Chinese Medical Journal 2013;126(23):4409-4416:10.3760/cma.j.issn.0366-6999.20132065
Prevalence, risk factors, clinical course, and outcome of acute kidney injury in Chinese intensive care units: a prospective cohort study

WEN Ying, JIANG Li, XU Yuan, QIAN Chuan-yun, LI Shu-sheng, QIN Tie-he, CHEN Er-zhen, LIN Jian-dong, AI Yu-hang, WU Da-wei, WANG Yu-shan, SUN Ren-hua, HU Zhen-jie, CAO Xiang-yuan, ZHOU Fa-chun, HE Zhen-yang, ZHOU Li-hua, AN You-zhong, KANG Yan, MA Xiao-chun, YU Xiang-you, ZHAO Ming-yan, XI Xiu-ming, DU Bin and China Critical Care Clinical Trial Group (CCCCTG)

Keywords
acute kidney injury; intensive care units; mortality; prevalence; risk factors
Abstract
Background Acute kidney injury (AKI) has been recognized as a major healthcare problem affecting millions of patients worldwide. However, epidemiologic data concerning AKI in China are still lacking. The objectives of this study were to characterize AKI defined by RIFLE criteria, assess the association with hospital mortality, and evaluate the impact of AKI in the context of other risk factors.
Methods This prospective multicenter observational study enrolled 3,063 consecutive patients from 1 July 2009 to 31 August 2009 in 22 ICUs across mainland China. We excluded patients who were admitted for less than 24 hours (n=1623), younger than 18 years (n=127), receiving chronic hemodialysis (n=29), receiving renal transplantation (n=1) and unknown reasons (n=28). There were 1255 patients in the final analysis. AKI was diagnosed and classified according to RIFLE criteria.
Results There were 396 patients (31.6%) who had AKI, with RIFLE maximum class R, I, and F in 126 (10.0%), 91 (7.3%), and 179 (14.3%) patients, respectively. Renal function deteriorated in 206 patients (16.4%). In comparison with non AKI patients, patients in the risk class on ICU admission were more likely to progress to the injury class (odds ratio (OR) 3.564, 95% confidence interval (CI) 1.706 – 7.443, P = 0.001], while patients in the risk class (OR 5.215, 95% CI 2.798–9.719, P <0.001) and injury class (OR 13.316, 95% CI 7.507–23.622, P <0.001) had a significantly higher probability of deteriorating into failure class. The adjusted hazard ratios for 90-day mortality were 1.884 for the risk group, 3.401 for the injury group, and 5.306 for the failure group.
Conclusions The prevalence of AKI was high among critically ill patients in Chinese ICUs. In comparison with non-AKI patients, patients with RIFLE class R or class I on ICU admission were more susceptibility to progression to class I or class F. The RIFLE criteria were robust and correlated well with clinical deterioration and mortality.
Acute kidney injury (AKI) has been recognized as a major healthcare problem affecting millions of patients worldwide; it is associated with a substantial increase in morbidity and mortality.1-9In 2004, the Acute Dialysis Quality Initiative (ADQI) formulated the risk, injury, failure, loss, and end-stage kidney disease (RIFLE) classification.10RIFLE defines three grades of increasing AKI severity: risk (class R), injury (class I), and failure (class F), and two outcome classes (loss (L) and end-stage kidney disease (E)). AKI occurs in approximately 7% of all hospitalized patients,1,2and between 11% and 67% of critically ill patients.3,9A systematic review identified 24 studies with over 71 000 patients using the RIFLE classification to define AKI.11Compared with non-AKI, the authors reported a stepwise increase in relative risk (RR) for death going from risk (RR 2.40) to injury (RR 4.15) to failure (RR 6.37,P<0.001 for all). In addition to the predictive ability of RIFLE classification, Hoste and colleagues3reported that about 22% of patients already had AKI on intensive care unit (ICU) admission, whereas 67% of all patients had an episode of AKI during their ICU stay. There were few studies described the progression between classification stages and provided information about the natural history and clinical course of AKI.
However, epidemiologic data concerning AKI in China are still lacking. Therefore, we performed a prospective cohort study of all adult hospitalizations in 22 ICUs in China.12The objectives of this study were to characterize AKI defined by RIFLE criteria, assess the association with hospital mortality, and evaluate the impact of AKI in the context of other risk factors.
METHODS
Study design
This was a prospective cohort study covering a 2-month period (July 1, 2009 to August 31, 2009) in 22 ICUs in mainland China. All participating centers were closed ICUs of tertiary teaching hospitals in metropolitan cities of 19 provinces, managed by full-time ICU directors. Nineteen ICUs were in university-affiliated hospitals and the other three were in public hospitals.12
Ethics statement
The study was approved by the institutional review boards of Fuxing Hospital, Capital Medical University and all other participating hospitals. The institutional review board specifically approved the informed consent waiver due to the anonymous and non-interventional nature of the study.
Study population
In the original study, all patients admitted to any of the 22 ICUs were screened for eligibility. We only considered the first admission when patients were readmitted to ICU during the study period. Exclusion criteria included patients younger than 14 years of age and ICU length of stay less than 24 hours. For the purpose of this study, we excluded patients younger than 18 years, those receiving chronic hemodialysis, and those receiving renal transplantation.
Data collection
All researchers were selected and well-trained by each participating ICU. The assigned study coordinator in each participating ICU was responsible for strictly supervised daily patient screening, enrollment, and data collection. Demographic data, comorbidities, and admission status were recorded in an electronic case report form (CRF). Blood chemistry, complications, interventions and treatment were recorded until 28 days after ICU admission, discharge or death, whichever occurred earlier.
