- Email: [email protected]
Effect of dialysis membranes and middle molecule removal on chronic hemodialysis patient survival
Effect of Dialysis Membranes and Middle Molecule Removal on Chronic Hemodialysis Patient Survival John K. Leypoldt, PhD, Alfred K. Cheung, MD, Caitlin E. Carroll, BS, David C. Stannard, BS, Brian J.G. Pereira, MD, Lawrence Y. Agodoa, MD, and Friedrich K. Port, MD, MS ● The type of dialysis membrane used for routine therapy has been recently shown to correlate with the survival of chronic hemodialysis patients. We examined whether this effect of dialysis membrane could be explained by differences in dialyzer removal of middle molecules using data from the 1991 Case Mix Adequacy Study of the United States Renal Data System. The sample analyzed included patients who had been treated by hemodialysis for 1 year or more, who were dialyzed with the 19 most commonly used dialyzers in 1991, and for whom delivered urea Kt/V could be calculated from predialysis and postdialysis blood urea nitrogen concentrations. Vitamin B12 (1,355 daltons) was used as a marker for middle molecules, and the clearance of vitamin B12 was estimated based on in vitro data. After adjustments for case mix, comorbidities, and urea Kt/V, the relative risk of mortality for a 10% higher calculated total cleared volume of vitamin B12 was 0.953 (P F 0.0001 v 1.000). Similar results were obtained when middle molecule removal was adjusted for body size. We conclude that both small and middle molecule removal indices appear to be independently associated with the risk of mortality in chronic hemodialysis patients. Differences in mortality when using different types of dialysis membrane may be explained by differences in middle molecule removal. This is a US government work. There are no restrictions on its use. INDEX WORDS: Clearance; dialyzer; hemodialysis patient survival; middle molecules; vitamin B12.
R
ECENT REPORTS from the 1991 Case Mix Adequacy Study of the United States Renal Data System (USRDS) have suggested that two dialysis prescription parameters, dialysis dose as defined by urea Kt/V (or the urea reduction ratio)1 and the type of dialysis membrane,2 play an important role in determining the survival of chronic hemodialysis (HD) patients. Higher values of delivered urea Kt/V (calculated using a single-pool model), up to approximately 1.3, were associated with progressively better survival,1 an effect consistent with that previously reported in other retrospective analyses.3-6 Although the magnitude of the optimal dose of dialysis is debatable,7 the importance of small solute removal, as quantified by urea Kt/V, is well established. The role of dialysis membrane on the survival of chronic HD patients in the USRDS Case Mix Adequacy Study was evaluated by categorizing dialyzers into three groups based on the type of membrane material2: unsubstituted cellulose, modified cellulose, and synthetic. Membrane material was chosen as the classification parameter because of its association with complement activation: unsubstituted cellulose membranes activate complement to a high degree; modified cellulose membranes activate complement to an intermediate degree; and synthetic membranes activate complement to a low degree. It should be noted, however, that dialyzer reprocessing is
common in dialysis centers in the United States,8 and the relationship between membrane material and complement activation is maintained only for dialyzers that are reprocessed using bleach.9-11 Because approximately 50% of all HD centers in the calendar year 1991 reprocessed dialyzers using Renalin (Renal Systems, Plymouth, MN) as the germicide without bleach,8 the proposed relationship between membrane material and
From Research and Medical Services, Salt Lake City Veterans Affairs Medical Center; the Division of Nephrology & Hypertension and Department of Bioengineering, University of Utah, Salt Lake City; United States Renal Data System Coordinating Center and Departments of Internal Medicine and Epidemiology, University of Michigan, Ann Arbor; Division of Nephrology, New England Medical Center, Boston, MA; National Institute of Diabetes and Digestive and Kidney Diseases, National Institutes of Health, Bethesda, MD. Received April 1, 1998; accepted in revised form July 31, 1998. The data reported here have been supplied by The United States Renal Data System (USRDS). The interpretation and reporting of these data are the responsibility of the authors and in no way should be seen as an offıcial policy or interpretation of the United States Government. Supported by USRDS, DVA Medical Research Funds and the Dialysis Research Foundation, Ogden, UT. Address reprint requests to John K. Leypoldt, PhD, Dumke Building 535, University of Utah, Salt Lake City, UT 84112. E-mail: [email protected] This is a US government work. There are no restrictions on its use. 0272-6386/99/3302-0018$0.00/0
American Journal of Kidney Diseases, Vol 33, No 2 (February), 1999: pp 349-355
349
350
LEYPOLDT ET AL
complement activation may not be consistently valid. If the observed effect of dialysis membrane on patient survival was attributable to poor membrane biocompatibility (attributed to complement activation and associated events), then the true effect of dialysis membrane on survival would have been underestimated in that study.2 Another potential explanation for the difference in patient survival associated with the use of various dialysis membranes is their difference in water flux, because modified cellulose and synthetic membranes more often have higher ultrafiltration coefficients than unsubstituted cellulose membranes. The use of ultrafiltration control machines, which are required for dialysis membranes with high ultrafiltration coefficients, may lead to better achievement of dry weight and may explain the better outcomes. Furthermore, the possibility that HD centers who prescribe dialyzers containing modified cellulose or synthetic membranes also provide superior medical care cannot be excluded. Indeed, centers that use dialyzers containing either modified cellulose or synthetic membranes have been shown to commit greater financial resources to direct patient care than centers that use dialyzers containing unsubstituted cellulose membranes.12 Yet another potential explanation for the difference in patient survival relates to the solute clearance profiles of the dialysis membranes. The dialysis membranes that were used on patients in the 1991 Case Mix Adequacy Study of the USRDS differed not only in biocompatibility and flux characteristics but also in their removal rates for middle molecules (solutes substantially larger than urea). In the current study, we have developed an approach for calculating middle molecule clearances or removal rates using in vitro transport data supplied by the dialyzer manufacturer and determined the association of calculated middle molecule removal with the survival of long-term HD patients. METHODS
Middle Molecule Removal We have chosen vitamin B12 as the molecular marker for evaluating middle molecule removal for both historical and practical reasons. Vitamin B12 (1,355 daltons) has long been used as a marker for middle molecules in a number of in vitro and clinical studies,13 and manufacturers frequently report dialyzer clearances for vitamin B12 determined in vitro as a measure of the porosity of the dialysis membrane.
