Portal de Periódicos da CAPES

Artigo Acesso aberto Produção Nacional Revisado por pares

BIBTEX RIS

Oliguria is an early predictor of higher mortality in critically ill patients

2011; Elsevier BV; Volume: 80; Issue: 7 Linguagem: Inglês

10.1038/ki.2011.150

1523-1755

Etienne Macedo, Rakesh Malhotra, Josée Bouchard, Susan Wynn, Ravindra L. Mehta,

Trauma, Hemostasis, Coagulopathy, Resuscitation

Oliguria is a valuable marker of kidney function and a criterion for diagnosing and staging acute kidney injury (AKI). However, the utility of urine output as a specific metric for renal dysfunction is somewhat controversial. To study this issue further we tested whether urine output is a sensitive, specific, and early measure for diagnosing and staging AKI in 317 critically ill patients in a prospective observational study. Urine output was assessed every hour and serum creatinine every 12 to 24 h. The sensitivity and specificity of different definitions of oliguria for the diagnosis of AKI were compared with the Acute Kidney Injury Network serum creatinine criterion. The incidence of AKI increased from 24%, based solely on serum creatinine, to 52% by adding the urine output as a diagnostic criterion. Oliguric patients without a change in serum creatinine had an intensive care unit mortality rate (8.8%) significantly higher than patients without AKI (1.3%), and similar to oliguric patients with an increase in serum creatinine (10.4%). The diagnosis of AKI occurred earlier in oliguric than in non-oliguric patients. Oliguria of more than 12 h and oliguria of 3 or more episodes were associated with an increased mortality rate. Thus, urine output is a sensitive and early marker for AKI and is associated with adverse outcomes in intensive care unit patients. Oliguria is a valuable marker of kidney function and a criterion for diagnosing and staging acute kidney injury (AKI). However, the utility of urine output as a specific metric for renal dysfunction is somewhat controversial. To study this issue further we tested whether urine output is a sensitive, specific, and early measure for diagnosing and staging AKI in 317 critically ill patients in a prospective observational study. Urine output was assessed every hour and serum creatinine every 12 to 24 h. The sensitivity and specificity of different definitions of oliguria for the diagnosis of AKI were compared with the Acute Kidney Injury Network serum creatinine criterion. The incidence of AKI increased from 24%, based solely on serum creatinine, to 52% by adding the urine output as a diagnostic criterion. Oliguric patients without a change in serum creatinine had an intensive care unit mortality rate (8.8%) significantly higher than patients without AKI (1.3%), and similar to oliguric patients with an increase in serum creatinine (10.4%). The diagnosis of AKI occurred earlier in oliguric than in non-oliguric patients. Oliguria of more than 12 h and oliguria of 3 or more episodes were associated with an increased mortality rate. Thus, urine output is a sensitive and early marker for AKI and is associated with adverse outcomes in intensive care unit patients. Acute kidney injury (AKI) is associated with a 40–80% mortality rate, increased length of hospital stay, and high costs in critically ill patients.1.Chertow G.M. Burdick E. Honour M. et al.Acute kidney injury, mortality, length of stay, and costs in hospitalized patients.J Am Soc Nephrol. 2005; 16: 3365-3370Crossref PubMed Scopus (2256) Google Scholar, 2.Hoste E.A. Kellum J.A. Acute kidney injury: epidemiology and diagnostic criteria.Curr Opin Crit Care. 2006; 12: 531-537Crossref PubMed Scopus (156) Google Scholar, 3.Silvester W. Bellomo R. Cole L. Epidemiology, management, and outcome of severe acute renal failure of critical illness in Australia.Critical Care Med. 2001; 29: 1910-1915Crossref PubMed Scopus (271) Google Scholar, 4.Brivet F.G. Kleinknecht D.J. Loirat P. et al.Acute renal failure in intensive care units--causes, outcome, and prognostic factors of hospital mortality; a prospective, multicenter study. French Study Group on