J Vet Intern Med 2012

Acute Azotemia as a Predictor of Mortality in Dogs and Cats

E. Harison, C. Langston, D. Palma, and K. Lamb
Background: Acute kidney injury (AKI) has been shown to be a predictor of mortality in human medicine. Published studies in the veterinary literature evaluating relative changes in serum creatinine concentration as a prognostic factor are limited. Objective: To evaluate an AKI grading system based on serum creatinine concentration to determine if it correlates with outcome prediction in dogs and cats. Animals: Six hundred forty-ve dogs and 209 cats that had at least 2 serum creatinine concentration measurements measured within 7 days. Methods: Retrospective study. Dogs and cats with an initial serum creatinine concentrations of  1.6 mg/dL and that had more than 1 concentration measured within 2, 3, and 7 days were placed into levels (02) based on absolute changes. Mortality then was determined at 30 and 90 days. Results: Based on odds ratios calculated with a 95% condence interval, dogs placed in level 1 within 2 days were approximately 3 times more likely to die within 90 days. Dogs placed in level 2 within 2, 3, or 7 days were approximately 3 times more likely to die within 30 or 90 days. Cats placed in level 2 within 3 or 7 days were approximately 3 times more likely to die at 30 days and 4 times more likely to die if placed in this level within 7 days. If placed in level 2 within 2 or 3 days, cats were approximately 3 times more likely to die within 90 days. Conclusions and Clinical Importance: Detecting increasing severity of azotemia helps predict mortality in dogs and cats. Key words: Acute renal failure; Azotemia; Intensive care medicine; Renal function.

here have been several dierent denitions of acute kidney injury (AKI) in human medicine, but it is generally dened as an acute and abrupt decrease in kidney function.1,2 Although serum creatinine concentration (SCr) is not a sensitive marker of kidney function, it is widely used as a surrogate marker because it is rapidly and easily measured. AKI commonly is divided into prerenal, intrarenal, and postrenal causes.3 Prerenal azotemia causes a functional decrease in glomerular ltration without structural kidney injury and typically is more rapidly reversible than AKI associated with structural damage.3 In human medicine, 2 classication systems recently have been validated that use evidence of AKI to predict outcome in patients in the critical care setting.2,4,5 The RIFLE (R-risk, I-injury, F-failure, L-loss, and E-end stage kidney disease) and AKIN (Acute Kidney Injury Network) classication systems have been compared in human literature and show no statistical dierence in predicting mortality (Table 1).4,5 Both systems have documented an increased risk of mortality in patients with increasing severity of kidney injury as indicated by increasing SCr, decreasing urine output, or both. Most studies evaluated SCr concentrations within a

Abbreviations:
AKI AKIN APACHE II CI CKD GFR HA-AKI ICU OR RIFLE SCr SOFA acute kidney injury Acute Kidney Injury Network Acute Physiology and Chronic Health Evaluation condence interval chronic kidney disease glomerular ltration rate hospital-acquired acute kidney injury intensive care unit odds ratio risk injury failure loss end-stage kidney disease serum creatinine concentration Sequential Organ Failure Assessment

From the Internal Medicine, Animal Medical Center, New York, NY (Harison); the Animal Medical Center, New York, NY (Langston, Palma); and the Lamb Consulting, Gainesville, FL (Lamb). This study was not supported by a grant and was privately funded. Presented as a poster at the 2011 ACVIM Forum, Denver, CO. Corresponding author: Emily Harison, DVM, Animerge, 21 US HW 206, Raritan, NJ 08869; e-mail: Emily.harison@gmail.com

Submitted October 4, 2011; Revised May 24, 2012; Accepted June 27, 2012.
Copyright 2012 by the American College of Veterinary Internal Medicine 10.1111/j.1939-1676.2012.00985.x

