Extremes of glycemic control (HbA1c) increase hospitalization risk in diabetic hemodialysis patients in the USA.

BACKGROUND/AIMS: Because the relation between glycemic control and clinical outcomes found in the general diabetic population has not been established in diabetic hemodialysis patients, we evaluated the association between glycemic control and hospitalization risk. METHODS: We performed a primary retrospective data analysis on 23,829 hemodialysis patients with diabetes mellitus. Hemoglobin A(1c) at baseline and hospitalization events over the subsequent 12 months were analyzed and logistic regression models constructed for unadjusted, case mix-adjusted and case mix plus lab- adjusted data. Models were also constructed for cardiovascular, vascular access and sepsis hospitalizations. RESULTS: Eighty percent had type 2 DM, 5% type 1 and 14% not specified. The groups had similar mean HbA(1c) levels, 6.8 +/- 1.6%. Among all patients, the mean HbA(1c) values were >7% in 35%. The odds ratio of hospitalizations grouped by baseline HbA(1c) was significant at extremes of <5% and >11%. Similar relationships were evident for the subset of type 2 DM and in the analysis for hospitalizations due to sepsis. CONCLUSION: Extremely high and low HbA(1c) values are associated with hospitalization risk in diabetic hemodialysis patients. Prospective studies are needed to determine whether meeting recommended HbA(1c) targets might improve outcomes without posing additional risks in this population.

Plasma F2-isoprostane levels are elevated in chronic hemodialysis patients.

AIMS: Cardiovascular mortality has been reported to be 10- to 20-fold higher in chronic dialysis patients than in the age-matched general population. It has been suggested that increased oxidant stress and resulting vascular wall injury due to uremia and the hemodialysis procedure may be one of the mechanisms predisposing to these cardiovascular complications. Further, hemodialysis membrane bioincompatibility can contribute to increased oxidative stress and prevalence of inflammation. MATERIALS: We studied 18 chronic hemodialysis (CHD) patients (age 62.8 +/- 14.7 years, 39% male, 61% African-American, 44% insulin-dependent diabetic, 61% smokers, 61% with documented coronary artery disease) during hemodialysis with 2 membranes with different flux and complement activating properties. METHODS: We have measured free and phospholipid-bound F2-isoprostane (F2-IsoP) levels, a sensitive marker of oxidative stress, in CHD patients and compared them to levels in healthy subjects. We have also examined the acute effects of the hemodialysis procedure using both biocompatible and bioincompatible membranes on F2-IsoP levels. RESULTS: The results indicated that, compared to controls, both free (96.2 +/- 48.8 pg/ml versus 37.6 +/- 17.2 pg/ml) and bound F2-IsoP (220.4 +/- 154.8 pg/ml versus 146.8 +/- 58.4 pg/ml) levels were significantly higher (p < 0.05 for both). There was a statistically significant decrease in free F2-IsoP concentrations at 15 and 30 minutes of HD, which rebounded to baseline levels at the completion of the procedure. There were no significant differences in F2-IsoP concentrations between the 2 study dialyzers at any time point. Age, smoking status, diabetes mellitus and presence of cardiovascular disease were also not correlated with F2-IsoP levels in this patient population. There was a significant association between predialysis F2-IsoP and C-reactive protein concentrations. CONCLUSION: Using a sensitive and specific assay for the measurement of F2-IsoP, we demonstrated that CHD patients are under increased oxidative stress. During a single hemodialysis treatment, the hemodialysis membrane appears to have no discernable effect on oxidative stress status. Measurement of F2-isoprostanes may be a useful biomarker of oxidative stress status as well as in developing new therapeutic strategies to ameliorate inflammatory and oxidative injury in this patient population.

Recombinant human growth hormone in patients with acute renal failure.

