Prescription and delivery of hemodialysis (HD) in acute renal failure (ARF) may be affected by patient-related factors such as hemodynamic instability, catabolism, variable extracellular fluid volume, and coagulation disturbances. This study was undertaken in a cohort of patients with ARF requiring HD, to quantify patient- and dialysis-related variables that influence dialysis delivery. The urea reduction ratio (URR) was calculated for each HD session. Patient-related variables included age, gender, weight, mean arterial pressure, and Acute Physiological and Chronic Health Evaluation (APACHE) II and Multiple Organ Failure (MOF) scores. Dialysis-related variables were dialyzer characteristics (membrane type, surface area, KoA, and K(UF)), blood flow rate (Qb), session length, anticoagulant use, vascular access, and ultrafiltration volume. The analysis of factors associated with URR was performed using mixed linear statistical models. The cohort consisted of 81 adult patients with ARF who underwent 419 consecutive dialysis sessions. Mean (+/- SD) age was 60 +/- 18 years; 57% were male. At dialysis initiation, APACHE II score was 23 +/- 8 and MOF score 2 +/- 1. The number of HD treatments averaged 5.5 +/- 6.1/patient and 0.8 +/- 0.2/patient/day, mean URR was 54 +/- 15%, and session length 3.2 +/- 1.1 h; 58% used a femoral venous catheter, and 92% polysulfone dialyzers. Among patient-related variables, the only independent predictor of delivered dose of dialysis, as measured by URR was the predialysis weight (P < 0.01). Among the dialysis-related variables, treatment time (P < 0.01), dialyzer surface area (P < 0.01), dialyzer K(UF) (P = 0.04), blood flow rate (P < 0.01), and the use of a femoral venous catheter (P < 0.01) were also independently associated with URR. An interaction between vascular access site and blood flow rate was also found to be significant (P < 0.01). This study underscores the importance of the dialysis prescription parameters and vascular access site in influencing the dialysis dose in critically ill patients, and argues against the importance of patient-related characteristics such as disease severity.