Acute Renal Failure.Crit Care Med. 1996; 24: 192-198Crossref PubMed Scopus (729) Google Scholar Even minor changes in renal function, such as a serum creatinine (sCr) increase of 0.3 mg/dl, are associated with short-term and long-term mortality and the development of chronic kidney disease.5.Chertow G.M. Levy E.M. Hammermeister K.E. et al.Independent association between acute renal failure and mortality following cardiac surgery.Am J Med. 1998; 104: 343-348Abstract Full Text Full Text PDF PubMed Scopus (992) Google Scholar It is still unclear whether these changes are a marker for AKI severity or direct mediators of the adverse outcomes. Regardless, the emerging role of declining renal function as a risk factor for mortality advocates an urgent need for early recognition and management of AKI. These data have catalyzed the search for novel, more sensitive, and specific biomarkers to detect AKI earlier. Several candidates are undergoing prospective validation: neutrophil gelatinase-associated lipocalin, kidney injury molecule-1, liver fatty acid-binding protein, α- and π-glutathione S-transferase.6.Han W.K. Waikar S.S. Johnson A. et al.Urinary biomarkers in the early diagnosis of acute kidney injury.Kidney Int. 2008; 73: 863-869Abstract Full Text Full Text PDF PubMed Scopus (432) Google Scholar, 7.Bagshaw S.M. Gibney R.T. Conventional markers of kidney function.Crit Care Med. 2008; 36: S152-S158Crossref PubMed Scopus (100) Google Scholar, 8.Parikh C.R. Abraham E. Ancukiewicz M. et al.Urine IL-18 is an early diagnostic marker for acute kidney injury and predicts mortality in the intensive care unit.J Am Soc Nephrol. 2005; 16: 3046-3052Crossref PubMed Scopus (416) Google Scholar, 9.Westhuyzen J. Endre Z.H. Reece G. et al.Measurement of tubular enzymuria facilitates early detection of acute renal impairment in the intensive care unit.Nephrol Dial Transplant. 2003; 18: 543-551Crossref PubMed Scopus (263) Google Scholar Although oliguria is a frequent event in intensive care unit (ICU) patients and urine flow has been proposed as a diagnostic and staging criterion for AKI, none of the biomarker studies have focused on urine output (UO) as a renal biomarker. Most of the previous studies were derived from analysis of retrospective data, and the UO criterion, if not omitted, was modified, confirming that its application is challenging and that further studies are required to validate this criterion. We believe that urine flow rate is a sensitive and specific biomarker that provides an early warning signal for impending renal dysfunction. We hypothesized that the UO criteria would increase the sensitivity and specificity of the Acute Kidney Injury Network (AKIN) classification system, as oliguric patients without sCr change would have an increased mortality, dialysis requirements, and longer lengths of ICU stay than non-AKI patients. In addition, we hypothesized that changes in urine volume would be an earlier marker of AKI in comparison with the sCr criterion. We assessed different definitions of oliguria to determine whether the current standard for AKI diagnosis and staging is optimal in comparison with alternate criteria. This study included 317 patients. We assessed the incidence of oliguria using different definitions as presented in Figure 1. One hundred and fifty one patients (47%) had an episode of oliguria during their ICU stay. More patients were classified as oliguric using the total urine volume over a 6-h period (olig6-moving blocks (mblock) and olig6-fixed blocks (fblock)) compared with the current AKIN UO definition, which requires 6 consecutive hours with <0.5 ml/kg (olig6 consecutive (consec)). Patients who met olig6-consec definition had the highest rate of progression to AKIN stage 2 based on UO criterion, 74/94 (79%). Of these 74, 15 (20%) progressed to olig24 (UO <7.2 ml/kg in 24 h or anuria for 12 h). Olig6-mblock was the most sensitive definition, but had the lowest rate of progression: 70/150 (52%) to stage 2 and 15 (19%) to olig24. The rate of progression of patients who had met olig6-fblock was 64% to olig12 and 