aki commentary 2013

8/11/2019 AKI Commentary 2013

1/24

KDOQI Commentary

KDOQI US Commentary on the 2012 KDIGO Clinical PracticeGuideline for Acute Kidney Injury

Paul M. Palevsky, MD, 1,2 Kathleen D. Liu, MD, PhD, 3 Patrick D. Brophy, MD, 4

Lakhmir S. Chawla, MD, 5 Chirag R. Parikh, MD, PhD, 6,7 Charuhas V. Thakar, MD, 8,9

Ashita J. Tolwani, MD, 10 Sushrut S. Waikar, MD, 11 and Steven D. Weisbord, MD 1,2

In response to the recently released 2012 KDIGO (Kidney Disease: Improving Global Outcomes) clinicalpractice guideline for acute kidney injury (AKI), the National Kidney Foundation organized a group of USexperts in adult and pediatricAKI and critical care nephrology to review the recommendationsand comment ontheir relevancy in the context of current US clinical practice and concerns. The rst portion of the KDIGOguideline attempts to harmonize earlier consensus denitions and staging criteria for AKI. While the expertpanel thought that the KDIGO denition and staging criteria are appropriate for dening the epidemiology ofAKI and in the design of clinical trials, the panel concluded that there is insufcient evidence to support theirwidespread application to clinical care in the United States. The panel generally concurred with the remainderof the KDIGO guidelines that are focused on the prevention and pharmacologic and dialytic management ofAKI, although noting the dearth of clinical trial evidence to provide strong evidence-based recommendationsand the continued absence of effective therapies beyond hemodynamic optimization and avoidance ofnephrotoxins for the prevention and treatment of AKI.Am J Kidney Dis. 61(5):649-672. Published by Elsevier Inc. on behalf of the National Kidney Foundation, Inc.This is a US Government Work. There are no restrictions on its use.

Editorial, p. 686

INTRODUCTION

KDIGO (Kidney Disease: Improving Global Out-comes) is an international initiative to develop andimplement clinical practice guidelines for patients withkidney disease. In March 2012, KDIGO published itsguideline for the eva luation and management of acutekidney injury (AKI). 1 This guideline covers numeroustopics, including thedenition and classication ofAKI,the prevention and treatment of AKI in general withspecic recommendations for the prevention of contrast-inducedAKI, and the management of renal replacementtherapy (RRT) in patients with AKI. Because interna-tional guidelines need to be adapted for the UnitedStates, the National Kidney FoundationKidney DiseaseOutcomes Quality Initiative (NKF-KDOQI) convened amultidisciplinary work group with expertise in adult andpediatric nephrology and critical care medicine to com-ment on the applicability and implementation of theKDIGO AKI guideline in the United States. This com-mentary provides a summary of the KDIGO recommen-dation statements along with the supporting rationalesand comments on their applicability to clinical practicein theUnitedStates. The KDOQI WorkGroup congratu-lates KDIGO, the members of the AKI Guideline Work Group, and theevidence review team for producingsucha comprehensive document and believes that this guide-line will be of great value to health professionals and willadvance both current clinical care of patients with AKIand future clinical research.

AKI represents the sudden loss of kidney function,generally occurring over the course of hours to daysand resulting in the retention of metabolic wasteproducts and dysregulation of uid, electrolyte, andacid-base homeostasis. During the past decade, thisacute loss of kidney function, previously referred to asacute renal failure, has been the subject of signicantre-examination, with increased recognition of the impor-

Originally published online March 18, 2013From the 1 Renal Section, VA Pittsburgh Healthcare System;

2 Renal-Electrolyte Division, Department of Medicine, Universityof Pittsburgh School of Medicine, Pittsburgh, PA; 3 Nephrology Division, Department of Medicine, University of California, SanFrancisco, CA; 4 Nephrology Division, Department of Pediatrics,University of Iowa Carver College of Medicine, Iowa City, IA;5 Department of Anesthesiology and Critical Care Medicine and Division of Renal Diseases and Hypertension, Department of Medicine, George Washington University, Washington, DC; 6 Re-nal Section, VA Connecticut Healthcare System, West Haven;7 Section of Nephrology, Department of Medicine, Yale UniversitySchool of Medicine, New Haven, CT; 8 Renal Section, CincinnatiVA Medical Center; 9 Division of Nephrology and Hypertension, Department of Internal Medicine, University of Cincinnati, Cincin-nati, OH; 10 Division of Nephrology, Department of Medicine,University of Alabama at Birmingham, Birmingham, AL; and 11 Renal Division, Department of Medicine, Brigham and Womens Hospital, Boston, MA.

Address correspondence to Paul M. Palevsky, MD, Rm 7E123(111F-U), VA Pittsburgh Healthcare System, University Dr, Pitts-burgh, PA 14240. E-mail: [email protected]

Published by Elsevier Inc. on behalf of the National KidneyFoundation, Inc. This is a US Government Work. There are norestrictions on its use.

0272-6386/$0.00http://dx.doi.org/10.1053/j.ajkd.2013.02.349

Am J Kidney Dis. 2013;61(5):649-672 649
mailto:[email protected]:[email protected]://dx.doi.org/10.1053/j.ajkd.2013.02.349http://dx.doi.org/10.1053/j.ajkd.2013.02.349http://dx.doi.org/10.1053/j.ajkd.2013.02.349mailto:[email protected]


2/24

tance of relatively small changes in kidney fun ctio n onboth short- and longer term clinical outcomes. 2-6 Thishas resulted in the change in terminology from acuterenal failure, for which the focus generally was limitedto the most severe episodes with complete or near-complete loss of kidney function, to thecurrent terminol-ogy ofAKI, with i ncreased focus on smaller decrementsin kidney function. 7-9

AKI may developin a wide variety of settings, includ-ing in ambulatory outpatients, hospitalized patients, and,in particular, critically ill patients, for whom AKI repre-sentsa common complicationof both underlyingdiseaseand treatment.AKIis associated withsubstantial morbid-ity and mortality. For example, severe AKI occurs in

