PIN66 VACCINE PREGNANCY REGISTRIES: METHODS AND IMPACT ON FINDINGS

Abstract

Pregnancy registries offer a prospective means of assessing the effect of vaccine exposure on pregnancy outcome. Multiple vaccine pregnancy registries exist; however, research methods may vary between registries. We conducted a targeted review of these registries to examine their methods and whether these methods were robust enough to support the registries’ findings. Registries were identified from multiple sources: the FDA pregnancy registry website, http://clinicaltrials.gov, and PubMed. Data were gathered on exposures and outcomes of interest, sample size, proportion lost to follow-up, data source(s), comparator(s), and results. Data were analyzed descriptively. Our search identified a total of 15 vaccine pregnancy registries: 1 surveillance system for multiple vaccines and 14 individual registries examining exposures to influenza (n=3), meningitis (n=3), HPV (n=2), adenovirus (n=1), anthrax (n=1), hepatitis B (n=1), pertussis (n=1), smallpox (n=1), varicella (n=1) vaccines. Results were available for 6 (40%) registries. None of these registries found an elevated risk of fetal death, spontaneous abortion, or preterm birth. The multi-vaccine surveillance system reported an elevated risk of major congenital malformations (MCMs) associated with influenza vaccine exposure during pregnancy but no evidence of a pattern suggestive of possible teratogenesis. None of the other registries noted an elevated risk of MCMs. All but 1 of the registries had sufficient sample size and power to assess the outcomes of interest. However, the multi-vaccine surveillance system was the only registry that used multiple, reliable data sources (maternal interview and medical record review), an internal comparator group, and a low proportion of patients lost to follow-up (2% versus >30%). The findings from these 6 vaccine pregnancy registries did not suggest a teratogenic effect on the fetus or elevated risk of fetal death, spontaneous abortion, or preterm birth. However, only 1 of these registries had robust enough methods to support its findings.