19% to olig24. Table 1 shows the sensitivity, specificity, and positive and negative predictive values of the different oliguria definitions based on AKIN sCr criterion as the gold standard for AKI diagnosis. Olig6-mblock had the highest sensitivity (0.54) and olig6-consec the highest specificity (0.71). Increasing the time frame to detect oliguria from 6 to 12 h (olig12) and to 24 h (olig24) increased the specificity, but markedly decreased the sensitivity compared with the AKIN sCr definition. Within any 6-h interval, olig6-fblock had the best positive and negative predictive values. We used olig6-fblock (i.e., <3 ml/kg during a 6-h fblock) to further classify the patients into four groups: non-AKI, non-oliguric AKI, oliguric with sCr change (Type A), and oliguric without sCr change (Type B).Table 1Sensitivity, specificity, positive, and negative predictive values for UO criteria based on serum creatinine as the standard diagnostic criteriaaAKIN stage 1 serum creatinine criterion: absolute 0.3mg/dl or relative 50% increase from reference serum creatinine within 48h.AKIN Stage 1Stage 2 Olig12Stage 3 Olig24cUrine output <7.2ml/kg in 24h or anuria for 12h.Olig6-consecbAKIN stage 1 urine output criterion.Olig6-mblockOlig6-fblockNo. of patients941501267915Sensitivity0.34 (0.24–0.45)0.53 (0.41–0.64)0.47 (0.35–0.58)0.30 (0.20–0.41)0.08 (0.03–0.17)Specificity0.71 (0.65–0.77)0.54 (0.48–0.61)0.62 (0.56–0.68)0.76 (0.70–0.81)0.96 (0.92–0.98)Positive predictive value0.270.270.280.290.37Negative predictive value0.770.780.790.770.76Abbreviations: AKIN, Acute Kidney Injury Network; consec, consecutive; fblock, fixed block; mblock, moving block; Olig, oliguria; UO, urine output.a AKIN stage 1 serum creatinine criterion: absolute 0.3 mg/dl or relative 50% increase from reference serum creatinine within 48 h.b AKIN stage 1 urine output criterion.c Urine output 70 years18 (25)19 (30)3 (20)0.70 Diabetes14 (20)24 (37)9 (60)0.003 Hypertension21 (30)26 (40)5 (33)0.40 Obesity (BMI >30)21 (30)24 (37)8 (53)0.19 Chronic liver disease8 (11)9 (14)7 (43.8)0.003 Chronic lung disease10 (14)4 (6)7 (46)<0.001 Congestive heart failure5 (7)8 (12)6 (40)0.002 Reference sCr (mg/dl), median (IQR)0.9 (0.7–1.1)0.9 (0.8–1.3)1.0 (0.7–1.4)0.42 Atherosclerotic disease8 (11)11 (17)7 (46)0.004Acute risk factors Severe infection/sepsis11 (15)9 (14)7 (46)0.009 Hypotension23 (32)18 (28)12 (80)0.001 Mechanical ventilation31 (43)34 (53)12 (80)0.03 pH value ≤7.306 (8)13 (20)6 (40)0.007 Nephrotoxin exposure10 (14)11 (17)6 (40)0.06Abbreviations: AKIN, Acute Kidney Injury Network; BMI, body mass index; IQR, interquartile range; mblock, moving block; olig, oliguria; sCr, serum creatinine. Open table in a new tab Abbreviations: AKIN, Acute Kidney Injury Network; BMI, body mass index; IQR, interquartile range; mblock, moving block; olig, oliguria; sCr, serum creatinine. At AKI diagnosis of the 167 patients, 56 were non-oliguric (sCr changes alone) and 111 were oliguric (olig6-fblock definition; Figure 3). One hundred and sixty five patients were classified as AKIN stage 1 and two patients were classified as AKIN stage 2 (by the sCr criteria). Fifteen non-oliguric patients had an episode of oliguria after AKI diagnosis and were classified as Type A group (sCr and oliguria by olig6-fblock), based on maximum stage during ICU stay. Of the oliguric patients, 21 subsequently had sCr changes and were also classified as Type A (sCr and oliguria by olig6-fblock). Based on maximum AKIN stage reached, 41 patients were non-oliguric, 36 had oliguric AKI with sCr elevation (Type A) and 90 were oliguric without change in sCr (Type B). Fifteen patients progressed to AKIN stage 2 exclusively by sCr criterion, 8 by both sCr and oliguria criteria, and 71 exclusively by oliguria criteria. A total of 22 patients reached Stage 3 criteria; 7 exclusively by sCr, 1 by both criteria, and 14 exclusively by oliguria criteria (Figure 3). Table 3 compares the demographic and patient characteristics at ICU admission of non-AKI patients, non-oliguric patients, and Type A and B oliguric patients. There was no difference