Although the validity of projecting in vitro clearances of vitamin B12 to the clinical HD setting is unknown, calculated clearances of vitamin B12 appear reasonable. Furthermore, for certain low-flux and high-flux dialyzers, the calculated values correlate with clearances of other middle molecules such as 2-microglobulin measured during clinical HD.14 Solute clearance during HD consists of both diffusive and convective contributions. The evaluation of separate diffusive and convective contributions to total clearance is complex, and this topic has been recently reviewed.15,16 Experimental data suggest that the total clearance of vitamin B12 (KB12) during HD can be expressed as a diffusive term plus a convective term by an equation simplified from that recently proposed by Jaffrin16: KB12 ⫽ KB0 12 ⫹ 0.43 ⫻ Qf
(1)
where KB0 12 denotes the diffusive clearance of vitamin B12 in the absence of ultrafiltration and Qf denotes the ultrafiltration rate. Equation 1 has been shown to be approximately valid for sodium, urea, creatinine, vitamin B12, and myoglobin over a range of ultrafiltration rates less than 20 mL/ min.16,17 The ultrafiltration rate averaged over the entire treatment can be calculated from the predialysis body weight (BWpre) and postdialysis body weight (BWpost) as Qf ⫽ (BWpre ⫺ BWpost)/t
(2)
where t denotes treatment time. Thus, the convective component to total vitamin B12 clearance can be calculated using equations 1 and 2 from measurements of predialysis and postdialysis body weight. Diffusive clearance of vitamin B12 is usually reported by dialyzer manufacturers for only a limited number of blood and dialysate flow rate combinations. To calculate KB0 12 under any treatment condition for a given dialyzer, we must first calculate the mass transfer-area coefficient for vitamin B12 (K o AB12) from vitamin B12 clearance at the blood (Qb) and dialysate (Qd) flow rates reported by the manufacturer, using the following equation18: K o AB12 ⫽
Qb 1 ⫹ Qb /Qd
ln
3
1 ⫺ (1 ⫹
1 KB0 12/Qb)
4
⫻ Qb /Qd
(3)
Table 1 shows the calculated values of K o AB12 for the 19 most commonly employed dialyzers during 1991 using diffusive vitamin B12 clearances provided by the manufacturers and reported in recent reviews.19-21 Using the calculated values in this table, equation 3, after rearrangement, can then be used to calculate diffusive vitamin B12 clearance in the absence of ultrafiltration at any blood and dialysate flow rate combination. The use of the equations 1 through 3 for calculating total dialyzer clearance of vitamin B12 is similar to that for calculating urea clearance under any conditions of blood, dialysate, and ultrafiltration flow rates.22 Dialyzer urea K o A values used in this study were identical to those used in a previous study.1
Middle Molecule Removal Parameters Two different parameters were used as measures of middle molecule removal in the current study. The first parameter
MIDDLE MOLECULES AND HD PATIENT SURVIVAL
351
Table 1. Mass Transfer-Area Coefficients (KoA) for Vitamin B12 for the Dialyzers Most Commonly Used in 1991
is directly proportional to the ratio of their clearances. The ratio of vitamin B12 clearance to urea clearance for any dialyzer can be estimated from the respective K o A values and the blood, dialysate, and ultrafiltration rates by using equations 1 through 3. Combining equations 1 and 5 for both vitamin B12 and urea yields an expression for vitamin B12 Kt /V:
Dialyzer
Vitamin B12 KoA (mL/min)
F80 Focus 120 CF 15-11 Focus 90 CF 23-08 F60 CA 110 TAF 175 TAF 12 CA 210 C121L C101L TAF 10 CA 90 CF 12-11 CT 190 F8 ICL-M151L CA 170
289.4 54.2 53.1 49.3 62.3 215.3 63.8 103.6 65.3 107.6 77.9 66.8 53.5 49.3 38.6 278.3 104.4 69.9 91.0
KB12 t /V ⫽
(4)
There are two concerns when using this parameter to estimate middle molecule removal. First, this parameter is not normalized for body size. We have, however, also evaluated TCVB12 normalized by postdialysis body weight or by total body water calculated using anthropometric formulas23 as alternative measures of total middle molecule removal and obtained similar results (not shown). Second, the treatment time that is available from the USRDS Case Mix Adequacy Study is the prescribed, not the delivered, treatment time, and the use of prescribed treatment time in equation 4 may significantly overestimate the true treatment time and therefore middle molecule removal. The second parameter evaluated as a measure of middle molecule removal is Kt /V for vitamin B12 (KB12t /V) and corrects for both of the above concerns. Delivered urea Kt /V can be calculated from predialysis and postdialysis blood urea nitrogen concentrations.1 Assuming that the distribution volumes for vitamin B12 and urea are equal, the following relationship can be established:
Kut /V
⫽
KB12 Ku
⫻ Kut /V
(6)
Data Source
was the total cleared volume of vitamin B12 (TCVB12) and is an estimate of middle molecule removal for an entire dialysis treatment. It can be calculated as
KB12t /V
Ku0 ⫹ 0.43 ⫻ Qf
where Ku0 denotes urea clearance at zero ultrafiltration rate. It should be noted that the volume of distribution for vitamin B12 is unlikely equal to that for urea; nevertheless, the above parameter is still a valid measure of vitamin B12 removal if the ratio of the volume of distribution for vitamin B12 to that for urea is constant. Both TCVB12 and KB12t /V were used separately in this study as measures of middle molecule removal for predicting patient survival in the 1991 Case Mix Adequacy Study.
NOTE. Dialyzers are listed in descending order based on their frequency of use in the United States during 1991. Unlisted dialyzers were used by 27% of patients.
TCVB12 ⫽ KB12 ⫻ t
KB0 12 ⫹ 0.43 ⫻ Qf
(5)
where Ku denotes dialyzer urea clearance. This equation states that the ratio of Kt /V values for vitamin B12 and urea
The data used in these analyses originated from the Special Study of Case Mix Adequacy of the USRDS, with additional data supplied by the standard USRDS data base. The Case Mix Adequacy Study had a historical prospective design and was based on a random sample of chronic HD patients who were alive on December 31, 1990. There were approximately 7,000 dialysis patients from 523 dialysis units across the United States in this random sample. Patient data that were collected included age, gender, race, primary cause of end-stage renal disease (ESRD), date of first dialysis treatment, date of death, date of any transplant, selected patient comorbid and risk factors, dialysis treatment parameters, psychosocial information, kidney transplant information, and the results of routine laboratory tests. Details of the protocol for data collection during this study have been previously reported.1,2,24 Certain patients were excluded from the analyses in the current study. Thirty-two percent of the patients were excluded because information necessary for calculating urea Kt /V (discussed later) was not available. Another 1% of the patients were excluded because urea Kt /V was less than 0.4 per treatment or greater than 2.0 per treatment, and 2% of the patients were excluded because they were not dialyzed three times per week. To minimize the effects of residual renal function (not measured) on this analysis, patients who had been dialyzed for less than 1 year also were excluded. This latter criterion resulted in the exclusion of 27% of the patients. An additional 3% of patients were excluded because they used acetate as the dialysate base, and 1% of patients were excluded because there was no information on more than half of the comorbid conditions. Last, patients were excluded if they were not treated using one of the 19 most common dialyzers employed during 1991 (Table 1). The total sample size was therefore reduced to 1,771 patients.
352
LEYPOLDT ET AL
The delivered, single-pool urea Kt /V was calculated using the following formula1: Kut /V ⫽ ⫺ln [BUNpost/BUNpre ⫺ 0.008 ⫻ t] ⫹ (4 ⫺ 3.5 ⫻ BUNpost/BUNpre) ⫻ Qf ⫻ t /BWpost
(7)
where BUNpre and BUNpost denote the predialysis and postdialysis blood urea nitrogen (BUN). The timing of the postdialysis sample was not clearly specified, but most samples were likely drawn immediately at the end of the dialysis treatment. Thus, this estimate of the dose of dialysis would not usually account for access or cardiopulmonary recirculation nor for compartmentalization of urea into different pools or tissues.
