Developing methods of simultaneously analysing occurrence of several birth defects to improve identification of teratogenic medications

Many pregnant women take prescribed or over-the counter medications in the 1st trimester of their pregnancy. In many cases, the medication is imperative to the health of the woman and cannot be avoided. In some cases, a woman who is taking medication may not know she is pregnant until after organogenesis has begun. Some medications are known to be teratogenic when taken in the 1st trimester of pregnancy, however, the risk to the fetus of most drugs is uncertain. Clinical trials do not assess possible effects of new medications on a fetus as pregnant women are purposefully excluded, and the effects of medications on animal models are not always comparative to humans due to differences in developmental pathways. Identification of teratogenic medications, therefore, requires careful analysis of observational data on congenital anomalies occurring in exposed pregnancies in the clinical setting.

Although around 2-3% of pregnancies are affected by a birth defect, specific birth defects are rare. Therefore, the analysis of very large numbers of exposures is required to identify statistically significant effects for specific birth defects. Previous statistical methods for teratogen identification have focused on identifying increased occurrence of single defects. However, teratogen exposure often results in constellations of defects dependant on timing, dosage, and genetic interactions. Analysis of a group of birth defects may produce a statistically significant result when separate analysis of the single defects did not, implicating the medication as associated to the pattern of defects.

EUROmediCAT has established a unique database containing information on the medications 33,000 mothers were prescribed during the first trimester of pregnancy, and the congenital anomalies the fetus had, from 15 countries in Europe from 1995-2016. This project will build upon the quantitative skills and knowledge developed during the Medical Statistics MSc (MRC-LID funded). Statistical and computational methods will be explored and applied to the EUROmediCAT dataset, in order to produce a methodology with increased ability to identify teratogenic medications, through the analysis of groups of birth defects. This will provide information on potential teratogens and will therefore influence targeted evidence gathering for these medications. This evidence can be used to help women and healthcare providers make better decisions when balancing the risk and benefit of medications taken during pregnancy.