Pharmacokinetic analysis of creatinine generation discrepancy as an index of noncompliance in CAPD

The limitations of the ratio of estimated-over-predicted creatinine generation (E/P) as an index of noncompliance were investigated in hypothetical continuous ambulatory peritoneal dialysis (CAPD) populations with steady-state E/ P ranging between 0.5 and 1.5 (mean 1.0), anuria or urine creatinine clearance of 50 L weekly, and omission during the noncompliant state of one, two, or three daily exchanges from a schedule of four daily exchanges, which would provide, in the compliant state, a peritoneal creatinine clearance of 50 L weekly. Standard pharmacokinetic methods were applied to estimate both serum creatinine concentration at the end of the first 24-hour period after resumption of the schedule with four daily exchanges and the amount of creatinine lost from the body in the same nonsteady-state period. The shift of the E/P frequency distribution curve to higher values between the steady and nonsteady states was sufficient to create a clear separation of these curves only in anuria with three-exchange daily noncompliance (nonsteady state: E/P 1.02 - 3.04, with a mean of 2.03). In all other cases, the shift of the E/P curve was small and did not create a clear separation of the E/P frequency distribution curves. Because of the wide variation of the E/P values in the steady state and the relatively small shift of E/P distribution to higher values in the nonsteady state after a sudden increase in total creatinine clearance, E/P is not a suitable test for noncompliance in cross-sectional studies.