Acute renal failure (ARF) occurs in up to 30% of all critically ill patients treated in the intensive care unit (ICU) and is associated with a particularly poor prognosis and high healthcare costs [1-6]. Of the patients who survive ARF in the ICU, 8-30% remain on long-term dialysis after hospital discharge [3,7,8]. During ARF, both intermittent and continuous renal replacement therapy effectively remove toxic substances and replace kidney function. Despite its frequent usage, however, there is still no consensus on the proper timing, dose and technique of renal replacement therapy. There is evidence that an early start of renal replacement therapy and treatment with high-filtrate flow rates improve the overall outcome and survival rate [9,10]. However, this was not confirmed by other investigators . Since ARF is the common end-point of a variety of diseases, it is important to identify those patients with a high risk of adverse outcome in the course of ARF. This might contribute to a better definition of patients who would benefit from optimised renal replacement therapy [12,13]. Here, we retrospectively analysed a mixed surgical population without pre-existing chronic renal failure who developed ARF. We assigned the patients to one of the six groups according to their underlying disease at the time of ICU admission. The aim of the study was to evaluate the outcome of ARF in each of the groups.
This was assessed by analysing the correlation between (a) the outcome and (b) a variety of potentially prognostic factors: initial diagnosis, APACHE II score, ICU length of stay, use of catecholamines, time of onset of ARF, age and gender.
In this retrospective study, we analysed all patients who were admitted to a surgical ICU without pre-existing renal insufficiency and then developed ARF, which was treated by renal replacement therapy.
Data were recorded from the charts of 306 patients over a 6-yr period. In general, renal replacement therapy was initiated in patients with oliguria (<20 mL h−1), an increase in serum creatinine concentration (>30-50% compared to baseline value), a decrease in creatinine clearance, uncontrolled uraemia, or severe electrolyte or acid-base imbalance. For purposes of analysis, the patients were assigned to one of the following six groups according to their underlying disease on ICU admission: haemorrhagic shock, post-cardiac surgery, post-liver transplantation, trauma, severe sepsis and miscellaneous. Patients were assigned to the haemorrhagic shock group after abdominal surgery (e.g. bleeding gastro-intestinal ulcers, major liver surgery, major colo-rectal surgery), major vascular surgery (e.g. aortic aneurysm surgery) or gynaecology/obstetrics surgery (e.g. atonic uterus bleeding, placenta accreta). Patients admitted after major traumatic events, such as car/motorcycle accidents or falls, were assigned to the trauma group. Severe sepsis patients fulfilled the criteria set by the Association of Critical Care Physicians and Society of Critical Care Medicine consensus conference (e.g. pneumonia, bacteraemia, severe infective fourquadrant peritonitis) (Table 1). Patients were irrevocably assigned to one of the groups at the time of ICU entry.
Patients were treated by continuous veno-venous haemodiafiltration according to the definition of Kellum and colleagues . Dialysis was performed with synthetic high-flux membranes (polysulfone or polyacrylonitrile; effective surface area of 0.9-1.2 m2) with a flow rate of 20-30 mL kg−1 h−1 (bicarbonatebuffered dialysis solution). Ultrafiltration volumes (1-4 mL kg−1 h−1) were optimised to achieve desired weight loss and enhance solute clearance by convection. In cases where renal function did not recover, renal replacement therapy was switched to intermittent dialysis.
Patient characteristics data, admission type (scheduled or emergency admission) and the APACHE II score were recorded at ICU entry. Renal function parameters, onset of ARF, start of renal replacement therapy and outcome were recorded throughout the ICU stay. Use of catecholamines was defined as continuous application of epinephrine and/or norepinephrine and/or dopamine (dopamine >4 μg kg−1 min−1). The primary outcome variable was death in the ICU.
Variables were described as mean ± SD, median with 25-75% quartiles or as frequencies. Differences between the initial diagnosis groups were evaluated by univariate comparisons: analysis of variance for continuous variables and χ2-tests for discrete data. Non-parametric tests (Wilcoxon and Kruskal-Wallis) were used, where distributions were skewed. Multiple logistic regression analysis was performed, taking the ICU mortality as the dependent variable. Variables found to be independent from the initial diagnosis grouping were taken as independent variables. The odds ratios were used to estimate the association between the covariates and the dependent variable 'death in the ICU' (univariate/multivariate logistic regression). The accuracy of the model's prediction of ICU mortality was described by a receiver operating characteristic curve.
