RIFLE Criteria for Cardiac Surgery–Associated Acute Kidney Injury: Risk Factors and Outcomes
Augusto D’Onofrio, MD, Division of Cardiac Surgery, San Bortolo Hospital, Viale Rodolfi, 37, 36100 Vicenza, Italy
The aims of this study were to identify risk factors and evaluate the association with clinical outcomes of postoperative cardiac surgery–associated acute kidney injury (CSA-AKI). Data from 2488 consecutive adult patients were analyzed. Patients were classified as having CSA-AKI based on the risk, injury, failure, loss of kidney function, and end-stage kidney disease (RIFLE) criteria using peak postoperative creatinine in the postoperative intensive care unit (ICU). Multiple stepwise logistic regression analysis was used to identify independent risk factors for CSA-AKI. CSA-AKI occurred in 584 patients (23.5%). CSA-AKI patients had significantly longer aortic cross-clamp and cardiopulmonary bypass times. Furthermore, CSA-AKI patients had higher hospital mortality (5.5% vs 1.5%, P<.001) and significantly longer ICU and hospital stays. Independent risk factors for CSA-AKI were age, peripheral vascular disease, hypertension, left ventricular ejection fraction, cardiopulmonary bypass time, and surgery on the thoracic aorta. In conclusion, patients who develop CSA-AKI have a higher preoperative risk profile, more complex surgery, and worse clinical outcomes. Congest Heart Fail. 2010;16(4)(suppl 1):S32–S36. ©2010 Wiley Periodicals, Inc.
Acute kidney injury (AKI) is a severe postoperative complication after cardiac surgery. It is associated with increased short-term mortality, morbidity, and intensive care unit (ICU) stay, especially if there is requirement for renal replacement therapy (RRT).1,2
In 2004, the Acute Dialysis Quality Initiative (ADQI) workgroup3 proposed a multilevel classification system for AKI identified by the acronym RIFLE (risk, injury, failure, loss of kidney function, end-stage kidney disease). The RIFLE criteria system stratifies renal failure into 3 grades of increasing severity of AKI (risk, injury, failure) and 2 outcome classes (loss of kidney function, end-stage kidney disease). The 3 grades of severity for AKI can be based on changes in either plasma creatinine or urine output from baseline or both. RIFLE criteria have been validated in cardiac surgery,4,5 but epidemiology, pathophysiology, prevention, therapy, and consequences of cardiac surgery–associated (CSA) AKI are still not well-established. Furthermore, risk factors for CSA AKI defined with the RIFLE criteria still have not been identified. Aims of this single-center retrospective study were to identify independent risk factors for AKI and to evaluate its association with clinical outcomes in patients undergoing cardiac surgery.
Patients and Methods
This study was approved by our institutional review committee and all patients gave informed consent.
We analyzed data from 2715 consecutive patients who underwent scheduled or emergency surgery for acquired cardiovascular diseases at our center from January 2003 to March 2008. After the exclusion of 208 patients for incompleteness of data and 19 for preoperative chronic dialysis therapy, 2488 patients were considered eligible and included in the study.
Data were prospectively included in our institutional database and retrospectively analyzed. All procedures, except for off-pump coronary artery bypass grafting, were performed with moderately hypothermic cardiopulmonary bypass (roller pump), and cardioplegic arrest of the heart was obtained with antegrade and retrograde intermittent cold cristalloid cardioplegia (doses repeated every 20 minutes). Standard cannulation (right atrium or venae cavae and ascending aorta) was used in coronary artery bypass grafting (CABG) and valve operations, while surgery on the ascending aorta/aortic arch arterial cannulation was mainly achieved via the right axillary artery.
Patients were classified as having CSA AKI based on the RIFLE criteria using the peak postoperative plasma creatinine in the ICU (Table I). For baseline creatinine, we used that of hospital admission. Patients who met the RIFLE criteria for CSA AKI were classified as “AKI,” whereas those who did not were classified as “no AKI.” Patients with CSA AKI were stratified according to the maximum RIFLE class (risk, injury, failure) reached during the hospital stay. RIFLE classes loss of kidney function and end-stage kidney disease were not evaluated since they reflect late outcomes and these data were not available in our database.