5% of critically ill patie nts and is associated withmortality rates of 40%-70%. 10-12 Although recovery of kidney functionoccurs in themajorityof patients surviv-ing an episode of AKI, many patients remain dialysisdependent or are left with severe renal impairment. Morerecently, it has been recognized that even patients whohave complete or near-complete recovery of kidneyfunction are at increased risk of progressive chronickidney disease (CKD) and that superimposition of AKI

on CKD is associated with acc elerat ion in the rate of progression to end-stage disease. 13-17

Our understanding of the epidemiology of AKI andinterpretation of results across clinical trials has beenhindered by the prior absence of a broadly acceptedclinical denition, with more than 30 o perationaldenitions of AKI used in published studies. 18 Duringthe past decade, there has been a considerable effort toforge a consensus denition. The rst attempt atdeveloping a consensus denition, known as theRIFLE criteria, was developed by the Acute Dia lysisQuality Initiative (ADQI) in 2002 ( Table 1 ).19 Thisdenition considered 3 strata of severity (risk, injury,and failure) based on the magnitude of increase inserum creatinine level and/or the duration of oliguria,as well as 2 outcome stages (loss of kidney functionand end-stage kidney disease). The risk, injury, andfailure categories were constructed to provide grada-tions in severity of kidney dysfunction, with greatersensitivity associated with risk and greater specicitywith failure. The RIFLE criteria subsequently weremodied by theAKI Network (AKIN) by the additionof an absolute increase in serum creatinine level 0.3

Table 1. RIFLE and AKIN Criteria for Diagnosis and Classication of AKI

RIFLE

Urine Output(common to both)

AKIN

Class SCr a Stage SCr b

Risk Increased SCr to 1.5 baseline Urine output 0.5 mg/kg/h

for 6 h

1 Increase in SCr 0.3 mg/dL or

increase in SCr to 150%-200% of baseline

Injury Increased SCr to 2 baseline Urine output 0.5 mg/kg/hfor 12 h

2 Increase in SCr to 200%-300% of baseline

Failure Increased SCr to 3 baseline; or anincrease of 0.5 mg/dL to a valueof 4 mg/dL

Urine output 0.3 mg/kg/hfor 12 h or anuria for

12 h

3 Increase in SCr to 300% ofbaseline; or to 4 mg/dLwith an acute increase of

0.5 mg/dL; or on RRTLoss Need for RRT for 4 wkEnd Stage Need for RRT for 3 mo

Abbreviations: AKI, acutekidney injury; AKIN,AcuteKidney Injury Network; RIFLE, risk, injury, failure, loss, end-stage disease; RRT,renal replacement therapy; SCr, serum creatinine.

a For RIFLE, the increase in SCr should be both abrupt (within 1-7 days) and sustained ( 24 hours).b For AKIN, the increase in SCr must occur in less than 48 hours.

Table 2. Pediatric Modied RIFLE (pRIFLE) Criteria for Diagnosis and Classication of AKI in Children

Class eCCr Urine Output

Risk eCCr decrease by 25% Urine output 0.5 mL/kg/h for 8 hInjury eCCr decrease by 50% Urine output 0.5 mL/kg/h for 16 hFailure eCCr decrease by 75%; or eCCr 35 mL/min/1.73 m 2 Urine output 0.3 mL/kg/h for 12 h; or anuria for 12 hLoss Persistent failure for 4 wkEnd Stage Persistent failure for 3 mo

Abbreviations and denitions: AKI, acute kidney injury; eCCr: estimated creatinine clearance using the Schwartz formula; RIFLE,risk, injury, failure, loss, end-stage disease.

Adapted and reproduced from Akcan-Arikan et al, 20 with permission of MacMillan Publishers Ltd.

Am J Kidney Dis. 2013;61(5):649-672650

Palevsky et al


3/24

mg/dL to the denition of AKI, a shortening of thetime for the increase in serum creatinine level from 7days to no more tha n 48 hou rs, and elimination of the2 outcome criteria ( Table 1 ).7,8 A modication of theRIFLE criteria for use in pediatri c patients (pRIFLE)has also been developed ( Table 2 ).20 Validation stud-

ies using these denitions have demonstrated in-creased mortality risk associated with progressivelymore severe stages of AKI. The KDIGO AKI guide-line builds upon these earlier efforts in dening AKI,with a full section of the guideline devoted to thedenition of AKI.

The development of successful therapeutic strate-gies for the prevention and treatment of most forms of AKI has been disappointing. Although numerousagents have shown promise in experimental models,none has demonstrated utility in clinical care. Aparticular area of focus has been in the prevention of contrast-induced AKI. This common cause ofAKI hasbeen the focus of multiple preventative interventionsbecause individuals at high risk of contrast-inducedAKI can be readily identied and the timing of exposure can be predetermined, allowing an opportu-nity for intervention. In the absence of effective phar-macologic therapy, the management of establishedAKI is predominantly supportive care, with the use of RRT in severe AKI. There has been tremendousadvancement in the technology and modalities avail-able for providing RRT; however, the optimal ap-proach to management of RRT in this setting remainscontroversial. The KDIGO AKI guideline providesspecic recommendations, based on the current litera-ture, for best practices in the prevention and manage-ment of AKI. In this KDOQI commentary, we haveattempted to place this international guideline in thecontext of care practices in the United States. How-ever, our commentary should be viewed in conjunc-tion with the full KDIGO document when makingclinical decisions.

REVIEW AND APPROVAL PROCESS FORTHIS COMMENTARY

This commentary was developed by a Work Groupconvened by the NKF-KDOQI, beginning with se-lecting Co-Chairs by the KDOQI steering commit-tee and individual members selected based on theirclinical expertise and interest in the guideline pro-cess. Teleconferences took place during 2011 and2012 to determine the specic areas for focus forthis commentary. Individual sections focusing oneach of the topical areas of the KDIGO AKI guide-line were drafted by groups of coauthors based ondetailed review of the particular KDIGO chaptersupplemented by additional literature review, asneeded. Because this was a commentary, no specic

voting occurred; rather, consensus among coau-thors was achieved through discussion. The docu-ment was reviewed and approved by all coauthorsand by the KDOQI leadership.

