Electronic Thesis and Dissertation Repository


Doctor of Philosophy


Physiology and Pharmacology


Dr. Michael J Rieder

2nd Supervisor

Dr. Gideon Koren

Joint Supervisor


There is inadequate information on the fetal safety of drugs during pregnancy for the majority of marketed drugs. It is challenging to examine the safety and efficacy of drugs during pregnancy due to the ethical issues of exposing unborn babies to these chemicals. It often takes many years before associations between a drug and its safety, efficacy, and toxicity in pregnancy can be established. This thesis will examine strategies in signal detection of the effects of drug exposures during pregnancy.

Meta-analyses have become useful in the area of clinical teratology. Observational studies provide the main source of information in these meta-analyses. Although the quality of meta-analysis of small observational studies is challenging, it is an effective strategy, as shown in the present study, in predicting correct signals to estimate teratogenicity years before large cohort studies become available.

Results of retrospective pregnancy registries are commonly reported in regulatory documentations. However, little data are available on the precision of the estimates from such registries. The present study confirms a consistent bias against the null hypothesis in a retrospective registry which needs to be considered when interpreting such data as a strategy in generating safety/risk signals of new drugs.

H1 antihistamines are used for the treatment of nausea and vomiting during pregnancy as well as the symptomatic relief of allergy. Although they are felt to be safe, several studies have challenged this assumption. By using meta-analysis, the safety of antihistamines has been confirmed in this thesis with over 1.3 million exposed and control subjects.

Typically, after experimental animal studies, novel therapeutic modalities are tested by randomized controlled trials. Cumulative meta-analysis is an effective strategy to detect a possible time- dependent effect and potential bias against the null hypothesis, whether antioxidant treatment decreases the rates of preeclampsia. I have shown that the initial favorable effect seen in the first studies is nullified as the sample sizes and number of studies is increased.

There is a need to continue using and developing the above strategies to study the safety and efficacy of drugs to improve maternal-fetal health.