Frailty as a novel predictor of mortality and hospitalization in individuals of all ages undergoing hemodialysis

Abstract

Objectives: To quantify the prevalence of frailty in adults of all ages undergoing chronic hemodialysis, its relationship to comorbidity and disability, and its association with adverse outcomes of mortality and hospitalization.

Design: Prospective cohort study.

Setting: Single hemodialysis center in Baltimore, Maryland.

Participants: One hundred forty-six individuals undergoing hemodialysis enrolled between January 2009 and March 2010 and followed through August 2012.

Measurements: Frailty, comorbidity, and disability on enrollment in the study and subsequent mortality and hospitalizations.

Results: At enrollment, 50.0% of older (≥ 65) and 35.4% of younger (<65) individuals undergoing hemodialysis were frail; 35.9% and 29.3%, respectively, were intermediately frail. Three-year mortality was 16.2% for nonfrail, 34.4% for intermediately frail, and 40.2% for frail participants. Intermediate frailty and frailty were associated with a 2.7 times (95% confidence interval (CI) = 1.02-7.07, P = .046) and 2.6 times (95% CI = 1.04-6.49, P = .04) greater risk of death independent of age, sex, comorbidity, and disability. In the year after enrollment, median number of hospitalizations was 1 (interquartile range 0-3). The proportion with two or more hospitalizations was 28.2% for nonfrail, 25.5% for intermediately frail, and 42.6% for frail participants. Although intermediate frailty was not associated with number of hospitalizations (relative risk = 0.76, 95% CI = 0.49-1.16, P = .21), frailty was associated with 1.4 times (95% CI = 1.00-2.03, P = .049) more hospitalizations independent of age, sex, comorbidity, and disability. The association between frailty and mortality (interaction P = .64) and hospitalizations (P = .14) did not differ between older and younger participants.

Conclusions: Adults of all ages undergoing hemodialysis have a high prevalence of frailty, more than five times as high as community-dwelling older adults. In this population, regardless of age, frailty is a strong, independent predictor of mortality and number of hospitalizations.

Figure 1. Distribution of Frailty Score in Study Population

The percentage of nonfrail, intermediately frail, and frail participants are listed above the cutoffs.

Figure 2. Overlap of Frailty, Comorbidity, and Disability

Frailty is defined as 3 or more components as defined by Fried. Disability is defined as the need for assistance in 2 or more activities of daily living categories. Comorbidity is defined as 4 or more conditions as specified in the methods. The total represents 166 study participants on hemodialysis. The n of each subgroup indicated in parentheses.

Figure 3. Estimated Cumulative Incidence of Mortality, by Frailty

Kaplan-Meier method was used to estimate curves; the log-rank test was statistically significant (P=0.047).