Peritoneal glucose kinetics were evaluated in the anaesthetized rat, to assess whether the peritoneal cavity would be a suitable site for the implantation of membrane-protected islets of Langerhans (bioartificial pancreas) or the glucose sensor of an artificial B cell. Glucose was measured in peritoneal fluid samples aspirated by needle puncture. Basal peritoneal and blood glucose concentrations were identical in 16 h fasted (n=4) and non fasted (n=3) animals. After 10 min of an i.v. glucose infusion (n=15) the increment in peritoneal glucose concentration was 63±3% of the increment in blood glucose concentration and both values were significantly correlated (r=0.92; p<0.001). After 10 min of glucose clamping (12.6±0.8 mmol/l), the increment in peritoneal glucose concentration was 69±3% (n=5; p<0.05) of the increment in blood glucose concentration. In three additional experiments it was 93±3% of the increment in blood glucose concentration (NS), after 30 min of glucose clamping (8.0±0.5 mmol/l). Peritoneal glucose concentration monitored by a glucose sensor: (a) followed blood glucose sluggishly during a glucose clamp (n=5), confirming the data shown above, (b) followed blood glucose with a 5 min delay and reached the same plateau after the intravenous injection of 1U insulin (n=3; NS). We conclude that peritoneal glucose reflects blood glucose at basal state and during variations of glycaemia, nevertheless, presenting heterogeneous kinetics. These kinetics might be appropriate for a bioartificial pancreas but not for an in vivo calibration procedure, of a peritoneally implanted glucose sensor.