Data Analysis Analyses were adjusted for the following covariates as described previously1: age, gender, race, Hispanic origin, primary cause of ESRD (diabetes, hypertension, other), preexisting comorbid conditions (obesity, undernourished, nonambulatory, left ventricular hypertrophy, coronary artery disease, congestive heart failure, cirrhosis, chronic obstructive pulmonary disease, cardiomegaly by x-ray examination, neoplasm, and peripheral vascular disease), and baseline laboratory values (serum albumin, serum bilirubin, and serum cholesterol). The dependent variable analyzed was time to death from the study start date, and a proportional hazards regression analysis25 was used to estimate the relationship between middle molecule removal parameters and dose of dialysis with all-cause mortality. Patients were censored (removed from the analysis alive) at the time of transplantation, 60 days after a switch to peritoneal dialysis, or after completion of the study. Separate analyses were performed for both middle molecule removal parameters, with and without including area Kt /V as a covariate.
RESULTS
The primary results from these analyses are shown in Tables 2 and 3 where the relative mortality risk was calculated for higher rates of middle molecule removal as assessed by either TCVB12 or KB12t/V, respectively. Table 2 shows that patients treated with a 10% higher calculated TCVB12 had an approximately 5% lower risk of mortality (RR ⫽ 0.953, P ⬍ 0.0001 v 1.000) Table 2. The Relative Risk of Mortality (RR) for a 10% Higher Total Middle Molecule Removal as Assessed by Total Cleared Volume of Vitamin B12 (TCVB12) and the RR for a 0.1 Unit Higher Single-Pool Urea Kt /V Adjustments for Urea Kt /V
Without With
RR ⫽ RR ⫽
*Less than 1.000 (P ⬍ 0.0001). †Less than 1.000 (P ⬍ 0.001).
TCVB12
Urea Kt /V
0.944* 0.953*
— 0.925†
Table 3. The Relative Risk of Mortality (RR) for a 10% Higher Total Middle Molecule Removal as Assessed by KB12t /V and the RR for a 0.1-Unit Higher Single-Pool Urea Kt /V Adjustments for Urea Kt /V
Without With
RR ⫽ RR ⫽
KB12t /V
Urea Kt /V
0.940* 0.952*
— 0.949†
*Less than 1.000 (P ⬍ 0.0001). †Less than 1.000 (P ⬍ 0.05).
when urea Kt/V remained constant. For comparison, patients treated with a 0.1 unit higher value of urea Kt/V had a 7.5% lower risk of mortality when TCVB12 remained constant. The latter calculated relative mortality risk with an increase in urea Kt/V was virtually identical to that described previously without considering middle molecule removal.1 TCVB12 correlated only weakly with urea Kt/V (R ⫽ 0.230, P ⬍ 0.0001). Table 3 shows that patients treated with a 10% higher vitamin B12 Kt/V had an approximately 5% lower risk of mortality (RR ⫽ 0.952, P ⬍ 0.0001 v 1.000) when urea Kt/V remained constant. In this analysis, patients treated with a 0.1 higher value of urea Kt/V had a 5% lower risk of mortality. The lesser impact of urea Kt/V on mortality in this latter analysis could be because KB12t/V correlated better with urea Kt/V (R ⫽ 0.527, P ⬍ 0.0001) than did TCVB12. DISCUSSION
The results of the analyses performed in this study show that the use of a dialyzer with high calculated middle molecule removal rates is associated with a reduced risk of mortality in chronic HD patients independent of urea Kt/V. To minimize bias in these analyses, patients were excluded from the study sample when we could not accurately evaluate parameters that have been previously shown to influence patient survival, such as urea Kt/V. Furthermore, we excluded patients who had been on dialysis for 1 year or less because these patients were likely to have substantial native renal clearances of small and large solutes that we could not evaluate. It should be noted that the correlations described here remain valid in spite of these exclusions, because our conclusions do not require that the study
MIDDLE MOLECULES AND HD PATIENT SURVIVAL
sample be representative of all chronic HD patients in the United States. The practical significance of our findings can best be shown by comparing the expected middle molecule removal parameter TCVB12 for three hypothetical dialyzers using treatments that deliver approximately the same dose of urea Kt/V. Use of a dialyzer containing a membrane of small surface area and low porosity that achieves a clearance for vitamin B12 of 40 mL/min in a 250-minute treatment has a TCVB12 of 10 liters. Use of a dialyzer containing a larger surface area, low-porosity membrane that can achieve a clearance for vitamin B12 of 75 mL/min in a 220-minute treatment has a TCVB12 of approximately 16 liters. Finally, use of a dialyzer with a high-porosity membrane that can achieve a clearance for vitamin B12 of 150 mL/min in a 175minute treatment has TCVB12 of approximately 26 liters. Compared with treatment with the dialyzer containing a membrane of small surface area and low porosity, the mortality risk for the above treatment with a large surface area, lowporosity membrane is 0.786, and that with the high-porosity membrane is 0.618. Similar predictions would be obtained without considering urea Kt/V (Table 2). Middle molecules have been defined as molecules that are substantially larger than (or substantially different in charge from) urea and, consequently, have kinetics that do not follow that of urea during HD using membranes with low porosity.26 A number of middle molecules with toxicity have been completely or partially isolated from uremic plasma or its ultrafiltrate in past and recent years, including parathyroid hormone, 2-microglobulin,27 erythropoiesis inhibitor,28 several granulocyte inhibitory proteins,29,30 lipase inhibitors,31-33 advanced glycosylation end products,34 and appetite-suppressing molecules.35 Although the accumulation of these molecules in plasma is less likely to increase acute mortality than certain small solutes for which urea serves as a marker (eg, potassium or hydrogen ion), it is entirely conceivable that these middle molecules could have adverse long-term effects. The use of middle molecule removal to assess the importance of dialysis membrane on HD patient survival has two additional advantages over the use of membrane material. First, the middle molecule removal parameter is a continu-
353
ous and not a categorical variable; it therefore permits a continuous scale to gauge the importance of dialysis membrane on patient outcome. Second, it permits unique classification of certain dialysis membranes that are low flux but are made from a material that is considered biocompatible by criteria such as complement activation. Several caveats are worthy of note when using calculated vitamin B12 removal rates to assess middle molecule removal. First, calculated vitamin B12 removal rates based on clearances determined in vitro may not be accurate in the clinical setting because of membrane fouling by contact with blood proteins, unequal distribution of solutes between plasma water and within blood cells, or binding of solutes to plasma proteins.36 Second, the calculated removal rates do not account for postdialysis rebound of middle molecules.37 These caveats are important when comparing results from HD therapy with other forms of renal replacement, such as peritoneal dialysis, but likely do not present a significant limitation to the above analyses. Finally, to what extent the kinetics of vitamin B12 inside the patient’s body or in the extracorporeal circuit resemble those of other middle molecules is unclear. Few studies have assessed middle molecule removal as a predictor of ESRD patient outcome. Although it is often claimed that the National Cooperative Dialysis Study (NCDS)38 disproved the middle molecule hypothesis, it is clear from the report by Wineman39 that the NCDS did not specifically test the importance of middle molecule removal. Indeed, because membranes with very low porosity were used exclusively in the NCDS, the usefulness of variations in treatment time to assess middle molecule removal was very limited. Therefore, the results of the current analysis are not inconsistent with the result from the NCDS. The current analyses provide an alternative explanation to the previously reported observation that mortality risk varied with the type of dialysis membrane; however, the results reported here do not exclude other possible explanations such as membrane biocompatibility, ultrafiltration control machines, or a center/medical team effect. Both small and middle molecule removal indices appear to be independently associated with risk of mortality in long-term HD patients.