The severity of illness, as measured by the acute physiology and chronic health evaluation (APACHE) II score13and sequential organ failure assessment (SOFA) score,14was calculated based on the lowest variables recorded during the first 24 hours of ICU admission. Chronic organ dysfunction was assessed based on the APACHE II score.13Severe sepsis and septic shock were defined according to the American College of Chest Physicians/Society of Critical Care Medicine consensus conference definitions.15Both acute lung injury (ALI) and acute respiratory distress syndrome (ARDS) were defined according to the American-European Consensus Conference criteria.16Chronic kidney insufficiency (i.e. glomerular filtration rate <60 ml·min-1·1.73 m-2) was defined according to K/DOQI clinical practice guidelines for chronic kidney disease.17All patients were followed up until one of the following situations occurred: discharge from the current hospital admission, death during current hospital admission, or 90 days after study entry. Patients that were still hospitalized at November 30, 2009 were defined as survivors.
All completed CRFs were sent to the corresponding author on daily basis by e-mail. These were checked for the integrity of data collection; any queries were sent to the source hospital for resolution.
In the participating ICUs, serum creatinine (SCr) level was measured at least once daily for every patient as part of the clinical routine. In addition, hourly urine output (UO) was checked and recorded by the study coordinator, and categorized according to UO criteria,10based on either actual or estimated body weight. For patients without a reported history of chronic renal insufficiency, the creatinine value was estimated by the Modification of Diet in Renal Disease (MDRD) equation (SCr-MDRD) and customized for the Chinese population.18This value assumed a glomerular filtration rate of 75 ml·min-1·1.73 m-2.3,6
The baseline SCr level was determined as the lowest value among SCr levels on hospital admission and the SCr levels on ICU admission, SCr-MDRD.3Approximately 27.6% of all patients were classified with a baseline SCr-MDRD. For patients with chronic renal insufficiency but not on chronic dialysis, we used the SCr level measured on hospital admission as the baseline figure.
AKI was diagnosed and classified according to either SCr or UO RIFLE criteria.10This comprised criteria that led to the worst possible classification during each ICU stay. The diagnosis of AKI according to either SCr or UO criteria may differ depending on the day. Patients were classified into four mutually exclusive groups according to the worst RIFLE (RIFLEmax) class during ICU stay: non-AKI, risk group, injury group, or failure group.
Statistical analysis
In our study, we performed both population rate estimation and power analysis. Normally distributed or near-normally distributed continuous variables were presented as the mean ± standard deviation (SD). Student’st-test or analysis of variance (ANOVA) with Bonferroni correction was used for comparison. Non-normally distributed continuous variables were presented as median with interquartile range (IQR), and compared by Mann-WhitneyU-test, or Kruskal-Wallis ANOVA test with Bonferroni correction. Categorical variables were reported as either a percentage of the group from which they were derived, and then compared using either the chi-square test or Fisher’s exact test when appropriate.
We performed multivariable logistic regression to assess the independent predictors for AKI development. Different variables including demographics, comorbidities, severity of illness, admission status and complications, were added into the model using stepwise conditional backward entry, (i.e.P<0.1 in univariate analysis). Collinearity was then analyzed by assessing correlation between covariates.
The Sequential Organ Failure Assessment score without points for renal insufficiency (SOFA nonrenal) was chosen as a covariate to control for multicollinearity between the RIFLE classification and scoring systems that included points for SCr levels. The model fit was assessed by the Hosmer-Lemeshow goodness-of-fit test. Kaplan-Meier survival analysis was used to compare 90-day mortality. The log-rank statistic tested differences between groups. Patients alive at hospital discharge were censored. The study performed a Cox proportional hazards regression analysis in order to examine whether the RIFLE max class alone, or the RIFLE max class based on SCr or UO criteria were associated with mortality adjustment for baseline severity of illness, gender and age.
Analysis was performed using the statistical software package SPSS 17.0.1 (SPSS Inc., Chicago, IL, USA). A double-sidedPvalue <0.05 was considered statistically significant.
RESULTS
Among the 3063 patients included in this study, we excluded patients who were admitted for less than 24 hours (n=1623), younger than 18 years (n=127), receiving chronic hemodialysis (n=29), receiving renal transplantation (n=1) and unknown reasons (n=28). This left 1255 patients for the final analysis. Baseline characteristics of patients according to RIFLEmax class were presented in Table 1.
Clinical course and risk factor for the development of AKI
There were 396 patients (31.6%) who fulfilled RIFLE criteria for AKI. In this group, 126 patients (10.0%) were classified as the risk group, 91 patients (7.3%) as the injury group, and 179 patients (14.3%) as the failure group; 859 patients (68.4%) had no evidence of AKI according to the RIFLE classification.