48-hour period. As the retrospective nature of these studies, urine output was not commonly available. More recently, transient azotemia (TA) has been described as a rapidly reversible form of AKI that is an independent predictor of mortality.3 In addition to mortality, the median duration of hospitalization increases with increasing severity of AKI in humans.6,7 Acute Physiology and Chronic Health Evaluation scores (APACHE II) and Sequential Organ Failure Assessment (SOFA) scores are systems used to evaluate critically ill human patients. Patients with AKI were shown to have higher APACHE II and SOFA scores or be more likely to present with multiple organ dysfunction syndrome than patients without AKI.7 Some studies have documented that changes in SCr, both increasing and decreasing, may be signicant clinically relevant with regard to mortality.6,7 For example, higher intensive care unit (ICU) mortality has been documented in the patients with increasing SCr when compared to those with increases in SCr on presentation, that later decreased.7 There is limited information regarding hospital acquired acute kidney injury (HA-AKI) in

Harison et al

Table 1. Acute kidney injury network (AKIN) and risk injury failure loss end-stage kidney disease (RIFLE) criteria.2,4
AKIN Criteria Serum Creatinine Criteria Stage 1 Stage 2 Stage 3  0.3 mg/dL or >150% increase from baseline serum creatinine  200% increase from baseline serum creatinine  300% increase from baseline serum creatinine or absolute serum creatinine  4.0 mg/dL with an acute rise of  0.5 mg/dL RIFLE Criteria Serum Creatinine and Glomerular Filtration Rate (GFR) Criteria Risk (R) Injury (I) Failure (F) 1.5 times increase in creatinine or >25% decrease in GFR 2 times increase in creatinine or >50% decrease in GFR 3 times increase in creatinine or >75% decrease in GFR or an absolute creatinine  4 mg/dL with an acute rise  0.5 mg/dL Persistent acute renal failure = complete loss of renal function for >4 weeks End stage renal disease Urine Output Criteria <0.5 mL/kg/h for  6 hours <0.5 mL/kg/h for  12 hours <0.3 mL/kg/h for  24 hours or anuria for  12 hours Urine Output Criteria <0.5 mL/kg/h for  6 hours <0.5 mL/kg/h for  12 hours <0.3 mL/kg/h for  24 hours or anuria  12 hours

Table 2. Acute kidney injury staging criteria.

Level Level 0 Level 1 Level 2 Highest Additional Creatinine (mg/dL)  1.6 and change of <0.3  1.6 and change of  0.3 >1.6 and change of  0.3

2 SCr results within 3 days and then as well as within 2 days. All patients in the 2-day group also were included in the 3-day group, and all patients in the 2-day group and 3-day group were included in the 7-day group, even if no subsequent SCr measurements were obtained after 2 or 3 days. Patients were recorded as alive or dead at 30 and 90 days. If this information was not available in the medical record, referring veterinarians and, if necessary, owners were contacted by phone or mail to determine status. All SCr measurements were performed in the AMCs clinical pathology laboratory by kinetic modication of the Jae reaction (alkaline picrate).a The coecient for variation for the machine was 0.9%.a Patients were not necessarily admitted to the hospital and were not selected based on disease process. Patients receiving diuretic therapy were included in the study.

Exclusion Criteria
Patients receiving dialysis were excluded because of the uctuations of the SCr values between and during treatments. Patients with an initial SCr of >1.6 mg/dL also were excluded.

Clinical Staging
An acute increase in SCr of  0.3 mg/dL within the described time frame (2, 3, or 7 days) was used as evidence of AKI. Although the Cowgill staging system species an acute increase in 48 hours, we extended that interval to 3 and 7 days to evaluate the utility of a longer evaluation period.9 Each patient was placed into an AKI level ranging from 0 to 2 based on SCr as described in Table 2. Initially, we categorized patients based on 5 stages of AKI adapted from the Cowgill staging system, but because of the low number of patients in Cowgill stages 25, we combined those stages into stage 2 for this study.9 If a patient had more than 2 SCr deteriminations within 2 days, the initial SCr was compared only with the highest subsequent SCr within the specied time frame. The same protocol was followed for patients that had 2 SCr determinations within 3 and 7 days.

Loss (L)

End stage renal disease (E)

veterinary patients to date, and the impact of minor changes in SCr is rarely considered. The 1 published study evaluating a staging system based on the AKIN system showed an increase in mortality for every stage of HA-AKI.8 A new staging system based on absolute SCr values has been proposed for veterinary use by Cowgill.9 An advantage of this staging system is that patients with AKI can be staged even if baseline SCr value remains unknown.9 Our goal was to evaluate this grading system to see if it correlated with outcome prediction in dogs and cats. The purpose of this study was to determine if an increase in SCr after presentation to the hospital was associated with increased risk of mortality. An additional goal was to determine if there is an advantage of measuring changes in SCr within 3 dierent time frames.

