�������� ����� ��

Early reversible acute kidney injury is associated with improved survival inseptic shock

Manish M. Sood, Leigh Anne Shafer, Julie Ho, Martina Reslerova, GregMartinka, Sean Keenan, Sandra Dial, Gordon Wood, Claudio Rigatto, AnandKumar

PII: S0883-9441(14)00140-3DOI: doi: 10.1016/j.jcrc.2014.04.003Reference: YJCRC 51494

To appear in: Journal of Critical Care

Received date: 24 November 2013Revised date: 11 March 2014Accepted date: 9 April 2014

Please cite this article as: Sood Manish M., Shafer Leigh Anne, Ho Julie, ReslerovaMartina, Martinka Greg, Keenan Sean, Dial Sandra, Wood Gordon, Rigatto Claudio,Kumar Anand, Early reversible acute kidney injury is associated with improved survivalin septic shock, Journal of Critical Care (2014), doi: 10.1016/j.jcrc.2014.04.003

This is a PDF file of an unedited manuscript that has been accepted for publication.As a service to our customers we are providing this early version of the manuscript.The manuscript will undergo copyediting, typesetting, and review of the resulting proofbefore it is published in its final form. Please note that during the production processerrors may be discovered which could affect the content, and all legal disclaimers thatapply to the journal pertain.

ACC

EPTE

D M

ANU

SCR

IPT

ACCEPTED MANUSCRIPT

Early reversible acute kidney injury is associated with improved

survival in septic shock

(Running head: Reversible AKI improves survival)

Manish M Sood1, Leigh Anne Shafer

2, Julie Ho

2, Martina Reslerova

2, Greg Martinka

3, Sean

Keenan4, Sandra Dial

5, Gordon Wood

6, Claudio Rigatto

2 and Anand Kumar

7 for the Cooperative

Antimicrobial Therapy in Septic Shock (CATSS) Database Research Group.

1Ottawa Hospital Research Institute, University of Ottawa, Ottawa, Canada

2Section of Nephrology, University of Manitoba, Manitoba, Canada

3Richmond General Hospital, Vancouver, British Columbia, Canada

4Royal Columbian Hospital, Vancouver, British Columbia, Canada

5McGill University, Montreal Quebec, Canada

6Royal Jubilee Hospital/Victoria General Hospital, Victoria British Columbia, Canada

7Section of Critical Care Medicine, University of Manitoba, Manitoba, Canada

Word Count: Abstract 365, Body 3330, Figures 6, Tables 2

References: 36

All authors approved of this manuscript.

There are no conflicts of interest.

Corresponding Author:

Anand Kumar

Health Sciences Centre

JJ399

700 William Ave

Winnipeg MB

R3A-1R9

ACC

EPTE

D M

ANU

SCR

IPT

ACCEPTED MANUSCRIPT

Abstract

Introduction: The fact that acute kidney injury (AKI) is associated with worse clinical outcomes

forms the basis of most AKI prognostic scoring systems. However, early reversibility of renal

dysfunction in acute illness is not considered in such systems. We sought to determine whether

early (≤24 hrs after shock documentation) reversibility of AKI was independently associated

with in-hospital mortality in septic shock.

Methods: Patient information was derived from an international database of septic shock cases

from 28 different institutions in Canada, the United States and Saudi Arabia. Data from a final

cohort of 5443 patients admitted with septic shock between Jan 1996 and Dec 2009 was

analyzed. The following four definitions were used in regards to AKI status: 1) reversible AKI =

AKI of any RIFLE severity prevalent at shock diagnosis or incident at 6 hours post-diagnosis

that reverses by 24 hours 2) persistent AKI = AKI prevalent at shock diagnosis and persisting

during the entire 24 hours post-shock diagnosis, 3) new AKI = AKI incident between 6-24 hours

post-shock diagnosis and 4) improved AKI = AKI prevalent at shock diagnosis or incident at 6

hrs post followed by improvement of AKI severity across at least one RIFLE category over the

first 24 hours. Cox proportional hazards were used to determine the association between AKI

status and in-hospital mortality.

Results: During the first 24 hours, reversible AKI occurred in 13.0%, persistent AKI in 54.9%,

new AKI in 11.7% and no AKI in 22.4%. In adjusted analyses, reversible AKI was associated

with improved survival (HR 0.64 95%CI 0.53-0.77) compared to no AKI (referent), persistent

AKI (HR 0.99 95% CI 0.88-1.11) and new AKI (HR 1.41 95% CI 1.22-1.62). Improved AKI

ACC

EPTE

D M

ANU

SCR

IPT

ACCEPTED MANUSCRIPT

occurred in 19.1% with improvement across any RIFLE category associated with a significant

decrease in mortality (0.53 (95% CI 0.45-0.63). More rapid antimicrobial administration, lower

Apache II score, lower age, and a smaller number of failed organs (excluding renal) on the day

of shock as well as community-acquired infection were independently associated with reversible

AKI.

