Sources of heterogeneity when studying the cardiovascular effects of fetal growth restriction: an overview of the issues. (3rd April 2022)

• Record Type: Journal Article
• Authors: Rodriguez-Lopez, Merida; Vergara-Sanchez, Carlos; Crispi, Fatima; Cepeda, Ivan Leonardo
• Is Part Of: Journal of maternal-fetal & neonatal medicine. Volume 35:Number 7(2022)
• Page Start: 1379
• Page End: 1385
• Publication Date: 2022-04-03
• Subjects: Fetal growth retardation -- fetal programing -- prevention -- cardiovascular diseases -- statistical analysis -- heterogeneity; Obstetrics -- Periodicals; Perinatology -- Periodicals; Infants (Newborn) -- Diseases -- Periodicals; Neonatology -- Periodicals; 618.2
• DOI: 10.1080/14767058.2020.1749592
• Languages: English
• ISSNs: 1476-7058

Abstract

Abstract: Intrauterine growth restriction (IUGR) has been repeatedly identified as a risk factor for cardiovascular disease (CVD). A possible explanation for this association is the effect of IUGR on cardiovascular structure and function. However, the specific changes observed are not consistent among studies. In this paper, we analyze several sources of heterogeneity within and between studies related to exposure, outcome and co-variables. A broad IUGR definition might include different phenotypes, expressing heterogeneity as an outcome. Outcome heterogeneity may also be the result of the postnatal effect modification that can be explored within studies. In order to do so, it is important to move beyond mean differences between groups, for example using unsupervised, stratified or interaction analysis. Different definitions of IUGR and the inclusion of different postnatal variables as confounders are potential sources of heterogeneity between studies. Researchers should be aware that postnatal variables may play different roles throughout a person's life and are not limited to behave as confounders. Therefore, their inclusion in the statistical model needs to be carefully considered. We discuss when sources of heterogeneity need to be controlled, and when they need to be identified and shown as a result.