Statistical Methods
Logistic regression was conducted with primary endpoints of 30 and 90 day mortality. Clinical groups were stratied for analysis by species and time frame in which SCr determinations were made (within 2, 3 or 7 days), and designated as survival type. The independent variable being measured was level with 0 as the reference group. To determine whether the levels of stratication were synergistic, an interaction model was generated for both 30 and 90 day mortality, including stage, species, and survival type, but, only the main eects models are presented in the tables. Odds ratios with a condence interval of 95% were calculated to determine signicance of level as a predictor of mortality. A P value <.05 was considered signicant.

Materials and Methods

Criteria for Selection of Cases
Laboratory records of the Animal Medical Center (AMC) from January 2007 through December 2009 were searched to identify any dog or cat that had at least 2 SCr measurements within 7 days. Patients then were subdivided into those that had

Results
Four hundred dogs and 128 cats met the inclusion criteria and  2 or more SCr determinations within

Azotemia as a Predictor

Table 3. Stratication of acute kidney injury (AKI) in dogs and cats monitored for 90 days.*
Level of AKI 2-day 0 1 2 3-day 0 1 2 7-day 0 1 2 Dogs N = 400 339 (85%) 29 (7%) 32 (8%) N = 476 400 (84%) 38 (8%) 38 (8%) N = 645 520 (80%) 70 (11%) 55 (9%) Cats N = 128 101 (79%) 6 (5%) 21 (16%) N = 164 131 (80%) 10 (6%) 23 (14%) N = 209 163 (78%) 14 (7%) 32 (15%)

*2, 3, and 7-day indicate the time frame during which creatinine values were evaluated.

2 days. Four hundred seventy-six dogs and 164 cats had 2 or more concentrations measured within 3 days. Six hundred forty-ve dogs and 209 cats had  2 SCr determinations within 7 days. Staging results for measurement within 2, 3, and 7 days are presented in Table 3. Seven percent of dogs were categorized as level 1 AKI within the 2-day evaluation period and 8% reached level 2 AKI. For cats, 5% were categorized as level 1 AKI and 16% reached level 2 AKI within the 2-day evaluation period. Eight percent of dogs were placed in both level 1 and level 2 within the 3-day period. For cats, 6% were placed in level 1 and 14% were placed in level 2 for the 3-day period. Eleven percent of dogs were placed in level 1 and 9% of dogs were placed in level 2 for the 7-day period. Seven percent of cats were placed in level 1 and 15% were placed in level 2 for the 7-day period. Odds ratios of mortality, condence intervals, and P values are presented in Table 4 and complete mortality

Table 4. Odds ratios and mortality rates for dogs and cats developing acute kidney injury at 30 and 90 days.
2-Day OR (CI 95%) Dogs, level 1 30 day mortality 90 day mortality Dogs, level 2 30 day mortality 90 day mortality Cats, level 1 30 day mortality 90 day mortality Cats, level 2 30 day mortality 90 day mortality 1.83 (0.834.02) 2.98 (1.386.43) 2.99 (1.436.23) 3.54 (1.697.45) 2.13 (0.4111.12) 1.56 (0.38.14) 2.34 (0.96.07) 3.13 (1.168.44) P < .05 .135 .004 .004 .001 .372 .595 .081 .02 3-Day OR (CI 95%) 1.40 (0.682.89) 1.94 (0.993.80) 3.38 (1.726.64) 3.66 (1.857.2) .91 (0.223.69) .65 (0.162.63) 2.75 (1.126.79) 3.47 (1.349.02) P < .05 .357 .055 .001 .001 .893 .547 .03 .01 7-Day OR (CI 95%) 1.29 (0.752.22) 1.32 (0.792.22) 3.20 (1.845.55) 3.65 (2.066.50) 1.46 (0.474.58) .89 (0.292.78) 3.13 (1.466.70) 3.13 (1.168.44) P < .05 .362 .300 <.001 <.001 .518 .841 .003 .001

OR, Odds ratio with level 0 being the referent group; CI, condence interval; bold indicates signicant ndings.