OBJECTIVE: The objective of this study is to determine the impact of recombinant human growth hormone (rhGH) on metabolic and nutritional parameters in malnourished patients with acute renal failure. DESIGN: The design is an open-labeled pilot trial examining the effects of rhGH administration in a small group of highly catabolic, malnourished patients with acute renal failure. Each patient served as his or her own control. SETTING: An intensive care unit in a tertiary care medical institution. PATIENTS: Five patients with established acute renal failure in a critical care unit. Entry criteria included clinical evidence of malnutrition: a serum albumin level of <3.2 g/dL, a prealbumin level of < or = 20 mg/dL, and an insulin-like growth factor IGF 1 level <200 ng/mL. The study consisted of 3 periods: phase I, 3 day baseline; phase II, 6 day treatment; and phase III, 3 day washout. During the entire study, blood and urine samples were obtained daily to calculate normalized protein catabolic rate, total nitrogen appearance rate (TNA), and nitrogen balance. Additional data were collected to measure metabolic and inflammatory parameters. INTERVENTION: The intervention consisted of administering 100 microg/kg/d of rhGH for 6 days. RESULTS: There were significant changes in TNA, normalized protein catabolic rate, and nitrogen balance during the 3 study phases. TNA decreased from 43.3 +/- 24.4 g/d in phase I, to 25.2 +/- 16.5 g/d during phase II (P <.001). There was a further decrease in TNA to 16.2 +/- 8.3 g/d during phase III (P <.001 v phase I). Nitrogen balance improved from - 31.8 +/- 21.4 g/d during phase I, to - 12.9 +/- 10.3 g/d during phase II (P <.001), and further improved to - 4.1 +/- 4.0 g/d in phase III (P <.001 v phase I). Significant changes were also noted in levels of blood urea nitrogen, phosphorous, serum growth hormone, IGF-1, and serum leptin levels after growth hormone administration. A statistically significant increase in serum albumin was noted in phase III (3.1 g/dL) versus phase I (2.7 +/- 0.7 g/dL). CONCLUSIONS: Administration of rhGH to critically ill patients with acute renal failure resulted in improvements in negative nitrogen balance and a significant decrease in total nitrogen appearance rate. These changes corresponded to increases in serum growth hormone, IGF-1, IGF-1 binding protein 3, and leptin levels after growth hormone administration.

Vascular access blood flow monitoring reduces access morbidity and costs.

BACKGROUND: Vascular access morbidity results in suboptimal patient outcomes and costs more than $8000 per patient-year at risk, representing approximately 15% of total Medicare expenditures for ESRD patients annually. In recent years, the rate of access thrombosis has improved following the advent of vascular access blood flow monitoring (VABFM) programs to identify and treat stenosis prior to thrombosis. To define further both the clinical and financial impact of such programs, we used the ultrasound dilution method to study the effects of VABFM on thrombosis-related morbid events and associated costs, compared with both dynamic venous pressure monitoring (DVPM) and no monitoring (NM) in arteriovenous fistulas (AVF) and grafts. METHODS: A total of 132 chronic hemodialysis patients were followed prospectively for three consecutive study phases (I, 11 months of NM; II, 12 months of DVPM; III, 10 months of VABFM). All vascular access-related information (thrombosis rate, hospitalization, angiogram, angioplasty, access surgery, thrombectomy, catheter placement, missed treatments) was collected during the three study periods. RESULTS: During the three study phases, graft thrombosis rate was reduced from 0.71 (phase I), to 0.67 (phase II), to 0.16 (phase III) events per patient-year at risk (P < 0.001 phase III vs. phases I and II). Similarly, hospital days, missed treatments, and catheter use related to thrombotic events were significantly reduced during phase III compared to phases I and II. Hospital days related to vascular access morbidity and adjusted for patient-year at risk were 1.8, 1.6, and 0.4 and missed dialysis treatments were 0.98, 0.86, and 0.26 treatments per patient-year at risk for phases I, II, and III, respectively (P < 0.001 for phase III vs. phases I and II). Catheter use was also significantly reduced during phases II and III, from 0.29 (phase I) to 0.17 and further to 0.07 catheters per patient-year at risk, respectively (P < 0.05 for phase III vs. phase I). Percutaneous angioplasty procedures increased during phases II and III from 0.09 to 0.32 to 0.54 procedures per patient-year at risk for phases I, II, and III, respectively (P < 0.01 for phase III vs. phase I). When the total cost of treatment for thrombosis-related events for grafts was estimated, it was found that during phase III, the adjusted yearly billed amount was reduced by 49% versus phase I and 54% versus phase II to $158,550. Similar trends in reduced thrombosis-related morbid events and cost were observed for AVFs. CONCLUSIONS: VABFM for early detection of vascular access malfunction coupled with preventive intervention reduces thrombosis rates in both polytetrafluoroethylene (PTFE) grafts and native AVFs. While there was a significant increase in the number of angioplasties done during the flow monitoring phase, the comprehensive cost is markedly reduced due to the decreased number of hospitalizations, catheters placed, missed treatments, and surgical interventions. Vascular access blood flow monitoring along with preventive interventions should be the standard of care in chronic hemodialysis patients.

Serum transferrin and serum prealbumin are early predictors of serum albumin in chronic hemodialysis patients.

