| アブストラクト | BACKGROUND: Frailty, reflecting increased vulnerability to adverse outcomes, is commonly measured by the accumulation of health deficits. We aimed to determine whether periods of accelerated deficit accumulation involve specific deficits and whether these periods provide additional prognostic value for adverse outcomes beyond total deficit count. METHODS: In this longitudinal cohort study, we analysed primary care records for 1 118 843 people aged 50 years and older in the Clinical Practice Research Datalink Aurum. Deficit accumulation was measured using 36 electronic Frailty Index deficits, recorded on the dates they appeared in primary care records from 2009-17. Piecewise linear regression was used to identify changes in the rate of deficit accumulation for each individual, and we estimated up to five linear segments per trajectory, with data-driven breakpoints indicating changes in the rate of deficit accumulation. We used permutation tests to assess specific deficits that were disproportionately represented (ie, occurred more often or less often than expected by chance) during accumulation periods and at the onset of deficit accumulation. Finally, Cox proportional hazards models were used to examine associations of accelerated accumulation with adverse outcomes in 2018-19. We examined the following four adverse outcomes from Jan 1, 2018, to Dec 31, 2019: unplanned hospital admissions (emergency or urgent), falls requiring hospitalisation, hip fractures, and death (Office for National Statistics mortality registration); falls and hip fractures were identified in Health Episode Statistics-Admitted Patient Care data using ICD-10 codes. All outcomes were analysed as time-to-event variables, with follow-up ending at the first event, deregistration, death (for non-death outcomes), or Dec 31, 2019. FINDINGS: 621 835 (55.6%) individuals were female and 497 008 (44.4%) individuals were male, with a mean age of 75.1 years (SD 9.2) in 2018. 100 373 (9.0%) individuals showed a change in their rate of deficit accumulation, indicated by two or more linear segments in their trajectories. Across all events, the mean age at onset was 70.9 years (SD 10.0). During accelerated accumulation, individuals accrued an average of 4.3 deficits (1.6) and the average duration of accelerated accumulations was 530 days (412). People with accelerated accumulation had more deficits at both the start and end of the observation period and accumulated them faster (mean 0.7 deficits per year, SD 0.2), compared with a mean rate of 0.4 deficits per year (0.1) in those without accelerated accumulation. These episodes were disproportionately linked with polypharmacy, heart failure, atrial fibrillation, dyspnoea, and mobility and transfer problems, but less likely to be associated with hypertension and arthritis, and respiratory disease. In models adjusted for age, sex, and total deficit count (on Dec 31, 2017), a time interval since accelerated accumulation onset of less than 3 years was associated with significantly higher risk of all outcomes: death (HR 1.49, 95% CI 1.44-1.55), unscheduled hospital admission (1.20, 1.17-1.23), falls (1.21, 1.17-1.26), and hip fracture (1.30, 1.18-1.42). Elevated risks persisted, though attenuated, at 3-6 years after onset, but risks at 6 years and beyond HRs were no longer significantly elevated. INTERPRETATION: In some individuals, accrual of frailty deficits includes episodes of accelerated accumulation, linked to specific deficits and signalling heightened vulnerability. Both the timing and rate of accumulation-not just the cumulative burden-might shape future health risks. FUNDING: Legal & General Group and National Institute for Health and Care Research. |
|---|
| 組織名 | Advanced Care Research Centre, University of Edinburgh, Edinburgh, UK.;Advanced Care Research Centre, University of Edinburgh, Edinburgh, UK; Centre for;Cardiovascular Science, University of Edinburgh, Edinburgh, UK.;Advanced Care Research Centre, University of Edinburgh, Edinburgh, UK; School of;Social and Political Science, University of Edinburgh, Edinburgh, UK.;Informatics, University of Edinburgh, Edinburgh, UK. Electronic address:;sohan.seth@ed.ac.uk. |
|---|
