IVF/ICSI cycles: a retrospective analysis in a tertiary care centre

Background: The most important predictor of oocyte yield in ART cycles is female age, but other biochemical and ultrasonographic markers done before controlled ovarian stimulation may predict the oocyte yield in women undergoing COS in IVF cycles.Methods: The main aim of the study was to evaluate ovarian reserve markers before COS which can help to individualise treatment protocols to achieve optimal response and minimise risk of complications. It is retrospective observational study, 1924 women undergoing COS in IVF/ICSI cycles in tertiary care centre in India, from January 2010 to June 2017 were included. Results: Univariate analysis showed that age, D2FSH, AMH, D2AFC and E2 on the day of trigger were significant predictors of oocyte yield (p<0.05). E2 on day of trigger with ROC (0.81), indicating good discriminating potential for predicting poor ovarian response, followed by age and D2 FSH. The formula to calculate, number of oocytes retrieved=18.46+(0.174×AFC)+(0.092×AMH)-(0.123×age)-(1.19×FSH), FSH was formulated, with r2=0.2486 (p<0.001). ROC curve analysis shows that FSH has statistically significant discriminability to detect poor response than age [AUC (95% CI) FSH 0.77 (0.74, 0.81), age 0.56 (0.52, 0.60), (p<0.05)]. FSH >7.82 IU/ml was predictor of poor response (sensitivity 78.13%, specificity 79.53%).Conclusions: A combination of predictors demonstrated superior ability of predicting oocyte yield after controlled ovarian stimulation than compared with any single endocrine marker. D2 FSH though thought to be obsolete, but we found significant predictive ability in terms of oocyte yield in the Indian population.