354
LEYPOLDT ET AL
Differences in mortality when using different types of dialysis membranes may be explained by differences in middle molecule removal. REFERENCES 1. Held PJ, Port FK, Wolfe RA, Stannard DC, Carroll CE, Daugirdas JT, Bloembergen WE, Greer JW, Hakim RM: The dose of hemodialysis and patient mortality. Kidney Int 50:550-556, 1996 2. Hakim RM, Held PJ, Stannard DC, Wolfe RA, Port FK, Daugirdas JT, Agodoa L: Effect of the dialysis membrane on mortality of chronic hemodialysis patients. Kidney Int 50:566-570, 1996 3. Owen WF Jr, Lew NL, Liu Y, Lowrie EG, Lazarus JM: The urea reduction ratio and serum albumin concentration as predictors of mortality in patients undergoing hemodialysis. N Engl J Med 329:1001-1006, 1993 4. Collins AJ, Ma JZ, Umen A, Keshaviah P: Urea index and other predictors of hemodialysis patient survival. Am J Kidney Dis 23:272-282, 1994 5. Hakim RM, Breyer J, Ismail N, Schulman G: Effects of dose of dialysis on morbidity and mortality. Am J Kidney Dis 23:661-669, 1994 6. Parker TF III, Husni L, Huang W, Lew N, Lowrie EG, Dallas Nephrology Associates: Survival of hemodialysis in the United States is improved with a greater quantity of dialysis. Am J Kidney Dis 23:670-680, 1994 7. Gotch FA, Levin NW, Port FK, Wolfe RA, Uehlinger DE: Clinical outcome relative to the dose of dialysis is not what you think: the fallacy of the mean. Am J Kidney Dis 30:1-15, 1997 8. Tokars JI, Alter MJ, Favero MS, Moyer LA, Bland LA: National surveillance of dialysis associated diseases in the United States, 1991. ASAIO J 39:966-975, 1993 9. Chenoweth DE, Cheung AK, Ward DM, Henderson LW: Anaphylatoxin formation during hemodialysis: Comparison of new and re-used dialyzers. Kidney Int 24:770-774, 1983 10. Dumler F, Zasuwa G, Levin NW: Effect of dialyzer reprocessing methods on complement activation and hemodialyzer-related symptoms. Artif Organs 11:128-131, 1987 11. Walton DF, Cheung AK: Membrane biocompatibility, in Nissenson AR, Fine RN, Gentile DE (eds): Clinical Dialysis (ed 3). Norwalk, CT, Appleton & Lange, 1995, pp 93-120 12. Orzol SM, Carroll CE, Hirth RA, Pereira BJG, Port FK, Held PJ: Differences in the cost of hemodialysis by dialyzer membrane. J Am Soc Nephrol 7:1460, 1996 (abstr) 13. Bergstro¨m J, Wehle B: Clinical implications of middle and larger molecules, in Nissenson AR, Fine RN, Gentile DE (eds): Clinical Dialysis (ed 3). Norwalk, CT, Appleton & Lange, 1995, pp 204-234 14. Leypoldt JK, Cheung AK, Deeter RB: Single compartment models for evaluating 2-microglobulin clearance during hemodialysis. ASAIO J 43:904-909, 1997 15. Werynski A, Waniewski J: Theoretical description of mass transport in medical membrane devices. Artif Organs 19:420-427, 1995
16. Jaffrin MY: Convective mass transfer in hemodialysis. Artif Organs 19:1162-1171, 1995 17. Waniewski J, Werynski A, Ahrenholz P, Lucjanek P, Judycki W, Esther G: Theoretical basis and experimental verification of the impact of ultrafiltration on dialyzer clearance. Artif Organs 15:70-77, 1991 18. Colton CK, Lowrie EG: Hemodialysis: physical principles and technical considerations, in Brenner BM, Rector FC Jr (eds): The Kidney (ed 2). Philadelphia, PA, Saunders, 1981, pp 2425-2489 19. Sigdell JE: Operating characteristics of hollow-fiber dialyzers, in Nissenson AR, Fine RN, Gentile DE (eds): Clinical Dialysis (ed 2). Norwalk, CT, Appleton & Lange, 1990, pp 97-117 20. Van Stone JC: Hemodialysis apparatus, in Daugirdas JT, Ing TS (eds): Handbook of Dialysis (ed 2). Boston, MA, Little, Brown, 1994, pp 30-52 21. Mujais SK, Schmidt B: Operating characteristics of hollow fiber dialyzers, in Nissenson AR, Fine RN, Gentile DE (eds): Clinical Dialysis (ed 3). Norwalk, CT, Appleton & Lange, 1995, pp 77-92 22. Sargent JA, Gotch FA: Principles and biophysics of dialysis, in Jacobs C, Kjellstrand CM, Koch KM, Winchester JF (eds): Replacement of Renal Function by Dialysis (ed 4). Dordrecht, Kluwer Academic, 1996, pp 34-102 23. Watson PE, Watson ID, Batt RD: Total body water volumes for adult males and females estimated from simple anthropometric measurements. Am J Clin Nutr 33:27-39, 1980 24. United States Renal Data System: USRDS 1996 Annual Data Report. Bethesda, MD, National Institutes of Health, National Institute of Diabetes and Digestive and Kidney Diseases, 1996 25. Cox D: Regression models and life tables. J R Stat Soc 34:187-210, 1972 26. Cheung AK: Quantitation of dialysis: the importance of membrane and middle molecules. Blood Purif 12:42-53, 1994 27. Gejyo F, Yamada T, Odani S, Nakagawa Y, Arakawa M, Kunitomo T, Kataoka H, Suzuki M, Hirasawa Y, Shirahama T, Cohen AS, Schmid K: A new form of amyloid protein associated with chronic hemodialysis was identified as 2-microglobulin. Biochem Biophys Res Commun 129: 701-706, 1985 28. Saito A, Suzuki I, Chung TG, Okamoto T, Hotta T: Separation of an inhibitor of erythropoiesis in ‘‘middle molecules’’ from hemodialysate from patients with chronic renal failure. Clin Chem 32:1938-1941, 1986 29. Ho¨rl WH, Haag-Weber M, Georgopoulos A, Block LH: Physicochemical characterization of a polypeptide present in uremic serum that inhibits the biological activity of polymorphonuclear cells. Proc Natl Acad Sci U S A 87:63536357, 1990 30. Haag-Weber M, Ho¨rl WH: Uremia and infection: Mechanisms of impaired cellular host defense. Nephron 63:125-131, 1993 31. Attman P-O, Samuelsson O, Alaupovic P: Lipoprotein metabolism and renal failure. Am J Kidney Dis 21:573592, 1993
MIDDLE MOLECULES AND HD PATIENT SURVIVAL
32. Seres DS, Strain GW, Hashim SA, Goldberg IJ, Levin NW: Improvement of plasma lipoprotein profiles during high-flux dialysis. J Am Soc Nephrol 3:1409-1415, 1993 33. Cheung AK, Parker CJ, Ren K, Iverius P-H: Increased lipase inhibitor in uremia: Identification of pre-HDL as a major inhibitor in normal and uremic plasma. Kidney Int 49:1360-1371, 1996 34. Makita Z, Bucala R, Rayfield EJ, Friedman EA, Kaufman AM, Korbet SM, Barth RH, Winston JA, Fuh H, Manogue KR, Cerami A, Vlassara H: Reactive glycosylation endproducts in diabetic uraemia and treatment of renal failure. Lancet 343:1519-1522, 1994 35. Anderstam B, Mamoun A-H, So¨dersten P, Bergstro¨m
