2007 JASN IMPACT FACTOR 7.111	HOME   AUTHOR INFO   EDITORIAL BOARD   SUBSCRIBE   FEEDBACK   ALERTS   HELP
advanced	CURRENT ISSUE	ARCHIVES	JASN Express	ONLINE SUBMISSION

This Article Abstract Full Text (PDF) All Versions of this Article: ASN.2006030268v1 17/11/3132    most recent Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Cho, K. C. Articles by Chertow, G. M. Search for Related Content PubMed PubMed Citation Articles by Cho, K. C. Articles by Chertow, G. M. Related Collections Related Articles

Survival by Dialysis Modality in Critically Ill Patients with Acute Kidney Injury

Kerry C. Cho^*, Jonathan Himmelfarb, Emil Paganini, T. Alp Ikizler, Sharon H. Soroko^||, Ravindra L. Mehta^|| and Glenn M. Chertow^*

^* Division of Nephrology, Department of Medicine, University of California San Francisco, San Francisco, California; Division of Nephrology, Department of Medicine, Maine Medical Center, Portland, Maine; Division of Nephrology, Department of Medicine, Cleveland Clinic Foundation, Cleveland, Ohio; Division of Nephrology, Department of Medicine, Vanderbilt University, Nashville, Tennessee; and ^|| Division of Nephrology, Department of Medicine, University of California San Diego, San Diego, California

Address correspondence to: Dr. Glenn M. Chertow, University of California San Francisco, Department of Medicine Research, UCSF Laurel Heights Suite 430, 3333 California Street, San Francisco, CA 94118. Phone: 415-476-2173; Fax: 415-476-1700; chertowg{at}medicine.ucsf.edu

Received for publication March 24, 2006. Accepted for publication August 21, 2006.

	Abstract

Top Abstract Introduction Materials and Methods Results Discussion References

 
Among critically ill patients, acute kidney injury (AKI) requiring dialysis is associated with mortality rates generally in excess of 50%. Continuous renal replacement therapies (CRRT) often are recommended and widely used, although data to support its superiority over intermittent hemodialysis (IHD) are lacking. Data from the Program to Improve Care in Acute Renal Disease (PICARD), a multicenter observational study of AKI, were analyzed. Among 398 patients who required dialysis, the risk for death within 60 d was examined by assigned initial dialysis modality (CRRT [n = 206] versus IHD [n = 192]) using standard Kaplan-Meier product limit estimates, proportional hazards ("Cox") regression methods, and a propensity score approach to account for selection effects. Crude survival rates were lower for patients who were treated with CRRT than IHD (survival at 30 d 45 versus 58%; P = 0.006). Adjusted for age, hepatic failure, sepsis, thrombocytopenia, blood urea nitrogen, and serum creatinine and stratified by site, the relative risk for death associated with CRRT was 1.82 (95% confidence interval 1.26 to 2.62). Further adjustment for the propensity score did not materially alter the association (relative risk 1.92; 95% confidence interval 1.28 to 2.89). Among critically ill patients with AKI, CRRT was associated with increased mortality. Although the results could reflect residual confounding by severity of illness, these data provide no evidence for a survival benefit afforded by CRRT. Larger, prospective, randomized clinical trials to compare CRRT and IHD in severe AKI are needed.

	Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

 
Acute kidney injury (AKI) frequently complicates critical illness and is associated with considerable mortality and morbidity. When severe enough to require dialysis, mortality rates in excess of 50% have been reported in most studies (1–4). Since its introduction in the late 1970s (5), continuous renal replacement therapy (CRRT), including hemofiltration and hemodiafiltration, has gained widespread acceptance in the treatment of dialysis-requiring AKI (6–10). Several clinical trials have demonstrated beneficial effects of CRRT over intermittent hemodialysis (IHD) on hemodynamic stability, solute clearance, and ultrafiltration capacity (11–16). Direct comparisons of CRRT and IHD using observational data are problematic, because patients who are hemodynamically unstable are more likely to be treated with CRRT. Attempts to account for underlying severity of illness and comorbidity have yielded disparate conclusions (17,18). Results from underpowered randomized clinical trials of CRRT and IHD have been limited and equivocal (19–21).

