BACKGROUND: Starting renal replacement therapy (RRT) for acute renal failure in critically ill patients with haematological malignancies is controversial because of the poor outcome and high costs. The aim of this study was to compare the outcome between critically ill medical patients with and without haematological malignancies who received RRT for acute renal failure. METHODS: We retrospectively collected data on all consecutive patients who received RRT for acute renal failure at the Medical Intensive Care Unit (ICU) of a University Hospital between 1997 and 2002, and assessed the impact of the presence of a haematological malignancy on the survival within 6 months after ICU admission by Cox proportional hazard models. RESULTS: Fifty of the 222 (22.5%) consecutive patients with haematological malignancies admitted to the ICU over the study period received RRT for acute renal failure compared with 248 of the 4293 (5.8%) patients without haematological malignancies (P<0.001). Among patients who received RRT, those with haematological malignancies had higher crude ICU (79.6 vs 55.7%, P=0.002) and in-hospital (83.7 vs 66.1%, P=0.016) mortality rates, and a higher mortality at 6 months (86 vs 72%, P=0.018) by Kaplan-Meier estimates compared with those without haematological malignancies. However, after adjustment for the severity of illness and the duration of hospitalization before ICU admission, haematological malignancy by itself was no longer associated with a higher risk of death (hazard ratio 1.04; 95% confidence interval, 0.73-1.54, P=0.78). CONCLUSIONS: Medical ICU patients with haematological malignancies have a higher rate of occurrence of acute renal failure treated with RRT and a higher mortality, compared with those without haematological malignancies. However, the presence of a haematological malignancy by itself is not a reason to withhold RRT in medical ICU patients with acute renal failure.