Twenty-one year mortality in a dialysis unit: changing effect of withdrawal from dialysis.

To characterize the factors affecting the decision to withdraw from dialysis, the authors compared patients withdrawing from dialysis (n=62) with patients dying from all other causes (n=242) over 21 years (1976-1996) in a single dialysis unit. Compared with those who died from other causes, patients who withdrew were older (67+/-11 vs 61+/-11 years); were more likely to have severe physical impairment (87% vs 62%) and severe restriction of activities of daily living (77% vs 46%); and had higher frequencies of congestive heart failure (81 % vs 62%), myocardial infarction (60% vs 42%), peripheral vascular disease (71 % vs 40%), and diabetes mellitus (66% vs 36%) (p < or = 0.014). Dialysis modality; duration of dialysis; the degree of family support; index of disease severity; the use of tobacco, alcohol, or illicit drugs; and the frequency of ischemic heart disease, dysrhythmia, pericarditis, cardiac arrest, cerebrovascular accident, hypertension, obstructive lung disease, cancer, and human immunodeficiency virus did not differ between the two groups. Stepwise logistic regression showed that dialysis during 1990-1996, severe limitation of activities of daily living, and diabetes mellitus were independent risk factors for withdrawal. During 1990-1996, 44% of the deaths were caused by withdrawal from treatment. In addition to other factors, dialysis in the 1990s is a strong predictor of withdrawal from dialysis. The reasons for the increased rate of withdrawal from dialysis in recent years, and the effect of this increased rate of withdrawal on mortality, need further evaluation.