Intermittent hemodialysis is considered the modality of choice when enhanced lithium removal is indicated. However, postdialysis rebound in serum lithium concentration is frequently observed after the dialysis sessions and results from incomplete intracellular removal. Continuous renal replacement therapy could provide a more gradual and complete lithium removal since it is performed over longer time periods, thus avoiding rebound following therapy. Seven patients presenting with symptomatic lithium intoxication were treated by continuous renal replacement therapy (continuous arteriovenous and venovenous hemodiafiltration [CAVHDF and CVVHDF]). For CAVHDF, the dialysate flow rate was increased to 4 L/hr to optimize solute clearances. Five intoxicated patients (four acute and one chronic) were treated by high dialysate flow rate (HDFR) (4 L/hr) CAVHDF and two patients with chronic poisoning were treated by CVVHDF, one with a dialysate flow rate of 1 L/hr and one with a dialysate flow rate of 2 L/hr. Serum lithium concentrations for the four acute poisoning cases were 4.0, 4.6, 4.4, and 3.2 mEq/L, at initiation of HDFR CAVHDF, and decreased respectively to 1.2, 0.8, 1.2, and 1.1 mEq/L after 15, 19, 35, and 21 hours of treatment. No lithium rebound was observed over 24 to 36 hours following CAVHDF. For the three chronic intoxication cases, serum lithium concentrations dropped from 1.7, 2.2, and 3.8 mEq/L to 0.7, 0.17, and 0.4 mEq/L, respectively, after 18, 42, and 44 hours of HDFR CAVHDF or CVVHDF. The chronic case treated for only 18 hours presented a slight rebound in lithium level (0.3 mEq/L), whereas no significant rebound was observed for the two other cases treated for longer periods. Mean +/- SEM dialyser urea, lithium, and creatinine clearance during HDFR CAVHDF were 50.5 +/- 5.0, 41.4 +/- 4.6, and 37.6 +/- 3.7 mL/min, respectively (number of measurements = 41). Dialyser lithium clearance during CVVHDF was 48.4 +/- 1.4 mL/min (n = 10) and 61.9 +/- 2.3 mL/min (n = 7), with dialysate flow rates of 1 and 2 L/hr, respectively. Mean dialyzer lithium removal for the seven cases was 106.4 mEq, while mean renal lithium removal was 21.5 mEq during the same period. We conclude that HDFR CAVHDF and CVVHDF are effective alternatives to intermittent hemodialysis for treatment of lithium poisoning. They provide excellent lithium clearances (60 to 85 L/d); in addition, because of their continuous nature, they prevent posttherapy lithium rebound by allowing a more gradual and complete removal from intracellular compartments, and they may be particularly useful in chronic poisoning in which intracellular lithium accumulation is more extensive.