Conclusions about the effects on fertilization of time from aspiration to incubation and blood in the aspirate depend on the use of appropriate statistical techniques

Objective: To examine the effects of exposure to blood and time from aspiration to incubation on oocyte fertilization. Design: Retrospective, cohort study. Setting: Hospital-based IVF practice. Patient(s): Eighty-two women undergoing IVF procedures. Intervention(s): Patients received a standard gonadotropin protocol or a microdose flare protocol. Main Outcomes Measure(s): Logistic regression was used to estimate odds of nonfertilization, adjusting for maternal age, time from aspiration to incubation, presence of blood in the aspirate, and a time-by-blood interaction. Two statistical methods were used: an analysis that assumed oocytes were independent and an analysis that adjusted for the lack of independence of oocytes within mothers (i.e., cluster-correlated data). Result(s): The 82 women contributed 1,093 oocytes. In a model with blood dicotomized as present or absent, and with time elapsed and maternal age categorized into quartiles, both time and maternal age were statistically significant when conventional logistic regression was used. In analyses that accounted for clustering within mother, retrieval-to-incubation times more than 4 minutes were associated with an increased risk of nonfertilization, but age and blood were not; there was no evidence of an interaction between time elapsed and blood. Conclusion(s): If appropriate statistics are applied, retrieval-to-incubation times more than 4 minutes are associated with increased risk of nonfertilization.