ASSESSMENT OF GUIDELINE QUALITY

The KDIGO AKI guideline contains 87 individualrecommendations, of which 26 (30%) are ungraded,39 (45%) are level 2 recommendations, and only 22(25%) are level 1 recommendations, reecting therelative paucity of high-level data guiding the manage-ment of AKI ( Fig 1 ). In addition, many of the level 1recommendations advise against the use of specicagents or therapeutic interventions.

All members of the Work Group completed theAGREE (Appraisal of Guideline for Research and

Evaluation) II instrument to assess the quality of theKDIGO AKI guideline. 21-23 Mean Work Group scoresfor each of the 23 elements and scaled domain scoresfor the 6 domains are provided in Table 3 , along withthe scoring range for domain and overall scores. Thedomain scores ranged from 0.53 for applicability to0.85 for editorial independence. The mean overallscore provided was 5.3 1.0 on a 7-point scale. AllWork Group members recommended the KDIGOAKI guideline for use, although the majority sug-gested that modications could improve the guide-line.

Figure 1. Strength of recommendation and level of evi-dence of the KDIGO Clinical Practice Guideline for Acute

Kidney Injury recommendations. Level 1 corresponds to arecommendation statement of we recommend; Level 2, to astatement of we suggest; Not Graded was used to provideguidance based on common sense or when the topic does notallow adequate application of evidence. The quality of support-ing evidence is graded from A to D , with letter grades corre-sponding to high, moderate, low, and very low quality ofevidence, respectively.

Am J Kidney Dis. 2013;61(5):649-672 651

KDOQI Commentary


4/24

REVIEW OF KDIGO AKI RECOMMENDATIONS

DenitionandClassicationof AKICommentary Overview . The rst group of recommendation state-

ments in the KDI GO A KI Guideline addresses thedenition of AKI ( Box 1 ). Of these 13 recommenda-tion statements, 12 are not graded. The KDIGO Work Group began by dening AKI b y harmonizing theprior RIFLE and AKIN criteria. 7,19 In the RIFLE

criteria ( Table 1 ), AKI was dened based on a 50%increase in serum creatinine level occurring over 1-7days or the presence of oliguria for more than 6 hours.The AKIN criteria added an absolute increase inserum creatinine level of 0.3 mg/dL and reduced thetimeframe for the increase in serum creatinine level to48 hours ( Table 1 ). The KDIGO Work Group harmo-nized these 2 denitions, keeping the absolute in-crease in serum creatinine level of 0.3 mg/dL within48 hours from the AKIN denition, but returning to

Table 3. KDOQI Work Groups Appraisal of the KDIGO AKI Guideline Using the AGREE II Instrument

Domain Element Score aScaled Domain

Score b

Scope and purpose 1. The overall objective(s) of the guideline is (are) specically described. 6.0 0.9 0.832. The health question(s) covered by the guideline is (are) specically

described.

6.1 0.8

3. The population (patients, public, etc) to whom the guideline is meantto apply is specically described.

5.8 1.3

Stakeholder involvement 4. The guideline development group includes individuals from allrelevant professional groups.

5.7 1.6 0.58

5. The views and preferences of the target population (patients, public,etc) have been sought.

2.9 1.8

6. The target users of the guideline are clearly dened. 4.9 1.1Rigor of development 7. Systematic methods were used to search for evidence. 6.0 1.0 0.76

8. The criteria for selecting the evidence are clearly described. 6.1 0.99. The strengths and limitations of the body of evidence are clearly

described.6.1 0.9

10. The methods for formulating the recommendations are clearly

described.

5.9 0.9

11. The health benets, side effects, and risks have been considered informulating the recommendations.

5.6 1.5

12. There is an explicit link between the recommendations and thesupporting evidence.

5.4 0.9

13. The guideline has been externally reviewed by experts prior to itspublication.

4.9 2.1

14. A procedure for updating the guideline is provided. 4.4 2.2Clarity of presentation 15. The recommendations are specic and unambiguous. 4.9 0.9 0.73

16. The different options for management of the condition or healthissue are clearly presented.

5.1 1.5

17. Key recommendations are easily identiable. 6.1 0.8Applicability 18. The guideline describes facilitators and barriers to its application. 4.6 1.8 0.53

19. The guideline provides advice and/or tools on how therecommendations can be put into practice. 4.1 1.5

20. The potential resource implications of applying therecommendations have been considered.

4.4 1.3

21. The guideline presents monitoring and/or auditing criteria. 3.7 1.9Editorial independence 22. The views of the funding body have not inuenced the content of the

guideline.6.1 0.9 0.85

23. Competing interests of guideline development group members havebeen recorded and addressed.

6.1 0.9

Abbreviation: AGREE, Appraisal of Guideline for Research and Evaluation.a Given as mean standard deviation. Individual element statement is scored on a 7-point scale where 1 represents strongly

disagree and 7 represents strongly agree.b Domain scores are reported on a scale from 0-1, where a score of 0 would represent all reviewers providing the minimum possible

score for each element in the domain and a score of 1 indicating that allreviewers gave the highest possible score for allelements in the

domain.

Am J Kidney Dis. 2013;61(5):649-672652

Palevsky et al


5/24

the 7-day timeframe for the 50% increase in serumcreatinine level. With regard to pediatric AKI, theKDIGO denition did not adopt the smaller changesin estimated creatinine clearance proposed in thepRIFLE criteria. Rather, the KDIGO Work Groupsuggested that the absolute change in creatinine levelof 0.3 mg/dL that is part of the KDIGO AKIdenition would capture most pediatric AKI. TheKDIGO denition considers a decrease in estimatedglomerular ltration rate (eGFR) to 35 mL/min/ 1.73 m 2 to be stage 3 AKI in pediatric patients, asinitially proposed in the pRIFLE criteria. 20

Overall, while our KDOQI Work Group concurredwith the efforts on the partof KDIGO to meld the RIFLEand AKIN criteria into a uniform denition of AKI, wehave signicant reservations regarding the applicabilityof this denition and staging system to clinical practice.The RIFLE and AKIN denitions and staging systemshave substantially enhanced theability to conduct epide-miologic studies, and we support the use of the KDIGOAKI guideline denition and staging strategy for futureepidemiologic research until novel biomarkers facilitate

the development of better criteria. We concur that therigorous application of a standard denition will permitepidemiologic comparisons across populations and overtime. In addition, the denitions as described have valuein standardization of entry criteria and endpoints inclinical trials of AKI. However, we question whetherthese criteria are currently appropriate to guide clinicalmanagement of adult patients.