Conclusion: In septic shock, reversible AKI within the first 24 hours of admission confers a

survival benefit compared to no, new, or persistent AKI. Prognostic AKI classification schemes

should consider integration of early AKI reversibility into the scoring system.

ACC

EPTE

D M

ANU

SCR

IPT

ACCEPTED MANUSCRIPT

Introduction

Acute kidney injury (AKI) is associated with adverse outcomes, universally increasing

mortality, length of hospital stay and the risk of long term chronic kidney disease and kidney

failure (1-10). In septic shock, AKI is especially common with the risk of worse outcomes

increasing with the severity of injury (2, 4, 11-16).

Current classification schemes for defining AKI, such as the Risk, Injury, Failure, Loss of

kidney function, and End-stage kidney disease (RIFLE) system and the Acute Kidney Injury

Network (AKIN) system define the stage and severity of AKI using criteria based on declining

urine output and changes in serum creatinine compared to baseline (17-21). AKI is then

classified according to the most severe stage achieved at any time point, regardless of

reversibility. Although this classification has been validated in large international AKI datasets,

little is known regarding the impact of earl (≤24 hours) reversibility of AKI on outcomes(17-19).

In this retrospective analysis, we examined the effect of early reversibility of AKI on in-hospital

mortality in septic shock.

Methods

Study Population

The Cooperative Antimicrobial Therapy of Septic Shock (CATSS) database is an

international, multicenter database of patients admitted with septic shock to an intensive care

unit. The CATSS database, which has been described in detail previously, uses standardized case

definitions and includes repeat serum creatinine measurements during the first 24 hours of

admission (13, 22, 23). The database captures information on consecutive adult (> 18 years old)

patients admitted with septic shock from 28 medical institutions in Canada, the United States and

ACC

EPTE

D M

ANU

SCR

IPT

ACCEPTED MANUSCRIPT

Saudi Arabia. Patients from discrete periods between Jan. 1996 and Dec. 2008 were screened

and included if they meet the criteria for septic shock as defined by the ACCP/SCCM consensus

conference guidelines (N=7,390) (23, 24). For the current study, we excluded patients who had a

history of chronic kidney disease (N=566) or had dialysis therapy prior to ICU admission

(N=563). Patients without serum creatinine levels at each assessment time point (N=818) were

also excluded. This left 5,443 patients in the study population (Figure 1). Chronic kidney disease

(CKD) was pre-defined at database creation as a stable creatinine > 160 µmol/L (1.5 X normal)

prior to shock occurrence. Dialysis status on admission was defined as the need for renal

replacement therapy (either peritoneal or hemodialysis) as an outpatient immediately prior to

hospitalization. This study was approved by the Health Research Ethics Board at the University

of Manitoba and all participating institutions.

Data Collection

Data collection definitions and methodology have been outlined previously (22, 23). All

patients required vasopressor therapy for at least 3 hours. Trained research personnel

prospectively collected data on patient demographics, co-morbidities, physiological

characteristics, ICU treatments and ICU and in-hospital outcomes. Serum creatinine

measurements were taken at baseline (N= 7,390), approximately 6 hours (range 4-8, N=6385)

and approximately 24 hours (range 20-28, N=6971). Acute Physiology and Chronic Health

Evaluation (APACHE) II scores were calculated based on the most aberrant values within 24

hours of the diagnosis of septic shock(25). Similarly, the number of organ failures was inclusive

of occurrence with the first 24 hours after the diagnosis (ie Day 1).

ACC

EPTE

D M

ANU

SCR

IPT

ACCEPTED MANUSCRIPT

Definitions

AKI was defined and classified according to the RIFLE criteria. As pre-ICU creatinine values

were not consistently available in our study cohort, we assumed a baseline renal function for all

patients after excluding those with a history of CKD and dialysis (based on hospital, clinic and

external records). Patients without a history of CKD or dialysis were assumed to have a baseline

eGFR of 75 mL/min/m2. We calculated the eGFR using the Modification of Diet, in Renal

Disease (MDRD) formula(26) for all patients using the creatinine at ICU admission and at 6 and

24 hours post ICU admission. AKI was then determined using RIFLE criteria for eGFR

changes. As urine output changes were not captured in our cohort, only eGFR-based changes

were used in the determination of AKI. Patients were classified as AKI if they experienced the

injury upon any measure of renal function.

The following four definitions were used in regards to AKI status: 1) reversible AKI =

AKI of any RIFLE severity prevalent at shock diagnosis or incident at 6 hours post-diagnosis

that reverses by 24 hours 2) persistent AKI = AKI prevalent at shock diagnosis and persisting

during the entire 24 hours post-shock diagnosis, 3) new AKI = AKI incident between 6-24 hours

post-shock diagnosis and 4) improved AKI = AKI prevalent at shock diagnosis or incident at 6

hrs post followed by improvement of AKI severity across at least one RIFLE category over the

first 24 hours. Categories were not mutually exclusive. For example if a patient upon shock

diagnosis had AKI RIFLE class failure and then at 24 hours AKI RIFLE class risk they were

categorized as both persistent AKI and improved AKI.