Table I. Description of RIFLE Classification for AKI Used in This Study (Creatinine Criteria Only)
|Risk||Increased creatinine × 1.5 (of baseline)|
|Injury||Increased creatinine × 2.0 (of baseline)|
|Failure||Increased creatinine × 3.0 (of baseline)|
|Loss||Persistent acute renal failure = complete loss of kidney function >4 weeks|
|ESKD||ESKD (>3 months)|
Categoric variables are expressed as percentages and continuous variables are expressed as mean ± SD. All statistical analyses were performed with SPSS software (SPSS Inc, Chicago, IL). Mann–Whitney U test was used to compare continuous variables; chi-square or Fisher exact test were used to compare proportions between groups. Stepwise multiple logistic regression analysis was performed to identify factors that were independently associated with the development of postoperative CSA AKI; in this analysis, the use of quartiles as continuous covariables was made when appropriate. Area under the receiver operating characteristic (ROC) curve was determined to confirm the usefulness of the risk predicting model. An area >0.7 was considered acceptable. A P value <.05 was considered statistically significant.
Mean age of the overall population was 67.9±11 years, and 26.5% of patients were female. Preoperative clinical characteristics of the general population are listed in Table II. Isolated CABG was performed in 1241 patients (49.9%); of these, 246 were off-pump. Isolated valve operations were performed in 480 patients (19.3%), and surgery of the ascending aorta and of the left ventricular wall were performed in 306 (12.3%) and 40 (1.6%) patients, respectively. Other procedures, including associated operations, were performed in 420 patients (16.9%).
Table II. Clinical Variables of the Overall Population and of the AKI and No AKI Populations
|EuroSCORE risk factors|
| Age, y, mean±SD|| 67.9±11|| 70.3±10||67.9±11||<.001|
| Female, %||26.5||24.8||27||ns|
| Chronic pulmonary disease, %||2.2||2.9||2||ns|
| Extra cardiac arteriopathy, %||20.4||26.2||18.7||<.001|
| Neurologic dysfunction, %||2||1.7||2||ns|
| Previous cardiac surgery, %||6.6||7.2||6.5||ns|
| Plasma creatinine >200 μmol/L, %||2.4||2.9||2.2||ns|
| Active endocarditis, %||0.5||0.3||0.5||ns|
| Critical preoperative state, %||4.6||5||4.5||ns|
| Unstable angina, %||8.5||9||8.3||ns|
| LVEF 30%–50%, %||24.5||26.9||23.8||ns|
| LVEF <30%, %||2.6||2.6||2.6||ns|
| Recent myocardial infarction, %||16.8||16.3||18.7||ns|
| Pulmonary hypertension, %||2.1||2.7||1.9||ns|
| Emergency, %||7.3||10.3||6.4||.002|
| Surgery on thoracic aorta, %||11.8||17.6||10||<.001|
| Postinfarct septal rupture, %||0.2||0.5||0.2||ns|
| Logistic EuroSCORE, mean±SD|| 8.0±9|| 10.4±11|| 7.2±8.8||<.001|
| Additive EuroSCORE, mean±SD|| 5.8±3.2|| 6.9±3.1|| 5.5±3.2||<.001|
|Other risk factors|
| Diabetes, %||23.4||24||23.2||ns|
| Arterial hypertension, %||77.4||83||75.6||<.001|
| CVA, %||9.1||10.4||8.6||ns|
| Previous AMI, %||34.2||36.2||33.3||ns|
| LVEF, mean±SD|| 57.0±13|| 55.6±14|| 57±13||.006|
| Aortic dissection, %||3.9||7||2.9||<.001|
| OPCAB, %||9.9||8.7||10.3||ns|
| Aortic cross-clamp time (min), mean±SD|| 58.0±35|| 64.8±38||56.0±34||<.001|
| CPB time, mean±SD||103.2±54||116.7±61||99.2±52||<.001|
|Nephrologic risk factors|
| Preoperative creatinine, mean±SD|| 1.2±0.7|| 1.1±0.4|| 1.2±0.7||ns|
| GFR, mean±SD|| 68±19|| 69±21|| 68±18||ns|
| CKD, %|
| Stage 1||7.7||8.7||7.3|| |
| Stage 2||56.5||54.4||57.1|| |
| Stage 3||33||33.7||32.8||ns|
| Stage 4||1.9||2.6||1.7|| |
| Stage 5||0.9||0.5||1|| |
| Isolated CABG, %||49.9||40.6||52.7||<.001|
| Isolated valve, %||19.3||19.7||19.2||ns|
| Great vessels, %||12.3||18||10.6||<.001|
| Surgical repair of left ventricular wall, %||1.6||1.7||1.5||ns|
| Other, %||16.9||20||16||ns|
Postoperative CSA AKI occurred in 584 patients (23.5%). The peak RIFLE class was reached as follows: risk, 228 (9.2%) patients; injury, 248 (9.7%) patients; and failure, 108 (4.3%) patients. Patients were divided into 2 groups according to the development of CSA AKI (AKI, no AKI). The preoperative characteristics of the 2 groups are listed in Table II.
