TROUGH SERUM VANCOMYCIN LEVELS PREDICT THE RELAPSE OF GRAM-POSITIVE PERITONITIS IN PERITONEAL-DIALYSIS PATIENTS

We reviewed 31 episodes of gram-positive peritonitis that occurred in our peritoneal dialysis population between 1990 and 1993 in an attempt to identify the risk factor(s) for peritonitis relapse. All patients were treated with 4 weekly doses of intravenous vancomycin. Vancomycin doses no. 1 and 2 were based on body weight (15 mg/kg with a 1-g mini mum); vancomycin doses no. 3 and 4 were adjusted in an attempt to main tain the trough serum vancomycin level at greater than 12 mg/L. Nine p eritonitis episodes complicated by a relapse were identified. Peritoni tis episodes preceding a relapse were similar to relapse-free episodes with respect to patient age, diabetes, peritoneal dialysis modality, duration of peritoneal dialysis treatment, residual urea clearance, pe ritoneal fluid cell count, causative organism, and weekly vancomycin d ose. However, cumulative 4-week mean trough vancomycin levels were con sistently lower during peritonitis episodes preceding a relapse (7.8 /- 0.6 mg/L during relapse-prone episodes v 13.7 +/- 0.9 mg/L during r elapse-free episodes; P = 0.0004). Furthermore, relapses developed dur ing nine of 14 peritonitis episodes demonstrating a 4-week mean trough vancomycin level less than 12 mg/L compared with zero of 17 episodes with a 4-week trough level greater than 12 mg/L (P < 0.05). The detect ion of a low initial 7-day trough vancomycin level also was a useful m arker for subsequent peritonitis relapse. In 13 peritonitis episodes a ssociated with an initial trough level less than 9 mg/L, nine were com plicated by a relapse. The mean trough vancomycin level persisted belo w 9 mg/L over the final 3 weeks of therapy in all nine relapse-prone e pisodes as opposed to only one of four relapse-free episodes. All peri tonitis relapses were treated successfully with 4 weeks of intravenous vancomycin therapy. This coincided with a significantly higher 4-week mean trough vancomycin level during treatment of the relapses (14.7 /- 0.7 mg/L during the relapse v 7.8 +/- 0.6 mg/L immediately precedin g the relapse; P < 0.05) despite similar mean weekly vancomycin dosage s during both periods (19.4 +/- 1.0 mg/kg during the relapse v 17.8 +/ - 1.2 mg/kg prior to the relapse; P = NS). In patients with uncomplica ted peritoneal dialysis-related gram-positive peritonitis treated with weekly intravenous vancomycin, we conclude that (1) a suboptimal trou gh serum vancomycin level (cumulative 4-week trough serum vancomycin l evel <12 mg/L or an initial 7-day trough serum level <9 mg/L) is the o nly clinical parameter that identified peritonitis episodes at risk fo r relapse, and that (2) on detection of a suboptimal initial trough le vel (<9 mg/L), prevention of a subsequent peritonitis relapse may be p ossible if adequate trough vancomycin levels (>9 mg/L and preferably > 12 mg/L) can be maintained for the remainder of therapy. (C) 1995 by t he National Kidney Foundation, Inc.