355
J: Middle-sized molecule fractions isolated from uremic ultrafiltrate and normal urine inhibit ingestive behavior in the rat. J Am Soc Nephrol 7:2453-2460, 1996 36. Leypoldt JK, Cheung AK: Characterization of molecular transport in artificial kidneys. Artif Organs 20:381-389, 1996 37. Odell RA, Slowiaczek P, Moran JE, Schindhelm K: Beta2-microglobulin kinetics in end-stage renal failure. Kidney Int 39:909-919, 1991 38. Lowrie EG: The National Cooperative Dialysis Study. Kidney Int 23:S-1-S-122, 1983 (suppl 13) 39. Wineman RJ: Rationale of the National Cooperative Dialysis Study. Kidney Int 23:S-8-S-10, 1983 (suppl 13)
R
ECENT REPORTS from the 1991 Case Mix Adequacy Study of the United States Renal Data System (USRDS) have suggested that two dialysis prescription parameters, dialysis dose as defined by urea Kt/V (or the urea reduction ratio)1 and the type of dialysis membrane,2 play an important role in determining the survival of chronic hemodialysis (HD) patients. Higher values of delivered urea Kt/V (calculated using a single-pool model), up to approximately 1.3, were associated with progressively better survival,1 an effect consistent with that previously reported in other retrospective analyses.3-6 Although the magnitude of the optimal dose of dialysis is debatable,7 the importance of small solute removal, as quantified by urea Kt/V, is well established. The role of dialysis membrane on the survival of chronic HD patients in the USRDS Case Mix Adequacy Study was evaluated by categorizing dialyzers into three groups based on the type of membrane material2: unsubstituted cellulose, modified cellulose, and synthetic. Membrane material was chosen as the classification parameter because of its association with complement activation: unsubstituted cellulose membranes activate complement to a high degree; modified cellulose membranes activate complement to an intermediate degree; and synthetic membranes activate complement to a low degree. It should be noted, however, that dialyzer reprocessing is
common in dialysis centers in the United States,8 and the relationship between membrane material and complement activation is maintained only for dialyzers that are reprocessed using bleach.9-11 Because approximately 50% of all HD centers in the calendar year 1991 reprocessed dialyzers using Renalin (Renal Systems, Plymouth, MN) as the germicide without bleach,8 the proposed relationship between membrane material and
From Research and Medical Services, Salt Lake City Veterans Affairs Medical Center; the Division of Nephrology & Hypertension and Department of Bioengineering, University of Utah, Salt Lake City; United States Renal Data System Coordinating Center and Departments of Internal Medicine and Epidemiology, University of Michigan, Ann Arbor; Division of Nephrology, New England Medical Center, Boston, MA; National Institute of Diabetes and Digestive and Kidney Diseases, National Institutes of Health, Bethesda, MD. Received April 1, 1998; accepted in revised form July 31, 1998. The data reported here have been supplied by The United States Renal Data System (USRDS). The interpretation and reporting of these data are the responsibility of the authors and in no way should be seen as an offıcial policy or interpretation of the United States Government. Supported by USRDS, DVA Medical Research Funds and the Dialysis Research Foundation, Ogden, UT. Address reprint requests to John K. Leypoldt, PhD, Dumke Building 535, University of Utah, Salt Lake City, UT 84112. E-mail: [email protected] This is a US government work. There are no restrictions on its use. 0272-6386/99/3302-0018$0.00/0
American Journal of Kidney Diseases, Vol 33, No 2 (February), 1999: pp 349-355
349
350
LEYPOLDT ET AL
complement activation may not be consistently valid. If the observed effect of dialysis membrane on patient survival was attributable to poor membrane biocompatibility (attributed to complement activation and associated events), then the true effect of dialysis membrane on survival would have been underestimated in that study.2 Another potential explanation for the difference in patient survival associated with the use of various dialysis membranes is their difference in water flux, because modified cellulose and synthetic membranes more often have higher ultrafiltration coefficients than unsubstituted cellulose membranes. The use of ultrafiltration control machines, which are required for dialysis membranes with high ultrafiltration coefficients, may lead to better achievement of dry weight and may explain the better outcomes. Furthermore, the possibility that HD centers who prescribe dialyzers containing modified cellulose or synthetic membranes also provide superior medical care cannot be excluded. Indeed, centers that use dialyzers containing either modified cellulose or synthetic membranes have been shown to commit greater financial resources to direct patient care than centers that use dialyzers containing unsubstituted cellulose membranes.12 Yet another potential explanation for the difference in patient survival relates to the solute clearance profiles of the dialysis membranes. The dialysis membranes that were used on patients in the 1991 Case Mix Adequacy Study of the USRDS differed not only in biocompatibility and flux characteristics but also in their removal rates for middle molecules (solutes substantially larger than urea). In the current study, we have developed an approach for calculating middle molecule clearances or removal rates using in vitro transport data supplied by the dialyzer manufacturer and determined the association of calculated middle molecule removal with the survival of long-term HD patients. METHODS
Middle Molecule Removal We have chosen vitamin B12 as the molecular marker for evaluating middle molecule removal for both historical and practical reasons. Vitamin B12 (1,355 daltons) has long been used as a marker for middle molecules in a number of in vitro and clinical studies,13 and manufacturers frequently report dialyzer clearances for vitamin B12 determined in vitro as a measure of the porosity of the dialysis membrane.