Patients in whom CSA AKI developed were older (70.3±10 vs 67.9±11 years; P<.001) and had higher incidence of extracardiac arterial disease (26.2% vs 18.7%; P<.001) and higher logistic and additive EuroSCORE values (logistic EuroSCORE, 10.4%±11% vs 7.2%±8.8%; P<.001). Conversely, isolated CABG surgery was more common in the no AKI group (52.7% vs 40.6%, P<.001). Furthermore, patients with AKI had a higher incidence of emergency surgery (10.3% vs 6.4%, P=.002) and surgery on the aorta (17.6% vs 10%; P<.001). Aortic cross-clamp time (64.8±38 vs 56±34 minutes; P<.001) and cardiopulmonary bypass time (116.7±61 vs 99.2±52 minutes; P<.001) were longer in CSA AKI patients. Hospital mortality was 1.5% in patients without CSA AKI and 5.5% in patients with CSA AKI (P<.001). In particular, in RIFLE classes risk, injury, and failure, hospital mortality occurred in 10 (4.4%), 13 (5.2%), and 9 (8.3%) patients, respectively (P=.007). Postoperative dialysis was necessary in 9 (3.9%), 9 (3.6%), and 14 (12.9%) of patients in classes risk, injury, and failure, respectively. Furthermore, patients with CSA AKI had longer ICU stay (88.3±101.9 vs 55.9±66.8 hours; P<.001) and longer hospitalization (9.5±7.6 vs 8.2±9.2 days; P<.001) when compared with no AKI patients.
In multivariate analysis, RIFLE class failure was identified as an independent predictor of hospital mortality in our population (odds ratio [OR], 3.8; 95% confidence interval [CI], 1.39–10.13; P=.009; area under the ROC, 0.86). Age, extracardiac arterial disease, systemic arterial hypertension, lower left ventricular ejection fraction, cardiopulmonary bypass time, and surgery on the aorta were independent risk factors for CSA AKI (area under ROC, 0.71) (Table III). Independent risk factors for severe CSA AKI (class failure) were pulmonary hypertension, defined as a systolic pulmonary artery pressure >60 mm Hg (OR, 2.6; 95% CI, 1.08–6.43; P=.03); systemic arterial hypertension (OR, 1.86; 95% CI, 1.02–3.38; P=.04); cardiopulmonary bypass time (minutes, continuous variable) (OR, 1.48; 95% CI, 1.23–1.78; P<.001); and preoperative chronic kidney disease (stage 4 and 5) (OR, 6.31; 95% CI, 3.40–11.72; P<.001).
