In computer simulations, according to the three-pore model of peritoneal transport, neutralization of conventional acidic peritoneal dialysis fluids is predicted to produce an improved ultrafiLtration (UF). However, in a previous study, a two-compartment peritoneal dialysis system with a minimum of glucose degradation products (GDP), PD-Bio, having a pH of 6.3 and being conventionalLy lactate buffered, did not produce an increased UF. We tested a newly formulated, glucose-based, GDP-reduced solution, denoted "N" for "neutral," containing a mixture of lactate (30 mmol/L) and bicarbonate (10 mmol/L) as buffer system, and having a pH of 7.2. This new formulation was compared with Gambrosol trio (GT) (identical in composition to PD-Bio, but delivered in a three-compartment system; both by Gambro Lundia AB, Lund, Sweden) in an open, prospective controlled study of 13 patients. Each of the 13 patients used GT for 14 days, followed by 14 days of N. All bags were weighed on a digital scale before instillation and after drainage to assess the UF in each dwell (and during 24 hours). Glucose concentration in each bag was noted. In the morning and night dwells, dialysis fluid glucose concentration was standardized to 2.5%. Body weight was measured every morning (empty abdomen). In the middle of each 14-day period, a 4-hour standardized ("study day") dwell was performed, using 125I-albumin (RISA) as volume marker, during which blood and dialysate samples were taken repeatedly and anaLyzed for RISA, creatinine, urea, phosphate, glucose, standard bicarbonate, lactate, and pH. The permeability surface area product (PS) for small solutes (and A0/deltaX; "area parameter") was calculated. Clearance (Cl) of RISA to plasma (P) (C-->P), "direct lymphatic absorption," and total Cl of RISA out of the peritoneal cavity (Cl(out)) were also determined. The 13 patients using N, compared to GT, displayed an increased daily UF, the difference being 233 mL (p < 0.05). The pH values of N were higher during the first 90 minutes of the 4-hour dwell compared to the pH values of GT. Neither the small solute PS values nor RISA determined UF, nor did body weight differ significantly between the GT and the N periods. A new bicarbonate/lactate-buffered solution, N, with neutral pH (of 7.2) and low in GDP seems to produce an improved UF compared to a lactate-buffered solution with a pH of 6.3, equally low in GDP, partly in agreement with our earlier predictions. A dialysis solution with a neutral pH combined with a reduced lactate concentration, partially replaced by bicarbonate, evidently increases UF, conceivably by causing less peritoneal vasodilatation than solutions buffered by lactate or high concentrations of bicarbonate alone.