In comparison with patients without AKI, patients with any degree of AKI were older and more severely ill, as suggested by higher APACHE II, SOFA, and SOFA nonrenal scores. Patients with AKI were more likely to have comorbidities such as coronary heart disease, hypertension, diabetes, COPD, hematological malignancy, autoimmune disease, and chronic kidney insufficiency (Table 1). Patients with AKI were more likely to be admitted to the hospital due to medical reasons.
The development of AKI during ICU stay as correlated to the RIFLEmax class was shown in Figure 1. On ICU admissions, 1013 out of 1255 patients (80.7%) did not have AKI, according to the RIFLE criteria. In comparison, 82 patients (6.5%) were classified as a risk class, 68 patients (5.4%) as an injury class, and 92 patients (7.3%) as a failure class. Among the 1013 patients with non-AKI on ICU admission, 154 (15.2%) developed AKI during their ICU stay, including 71 patients (7.0%) in the risk group, 38 patients (3.8%) in the injury group, and 45 patients (4.4%) in the failure group. In comparison, patients in the risk class on ICU admission were more likely to progress to the injury group (26/82 (31.7%) vs. 38/1013 (3.8%),OR3.564, 95% confidence interval (CI) 1.706 to 7.443,P= 0.001), while patients in the risk class (19.5% vs. 4.5%,OR5.215, 95%CI2.798 to 9.719,P<0.001) and the injury class (38.2% vs. 4.5%,OR13.316, 95%CI7.507 to 23.622,P<0.001) had a significantly higher probability to deteriorate into the failure group (Figures 1 and 2). There were 187 patients (14.9%) who met the RIFLE criteria for the risk class, 29 (15.5%) and 32 (17.1%) whom ultimately progressed to both injury and failure classes, respectively. In addition, there were 147 patients (11.7%) who ever met the criteria for the injury class; 58 (39.5%) of those ultimately progressed to the failure class.
According to logistic regression analysis, increasing age (OR1.015, 95%CI1.006 to 1.023,P<0.001), higher SOFA nonrenal score (OR1.130, 95%CI1.089 to 1.172,P<0.001), autoimmune disease (OR2.310, 95%CI1.076 to 4.959,P=0.032), presence of severe sepsis/septic shock (OR2.038, 95%CI1.557 to 2.667,P<0.001), and ICU admission due to renal disorders (OR8.839, 95%CI4.520 to 17.284,P<0.001) were independent risk factors for development of AKI. In comparison, ICU admission due to trauma (OR0.571, 95%CI0.337 to 0.969,P=0.038) was independently associated with less probability of AKI (Table 2).
Clinical outcome and risk factors for mortality
During the ICU stay, AKI patients were more likely to receive intensive treatment, including renal replacement therapy (26.5% vs. 4.8%,P< 0.001), mechanical ventilation (80.3% vs. 70.3%,P< 0.001), vasoactive agents (59.3% vs. 24.4%,P< 0.001), and packed red blood cell transfusion (57.1% vs. 41.3%,P< 0.001). Invasive monitoring measures (such as arterial line, central venous catheterization, and cardiac output monitoring) were more commonly used in AKI patients (Table 3).
Both overall ICU mortality and 90-day mortality were significantly higher in patients with AKI (35.9% vs. 6.4%,P< 0.001; and 41.9% vs. 9.4%,P< 0.001, respectively). Increasing AKI severity was associated with significantly higher 90-day mortality, 24.6% in the RIFLEmax risk group, 39.6% in the injury group, and 55.3% in the RIFLEmax failure group (Table 3, Figure 3). The adjusted 90-day mortality hazard ratios were 1.884 (95%CI1.234 to 2.877,P= 0.003) for the RIFLEmax risk group, 3.401 (95%CI2.257 to 5.124,P< 0.001) for the injury group, and 5.306 (95%CI3.865 to 7.286,P< 0.001) for the failure group. Other independent risk factors for 90-day mortality included increasing age, greater SOFA nonrenal score, immunosuppression (as defined by chronic organ dysfunction in APACHE II score), ICU admissions due to neurological disorders, the presence of severe sepsis/septic shock and ALI/ARDS. These were analyzed according to Cox proportional hazards regression analysis (Table 4).
Value of serum creatinine criteria and urine output
criteria
The RIFLEmax class was based on SCr criteria only in 232 patients (58.6%) and UO criteria in 77 patients (19.4%). Concordance between the SCr and UO criteria was found in 87 cases (22.0%) (Table 1 and Table 5). In addition, both SCr and UO criteria were identified as independent risk factors for 90-day mortality, according to Cox regression analysis.
DISCUSSION
To our knowledge, this is the first prospective, multicenter cohort study measuring AKI in adult ICUs in mainland China. The goal of our study was to investigate the AKI in critically ill patients.
Prevalence of AKI in the ICU has varied significantly across different studies, ranging from between 11% to 67%.3-9
Such differences may be explained by study design, the population studied, differentiation of SCr versus UO criteria, and determination of baseline creatinine level. For example, Hoste and colleagues reported the highest AKI prevalence (i.e. 67%) among 5383 ICU admissions in a retrospective, single center cohort study.3Bagshaw and colleagues, in a retrospective analysis of prospectively collected data, reported a 36.1% AKI prevalence among 120 123 patients in the Australian New Zealand Intensive Care Society Adult Patient Database.6Similarly, when using the Simplified Acute Physiology Score 3 database, Joannidis found 35.5% of patients developed AKI.8Both results were comparable to our study. However, they only defined cases with early AKI, due to laboratory results seen beyond the first 24 or 48 hours of ICU admission. Moreover, some investigators used only SCr or glomerular filtration rate criteria to determine the RIFLE category, due to the lack of 6- or 12-hour UO data.7In comparison, in a 3-month Italian prospective multicenter observational study, only 10.8% (95%CI9.5 to 12.1%) of patients developed AKI.5Another 3-month prospective multicenter study in Greece, using the same study design, observed that 170 patients developed AKI, which corresponding to a prevalence of 16%.19Nevertheless, we believed that the above results were actually comparable to our study. We excluded patients with an ICU length of stay of less than 24 hours, while they were included in the Italian and Greek studies.