Although the validity of projecting in vitro clearances of vitamin B12 to the clinical HD setting is unknown, calculated clearances of vitamin B12 appear reasonable. Furthermore, for certain low-flux and high-flux dialyzers, the calculated values correlate with clearances of other middle molecules such as 2-microglobulin measured during clinical HD.14 Solute clearance during HD consists of both diffusive and convective contributions. The evaluation of separate diffusive and convective contributions to total clearance is complex, and this topic has been recently reviewed.15,16 Experimental data suggest that the total clearance of vitamin B12 (KB12) during HD can be expressed as a diffusive term plus a convective term by an equation simplified from that recently proposed by Jaffrin16: KB12 ⫽ KB0 12 ⫹ 0.43 ⫻ Qf
(1)
where KB0 12 denotes the diffusive clearance of vitamin B12 in the absence of ultrafiltration and Qf denotes the ultrafiltration rate. Equation 1 has been shown to be approximately valid for sodium, urea, creatinine, vitamin B12, and myoglobin over a range of ultrafiltration rates less than 20 mL/ min.16,17 The ultrafiltration rate averaged over the entire treatment can be calculated from the predialysis body weight (BWpre) and postdialysis body weight (BWpost) as Qf ⫽ (BWpre ⫺ BWpost)/t
(2)
where t denotes treatment time. Thus, the convective component to total vitamin B12 clearance can be calculated using equations 1 and 2 from measurements of predialysis and postdialysis body weight. Diffusive clearance of vitamin B12 is usually reported by dialyzer manufacturers for only a limited number of blood and dialysate flow rate combinations. To calculate KB0 12 under any treatment condition for a given dialyzer, we must first calculate the mass transfer-area coefficient for vitamin B12 (K o AB12) from vitamin B12 clearance at the blood (Qb) and dialysate (Qd) flow rates reported by the manufacturer, using the following equation18: K o AB12 ⫽
Qb 1 ⫹ Qb /Qd
ln
3
1 ⫺ (1 ⫹
1 KB0 12/Qb)
4
⫻ Qb /Qd
(3)
Table 1 shows the calculated values of K o AB12 for the 19 most commonly employed dialyzers during 1991 using diffusive vitamin B12 clearances provided by the manufacturers and reported in recent reviews.19-21 Using the calculated values in this table, equation 3, after rearrangement, can then be used to calculate diffusive vitamin B12 clearance in the absence of ultrafiltration at any blood and dialysate flow rate combination. The use of the equations 1 through 3 for calculating total dialyzer clearance of vitamin B12 is similar to that for calculating urea clearance under any conditions of blood, dialysate, and ultrafiltration flow rates.22 Dialyzer urea K o A values used in this study were identical to those used in a previous study.1
Middle Molecule Removal Parameters Two different parameters were used as measures of middle molecule removal in the current study. The first parameter
MIDDLE MOLECULES AND HD PATIENT SURVIVAL
351
Table 1. Mass Transfer-Area Coefficients (KoA) for Vitamin B12 for the Dialyzers Most Commonly Used in 1991
is directly proportional to the ratio of their clearances. The ratio of vitamin B12 clearance to urea clearance for any dialyzer can be estimated from the respective K o A values and the blood, dialysate, and ultrafiltration rates by using equations 1 through 3. Combining equations 1 and 5 for both vitamin B12 and urea yields an expression for vitamin B12 Kt /V:
Dialyzer
Vitamin B12 KoA (mL/min)
F80 Focus 120 CF 15-11 Focus 90 CF 23-08 F60 CA 110 TAF 175 TAF 12 CA 210 C121L C101L TAF 10 CA 90 CF 12-11 CT 190 F8 ICL-M151L CA 170
289.4 54.2 53.1 49.3 62.3 215.3 63.8 103.6 65.3 107.6 77.9 66.8 53.5 49.3 38.6 278.3 104.4 69.9 91.0
KB12 t /V ⫽
(4)
There are two concerns when using this parameter to estimate middle molecule removal. First, this parameter is not normalized for body size. We have, however, also evaluated TCVB12 normalized by postdialysis body weight or by total body water calculated using anthropometric formulas23 as alternative measures of total middle molecule removal and obtained similar results (not shown). Second, the treatment time that is available from the USRDS Case Mix Adequacy Study is the prescribed, not the delivered, treatment time, and the use of prescribed treatment time in equation 4 may significantly overestimate the true treatment time and therefore middle molecule removal. The second parameter evaluated as a measure of middle molecule removal is Kt /V for vitamin B12 (KB12t /V) and corrects for both of the above concerns. Delivered urea Kt /V can be calculated from predialysis and postdialysis blood urea nitrogen concentrations.1 Assuming that the distribution volumes for vitamin B12 and urea are equal, the following relationship can be established:
Kut /V
⫽
KB12 Ku
⫻ Kut /V
(6)
Data Source
was the total cleared volume of vitamin B12 (TCVB12) and is an estimate of middle molecule removal for an entire dialysis treatment. It can be calculated as
KB12t /V
Ku0 ⫹ 0.43 ⫻ Qf
where Ku0 denotes urea clearance at zero ultrafiltration rate. It should be noted that the volume of distribution for vitamin B12 is unlikely equal to that for urea; nevertheless, the above parameter is still a valid measure of vitamin B12 removal if the ratio of the volume of distribution for vitamin B12 to that for urea is constant. Both TCVB12 and KB12t /V were used separately in this study as measures of middle molecule removal for predicting patient survival in the 1991 Case Mix Adequacy Study.
NOTE. Dialyzers are listed in descending order based on their frequency of use in the United States during 1991. Unlisted dialyzers were used by 27% of patients.
TCVB12 ⫽ KB12 ⫻ t
KB0 12 ⫹ 0.43 ⫻ Qf
(5)
where Ku denotes dialyzer urea clearance. This equation states that the ratio of Kt /V values for vitamin B12 and urea
The data used in these analyses originated from the Special Study of Case Mix Adequacy of the USRDS, with additional data supplied by the standard USRDS data base. The Case Mix Adequacy Study had a historical prospective design and was based on a random sample of chronic HD patients who were alive on December 31, 1990. There were approximately 7,000 dialysis patients from 523 dialysis units across the United States in this random sample. Patient data that were collected included age, gender, race, primary cause of end-stage renal disease (ESRD), date of first dialysis treatment, date of death, date of any transplant, selected patient comorbid and risk factors, dialysis treatment parameters, psychosocial information, kidney transplant information, and the results of routine laboratory tests. Details of the protocol for data collection during this study have been previously reported.1,2,24 Certain patients were excluded from the analyses in the current study. Thirty-two percent of the patients were excluded because information necessary for calculating urea Kt /V (discussed later) was not available. Another 1% of the patients were excluded because urea Kt /V was less than 0.4 per treatment or greater than 2.0 per treatment, and 2% of the patients were excluded because they were not dialyzed three times per week. To minimize the effects of residual renal function (not measured) on this analysis, patients who had been dialyzed for less than 1 year also were excluded. This latter criterion resulted in the exclusion of 27% of the patients. An additional 3% of patients were excluded because they used acetate as the dialysate base, and 1% of patients were excluded because there was no information on more than half of the comorbid conditions. Last, patients were excluded if they were not treated using one of the 19 most common dialyzers employed during 1991 (Table 1). The total sample size was therefore reduced to 1,771 patients.
352
LEYPOLDT ET AL
The delivered, single-pool urea Kt /V was calculated using the following formula1: Kut /V ⫽ ⫺ln [BUNpost/BUNpre ⫺ 0.008 ⫻ t] ⫹ (4 ⫺ 3.5 ⫻ BUNpost/BUNpre) ⫻ Qf ⫻ t /BWpost
(7)
where BUNpre and BUNpost denote the predialysis and postdialysis blood urea nitrogen (BUN). The timing of the postdialysis sample was not clearly specified, but most samples were likely drawn immediately at the end of the dialysis treatment. Thus, this estimate of the dose of dialysis would not usually account for access or cardiopulmonary recirculation nor for compartmentalization of urea into different pools or tissues.
Data Analysis Analyses were adjusted for the following covariates as described previously1: age, gender, race, Hispanic origin, primary cause of ESRD (diabetes, hypertension, other), preexisting comorbid conditions (obesity, undernourished, nonambulatory, left ventricular hypertrophy, coronary artery disease, congestive heart failure, cirrhosis, chronic obstructive pulmonary disease, cardiomegaly by x-ray examination, neoplasm, and peripheral vascular disease), and baseline laboratory values (serum albumin, serum bilirubin, and serum cholesterol). The dependent variable analyzed was time to death from the study start date, and a proportional hazards regression analysis25 was used to estimate the relationship between middle molecule removal parameters and dose of dialysis with all-cause mortality. Patients were censored (removed from the analysis alive) at the time of transplantation, 60 days after a switch to peritoneal dialysis, or after completion of the study. Separate analyses were performed for both middle molecule removal parameters, with and without including area Kt /V as a covariate.
