Background: In regions where prenatal screening for anencephaly and spina bifida is widespread, many cases of these defects are diagnosed prenatally. The purpose of this study was to estimate the frequency of termination of pregnancy (TOP) following prenatal diagnosis of anencephaly or spina bifida and to investigate factors associated with TOP that might lead to selection bias in epidemiologic studies.
Methods: We included articles indexed in Medline or Embase between 1990 and May 2012 reporting the frequency of TOP following prenatal diagnosis of anencephaly or spina bifida with English-language abstracts, 20 or more prenatally diagnosed cases, and at least half of the study years in 1990 or later. We summarized the frequency of TOP across studies using random-effects metaanalysis and stratified results by fetal and study characteristics.
Results: Among the 17 studies identified, 9 included anencephaly and 15 included spina bifida. Nine were from Europe, six were from North America, and one each was from South America and Asia. The overall frequency of TOP following prenatal diagnosis was 83% for anencephaly (range, 59-100%) and 63% for spina bifida (range, 31-97%). There were insufficient data to stratify the results for anencephaly; TOP for spina bifida was more common when the prenatal diagnosis occurred at less than 24 weeks' gestation, with defects of greater severity, and in Europe versus North America.
Conclusions: Because underascertainment of birth defects might be more likely when the pregnancy ends in TOP and TOP is associated with fetal characteristics, selection bias is possible in epidemiologic studies of anencephaly or spina bifida.
Copyright © 2012 Wiley Periodicals, Inc.