Outcome

The primary outcome of interest was in-hospital mortality.

Statistical Analysis

ACC

EPTE

D M

ANU

SCR

IPT

ACCEPTED MANUSCRIPT

Continuous variables of interest were summarized as mean or medians with standard

deviation or inter-quartile range as appropriate. Differences in baseline characteristics were

determined by student’s t-test or one-way ANOVA for continuous variables and chi-square for

dichotomous variables. All analyses were conducted using PASW v. 18

(www.ibm.com/SPSS_Statistics) and Stata v.11.2 (StatCorp LP).

We examined the impact of AKI status on in-hospital mortality by the Kaplan Meier

(KM) method and Cox proportional hazards model. Statistical significance was determined by

the log rank method for the KM. The assumptions of proportionality for the Cox model were

assessed by examining log-minus-log survival plots. Cox models were adjusted for

demographics (age, sex), co-morbidity (cancer, immunosuppression, , congestive heart failure,

coronary artery disease, chronic obstructive pulmonary disease, diabetes mellitus, alcohol or

intravenous recreational drug abuse, surgical status), illness severity (APACHE II score, number

of day 1 organ failures, community vs nosocomial infection), and treatment (appropriate empiric

antimicrobials administered).

We investigated whether improvements across any category of AKI severity (both AKI

reversible and AKI improved groups) impacted mortality by including any AKI improvement as

a covariate in our models. We further categorized each possible type of improvement (AKI

failure to injury, AKI failure to risk, AKI failure reversal to no AKI, AKI injury to risk, AKI

injury reversal to no AKI, AKI risk reversal to no AKI) and compared their association with

mortality to individuals with no AKI (at baseline and throughout) and those with non-improved

AKI failure, AKI injury and AKI risk.

Exploratory and Sensitivity Analyses

ACC

EPTE

D M

ANU

SCR

IPT

ACCEPTED MANUSCRIPT

With the use of an imputed baseline creatinine of 75 mls/min/m2, there is a significant

potential of misclassifying individuals with CKD as AKI. We thereby performed a series of

sensitivity analyses investigating the change in creatinine values (Δcr) between the first available

(baseline) at documented onset of hypotension and 6 hours (Δcr6) or 24 hours after hypotension

documentation (Δcr24). All models were adjusted for demographics (age, sex), co-morbidity

(cancer, congestive heart failure, coronary artery disease, diabetes mellitus, surgical status),

illness severity (APACHE, number of organ failures, nosocomial infection), treatment

(appropriate antimicrobials administered) and first available creatinine value. Both Δcr6 and

Δcr24 were investigated as continuous variables and categorized for illustrative purposes.

To investigate whether time varying changes altered the impact of AKI and in-hospital

mortality, we repeated the crude and adjusted Cox proportional hazards models accounting for

repeat measures of renal function upon documented hypotension onset and after 6 and 24 hours.

There were no qualitative differences in the point estimates or statistical significance in the crude

or adjusted analyses.

In an exploratory analysis, we analyzed variables associated with reversible AKI using

backward logistic regression limiting the analysis to individuals with AKI only. Time to

initiation of appropriate antimicrobial administration (from documentation of initial hypotension)

was divided into tertiles. Utilizing unadjusted and adjusted logistic regression models, we

examined the association between reversible AKI and time to antimicrobial administration. The

model was adjusted for the previously mentioned variables.

Results:

ACC

EPTE

D M

ANU

SCR

IPT

ACCEPTED MANUSCRIPT

During the study period, 709 (13.0%) exhibited reversible AKI, 2,878 (54.9%) persistent

AKI, 635 (11.7%) new AKI and 1,221 (22.4%) no AKI. Improved AKI occurred in

1,041(19.1%) individuals. ICU and in-hospital morality occurred in 1851 (34.0%) and 2,477

(45.5%), respectively. In hospital mortality occurred in 150 (21.2%) with reversible AKI, 1524

(53.0%) with persistent AKI, 389 (61.3%) with new AKI and 414 (33.9%) with no AKI. Table 1

outlines the study characteristics stratified by AKI status. Patients with persistent AKI were older

and more likely to be female. Persistent AKI patients were more likely to have CHF, CAD and

diabetes mellitus. In general, patients with persistent AKI had higher APACHE II scores and

number of failed organs on the first day of shock. Patients with reversible AKI were more likely

to have community-acquired infection as a cause of septic shock and to have received

appropriate initial empiric antimicrobials.

In-hospital survival stratified by AKI status is presented in Figure 2. Survival was

significantly different among the groups (p<0.001). The crude hazard ratio for mortality of

reversible AKI was 0.61 (95% CI 0.51-0.73) compared to 1.92 (95% CI 1.73-2.13) with

persistent AKI, 2.09 (95% CI 1.83-2.39) with new AKI and with no AKI used as the referent.

