A consensus report on the dose of dialysis: Quantity and time

Abstract

IN THIS ISSUE, Owen (see p S39) and Port et al (see p S34) discuss their views of how changes in the dose of hemodialysis during the past decade have impacted on outcome in our patients. Port et al conclude, based on the data compiled by the US Renal Data System in their special studies, that changes in the hemodialysis dose prescription itself have resulted in substantial improvements in the survival of patients. The delivered dose of hemodialysis, as determined by Kt/V, has increased, as has the use of bicarbonate-buffered dialysate and the use of biocompatible membranes. These trends have been associated with lower mortality rates. Of interest, the improvement in Kt/V has been caused by higher clearances (K) because of the introduction of higher blood flows and dialyzers with a higher urea KoA. The length of the treatment session has only increased to a minor extent. The issue of time as an independent predictor of survival has yet to be shown, but it is a variable that may eventually be shown to impact on survival in subsequent talks by Charra, Pierratos, and Levin. Owen also examines survival in a large database and shows increasing hemodialysis dose as measured by the urea reduction ratio. However, he points out that attention to other indices not directly associated with the hemodialysis prescription may have also been responsible for improved patient outcome. The monitoring of serum albumin levels and improvement in hematocrits because of the use of erythropoietin have also been a feature of management of the hemodialysis patient that has occurred in the past decade. In reality, the conclusions expressed by Owen and Port et al do not conflict. The data on hemodialysis dose have been developed from cross-sectional data. Port et al correctly point out that observational data cannot prove causation. Yet the findings they present are compelling and have led to the development of the HEMO Study, which will examine hemodialysis dose and the membrane effect in a prospective fashion. In addition, it is obvious that hemodialysis itself is an imperfect form of renal replacement therapy, and the other measures suggested by Owen will also impact on our patients. As for time of the dialysis session, the way that a given prescription of hemodialysis is delivered is quite variable. Dialysis time can vary from as little as 2.5 hours to as long as 8 hours. Patients can receive their treatments three times weekly or daily/nightly. Blood flow rates vary from 150 to 200 mL/min to greater than 400 mL/min. In the extreme, Charra et al (see p S63) have traditionally delivered hemodialysis for 3 hours three times each week at blood flows of 200 mL/min or greater using a 1-m2 Kiil dialyzer. The group only began to switch to bicarbonate-buffered dialysate in 1995 and adopted a shorter hemodialysis regimen of 5 hours/wk at a blood flow of 300 mL/min. Their average Kt/V was 1.97 on the former regimen and 1.78 on the latter. Their patient survival is enviable. Charra et al point out that long dialysis times improve middle-molecule clearance and allow volume and blood pressure control without the need for antihypertensive agents. At the very high dose of hemodialysis, Kt/V had no effect on survival with multivariate analysis, whereas middle-molecule clearance showed a significant effect. Predialysis mean arterial pressure and serum albumin levels were the strongest predictors of mortality. Inorganic phosphorous levels were also more easily controlled. These investigator’s findings with respect to hypertension and inorganic phosphorous control have been confirmed by patients being dialyzed nightly at low blood and dialysate flows, but for long treatment sessions, as reported by Pierratos in this issue (see p S76). At the other end of the spectrum, Levin (see p S83) states that long dialysis sessions are not required for all patients. He cites the results of Gotch et al, who have reported excellent survival in patients with average treatments of only 2.5 From the Division of Nephrology, Vanderbilt University Medical Center, Nashville, TN. Address reprint requests to Gerald Schulman, VUMC, Division of Nephrology, S 3223 Medical Center North, Nashville, TN 37232. E-mail: gerald.schulman@mcmail. vanderbilt.edu 1998 by the National Kidney Foundation, Inc. 0272-6386/98/3206-0415$3.00/0