Peritonitis complicating continuous ambulatory peritoneal dialysis (CAPD) can be used as an in vivo model to study the contribution of mediators in dialysate to the regulation of peritoneal permeability. Previously we reported that changes in the peritoneal appearance rates of the cytokines interleukin-6 (IL-6) and tumor necrosis factor-alpha (TNF alpha) were related to alterations in the effective peritoneal surface area. Changes in the intrinsic peritoneal permeability were mainly related to those in the peritoneal appearance rate of the prostanoid prostaglandin E2 (PGE2) and partly also to that of IL-6. In this intervention study the role of these mediators was further analyzed. Eleven peritonitis episodes were followed on 8 consecutive days from the start of the infection and once after recovery. Indomethacin was given intraperitoneally during the first 3 days. beta 2-Microglobulin clearance was used as indicator of the effective peritoneal surface area. The intrinsic peritoneal permeability was characterized functionally by the restriction coefficient. The 15 peritonitis episodes studied previously served as the control group. This study supports the formerly obtained relationships in two ways. First, significant reductions were observed for peritoneal PGE2, 6-keto-PGF1 alpha, and TxB2 during cyclooxygenase inhibition to 6%, 0.6%, and 9% of the values on day 1, whereas simultaneously the intrinsic permeability was less increased. This indomethacin effect on intrinsic permeability was not entirely significant, probably because of the additional role of IL-6, which was not influenced by indomethacin. Also, the appearance rate of TNF alpha in the effluent was not affected by cyclooxygenase inhibition. Accordingly, the changes in the effective surface area were similar to those in the control group. Second, in 8 of the 11 cases, new rises both in peritoneal PGE2 and in intrinsic permeability occurred after discontinuation of indomethacin. Rebounds were not seen for TNF alpha or IL-6, and, consistently, not for the effective surface area. In conclusion, local cyclooxygenase inhibition results in a less-increased intrinsic permeability during peritonitis but has no effect on the effective surface area. These data support our previous finding that IL-6 and TNF alpha contribute to alterations in surface area, whereas PGE2 is more involved in intrinsic peritoneal permeability changes.