After adjustment for demographics, co-morbidities, illness severity and initiation of microbially

appropriate empiric antimicrobial therapy, the risk of mortality for reversible AKI remained

similar (adjusted HR 0.63 95%CI 0.52-0.76) (Figure 3). Furthermore, AKI improvement across

any category was significantly associated with a survival benefit with a crude and adjusted HR of

0.44 (95% CI 0.37-0.53) and 0.53 (95% CI 0.45-0.63). The survival benefit was greatest for

patients with complete reversal of AKI (reversible AKI group) with a trend towards greater

benefit in those who completely resolved from the most severe injury (RIFLE failure resolution

HR 0.31 95% CI 0.23-0.43, injury resolution HR 0.33 95% CI 0.22-0.48, risk resolution HR 0.51

ACC

EPTE

D M

ANU

SCR

IPT

ACCEPTED MANUSCRIPT

95% CI 0.40-0.64) (Figure 4). The results were similar after adjustment for demographics, co-

morbidities, illness acuity and therapeutic interventions. In a sensitivity analysis investigating

changes in creatinine between documented hypotension onset and both 6 (Figure 5a) and 24

(Figure 5b) hours, the largest declines in creatinine were associated with greater improvements in

mortality.

Increasing antimicrobial delay, higher illness severity (APACHE and number of organ

failures) and greater age as well as nosocomial infection as the cause of septic shock were

independently associated with decreased probability of reversible AKI (Table 2). The

relationship between time to antimicrobial administration and reversible AKI is presented in

Figure 6. Reversible AKI was significantly associated with a shorter time to antimicrobial

administration in both unadjusted and adjusted models (time < 2.5 hours adjusted OR referent,

2.5 -7.8 hours adjusted OR 0.84 95%CI 0.68-1.03, > 7.8 hours adjusted OR 0.52 95%CI 0.40-

0.66 for reversible AKI).

Discussion:

In this large, international, observational cohort study of patients with septic shock

admitted to the ICU, reversibility of AKI was associated with a decreased risk of mortality.

Even without complete reversal of AKI, a survival benefit was noted with any improvement in

AKI severity. Furthermore independent of AKI classification, individuals with the largest decline

in creatinine with the first 24 hours of ICU admission demonstrated a similar survival benefit.

These findings suggest that reversibility of AKI has important prognostic implications.

ACC

EPTE

D M

ANU

SCR

IPT

ACCEPTED MANUSCRIPT

Few studies have investigated whether reversibility of AKI specifically alters outcomes

(27-29). All were retrospective, observational studies with two in diverse hospitalized

populations and another investigating reversible AKI in post-acute coronary syndrome. All

studies found reversible AKI to be associated with increased mortality compared to no AKI, a

finding contrasting our own. One potential reason for this discrepancy may be differences in the

study populations including the exclusion of patients with pre-existing CKD in our study. We

excluded patients with CKD and previous dialysis to reduce the possibility of a misclassification.

The accuracy of the RIFLE classification for predicting outcomes has been shown to

significantly improve with exclusion of patients with pre-existing CKD (27). In addition, the

cohorts involving hospitalized populations in the previous studies were heterogeneous with

differing etiologies of admission and treatments. In contrast, our study had a relatively

homogenous, well defined etiology, and as such subsequent interventions and therapies more

consistent.

Perhaps more important are differences with our study in respect to the time frame over

which AKI reversibility was assessed. The previous studies assessed reversibility over longer

time periods (48-72 hours) whereas we assessed reversibility within the first 24 hours. The

length of time required to reverse to normal renal function may be a surrogate marker of a more

severe injury such as acute tubular necrosis (ATN) or delays in treatment, factors associated with

an increased mortality (22, 30). Rapid (<24 hr) reversal of AKI denotes both adequacy of

resuscitative efforts including antimicrobials and absence of fixed renal injury such as ATN.

This may be reflected by the lower APACHE II score and fewer day 1 organ failures in this

group compared to those with new AKI or persistent AKI (Table 1).

ACC

EPTE

D M

ANU

SCR

IPT

ACCEPTED MANUSCRIPT

In our study, we found rapid reversal of AKI was significantly and consistently

associated with a decrease in mortality. In-hospital mortality increased substantially from

reversible AKI (decreased HR) through persistent AKI to new AKI (increased HR) compared to

the absence of AKI by RIFLE criteria at any point (Figure 2 and 3). In addition, the survival

benefit was apparent in patients who experienced any improvement in the severity of their AKI

(Figure 3). Among individuals with reversible AKI, there was a trend towards decreased

mortality with greater degrees of reversal (Figure 4). However, this also means that those

patients with the highest degree of AKI reversal also start with the highest degree of AKI (ie

revert from failure at presentation to normal function at 24 hours). To examine this issue more

closely, we examined changes in serum creatinine over time (Δcr) from shock presentation and

found consistent results in magnitude and effect. Sequentially larger decreases in serum

creatinine over 6 and 24 hour periods were associated with a higher probability of survival

(Figure 5a and 5b). In addition, an intriguing finding is that faster administration of appropriate

antimicrobial therapy was associated with a higher probability of early AKI reversibility (Figure

6). This could reflect covariance with early, aggressive fluid resuscitation or could indicate that

early antimicrobials are intrinsically protective.