In this study, we analyzed the subcohort of patients from the Program to Improve Care in Acute Kidney Disease (PICARD) who required dialysis (n = 398), evaluating clinical characteristics and outcomes that were associated with the initial assigned dialysis modality (CRRT versus IHD). We hypothesized that unadjusted results would show a survival advantage to IHD and that results adjusted for confounding and selection effects would show no significant difference between assigned modality groups.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

 
Study Participants
The PICARD network is composed of five academic medical centers in the United States: University of California San Diego (Coordinating Center), Cleveland Clinic Foundation (CCF), Maine Medical Center, Vanderbilt University, and University of California San Francisco. During a 31-mo period (February 1999 to August 2001), all patients who were consulted for AKI in the intensive care unit (ICU) were evaluated by PICARD study personnel for potential study participation. Given the large number of ICU beds at CCF, one in six patients with AKI were randomly assigned for possible study inclusion to avoid single-center overrepresentation. For the PICARD study, AKI was defined as an increase in serum creatinine 0.5 mg/dl with baseline serum creatinine <1.5 mg/dl or an increase in serum creatinine 1.0 mg/dl with baseline serum creatinine 1.5 mg/dl and <5.0 mg/dl. Patients with a baseline serum creatinine 5.0 mg/dl were not considered for study inclusion. Baseline chronic kidney disease was defined as an estimated GFR <30 ml/min per 1.73 m² (corresponding to National Kidney Foundation Kidney Disease Outcomes Quality Initiative [K/DOQI] stage IV chronic kidney disease).

A detailed description of PICARD inclusion and exclusion criteria, data elements, and data collection and management strategies are described elsewhere (22). Patients who were contacted by study personnel and who signed (or whose proxy signed) informed consents were enrolled in the study cohort. The reason for nonenrollment was determined for patients who did not sign informed consent, although no additional data were collected for privacy considerations (23). The Committees on Human Research at each participating clinical site approved the study protocol and informed consent. The timing of initiation, modality, frequency, and dose of dialysis were determined by the treating physician with no influence from study personnel.

Statistical Analyses
Continuous variables were expressed as means ± SD or median and compared (by assigned modality) using t test or the Wilcoxon rank sum test, where appropriate. Categorical variables were expressed as proportions and compared with the Cochran-Mantel-Haenszel ² test or Fisher Exact test. We examined the time to death within 60 d using the Kaplan-Meier product limit estimate, and compared survival curves with the log rank test.

We created a propensity score using assigned modality as the dependent variable (24). Using multiple logistic regression, we considered as candidate variables all demographic, clinical, and laboratory factors that were associated with assigned modality on univariate analysis. We retained all variables with P < 0.20 in the propensity score. We then ranked patients by their estimated propensity score and grouped patients into tertiles. By considering outcomes within propensity categories, comparisons are closer to what might be expected if assignment were randomized (25). Discrimination of the propensity score model was assessed using the area under the receiver operating characteristic curve (26), with higher values indicating better discrimination. Calibration was assessed using the Hosmer-Lemeshow goodness-of-fit test (27). The Hosmer-Lemeshow test compares model performance (observed versus expected) across deciles of risk to test whether the model is biased (i.e., performs differentially at the extremes of risk). A nonsignificant value for the Hosmer-Lemeshow ² suggests an absence of such bias.

Proportional hazards ("Cox") regression was used to determine the associations of modality assignment and other covariates measured at the time of dialysis initiation, stratified by site (28). We included as covariates factors that were associated with mortality on the day of dialysis initiation (29). Hazard ratios (relative risks [RR]) and 95% confidence intervals (CI) were calculated from model parameter coefficients and SE, respectively. Plots of log (–log [survival rate]) against log (survival time) were performed to establish the validity of the proportionality assumption (30). We fitted models adjusted for covariates only, the propensity score only, and a combination of covariates plus the propensity score. We also fitted models within tertiles of propensity score to evaluate the consistency of the results across the range of likelihood of modality assignment.