RESULTS
The primary results from these analyses are shown in Tables 2 and 3 where the relative mortality risk was calculated for higher rates of middle molecule removal as assessed by either TCVB12 or KB12t/V, respectively. Table 2 shows that patients treated with a 10% higher calculated TCVB12 had an approximately 5% lower risk of mortality (RR ⫽ 0.953, P ⬍ 0.0001 v 1.000) Table 2. The Relative Risk of Mortality (RR) for a 10% Higher Total Middle Molecule Removal as Assessed by Total Cleared Volume of Vitamin B12 (TCVB12) and the RR for a 0.1 Unit Higher Single-Pool Urea Kt /V Adjustments for Urea Kt /V
Without With
RR ⫽ RR ⫽
*Less than 1.000 (P ⬍ 0.0001). †Less than 1.000 (P ⬍ 0.001).
TCVB12
Urea Kt /V
0.944* 0.953*
— 0.925†
Table 3. The Relative Risk of Mortality (RR) for a 10% Higher Total Middle Molecule Removal as Assessed by KB12t /V and the RR for a 0.1-Unit Higher Single-Pool Urea Kt /V Adjustments for Urea Kt /V
Without With
RR ⫽ RR ⫽
KB12t /V
Urea Kt /V
0.940* 0.952*
— 0.949†
*Less than 1.000 (P ⬍ 0.0001). †Less than 1.000 (P ⬍ 0.05).
when urea Kt/V remained constant. For comparison, patients treated with a 0.1 unit higher value of urea Kt/V had a 7.5% lower risk of mortality when TCVB12 remained constant. The latter calculated relative mortality risk with an increase in urea Kt/V was virtually identical to that described previously without considering middle molecule removal.1 TCVB12 correlated only weakly with urea Kt/V (R ⫽ 0.230, P ⬍ 0.0001). Table 3 shows that patients treated with a 10% higher vitamin B12 Kt/V had an approximately 5% lower risk of mortality (RR ⫽ 0.952, P ⬍ 0.0001 v 1.000) when urea Kt/V remained constant. In this analysis, patients treated with a 0.1 higher value of urea Kt/V had a 5% lower risk of mortality. The lesser impact of urea Kt/V on mortality in this latter analysis could be because KB12t/V correlated better with urea Kt/V (R ⫽ 0.527, P ⬍ 0.0001) than did TCVB12. DISCUSSION
The results of the analyses performed in this study show that the use of a dialyzer with high calculated middle molecule removal rates is associated with a reduced risk of mortality in chronic HD patients independent of urea Kt/V. To minimize bias in these analyses, patients were excluded from the study sample when we could not accurately evaluate parameters that have been previously shown to influence patient survival, such as urea Kt/V. Furthermore, we excluded patients who had been on dialysis for 1 year or less because these patients were likely to have substantial native renal clearances of small and large solutes that we could not evaluate. It should be noted that the correlations described here remain valid in spite of these exclusions, because our conclusions do not require that the study
MIDDLE MOLECULES AND HD PATIENT SURVIVAL
sample be representative of all chronic HD patients in the United States. The practical significance of our findings can best be shown by comparing the expected middle molecule removal parameter TCVB12 for three hypothetical dialyzers using treatments that deliver approximately the same dose of urea Kt/V. Use of a dialyzer containing a membrane of small surface area and low porosity that achieves a clearance for vitamin B12 of 40 mL/min in a 250-minute treatment has a TCVB12 of 10 liters. Use of a dialyzer containing a larger surface area, low-porosity membrane that can achieve a clearance for vitamin B12 of 75 mL/min in a 220-minute treatment has a TCVB12 of approximately 16 liters. Finally, use of a dialyzer with a high-porosity membrane that can achieve a clearance for vitamin B12 of 150 mL/min in a 175minute treatment has TCVB12 of approximately 26 liters. Compared with treatment with the dialyzer containing a membrane of small surface area and low porosity, the mortality risk for the above treatment with a large surface area, lowporosity membrane is 0.786, and that with the high-porosity membrane is 0.618. Similar predictions would be obtained without considering urea Kt/V (Table 2). Middle molecules have been defined as molecules that are substantially larger than (or substantially different in charge from) urea and, consequently, have kinetics that do not follow that of urea during HD using membranes with low porosity.26 A number of middle molecules with toxicity have been completely or partially isolated from uremic plasma or its ultrafiltrate in past and recent years, including parathyroid hormone, 2-microglobulin,27 erythropoiesis inhibitor,28 several granulocyte inhibitory proteins,29,30 lipase inhibitors,31-33 advanced glycosylation end products,34 and appetite-suppressing molecules.35 Although the accumulation of these molecules in plasma is less likely to increase acute mortality than certain small solutes for which urea serves as a marker (eg, potassium or hydrogen ion), it is entirely conceivable that these middle molecules could have adverse long-term effects. The use of middle molecule removal to assess the importance of dialysis membrane on HD patient survival has two additional advantages over the use of membrane material. First, the middle molecule removal parameter is a continu-
353
ous and not a categorical variable; it therefore permits a continuous scale to gauge the importance of dialysis membrane on patient outcome. Second, it permits unique classification of certain dialysis membranes that are low flux but are made from a material that is considered biocompatible by criteria such as complement activation. Several caveats are worthy of note when using calculated vitamin B12 removal rates to assess middle molecule removal. First, calculated vitamin B12 removal rates based on clearances determined in vitro may not be accurate in the clinical setting because of membrane fouling by contact with blood proteins, unequal distribution of solutes between plasma water and within blood cells, or binding of solutes to plasma proteins.36 Second, the calculated removal rates do not account for postdialysis rebound of middle molecules.37 These caveats are important when comparing results from HD therapy with other forms of renal replacement, such as peritoneal dialysis, but likely do not present a significant limitation to the above analyses. Finally, to what extent the kinetics of vitamin B12 inside the patient’s body or in the extracorporeal circuit resemble those of other middle molecules is unclear. Few studies have assessed middle molecule removal as a predictor of ESRD patient outcome. Although it is often claimed that the National Cooperative Dialysis Study (NCDS)38 disproved the middle molecule hypothesis, it is clear from the report by Wineman39 that the NCDS did not specifically test the importance of middle molecule removal. Indeed, because membranes with very low porosity were used exclusively in the NCDS, the usefulness of variations in treatment time to assess middle molecule removal was very limited. Therefore, the results of the current analysis are not inconsistent with the result from the NCDS. The current analyses provide an alternative explanation to the previously reported observation that mortality risk varied with the type of dialysis membrane; however, the results reported here do not exclude other possible explanations such as membrane biocompatibility, ultrafiltration control machines, or a center/medical team effect. Both small and middle molecule removal indices appear to be independently associated with risk of mortality in long-term HD patients.