in the number of patients with chronic kidney disease or the reference sCr level among groups. However, Type A oliguric patients were older and had more co-morbidities, with a higher incidence of diabetes mellitus, hypertension, and obesity. Type A oliguric patients also had higher incidence of sepsis and hypotension, required mechanical ventilation and were exposed to nephrotoxic drugs more often.Table 3Patient characteristics by AKI diagnostic criteriaNumber of patients, n (% total)Non-AKI 151 (47)Non-oliguric AKI 41 (13)Oliguric with sCr change (Type A) 36 (12)Oliguric without sCr change (Type B) 90 (28)P-valueDemographic characteristics Age (years), median (IQR)42 (27–61)45 (31–58)54 (43–76)59 (48–73)<0.0001 Race—Caucasian, n (%)71 (47.0)21 (51.2)18 (50.0)53 (58.2)0.262 Gender, male95 (62.9)26 (63.4)24 (66.7)55 (60.4)0.568 BMI (kg/m2), median (IQR)25 (22–29)24 (22–27)28 (24–35)27 (24–34) 70 years19 (12.6)5 (12.2)10 (27.8)25 (27.5)0.009 Diabetes mellitus19 (12.6)5 (12.2)17 (47.2)29 (31.9)<0.0001 Hypertension27 (17.9)9 (22.0)19 (52.8)25 (27.5)<0.0001 Morbid obesity33 (21.9)4 (9.8)14 (38.9)32 (35.2)0.003 Chronic liver disease5 (3.3)7 (17.1)11 (30.6)11 (12.1)<0.0001 Congestive heart failure3 (2.0)2 (4.9)11 (30.6)7 (7.7)<0.0001 Chronic lung disease7 (4.6)7 (17.1)10 (27.8)9 (9.9)<0.0001 Cerebrovascular accident28 (18.5)9 (22.0)8 (22.2)22 (24.2)0.76 Chronic kidney disease5 (3.3)2 (4.9)1 (2.8)1 (1.1)0.69Acute risk factors—n (%) Severe infection/sepsis9 (6.0)6 (14.6)15 (41.7)10 (11.0)<0.0001 Hypotension26 (17.2)11 (26.8)17 (47.2)30 (33.0)0.001 Mechanical ventilation57 (37.7)19 (46.3)25 (69.4)43 (47.3)0.007 pH value ≤7.3013 (8.6)5 (12.2)12 (33.3)11 (12.1)0.001 Nephrotoxin exposure17 (11.3)9 (22.0)11 (30.6)11 (12.1)0.014Abbreviations: AKI, acute kidney injury; BMI, body mass index; IQR, interquartile range; sCr, serum creatinine.Continuous variables are expressed as median (IQR) and categorical variables as frequency (n) and percentage (%). P-value for the difference among groups. Open table in a new tab Abbreviations: AKI, acute kidney injury; BMI, body mass index; IQR, interquartile range; sCr, serum creatinine. Continuous variables are expressed as median (IQR) and categorical variables as frequency (n) and percentage (%). P-value for the difference among groups. We calculated the time to reach AKI diagnosis from ICU admission by oliguria and by sCr criteria. Type B oliguric patients were diagnosed earlier than non-oliguric AKI (UO 12 h (interquartile ratio (IQR) 6–24) vs. sCr 24 h (IQR 12–37); P=0.008; Figure 4). In patients who met both criteria (Type A), there was no significant difference in time to reach UO or sCr diagnosis; UO 12 h (IQR 6–27) vs. sCr 14 h (IQR 7–23); P=0.36. The frequency to require renal replacement therapy was different among the groups of AKI diagnosis, but not significantly different between Type A (2.2%) and Type B (13.9%; P=0.20) oliguric patients. The lengths of ICU and hospital stay were shorter in non-AKI than in Type B oliguric patients. The ICU length of stay was 2 (IQR 2–3) in non-AKI versus 3 (IQR 2–7) in Type B (P<0.001). Hospital length of stay was 6 (IQR 3–13) in non-AKI versus 8 (IQR 4–17) in Type B (P=0.02). The overall mortality rate was 5.6% and varied in each group (Figure 5). The mortality rate in AKI patients was 9.5%, compared with 1.3% in non-AKI patients (P=0.001). Type B oliguric patients had a significantly higher mortality rate than non-AKI patients (non-AKI 1.3 vs. 8.8% olig6-fblock; P=0.007). Applying the UO criterion in addition to the sCr criterion, increases the area under the curve to predict mortality for AKI from 0.60 (95% confidence interval 0.46–0.75; P=0.12) to 0.69 (95% confidence interval 0.58–0.79; P=0.006). We evaluated the association of AKI diagnosis and mortality by adjusting for age, cumulative fluid balance (from ICU admission to the day of AKI diagnosis by UO or sCr), sepsis, and need for mechanical ventilation. Table 4 shows that the association of AKI diagnosis with mortality was maintained in non-oliguric, and Type A and B oliguric patients after multivariate