The successive stages of a serum creatininebasedAKI denition reect a tradeoff between sensitivity andspecicity. At low stages, sensitivity is high (ie, almostevery patient with true AKI is identied), but specicityis low (ie, many patients without true AKI are misidenti-ed as having AKI). As part of the decision to recom-mend a cutoff value based on a biomarker such as serumcreatinine or physiologic variable such as urine output,the implications of the cutoff need to be carefullyconsid-ered. We thought that particularly in the United States,where defensive medicine is common due to the highlylitigious environment in which many physicians prac-tice, the KDIGO denition may result in a dramaticincrease in nephrology consultations and have the unin-

Box 1. Summary of KDIGO Recommendation Statements: Denition of AKI.

2.1.1: AKI is dened as any of the following ( Not Graded ): Increase in SCr by 0.3 mg/dL ( 26.5 mol/L) within 48 hours; or Increase in SCr to 1.5 times baseline, which is known or presumed to have occurred within the prior 7 days; or Urine volume 0.5 mL/kg/h for 6 hours.

2.1.2: AKI is staged for severity according to the following criteria. a (Not Graded )Stage 1: Increase in SCr by 1.5-1.9 times baseline; OR

Increase in sSCr by 0.3 mg/dL ( 26.5 mol/L); ORUrine output 0.5 mL/kg/h for 6-12 hours

Stage 2: Increase in SCr by 2.0-2.9 times baseline; ORUrine output 0.5 mL/kg/h for 12 hours

Stage 3: Increase in SCr by 3.0 times baseline; ORIncrease in SCr to 4.0 mg/dL (353.6 mol/L); ORInitiation of renal replacement therapy; ORIn patients 18 years, decrease in eGFR to 35 mL/min/1.73 m 2 ; ORUrine output 0.3 mL/kg/h for 24 hours; ORAnuria for 12 hours

2.1.3: The cause of AKI should be determined whenever possible. ( Not Graded )2.2.1: We recommend that patients be stratied for risk of AKI according to their susceptibilities and exposures. ( 1B )2.2.2: Manage patients according to their susceptibilities and exposures to reduce the risk of AKI (see relevant guideline sections).

(Not Graded )2.2.3: Test patients at increased risk for AKI with measurements of SCr and urine output to detect AKI. ( Not Graded ) Individualize

frequency and duration of monitoring based on patient risk and clinical course. ( Not Graded )2.3.1: Evaluate patients with AKI promptly to determine the cause, with special attention to reversible causes. ( Not Graded )2.3.2: Monitor patients with AKI with measurements of SCr and urine output to stage the severity, according to Recommendation

2.1.2. ( Not Graded )2.3.3: Manage patients with AKI according to the stage (Fig 2) and cause. (Not Graded )2.3.4: Evaluate patients 3 months after AKI for resolution, new onset, or worsening of pre-existing CKD. ( Not Graded )

If patients have CKD, manage these patients as detailed in the KDOQI CKD Guideline (Guidelines 7-15). ( Not Graded ) If patients do not have CKD, consider them to be at increased risk for CKD and care for them as detailed in the KDOQI CKD

Guideline 3 for patients at increased risk for CKD. ( Not Graded )

Note: Conversion factor for SCr mg/dL to mol/L, 88.4.Abbreviations: AKI, acute kidney injury; CKD, chronic kidney disease; eGFR, estimated glomerular ltration rate; SCr, serum

creatinine.Reproduced with permission of KDIGO from the KDIGO Clinical Practice Guideline for Acute Kidney Injury .1a The staging information is adapted from a table presented in the guideline.

Am J Kidney Dis. 2013;61(5):649-672 653

KDOQI Commentary


6/24

tended adverse consequence of diverting limited re-sources toward a large number of patients for whomspecialty referral may have uncertain benet.

However, we recognize that the clinical applicationof the KDIGO AKI denition could increase the earlyrecognition of AKI. Given the current lack of effec-

tive pharmacologic interventions forAKI, early recog-nition based on small changes in serum creatininelevel or urine output could allow implementation of simple interventionssuch as avoiding potentiallynephrotoxic medications or diagnostic or therapeuticprocedures that are associated with increased risk of AKI unless clearly clinically indicated, or judiciousprotocol-driven volume resuscitationthat could re-duce the risk of more severe kidney injury.

In addition to these more global concerns, our Work Group identied several specic issues regarding thedenition and staging criteria that we believe require

specic comment, as discussed next. Inclusion of urine output into the denition of AKI.Both RIFLE and AKIN included duration of oliguriain their denitions and staging criteria ( Table 1 ).Several criticisms have been offered of these urineoutput criteria, many of which are acknowledged bythe KDIGO AKI Guideline Work Group but deservefurther mention. Only a small number of studies haveexamined the urine output criteria for AKI and corre-lated them with adverse clinical outcomes, in contrastto the numerous studies that have focused on theserum creatinine criteria. The studies that have evalu-ated the urine output component of the criteria havefound poor calibration between these criteria and theserum creatinine criteria. 24 Specically, when as-sessed, there was poor prognostic correlation betweenthe briefer durations of oliguria and small changes inserum creatinine level. Second, oliguria may be anappropriate response to volume depletion and hencereect under-resuscitation rather than injury to thekidney, which is implicit in the term AKI. There-fore, even demonstrating an association between oli-guria and adverse clinical outcomes is insufcient to justify its incorporation into the denition of AKI.Third, the use of a weight-based denition for AKIlimits its use in the obese due to the nonlinear relation-ship between body weight and urine output. Under thecurrent denition, urine output of 40 mL/h in a 90-kgpatient for 12 hours would lead to classication asstage 2 AKI. Fourth, diuretic administration has beenshown to change RIFLE classication by urine outputcriteria, 24 as would be expected when a denition isbased on a physiologic variable that can be manipu-lated pharmacologically. Finally, the use of oliguria ina denition of AKI may promote its use in clinicalpractice as a surrogate end point. We know fromseveral studies that pharmacologic agents that can