The use of serum creatinine changes illustrates the importance of improvement in renal

function within the first 24 hours of ICU admission independent of pre-existing renal function.

These are novel observations as previous AKI literature often classifies patients based on their

most severe degree of AKI associating that with outcomes such as mortality and ESRD (2, 12,

16, 31). We have shown that even small degrees of improvement, for example from AKI RIFLE

failure to AKI RIFLE injury are associated with an improvement in mortality.

ACC

EPTE

D M

ANU

SCR

IPT

ACCEPTED MANUSCRIPT

Our data illustrates the importance of evolving changes over time in renal function and

that determining mortality solely based on the most severe degree of AKI may not be appropriate

as it would significantly overestimate the mortality risk in individuals with AKI that improves.

Currently none of the existing AKI classifications schemes such as AKIN, RIFLE and the

pending KDIGO recommendations account for reversibility of AKI, a future consideration if our

findings are validated in other cohorts (20, 21).

The etiology of reversible AKI may be multifactorial. Physiologically, rapid reversibility

may be explained by early effective resuscitation of volume depletion/decreased effective

circulating volume (pre-renal azotemia) or alleviation of renal obstruction. With respect to

terminology, we prefer reversible AKI as epidemiological studies of AKI often do not specify

distinct etiologies of AKI and distinguishing them may be problematic. Patients who experience

reversibility may have less severe illness in general or may be responding to therapies which

have been shown to improve outcomes (13, 22, 32). In our investigation, we observed that

delays in antimicrobial administration were associated with a deleterious effect and a lower

likelihood of developing reversible AKI (Figure 6). It is unlikely that all of the antimicrobials

used in septic shock are directly reno-protective. As part of the pathophysiology of septic shock,

intense vasodilation occurs secondary to a systemic inflammatory response induced by microbial

toxins and/or cytokine release (33). Given this pathophysiology, it seems more likely that early

antimicrobials slow the progression of and severity of the septic inflammatory response leading

to less severe septic shock. In addition, early appropriate antimicrobial administration may

represent a surrogate marker for early aggressive non-antimicrobial therapy elements including

fluid resuscitation and pressor initiation.

ACC

EPTE

D M

ANU

SCR

IPT

ACCEPTED MANUSCRIPT

Strengths of our study include repeat measures of renal function with the first 24 hours of

ICU admission, a large international well defined cohort and consistency of findings. The

CATSS database has well ascertained and detailed data allowing us to appropriately adjust for

comorbid conditions with little missing data.

Our results have some potential limitations. One is that our dataset was limited to septic

shock and the mortality benefits with reversible AKI may or may not be applicable to other

populations. We do not have sequential data on renal function beyond 24 hours and are therefore

unable to ascertain the occurrence of late deterioration or improvement of renal function. For

example, we do not know whether patients with reversible AKI had sustained improvement

(apart from improved survival). In addition, in this study we lacked data on urine output. This

limited our ability to diagnose AKI according to the full RIFLE criteria. However, the

importance the urine output aspect of the RIFLE criteria for AKI has recently been called in to

question as it is a poor predictor of subsequent rises in creatinine and is associated with lower

mortality than the creatinine criteria(34, 35). Another issue is that the lack of pre-ICU creatinine

values for all of the patients in the cohort may have resulted in misclassification of in some

individuals with CKD as AKI. We attempted to reduce this bias by excluding all patients with

known CKD and/or persistent pre-ICU serum creatinine elevation. However the original design

of the CATSS database defined CKD based on serum creatinine criteria, as opposed to eGFR

criteria, and was arbitrarily defined as a pre-existing serum creatinine ≥ 160 µmol/L. If no

previous history of CKD was available, they were assumed to have a baseline eGFR of 75

mL/min/m2. Use of an estimated baseline GFR of 75 mL/min/m

2 is accepted method of

assessing AKI in renal research with considerable improvement in accuracy if patients with

known CKD are excluded (36, 37). Results of our sensitivity analyses examining changes in

ACC

EPTE

D M

ANU

SCR

IPT

ACCEPTED MANUSCRIPT

serum creatinine (Figure 5) avoided this potential misclassification and the results were

consistent in magnitude and effect. None-the-less, the potential misclassifications may account in

part for the unexpectedly high mortality of the no AKI group (compared to the lower mortality to

the reversible AKI group).

Another potential limitation is the very use of RIFLE criteria for renal dysfunction in the

study. Because we have examined changes in renal function over only the initial 24 hours after

documentation of hypotension in patients with septic shock, changes in serum creatinine will be

limited by the time available for creatinine rise (either 6 or 24 hours of shock). Serum creatinine

increases at a finite rate even with anephric level kidney injury. Within 24 hours of shock onset,

creatinine rises may be limited with even with severe dysfunction. We classified patients as

RIFLE risk if the eGFR decline was >25%. However, it is well established that smaller declines

are associated with mortality(9). The no AKI group therefore is likely to be very heterogenous

in terms of actual kidney dysfunction. Regardless our observations with respect to other groups

will be unchanged despite the potential heterogeneity of the reference group.