Protein-calorie malnutrition is a known risk factor for increased morbidity and mortality in maintenance hemodialysis patients (MHD). Serum albumin is the most commonly measured nutritional index in MHD patients because of its easy routine availability and association with outcomes of interest. However, its long half-life of approximately 20 days makes it a late index of nutritional status, and its exclusive use may delay implementation of appropriate nutritional interventions. Serum prealbumin and transferrin have been proposed as earlier nutritional markers. However, the temporal associations among these indices and serum albumin have not been well documented. To assess the ability of serum prealbumin and serum transferrin to predict changes in serum albumin over time, we prospectively analyzed these parameters in 105 MHD patients every month for 6 consecutive months. The mixed model analysis showed that early changes in either serum transferrin or prealbumin predicted and were significantly associated with changes in serum albumin (P<.0001). Using a prototype equation, a change of 0.12 g/dL in serum albumin concentration can be predicted by a 10% change in the same direction of serum transferrin and prealbumin. We conclude that clinically significant changes in albumin can be reliably predicted by earlier changes in serum transferrin and prealbumin.

The current and future state of the ESRD program: a provider roundtable.

The decade of the 1990s have seen substantial consolidation of services in the dialysis industry in the United States. A small number of horizontally and/or vertically integrated companies oversee the care of over two-thirds of dialysis patients. There are many questions regarding this trends as well as the vision of these large organizations regarding the future of the ESRD program. The senior physicians in the four largest such organizations agreed to participate in a provider roundtable to share their thoughts on the following issues: What are the advantages and disadvantages of industry consolidation?; What steps has your organization taken to succeed?; What are the key issues facing this industry in the next decade?; What policy changes by the Federal Government do you anticipate?; What policy changes would you like to see? Although significant differences in specifics are clear in the responses, a recurrent theme relates to how value will be maintained in the program-the balance between high-quality outcomes and the costs of achieving these outcomes. This is clearly the challenge in the years ahead.

Transforming growth factor-beta is involved in the pathogenesis of dialysis-related amyloidosis.

BACKGROUND: Advanced glycation end product-modified beta2-microglobulin (AGE-beta2m) is an important component of dialysis-related amyloidosis (DRA). Its presence induces monocyte chemotaxis and the release of the proinflammatory cytokines through macrophage activation. Transforming growth factor-beta (TGF-beta) is a multifunctional cytokine that also has chemotactic activity for monocytes at very low (0.1 to 10 pg/mL) concentrations and inhibits proinflammatory cytokine production of macrophages. In this study, we investigated the role of TGF-beta in the pathogenesis of DRA. METHODS: We performed an immunohistochemical study of DRA tissues (8 cases) to confirm the existence of TGF-betas and their receptors; we also performed a chemotaxis assay of human monocytes as well as enzyme-linked immunosorbent assay (ELISA) of TGF-beta1, tumor necrosis factor-alpha (TNF-alpha), interleukin-1beta (IL-1beta), and interleukin-1 receptor antagonist (IL-1Ra) in the supernatant of human monocyte-derived macrophage cell culture under varying conditions of incubation with TGF-beta1, AGE-beta2m, and TGF-beta1 antibody additions. RESULTS: There was positive staining for TGF-betas (types 1, 2, and 3) and their receptors (types I, II, and III) in infiltrated macrophages (CD68+), synovial lining cell, as well as vascular walls around amyloid deposition. AGE-beta2m also induced TGF-beta1 production by macrophages in a dose-dependent manner (410 +/- 80 pg/mL at 12.5 microg/mL, 621 +/- 62 pg/mL at 25 microg/mL, and 776 +/- 62 pg/mL at 50 microg/mL of AGE-beta2m). AGE-beta2m induced significant TNF-alpha and IL-1Ra production by macrophage. The addition of exogenous TGF-beta1 (0.1 to 10 ng/mL) decreased AGE-beta2m-induced TNF-alpha production and increased IL-1Ra production in a dose-dependent fashion. IL-1beta production was not effected by any experimental conditions. In chemotaxis assay, anti-TGF-beta1 antibody (0.1 to 10 microg/mL) attenuated AGE-beta2m-induced monocyte chemotaxis. CONCLUSIONS: These results provide the first evidence to our knowledge for the presence of TGF-beta in DRA tissue, as well as the stimulatory action of AGE-beta2m on tissue macrophages. In turn, TGF-beta suppresses the proinflammatory activation of macrophages, suggesting a dual role for TGF-beta in the inflammatory process of DRA. These observations may provide a pathophysiologic link between TGF-beta and DRA.