Table 5.

Numbers of surviving dogs and cats based on acute kidney injury level.
Dogs Cats 90 day 30 day 90 day

30 day

Alive 2 day Level Level Level Total 3 day Level Level Level Total 7 day Level Level Level Total 0 1 2 254 18 16 400 301 26 18 476 391 78 29 645

Dead 85 11 16

% Alive 74.9 62.1 50.0

Alive 240 13 13 400 282 21 15 476 360 44 21 645

Dead 99 16 19

% Alive 70.8 44.8 40.6

Alive 68 3 10 128 89 7 10 164 115 9 14 209

Dead 33 3 11

% Alive 67.3 50.0 47.6

Alive 61 3 7 128 79 7 7 164 100 9 9 209

Dead 40 3 14

% Alive 60.4 50.0 33.3

0 1 2

99 12 20

75.3 68.4 47.4

118 17 23

70.5 55.3 39.5

42 3 14

67.9 70.0 41.7

52 3 16

60.3 70.0 30.4

0 1 2

129 21 20

75.2 78.8 59.2

160 26 34

69.2 62.9 38.2

48 5 18

70.6 64.3 43.8

63 5 24

61.3 64.3 27.3

Harison et al

counts for each measurement period are presented in Table 5, respectively. Both dogs and cats developing level 2 AKI had a signicantly increased risk of mortality at both 30 and 90 days. The exception was cats surviving to 30 days that developed level 2 AKI within the 2-day time frame. These cats did not show a significantly increased risk of mortality. Dogs with level 1 AKI had an increased mortality rate if the change in SCr occurred within 2 days. All other patients with level 1 AKI did not show an increased risk of mortality within 30 and 90 days.

Discussion
Results of this study indicate that a relationship exists between levels of AKI and mortality at 30 and 90 days. This was especially evident in the 90-day survival rates. Dogs that developed AKI and were placed within level 2 within 7 days were 3.6 times more likely to die before 90 days and 3.2 times more likely to die within 30 days. Cats placed in level 2 within 7 days were 4 times more likely to die within 90 days and 3.1 times more likely to die within 30 days. Only dogs that were placed in level 1 within 2 days had an increased chance of dying within 90 days. However, a major limitation of this study was that it was underpowered and we cannot conclude that development of level 1 AKI (nonazotemic increase in SCr) does not impact mortality, whereas it is clear that development of even modest azotemia increases the risk of death 34 times. Although an increase in SCr is correlated with mortality, it does not imply that patients are dying from renal failure. This study parallels studies in human patients in that higher levels of SCr are associated with worse outcomes.13,6,8,10,11 The RIFLE and AKIN grading systems are based on relative changes in SCr and absolute changes in urine output (Tables 1 and 2).12 This classication system is based on absolute SCr, making it simple to apply, and it can be applied with a single SCr measurement if other indicators of AKI are present. For simplicity, we used the same entry criteria and scale for both dogs and cats, although the reference range for creatinine in cats typically is higher than in dogs. We did not include urine output because it was not accurately measured in the majority of included patients. An increase of  0.3 mg/dL for SCr values as a denition of AKI is well established in AKIN studies.3,57,1013 An increase in SCr does not necessarily dene AKI, but is an easy screening tool to identify patients that likely suered a renal insult. A more precise way to evaluate AKI would include other objective criteria, including glomerular ltration rate (GFR) studies and urine output. Other ways to evaluate AKI would include urine sediment evaluation for casts, measurement of proteinuria, serum or urine biomarkers, and renal histopathology. A recent study by Thoen et al looked at AKI in dogs hospitalized in an ICU.8 These patients had an initial normal SCr and must have had at least 2 SCr