In conclusion, our study demonstrates that reversibility of AKI is associated with

improved survival in patients with septic shock. Even partial reversibility (in those with AKI and

any improvement in severity within the first 24 hours) is associated with reduced mortality.

Importantly, delays in antimicrobial administration appear to decrease the likelihood of

reversibility in AKI. Our results illustrate the heterogeneity in the dynamics of septic AKI and

suggest that at least in septic AKI, AKI classification schemes should ideally include criteria to

account for early reversibility of AKI.

ACC

EPTE

D M

ANU

SCR

IPT

ACCEPTED MANUSCRIPT

Conflict of Interest Statement:

Anand Kumar received unrestricted grant funding from Pfizer, Lilly, Astellas, Bayer, and Merck

for the initial development of the CATSS Database. Additional grant funding has been provided

by the Manitoba Research Council, the Health Sciences Foundation and the Deacon Foundation.

No other author has significant conflict of interest. This specific analysis has not been supported.

Manish Sood has salary support through the Jindal Research Chair for the Prevention of Kidney

Disease at the University of Ottawa

ACC

EPTE

D M

ANU

SCR

IPT

ACCEPTED MANUSCRIPT

References:

1. Bagshaw S, Laupland K, Doig C, Mortis G, Fick G, Mucenski M, et al. Prognosis for long-term

survival and renal recovery in critically ill patients with severe acute renal failure: A population based

study. Crit Care. 2005;9:R700 - R9.

2. Bagshaw S, 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 -

9.

3. Bellomo R, Ronco C, Kellum J, Mehta R, Palevsky P, workgroup ADQI. 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 - R12.

4. Hoste E, Lameire N, Vanholder R, Benoit D, Decruyenaere J, Colardyn F. Acute renal failure in

patients with sepsis in a surgical ICU: predictive factors, incidence, comorbidity, and outcome. J Am Soc

Nephrol. 2003;14:1022 - 30.

5. Levy E, Viscoli C, Horwitz R. The effect of acute renal failure on mortality. A cohort analysis.

JAMA. 1996;275:1489 - 94.

6. Uchino S, Kellum JA, Bellomo R, Doig GS, Morimatsu H, Morgera S, et al. Acute Renal Failure

in Critically Ill Patients. JAMA: The Journal of the American Medical Association. 2005 August 17,

2005;294:813-8.

7. Barrantes FMD, Tian JMDMPH, Vazquez RMD, Amoateng-Adjepong YMDP, Manthous

CAMD. Acute kidney injury criteria predict outcomes of critically ill patients. Critical Care Medicine.

2008;36:1397-403.

8. Coca SG, Yusuf B, Shlipak MG, Garg AX, Parikh CR. Long-term Risk of Mortality and Other

Adverse Outcomes After Acute Kidney Injury: A Systematic Review and Meta-analysis. American

Journal of Kidney Diseases. 2009;53:961-73.

9. Coca S, Yusuf B, Shlipak M, Garg A, Parikh C. Long-term risk of mortality and other adverse

outcomes after acute kidney injury: a systematic review and meta-analysis. Am J Kidney Dis. 2009;6:961

- 73.

10. Wald R, Quinn RR, Luo J, Li P, Scales DC, Mamdani MM, et al. Chronic Dialysis and Death

Among Survivors of Acute Kidney Injury Requiring Dialysis. JAMA. 2009;302:1179-85.

11. Bagshaw S, George C, Bellomo R, Committee ADM. Early acute kidney injury and sepsis: a

multicentre evaluation. Crit Care. 2008;12:R47.

12. Bellomo R, Wan L, Langenberg C, May C. Septic acute kidney injury: new concepts. Nephron

Exp Nephrol. 2008;109:e95 - 100.

13. Bagshaw S, Lapinsky S, Dial S, Arabi Y, Dodek P, Wood G, et al. Acute kidney injury in septic

shock: clinical outcomes and impact of duration of hypotension prior to initiation of antimicrobial

therapy. Intensive Care Medicine. 2009;35:871-81.

14. Lopes J, Jorge S, Resina C, Santos C, Pereira A, Neves J, et al. Acute kidney injury in patients

with sepsis: a contemporary analysis. Int J Infect Dis. 2009;13:176 - 81.

15. Lopes JA, Fernandes P, Jorge S, Resina C, Santos C, Pereira A, et al. Long-term risk of mortality

after acute kidney injury in patients with sepsis: a contemporary analysis. BMC Nephrology. 2010;11:9.

16. Kim WY, Huh JW, Lim C-M, Koh Y, Hong S-B. Analysis of progression in risk, injury, failure,

loss, and end-stage renal disease classification on outcome in patients with severe sepsis and septic shock.

Journal of Critical Care 2014 (in Press).

17. Ostermann M, Chang RWS. Acute kidney injury in the intensive care unit according to RIFLE.

Critical Care Medicine. 2007;35:1837-43.