Risk factors for hospitalization in well-dialyzed chronic hemodialysis patients.

BACKGROUND/AIMS: Hemodialysis (HD) patients are hospitalized more frequently than patients with other chronic diseases, averaging 11.5 hospital days/patient/year. Hospital costs attributable to renal failure in the US exceed $2 billion per year. The present healthcare climate continues to force dialysis providers to focus on these issues in order to optimize patient care while limiting cost. METHODS: We used a novel method for analyzing hospitalization risk, a multiple-event Cox proportional hazards model, to identify factors that influenced hospitalization in a HD unit population over a two-year period. This model allows individual patients to contribute multiple failure events to the model while controlling for the serial dependency of events. RESULTS: 178 HD patients were retrospectively examined. There were 381 hospitalizations during the study period, averaging out to 1.9 hospitalizations and 10.5 hospital days/patient-year. Substance abuse and diabetes conveyed the largest risks for hospitalization (diabetes RR: 2.09; substance abuse RR: 2.24) in the study cohort, exposing the necessity for examining practice patterns and behavioral interventions as means for improving HD patient care. CONCLUSION: Despite the small numbers of patients in this single-center HD population, the model achieved adequate statistical power. Therefore, it has the potential to serve as a continuous quality improvement (CQI) tool in particular HD patient sub-groups, or in individual HD units.

Association of morbidity with markers of nutrition and inflammation in chronic hemodialysis patients: a prospective study.

BACKGROUND: Numerous studies suggest a strong association between nutrition and clinical outcome in chronic hemodialysis (CHD) patients. Nevertheless, the pathophysiological link between malnutrition and morbidity remains to be clarified. In addition, recent evidence suggests that nutritional indices may reflect an inflammatory response, as well as protein-calorie malnutrition. In this study, we prospectively assessed the relative importance of markers of nutritional status and inflammatory response as determinants of hospitalization in CHD patients. METHODS: The study consisted of serial measurements of concentrations of serum albumin, creatinine, transferrin, prealbumin, C-reactive protein (CRP), and reactance values by bio-electrical impedance analysis (BIA) as an indirect measure of lean body mass every 3 months over a period of 15 months in 73 CHD patients. Outcome was determined by hospitalizations over the subsequent three months following each collection of data. RESULTS: Patients who required hospitalization in the three months following each of the measurement sets had significantly different values for all parameters than patients who were not hospitalized. Thus, serum albumin (3.93 +/- 0.39 vs. 3.74 +/- 0.39 g/dl), serum creatinine (11.0 +/- 3.7 vs. 9.1 +/- 3.5 mg/dl), serum transferrin (181 +/- 35 vs. 170 +/- 34 mg/dl), serum prealbumin (33.6 +/- 9.2 vs. 30.0 +/- 10.1 mg/dl), and reactance (50.4 +/- 15.6 vs. 43.0 +/- 13.0 ohms) were higher for patients not hospitalized, whereas CRP (0.78 +/- 0.89 vs. 2.25 +/- 2.72 mg/dl) was lower in patients who were not hospitalized. All differences were statistically significant (P < 0.05 for all parameters). When multivariate analysis was performed, serum CRP and reactance values were the only statistically significant predictors of hospitalization (P < 0.05 for both). When a serum CRP concentration of 0.12 mg/dl was considered as a reference range (relative risk 1.0), the relative risk for hospitalization was 7% higher (relative risk = 1.07) for a CRP concentration of 0.92 mg/dl and was 30% (relative risk = 1.30) higher for a CRP concentration of 3.4 mg/dl. When a reactance value of 70 ohms was considered as a reference range with a relative risk of 1.0, the relative risk of hospitalization increased to 1.09 for a reactance value of 43 ohms and further increased to 1.14 for a reactance value of 31 ohms. CONCLUSIONS: The results of this study strongly indicate that both nutritional status and inflammatory response are independent predictors of hospitalization in CHD patients. CRP and reactance values by BIA are reliable indicators of hospitalization. Visceral proteins such as serum albumin, prealbumin, and transferrin are influenced by inflammation when predicting hospitalization. When short-term clinical outcomes such as hospitalizations are considered, markers of both inflammation and nutrition should be evaluated.

Measurement of the delivery of dialysis in acute renal failure.

