Observational epidemiology has repeatedly reported associations between greater parity and elevated cardiovascular disease (CVD) risk. It is uncertain whether these associations reflect causal effects of childbearing on later cardiometabolic health or are driven by confounding, including socioeconomic factors. This study evaluated whether the number of children ever born (NEB) has a causal relationship with a range of long-term cardiometabolic traits using genetic instrumental variable approaches, including sex-stratified two-sample Mendelian randomization (MR) and a novel spousal MR framework designed to reduce bias from horizontal pleiotropy.
Analyses used UK Biobank data comprising up to 172,122 females and 138,390 males for associations between NEB and 16 cardiometabolic traits. An independent subset of 53,237 UK Biobank spousal pairs was used for spousal analyses. Outcomes evaluated included anthropometric measures (for example, body mass index and body fat percentage), metabolic measures (basal metabolic rate, blood glucose), lipid measures (HDL cholesterol, ApoA1) and blood pressure, as well as type 2 diabetes risk.
The investigators conducted sex-stratified two-sample MR using genetic instruments for NEB. In addition to conventional MR applied to the individual's genotype, the study implemented a novel spousal MR approach in which an individual's spouse’s genotype serves as the instrumental variable. The spousal MR is intended to mitigate bias from horizontal pleiotropy by leveraging partner genotype correlations and shared household or environmental factors.
Multivariable regression models assessed observational associations between NEB and cardiometabolic traits with adjustment for socioeconomic status. Sex-stratified MR analyses tested for putative causal effects of genetically predicted NEB on each trait. The spousal MR analysis used the partner’s genotype as the instrument to provide an alternative causal estimate less susceptible to direct pleiotropic effects of the participant’s own genotype. The paper reports comparisons between conventional MR and spousal MR findings to explore robustness and potential pleiotropic bias.
In multivariable regression, NEB was associated with multiple cardiometabolic traits even after adjustment for socioeconomic status. The pattern and magnitude of associations differed between males and females, indicating sex-specific observational relationships between parity and later cardiometabolic measures.
Conventional two-sample MR provided evidence for several putative causal effects of higher NEB:
These MR estimates suggest some causal links between NEB and cardiometabolic traits, although the authors note that effect sizes are likely to be small.
The spousal MR analyses corroborated positive causal effects of NEB on female BMI and female BMR observed in conventional MR. In addition, spousal MR suggested inverse causal effects of NEB on female HDL cholesterol and ApoA1, and on male blood glucose, which were not uniformly observed in the traditional MR analyses. The divergence between individual-level MR and spousal MR findings highlights the influence of methodological choices and potential pleiotropic or environmental confounding on causal inference.
Taken together, the findings indicate a possible causal relationship between NEB and long-term cardiometabolic health, with some sex-specific effects. However, the authors emphasize that estimated causal effects are likely to be small. Differences between multivariable observational associations, conventional MR, and spousal MR underscore complexity in disentangling biological effects of parity from socioeconomic, behavioural, and pleiotropic genetic influences.
The report does not present detailed effect sizes or confidence intervals within this summary, and the full preprint should be consulted for numeric estimates and sensitivity analyses. As with all MR studies, validity depends on the strength and specificity of genetic instruments and on assumptions about pleiotropy and population structure.
This research used UK Biobank data and had relevant ethical approvals: UK Biobank ethical approval from the North West Multi-Centre Research Ethics Committee (MREC) and study-specific approvals listed in the manuscript. The project was conducted under UK Biobank application number 53641 and received institutional ethics approval at the University of Queensland. Sex-specific GWAS summary statistics generated in the study have been deposited in the University of Queensland eSpace repository and are publicly available. The work reports funding from multiple agencies, including the Australian Research Council, the Research Council of Norway, the National Health and Medical Research Council, and United Kingdom Research and Innovation Medical Research Council. The authors declared no competing interests.