Patients who die within the first 90 days after beginning dialysis are not included in mortality rates and may be absent from incidence counts. To identify factors associated with mortality within 3 months of the initiation of dialysis for ESRD and to estimate the impact of early deaths on ESRD incidence and mortality rates, this study investigated 15,245 patients who began dialysis in Georgia, North Carolina, and South Carolina over a 5-yr period. Data were collected by dialysis facility staff and reported to an ESRD registry. Six percent of all new patients died within 90 days of dialysis initiation (32% of all deaths occurring in the first year of treatment). Characteristics independently associated with increased risk of early death included older age, white race, male gender, physical and nutritional impairment, smoking, and the presence of cancer, congestive heart failure, clinical depression, and history of myocardial infarction. Depending on race-gender group, age-adjusted mortality rates based on this cohort were underestimated by 3 to 12% when patients who died early were excluded. These results suggest that certain patient characteristics-some potentially modifiable-confer increased risk of early death, and that the systematic exclusion of patients who die early from the U.S. national registry substantially influences ESRD mortality rates.
