The only widely accepted method to quantify the dose of dialysis is the fractional clearance of urea from body water or Kt/VU, where K is dialyzer urea clearance, t is treatment time and V is the urea distribution volume (KDOQI) which is considered to be a dose surrogate for removal of low molecular weight toxins. There have been two prospectively randomized, controlled trials (RCTs) of dialysis therapy [1, 2] and in both these trials the dialysis dose was tightly controlled with urea kinetic modeling (UKM). In contrast there have been no prospectively RCTs on the effect of treatment time on outcome. Over the years large observational studies (OSs) have not shown a consistent effect of treatment time on mortality until recently two fairly large OSs were report-
© 2007 S. Karger AG, Basel 0253-5068/07/0251-0018$23.50/0
Frank Gotch, MD
144 Belgrave Avenue
San Francisco, CA 94117 (USA)
Tel. +1 415 661 6191, Fax +1 415 731 7876, E-Mail [email protected]
ed which statistically show an association of longer treatment time with lower mortality [3, 4]. However, the validity of OSs, which might be characterized as 'guilt or innocence by association', have recently been seriously questioned . It is important to trace the historical roles that RCTs and OSs have played in understanding the dialysis dose to provide a perspective on the validity of these recent OSs correlating long treatment time with reduced mortality.
The National Cooperative Dialysis Study
160 140 120 100 80 60 40 20
Groups II, IV BUN 120 ± 15 and t = 4.5-5.5 h (II) t = 2.5-3.5 h (IV)
Groups I, III BUN 70 ± 15 and t = 4.5-5.5 h (I) t = 2.5-3.5 h (III)
Dialysis technology was changing rapidly in the 1960s and 1970s. The inefficient Kiil dialyzers used with 8- to 12-hour treatment times were being replaced by new dia-lyzers with smaller blood volumes and more reproducible and higher clearances developed through the stimulus of the Artificial Kidney Chronic Uremia Program of the NIH, but there was no agreement on a definition of adequate dialysis. A consensus was reached at the NIH Conference on Adequacy of Dialysis in 1974 that a prospective controlled trial of dialysis was required. The trial subsequently designed and carried out was the National Cooperative Dialysis Study (NCDS)  which was guided by UKM , and its basic design is shown in figure 1. UKM was used to control blood urea nitrogen (BUN) at 2 nominal levels of 70 8 10 and 120 8 10 mg/dl with protein intake (normalized protein catabolic rate or NPCR, g/kg) controlled to 0.80 < NPCR < 1.40. The protocol also called for short and long treatment times for both low and high BUN groups as shown in figure 1. The short treatment times were targeted at 2.5-3.5 h and long treatment times 4.5-5.5 h and the mean times achieved were 3.2 and 4.5 h, so a substantial difference in treatment time was in fact observed.
Clinical outcome in the NCDS for 10 groups of 16 patients each expressed as percent success or failure is plotted in figure 2A . Groups II and IV with high BUN had a 56% incidence of failure while groups I and III with low BUN had 13% failure which was thought to confirm the hypothesis that high BUN resulted in poor outcome. However, there was a 5th group (V) of patients with NPCR <0.80 that had a very high incidence of failure (75%) which was unrelated to BUN . In an attempt to understand group V, the UKM which was used to guide the study was solved for BUN over a wide range of spKt/ V and NPCR with results shown in figure 2B. When spKt/V is held constant, BUN is a linear function of NPCR as depicted in figure 2B for 0.4 < spKt/V < 2.0.
Fig. 1. The design of the NCDS. The protocol called for identical control of pretreatment BUN as a function of NPCR in the low and high BUN groups with short and long treatment times as described. t = Treatment time.
An analysis of NCDS data with the UKM grid in figure 2B is shown in figure 3 where it can be seen that spKt/V 61.0 clearly separated the patients in groups II, IV and V with a high failure rate from groups I and III with a low failure rate. The analysis in figure 3 is portrayed in figure 4 as the relative probability of failure as a function of spKt/V. These data were interpreted to show that the mechanism by which dialysis controls uremia can be quantified as the magnitude of dialyzer urea clearance (Kt) normalized to urea distribution (V) or Kt/V. It was concluded that an adequate dose of dialysis was achieved with Kt/V = 1.00 and no improvement was seen with higher doses up to spKt/V 1.40. The step function, outcome, shown in figure 4 represents the two levels of dosing in this prospectively randomized database.
Was this article helpful?