Two or Three Live Births Linked to Lower Osteoporosis Risk

Key Takeaways

• Among UK Biobank women, two or three live births were associated with lower osteoporosis risk than none; 5,390 women (4.5%) received osteoporosis diagnoses.
• Three live births were associated with significantly lower osteoporotic-fracture risk than no live births.
• Fully adjusted femoral-neck and lumbar-spine areal BMD associations with live-birth history were not statistically significant.

Pregnancy and lactation involve temporary skeletal changes, but the relationship between live-birth history and bone health later in life remains uncertain. For women entering middle and older age, osteoporosis diagnoses, fractures, and measured bone density may offer different perspectives on that relationship. Investigators examined these outcomes across categories of live-birth history.

Investigators conducted a prospective observational analysis of 119,969 women aged 40–69 years at baseline in the United Kingdom (UK) Biobank. Participants were recruited in 2006–2010 at sites in England, Scotland, and Wales. Self-reported live births were categorized by count, with no live births as the comparator. Exclusions included missing parity or covariate data, pre-existing osteoporosis, specified baseline illnesses, and previous osteoporosis-related prescriptions.

Linked diagnostic records supplied incident osteoporosis and osteoporotic-fracture outcomes; follow-up ended at the first relevant diagnosis, death, or November 30, 2022. Unweighted and inverse probability of treatment weighting (IPTW) Cox models accounted for demographic, genetic, lifestyle, reproductive, and clinical factors, including body mass index (BMI). Generalized linear models assessed femoral-neck and lumbar-spine areal bone mineral density (BMD) in 12,579 women; dual-energy X-ray absorptiometry (DXA) measurements began in 2014.

Compared with no live births, two were associated with lower incident osteoporosis risk in the weighted, fully adjusted model: hazard ratio (HR) 0.894 (95% confidence interval [CI] 0.825–0.970; P = 0.007). This was an association with recorded diagnoses, not evidence that childbirth prevented osteoporosis.

For three live births versus none, the corresponding weighted, fully adjusted osteoporosis estimate was HR 0.892 (95% CI 0.811–0.981; P = 0.019). Overall comparisons for one, four, and at least five live births were not statistically significant in the fully adjusted weighted analyses.

Osteoporotic-fracture risk was lower with three live births than with none in the weighted, fully adjusted model: HR 0.645 (95% CI 0.449–0.925; P = 0.017). Fully adjusted femoral-neck and lumbar-spine areal BMD associations were not significant. In age-stratified analyses, lower osteoporosis risk was observed for several parity groups among older women but not younger women. The two-birth association appeared in the normal-BMI group but not in the overweight or obesity groups; these patterns do not establish an interaction. Excluding early follow-up events retained the osteoporosis associations for two and three births and the fracture association for three births.

The investigators lacked bone microarchitecture and bone-turnover-marker measurements, as well as information on age at last childbirth, breastfeeding, and infant birth weight. Fertility-related genetic factors may have remained unmeasured. Excluding women with missing exposure or covariate information and relying on a smaller DXA subset also limited interpretation. Neither the null areal BMD findings nor proposed bone-quality explanations establish a BMD-independent mechanism, and observational associations do not demonstrate prevention. Although the authors proposed considering parity in future risk assessment, they did not validate a revised prediction, screening, or treatment approach.

The authors interpreted moderate parity as associated with lower later osteoporosis risk; the observed fracture association was narrower, occurring at three live births. The findings describe associations rather than effects of childbirth.

Clinician Questions

How was incident osteoporosis identified in the UK Biobank parity analysis?

Among UK Biobank women grouped by live-birth count, investigators identified incident osteoporosis from the first linked diagnosis coded M80–M82 under the International Classification of Diseases, Tenth Revision. This was a recorded diagnosis, not a case definition based solely on the later DXA subset.

Does zero live births mean a woman was never pregnant in the parity–osteoporosis comparison?

In the UK Biobank comparison of live-birth history and incident osteoporosis, parity was the self-reported count of live births, with zero as the reference category. Zero live births does not establish that a woman was never pregnant because the exposure counted live births rather than all pregnancies.

This content is intended for healthcare professionals for educational purposes. The article and its image were developed with AI assistance and reviewed by members of the ReachMD team prior to publication. Please refer to the original source linked for the complete and most up-to-date information.