Two-tailed P < 0.05 was considered significant. Statistical analyses were conducted using SAS 9.1 (SAS Institute, Cary, NC).

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

 
Of the 398 patients who required dialysis for severe AKI, 206 started on CRRT and 192 started on IHD. Table 1 shows demographic, historical, clinical, and selected laboratory values by initial dialysis modality. Modality assignment differed significantly by site: Initial assignment to CRRT ranged from 27% at Vanderbilt University to 61% at CCF and University of California San Diego. In general, patients who were assigned to CRRT had more organ system failure and more significant physiologic disturbances, including hypotension and tachycardia. Fewer than half of all patients had a pulmonary artery (PA) catheter in place on the day of dialysis initiation, although PA catheter use was more common, as expected, among patients who were started on CRRT (96 [47%] of 206 versus 43 [22%] of 192; P < 0.0001). Among patients with a PA catheter, there were no significant differences in mean PA systolic (P = 0.17) or diastolic (P = 0.89) pressure or pulmonary capillary wedge pressure (P = 0.54) by initial dialysis modality.

Within tertile 1 (patients whose clinical characteristics predicted a low likelihood of assignment to CRRT), 22 (17%) patients started CRRT and 110 (83%) started IHD. Within tertile 2, 71 (53%) patients started CRRT and 62 (47%) started IHD. Within tertile 3 (patients whose clinical characteristics predicted a high likelihood of assignment to CRRT), 113 (85%) patients started CRRT and 20 (15%) started IHD. Within all three tertiles, the risks for death were nominally higher with assignment to CRRT, although there was no significant difference in tertile 2, the group in which patients’ characteristics did not clearly predict one modality or another. The risks that were associated with initial assignment to CRRT within the three tertiles were as follows: Tertile 1 RR 2.88, 95% CI 1.47 to 5.66; tertile 2 RR 1.39, 95% CI 0.75 to 2.58; and tertile 3 RR 2.90, 95% CI 1.06 to 7.94).

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

 
A relatively large fraction of critically ill patients with severe AKI require dialysis during an ICU stay. The traditional approach to management is IHD, usually delivered three times per week for several hours per session, not unlike maintenance hemodialysis that is used in patients with ESRD. In the acute setting, peritoneal dialysis has fallen out of favor as a result of infectious and other (e.g., respiratory and metabolic) complications. In the past two decades, CRRT, including continuous venovenous hemodialysis (CVVHD), continuous venovenous hemofiltration, and continuous venovenous hemodiafiltration, have gained in popularity and usage, particularly for the treatment of hemodynamically unstable patients. Although several small studies have suggested improved physiologic parameters in response to CRRT relative to IHD (11–16), these data are not definitive. Evaluation of outcomes that are associated with modality choice are hampered by residual confounding and selection bias, because most programs tend to use CRRT for more severely, acutely ill patients (18,22,30). For example, Swartz et al. (18) compared mortality rates by modality assignment in 349 patients with dialysis-requiring AKI during 1995 through 1996 at the University of Michigan. The odds for death with assignment to CRRT was roughly twice that of IHD; however, after exclusion of patients with hypotension (systolic BP <90 mmHg), severe hyperbilirubinemia (>15 mg/dl), and a very short period of renal replacement (<48 h), the risk for death that was associated with CRRT no longer was significantly higher (RR 1.09; 95% CI 0.67 to 1.80). In a re-analysis of this cohort, Martin et al. (31) found that the CRRT-treated patients had an unexpected higher mortality, particularly in patients who were categorized as low risk by the Cleveland Clinic score. In a follow-up study during 2000 through 2001 (n = 383), Swartz et al. (32) found an increase in the risk for death with CRRT on unadjusted analysis and no significant difference after adjustment for comorbidity and severity of illness. Subgroup analyses were conducted, some of which suggested favorable trends with CRRT, although none was statistically significant and there was no consideration of multiple comparisons. Chang et al. (33) described 148 South Korean ICU patients with dialysis-requiring AKI. As expected, patients who were treated with continuous venovenous hemodiafiltration were more severely ill and had significantly lower survival rates (21 versus 46%; P = 0.002). On subgroup analysis, patients with APACHE III scores >103 and more than three organ failures had nominally higher survival with CRRT. Other observational studies have demonstrated conflicting findings, examining mortality and renal recovery after dialysis-requiring AKI.

