Background: Hemofiltration protocols using a citrate-buffered replacement solution offer the advantage of regional anticoagulation and a buffer effect. The role played by such fluids in clinical practice is not yet well established. The risk of electrolytic disorders, acid-base imbalance, or citrate accumulation should be clarified. We report on a renal therapy protocol based on a citrate isonatremic replacement solution. Method: We considered all patients needing renal replacement therapy admitted to our cardiovascular intensive care unit between January 2003 and June 2007. A citrate-buffered fluid was delivered in pre-dilution mode to a post-filter ionized calcium target <= 0.25 mmol/L. Extracorporeal blood flow was set at a constant of 140 +/- 10 ml/min. Blood calcemia was maintained by a 5% calcium-chloride solution infused into the patient. We recorded the patients' acid-base variables, ionized calcium, daily electrolytes, albumin, urea and filter life-span. Results: We observed 101 consecutive patients out of 2,523; incidence 4%, overall mortality was 57% at ICU discharge. Mean replacement rate was 2,554 +/- 475 ml/h corresponding to 34 5 ml/kg/h. Mean patient ionized calcium level was 1.07 +/- 0.04 mmolL, maintained by 13 +/- 2 ml/h of infused calcium-chloride. All other electrolytes remained in the normal range. The Stewart biophysical approach confirmed a strong anion gap of 3.1 +/- 3 meq/L. Acid-base balance showed a buffer effect. Mean filter life-span was 52 +/- 11 h. Conclusion: Penal replacement therapy based on citrate-buffered fluid may be useful in clinical practice. This methodology presented an adequate metabolic control and allowed regional anticoagulation. A sufficient calcium supply was mandatory to avoid hypocalcemia. The small strong ion gap suggested a modest citrate accumulation. (Int J Artif Organs 2008; 31: 937-43)