adjustment.Table 4Independent predictors for ICU mortalityOR95% CIP-valueNon-oliguric AKI Sepsis9.834.17–23.120.000 AKI diagnosis2.961.21–7.190.017Type B—oliguric AKI without sCr change (definition olig6-mblock) AKI diagnosis5.091.69–15.30.004 Sepsis10.915.29–22.50.000Type A—oliguric AKI with sCr change Age1.011.00–1.020.041 Sepsis6.362.44–16.50.000 AKI diagnosis5.561.22–25.40.027Abbreviations: AKI, acute kidney injury; CI, confidence interval; ICU, intensive care unit; mblock, moving block; OR, odds ratio; sCr, serum creatinine.Multivariate analysis adjusting the association of AKI diagnosis (non-oliguric, oliguric with sCr change, and oliguric without sCr change) and mortality for age, cumulative fluid balance (from ICU admission to the day of AKI diagnosis by urine output or sCr), sepsis diagnosis, and need for mechanical ventilation. Open table in a new tab Abbreviations: AKI, acute kidney injury; CI, confidence interval; ICU, intensive care unit; mblock, moving block; OR, odds ratio; sCr, serum creatinine. Multivariate analysis adjusting the association of AKI diagnosis (non-oliguric, oliguric with sCr change, and oliguric without sCr change) and mortality for age, cumulative fluid balance (from ICU admission to the day of AKI diagnosis by urine output or sCr), sepsis diagnosis, and need for mechanical ventilation. The median number of non-consecutive hours with urine volume <0.5 ml/kg was 10 h (IQR 3–22); 22 h (IQR 12–40) for patients with at least one episode of oliguria versus 4 h (IQR 2–7) in patients with no episode of oliguria. Non-survivors had a statistically significant higher non-consecutive number of hours with urine volume <0.5 ml/kg: 24.5 h (IQR 11.2–37) than survivors: 9 h (IQR 3–21); P=0.002. There was an increment in mortality rate in patients presenting more than 12 h of oliguria (Figure 6). Among patients presenting with oliguria (olig6-fblock), the median number of episodes was 2 (IQR 1–5). Patients with more than three episodes of oliguria presented a significant higher mortality rate than patients with less than three episodes of oliguria (30 vs. 6%; P=0.01; Figure 6). Oliguria is a frequent event in ICU patients, being the final pathway for several injuries to renal parenchyma; still oliguria is also the result of transitory changes in volume status or of external influences, such as drug administration. Although the UO is currently included as a criterion to diagnose and stage AKI, few studies have evaluated if decreased UO without elevation in sCr is a specific marker of AKI that correlates with outcomes. A major barrier to the application of the UO criterion is that accurate hourly UO measurements have been difficult to obtain. Urine flow measurements are time-consuming, as urine meters require manipulation, visual assessment, and manual data recording. In most ICUs, nurses empty the collection bag every 6 h. These difficulties in measuring, monitoring, and accurately recording UO have resulted in a lack of a standardized approach to assessing changes in UO and identifying of episodes of oliguria. While oliguria can be considered a marker of renal function and a criterion that correlates with outcomes, devices providing a continuous and accurate measurement of urine flow are not widely available. The hourly information on urine volume, with more frequent observations of the parameter, could allow an earlier detection of renal dysfunction. Treating urine flow as a continuous physiological variable instead of as interval parameter, would provide more time points for the detection of AKI. For intervention trials on prevention and treatment of AKI, accurate hourly monitoring of urine flow would provide more opportunities for intervention.10.Mehta R.L. Timed and target therapy for acute kidney injury: a glimpse of the future.Kidney Int. 2010; 77: 947-949Abstract Full Text Full Text PDF PubMed Scopus (17) Google Scholar On the other hand, for retrospective evaluations and prospective epidemiological studies, the assessment of total urine volume in a longer time intervals could facilitate the application of