Although several dosage adjustment regimens have been proposed, there is little quantitative information to guide the initiation of ceftazidime therapy in patients who are receiving continuous renal replacement therapy. To determine the clearance of ceftazidime by continuous venovenous hemofiltration (CVVH) and continuous venovenous hemodialysis (CVVHD), we performed controlled clearance studies with stable hemodialysis patients with three hemofilters: a 0.6-m² acrylonitrile copolymer (AN69; Hospal) filter, a 2.1-m² polymethylmethacrylate filter (PMMA; Toray) filter and a 0.65-m² polysulfone (PS; Fresenius) filter. Subjects received 1,000 mg of ceftazidime intravenously prior to the start of a clearance study. The concentration of ceftazidime in multiple plasma and dialysate or ultrafiltrate samples was determined by high-performance liquid chromatography. The diffusional clearances (CI_diffusion) and sieving coefficients of ceftazidime were compared by a mixed-model repeated-measures analysis of variance with filter and blood, dialysate inflow, or ultrafiltration rate as the main effect and the patient as a random effect. The fraction of ceftazidime bound to plasma proteins was 17% ± 7% (range, 10 to 25%). The clearances of ceftazidime, urea, and creatinine by CVVHD were essentially constant at blood flow rates of 75 to 250 ml/min for all three filters. Significant linear relationships (P < 0.0001) were observed between CI_diffusion of ceftazidime and clearance of urea for all three filters: AN69 (slope = 0.83), PMMA (slope = 0.89), and PS (slope = 1.03). Ceftazidime clearance was membrane independent during CVVH and CVVHD. CVVH and CVVHD can significantly augment the clearance of ceftazidime. Dosing strategies for initiation of ceftazidime therapy in patients receiving CVVH and CVVHD are proposed.