18. Hoste E, Clermont G, Kersten A, Venkataraman R, Angus D, 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.

ACC

EPTE

D M

ANU

SCR

IPT

ACCEPTED MANUSCRIPT

19. Lopes J, Fernandes P, Jorge S, Goncalves S, Alvarez A, Costae 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.

20. Mehta R, Kellum J, Shah S, Molitoris B, Ronco C, Warnock D, et al. Acute Kidney Injury

Network: report of an initiative to improve outcomes in acute kidney injury. Critical Care. 2007;11:R31.

21. Bellomo R, Ronco C, Kellum J, Mehta R, Palevsky P, et al. 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. Critical Care.

2004;8:R204 - R12.

22. Kumar A, Light RB, Parrillo JE, Maki DG, Simon D, Lapinsky S, et al. Early combination

antibiotic therapy yields improved survival compared with monotherapy in septic shock: A propensity-

matched analysis Critical Care Medicine. 2010;38:1773-85.

23. Kumar A, Ellis P, Arabi Y, Roberts D, Light B, Parrillo JE, et al. Initiation of Inappropriate

Antimicrobial Therapy Results in a Fivefold Reduction of Survival in Human Septic Shock. Chest. 2009

November 1, 2009;136:1237-48.

24. Bone RC BR, Cerra FB, et al:. ACCP/SCCM Consensus Conference: Definitions for sepsis and

organ failure and guidelines for use of innovative therapies in sepsis. Chest. 1992;101:1644-55.

25. Knaus WA, Draper EA. APACHE II: A severity of disease classification system. Critical Care

Medicine. 1985;13:818-29.

26. Levey AS, Bosch JP, Lewis JB, Greene T, Rogers N, Roth D. A more accurate method to

estimate glomerular filtration rate from serum creatinine: A new prediction equation. Annals of Internal

Medicine. 1999;130:461-70.

27. Goldberg A, Kogan E, Hammerman H, Markiewicz W, Aronson D. The impact of transient and

persistent acute kidney injury on long-term outcomes after acute myocardial infarction. Kidney Int.

2009;76:900-6.

28. Tian J, Barrantes F, Amoateng-Adjepong Y, Manthous CA. Rapid Reversal of Acute Kidney

Injury and Hospital Outcomes: A Retrospective Cohort Study. American Journal of Kidney Diseases.

2009;53:974-81.

29. Uchino S, Bellomo R, Bagshaw SM, Goldsmith D. Transient azotaemia is associated with a high

risk of death in hospitalized patients. Nephrology Dialysis Transplantation. 2010 June 1, 2010;25:1833-9.

30. Kiers HD, Griesdale DEG, Litchfield A, Reynolds S, Gibney RTN, Chittock D, et al. Effect of

early achievement of physiologic resuscitation goals in septic patients admitted from the ward on the

kidneys. Journal of Critical Care. 2010;25:563-9.

31. Kelly K. Distant effects of experimental renal ischemia/reperfusion injury. J Am Soc Nephrol.

2003;14:1549 - 58.

32. Rivers E, Nguyen B, Havstad S, Ressler J, Muzzin A, Knoblich B, et al. Early Goal-Directed

Therapy in the Treatment of Severe Sepsis and Septic Shock. New England Journal of Medicine.

2001;345:1368-77.

33. Schrier RW, Wang W. Acute Renal Failure and Sepsis. New England Journal of Medicine.

2004;351:159-69.

34. Joannidis M, Metnitz 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 Medicine. 2009;35:1692-702.

35. Prowle J, Liu Y-L, Licari E, Bagshaw S, Egi M, Haase M, et al. Oliguria as predictive biomarker

of acute kidney injury in critically ill patients. Critical Care. 2011;15.

36. Bagshaw S, 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 - 10.

37. Bagshaw S, Uchino S, Cruz D, Bellomo R, Morimatsu H, Morgera S, et al. A comparison of

observed versus estimated baseline creatinine for determination of RIFLE class in patients with acute

kidney injury. Nephrol Dial Transplant. 2009;24:2739 - 44.

ACC

EPTE

D M

ANU

SCR

IPT

ACCEPTED MANUSCRIPT

Figures:

Figure 1: Development of the study cohort.

Figure 2: In-hospital survival curves based on AKI status.

Improved survival is seen in those with reversible AKI (black line) compared to No AKI (dashed

line), Persistent AKI (dark grey line) and New AKI (light grey). Survival expression as a

fractional value. AKI= acute kidney injury

Figure 3: Crude and adjusted hazard ratio for in-hospital mortality.

Mortality is lower in patients with reversible AKI compared to No AKI (referent), Persistent

AKI and New AKI. Adjusted for age, sex, co-morbidities (cancer, immunosuppression, coronary

artery disease, hypertension, congestive heart failure, chronic obstructive pulmonary disease,

diabetes mellitus, surgical status, alcohol or recreational drug abuse), and illness severity and

treatment characteristics (APACHE II score, number of organ failures, use of empirically

appropriate antimicrobials, presence of nosocomial infection). AKI=acute kidney injury

Figure 4: Unadjusted hazard ratio for in-hospital mortality.

