In vitro fertilization: investigating the risk of ischemic placental disease and novel methods for quantifying success

• Main Author: Modest, Anna Merport
• Other Authors: Wise, Lauren A.
• Language: en_US
• Published: 2018
• Subjects: Epidemiology
• Online Access: https://hdl.handle.net/2144/27816

The use of in vitro fertilization (IVF) has increased over the last several decades in the United States. The influence of IVF on pregnancy complications is not well understood and current methods for quantifying overall success of IVF procedures may be flawed. The aims of this dissertation were to evaluate the extent to which various aspects of IVF treatment are associated with the risk of ischemic placental disease (IPD, defined as preeclampsia, placental abruption, or small for gestational age infant) and to apply novel methods for measuring live birth after IVF success. We used data from a large tertiary care center and an affiliated infertility treatment center. In Study 1, we found an increased risk of IPD in pregnancies conceived by donor and autologous IVF compared with non-IVF pregnancies (risk ratio (RR): 2.5, 95% confidence interval (CI): 2.1-3.1 and RR 1.8, 95% CI: 1.7-2.0, respectively), and the risks were consistently higher for donor IVF pregnancies than autologous IVF pregnancies. In Study 2, we found an increased risk of IPD in pregnancies in the highest and middle tertiles of serum progesterone levels among autologous IVF cycles compared with the lowest tertile (RR 2.0, 95% CI 1.3-3.3 and RR 1.6, 95% CI 0.6-2.6, respectively), although the results were not imprecise. In study 3, we used inverse probability-of-censoring weighting (IPCW) to address a potential violation of the uninformative censoring assumption of the Kaplan-Meier (KM) survival analysis to calculate the cumulative incidence of live birth after multiple cycles of IVF. The two approaches were similar (cumulative incidence of live birth 69.1% using IPCW and 73.9% using KM). However, additional information is needed to calculate better IPCW weights, which may be more important when investigating exposure/outcome relationships. Our results suggest that women undergoing IVF, particularly donor IVF, might benefit from counseling about the increased risk of IPD. Our results also suggest that IPCW methods offer an improvement over other methods for validly estimating cumulative incidence of live birth across multiple IVF cycles. The routine application of IPCW methods to estimating incidence of live birth in epidemiologic studies will allow patients to make better-informed decisions about whether to pursue treatment.