Several randomized clinical trials have compared CRRT and IHD in severe AKI, although none has been adequately powered. In the largest randomized clinical trial, Mehta et al. (19) compared CRRT and IHD in 166 critically ill patients with severe AKI and found a significantly higher ICU mortality rate in patients who were randomly assigned to CRRT (60 versus 42%; P = 0.02). However, despite randomization, patients who were assigned to CRRT were significantly more likely to have liver failure and had a higher overall severity of illness, as determined by APACHE III score. Adjustment for these factors attenuated the increased risk that was attributed to CRRT (OR 1.6; 95% CI 0.7 to 3.3). More recently, two (underpowered) randomized clinical trials that compared CRRT and IHD failed to show a significant difference in survival by dialysis modality (21,34). Meta-analyses also have been conducted and have concluded that there is no significant difference in survival by modality, although study quality and heterogeneity were not optimal (17,20).

This study extends previous work in this area using observational data by incorporating multiple sites, a larger sample size, multivariable regression analysis, and the propensity score approach to address residual confounding and selection effects. Although data collection in PICARD generally was comprehensive, all data elements were not collected in all patients (e.g., data from PA catheters), largely as a result of differences in clinical practice among and within sites. In addition, we could not control for other aspects of dialysis care (e.g., urea or other solute clearance) and other co-interventions that were not randomly assigned.

Although propensity scores cannot fully adjust for residual confounding and selection bias, the method has been widely used in observational studies that have examined the effectiveness of various interventions in nephrology and critical care. As with other conditions for which there is uncertainty as to the optimal therapeutic approach, there tends to be wide variation in practice by institution and individual physician in the choice of modality for dialysis-requiring AKI. A propensity score can help account for this variation, which may be unrelated to severity of illness or other biologic factors that influence outcomes. We previously used propensity scores to estimate the effects of the timing of consultation (35) and the use of diuretics (36) and dopamine (37) in AKI. Propensity scores also were used in studies that evaluated the effectiveness of albumin administration (38), blood transfusion (39), and right heart catheterization (40) in the critically ill.

It should be emphasized that the multivariable analyses (with or without the propensity score) support the crude (unadjusted) results that demonstrate an increased risk for death among patients who are assigned to CRRT. In the "naïve" multivariable analyses (not adjusted for the propensity score), we adjusted for significant predictors of death at dialysis initiation. Inclusion of other covariates (e.g., mechanical ventilation, clinical criteria for acute lung injury or adult respiratory distress syndrome, systolic BP) in addition to or in place of the core model covariates did not extinguish the association between assignment to CRRT and mortality. Moreover, when considering factors that were associated with assignment to CRRT but were not significant predictors of death (e.g., use of PA catheter, intake-output balance), we observed risk ratios in the same direction and of the same magnitude as in the unadjusted analysis.

There are several important limitations to this study. First, propensity scores can adjust only for the associations among observed covariates and the chosen treatment or strategy. Other unobserved covariates could influence the likelihood of treatment, and there is no guarantee that the correlation among observed and unobserved covariates is sufficiently high to account adequately for this deficiency. Second, the study was conducted at five academic tertiary care medical centers. Therefore, results may not be fully generalizable to other medical centers, particularly those where CRRT is applied less frequently. Third, we collected no information on long-term survival, functional status, or dialysis dependence for patients who survived hospitalization. Future observational studies and clinical trials in AKI should attempt to understand the long-term effects of dialysis-requiring AKI episodes. Finally, although an extensive number of variables were collected and were done so serially during patients’ ICU stays, we could not capture every aspect of intensive care, so residual confounding by severity of illness is likely.

