Associations of activities of daily living and their trajectories with the risk of diabetes-related lower-limb amputation: evidence from the HRS and ELSA longitudinal cohorts.
Diabetes-related lower-limb amputation (DLLA) is a severe complication of diabetes associated with considerable disability and mortality. However, the associations of activities of daily living (ADL) and their longitudinal trajectories with DLLA risk remain unclear.
Data from the Health and Retirement Study (HRS; n = 3007) and the English Longitudinal Study of Ageing (ELSA; n = 833) were analyzed. ADL trajectories were identified using Group-Based Trajectory Modeling. Associations of ADL levels and trajectories with diabetes-related lower-limb amputation (DLLA) were evaluated using Cox regression models, with competing-risk, nonlinear, subgroup, and mediation analyses performed as sensitivity and exploratory analyses.
During a median follow-up of approximately two survey waves (approximately four years), 986 DLLA events occurred in the HRS cohort and 254 events occurred in the ELSA cohort. Multivariable Cox regression analyses demonstrated that ADL scores were significantly associated with DLLA risk after adjustment for potential confounders (HRS: adjusted HR = 1.10, 95% CI: 1.07-1.14; ELSA: adjusted HR = 1.20, 95% CI: 1.12-1.28). Compared with the Stable-low trajectory group, the risks of DLLA were significantly higher in both the Stable-rise and Stable-high groups. The risk increased by approximately 51%-84% in the Stable-rise group and by 36%-1.31-fold in the Stable-high group. Kaplan-Meier survival analysis showed significant differences in DLLA incidence among the ADL trajectory groups (log-rank P < 0.001). The competing risk analyses yielded results consistent with those of the Cox models. RCS analysis indicated a significant nonlinear association between ADL and DLLA risk, with a threshold effect observed around ADL ≈ 2. Furthermore, mediation analysis suggested that depression partially mediated the association between ADL and DLLA risk.
ADL levels and their longitudinal trajectories were significantly associated with the risk of DLLA among individuals with diabetes. Individuals with persistently poor functional status or progressively declining function exhibited the highest risk. Dynamic changes in functional status may serve as useful indicators for identifying individuals at elevated risk of DLLA. Routine assessment of ADL may contribute to risk stratification and clinical monitoring in individuals with diabetes.
Data from the Health and Retirement Study (HRS; n = 3007) and the English Longitudinal Study of Ageing (ELSA; n = 833) were analyzed. ADL trajectories were identified using Group-Based Trajectory Modeling. Associations of ADL levels and trajectories with diabetes-related lower-limb amputation (DLLA) were evaluated using Cox regression models, with competing-risk, nonlinear, subgroup, and mediation analyses performed as sensitivity and exploratory analyses.
During a median follow-up of approximately two survey waves (approximately four years), 986 DLLA events occurred in the HRS cohort and 254 events occurred in the ELSA cohort. Multivariable Cox regression analyses demonstrated that ADL scores were significantly associated with DLLA risk after adjustment for potential confounders (HRS: adjusted HR = 1.10, 95% CI: 1.07-1.14; ELSA: adjusted HR = 1.20, 95% CI: 1.12-1.28). Compared with the Stable-low trajectory group, the risks of DLLA were significantly higher in both the Stable-rise and Stable-high groups. The risk increased by approximately 51%-84% in the Stable-rise group and by 36%-1.31-fold in the Stable-high group. Kaplan-Meier survival analysis showed significant differences in DLLA incidence among the ADL trajectory groups (log-rank P < 0.001). The competing risk analyses yielded results consistent with those of the Cox models. RCS analysis indicated a significant nonlinear association between ADL and DLLA risk, with a threshold effect observed around ADL ≈ 2. Furthermore, mediation analysis suggested that depression partially mediated the association between ADL and DLLA risk.
ADL levels and their longitudinal trajectories were significantly associated with the risk of DLLA among individuals with diabetes. Individuals with persistently poor functional status or progressively declining function exhibited the highest risk. Dynamic changes in functional status may serve as useful indicators for identifying individuals at elevated risk of DLLA. Routine assessment of ADL may contribute to risk stratification and clinical monitoring in individuals with diabetes.