Telomere length and risk of incident fracture and arthroplasty: findings from UK Biobank.
Curtis EM., Codd V., Nelson C., D'Angelo S., Wang Q., Allara E., Kaptoge S., Matthews P., Tobias J., Danesh J., Cooper C., Samani N., Harvey NC.
We investigated independent associations between telomere length and risk of fracture and arthroplasty in UK Biobank participants. Leucocyte telomere length (LTL) was measured in baseline samples using a validated PCR method. We used, in men and women separately, Cox proportional hazards models to calculate the hazard ratio for incident fracture (any, osteoporotic) or arthroplasty (hip or knee) over 1,186,410 person-years of follow-up. Covariates included age, white cell count, ethnicity, smoking, alcohol, physical activity and menopause (women). In further analyses we adjusted for either estimated bone mineral density from heel quantitative ultrasound, handgrip strength, gait speed, total fat mass (bioimpedance) or blood biomarkers, all measured at baseline (2006-2010). We studied 59,500 women and 51,895 men, mean(SD) age 56.4(8.0) and 57.0(8.3) years respectively. During follow-up there were 5,619 fractures; 5,285 hip and 4,261 knee arthroplasties. In confounder-adjusted models, longer LTL was associated with reduced risk of incident knee arthroplasty in both men [hazard ratio/SD (95%CI): 0.93 (0.88,0.97)] and women [0.92 (0.88,0.96)] and hip arthroplasty in men [0.91 (0.87,0.95)] but not women [0.98 (0.94,1.01)]. Longer LTL was weakly associated with reduced risk of any incident fracture in women [hazard ratio/SD (95% CI): 0.96 (0.93,1.00)] with less evidence in men [0.98 (0.93,1.02)]. Associations with incident outcomes were not materially altered by adjustment for heel estimated bone mineral density, grip strength, gait speed, fat mass or blood biomarker measures. In this, the largest study to date, longer LTL was associated with lower risk of incident knee or hip arthroplasty, but only weakly associated with lower risk of fracture. The relative risks were low at a population level, but our findings suggest that common factors acting on the myeloid and musculoskeleletal systems might influence later life musculoskeletal outcomes. This article is protected by copyright. All rights reserved.