The combined annual mortality and drop out rate for peritoneal dialysis (PD) patients is relatively uniform worldwide at approximately 35%. The level of PD therapy prescribed in clinical practice is largely empirical and typically consists of four 2-L exchanges daily. It might be speculated that the 35% annual attrition rate in PD may in part reflect underdialysis in some patients due to empirical rather than quantitative and individualized prescription of PD therapy. Urea kinetic modeling has been successfully used to quantitatively prescribe hemodialysis (HD) therapy and, in principle, it should be able to serve the same purpose in PD. Comparison of HD and PD is complicated, because peritoneal urea clearance is virtually continuous, while in HD clearance is provided only about 5% of the time and urea acculumlates over 95% of each treatment cycle. The blood urea nitrogen (BUN) in PD (BUNpd) is essentially constant and reflects the steady state, while in HD a sawtooth BUN profile results that reflects the short intermittent dialyses. The HD BUN profile can be characterized by either the predialysis level (BUNo) or the time-averaged concentration (TAC) over each treatment cycle. TAC is substantially lower than BUNo due to the sharp obligatory decrease in BUN during each short highclearance dialysis. The rate of clearance required in HD is approximately 30 times higher than in PD, and the total clearance (KT) required in HD is 50% higher than in PD to achieve BUNpd = BUNo (at identical normalized protein catabolic rate [NPCR]), which reflects the decreasing urea flux rate during HD due to the decreasing BUN. Intuitively, it would seem that the higher KT in HD is most likely simply an artifact of the intermittency of therapy with high clearances required during each dialysis to achieve BUNpd = BUNo. However, it has been argued both that (1) the higher KT in HD relects “more” and therefore better therapy and, conversely, (2) the lower KT in PD with similar clinical outcome must reflect higher removal rates of undefined middle molecular weight toxins by the peritoneal membrane. Kinetic analyses of the liminted data available indicate (1) that adequate dialysis in PD is in fact achieved with BUNpd = BUNo and requires a KT in PD that is only about two thirds of that required in HD; (2) that adequate daily peritoneal plus residual renal urea clearance (KprT) divided by body water (V), KprT/V, is 0.29 for NPCR less than or equal to 1.1; and (3) that in current clinical practice, KprT/V appears to be uniformly distributed over a wide range of 0.15 to 0.40. These results suggest there may be a substantial incidence of underdialysis in PD and indicate the need for a prospective, randomized study of quantified PD therapy to define adequate treatment. © 1993, National Kidney Foundation, Inc.. All rights reserved.
