Activity of Daily Living Disability and Dialysis Mortality: Better Prediction Using Metrics of Aging

To the Editor: The ability to predict mortality in individuals undergoing dialysis is limited, with predictive accuracy that is moderate at best (C-statistic = 0.67–0.76).1-4 Metrics of aging, such as activity of daily living (ADL) disability,5 may improve mortality prediction because individuals with end-stage renal disease (ESRD) experience physiologic decline similar to that sometimes seen in aging. Initiation of dialysis is associated with a decline in ADLs,6 although the association between ADL disability and subsequent mortality remains unknown in all age groups. The goal of this study was to estimate the association between ADL disability and mortality in a prospective cohort of individuals undergoing dialysis. Individuals undergoing prevalent dialysis were recruited from a dialysis center in Baltimore, Maryland (1/2009–3/2010). Participants provided a medical history and underwent physical examination at enrollment and were followed until November 15, 2011. Vital status and dates of death were obtained from the dialysis center and augmented by linkage to the National Death Index. Each of the six ADL domains was measured at enrollment (feeding, dressing, ambulation, grooming, using a toilet, and bathing).5 ADL disability was defined as inability to perform any of the domains without assistance.5 Similar to previous studies, the following nine comorbidities were captured and counted: peripheral vascular disease (PVD), rheumatoid arthritis, cancer, hypertension, chronic obstructive pulmonary disease (COPD), diabetes mellitus, congestive heart failure, angina pectoris, and myocardial infarction.7 Cox proportional hazards models were used to estimate adjusted relative mortality accordihg to ADL disability (censored at transplantation or end of study). Analyses were performed using Stata 11.0/MP for Linux (StataCorp., College Station, TX). The mean age of the 143 individuals undergoing dialysis in the study was 60.6 ± 13.4; 46% were female, 82% were African American, and 41% had at least one ADL disability. Average follow-up was 24 ± 6.9 months. Participants with an ADL disability were more likely to be older (65.3 vs 57.4, P < .001) and have diabetes mellitus (70.7% vs 49.4%, P = .01), PVD (46.6% vs 20.0%, P = .001), and more comorbid conditions (3.1 vs 2.4, P = .01). Time on dialysis was not associated with ADL disability (5.0 vs 4.4 months, P = .60). There were 33 deaths during the study; mortality was 17.7 per 100 person-months for participants with an ADL disability and 5.1 per 100 person-months for those without (P < .001) (Figure 1). After adjusting for age, sex, and number of comorbidities, ADL disability was associated with 3.37 times higher mortality (95% confidence interval (CI) = 1.57–7.22); results were similar when adjusting for comorbidities separately (diabetes mellitus, PVD, and COPD were statistically significant). This risk was similar in adults younger than 65 (hazard ratio (HR) = 3.31, 95% CI = 1.02–10.79) and those aged 65 and older (HR = 3.46, 95% CI = 1.22–9.77) (interaction HR = 0.94, P = .90). Results were also similar according to sex (interaction HR = 0.84, P = .80) and race (interaction HR = 1.16, P = .90). Inclusion of ADL disability in a model adjusted for age, sex, and number of comorbidities substantially improved the C statistic from 0.66 to 0.74 (P = .03). In this prospective study of individuals undergoing dialysis, ADL disability was independently associated with 3.37 times higher mortality. In addition to age, sex, and comorbidities, use of a measure of ADL disability substantially improved the ability to predict mortality in these individuals. The results did not differ according to age, suggesting that, although ADL disability is a metric of aging, it is predictive of mortality at all ages in adults with ESRD. The prevalence of ADL disability is strikingly higher in individuals with ESRD of all ages (41% in the current study) than in community-dwelling older adults (5.0–8.1% of noninstitutionalized adults aged ≥ 65).8 The findings of the current study were consistent with previous observations that ADL disability predicted risk of in-hospital mortality in older hospitalized individuals undergoing dialysis9 and extended these findings to individuals undergoing outpatient dialysis of all ages. Strengths of this study include its prospective design, direct measurement of ADL disability, granular ascertainment of comorbidities, and no loss to follow-up. The main limitation was that ADL disability was not established before dialysis initiation. This was a single-center study; direct inferences may be limited to the demographics of this study population if effect modification exists, although no effect modification was identified according to age, sex, or race. Measuring an individual's ability to perform ADLs may be an important tool for identifying individuals undergoing dialysis who are at high risk of dying. This information is easy to capture and does not require sophisticated tests. If ADL disability is confirmed to be on the causal pathway between ESRD and mortality, these findings may be important in identifying a potential target for intervention. We thank the study participants and the staff, including Colleen Reft, at the Renal Dialysis Program at Good Samaritan for their dedication to this study. This study was supported by a Clinical Scientist Development Award from the Doris Duke Charitable Foundation (Segev, PI). Conflict of Interest: The editor in chief has reviewed the conflict of interest checklist provided by the authors and has determined that the authors have no financial or any other kind of personal conflicts with this paper. Clinical Scientist Development Award was provided from the Doris Duke Charitable Foundation.