Wrist-worn accelerometer-derived activity fragmentation and cause-specific mortality in US adults: a population-based prospective cohort study

Wrist-worn accelerometers capture the continuity of everyday movement as well as its amount. We examined how activity fragmentation relates to specific causes of death and whether it adds prognostic information beyond clinical risk factors. We analysed 7098 adults aged ≥ 20 years from the National Health and Nutrition Examination Survey (NHANES) 2011–2014, linked to mortality through 2019. The active-to-sedentary transition probability (ASTP) was derived from usable waking minutes. Survey-weighted Cox and Fine–Gray models estimated mortality associations; 6586 participants had complete clinical covariates. Sensitivity analyses addressed nonresponse, early deaths, metabolic markers and movement amount. Exploratory prediction was assessed internally using paired model comparisons, 1000 bootstrap refits and five-year calibration. During median follow-up of 6.8 years, 522 deaths occurred. In the fully adjusted sample, each SD of ASTP was associated with all-cause mortality (hazard ratio [HR] 1.165, 95% confidence interval [CI] 1.042–1.304) and other-cause mortality (HR 1.293, 95% CI 1.099–1.521). Cardiovascular estimates were imprecise in both models; cancer showed an inverse subdistribution association but an imprecise cause-specific estimate. Nonresponse weighting produced a similar all-cause HR of 1.168. The 24-month landmark estimate was 1.097 (95% CI 0.959–1.253), and adjustment for movement amount attenuated or reversed the association. Adding ASTP changed apparent C-index by 0.0034 (bootstrap 95% CI 0.0006 to 0.0067); the optimism-corrected difference was 0.0032. Wrist-derived activity fragmentation, measured by ASTP, is an epidemiologic marker of movement pattern correlated with movement amount. Associations with all-cause and other-cause mortality after clinical adjustment do not establish an independent causal effect of activity continuity; the all-cause estimate was sensitive to movement amount and the 24-month landmark. The small prediction increment was internal and exploratory, with no established clinical utility. ASTP is distinct from accumulated brief moderate-to-vigorous activity.