the criteria, as most hospitals do not have digital monitors to record UO hourly. Balancing the practicality of using longer time intervals against ascertaining urine flow every hour to diagnose oliguria is challenging. In our study, we initially compared different definitions of oliguria to detect changes in UO. As hourly UO was recorded in the electronic medical record, we developed the olig6-consec calculation to correspond to the strictest interpretation of the UO AKIN criterion, requiring the urine volume to be <0.5 ml/kg every hour during 6 consecutive hours. The olig6-mblock definition was less strict in accepting a cumulative decrease in UO to <3 ml/kg over 6 consecutive hours (Figure 7). However, both these calculations required computation of urine volume in consecutive 6-h periods, requiring a moving reference point to define the 6-h period. This moving reference point adds some complexity to the calculations, and patients may not be classified as AKI if the decline in UO varies over a 6-h interval, as shown in Figure 7. We therefore tested the olig6-fblock and olig12 calculations to evaluate if applying the criterion in blocks of 6 or 12 h, matching a nurse's shift, would facilitate the application without decreasing the sensitivity and specificity of the criterion. We demonstrated that assessing the urine volume in a 6-h interval (olig6-mblock and olig6-fblock) resulted in an increased sensitivity as compared with olig6-consec. Although the assessment of the total urine volume over 6 h decreased the specificity of the criterion, the olig6-fblock definition had the best positive predictive value for AKI, based on sCr criterion (Table 1). We also assessed the stage progression based on UO criterion, and confirmed that patients classified by olig6-consec, the strictest definition, had a higher rate of stage progression, and those classified by olig6-mblock and olig6-fblock had an intermediate rate of progression: 79, 52, and 64%, respectively. As the olig6-fblock definition had the best overall performance, we used that for the subsequent analysis. Our study confirmed the finding of previous studies that applying the UO criterion in addition to the sCr criterion increases the number of patients diagnosed as AKI. Although the UO increases the sensitivity of the diagnostic criteria, the specificity of this criterion has not been defined.11.Joannidis M. Metnitz B. Bauer P. et al.Acute kidney injury in critically ill patients classified by AKIN versus RIFLE using the SAPS 3 database.Intensive Care Med. 2009; 35: 1692-1702Crossref PubMed Scopus (366) Google Scholar, 12.Haase M. Bellomo R. Matalanis G. et al.A Comparison of the RIFLE and Acute Kidney Injury Network classifications for cardiac surgery-associated acute kidney injury: a prospective cohort study.J Thoracic Cardiovasc Surg. 2009; 138: 1370-1376Abstract Full Text Full Text PDF PubMed Scopus (131) Google Scholar, 13.Hoste E.A. Kellum J.A. RIFLE criteria provide robust assessment of kidney dysfunction and correlate with hospital mortality.Crit Care Med. 2006; 34: 2016-2017Crossref PubMed Scopus (68) Google Scholar, 14.Cruz D.N. Bolgan I. Perazella M.A. 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: 418-425Crossref PubMed Scopus (199) Google Scholar, 15.Barrantes F. Tian J. Vazquez R. et al.Acute kidney injury criteria predict outcomes of critically ill patients.Crit Care Med. 2008; 36: 1397-1403Crossref PubMed Scopus (162) Google Scholar In our study, the incidence of AKI increased from 24 to 52% (using sCr criterion only) applying both criteria. Joannidis et al.,11.Joannidis M. Metnitz B. Bauer P. et al.Acute kidney injury in critically ill patients classified by AKIN versus RIFLE using the SAPS 3 database.Intensive Care Med. 2009; 35: 1692-1702Crossref PubMed Scopus (366) Google Scholar using SAPS 3 database with more than 16,000 ICU patients, assessed urine volume in a 24-h interval. In that study, they modified the AKIN UO criterion, classifying patients with 48 h prior to screening, transferred from another ICU, had a sCr >177 μmol/l ≤3 days before ICU