Although the major results described here could reflect residual confounding, the possibility that CRRT might cause harm still should be considered. CRRT requires continuous anticoagulation; may remove water-soluble vitamins, drugs (including antibiotics), and amino acids; and may result in clearance and/or adsorption of a variety of known and unknown modulators of the inflammatory and counterinflammatory response. Moreover, despite technological advances, it remains a complex intensive therapy that requires considerable nursing and physician expertise. Given the high incidence of AKI in the ICU and the morbidity, associated mortality, and costs that are associated with dialysis-requiring AKI, better evidence is needed to guide AKI treatment strategies.

There are numerous examples in which drugs, devices, and technologies have been introduced into medical practice on the basis of a sound rationale, yet subsequent clinical practice and research demonstrate that the use of these technologies is associated with no benefit or even harm. Recent examples include the use of PA catheters in the ICU (41,42), the use of certain antiarrhythmic agents in an attempt to prevent sudden cardiac death (43), and the use of selected inotropes and vasodilators in the treatment of congestive heart failure (44). When the results of the current observational study on dialysis-requiring AKI in the ICU are integrated with other observational studies and clinical trials, there is no evidence of the superiority of CRRT over IHD and some evidence to support possible inferiority.

At this time, the data that are presented here should be considered hypothesis generating and not definitive and should not be used to change clinical practice. However, in the context of increasing CRRT use, it now is imperative that a randomized clinical trial of adequate power be conducted to determine whether mortality rates that are associated with severe dialysis-requiring AKI can be reduced with the use of CRRT, IHD, or a hybrid technique, such as slow low-efficiency dialysis. Such a study will need to include patients who could be treated successfully with either modality (comparing "apples and apples"), potentially excluding patients with severe hypotension and hemodynamic instability, who may be poor candidates for traditional IHD, and should standardize key elements of therapy, including the timing of initiation, dosage of dialysis, and the expertise of personnel delivering the therapy.

View larger version (23K): [in this window] [in a new window] [as a PowerPoint slide]   Figure 1. Mortality within 60 d after acute kidney injury requiring dialysis: Continuous renal replacement therapies versus intermittent hemodialysis.

	Acknowledgments

	Footnotes

 
Published online ahead of print. Publication date available at www.jasn.org.

This controlled study by Cho and colleagues on the benefits of continuous versus intermittent dialysis in treating acute kidney injury relates to a Mini-Review by Van Biesen et al. in this month’s issue of CJASN (pp. 1314–1319) that discusses how acute kidney injury is currently defined and the use of the RIFLE criteria.

 

	References

Top Abstract Introduction Materials and Methods Results Discussion References

 

