The effect of maternal pre-/early-pregnancy BMI and pregnancy smoking and alcohol on congenital heart diseases: a parental negative control study

Abstract

Background: Congenital heart diseases (CHDs) are the most common congenital anomaly. The causes of CHDs are largely unknown, but intrauterine mechanisms appear to be important. Higher prenatal body mass index (BMI), smoking and alcohol consumption are associated with increased risk of CHDs. Whether these are causal is unclear. Methods: Seven European birth cohorts including 232,390 offspring were included. We applied negative exposure paternal control analyses to explore the intrauterine effects of maternal BMI, smoking and alcohol consumption during pregnancy, on offspring CHDs and CHD severity. We used logistic regression and combined estimates using a fixed-effects meta-analysis and compared maternal-specific confounder-adjusted associations to similarly adjusted associations of paternal exposures. If there is a causal intrauterine effect of a maternal pregnancy exposure, we would expect to see a maternal-specific association, with no (or weaker) association with the paternal exposure. Results: There were 2,469 CHD cases (1.1%). The association of mean BMI with CHD was null and similar in mothers (adjusted OR per 1kg/m2: 1.00 (95%CI: 0.99, 1.02)) and fathers (OR: 1.01 (0.99, 1.03)). The analyses of BMI categories resulted in similar increased odds of CHD in overweight (mothers OR: 1.15 (1.01, 1.31) and fathers 1.10 (0.96, 1.27)) and obesity (mothers OR: 1.12 (0.93, 1.36) and fathers 1.16 (0.90, 1.50)). Maternal smoking was associated with increased odds of CHD (OR: 1.11 (0.97, 1.25)) but paternal smoking was not (OR: 0.96 (0.85, 1.07)). The difference increased when removing offspring with genetic/chromosomal defects (mothers OR: 1.15 (1.01, 1.32) and fathers 0.93 (0.83, 1.05)). The positive association with maternal pregnancy smoking appeared to be driven by non-severe CHD cases (OR: 1.22 (1.04, 1.44)). Associations with maternal (OR: 1.16 (0.52, 2.58)) and paternal (OR: 1.23 (0.74, 2.06)) moderate/heavy pregnancy alcohol consumption were similar. Conclusions: We have shown that maternal, but not paternal, pregnancy smoking is associated with an increase in odds of offspring CHDs, primarily non-severe CHDs, providing evidence of an intrauterine effect. We found no strong evidence of an intrauterine effect of higher maternal BMI or alcohol consumption on offspring CHDs. Our findings provide further evidence for why smoking cessation is important during pregnancy.

Effect of Maternal Prepregnancy/Early‐Pregnancy Body Mass Index and Pregnancy Smoking and Alcohol on Congenital Heart Diseases: A Parental Negative Control Study

“…An analysis plan was written and published in October 2019, with any subsequent changes and their rationale documented in the publication. 35 All associations between exposures and CHDs were performed within participating studies using logistic regression (binary for main analyses and multinomial for CHD severity analyses). In the 2 largest cohorts (DNBC study and MoBa), we assessed deviation from linearity in our models in the BMI analyses by running our main confounder‐adjusted model with BMI split into fifths.…”

Ascertaining and classifying cases of congenital anomalies in the ALSPAC birth cohort

“…The prevalence of CHDs in ALSPAC is similar to other European birth cohorts. In recent work involving 7 European birth cohorts, we have shown that the prevalence of CHD was close to 1% in most cohorts, with the lowest with 0.4% and the highest with 1.4% 26 . Differences in case ascertainment could be one of a number of possible explanations for the slight differences in prevalence estimates.…”

Using linear and natural cubic splines, SITAR, and latent trajectory models to characterise nonlinear longitudinal growth trajectories in cohort studies

“…Even with adjustment, residual confounding (from using poorly measured confounders or not adjusting for important confounders) can bias results. Useful strategies for checking if residual confounding influences results include negative control variables and comparing cohorts with different confounding structures [38,[67][68][69].…”