During the study period, 13 191 patients were admitted to the ICU. Three hundred and seventy-two patients developed ARF requiring renal replacement therapy. Sixty-six patients were excluded from the analysis due to pre-existing renal failure, discharge or death within 24 h after admission. In the end, 306 patients were included in the study (2.3% of all patients admitted to the ICU). Among these patients, 71% were male. The overall median age was 60.4 yr. Patient characteristics data, APACHE II score in the first 24 h after ICU entry, use of catecholamines and baseline creatinine values did not differ significantly from one initial diagnosis group to the next (Table 2).
The mean length of stay in the ICU was 19.5 days (range: 1-129) and did not differ significantly between the initial diagnosis groups (P = 0.13, Kruskal-Wallis). The time between the onset of ARF and the actual start of renal replacement therapy was not significantly different between the initial diagnosis groups. The onset of ARF occurred after a mean of 4.3 (range: 1-36) days after ICU admission. Initiation of renal replacement therapy was on the same day as ARF diagnosis in 56% of the patients and on the day following ARF diagnosis in 31% (mean: 4.9; range: 1-37 days after admission) (Table 1). The initiation of renal replacement therapy was independent of the time of occurrence of ARF (Spearman's rank correlation), i.e. the initiation of renal replacement therapy was not delayed when ARF occurred later during the patient's ICU stay and vice versa.
A total of 51% of the patients with ARF died after the initiation of renal replacement therapy (overall mortality in the ICU during the study period was 7%). In the univariate regression analysis, the initial diagnosis on ICU admission was predictive of the outcome 'death in the ICU' (P = 0.0095). Patients with severe sepsis had a significantly higher mortality (68%) than the overall ARF patients (odds ratio: 0.35; P = 0.001). Besides the initial diagnosis, the APACHE II score was the only predictive variable present at ICU admission in the univariate analysis. After the onset of ARF, the use of catecholamines, prolonged length of ICU stay and delayed onset of ARF were risk factors for mortality (Table 3). A multivariate analysis was performed to adjust the predictors for confounding variables. Five predictors for a significantly higher risk of mortality were identified: the initial diagnosis, use of catecholamines, length of ICU stay, delayed onset of ARF, and the interaction between the length of ICU stay and the APACHE II score (Table 4). The predictors were validated by a receiver operating characteristics curve (area under curve = 0.80) (Fig. 1).
ARF remains a serious complication of critically ill patients and is associated with excess mortality [1,2]. It is difficult to define the exact contribution of renal failure to adverse outcome in patients with a complex variety of underlying diseases. In our study, we assigned patients from a single centre ICU with a heterogeneous surgical patient population, but uniform timing and technique of renal replacement therapy to one of six diagnosis groups. We retrospectively analysed our patient population for outcome and risk factors and found that the underlying diseases at the time of admission to the ICU, but before the actual occurrence of ARF, were independent risk factors for adverse outcome. The highest mortality was noted in patients with severe sepsis (68%). In addition to renal failure, circulatory failure occurred frequently in this group. Ninety-three percent of the patients with severe sepsis received circulatory support with epinephrine and/or norepinephrine, compared to 69% of all other study patients. In patients with severe sepsis, the development of ARF is the reflection of a complex state of multiorgan dysfunction that cannot be defined by severity of illness alone [12,14].
Recently, it was shown that intensive continuous renal replacement therapy improved the survival rate of critically ill patients with ARF , but the results could not be confirmed by other investigators . In addition, there is some controversy over the criteria used to decide when to initiate renal replacement therapy and whether continuous or intermittent renal replacement therapy would be more beneficial [11,12,15]. Due to the lack of consensus criteria, patient selection remains crucial [12,13]. Thus, it is important to select specific patient groups in studies comparing the efficacy of different treatment protocols. Here, we propose that the underlying disease could serve as an independent variable for patient selection in such studies. In view of the high mortality in patients with severe sepsis, the evaluation of alternative treatment protocols in this group is important for several reasons. First, these patients would benefit most dramatically from improved treatment protocols, and second, significant differences in outcome would be noted earliest in a group with high mortality.