354
LEYPOLDT ET AL
Differences in mortality when using different types of dialysis membranes may be explained by differences in middle molecule removal. REFERENCES 1. Held PJ, Port FK, Wolfe RA, Stannard DC, Carroll CE, Daugirdas JT, Bloembergen WE, Greer JW, Hakim RM: The dose of hemodialysis and patient mortality. Kidney Int 50:550-556, 1996 2. Hakim RM, Held PJ, Stannard DC, Wolfe RA, Port FK, Daugirdas JT, Agodoa L: Effect of the dialysis membrane on mortality of chronic hemodialysis patients. Kidney Int 50:566-570, 1996 3. Owen WF Jr, Lew NL, Liu Y, Lowrie EG, Lazarus JM: The urea reduction ratio and serum albumin concentration as predictors of mortality in patients undergoing hemodialysis. N Engl J Med 329:1001-1006, 1993 4. Collins AJ, Ma JZ, Umen A, Keshaviah P: Urea index and other predictors of hemodialysis patient survival. Am J Kidney Dis 23:272-282, 1994 5. Hakim RM, Breyer J, Ismail N, Schulman G: Effects of dose of dialysis on morbidity and mortality. Am J Kidney Dis 23:661-669, 1994 6. Parker TF III, Husni L, Huang W, Lew N, Lowrie EG, Dallas Nephrology Associates: Survival of hemodialysis in the United States is improved with a greater quantity of dialysis. Am J Kidney Dis 23:670-680, 1994 7. Gotch FA, Levin NW, Port FK, Wolfe RA, Uehlinger DE: Clinical outcome relative to the dose of dialysis is not what you think: the fallacy of the mean. Am J Kidney Dis 30:1-15, 1997 8. Tokars JI, Alter MJ, Favero MS, Moyer LA, Bland LA: National surveillance of dialysis associated diseases in the United States, 1991. ASAIO J 39:966-975, 1993 9. Chenoweth DE, Cheung AK, Ward DM, Henderson LW: Anaphylatoxin formation during hemodialysis: Comparison of new and re-used dialyzers. Kidney Int 24:770-774, 1983 10. Dumler F, Zasuwa G, Levin NW: Effect of dialyzer reprocessing methods on complement activation and hemodialyzer-related symptoms. Artif Organs 11:128-131, 1987 11. Walton DF, Cheung AK: Membrane biocompatibility, in Nissenson AR, Fine RN, Gentile DE (eds): Clinical Dialysis (ed 3). Norwalk, CT, Appleton & Lange, 1995, pp 93-120 12. Orzol SM, Carroll CE, Hirth RA, Pereira BJG, Port FK, Held PJ: Differences in the cost of hemodialysis by dialyzer membrane. J Am Soc Nephrol 7:1460, 1996 (abstr) 13. Bergstro¨m J, Wehle B: Clinical implications of middle and larger molecules, in Nissenson AR, Fine RN, Gentile DE (eds): Clinical Dialysis (ed 3). Norwalk, CT, Appleton & Lange, 1995, pp 204-234 14. Leypoldt JK, Cheung AK, Deeter RB: Single compartment models for evaluating 2-microglobulin clearance during hemodialysis. ASAIO J 43:904-909, 1997 15. Werynski A, Waniewski J: Theoretical description of mass transport in medical membrane devices. Artif Organs 19:420-427, 1995
16. Jaffrin MY: Convective mass transfer in hemodialysis. Artif Organs 19:1162-1171, 1995 17. Waniewski J, Werynski A, Ahrenholz P, Lucjanek P, Judycki W, Esther G: Theoretical basis and experimental verification of the impact of ultrafiltration on dialyzer clearance. Artif Organs 15:70-77, 1991 18. Colton CK, Lowrie EG: Hemodialysis: physical principles and technical considerations, in Brenner BM, Rector FC Jr (eds): The Kidney (ed 2). Philadelphia, PA, Saunders, 1981, pp 2425-2489 19. Sigdell JE: Operating characteristics of hollow-fiber dialyzers, in Nissenson AR, Fine RN, Gentile DE (eds): Clinical Dialysis (ed 2). Norwalk, CT, Appleton & Lange, 1990, pp 97-117 20. Van Stone JC: Hemodialysis apparatus, in Daugirdas JT, Ing TS (eds): Handbook of Dialysis (ed 2). Boston, MA, Little, Brown, 1994, pp 30-52 21. Mujais SK, Schmidt B: Operating characteristics of hollow fiber dialyzers, in Nissenson AR, Fine RN, Gentile DE (eds): Clinical Dialysis (ed 3). Norwalk, CT, Appleton & Lange, 1995, pp 77-92 22. Sargent JA, Gotch FA: Principles and biophysics of dialysis, in Jacobs C, Kjellstrand CM, Koch KM, Winchester JF (eds): Replacement of Renal Function by Dialysis (ed 4). Dordrecht, Kluwer Academic, 1996, pp 34-102 23. Watson PE, Watson ID, Batt RD: Total body water volumes for adult males and females estimated from simple anthropometric measurements. Am J Clin Nutr 33:27-39, 1980 24. United States Renal Data System: USRDS 1996 Annual Data Report. Bethesda, MD, National Institutes of Health, National Institute of Diabetes and Digestive and Kidney Diseases, 1996 25. Cox D: Regression models and life tables. J R Stat Soc 34:187-210, 1972 26. Cheung AK: Quantitation of dialysis: the importance of membrane and middle molecules. Blood Purif 12:42-53, 1994 27. Gejyo F, Yamada T, Odani S, Nakagawa Y, Arakawa M, Kunitomo T, Kataoka H, Suzuki M, Hirasawa Y, Shirahama T, Cohen AS, Schmid K: A new form of amyloid protein associated with chronic hemodialysis was identified as 2-microglobulin. Biochem Biophys Res Commun 129: 701-706, 1985 28. Saito A, Suzuki I, Chung TG, Okamoto T, Hotta T: Separation of an inhibitor of erythropoiesis in ‘‘middle molecules’’ from hemodialysate from patients with chronic renal failure. Clin Chem 32:1938-1941, 1986 29. Ho¨rl WH, Haag-Weber M, Georgopoulos A, Block LH: Physicochemical characterization of a polypeptide present in uremic serum that inhibits the biological activity of polymorphonuclear cells. Proc Natl Acad Sci U S A 87:63536357, 1990 30. Haag-Weber M, Ho¨rl WH: Uremia and infection: Mechanisms of impaired cellular host defense. Nephron 63:125-131, 1993 31. Attman P-O, Samuelsson O, Alaupovic P: Lipoprotein metabolism and renal failure. Am J Kidney Dis 21:573592, 1993
MIDDLE MOLECULES AND HD PATIENT SURVIVAL
32. Seres DS, Strain GW, Hashim SA, Goldberg IJ, Levin NW: Improvement of plasma lipoprotein profiles during high-flux dialysis. J Am Soc Nephrol 3:1409-1415, 1993 33. Cheung AK, Parker CJ, Ren K, Iverius P-H: Increased lipase inhibitor in uremia: Identification of pre-HDL as a major inhibitor in normal and uremic plasma. Kidney Int 49:1360-1371, 1996 34. Makita Z, Bucala R, Rayfield EJ, Friedman EA, Kaufman AM, Korbet SM, Barth RH, Winston JA, Fuh H, Manogue KR, Cerami A, Vlassara H: Reactive glycosylation endproducts in diabetic uraemia and treatment of renal failure. Lancet 343:1519-1522, 1994 35. Anderstam B, Mamoun A-H, So¨dersten P, Bergstro¨m
355
J: Middle-sized molecule fractions isolated from uremic ultrafiltrate and normal urine inhibit ingestive behavior in the rat. J Am Soc Nephrol 7:2453-2460, 1996 36. Leypoldt JK, Cheung AK: Characterization of molecular transport in artificial kidneys. Artif Organs 20:381-389, 1996 37. Odell RA, Slowiaczek P, Moran JE, Schindhelm K: Beta2-microglobulin kinetics in end-stage renal failure. Kidney Int 39:909-919, 1991 38. Lowrie EG: The National Cooperative Dialysis Study. Kidney Int 23:S-1-S-122, 1983 (suppl 13) 39. Wineman RJ: Rationale of the National Cooperative Dialysis Study. Kidney Int 23:S-8-S-10, 1983 (suppl 13)