increase urine output independent of augmentation of kidney function, such as loop diuretics or dopamine,may not be helpful and may even be harmful in somesettings. 25,26 While the inclusion of these criteria inthe AKI guideline may have the benecial effect of encouraging more rigorous documentation of urine

output and overall uid balance in settings outside theoperating room and critical care units, there are alsorisks of unintended consequences with dening tran-sient oliguria as AKI. In particular, we believe thatcurrent data are inadequate to support the reliance onoliguria as a surrogate end point in clinical trials or inperformance metrics. We note that pediatric, and inparticular neonatal, providers rely heavily on urineoutput and weight-based uid balance and thereforeinclusion of these parameters in the pediatric AKIcriteria is reasonable; furthermore, neonates in particu-lar are unlikely to have morbid obesity and therefore

the methodological issues with weight-based urineoutput are somewhat less problematic. Nonetheless,further research is needed in the pediatric populationto determine what metrics are best used to deneAKI,and practitioners should recognize that many of thesame limitations to the use of urine output criteriaapply in the pediatric population.

Use of small changes in serum creatinine to dene AKI. The inclusion of a small absolute change inserum creatinine level in the AKIN denition of AKIwas based on the demonstration by Chertow et al 4 thatminor uctuations in serum creatinine concentrationwere strongly associated with adverse outcomes in aretrospective analysis of data from individuals hospi-talized at a single medical center. Although the asso-ciation remained, albeit attenuated, after multivari-able adjustment, only administrative data wereavailable for construction of the multivariable mod-els, and residual confounding may have been present.Small changes in serum creatinine level have alsobeen identied to be of prognostic importance insubsequent studies. Even a 0.1-mg/dL increase inserum creatinine level appears to be associated withincreased risk compared to no change in serum creati-nine level. 27 Whether these small changes in serumcreatinine level reect clinically meaningful uctua-tions in kidney function that are causally linked toadverse outcomes or are merely a marker of underly-ing vasculardisease or diminished renal reserve, whichcould be the actual mediators of the adverse out-comes, remains unresolved. Furthermore, an implicitassumption underlying the incorporation of smallchanges in serum creatinine concentration into thedenition of AKI is that GFR in an individual isconstant over time and not subject to physiologicuctuation. In addition, biological uctuation in se-rum creatinine level may also result from variations in

Am J Kidney Dis. 2013;61(5):649-672654

Palevsky et al


7/24

diet, creatinine generation, tubular secretion, medica-tions, and variability in sodium and volume homeosta-sis that are within the physiologic range. Such uctua-tion in serum creatinine concentration is moreprominent in individuals with decreased renal reserveor overt CKD. For all these reasons, an absolute

change in serum creatinine level of 0.3 mg/dL mayrepresent a relatively inconsequential change in GFRin patients with underlying CKD and not reect super-imposed acute pathology. Despite these concerns inadult patients, in neonatal/pediatric patients, suchsmallchanges in serum creatinine level may represent rela-tively large changes in actual GFR that should not bedisregarded. Whether novel biomarkers of tubularinjury may enhance risk stratication of patients withsmall changes in serum creatinine level is a questionthat requires further investigation and may lead tofurther renement of AKI staging.

The independent role of duration of AKI. Severalrecent studies have suggested that duration of AKImay be a more important predictor of outcomes thanthe magnitude of change in serum creatinine level.For example, in one study, even a transient elevationin creatinine level ( 3 days) was associated withincreased risk of death, and the risk of mortality w asgreater in patients with prolonged duration of AKI. 28

After adjustment for duration of AKI, staging was nolonger predictive of adverse outcomes. Unfortunately,this important dimension can only be assessed retro-spectively and cannot be included in prospectivestaging criteria; however, consideration should begiven to including an assessment of duration of AKIin future criteria designed for epidemiologic studies.

Pediatric-specic concerns. The KDOQI Work Group also had several concerns regarding the appli-cation the KDIGO denition and staging of AKI to thepediatric population. One of the major concerns withboth the RIFLE and AKIN criteria is that they areproblematic for smaller (pediatric) patients with lowmuscle mass. The pRIFLE criteria were developed toaddress the issues of applying the adult criteria to thepediatric population. 20 For example, because abnor-mally elevated serum creatinine measurements areoften overlooked in pediatric patients (due to adultnormative data in laboratory readouts), utilization of the pRIFLE criteria might improve identication of those with true AKI. The pRIFLE criteria have beenused widely in pediatric clinical research and effortsare now being made to apply these criteria in clinicalpractice. However, there are signicant differencesbetween the KDIGO and pRIFLE criteria. Conse-quently, further validation will be required prior to theadoption of the KDIGO criteria into research andpractice. Specically, pediatric nephrologists will wantto better understand whether the differences in the 2

denitions result in the identication of differentpatient groups, and if so, what systematic biases maybe introduced by these differences. Thus, it was un-clear to the KDOQI Work Group whether pediatricnephrologists would widely embrace the new KDIGOAKI denition or continue to apply the pRIFLE

criteria.In addition, most pediatric patients will not neces-

sarily have had previous serum creatinine measure-ments and therefore baseline renal function determina-tion is often assumed to be normal, which may beproblematic. In addition, the neonatal population re-mains an enigma in terms of renal function determina-tion, making both the denition and identication of AKI in this population difcult. Many neonates havenot attained full renal mass development at birth andare obligatorily born with low GFR, which improveswith growth and development. AKI in neonates is

often associated with high urine output, which may gounrecognized given the paucity of serum creatininemeasurements in this patient cohort. Finally, immedi-ately after birth, serum creatinine level in neonatesoften reects maternal creatinine levels, which mustbe considered when evaluating renal function. How-ever, the recommendation of a 0.3-mg/dL increase inserum creatinine level (even in the context of baselinematernal creatinine level) should be sufcient to trig-ger concern, and this component of the denition isapplicable in the neonatal population.