Several studies have identified factors that are predictive of the outcome [1,16,17]. These factors have been separated into three categories: pre-existing patient health conditions, the actual clinical setting in which the ARF occurred and complications of ARF. Here, we focused on the actual clinical setting in which the ARF occurred. We found that the APACHE II score calculated within 24h of ICU admission was not a risk factor for patient outcome after adjusting for confounding factors. The value of the APACHE II score is well established, but has been a matter of controversy in the literature [2,3,18,19]. Differences might stem from the fact that the APACHE II score was measured at different time points (ICU admission as opposed to onset of ARF). In our model, we analysed the initial APACHE II score on ICU admission, i.e. before ARF actually occurred. In our study, patients who developed ARF after having a high APACHE II score on ICU arrival had an increasing risk of mortality with every additional day of ICU stay.
Interestingly, we found that patients who developed ARF after orthotopic liver transplantation had the lowest mortality (26%) of all the patients in our study. This contradicts previous studies, which showed that postoperative renal failure after liver transplantation had a strong negative impact on ARF outcome . However, transplant patients are exposed to nephrotoxic drugs. Compared to the acute tubular necrosis generally found in pre-renal failure [2,21], toxic ARF is associated with better renal recovery . This might in part explain the favourable outcome of liver-transplant patients in our study.
Brivet and colleagues found by logistic regression that delayed onset of ARF was a risk factor for mortality, but this finding was not confirmed by Guérin and colleagues [1,2]. Differences in the definitions of ARF and in the decision when to initiate renal replacement therapy may explain this discrepancy. In neither of the studies were patients selected by diagnosis. Our study could partially explain and resolve these different results. We observed that in the initial diagnosis group with the lowest mortality (post-liver transplantation), renal replacement therapy was started at the earliest time point (mean: 3.0 days after admission) of all initial diagnosis groups; whereas in the group with the highest mortality (severe sepsis), renal failure occurred at the latest time point (mean: 6.9 days). Overall, the delayed onset of ARF was a predictor for increased mortality in our study. Whether an early initiation of renal replacement therapy in patients with severe sepsis is beneficial must be evaluated in future studies.
Upon admission, patient characteristics data and clinical features were independent of the six diagnosis groups. However, we found a strong dependency of the type of admission (scheduled or emergency) on the underlying disease. Consequently, this variable was excluded from further analysis.
Our study has several limitations. The data were collected retrospectively, which means that the treatment of patients was not protocol driven, even though continuous veno-venous haemodiafiltration with flow rates between 1.5 and 2.0 L h−1 was a common practice in our ICU during the investigation period. Furthermore, we investigated exclusively surgical patients without pre-existing chronic renal failure, although it is commonly accepted that patients with altered renal function have an increased risk of developing ARF [23,24]. We elected to exclude these patients from the analysis because chronic renal failure is observed more often in a subset of our defined diagnosis groups. Several studies have shown an association between mechanical ventilation and ARF, while others have not [25-27]. Almost all patients admitted to our ICU were mechanically ventilated. Therefore, it is unlikely that mechanical ventilation could be identified as an independent risk factor in our patient population.
In conclusion, our retrospective analysis of 306 surgical patients with ARF identified the underlying diseases as the only independent predictor of mortality present at ICU admission. Patients with severe sepsis had the highest mortality. Patient selection according to the initial diagnosis could thus make a valuable contribution to study protocols assessing optimised timing and renal replacement therapy methods. We identified an additional five predictive factors for mortality in patients with ARF who are treated by renal replacement therapy. All factors, such as initial diagnosis, length of ICU stay, use of catecholamines, delayed onset of ARF, and the interaction between APACHE II and length of ICU stay, were found by multivariate analysis.
We thank Reinhard Vonthein PhD, Department of Medical Biometry, University of Tuebingen, for help with the statistical analysis. T. H. S. and M. H. contributed equally to this work.