BACKGROUND: Recent studies in patients with acute renal failure (ARF) have shown a relationship between the delivered dose of dialysis and patient survival. However, there is currently no consensus on the appropriate method to measure the dose of dialysis in ARF patients. In this study, the dose of dialysis was measured by blood- and dialysate-based kinetic methods in a group of ARF patients who required intermittent hemodialysis. METHODS: Treatments were performed using a Fresenius 2008E volumetric hemodialysis machine with the ability to fractionally collect the spent dialysate. Single-, double-pool, and equilibrated Kt/V were determined from the pre-, immediate post-, and 30-minute post-blood urea nitrogen (BUN) measurements. The solute reduction index was determined from the collected dialysate, as well as the single- and double-pool Kt/V. RESULTS: Forty-six treatments in 28 consecutive patients were analyzed. The mean prescribed Kt/V (1.11 +/- 0.32) was significantly greater than the delivered dose estimated by single-pool (0.96 +/- 0.33), equilibrated (0.84 +/- 0.28), and double-pool (0.84 +/- 0.30) Kt/V (compared with prescribed, each P < 0.001). There was no statistical difference between the equilibrated and double-pool Kt/V (P = NS). The solute removal index, as determined from the dialysate, corresponded to a Kt/V of 0.56 +/- 0.27 and was significantly lower than the single-pool and double-pool Kt/V (each P < 0.001). CONCLUSION: Blood-based kinetics used to estimate the dose of dialysis in ARF patients on intermittent hemodialysis provide internally consistent results. However, when compared with dialysate-side kinetics, blood-based kinetics substantially overestimated the amount of solute (urea) removal.

Relationship of dialysis membrane and cause-specific mortality.

A number of studies have suggested that type of dialysis membrane is associated with differences in long-term outcome of patients undergoing hemodialysis, both in terms of morbidity and mortality. The purpose of this study was to determine the relationship of membrane type and specific causes of death. Data from the United States Renal Data System Case Mix Adequacy Study, a national random sample of hemodialysis patients who were alive on December 31, 1990, were used. Our study was limited to patients in this data set who were undergoing dialysis for at least 1 year (n = 4,055). For the main analytic models, membrane type was classified into two categories: unmodified cellulose or MC/SYN (which combines modified cellulose [MC] and synthetic membranes [SYN]). The relationships of membrane type and major causes of mortality were analyzed using Cox proportional hazards models, which adjusted for multiple (21) covariates, including demographics, comorbidity, Kt/V, and other parameters. Patients were censored at transplantation or 60 days after a switch to peritoneal dialysis. Compared with patients dialyzed with unmodified cellulose membranes, the adjusted relative mortality risk (RR) from infection was 31% lower (RR = 0.69; P = 0.03) and from coronary artery disease was 26% lower (RR = 0.74; P = 0.07) for patients dialyzed with MC/SYN membranes. No statistically significant difference (all P > 0.1) was found in mortality risk from cerebrovascular disease (RR = 1.08), other cardiac causes (RR = 0.86), malignancy (RR = 0.90), or other known causes (RR = 0.82) between patients dialyzed with MC/SYN compared with unmodified cellulose membranes. These results offer support to reported experimental and observational clinical studies that have found that unmodified cellulose membranes may increase the risk for both infection and atherogenesis. Further studies are necessary to evaluate the possibility of confounding factors, compare more specific membrane types, and determine the pathophysiology linking membrane type to cause-specific mortality.

Proposed clinical trials in the evaluation of intradialytic parenteral nutrition.

New revised policies relating to intradialytic parenteral nutrition (IDPN) reimbursement by Medicare have made it difficult, if not impossible, to qualify malnourished hemodialysis patients for this potentially useful therapy. These policies were adopted by Medicare because of a lack of studies that provide clear documentation of the medical benefits of IDPN or their cost-effectiveness. We propose a limited study of the role of IDPN with and without anabolic growth factors to improve predefined nutritional parameters and document its cost-effectiveness by monitoring hospitalization and mortality. The proposed study will be multicenter, prospective, limited to severely malnourished patients, and will include a control group that would receive standard of care.

Change in access blood flow over time predicts vascular access thrombosis.