In addition to dening and staging AKI, the KDIGOguideline introduces the new term acute kidney dis-ease (AKD), dened as AKI, or GFR 60 mL/min/ 1.73 m 2 for less than 3 months, or a decrease in GFRby 35% or an increase in serum creatinine level by

50% for less than 3 months, or structural kidneydamage of less than 3 months duration. While weunderstand the nosological rationale for developingterminology to describe patients with kidney diseasethat does not meet the criteria for either AKI or CKD,we have concern that the introduction of this terminol-ogy may confuse clinicians and inappropriately divertattention away from diagnostic considerations. WhileCKD encompasses a number of pathophysiologicallyand histopathologically distinct diseases, arguably themajority of patients with CKD have diseases such ashypertensive nephrosclerosis and diabetic nephroscle-rosis, for which treatment approaches are similar.However, in the acute setting, distinctions betweencauses of disease have greater consequence. To lumpthe broad range of conditions that result in acute andsubacute kidney disease, ranging from acute tubularnecrosis to obstructive uropathy, atheroembolic dis-ease, and rapidly progressive glomerulonephritis, intothe single umbrella term of AKD runs the risk of promoting diagnostic laziness among clinicians who

Am J Kidney Dis. 2013;61(5):649-672 655

KDOQI Commentary


8/24

would have a convenient name to apply rather than anabnormal laboratory value to investigate. We would

have welcomed a more in-depth discussion of whatspecic kidney diseases are likely to segregate toAKD as opposed to AKI or CKD and how the intro-duction of a new term was thought by the authors tohelp in clinical practice. While AKD may be a usefulconstruct in epidemiologic studies, we believe the useof this term should be discouraged in clinical practice.

In contrast to our reservations regarding the intro-duction of the AKD terminology, we strongly concurwith the KDIGO recommendations that the cause of AKI should be determined whenever possible and thatpatients with AKI should be evaluated promptly todetermine the cause, with special attention to revers-ible causes. However, our Work Group had severalconcerns regarding several of the other recommenda-tions related to the evaluation and general manage-ment of patients with and at risk of AKI. Specically,our Work Group has concerns regarding the recom-mendation to manage patients with AKI according tothe stage and cause. We believe that the stage-basedmanagement recommendations ( Fig 2 ) are not ad-equately evidence based. These recommendations im-plicitly assume homogeneity within each AKI stageand successive increases in severity across stages. Asa result of the lack of correlation between serum

creatinine level and GFR in the acute setting, a pa-tients serum creatinine level may increase, resulting

in apparent progression inAKI stage, despite improve-ment in GFR. Clearly no guideline can accommodateall the subtleties of clinical practice or substitute forclinical judgment, but the recommendations for man-agement overall were relatively nonspecic and un-likely to help in daily clinical practice. We are espe-cially concerned that the development of clinicalaction plans based onAKI stage may result in inappro-priate protocolization of care. While recommenda-tions such as discontinuation of nephrotoxic agentswhen possible and ensuring volume status and perfu-sion pressure in high-risk patients or patients withAKI, waiting until stage 2AKI to check for changes indrug dosing implies that this need not be done earlier,while the recommendations for considering initiationof RRT and intensive care unit admission in stage 2AKI seem premature. Overall, we thought that theextreme heterogeneity of AKI and its lack of consis-tent mapping to stages 1, 2, and 3 make the proposedstage-based management of AKI clinically unhelpfuland inapplicable to many patients.

We agree with the emphasis on close postdischargeclinical evaluation of patients with moderate to severeAKI given the recent identication of an associationbetween AKI and long-term outcomes such as renal

Figure 2. Stage-based management of acute kidney injury (AKI). Shading of boxes indicates priority of actionsolid shading (withwhite lettering) indicates actions that are equally appropriate at all stages whereas graded shading (with black lettering) indicatesincreasing priority as intensity increases. Abbreviation: I CU, intensive care unit. Reproduced with permission of KDIGO from theKDIGO Clinical Practice Guideline for Acute Kidney Injury .1

Am J Kidney Dis. 2013;61(5):649-672656

Palevsky et al


9/24

function decline and mortality. However, we are con-cerned that clinicians may be overwhelmed by 3-monthfollow-up of many patients with stage 1 AKI, whowill constitute the majority of individuals identiedby the AKI denition due to transient oliguria or smallchanges in serum creatinine level. The risk of progres-

sive CKD after AKI is related to the severity of AKI, 16,29,30 suggesting that risk stratication on thebasis of AKI severity may be useful in guiding thetiming of outpatient follow-up. Those with stage 3AKI, for example, will likely require far earlier post-discharge follow-up. Furthermore, patients with AKIin the setting of pre-existing CKD or those whodevelop worsening CKD as a consequence of anepisode of AKI could represent a particularly high-risk group.

In the pediatric population, the 3-month time tofollow-up may be reasonable. The CKiD (ChronicKidney Disease in Children) national trial has identi-ed that a large number of the patients available forthat patient cohort were at-risk patients in the neonatalpopulation. 31 Although the evidence for transitionfrom AKI to CKD in the pediatric population is alsoobservational and further epidemiologic research isrequired, the number of pediatric patients with AKI ismuch smaller than in the adult population and thestakes of missing nascent CKD may be greater giventhe potential longer duration of follow-up than in theadult population. Thus, we would suggest that earlyfollow-up among pediatric patients with AKI is advis-able.