Due to the complexity of AKI pathogenesis, determination of exact mechanism(s) appeared rather difficult, even in prospective cohort studies.2,20Previous studies demonstrated severe sepsis/septic shock caused or contributed to AKI in 28% to 47.5% of patients.21-24Meanwhile, a cross-sectional one-day prevalence study in German ICUs reported that 166 (41.4%) of 401 patients with severe sepsis/septic shock had AKI.25Our result supported the previous findings. Among the 396 AKI patients, 212 (53.5%) had severe sepsis/septic shock, while 45.5% of patients with severe sepsis/septic shock developed AKI (Table 1). Logistic regression analysis also suggested severe sepsis/septic shock was an independent risk factor for AKI development, with anOR2.038 (P<0.001). These results reinforced the urgency of understanding the sepsis-related AKI mechanism, which might be crucial to clinical outcome Improvement.
Our data clearly demonstrated that patients with increasing renal dysfunction severity were also at higher risk of further deterioration. In comparison with non-AKI patients, those in the risk class on ICU admission were more likely to progress to either the injury (OR3.564) or the failure groups (OR5.215). Meanwhile, patients in the injury class on ICU admission were associated with greater than 13 times odds of deteriorating into the failure group. Few studies ever described AKI progression in ICU.3,9,26Similar to our study, Hoste and colleagues also observed increased risks of renal function deterioration associated with an increasing RIFLE class severity.3They found that more than 50% of patients with RIFLE class R progressed to class I or F, and more than one-third of patients with RIFLE class I progressed to class F. Therefore, how to prevent the renal function from deteriorating is crucially important.
The RIFLE classification was originally developed to standardize AKI definition and severity, rather than to predict mortality. Nevertheless, such a system might be anticipated to have better predictive validity. Consistent with existing literature,3-9,11we also found worsening RIFLE class also correlated with increasing linear trends in 90-day mortality. This persisted in multivariate analysis after adjustment for several covariates, including age, non-renal severity of illness, comorbidities, admission diagnosis, severe sepsis/septic shock, and ARDS. All the above variables confirmed the fact that RIFLE criteria provided a well-balanced and robust classification system for differentiating patients with different AKI severity, at least related to the risk of clinical deterioration, mortality, and the need for renal replacement therapy.
Recently, diagnostic value of UO criteria in the AKI setting has been challenged. Some investigators reported that UO criteria were valid to diagnose worse RIFLE classes in only 7.2% to 13% of cases.5,6,27In a case-control study, Solomon and colleagues observed that in a general ICU at a tertiary referral hospital, doctors were more likely oliguric than their patients (OR1.99, 95%CI1.08 to 3.68,P=0.03).28However, our data suggested that UO criteria presented diagnostic value in almost one-fifth of AKI cases, while worsening degrees of oliguria was independently associated with a stepwise hazard ratio increase for 90-day mortality, even more robust than SCr criteria. In addition, Andrikos and colleagues found that UO was independently associated with disease outcome (hazard ratio 0.99, 95%CI0.99 to 0.99,P= 0.009).19The most recent evidence suggested that UO criteria might increase the diagnostic sensitivity (from 24% to 52%),29identify higher percentages of AKI patients than SCr criteria (32% vs. 28%),30avoid underestimating AKI incidence and grade, as well as delayed diagnosis.31These findings support the importance of UO criteria in early AKI diagnosis and management.32
Our study was subject to several limitations. First, we did not compare RIFLE criteria with Acute Kidney Injury Network (AKIN) criteria.33However, Lopes and coworkers34concluded that, although AKIN criteria could improve AKI diagnosis sensitivity, it appeared not to improve on the RIFLE criteria ability in predicting in-hospital mortality of critically ill patients. In comparison Joannidis even observed greater robustness and higher detection AKI rates during the first 48 hours of ICU admission with RIFLE criteria.8Second, a lack of consensus in the determination of baseline creatinine levels might affect AKI classification and prognosis of AKI.35,36Within each RIFLE class, reported misclassification rates were as high as between 9.2% and 24.8%,35while the mortality rate varied between 76 and 13.9%,35,36depending which baseline creatinine level was used. In our study, extensive efforts determined the baseline creatinine level; these were the lowest values among SCr levels on hospital admission, SCr levels on ICU admission, or SCr-MDRD. Third, we did not investigate the exact AKI pathogenic factors, which appeared multifactorial and difficult to define.2,20Fourth, both the relatively small sample size in this study in comparison with studies conducted by countries with smaller population and data variation among centers are weaknesses of our study. Finally, similar to most medical literature, we did not evaluate outcome classes, such as loss and end-stage kidney disease.
In conclusion, in this prospective multicenter cohort study, we reported that AKI defined by RIFLE criteria had a high prevalence in Chinese ICUs. In comparison with non-AKI patients, patients with RIFLE class R or class I on ICU admission were more susceptibility to progression to class I or class F. The RIFLE criteria were robust and correlated well with both clinical deterioration and mortality.
REFERENCES
  • 1.