values assessed during their stay. The dogs were placed into 1 of 4 stages (03) based on percent relative change in SCr. Azotemia was dened as a SCr >1.6 mg/dL, and dogs with an initial SCr >1.6 mg/dL were excluded, which is similar to our study. Just as in our study, the majority of patients (85.3%) did not have evidence of AKI; however, of the patients with AKI, 54.2% died compared to a 15.7% mortality rate in the dogs without AKI. Patients meeting the criteria and placed in higher stages were less likely to survive until discharge than those that were placed in lower levels. This study documented that very minimal increases in SCr, even if within the reference range, impacted survival.8 This study was dierent from ours in that all patients were admitted to the ICU, whereas our patients were not limited to inpatients. We expanded the criteria of the Cowgill system to include SCr determinations made over a longer period (3 or 7 days) in addition to measurements made within 48 hours. The AKIN system species that the increase in SCr must occur within a 48-hour period, although the earlier RIFLE reports allowed for the increase in SCr to occur during the course of hospitalization.14 We did not nd a dierence in incidence of AKI or in mortality among the dierent observation periods (2 days versus 3 days versus 7 days), and thus there is no advantage to extending the window of observation. Unlike the Cowgill system, we excluded patients with an initially abnormal SCr.9 This excluded patients with pre-existing AKI, allowing us to focus exclusively on patients that manifested AKI while under veterinary care. However, because an increase in SCr may lag behind the initiation of AKI, we cannot exclude the possibility that the renal injury occurred shortly before presentation to the hospital. We also excluded patients with pre-existing chronic kidney disease (CKD). Humans with CKD are at higher risk for developing AKI and those developing AKI have higher mortality and need for ongoing dialysis.13 One of the potential advantages of this staging system is its ease of application for patients with pre-existing CKD, and evaluation of this system in that population may provide clinically useful information. Another veterinary study evaluated at HA-AKI in 29 dogs.15 Patients were included if they had normal blood urea nitrogen (BUN) concentration and SCr documented <2 weeks before, if prerenal and postrenal causes of azotemia were excluded, if SCr was >2.5 mg/ dL with an increase in BUN and SCr by a factor of at least 3 within 7 days or histologic evidence or clinical signs of AKI were present.15 The mortality rate was 62% for these patients and initial SCr was not dierent between survivors and nonsurvivors.15 AKI in dogs previously has been evaluated by Lee et al. using RIFLE-like criteria.16 Their study included dogs with community-acquired and hospital-acquired AKI.16 They evaluated the correlation between RIFLE criteria and mortality of dogs with AKI as well as prognosis using these criteria.16 Dogs were placed in 1 of 3 categories (risk, injury, and failure) by relative increases in SCr or percentage decrease in GFR calculated based

Azotemia as a Predictor

on SCr concentration.16 Using these criteria, they found a positive correlation between stage and mortality at 30 days.16 By including serum phosphorus concentration and diarrhea status, the criteria appeared to be more prognostic.16 A scoring system for AKI also was described by Segev et al.17 This scoring system was used for dogs undergoing hemodialysis.17 The authors evaluated clinical variables available on the 1st day of hospitalization for relevance to outcome.17 Predictive scores then were built by adding clinical variables.17 Higher scores were associated with decreased survival.17 This system, however, took into account various clinicopathologic data (eg, CBC, serum biochemistry, appetite, urinary abnormalities, and diarrhea frequency) and was not based on SCr or urine output alone.17 Serum creatinine concentration has been shown to be a predictor of mortality in many dierent disease processes in dogs, including heat stroke, sepsis, gastric dilatation and volvulus, dilated cardiomyopathy, chronic valvular disease, endocarditis, babesiosis, and dogs undergoing biliary surgery.1828 Some of these studies used the criterion of an increase in SCr by 0.5 mg/dL as the denition of acute kidney injury and others used absolute values, such as SCr concentration >2.5 mg/dL, for example.1828 We initially planned on including 6 levels of AKI (0-5); however, because of the low number of animals developing higher levels of AKI, we elected to collapse levels 2 through 5 into a single level (level 2). Overall, 816% of the dogs and cats included in this study developed level 2 AKI during the sampling period. This is similar to the incidence of AKI in the study of Thoen et al (14.6%), and the incidence in various studies of human patients (950%).2,3,8 This may represent an underestimation of the true incidence of AKI because many animals do not have serial evaluations of SCr or urine output. In addition, the incidence of AKI may be inuenced by inpatient versus outpatient status. The likelihood of AKI associated mortality in our study was similar to that in studies of human patients. A study of human patients using the AKIN criteria of Thakar et al reported an OR of 2.2 for those placed in stage I, an OR of 6.1 for those placed in stage II, and an OR of 8.7 for those placed in stage III.29 Our results were distributed similarly (as described in Table 3), but not as dramatically. This may be attributable to a dierence in our patient distribution in that all the human patients were admitted to the ICU. Reasons for lower numbers of patients placed in the higher levels include the possibility that the patients that would have had a higher second SCr determinations may not have survived to have a second blood sampling. In addition, some patients with lower muscle mass may have been underestimated with regard to severity of AKI if increases in SCr were not as dramatic as in an animal with normal or high muscle mass. Sampling conditions and illness severity were not standardized because patients were hospitalized in any