Table III. Independent Risk Factors for Postoperative AKI
|Age (increments of 10 y)||1.32||1.20–1.44||<.001||1.40||1.26–1.56||<.001|
|Extracorporeal circulation time (quartiles)||1.32||1.21–1.44||<.001||1.26||0.15–1.39||<.001|
AKI remains a major problem in critically ill patients and particularly in those undergoing heart surgery. In fact, observational studies have demonstrated that cardiac surgery is the second most common cause of AKI in the ICU.6–8 AKI after cardiac surgery is associated with worse outcomes in terms of short- and long-term survival, hospital stay,1,2,9 and financial costs.10 CSA AKI, defined as a deterioration of kidney function after cardiac surgery, has several pathophysiologic processes: exogenous and endogenous toxins, metabolic factors, ischemia-reperfusion, neurohormonal activation, inflammation, and oxidative stress. These processes act during all the epochs that characterize a cardiac surgery patient: preoperative, intraoperative, and postoperative period.11
More than 30 different definitions of AKI have been used in the literature, thus leading to a wide range of estimates regarding incidence of CSA AKI and putative risk factors as well as tremendous difficulty comparing between studies. The RIFLE classification for AKI was introduced in 2004 by the ADQI workgroup3 and arose from the need for a common definition and classification of one of the major intensive care syndromes. Furthermore, the RIFLE classification suggests that even mild increase of creatinine levels may indicate subtle but clinically important kidney dysfunction and negatively affect patients’ outcomes.12 The incidence of postoperative AKI following cardiac surgery defined with the RIFLE criteria reported in the literature ranges from 19% to 48%, according to different studies.4,5,9 Our analysis confirms these data, since CSA AKI in our population occurred in 23.5% of patients. We observed a progressive increase in mortality with worsening RIFLE class, specifically, patients with “AKI failure” had a mortality risk 2-fold higher than patients with “AKI risk” and almost 6-fold higher than no AKI patients. We identified that patients in whom CSA AKI develops are those with more preoperative comorbidities as summarized by the higher EuroSCORE value (both additive and logistic) and those who undergo more complex surgery, as indicated by the longer cardiopulmonary bypass times. This seems to identify a particular population with kidneys “at risk” after cardiac surgery. All efforts should be made to prevent the onset of CSA AKI in these patients and to diagnose in a timely way and aggressively treat patients with CSA AKI since RIFLE failure has been identified as an independent predictor for hospital mortality. Prevention strategies should be focused on hemodynamic management, pharmacologic, and nonpharmacologic strategies. While correction of hemodynamic impairment is crucial, pharmacologic strategies (loop diuretics, vasodilators, dopamine, N-acetyl-cysteine) do not seem to be effective in the prevention of CSA AKI.13 Lecomte and colleagues14 found that tight glycemic control was associated with a significant reduction in postoperative renal impairment and failure in nondiabetic patients after cardiac surgery. In our study, we specifically investigated independent risk factors for CSA AKI. Correction and/or optimization of those variables identified as independent risk factors for CSA AKI, such as systemic arterial hypertension and cardiopulmonary bypass duration are potential targets to improve patients’ outcomes. Furthermore, patients’ related risk factors, such as peripheral artery disease and older age should lead physicians to extremely strict monitoring of postoperative creatinine values in order to identify the onset of CSA AKI in a timely fashion. It is unknown whether specific cardiopulmonary bypass management such as optimization of bypass flow, perfusion pressure, and oxygen delivery strategies could be able to reduce the incidence of postoperative CSA AKI. Therapeutic strategies aimed at reducing kidney damage and accelerating function recovery could be nonpharmacologic (renal perfusion, avoidance of nephrotoxins), pharmacologic, and dialytic. Timing of renal replacement therapy and its method have not yet been clearly established,15 but there is a trend in considering early initiation of RRT beneficial as compared with late initiation,16 and researchers are increasingly recognizing the importance of fluid overload with high venous pressures playing an important role in the kidney dysfunction associated with cardiorenal syndrome.17
Limitations of this study are mainly related to its retrospective observational nature. We used only the RIFLE creatinine criteria and not the urine output criteria for the assessment of CSA AKI, meaning that there is a possibility for a mild underestimation of its incidence.
In conclusion, RIFLE classification is a reliable tool to classify CSA AKI. CSA AKI is more likely to develop in patients with a higher preoperative risk profile and more complex surgery. CSA AKI negatively affects outcomes in terms of survival and length of ICU and hospital stay. More advanced CSA AKI, in particular, RIFLE class failure, is an independent predictor for hospital mortality. Identification of independent risk factors for CSA AKI could potentially lead to the development of preventive measures and to timely and effective therapies.
Acknowledgments and disclosures: The authors are grateful to Andrew House, MD, and to Alexandra Chronopoulos, MD, for their kind and valuable manuscript revision and editing. Drs D’Onofrio, Bolgan, Cresce, and Fabbri have no financial disclosures for this paper. Dr Ronco serves as a speaker for Abbott, and has received honoraria from Gambro and Inverness. Dr Cruz serves as a speaker and has received honoraria from Biosite and Inverness Medical. Dr D'Onofrio received an honorarium funded by an unrestricted educational grant from Abbott Laboratories and Otsuka America Pharmacueticals for time and expertise spent in preparation of this article. The authors state that they had full control of the design of the study, methods used, outcome parameters, analysis of data, and production of the written report.