    Chertow GM, Soroko SH, Paganini EP, Cho KC, Himmelfarb J, Ikizler TA, et al. Mortality after acute renal failure: Models for prognostic stratification and risk adjustment. Kidney Int 2006; 70: 1120-1126.

  • 2.

    Nash K, Hafeez A, Hou S. Hospital-acquired renal insufficiency. Am J Kidney Dis 2002; 39: 930-936.

  • 3.

    Hoste EA, Clermont G, Kersten A, Venkataraman R, Angus DC, De Bacquer D, et al. RIFLE criteria for acute kidney injury are associated with hospital mortality in critically ill patients: a cohort analysis. Crit Care 2006; 10: R73.

  • 4.

    Uchino S, Bellomo R, Goldsmith D, Bates S, Ronco C. An assessment of the RIFLE criteria for acute renal failure in hospitalized patients. Crit Care Med 2006; 34: 1913-1917.

  • 5.

    Cruz DN, Bolgan I, Perazella MA, Bonello M, de Cal M, Corradi V, et al. North East Italian prospective hospital renal outcome survey on acute kidney injury (NEiPHROS-AKI): targeting the problem with the RIFLE criteria. Clin J Am Soc Nephrol 2007; 2: 418-425.

  • 6.

    Bagshaw SM, George C, Dinu I, Bellomo R. A multi-centre evaluation of the RIFLE criteria for early acute kidney injury in critically ill patients. Nephrol Dial Transplant 2008; 23: 1203-1210.

  • 7.

    Ostermann M, Chang RW. Acute kidney injury in the intensive care unit according to RIFLE. Crit Care Med 2007; 35: 1837-1843.

  • 8.

    Joannidis M, Metnitiz B, Bauer P, Schusterschitz N, Moreno R, Druml W, et al. Acute kidney injury in critically ill patients classified by AKIN versus RIFLE using the SAPS 3 database. Intensive Care Med 2009; 35: 1692-1702.

  • 9.

    Piccinni P, Cruz DN, Gramaticopolo S, Garzotto F, Dal Santo M, Aneloni G, et al. Prospective multicenter study on epidemiology of acute kidney injury in the ICU: a critical care nephrology Italian collaborative effort (NEFROINT). Minerva Anestesiol 2011; 77: 1072-1083.

  • 10.

    Bellomo R, Ronco C, Kellum JA, Mehta RL, Palevsky P; Acute Dialysis Quality Initiative workgroup. Acute renal failure — definition, outcome measures, animal models, fluid therapy and information technology needs: the Second International Consensus Conference of the Acute Dialysis Quality Initiative (ADQI) Group. Crit Care 2004; 8: R204-R212.

  • 11.

    Ricci Z, Cruz D, Ronco C. The RIFLE criteria and mortality in acute kidney injury: a systematic review. Kidney Int 2008; 73: 538-546.

  • 12.

    Du B, An Y, Kang Y, Yu X, Zhao M, Ma X, et al. Characteristics of critically ill patients in intensive care units in mainland China. Crit Care Med 2013; 41: 84-92.

  • 13.

    Knaus WA, Draper EA, Wagner DP, Zimmerman JE. APACHE II: A severity of disease classification system. Crit Care Med 1985; 13: 818-829.

  • 14.

    Vincent JL, Moreno R, Takala J, Willatts S, De Mendonca A, Bruining H, et al. The SOFA (Sepsis-related Organ Failure Assessment) score to describe organ dysfunction/failure. On behalf of the Working Group on Sepsis-Related Problems of the European Society of Intensive Care Medicine. Intensive Care Med 1996; 22: 707-710.

  • 15.

    Bone RC, Balk RA, Cerra FB, Dellinger RP, Fein AM, Knaus WA, et al. Definitions for sepsis and organ failure and guidelines for the use of innovative therapies in sepsis. The ACCP/SCCM Consensus Conference Committee. American College of Chest Physicians/Society of Critical Care Medicine. Chest 1992; 101: 1644-1655.

  • 16.

    Bernard GR, Artigas A, Brigham KL, Carlet J, Falke K, Hudson L, et al. The American–European Consensus Conference on ARDS. Definitions, mechanisms, relevant outcomes, and clinical trial coordination. Am J Respir Crit Care Med 1994; 149: 818-824.

  • 17.

    National Kidney Foundation. K/DOQI clinical practice guidelines for chronic kidney disease: evaluation, classification, and stratification. Am J Kidney Dis 2002; 39: S1-S266.

  • 18.

    Ma YC, Zuo L, Chen JH, Luo Q, Yu XQ, Li Y, et al. Modified glomerular filtration rate estimating equation for Chinese patients with chronic kidney disease. J Am Soc Nephrol 2006; 17: 2937-2944.

  • 19.

    Andrikos E, Tseke P, Balafa O, Cruz DN, Tsinta A, Androulaki M, et al. Epidemiology of acute renal failure in ICUs: a multi-center prospective study. Blood Purif 2009; 28: 239-244.

  • 20.

    Kohli HS, Bhaskaran MC, Muthukumar T, Thennarasu K, Sud K, Jha V, et al. Treatment-related acute renal failure in the elderly: a hospital-based prospective study. Nephrol Dial Transplant 2000; 15: 212-217.

  • 21.

    Uchino S, Kellum JA, Bellomo R, Doig GS, Morimatsu H, Morgera S, et al. Acute renal failure in critically ill patients: a multinational, multicenter study. JAMA 2005; 294: 813-818.

  • 22.

    Bagshaw SM, Uchino S, Bellomo R , Morimatsu H, Morgera S, Schetz M, et al. Septic acute kidney injury in critically ill patients: clinical characteristics and outcomes. Clin J Am Soc Nephrol 2007; 2: 431-439.

  • 23.

    Neveu H, Kleinknecht D, Brivet F, Loirat P, Landais P. Prognostic factors in acute renal failure due to sepsis: results of a prospective multicenter study. The French Study Group on Acute Renal Failure. Nephrol Dial Transplant 1996; 11: 293-299.

  • 24.

    Mehta RL, Bouchard J, Soroko SB, Ikizler TA, Paganini EP, Chertow GM, et al. Sepsis as a cause and consequence of acute kidney injury: Program to Improve Care in Acute Renal Disease. Intensive Care Med 2011; 37: 241-248.

  • 25.

