The association of diabetic status, age, and race to withdrawal from dialysis

Abstract

Higher rates of withdrawal from dialysis have been shown among young diabetic patients. This study examined the hypothesis that diabetic status has a Joint effect with age as the result of an age-related increase in comorbid conditions among nondiabetic ESRD patients. Using data collected during the 1980s by the Michigan Kidney Registry, the experience of 4.753 white and 2,988 black Michigan dialysis patients (13,863 patient years) is analyzed. During follow-up, 364 white and 80 black patients withdrew from dialysis. By use of a Cox analysis, adjusted rates of withdrawal were shown to be higher among white than black patients (rate ratio (RR) = 3.34; P < 0.001), without significant differences by diabetic status. Rates increased with age (RR = 7.9 for 80+ versus 20 to 35 yr; P < 0.001), without significantly different patterns by race. Adjusted withdrawal rates among diabetics were 2.3 times that of nondiabetics overall (P < 0.001), with a statistically significant joint effect with age, the RR being highest among young patients. Among white patients, nondiabetics using continuous ambulatory peritoneal dialysis had a lower relative withdrawal rate than those using hemodialysis (RR = 0.73; P = 0.07), whereas diabetics using continuous ambulatory peritoneal dialysis showed slightly higher although statistically insignificant rates than their counterparts using hemodialysis (RR = 1.21; P = 0.50). Between 1980 and 1989. withdrawal rates for white patients aged 20 to 34 yr tended to decrease by 20% per year (P = 0.05), whereas they increased by 11% per year for patients aged 50 to 69 yr (P < 0.01). These findings describe the joint effect of age, diabetic status, and specific time trends in withdrawal for cohorts initiating therapy during the 1980s while confirming the main effects of race and age.