Fig. 4. Two-sampled Mendelian Randomisation suggests a causal relationship between risk of low lumbar spine bone mineral density (ls-BMD) and height loss.

The effect of lumbar spine bone mineral density (ls-BMD) on longitudinal height-change using 30 ls-BMD associated SNPs from Estrada et al.²⁴ as instrument variables. The regression line from GSMR²² (b_xy = 0.011, s.e. 0.003, ${\chi }_1^2$ = 12.75, P = 3.5 × 10⁻⁴) indicates that increased osteoporosis risk (i.e. decreased ls-BMD) is causative for greater height loss with age. Each point represents the average effect of an individual SNP on either ls-BMD (x-axis, N = 31,800) or rate of height-change (y-axis, N = 50,072) with bars indicating the standard error of the estimates.