    Oppert M, Engel C, Brunkhorst FM, Bogatsch H, Reinhart K, Frei U, et al. Acute renal failure in patients with severe sepsis and septic shock – a significant independent risk factor for motatlity: results from the German Prevalence Study. Nephrol Dial Transplant 2008; 23: 904-909.

  • 26.

    Herget-Rosenthal S, Marggraf G, Husing J, Goring F, Pietruck F, Janssen O, et al. Early detection of acute renal failure by serum cystatin C. Kidney Int 2004; 66: 1115-1122.

  • 27.

    Bagshaw SM, George C, Gibney RT, Bellomo R. A multi-center evaluation of early acute kidney injury in critically ill trauma patients. Ren Fail 2008; 30: 581-589.

  • 28.

    Solomon AW, Kirwan CJ, Alexander ND, Nimako K, Jurukov A, Forth RJ, et al. Urine output on an intensive care unit: case-control study. BMJ 2010; 341: c6761.

  • 29.

    Macedo E, Malhotra R, Bouchard J, Wynn SK, Mehta RL. Oliguria is an early predictor of higher mortality in critically ill patients. Kidney Int 2011; 80: 760-767.

  • 30.

    Macedo E, Malhotra R, Claure-Del Granado R, Fedullo P, Mehta RL. Defining urine output criterion for acute kidney injury in critically ill patients. Nephrol Dial Transplant 2011; 26: 509-515.

  • 31.

    Wlodzimirow KA, Abu-Hanna A, Slabbekoorn M, Chamuleau RA, Schultz MJ, Bouman CS. A comparison of RIFLE with and without urine output criteria for acute kidney injury in critically ill patients. Crit Care 2012; 16: R200.

  • 32.

    Cerda J. Oliguria: an earlier and accurate biomarker of acute kidney injury? Kidney Int 2011; 80: 699-701.

  • 33.

    Mehta RL, Kellum JA, Shah SV, Molitoris BA, Ronco C, Warnock DG, et al. Acute kidney injury network: report of an initiative to improve outcomes in acute kidney injury. Crit Care 2007; 11: R31.

  • 34.

    Lopes JA, Fernandes P, Jorge S, Gonçalves S, Alvarez A, Costa e Silva Z, et al. Acute kidney injury in intensive care unit patients: a comparison between the RIFLE and the Acute Kidney Injury Network classifications. Crit Care 2008; 12: R110.

  • 35.

    Siew ED, Matheny ME, Ikizler TA, Lewis JB, Miller RA, Waitman LR, et al. Commonly used surrogates for baseline renal function affect the classification and prognosis of acute kidney injury. Kidney Int 2010; 77: 536-542.

  • 36.

    Cruz DN, Ricci Z, Ronco C. Clinical review: RIFLE and AKIN – time for reappraisal. Crit Care 2009; 13: 211.

