Continuous veno-venous haemodialysis with a novel bicarbonate dialysis solution: prospective cross-over comparison with a lactate buffered solution.
OBJECTIVE: To compare acid-base balance, lactate concentration and haemodynamic parameters during continuous veno-venous haemodialysis (CVVHD) using bicarbonate or a lactate buffered dialysate. METHODS DESIGN: prospective randomized cross-over design; SETTING: Multicentre combined adult surgical and medical intensive care units. Patients; 26 critically ill patients starting CVVHD for acute renal failure. INTERVENTIONS: Each patient to receive 48 h of bicarbonate dialysate and 48 h of lactate dialysate with the order of the 48 h block randomized at trial entry. RESULTS: The serum bicarbonate increased from baseline in both the lactate and bicarbonate groups over the first 48 h of treatment (16.3+/-1.53 to 22.2+/-1.41 mmol/l and 18.9+/-2.02 to 22.2+/-1.18 mmol/l, respectively) and continued to rise towards normal over the next 48 h after cross-over to the other dialysate. The H+ and pCO2 only trended higher in the lactate group. Unlike the acid base parameters, serum lactate levels varied depending on the dialysate composition. The patients initially randomized to the lactate dialysate had higher serum lactate levels and these tended to increase further after 48 h of dialysis from 2.4+/-0.8 to 2.6+/-0.4 mmol/l. However, in the following 48 h the lactate levels fell to 1.8+/-0.6 (P = 0.039) while patients were being treated with the bicarbonate dialysate. Similar results were seen in the patients initially randomized to the bicarbonate dialysate. Serum lactate remained stable over the first 48h (1.4+/-0.2 to 1.5+/-0.1 mmol/l) but after cross-over to the lactate dialysate increased to 3.1+/-0.7 mmol/l (P = 0.051). Overall, lactate levels were significantly higher during dialysis with lactate buffered solution than bicarbonate buffered solution (2.92+/-0.45 vs. 1.61+/-0.25 mmol/l P = 0.01). Mean arterial pressure trended higher during bicarbonate dialysis but did not reach statistical significance (lactate vs. bicarbonate; 71.1+/-3.1 vs. 81.3+/-5.8 mm Hg). Subgroup analysis of the patients with abnormal liver indices or increased lactate levels at initiation of dialysis (n = 15) revealed only a trend toward better bicarbonate control (lactate vs. bicarbonate; 22.00+/-1.73 vs. 22.86+/-1.09, P = 0.2). However, in this group with hepatic insufficiency elevations in serum lactate were even greater during lactate compared to the bicarbonate dialysis (3.39+/-0.68 vs. 1.78+/-0.42 P = 0.036). Patients who had elevations of lactate during lactate dialysis had a high mortality (6 of 7). These patients had an even greater disparity in lactate levels (4.3+/-1.4 vs. 1.3 +/-0.3) and blood pressure (68.0+/- 7.7 vs. 87.2+/-17.1) between lactate and bicarbonate dialysis. Due to small patient numbers these comparisons did not achieve statistical significance. CONCLUSION: During continuous veno venous haemodialysis a bicarbonate buffered dialysis solution provided equal acid-base control but maintained more normal lactate levels than a lactate buffered dialysis solution.