1. Himmelfarb J, Tolkoff Rubin N, Chandran P, Parker RA, Wingard RL, Hakim R: A multicenter comparison of dialysis membranes in the treatment of acute renal failure requiring dialysis. J Am Soc Nephrol 9 : 257 –266, 1998[Abstract]
2. Gastaldello K, Melot C, Kahn RJ, Vanherweghem JL, Vincent JL, Tielemans C: Comparison of cellulose diacetate and polysulfone membranes in the outcome of acute renal failure. A prospective randomized study. Nephrol Dial Transplant 15 : 224 –230, 2000[Abstract/Free Full Text]
3. Marshall MR, Golper TA, Shaver MJ, Alam MG, Chatoth DK: Sustained low-efficiency dialysis for critically ill patients requiring renal replacement therapy. Kidney Int 60 : 777 –785, 2001[CrossRef][Medline]
4. Bouman CS, Oudemans-Van Straaten HM, Tijssen JG, Zandstra DF, Kesecioglu J: Effects of early high-volume continuous venovenous hemofiltration on survival and recovery of renal function in intensive care patients with acute renal failure: A prospective, randomized trial. Crit Care Med 30 : 2205 –2211, 2002[CrossRef][Medline]
5. Kramer P, Wigger W, Rieger J, Matthaei D, Scheler F: Arteriovenous haemofiltration: A new and simple method for treatment of over-hydrated patients resistant to diuretics [in German]. Klin Wochenschr 55 : 1121 –1122, 1977[CrossRef][Medline]
6. Mehta RL, Letteri JM: Current status of renal replacement therapy for acute renal failure. A survey of US nephrologists. The National Kidney Foundation Council on Dialysis. Am J Nephrol 19 : 377 –382, 1999[CrossRef][Medline]
7. Silvester W, Bellomo R, Cole L: Epidemiology, management, and outcome of severe acute renal failure of critical illness in Australia. Crit Care Med 29 : 1910 –1915, 2001[CrossRef][Medline]
8. Wright SE, Bodenham A, Short AI, Turney JH: The provision and practice of renal replacement therapy on adult intensive care units in the United Kingdom. Anaesthesia 58 : 1063 –1069, 2003[CrossRef][Medline]
9. Hyman A, Mendelssohn DC: Current Canadian approaches to dialysis for acute renal failure in the ICU. Am J Nephrol 22 : 29 –34, 2002[CrossRef][Medline]
10. Ronco C, Zanella M, Brendolan A, Milan M, Zamperetti N, Bellomo R: Answers from the first international course on critical care nephrology questionnaire. Contrib Nephrol 1 : 196 –209, 2001
11. Davenport A, Will EJ, Davison AM: Continuous vs. intermittent forms of haemofiltration and/or dialysis in the management of acute renal failure in patients with defective cerebral autoregulation at risk of cerebral oedema. Contrib Nephrol 93 : 225 –233, 1991[Medline]
12. Davenport A, Will EJ, Davison AM: Effect of renal replacement therapy on patients with combined acute renal and fulminant hepatic failure. Kidney Int Suppl 41 : S245 –S251, 1993[Medline]
13. Clark WR, Mueller BA, Alaka KJ, Macias WL: A comparison of metabolic control by continuous and intermittent therapies in acute renal failure. J Am Soc Nephrol 4 : 1413 –1420, 1994[Abstract]
14. Misset B, Timsit JF, Chevret S, Renaud B, Tamion F, Carlet J: A randomized cross-over comparison of the hemodynamic response to intermittent hemodialysis and continuous hemofiltration in ICU patients with acute renal failure. Intensive Care Med 22 : 742 –746, 1996[Medline]
15. John S, Griesbach D, Baumgartel M, Weihprecht H, Schmieder RE, Geiger H: Effects of continuous haemofiltration vs intermittent haemodialysis on systemic haemodynamics and splanchnic regional perfusion in septic shock patients: A prospective, randomized clinical trial. Nephrol Dial Transplant 16 : 320 –327, 2001[Abstract/Free Full Text]
16. Tan HK, Bellomo R, M’Pisi DA, Ronco C: Phosphatemic control during acute renal failure: Intermittent hemodialysis versus continuous hemodiafiltration. Int J Artif Organs 24 : 186 –191, 2001[Medline]
17. Kellum JA, Angus DC, Johnson JP, Leblanc M, Griffin M, Ramakrishnan N, Linde-Zwirble WT: Continuous versus intermittent renal replacement therapy: A meta-analysis. Intensive Care Med 28 : 29 –37, 2002[CrossRef][Medline]
18. Swartz RD, Messana JM, Orzol S, Port FK: Comparing continuous hemofiltration with hemodialysis in patients with severe acute renal failure. Am J Kidney Dis 34 : 424 –432, 1999[Medline]
19. Mehta RL, McDonald B, Gabbai FB, Pahl M, Pascual MT, Farkas A, Kaplan RM: A randomized clinical trial of continuous versus intermittent dialysis for acute renal failure. Kidney Int 60 : 1154 –1163, 2001[CrossRef][Medline]
20. Tonelli M, Manns B, Feller-Kopman D: Acute renal failure in the intensive care unit: A systematic review of the impact of dialytic modality on mortality and renal recovery. Am J Kidney Dis 40 : 875 –885, 2002[CrossRef][Medline]
21. Augustine JJ, Sandy D, Seifert TH, Paganini EP: A randomized controlled trial comparing intermittent with continuous dialysis in patients with ARF. Am J Kidney Dis 44 : 1000 –1007, 2004[CrossRef][Medline]
22. Mehta RL, Pascual MT, Soroko SH, Savage BR, Himmelfarb J, Ikizler TA, Paganini EP, Chertow GM: Spectrum of acute renal failure in the intensive care unit: The PICARD experience. Kidney Int 66 : 1613 –1621, 2004[CrossRef][Medline]