admission, were prisoners, received dialysis within the 12 months prior to admission, had a functioning kidney transplant, were on anticoagulants or warfarin within the last 7 days, suffered from decompensated cirrhosis, had chronic kidney disease stage 5, were anemic (hemoglobin <90 g/l or hematocrit <27%) or were already enrolled in another research project. The Institutional Review Board approved a retrospective analysis of all patients screened for the study, and a waiver of individual authorization for use of Protected Health Information was granted as stipulated by the Health Insurance Portability and Accountability Act rules. We recorded demographic data, co-morbidities, clinical history, and lab studies from the day of ICU admission until ICU discharge. sCr was available at least once per 24 h. We applied the AKIN criteria to define AKI by sCr (creatinine change ≥0.3 mg/dl or ≥50% from reference within 48 h). We considered the first sCr measured at ICU admission as the reference sCr. We computed daily and cumulative fluid balance for each patient; however, we did not record details of the type and duration of fluid administration or use of diuretics and other medications. We classified patients by three different definitions of oliguria on a 6-h time interval (Table 1). Each of these oliguria definitions was compared with AKIN sCr criteria for diagnosing AKI. The oliguria duration over ICU stay was assessed by the number of olig6-fblock episodes and by the total number of hours during which the patient had a urine volume <0.5 ml/kg during the ICU stay. We classified patients at time of diagnosis as oliguric or non-oliguric. Patients reaching oliguria diagnosis before elevation of sCr were classified as being diagnosed by oliguria criterion, and those reaching sCr criterion first were classified as non-oliguric at diagnosis. We additionally assessed the maximum stage reached throughout the course of ICU stay. On the basis of this maximum AKIN stage (either UO or sCr), we stratified patients on one of three categories: non-AKI, non-oliguric AKI, oliguric with sCr change (Type A), and oliguric without sCr change (Type B). Patients with no AKI by either criterion were compared with non-oliguric AKI and with Type A (with sCr changes) and B (without sCr changes) oliguric patients. We compared demographics and risk factors for AKI in these groups of patients. We assessed the rate of progression to more severe stages of AKI, need of renal replacement therapy, length of ICU, and ICU mortality in these patients by AKI diagnosis. We calculated the time to reach AKI diagnosis from ICU admission by oliguria and by sCr criteria. In Type A oliguric AKI patients, the time to reach the criteria was compared. Continuous variables were expressed as mean±s.d. and analyzed by unpaired Student's t-test, or Wilcoxon rank-sum test, as appropriate. Non-parametric variables were expressed as median and 25–75 percentiles and analyzed by Mann–Whitney's test. Categorical variables were analyzed by Pearson's two-test or Fisher's exact test, whenever appropriate. The predictive value of UO on AKI diagnosis was performed using a receiver operator characteristic curve, and the area under the curve was computed. All statistical tests were two-sided, and P<0.05 was considered to be significant. A multivariate stepwise logistic regression was used to assess the association of AKI diagnosis and mortality. We evaluate the association of AKI diagnosis using only sCr criterion and including UO criterion. We adjusted the association of AKI diagnosis and mortality for age, cumulative fluid balance (from ICU admission to the day of AKI diagnosis by UO or sCr), sepsis, and need for mechanical ventilation. Statistical analyses were conducted using SPSS 17.0 (Chicago, IL). EM's work has been made possible through an International Society of Nephrology Fellowship and CNPq (Conselho Nacional de Desenvolvimento Científico e Tecnológico) support. JB is a recipient of a research fellowship from the Kidney Foundation of Canada. Research was supported through the O’Brien Center.