ward or evaluated on an outpatient basis, which represents a limitation of this study. Fasting state and hydration status were not recorded. The development of an increase in SCr in response to volume depletion (eg, associated with diuretic administration) may not be associated with AKI, and instead may represent an appropriate response of the kidney to the hemodynamic state. However, the development of transient azotemia (dened as resolving within 3 days) has an impact on mortality in people.3 The retrospective nature of this study precluded evaluation of the duration of azotemia. Inuence of breed or muscle mass was not taken into consideration with SCr determinations. Greyhounds have higher SCr than other breeds, likely because of higher muscle mass, and patients with lower muscle mass have lower SCr.30 The impact of these variations should be minimal because levels were determined based on a change from a normal initial SCr. Urine output, which may have provided a more thorough assessment of kidney function, was not included because of inconsistent quantitation of urine output. In summary, the extent of increase in SCr is predictive of mortality in dogs and cats. Dogs and cats have decreased survival time if placed in level 2 within 2, 3 or 7 days and in dogs if placed in level 2 within 2 days. Prospective studies are needed to determine the usefulness of this grading system if assessed exclusively in critical patients.

Footnote
a

Olympus AU 400, Olympus America, Center Valley, PA

Acknowledgments
Conict of Interest Declaration: Animerge is the current employer of Dr Emily Harison. She no longer works at the Animal Medical Center after nishing her residency. Dr Kenneth Lamb of Lamb Consulting is a private statistician who calculated statistics.

References
1. Hoste EA, Kellum JA. Incidence, classication, and outcomes of acute kidney injury. Contrib Nephrol 2007;156:3238. 2. Hoste EA, Clermont G, Goldstein D, et al. RIFLE criteria for acute kidney injury are associated with hospital mortality in critically ill patients: A cohort analysis. Crit Care Med 2006;10: R37. E pub. 3. Uchino S, Bellomo R, Bagshaw S, et al. Transient azotemia is associated with high risk death in hospitalized patients. Nephrol Dial Transplant 2010;25:18331839. 4. Mehta RL, Kellum JA, Shah SV, et al. Acute kidney injury network: Report of an initiative to improve outcomes in acute kidney injury. Crit Care 2007;11:R31. 5. Lopes J, Fernandes P, Jorge S, et al. Acute kidney injury in intensive care unit patients: A comparison between the RIFLE and the acute kidney injury network classications. Crit Care 2008;12:R110. E pub.