23. Chertow GM, Pascual MT, Soroko S, Savage BR, Himmelfarb J, Ikizler TA, Paganini EP, Mehta RL: Reasons for non-enrollment in a cohort study of ARF: The Program to Improve Care in Acute Renal Disease (PICARD) experience and implications for a clinical trials network. Am J Kidney Dis 42 : 507 –512, 2003[CrossRef][Medline]
24. Rosenbaum PR, Rubin DB: Reducing bias in observational studies using subclassification on the propensity score. J Am Stat Assoc 79 : 516 –524, 1984[CrossRef]
25. Rubin DB: Estimating causal effects from large data sets using propensity scores. Ann Intern Med 127 : 757 –763, 1997[Abstract/Free Full Text]
26. Hanley JA, McNeil BJ: The meaning and use of the area under a receiver operating characteristic (ROC) curve. Radiology 143 : 29 –36, 1982[Abstract/Free Full Text]
27. Hosmer DW, Lemeshow S: Applied Logistic Regression, New York, John Wiley & Sons, 1989
28. Cox DR: Regression models and life tables. J Royal Stat Soc [B] 74 : 187 –220, 1972
29. Chertow GM, Soroko SH, Paganini EP, Cho KC, Himmelfarb J, Ikizler TA, Mehta RL: Mortality after acute renal failure: Models for prognostic stratification and risk adjustment. Kidney Int 70 : 1120 –1126, 2006[CrossRef][Medline]
30. Guerin C, Girard R, Sellit JM, Ayzac L: Intermittent versus continuous renal replacement therapy for acute renal failure in intensive care units: Results from a multicenter prospective epidemiological survey. Intensive Care Med 28 : 1411 –1418, 2002[CrossRef][Medline]
31. Martin C, Saran R, Leavey S, Swartz RD: Predicting the outcome of renal replacement therapy in severe acute renal failure. ASAIO J 48 : 640 –644, 2002[CrossRef][Medline]
32. Swartz RD, Bustami RT, Daley JM, Gillespie BW, Port FK: Estimating the impact of renal replacement therapy choice on outcome in severe acute renal failure. Clin Nephrol 63 : 335 –345, 2005[Medline]
33. Chang JW, Yang WS, Seo JW, Lee JS, Lee SK, Park SK: Continuous venovenous hemodiafiltration versus hemodialysis as renal replacement therapy in patients with acute renal failure in the intensive care unit. Scand J Urol Nephrol 38 : 417 –421, 2004[CrossRef][Medline]
34. Uehlinger DE, Jakob SM, Ferrari P, Eichelberger M, Huynh-Do U, Marti HP, Mohaupt MG, Vogt B, Rothen HU, Regli B, Takala J, Frey FJ: Comparison of continuous and intermittent renal replacement therapy for acute renal failure. Nephrol Dial Transplant 20 : 1630 –1637, 2005[Abstract/Free Full Text]
35. Mehta RL, McDonald B, Gabbai F, Pahl M, Farkas A, Pascual MTA, Zhuang S, Kaplan R, Chertow GM: Nephrology consultation in acute renal failure: Does timing matter? Am J Med 113 : 456 –461, 2002[CrossRef][Medline]
36. Mehta RL, Pascual MT, Soroko SH, Chertow GM: Diuretic use, mortality and non-recovery of renal function in acute renal failure. JAMA 288 : 2547 –2553, 2002[Abstract/Free Full Text]
37. Chertow GM, Sayegh MH, Allgren RL, Lazarus JM: Is the administration of dopamine associated with adverse or favorable outcomes in acute renal failure? Am J Med 101 : 49 –53, 1996[CrossRef][Medline]
38. Vincent JL, Sakr Y, Reinhart K, Sprung CL, Gerlach H, Ranieri VM: Is albumin administration in the acutely ill associated with increased mortality? Results from the SOAP study. Crit Care 9 : R745 –R754, 2005[CrossRef][Medline]
39. Vincent JL, Baron JF, Reinhart K, Guttinoni L, Thijs L, Webb A, Meier-Hellman A, Nollet G, Peres-Bota D: Anemia and blood transfusion in critically ill patients. JAMA 288 : 1499 –1507, 2002[Abstract/Free Full Text]
40. Connors AF Jr, Speroff T, Dawson NV, Thomas C, Harrell FE Jr, Wagner D, Desbiens N, Goldman L, Wu AW, Califf RM, Fulkerson WJ Jr, Vidaillet H, Broste S, Bellamy P, Lynn J, Knaus WA: The effectiveness of right heart catheterization in the initial care of critically ill patients. JAMA 276 : 889 –897, 1996[Abstract]
41. Shah MR, Hasselblad V, Stevenson LW, Binanay C, O’Connor CM, Sopko G, Califf RM: Impact of the pulmonary artery catheter in critically ill patients: Meta-analysis of randomized clinical trials. JAMA 294 : 1664 –1670, 2005[Abstract/Free Full Text]
42. Wheeler AP, Bernard GR, Thompson BT, Schoedfeld D, Wiedemann HP, deBoisblanc B, Connors AF Jr, Hite RD, Harabin AL: Pulmonary-artery versus central venous catheter to guide treatment of acute lung injury. N Engl J Med 354 : 2273 –2274, 2006[CrossRef][Medline]
43. Echt DS, Liebson PR, Mitchell LB, Peters RW, Obias-Manno D, Barker AH, Arensberg D, Baker A, Friedman L, Greene HL: Mortality and morbidity in patients receiving encainide, flecainide, or placebo. The Cardiac Arrhythmia Suppression Trial. N Engl J Med 324 : 781 –788, 1991[Abstract]
44. Cohn JN, Goldstein SO, Greenberg BH, Lorell BH, Bourge RC, Jaski BE, Gottlieb SO, McGrew F 3rd, DeMets DL, White GBG: A dose-dependent increase in mortality with vesnarinone among patients with severe heart failure. Vesnarinone Trial Investigators. N Engl J Med 339 : 1810 –1816, 1998[Abstract/Free Full Text]