ImplicationsWithinUSHealthCare1. The KDIGO denition and staging system for

AKI provides an important tool for conducting epide-miologic studies and for the design of clinical trialsand helps increase awareness of the importance of small changes in kidney function in the acute setting.However, there is insufcient validation of this deni-tion and staging system for its use in the diagnosis andclinical management of patients. In particular, we donot believe that there are sufcient data to support useof the stage-based management approach proposed inthe KDIGO guideline. Rather, management of pa-tients with AKI should be based on assessment of overall clinical status, including specic cause of AKI, trends in kidney function over time, comorbidconditions, assessment of volume status, and concomi-tant acid-base and electrolyte disturbances.

2. While AKI is an important risk factor for thedevelopment and progression of CKD, the majority of patients with mild, readily reversible AKI (eg, thepatient with no baseline CKD who presents withreversible AKI in the setting of volume depletion) areat relatively low risk of progressive CKD. From a

public health standpoint in the United States, fol-low-up of kidney function after an episode of AKIshould be targeted to the highest risk populations,including neonatal/pediatric patients, individuals withbaseline CKD, and patients with severe AKI or whohave incomplete recovery of kidney function at hospi-

tal discharge.3. The KDIGO recommendations for denition and

staging of adult AKI should not be used for thedevelopment of clinical performance measures.

4. Administrative coding for AKI should not bebased solely on this denition and staging system.

Prevention andTreatment of AKICommentary The section of the KDIGO guideline on prevention

and treatment of AKI includes 7 level 1 r ecomm enda-tions and 22 level 2 recommendations ( Box 2 ). Our

KDOQI Work Group agreed with the 7 level 1 recom-mendations and thought that they were generallyapplicable in the United States. These focused on theuse of vasopressors and uids to treat patients inshock; avoidance of diuretics, dopamine, and recom-binant human IGF-1 (insulin-like growth factor 1) toprevent or treat AKI; therapeutic drug monitoringduring the use of aminoglycosides; and avoidance of nephrotoxic medications when possible. Importantly,avoidance of nephrotoxins is not always possible. Forexample, some pathogens are not effectively treatedby azoles or echinocandins: Candida krusei is intrinsi-

cally resistant to azoles, and Candida parapsilosis isfrequently resistant to echinocandins.The KDOQI Work Group largely agreed with the

recommendation to use isotonic crystalloids ratherthan colloids for volume expansion in patients at risk of AKI or with AKI. In addition, we noted that t hereappears to be harm with starch-containing uids. 32-36

Therefore, these solutions should be avoided in pa-tients at risk of AKI or with AKI. Along the samelines, albumin resuscitation has been associated withharm in patients with traumat ic brain injury andshould be avoided in that setting. 37,38

There are also specic settings in which albumin isappropriate for initial management of expansion of intravascular volume. Specically, in patients withliver disease, intravenous albumin administration ap-pears to be benecial for the prevention of renalfailure and death in patients with spontaneous bacte-rial peritonitis, as well as for the prevention of renalfai lure i n those undergoing large-volume paracente-sis. 39,40 Along the same lines, the most recent diagnos-tic criteria for hepatorenal syndrome include a lack of improvement in renal function after volume expan-sion with albumin (1 g/kg/d up to 100 g/d) for at least2 days and withdrawal of diuretic therapy. 41 Data

Am J Kidney Dis. 2013;61(5):649-672 657

KDOQI Commentary


10/24

published since the KDIGOAKIguideline rst appearedsuggest that there may be harm associated with theadministration of hyperchloremic intravenous solutions,including crystalloid 42,43 ; however, these studies haveall been observational and this premiseneeds to be testedin a randomized clinical trial given the enormous effectsize observed and the inability to control for othertemporal trends in care in the prior studies.

With regard to the use of vasopressors in addition tointravenous uids in patients with vasomotor shock,we noted that uid resuscitation is typically insuf-cient to fully restore blood pressure to normal levels,which is critical for the prevention and managementof AKI. However, whether one vasopressor is moreeffective for patients with or at risk of AKI is un-known. In a large multicenter clinical trial of patientswith shock comparing dopamine and norepinephrinefor rst-line vasopressor support, the use of dopamine

was associated with more adverse events in those withsepsis and with an increased risk of death in those withcardiogenic shock. 44 In a subsequent meta-analysis andsystematic review of patients with septic shock, dopa-mine was associated with an increased risk of death andarrhythmias. 45,46 Thus, we believe that dopamineshouldbe used with caution as the rst-line agent of choice inpatients with shock at present.

Our Work Group did not fully agree with therecommendation to use protocol-based managementof hemodynamic and oxygenation parameters to pre-vent the development or worsening of AKI in high-risk patients in the perioperative setting or in patientswith septic shock.The group noted that the recommen-dation for early goal-directed therapy in patients withseptic shock is based primarily on single-center ran-domized clinical trials 47,48 and observational stud-ies. 49-52 There was no assessment of renal outcomes

Box 2. Summary of KDIGO Recommendation Statements: Prevention and Treatment of AKI

3.1.1: In the absence of hemorrhagic shock, we suggest using isotonic crystalloids rather than colloids (albumin or starches) as initialmanagement for expansion of intravascular volume in patients at risk for AKI or with AKI. ( 2B )

3.1.2: We recommend the use of vasopressorsin conjunction with uids in patients with vasomotor shock with, or at risk for, AKI. ( 1C )3.1.3: We suggest using protocol-based management of hemodynamic and oxygenation parameters to prevent development or

worsening of AKI in high-risk patients in the perioperative setting ( 2C ) or in patients with septic shock ( 2C ).

