Comparisons of survival between dialysis and nondialysis care for older adults with kidney failure have been limited to those managed by nephrologists, and are vulnerable to lead and immortal time biases. So we compared time to all-cause mortality among older adults with kidney failure treated vs. not treated with chronic dialysis. Our retrospective cohort study used linked administrative and laboratory data to identify adults aged 65 or more years of age in Alberta, Canada, with kidney failure (2002-2012), defined by two or more consecutive outpatient estimated glomerular filtration rates less than 10 mL/min/1.73m2, spanning 90 or more days. We used marginal structural Cox models to assess the association between receipt of dialysis and all-cause mortality by allowing control for both time-varying and baseline confounders. Overall, 838 patients met inclusion criteria (mean age 79.1; 48.6% male; mean estimated glomerular filtration rate 7.8 mL/min/1.73m2). Dialysis treatment (vs. no dialysis) was associated with a significantly lower risk of death for the first three years of follow-up (hazard ratio 0.59 [95% confidence interval 0.46-0.77]), but not thereafter (1.22 [0.69-2.17]). However, dialysis was associated with a significantly higher risk of hospitalization (1.40 [1.16-1.69]). Thus, among older adults with kidney failure, treatment with dialysis was associated with longer survival up to three years after reaching kidney failure, though with a higher risk of hospital admissions. These findings may assist shared decision-making about treatment of kidney failure.
