The objective of this study was to determine the probabilities of specific morbid events or death among patients with end-stage renal disease (ESRD) treated by hemodialysis. A prospective cohort study was performed between March 1988 and September 1989 in 18 hemodialysis centers in 13 Canadian cities, representing about one third of the hemodialysis population in Canada. The inception cohort consisted of 496 patients entering hemodialysis who had survived 1 month. The few new hemodialysis patients who received erythropoietin (EPO) in the last 3 months of the study were excluded. Survival curves were compared using the Cox proportional hazards regression model. Older age and history of cardiovascular disease were independently associated with a greater probability of death. Age and history of cardiovascular disease were also associated with a greater probability of nonfatal circulatory events (myocardial infarction, angina requiring hospitalization, or stroke), while a serum albumin level less than or equal to 30 g/ L (3.0 g dL) was associated with an increased probability of pulmonary edema. The probability of surviving 12 months without receiving a blood transfusion was 47.2% for males and 27.5% for females. The incidence of non-A, non-13 hepatitis, as estimated by unexplained elevations in serum aspartate aminotransferase (AST) values, was not different between patients receiving and not receiving blood transfusions. The probability of hospitalization for any cause was greater for patients with grafts for vascular access than for those with fistulae, for those with a history of cardiovascular disease, for those with a serum albumin level less than or equal to 30 g/L, and for those with renal disease due to diabetes or vascular disease. Hospitalization due to circulatory disease was more likely among those with a history of cardiovascular disease and among those with a lower serum albumin level. Hospitalization for infectious disease was more likely among those with a lower serum albumin level and less likely among those with a fistula for vascular access. Among all patients receiving hemodialysis treatment for more than 6 months, there were 14.8 hospital days per year. The probability of any infection other than local access infection was greater for subjects with a prosthetic graft than for those with a fistula, for those with a serum albumin level less than or equal to 30 g/L, and for aboriginal compared with nonaboriginal people. The factors associated with septicemia were type of vascular access, serum albumin level, and race. For pneumonia, the relative risk was greater for those with a low serum albumin level. The probabilities of thrombosis or local access infection were determined among 347 hemodialysis patients with either a fistula or polytetrafluoroethylene (PTFE) graft as a first permanent vascular access. Follow-up commenced with the first successful use of the access. The probability of an access infection by 12 months was 5.5% for fistulae and 19.7% for grafts. The probability of graft thrombosis requiring surgical correction by 12 months was greater among subjects with grafts and those with a serum albumin value less than or equal to 30 g/L. For those with a serum albumin level less than or equal to 30 g/L, the probability of thrombosis was 18.4% for fistulae and 39.9% for grafts. For those with a serum albumin level greater than 30 g/L, the probabilities were 14.4% and 32.3%. The objective of this study was to determine the probabilities of specific morbid events or death among patients with end-stage renal disease (ESRD) treated by hemodialysis. A prospective cohort study was performed between March 1988 and September 1989 in 18 hemodialysis centers in 13 Canadian cities, representing about one third of the hemodialysis population in Canada. The inception cohort consisted of 496 patients entering hemodialysis who had survived 1 month. The few new hemodialysis patients who received erythropoietin (EPO) in the last 3 months of the study were excluded. Survival curves were compared using the Cox proportional hazards regression model. Older age and history of cardiovascular disease were independently associated with a greater probability of death. Age and history of cardiovascular disease were also associated with a greater probability of nonfatal circulatory events (myocardial infarction, angina requiring hospitalization, or stroke), while a serum albumin level less than or equal to 30 g/ L (3.0 g dL) was associated with an increased probability of pulmonary edema. The probability of surviving 12 months without receiving a blood transfusion was 47.2% for males and 27.5% for females. The incidence of non-A, non-13 hepatitis, as estimated by unexplained elevations in serum aspartate aminotransferase (AST) values, was not different between patients receiving and not receiving blood transfusions. The probability of hospitalization for any cause was greater for patients with grafts for vascular access than for those with fistulae, for those with a history of cardiovascular disease, for those with a serum albumin level less than or equal to 30 g/L, and for those with renal disease due to diabetes or vascular disease. Hospitalization due to circulatory disease was more likely among those with a history of cardiovascular disease and among those with a lower serum albumin level. Hospitalization for infectious disease was more likely among those with a lower serum albumin level and less likely among those with a fistula for vascular access. Among all patients receiving hemodialysis treatment for more than 6 months, there were 14.8 hospital days per year. The probability of any infection other than local access infection was greater for subjects with a prosthetic graft than for those with a fistula, for those with a serum albumin level less than or equal to 30 g/L, and for aboriginal compared with nonaboriginal people. The factors associated with septicemia were type of vascular access, serum albumin level, and race. For pneumonia, the relative risk was greater for those with a low serum albumin level. The probabilities of thrombosis or local access infection were determined among 347 hemodialysis patients with either a fistula or polytetrafluoroethylene (PTFE) graft as a first permanent vascular access. Follow-up commenced with the first successful use of the access. The probability of an access infection by 12 months was 5.5% for fistulae and 19.7% for grafts. The probability of graft thrombosis requiring surgical correction by 12 months was greater among subjects with grafts and those with a serum albumin value less than or equal to 30 g/L. For those with a serum albumin level less than or equal to 30 g/L, the probability of thrombosis was 18.4% for fistulae and 39.9% for grafts. For those with a serum albumin level greater than 30 g/L, the probabilities were 14.4% and 32.3%.