The team integrated two-sample Mendelian randomization, which uses naturally occurring genetic variants as proxies for an exposure to probe whether associations are likely to reflect causal effects rather than confounding, with multivariable MR, MRlap, reverse MR, and longitudinal UK Biobank analyses. The primary exposure was heel estimated bone mineral density measured in the UK Biobank, with lumbar spine and total body dual-energy X-ray absorptiometry measures evaluated as sensitivity exposures. The primary outcome came from the FinnGen R12 dataset, comprising 26,090 spinal stenosis cases and 353,224 controls. A secondary outcome, primary lumbar decompression surgery, was assessed using genome-wide association summary statistics generated within the UK Biobank from 1,215 surgical cases and 155,913 controls. The final genetic instruments included 462 independent variants with a mean F-statistic of 46.3, indicating the proxies were strong enough to support inference.

Genetically proxied higher heel estimated bone mineral density was associated with spinal stenosis, with an odds ratio of 1.097 (95% confidence interval 1.01 to 1.191, p = 0.028) in the primary analysis. The association persisted after MR-PRESSO outlier removal, at an odds ratio of 1.09 (95% confidence interval 1.010 to 1.175, p = 0.026), and after multivariable Mendelian randomization adjusted for body mass index and height, two traits that could otherwise account for the link. In the sensitivity analyses, lumbar spine bone mineral density showed a stronger association with stenosis, with an inverse-variance weighted odds ratio of 1.23 (95% confidence interval 1.14 to 1.33, p < 0.001). However, reverse Mendelian randomization also associated genetic liability to stenosis with higher lumbar bone mineral density, a bidirectional signal the authors interpret as possible local bone remodeling at the disease site or contamination of the lumbar measure by the disease process itself.

The picture differed for the clinical endpoint of surgery. Neither heel estimated bone mineral density nor lumbar spine bone mineral density was associated with primary lumbar decompression in the MRlap sensitivity analysis, which corrects for sample overlap and winner’s curse bias. Observationally, baseline heel estimated bone mineral density was modestly associated with incident stenosis in Cox models of 157,128 European-ancestry UK Biobank participants with valid baseline heel measures, with a hazard ratio of 1.04 (95% confidence interval 1.00 to 1.09, p = 0.037), but it was not associated with subsequent decompression surgery. Taken together, the results support genetically influenced bone density as a contributor to stenosis risk while highlighting decompression surgery as a downstream clinical endpoint. The observational hazard ratio was modest, the surgical cohort was comparatively small, and Mendelian randomization estimates reflect lifelong genetic liability rather than modifiable bone density in later life, so the findings do not by themselves imply changes in clinical management.

Subject of Research: Genetic bone mineral density and risk of lumbar spinal stenosis

Article Title: The relationship between genetic bone mineral density and spinal stenosis: a Mendelian randomization and cohort study

Article References: Adegbenro, T. M., Carr, M., He, J., Polimanti, R., & Margetis, K. (2026). The relationship between genetic bone mineral density and spinal stenosis: a Mendelian randomization and cohort study. Scientific Reports. https://doi.org/10.1038/s41598-026-76865-3

Image Credits: AI Generated

DOI: 10.1038/s41598-026-76865-3

Keywords: bone mineral density, spinal stenosis, Mendelian randomization, UK Biobank, FinnGen, lumbar decompression, genetic epidemiology, heel eBMD, skeletal remodeling, Scientific Reports, relationship, between