Sleep pattern clusters, physical function and fall risk: geriatric syndromes among older ambulatory women

Poor sleep is a suspected risk factor for lower physical functioning and frequent falling at older ages. We evaluated the relationship of multiple sleep and rest-activity rhythm (RAR) metrics with fall risk and physical functioning.

Older women (N = 4543) wore hip-worn accelerometers, recording their falls daily for 13 months following accelerometry. Uniform manifold approximation projection identified sleep-circadian clusters; K-means clustering further distinguished healthy and unhealthy sleep patterns. After cross-validation, we examined associations between sleep clusters and fall risk using adjusted, negative binomial models. Linear regression models estimated associations of sleep clusters with short physical performance battery (SPPB) score and its sub-scores. We evaluated whether SPPB status modified associations of sleep and RAR with fall risk.

Five sleep clusters were identified including C1 ("sleep disturbed," n = 1051), C2 ("healthy," n = 1043), C3 ("mild RAR, active," n = 1446), C4 ("earlier sleepers, n = 105), and C5 ("shorter, mildly disrupted, later sleeper," n = 898). Unhealthy sleep clusters C1 and C4 were associated with a higher fall risk compared to healthy cluster C2 after adjustment (C4, IRR: 1.76 (95% CI: 1.15-2.69)). These same clusters were also associated with lower balance scores (score: 0-4) after adjustment (C1, beta: -0.11 (95% CI: -0.21 to -0.01); C4, beta: -0.30 (95% CI: -0.55 to -0.05)).

Older women with unhealthier sleep-RAR patterns are more at risk for falling, which may be partially explained by the role of sleep on balance and physical functioning.