Association between a Deficit Accumulation Frailty Index and Mobility Outcomes in Older Adults: Secondary Analysis of the Lifestyle Interventions and Independence for Elders (LIFE) Study

Abstract

Frailty is a geriatric syndrome represented by susceptibility to precipitating health events and reduced functional reserve. Frailty can be difficult to measure in clinical practice and research. One approach to approximate frailty is based on a deficit accumulation approach, which assesses a larger number of less specific measures such as the presence of comorbidities, physical or cognitive assessments, and lab tests, and summarizes these as a frailty index. The objective of this study was to develop such an index using the Lifestyle Interventions and Independence for Elders (LIFE) Study and evaluate the validity of the frailty measure derived based on baseline information via its association with the primary outcomes of the trial, namely major mobility disability (MMD) and persistent MMD (pMMD). Further, this study aimed to evaluate the effectiveness of the physical activity intervention among participants based on their baseline frailty score. Subjects in the LIFE Study were evaluated at baseline for demographics, clinical history, and a battery of physical and cognitive functioning assessments. In total, 75 possible deficits were scored either as present (yes/no) or based on each score's quintiles for score-based assessments. The frailty index was measured as the total sum of deficits divided by the total number of possible deficits on a continuous scale between 0 and 100 (i.e., percent of deficits present). The frailty index was further divided into quintiles for comparison. A proportional hazards model was estimated for the MMD outcome controlling for other baseline information. A data driven approach was also used to determine relevant cut-offs in the frailty index where the trial intervention appeared to be modified. Among 1635 trial participants, the mean frailty index was 30.4 ± 6.6 and normally distributed. Over 2.5 years of average follow-up, 14.6%, 16.5%, 18.6%, 22.6%, and 27.6% of participants experienced MMD in quintiles 1-5, respectively. Each 1-unit increase in the frailty index increased the hazard of MMD by 4% (2-5%), and there was a nearly 2-fold increase in MMD between the highest and lowest frailty quintiles. Using log-rank criteria, a cut-point at the median was identified. Further, iterations tested for a frailty cut-off and indicated a subgroup beyond the 85th percentile wherein the physical activity intervention appeared to be no longer be effective. This internally derived deficit accumulation frailty index was uniquely able to identify individuals at higher risk of MMD and pMMD and showed that along the spectrum of frailty, the physical activity intervention remained effective for the majority of participants.

Keywords: LIFE Study; deficit accumulation; disability; frailty; healthy aging; mobility; older adults.

Figure 1

Distribution and summary statistics of the frailty index.

Figure 2

Kaplan-Meier plots of major mobility disability (MMD) (Panel A) and persistent MMD (pMMD) (Panel B) for quintile-stratified frailty index groups over 3 years of follow-up. Quintile 0 is equivalent to the group with the least frailty measures and Quintile 4 is the group with the most.

Figure 3

Cox proportional hazard regression results of the interaction between the LIFE Study intervention and a continuous frailty index for both major mobility disability and persistent major mobility disability. * Results show the overall effect of the physical activity (PA) intervention compared with the health education (HE) control arm at the mean frailty index. ^# Per 1-unit change in frailty index. ^#\ Results for the overall pooled cohort.

Figure 4

Cox proportional hazard regression results for the interaction between the LIFE Study intervention and quintile stratification of the frailty index for major mobility disability. The topmost results compare the physical activity (PA) intervention versus the health education (HE) control arm stratified by frailty quintiles. The stratified and pooled associations of the quintiles are also shown.

Figure 5

Cox proportional hazard regression results for the interaction between the LIFE Study intervention and quintile stratification of the frailty index for persistent major mobility disability. The topmost results compare the physical activity (PA) intervention versus the health education (HE) control arm stratified by frailty quintiles. The stratified and pooled associations of the quintiles are also shown.

Figure 6

Cox proportional hazards regression results testing frailty index cut-points as strata to identify subgroups in which the intervention was no longer effective in preventing major mobility disability.

Figure 7

Cox proportional hazards regression results testing frailty index cut-points as strata to identify subgroups in which the intervention was no longer effective in preventing persistent major mobility disability.