This article has been cited by other articles:

				B. Ponte, C. Felipe, A. Muriel, M. T. Tenorio, and F. Liano Long-term functional evolution after an acute kidney injury: a 10-year study Nephrol. Dial. Transplant., December 1, 2008; 23(12): 3859 - 3866. [Abstract] [Full Text] [PDF]

				S. S. Waikar, K. D. Liu, and G. M. Chertow Diagnosis, Epidemiology and Outcomes of Acute Kidney Injury Clin. J. Am. Soc. Nephrol., May 1, 2008; 3(3): 844 - 861. [Abstract] [Full Text] [PDF]

				C. Ronco, J. A. Kellum, R. Bellomo, and A. A. House Potential Interventions in Sepsis-Related Acute Kidney Injury Clin. J. Am. Soc. Nephrol., March 1, 2008; 3(2): 531 - 544. [Abstract] [Full Text] [PDF]

				C. Ronco and R. Bellomo Dialysis in Intensive Care Unit Patients with Acute Kidney Injury: Continuous Therapy is Superior Clin. J. Am. Soc. Nephrol., May 1, 2007; 2(3): 597 - 600. [Full Text] [PDF]

				J. Himmelfarb Continuous Renal Replacement Therapy in the Treatment of Acute Renal Failure: Critical Assessment Is Required Clin. J. Am. Soc. Nephrol., March 1, 2007; 2(2): 385 - 389. [Abstract] [Full Text] [PDF]

This Article Abstract Full Text (PDF) All Versions of this Article: ASN.2006030268v1 17/11/3132    most recent Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Cho, K. C. Articles by Chertow, G. M. Search for Related Content PubMed PubMed Citation Articles by Cho, K. C. Articles by Chertow, G. M. Related Collections Related Articles