(Received August 4, 2013)
Edited by SUN Jing
Table 1.Patient characteristics according to maximum RIFLE**class
Parameters
All patients
(n= 1255)
Non-AKI*
(n= 859)
Any AKI*
(n= 396)
Pvalues
Maximum RIFLE**Classification
Risk
(n= 126)
Injury
(n= 91)
Failure
(n= 179)
Pvalues
Male sex (n(%)) 798 (63.6) 541 (63.0) 257 (64.9) 0.512 81 (64.3) 60 (65.9) 116 (64.8) 0.968
Age (median (IQR¶)) 62 (44–74) 59 (41–73) 68 (53–78) < 0.001 67 (51–77) 69 (56–79) 68 (46–78) 0.911
APACHE II‡score (median (IQR¶)) 17 (12–23) 15 (10–20) 22 (17–28) < 0.001 18 (14–24) 21 (17–28) 24 (19–31) < 0.001
SOFA‡‡score (median (IQR¶)) 6 (3–9) 5 (3–8) 8 (5–11) < 0.001 7 (4–9) 8 (4–11) 9 (7–12) < 0.001
SOFA nonrenal§§score (median (IQR¶)) 5 (3–8) 5 (3–7) 7 (4–10) < 0.001 6 (4–8) 7 (4–9) 7 (5–10) 0.153
Comorbidities (n(%)) 721 (57.5) 451 (52.5) 270 (68.2) < 0.001 84 (66.7) 64 (70.3) 122 (68.2) 0.849
Coronary heart disease 225 (17.9) 141 (16.4) 84 (21.2) 0.039 25 (19.8) 22 (24.2) 37 (20.7) 0.722
Hypertension 407 (32.4) 251 (29.2) 156 (39.4) < 0.001 47 (37.3) 39 (42.9) 70 (39.1) 0.707
Diabetes 191 (15.2) 111 (12.9) 80 (20.2) 0.001 25 (19.8) 22 (24.2) 33 (18.4) 0.536
COPD§ 126 (10.0) 74 (8.6) 52 (13.1) 0.013 14 (11.1) 15 (16.5) 23 (12.8) 0.507
Solid tumor 144 (11.5) 92 (10.7) 52 (13.1) 0.211 15 (11.9) 17 (18.7) 20 (11.2) 0.199
Hematological malignancy 18 (1.4) 8 (0.9) 10 (2.5) 0.027 2 (1.6) 1 (1.1) 7 (3.9) 0.273
Autoimmune disease 31 (2.5) 15 (1.7) 16 (4.0) 0.015 4 (3.2) 3 (3.3) 9 (5.0) 0.662
Chronic kidney insufficiency 54 (4.3) 24 (2.8) 30 (7.6) < 0.001 6 (4.7) 3 (3.3) 21 (11.7) 0.016
Chronic organ dysfunction (n(%)) 261 (20.8) 155 (18.0) 106 (26.8) < 0.001 24 (19.0) 28 (30.8) 54 (30.2) 0.060
Heart 95 (7.6) 49 (5.7) 46 (11.6) < 0.001 8 (6.3) 14 (15.4) 24 (13.4) 0.073
Respiratory 114 (9.1) 70 (8.1) 44 (11.1) 0.090 13 (10.3) 11 (12.1) 20 (11.2) 0.919
Liver 22 (1.8) 12 (1.4) 10 (2.5) 0.157 2 (1.6) 3 (3.3) 5 (2.8) 0.697
Immune 75 (6.0) 43 (5.0) 32 (8.1) 0.033 6 (4.8) 6 (6.6) 20 (11.2) 0.108
Admission status (n(%))                
Medical 836 (66.6) 535 (62.3) 301 (76.0) < 0.001 84 (66.7) 68 (74.7) 149 (83.2) 0.004
Elective surgery 263 (21.0) 210 (24.4) 53 (13.4) < 0.001 26 (20.6) 13 (14.3) 14 (7.8) 0.005
Emergency surgery 152 (12.1) 111 (12.9) 41 (10.4) 0.195 16 (12.7) 10 (11.0) 15 (8.4) 0.464
Reasons for ICU||admission (n(%))                
Respiratory 423 (33.7) 273 (31.8) 150 (37.9) 0.034 46 (36.5) 40 (44.0) 64 (35.8) 0.392
Cardiovascular 160 (12.7) 101 (11.8) 59 (14.9) 0.121 16 (12.7) 13 (14.3) 30 (16.8) 0.607
Trauma 129 (10.3) 109 (12.7) 20 (5.1) < 0.001 2 (1.6) 4 (4.4) 14 (7.8) 0.047
Neurological 188 (15.0) 139 (16.2) 49 (12.4) 0.079 24 (19.0) 10 (11.0) 15 (8.4) 0.019
Gastrointestinal 238 (19.0) 173 (20.1) 65 (16.4) 0.118 30 (23.8) 14 (15.4) 21 (11.7) 0.019
Renal 53 (4.2) 13 (1.5) 40 (10.1) < 0.001 1 (0.8) 6 (6.6) 33 (18.4) < 0.001
Other 64 (5.1) 51 (5.9) 13 (3.3) 0.047 7 (5.6) 4 (4.4) 2 (1.1) 0.080
Complications on admission (n(%))                
Infection 592 (47.2) 360 (41.9) 232 (58.6) < 0.001 60 (47.6) 55 (60.4) 117 (65.4) 0.008
Severe sepsis/septic shock 407 (32.4) 229 (26.7) 178 (44.9) < 0.001 41 (32.5) 41 (45.1) 96 (53.6) 0.001
Complications during ICU||stay (n(%))                
Severe sepsis /septic shock 466 (37.1) 254 (29.6) 212 (53.5) < 0.001 52 (41.3) 48 (52.7) 112 (62.6) 0.001
ALI/ARDS† 335 (26.7) 172 (20.0) 163 (41.2) < 0.001 44 (34.9) 34 (37.4) 85 (47.5) 0.063
ICU||-acquired infection 145 (11.6) 98 (11.4) 47 (11.9) 0.813 19 (15.1) 11 (12.1) 17 (9.5) 0.331
Time to RIFLE**max, median (IQR¶)     1 (0–3)   1 (0–2) 1 (0–4) 0 (0–3) 0.073
Determination of RIFLE**class (n(%))                
SCr††criteria only     232 (58.6)   95 (75.4) 67 (73.6) 70 (39.1) <0.001
UO||||criteria only     77 (19.4)   17 (13.5) 15 (16.5) 45 (25.1) <0.001
Both criteria     87 (22.0)   14 (11.1) 9 (9.9) 64 (35.8) <0.001
*AKI: acute kidney injury;†ALI/ARDS: acute lung injury/acute respiratory distress syndrome;‡APACHE II: Acute Physiology and Chronic Health Evaluation, version II;§COPD: chronic obstructive pulmonary disease;||ICU: intensive care unit;¶IQR: interquartile range;**RIFLE: risk, injury, failure, loss and end-stage kidney class;††SCr: serum creatinine level;‡‡SOFA: Sequential Organ Failure Assessment score;§§SOFA nonrenal, sequential organ failure assessment score without points for renal insufficiency;||||UO: urine output.
Table 2.Independent risk factors for development of acute kidney injury by multivariate logistic regression analysis
Variables OR† 95%CI* Pvalues
Age 1.015 1.006 to 1.023 < 0.001
SOFA nonrenal‡ 1.130 1.089 to 1.172 < 0.001
Comorbidity      
Autoimmune disease 2.310 1.076 to 4.959 0.032
Reasons for admission      
Renal 8.839 4.520 to 17.284 < 0.001
Trauma 0.571 0.337 to 0.969 0.038
Complication      
Severe sepsis/septic shock 2.038 1.557 to 2.667 < 0.001
*CI: confidence interval;†OR: odds ratio;‡SOFA nonrenal: Sequential organ failure assessment score without points for renal insufficiency.§Goodness of fit of the multivariable regression model was tested by the Hosmer-Lemeshow statistic:P=0.431 for the model with acute kidney injury as the endpoint.

