Advanced glycation end products (AGEs) may contribute to peritoneal and cardiovascular damage in peritoneal dialysis (PD) patients, possibly in part by over-expression of vascular cell adhesion molecule-1 (VCAM-1). It has been suggested that peritoneal glucose load, oxidative stress, as well as the uraemic state itself may lead to an increased formation of AGEs. Aims of the present study were first to investigate the relation between residual glomerular filtration rate (rGFR), malondialdehyde (MDA) as a marker of lipid peroxidation, and peritoneal glucose prescription and absorption with serum levels of VCAM-1 and with the well characterized AGEs N(epsilon)-(carboxymethyl)lysine (CML) and N(epsilon)-(carboxyethyl)lysine (CEL), as well as with CML and CEL in peritoneal effluent. CML and CEL were measured by tandem mass spectroscopy, MDA by HPLC, and VCAM-1 by ELISA in 37 stable PD patients (age 54 +/- 12 years; time on PD 25 +/- 18 months). CML and CEL were also measured after a 4-month interval. rGFR was independently related to CML both in serum (r = -0.66; P<0.001) and effluent (r = -0.62; P<0.001), whereas peritoneal glucose prescription and absorption were, respectively, related to CML in serum and effluent (r = 0.49; P<0.001 and r = 0.44; P<0.01). Relationships were comparable when assessed after the follow-up period. Peritoneal glucose absorption (r = 0.37; P<0.05), but not rGFR, was related to CEL in serum. The relation between peritoneal glucose prescription and CML in effluent lost significance when rGFR was added to the multi-regression model. Both rGFR (r = -0.40; P<0.05) and peritoneal glucose absorption (r = 0.37; P<0.05) were associated with VCAM-1 expression, which was itself weakly related only to CML in effluent (r = 0.38; P<0.05). MDA was not related to any parameter. Peritoneal glucose prescription and absorption, as well as rGFR are related to serum and effluent levels of CML and to VCAM-1 expression in serum, whereas peritoneal glucose absorption was related to serum levels of CEL. Still, the effect of rGFR, which does not appear to be mediated through lipid peroxidation pathways, on effluent levels of CML appears to outweigh the effect of the PD treatment. Even small differences in residual renal function in patients already on dialysis therapy are related to large variations of CML in serum and the peritoneal cavity.