Improved survival is noted with any degree of AKI improvement with the greatest benefit in

those with reversible AKI.

AKI= acute kidney injury, HR= hazard ratio, CI=confidence interval, NO AKI = referent

Figure 5: Sensitivity analysis depicting adjusted hazard ratios for in-hospital mortality

stratified by changes in serum creatinine (Δcr) between initial documentation of

hypotension and 6 hours (a) and 24 hours (b) following documentation.

A negative value of Δcr denotes a decrease in creatinine values between admission and 6 or 24

hour measurement. Δcr -4 to 2 µmol/L = referent

All models were adjusted for demographics (age, sex), co-morbidity (cancer,

immunosuppression, congestive heart failure, coronary artery disease, diabetes mellitus, surgical

status), illness severity (APACHE II score, number of organ failures, nosocomial infection), use

of appropriate initial empiric antimicrobial therapy and first available creatinine value.

ACC

EPTE

D M

ANU

SCR

IPT

ACCEPTED MANUSCRIPT

Figure 6: The crude and adjusted odds ratio for developing reversible AKI.

Probability of reversible AKI decreases with increased delays in antimicrobial administration.

All models were adjusted for demographics (age, sex), co-morbidity (cancer,

immunosuppression, congestive heart failure, coronary artery disease, diabetes mellitus, surgical

status), illness severity (APACHE II score, number of organ failures, nosocomial infection) and

use of appropriate initial empiric antimicrobial therapy

AKI=acute kidney injury

ACC

EPTE

D M

ANU

SCR

IPT

ACCEPTED MANUSCRIPT

Fig. 1

ACC

EPTE

D M

ANU

SCR

IPT

ACCEPTED MANUSCRIPT

Fig. 2

ACC

EPTE

D M

ANU

SCR

IPT

ACCEPTED MANUSCRIPT

Fig. 3

ACC

EPTE

D M

ANU

SCR

IPT

ACCEPTED MANUSCRIPT

Fig. 4

ACC

EPTE

D M

ANU

SCR

IPT

ACCEPTED MANUSCRIPT

Fig. 5

ACC

EPTE

D M

ANU

SCR

IPT

ACCEPTED MANUSCRIPT

Fig. 6

ACC

EPTE

D M

ANU

SCR

IPT

ACCEPTED MANUSCRIPT

Table 1: Characteristics of septic shock patients stratified by AKI status.

Reversible AKI

Persistent AKI

New AKI No AKI

P value

Variable: N = 709

(13.0%)

N = 2878

(52.9%)

N = 635

(11.7%)

N = 1221

(22.4%)

Sex (%F) 41.7 47.4 38.1 37.8 <0.0001

Age 60.3±16.8 65.6±15.4 60.0±17.6 57.0±17.8 <0.0001

Co-morbidities (%):

Cancer 15.7 18.6 23.9 20.7 0.001

Immunosuppression 12.0 14.9 18.0 12.4 0.02

CHF 8.3 10.6 6.9 6.1 <0.0001

CAD 11.4 12.7 9.3 9.3 0.004

HTN 16.4 17.0 10.4 12.9 <0.0001

COPD 15.7 13.2 15.6 17.2 0.008

DM 21.4 26.0 19.7 18.3 <0.0001

Surgery 18.3 19.7 25.8 24.7 <0.0001

Alcohol abuse 15.7 12.8 16.5 17.2 0.001

ICU characteristics:

APACHE II score 21.2±6.9 26.5±7.8 24.5±7.3 20.6±6.5 <0.0001

Number of failed

organ(s)

3.4±1.4 4.3±1.5 4.1±1.5 3.3±1.4 <0.0001

Community

infection (%)

26.7 36.6 50.4 43.0 <0.0001

Appropriate

empiric antimicrobials

90.6 84.1 80.6 87.2 <0.0001

Heart rate and WBC values are most aberrant values within 24 hours of the diagnosis of septic

shock. AKI acute kidney injury, F female, BMI body mass index, AIDS acquired

immunodeficiency syndrome, CHF congestive heart failure, CAD coronary artery disease, HTN

hypertension, COPD chronic obstructive pulmonary disease, DM diabetes mellitus, ICU

intensive care unit, WBC white blood cell, % percentage, APACHE acute physiology and

chronic health evaluation; continuous data presented as mean ± standard deviation

ACC

EPTE

D M

ANU

SCR

IPT

ACCEPTED MANUSCRIPT

Table 2: Characteristics associated with reversible AKI.

Characteristics: Odds Ratio 95% Confidence Interval

Antimicrobial delay <2.5 hours (referent) 1

2.5 – 7.8 hours 0.81 0.66-1.00

> 7.8 hours 0.48 0.38-0.62

Age (per year) 0.99 0.98-1.00

APACHE II score (per point) 0.95 0.94-0.96

Number of organ failure(s) 0.87 0.82-0.93

Nosocomial infection 0.65 0.53-0.80

APACHE; Acute Physiology and Chronic Health Evaluation