view in a new window
Figure 1. Clinical course of acute kidney injury according to the RIFLE criteria. Data were expressed as patient numbers who were identified at each level. “Ever risk” and “ever failure” represent the patients who could be identified at this stage. AKI: acute kidney injury; ICU: intensive care unit; RIFLE: risk, injury, failure, loss, and end-stage kidney disease.
Figure 2. Progression to RIFLEmax class by severity of RIFLE on ICU admission.
Figure 3. Kaplan-Meier curve for survival by RIFLE class; Log-rank statistic P < 0.0001.
Table 3.Treatment and clinical outcome of patients according to the maximum RIFLE**class
Parameters
All patients
(n=1255)
Non-AKI*
(n=859)
Any AKI*
(n=396)
Pvalues
Maximum RIFLE**classification
Risk
(n=126)
Injury
(n=91)
Failure
(n=179)
Pvalues
Treatment during ICU‡stay                
Renal replacement therapy 146 (11.6) 41 (4.8) 105 (26.5) <0.001 5 (4.0) 8 (8.8) 92 (51.4) < 0.001
CRRT† 130 (10.4) 32 (3.7) 98 (24.7) <0.001 4 (3.2) 7 (7.7) 87 (48.6) < 0.001
IHD§ 16 (1.3) 9 (1.0) 7 (1.8) 0.002 1 (0.8) 1 (1.1) 5 (2.8) 0.004
Transfusion 581 (46.3) 355 (41.3) 226 (57.1) <0.001 65 (51.6) 44 (48.4) 117 (65.4) 0.009
Mechanical ventilation 922 (73.5) 604 (70.3) 318 (80.3) <0.001 97 (77.0) 71 (78.0) 150 (83.8) 0.278
Noninvasive ventilation 150 (12.0) 100 (11.6) 50 (12.6) 0.617 14 (11.1) 13 (14.3) 23 (12.8) 0.780
Invasive ventilation 864 (68.8) 562 (65.4) 302 (76.3) <0.001 93 (73.8) 65 (71.4) 144 (80.4) 0.190
Vasopressor 445 (35.5) 210 (24.4) 235 (59.3) <0.001 63 (50.0) 61 (56.0) 121 (67.6) 0.007
Inotrope 201 (16.0) 79 (9.2) 122 (30.8) <0.001 34 (27.0) 28 (30.8) 60 (33.5) 0.812
Antiarrhythmias 171 (13.6) 79 (9.2) 92 (23.2) <0.001 28 (22.2) 21 (23.1) 43 (24.0) 0.934
Vasodilator 278 (22.2) 172 (20.0) 106 (26.8) 0.007 41 (32.5) 18 (19.8) 47 (26.3) 0.109
Arterial line 349 (27.8) 210 (24.4) 139 (35.1) <0.001 42 (33.3) 28 (30.8) 69 (38.5) 0.395
Central venous catheter 891 (71.0) 577 (67.2) 314 (79.3) <0.001 101 (80.2) 65 (71.4) 148 (82.7) 0.094
Cardiac output monitoring 45 (3.6) 15 (1.7) 30 (7.6) <0.001 3 (2.4) 7 (7.7) 20 (11.2) 0.017
Clinical outcome                
ICU‡mortality (n(%)) 197 (15.7) 55 (6.4) 142 (35.9) <0.001 22 (17.5) 28 (30.8) 92 (51.4) < 0.001
90-day mortality (n(%)) 247 (19.7) 81 (9.4) 166 (41.9) <0.001 31 (24.6) 36 (39.6) 99 (55.3) < 0.001
ICU‡LOS¶(days) (median (IQR||)) 5 (3–10) 4 (3–8) 7 (3–16) <0.001 7 (3–18) 6 (3–15) 7 (3–15) 0.796
Hospital LOS¶(days) (median (IQR||)) 21 (12–39) 22 (12–38) 20 (9–42) 0.346 26 (12–46) 19 (10–41) 16 (8–38) 0.013
*AKI: acute kidney injury;†CRRT: continuous renal replacement therapy;‡ICU: intensive care unit;§IHD: intermittent hemodialysis;||IQR: interquartile range;¶LOS: length of stay;**RIFLE: risk, injury, failure, loss and end-stage kidney class
Table 4.Independent risk factor for 90-day mortality by Cox hazard regression analysis
Variables HR§ 95%CI‡ Pvalues
Age 1.009 1.002 to 1.017 0.017
SOFA nonrenal** 1.074 1.036 to 1.112 < 0.001
Chronic organ dysfunction      
Immunosuppression 1.574 1.058 to 2.342 0..025
Reasons for ICU||admission      
Neurological 2.103 1.457 to 3.034 < 0.001
Complication during ICU||stay      
Severe sepsis/septic shock 1.393 1.038 to 1.870 0.027
ALI/ARDS† 1.718 1.295 to 2.280 < 0.001
RIFLEmax¶class      
No AKI* 1.000
Risk 1.884 1.234 to 2.877 0.003
Injury 3.401 2.257 to 5.124 < 0.001
Failure 5.306 3.865 to 7.286 < 0.001
*AKI: acute kidney injury;†ALI/ARDS: acute lung injury/acute respiratory distress syndrome;‡CI: confidence interval;§HR: hazard ratio;||ICU: intensive care unit;¶RIFLEmax class: the worst risk, injury, failure, loss and end-stage kidney class;**SOFA nonrenal: Sequential Organ Failure Assessment score without points for renal insufficiency.
Table 5.RIFLE†classification by serum creatinine criteria and urine output criteria
Serum creatinine criteria No AKI* Risk Injury Failure Total
No AKI* 859 95 57 42 1053
Risk 17 14 10 14 55
Injury 11 4 9 14 38
Failure 17 9 19 64 109
Total 904 122 95 134 1255
*AKI: acute kidney injury;†RIFLE: risk, injury, failure, loss and end-stage kidney class.
  1. a grant from the Beijing Municipal Science & Technology Commission (BSTC) (No. 101100050010058).