1. Brivet FG, Kleinknecht DJ, Loirat P, Landais PJM, the French study group on acute renal failure. Acute renal failure in intensive care
units - causes, outcome, and prognosis factors of hospital mortality: a prospective, multicenter study. Crit Care Med
2. Guérin C, Girard R, Selli J, Perdrix J, Ayzac L, the Rhone-Alpes Area study group on acute renal failure. Initial versus delayed acute renal failure in the intensive care
unit. Am J Respir Crit Care Med
3. Chertow GM, Christiansen CL, Cleary PD, Munro C, Lazarus M. Prognostic stratification in critically ill patients with acute renal failure requiring dialysis. Arch Intern Med
4. Liano F, Junco E, Pascual J, Madero R, Verde E. The spectrum of acute renal failure in the intensive care
unit compared with that seen in other settings. The Madrid acute renal failure study group. Kidney Int
5. Kresse S, Schlee H, Deuber HJ, Koall W, Osten B. Influence of renal replacement therapy
on outcome of patients with acute renal failure. Kidney Int
6. Druml W. Prognosis of acute renal failure. 1975-1995. Nephron
7. Silvester W, Bellomo R, Cole L. Epidemiology, management, and outcome of severe acute renal failure of critical illness in Australia. Crit Care Med
8. Cosentino F, Chaff C, Piedmonte M. Risk factors influencing survival in ICU acute renal failure. Nephrol Dial Transplant
9. Gettings LG, Reynolds HN, Scalea T. Outcome in posttraumatic acute renal failure when continuous renal replacement therapy
is applied early vs. late. Intensive Care Med
10. Ronco C, Bellomo R, Homel P, et al.
Effects of different doses in continuous veno-venous haemofiltration on outcomes of acute renal failure: a prospective randomised trial. Lancet
11. 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
12. Kellum JA, Metha RL, Angus DC, Palevsky P, Ronco C for the ADQI workgroup. The first international consensus conference on continuous renal replacement therapy
. Kidney Int
13. 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
14. Metnitz PGH, Krenn CG, Steitzer H, et al.
Effect of acute renal failure requiring renal replacement therapy
on outcome in criticall ill patients. Crit Care Med
15. Chertow GM, Levy EM, Hammermeister KE, Grover F, Daley J. Independent association between acute renal failure and mortality following cardiac surgery. Am J Med
16. Bent P, Tan HK, Bellomo R, et al.
Early and intensive continuous hemofiltration for severe renal failure after cardiac surgery. Ann Thorac Surg
17. Lins RL, Elseviers M, Daelemans R, et al.
Prognostic value of a new scoring system for hospital mortality in acute renal failure. Clin Nephrol
18. Lins RL, Elseviers M, Daelemans R, De Broe ME. Problems in the development, validation, and adaptation of prognostic models for acute renal failure. Nephrol Dial Transplant
19. Chen Y, Hsu H, Kao K, Fang J, Huang C. Outcomes and APACHE II predictions for critically ill patients with acute renal failure requiring dialysis. Ren Fail
20. Krueger WA, Lenhart FP, Neeser G, et al.
Influence of combined intravenous and topical antibiotic prophylaxis on the incidence of infections, organ dysfunctions, and mortality in critically ill surgical patients: a prospective, stratified, randomized, double-blind, placebo-controlled clinical trial. Am J Respir Crit Care Med
21. Gonwa TA, Mai MLM, Melton LB, et al. Renal replacement therapy
and orthotopic liver transplantation: the role of continuous veno-venous hemodialysis. Transplantation
22. Metha RL, McDonald B, Gabbai FB, et al.
A randomized clinical trial of continuous versus intermittent dialysis for acute renal failure. Kidney Int
23. Platz KP, Mueller AR, Blumhardt G, et al.
Nephrotoxicity after orthotopic liver transplantation in cyclosporin A and F 506 treated patients. Transpl Int
24. Rasmussen HH, Pitt EA, Ibels LS, McNeil DR. Prediction of outcome in acute renal failure by discriminant analysis of clinical variables. Arch Intern Med
25. Conlon PJ, Stafford-Smith M, White WD, et al.
Acute renal failure following cardiac surgery. Nephrol Dial Transplant
26. Metnitz PGH, Krenn CG, Steltzer H, et al.
Effect of acute renal failure requiring renal replacement therapy
on outcome in critically ill patients. Crit Care Med
27. Hoste EA, Lameire NH, Vanholder RC, Benoit DD, Decruyenaere JM, Colardyn FA. Acute renal failure in patients with sepsis in a surgical ICU: predictive factors, incidence, comorbidity, and outcome. J Am Soc Nephrol