Harison et al 19. Sykes JE, Kittleson MD, Chomel BB, et al. Clinicopathologic ndings and outcome in dogs with infective endocarditis: 71 cases (19922005). J Am Vet Med Assoc 2006;228:17351747. 20. Nicolle AP, Chetboul V, Allerheiligen T, et al. Azotemia and glomerular ltration rate in dog with chronic valvular disease. J Vet Intern Med 2007;21:943949. 21. Amsellem PM, Seim HB, MacPhail CM, et al. Long-term survival and risk factors associated with biliary surgery in dogs: 34 cases (19942004). J Am Vet Med Assoc 2006;229: 14511457. 22. Beck JJ, Staatz AJ, Pelsue DH, et al. Risk factors associated with short-term outcome and development of perioperative complications in dogs undergoing surgery because gastric dilatation volvulus: 166 cases (19922003). J Am Vet Med Assoc 2006;229:19341939. 23. Buber T, Saragusty J, Ranan E, et al. Evaluation of lidocaine treatment and risk factors for death associated with gastric dilatation and volvulus in dogs: 112 cases (19972005). J Am Vet Med Assoc 2007;230:13341339. 24. Aroch I, Segev G, Loeb E, Bruchim Y. Peripheral nucleated red blood cells as a prognostic indicator in heatstroke in dogs. J Vet Intern Med 2009;23:544551. 25. Kenney EM, Rozanski RE, Rush JE, et al. Association between outcome and organ system dysfunction in dogs with sepsis: 114 cases (20032007). J Am Vet Med Assoc 2010;236:8387. 26. Camacho AT, Guitian EJ, Pallas E, et al. Azotemia and mortality among Babesia microti-like infected dogs. J Vet Intern Med 2004;18:141146. 27. Bruchim Y, Klement E, Saragusty J, et al. Heat stroke in dogs: A retrospective study of 54 cases (19992004) and analysis of risk factors for death. J Vet Intern Med 2006;20:3846. 28. Maddens B, Heiene P, et al. Evaluation of kidney injury in dogs with pyometra based on proteinuria, renal histomorphology, and urinary biomarkers. J Vet Intern Med 2011; 25:10751083. 29. Thakar CV, Christianson A, et al. Incidence and outcomes of acute kidney injury in intensive care units: A Veterans administration study. Crit Care Med 2009;37:25522558. 30. Feeman WE, Couto CG, Gray TL. Serum creatinine concentration in retired racing greyhounds. Vet Clin Pathol 2003;32:4042.

6. Hoste EA, Kellum JA. RIFLE criteria provide robust assessment of kidney dysfunction and correlate with hospital mortality. Crit Care Med 2006;34:20162017. 7. Ostermann M, Chang R, Riyadh ICU Program Users Group. Correlation between the AKI classication and outcome. Crit Care 2008;12:R144. E pub. 8. Thoen ME, Kerl ME, et al. Characterization of acute kidney injury in hospitalized dogs and evaluation of a veterinary acute kidney injury staging syste. Vet Emerg and Crit Care 2011;21:648657. 9. Cowgill LD, Langston CE. Acute kidney insuciency. In: Bartges J, Polzin D, eds. Nephrology and Urology of Small Animals. West Sussex: Blackwell Publishing; 2011:472532. 10. Abosaif NY, Tolba YA, Heap M, et al. The outcome of acute renal failure in the intensive care unit according to RIFLE: Model application, sensitivity, and predictability. Am J Kidney Dis 2005;46:10381048. 11. Morgera S, Schneider M, Neumayer HH. Long-term outcomes after acute kidney injury. Crit Care Med 2008;36: S193S197. 12. Cruz D, Ricci Z, Ronco C. Clinical review: RIFLE and AKIN time for reappraisal. Crit Care 2009;13:211. Available as E-Pub: http://www.ncbi.nlm.nih.gov/pmc/articles/PMC2717405/? tool=pubmed. 13. Wu VC, Huang TM, Lai CF, et al. Acute-on-chronic kidney injury at hospital discharge is associated with long-term dialysis and mortality. Kidney Int 2011;80:12221230. 14. Chang C, Lin C, Tian YC, et al. Acute kidney injury classication: Comparison of AKIN and RIFLE criteria. Shock 2010;33:247252. 15. Behrend EH, Grauer GF, et al. Hospital-acquired acute renal failure in dogs: 29 cases (19831992). J Am Vet Med Assoc 1996;208:537541. 16. Lee YJ, Chang CC, Chan JP, et al. Prognosis of acute kidney injury in dogs using RIFLE (risk, injury, failure, loss, and end-stage renal failure)-like criteria. Vet Rec 2011;168:264. 17. Segev G, Kass PH, Francey T, Cowgill LD. A novel clinical scoring system for outcome prediction in dogs with acute kidney injury managed by hemodialysis. J Vet Intern Med 2008;22:301308. 18. Martin MW, Staord Johnson MJ, Strehlau G, King JN. Canine dilated cardiomyopathy: A retrospective study of prognostic ndings in 367 clinical cases. J Small Anim Pract 2010;51:428436.