To investigate the biocompatibility of "new" peritoneal dialysis (PD) solutions with bicarbonate/lactate buffer, non glucose osmotic agents (icodextrin or amino acids), neutral pH, and low levels of glucose degradation products (GDPs). Using M199 culture medium as a control, we compared conventional and new PD solutions with respect to their effects on the viability of human peritoneal mesothelial cells (HPMCs) [using lactate dehydrogenase (LDH) release], on DNA damage in HPMCs [using single-cell gel electrophoresis (Comet assay)], and on HPMC proliferation (using [3H]-thymidine incorporation). The experiments were performed after cell growth was synchronized by incubation with serum-free media for 24 hours. The PD solutions tested included commercial 1.5% glucose and 4.25% glucose solutions with 40 mmol/L lactate (D 1.5 and D 4.25, respectively), 7.5% icodextrin (E), 1.1% amino acid (N), 1.5% glucose solution in a triple-chambered bag (Bio 1.5), 1.5% glucose solution in a dual-chambered bag with neutral pH (Bal 1.5), and 1.5% glucose and 4.25% glucose solution containing 25 mmol/L bicarbonate and 15 mmol/L lactate (P 1.5 and P 4.25, respectively). When HPMCs were continuously exposed to undiluted PD solutions, D 1.5, D 4.25, P 4.25, and E increased LDH release by more than 60% at 24 hours. All PD solutions tested increased LDH release by more than 75% at 96 hours. With 2-fold diluted PD solutions, only D 4.25 significantly increased LDH release at 96 hours, though not at 24 hours. When cells were exposed to undiluted PD solutions for 60 min and allowed to recover in M199 for up to 96 hours, LDH release was significantly higher at 24-96 hours in E (55%-69%) and D 1.5 (48%-72%) as compared with control [M199 (18%)]. Release of LDH was significantly lower with PD solutions containing lower levels of GDPs than those in D 1.5, suggesting that GDPs may have a role in cell viability. The D solutions (D 1.5 and D 4.25) and E solution also induced significant DNA damage. Both LDH release and DNA damage by D and E were significantly attenuated by adjusting the solution pH to 7.4, suggesting that low pH may be implicated in PD solution-induced DNA damage and cell death. When diluted 2-fold, D 1.5, D 4.25, and P 4.25 decreased [3H]-thymidine incorporation to 43%, 34%, and 41% of control, respectively, at 24 hours and to 45%, 26%, and 35% of control, respectively, at 96 hours. When cells were exposed to undiluted PD solutions for 5 minutes and allowed to recover in M199 for up to 96 hours, D1.5 and P 4.25--but not D 4.25--significantly inhibited cell proliferation at 24 hours. This effect was sustained up to 96 hours. The present in vitro data demonstrate that PD solutions with low pH, or high levels of GDPs, or both, promote HPMC death and DNA damage, and that PD solutions with high osmolality inhibit cell proliferation. Solutions with neutral pH, amino acids, and "low GDPs" appear to be more biocompatible than conventional PD solutions. These results require confirmation in in vivo animal and clinical studies.