BACKGROUND: Vascular access thrombosis accounts for at least $1 billion dollars in annual expenses and 25% of hospitalizations for chronic hemodialysis patients. Low vascular access blood flow (less than 800 ml/min) has been shown to modestly increase the relative risk for thrombosis in the subsequent three months. In this study, it is hypothesized that a time-dependent decrease in vascular access blood flow may be more predictive of subsequent thrombosis especially in vascular accesses with flows more than 800 ml/min, since it would indicate the development of a critical outlet stenosis in the graft. METHODS: Ninety-five accesses in 91 CHD patients were prospectively followed over 18 months. Vascular access blood flow was measured every six months by the ultrasound dilution technique. Thrombotic events were recorded during the three study periods. RESULTS: A total of 34 thrombotic events in 95 accesses were documented through the total study duration. Accesses that thrombosed had a 22% decrease in vascular access blood flow during the first observation period and a further 41% decrease during the second observation period as compared to 4% drop and 15% increase during the first and second observation periods, respectively, for accesses that did not thrombose. There was an estimated 13.6-fold (95%, confidence interval 2.68 to 69.16) increase in the relative risk of thrombosis for accesses with more than 35% decrease in vascular access blood flow compared to those accesses with no change in blood flow. There was no statistical difference in the average vascular access blood flow of all patients over the study period. CONCLUSIONS: Accesses that show a large (>15%) decrement in vascular access blood flow are associated with a high risk of thrombosis. Serial measurements of vascular access blood flow predict access thrombosis.

Prescribed versus delivered dialysis in acute renal failure patients.

The current study was designed first to determine separately the prescribed and delivered dose of dialysis and, second, to determine what factors lead to failure to deliver the prescribed dose of dialysis in patients with acute renal failure (ARF). Forty patients, who collectively underwent 136 dialysis treatments, were studied prospectively at two institutions. The results showed that almost half the prescriptions (49%) were for a Kt/V less than 1.2 and, more importantly, nearly 70% of the treatments delivered a Kt/V less than 1.2, the minimally acceptable dose defined in the Dialysis Outcomes Quality Initiative (DOQI) guidelines for chronic hemodialysis (CHD) patients. Patient predialysis weight was the most important variable associated with a low prescribed and delivered dose of dialysis, as well as lack of delivery of the prescribed dose of dialysis. From the statistical model, it is estimated that for every 10-kg increase in predialysis weight, the chance of prescribing or delivering a Kt/V less than 1.2 increased 4.6- and 1.95-fold, respectively. The lower than prescribed blood flow achieved by the temporary catheters and patients not receiving anticoagulation were variables also associated with not receiving the prescribed Kt/V. It is concluded that patients with ARF are prescribed and receive a dose of dialysis that would be considered inadequate for CHD patients. Until the association between dose of dialysis and outcome is better defined, it would be prudent that both the dialysis prescription and the delivery of dialysis to patients with ARF should be performed with the same care and goals as that currently received by patients with end-stage renal disease (ESRD).

Prognosis of patients with acute renal failure requiring dialysis: results of a multicenter study.

Despite several decades of clinical experience, the mortality rate for patients with acute renal failure (ARF) requiring dialysis remains high, and the evaluation of the patients prognosis has been difficult. To date, the Acute Physiology and Chronic Health Evaluation II (APACHE II) scoring system has been used more frequently for prediction in studies of ARF than any other scoring system, but has not been prospectively validated in controlled multicenter studies of this entity. In a multicenter, prospective, controlled trial evaluating the use of biocompatible hemodialysis membranes (BCMs) in patients with ARF, we evaluated the extent to which the APACHE II scoring system, based on the physiological variables in the 24 hours before the onset of dialysis and the presence or absence of oliguria, is predictive of outcome. Analysis of survival and recovery of renal function for the 153 patients treated in this study show that APACHE II scores are predictive both of survival and recovery of renal function, whether analyzed separately by type of dialysis membrane used (BCM or bioincompatible [BICM]) or for both groups combined (all P < 0.01). There was no evidence of a significant center effect or interaction of APACHE II score with dialysis membrane in our study. After adjusting for the APACHE II score, there was a positive effect of the BCM on both probability of survival (P < 0.05) and recovery of renal function (P < 0.01). In patients dialyzed with BCMs, oliguria at onset of dialysis had an adverse effect on both survival and recovery of renal function (both P < 0.01). Receiver operator curves (ROCs) using APACHE II score and the use of BCMs in nonoliguric patients yielded a statistically significant improvement versus the use of APACHE II score alone in the area under the curve (AUC) for survival (0.747 to 0.801; P < 0.05) and recovery of renal function (0.712 to 0.775; P < 0.05). We conclude that the use of the APACHE II score determined at the time of initiation of dialysis for patients with ARF is a statistically significant predictor of patient survival and recovery of renal function. The use of the APACHE II score measured at the time of dialysis initiation, especially when modified by the presence or absence of oliguria, should help in predicting outcome when evaluating interventions for patients with ARF.