3.3.1: In critically ill patients, we suggest insulin therapy targeting plasma glucose 110-149 mg/dL (6.1-8.3 mmol/L). ( 2C )3.3.2: We suggest achieving a total energy intake of 20-30 kcal/kg/d in patients with any stage of AKI. ( 2C )3.3.3: We suggest to avoid restriction of protein intake with the aim of preventing or delaying initiation of RRT. ( 2D )3.3.4: We suggest administering 0.8-1.0 g/kg/d of protein in noncatabolic AKI patients without need for dialysis ( 2D ), 1.0-1.5 g/kg/d in

patients with AKI on RRT ( 2D ), and up to a maximum of 1.7 g/kg/d in patients on continuous renal replacement therapy (CRRT)and in hypercatabolic patients. ( 2D )

3.3.5: We suggest providing nutrition preferentially via the enteral route in patients with AKI. ( 2C )3.4.1: We recommend not using diuretics to prevent AKI. ( 1B )3.4.2: We suggest not using diuretics to treat AKI, except in the management of volume overload. (2C )3.5.1: We recommend not using low-dose dopamine to prevent or treat AKI. ( 1A)3.5.2: We suggest not using fenoldopam to prevent or treat AKI. ( 2C )3.5.3: We suggest not using atrial natriuretic peptide (ANP) to prevent ( 2C ) or treat ( 2B ) AKI.3.6.1: We recommend not using recombinant human (rh)IGF-1 to prevent or treat AKI. ( 1B )3.7.1: We suggest that a single dose of theophylline may be given in neonates with severe perinatal asphyxia, who are at high risk of

AKI. (2B )

3.8.1: We suggest not using aminoglycosides for the treatment of infections unless no suitable, less nephrotoxic, therapeuticalternatives are available. ( 2A)

3.8.2: We suggest that, in patients with normal kidney function in steady state, aminoglycosides are administered as a single dosedaily rather than multiple-dose daily treatment regimens. ( 2B )

3.8.3: We recommend monitoring aminoglycoside drug levels when treatment with multiple daily dosing is used for more than 24hours. ( 1A)

3.8.4: We suggest monitoring aminoglycoside drug levels when treatment with single-daily dosing is used for more than 48 hours.(2C )

3.8.5: We suggest using topical or local applications of aminoglycosides (e.g., respiratory aerosols, instilled antibiotic beads), ratherthan i.v. application, when feasible and suitable. ( 2B )

3.8.6: We suggest using lipid formulations of amphotericin B rather than conventional formulations of amphotericin B. ( 2A)3.8.7: In the treatment of systemic mycoses or parasitic infections, we recommend using azole antifungal agents and/or the

echinocandins rather than conventional amphotericin B, if equal therapeutic efcacy can be assumed. ( 1A)3.9.1: We suggest that off-pump coronary arterybypass graft surgery not be selected solely for the purpose of reducing perioperative

AKI or need for RRT. (2C) 3.9.2: We suggest not using NAC to prevent AKI in critically ill patients with hypotension. (2D) 3.9.3: We recommend not using oral or i.v. NAC for prevention of postsurgical AKI. ( 1A)

Abbreviations: AKI, acute kidney injury; IGF-1, insulin-like growth factor 1; i.v., intravenous; NAC, N -acetylcysteine; RRT, renalreplacement therapy.

Reproduced with permission of KDIGO from the KDIGO Clinical Practice Guideline for Acute Kidney Injury .1

Am J Kidney Dis. 2013;61(5):649-672658

Palevsky et al


11/24

in the original randomized clinical trial, so it is un-clear if the intervention had any impact on renaloutcomes. Similarly, perioperative protocols have simi-larly shown aggregate improvement when assessed ina meta-analysis, but individual studies have beenlimited by small numbers and single-center design. 53

Furthermore, our Work Group thought that the proto-col must be specied. That is, any protocol that isinstituted based on this recommendation should beone that has been previously studied and validated.For example, previous studies of increasing oxygentransport have demonstrated that the use of dobut-amine to increase oxygen trans por t can worsen mortal-ity, 54 although the Rivers et al 47 study of early goal-directed therapy suggests that dobutamine in anappropriate clinical context may not be harmful andmay have benet. Thus, medical centers adoptingprotocolized care should only adopt protocols that

have been previously shown to be helpful and demon-strated no harm. If de novo protocols are developedfor use, they should be used in the connes of aclinical trial to ensure no harm.

With regard to glycemic control, we agreed withthe recommended target of 110-149 mg/dL, but notedthat emerging evidence suggests that rapid and sus-tained correction of glycemia in diabetic patients withprevious poor glycemic control may have worse out-comes. 55 OurWork Group agreed with the recommen-dations for nutritional support in AKI, but noted thatthe evidence base for these recommendations is lim-ited. Specically, the guideline recommends total en-ergy intake of 20-30 kcal/kg/d in patients with anystage of AKI with administration of 0.8-1.0 g/kg/d of protein in noncatabolic patients with AKI withoutneed for dialysis, 1.0-1.5 g/kg/d in patients with AKIon RRT, and up to a maximum of 1.7 g/kg/d inpatients on continuous RRT (CRRT) and in hypercata-bolic patients. The provision of protein to patientswith AKI is intended to avoid marked net negativenitrogen balance. Patients with AKI on RRT will haveadditional protein loss as amino acids are removed.Therefore, these patients require additional proteincompared with patients with AKI who are not on RRT.Similarly, patients who are hypercatabolic will re-quire additional protein to avoid net negative nitrogenbalance. Along the same lines, the KDOQI Work Group agreed with the recommendation to avoidrestriction of protein intake with the aim of preventingor delaying initiation of RRT. We also agreed with therecommendation to provide nutrition preferentially bythe enteral route in patients with AKI. The Work Group noted that although evidence is more limited inthe setting of AKI, 56 in a recent large randomizedclinical trial of critically ill patients, early parenteralnutrition was associated with a higher rate of compli-

cations, primarily infectious complications. 57 Alongthe same lines, recent trials suggest that enteral feed-ing is t olera ted by either the nasogastric or nasojejunalroutes. 58,59

Nutritional support in pediatric patients must in-clude the recognition that these patients are generally

in a developmental growth phase; thus, their require-ments may be increased compared with adult patients.Recent estimates have been published in terms of estimated protein/amino acid supplementation in criti-cally ill children and suggest that protein require-ments are on the order of 2-3 g/kg/d for children aged0-2 years, 1.5-2.0 g/kg/d for children aged 2-13 yea rs,and 1.5 g/kg/d for adolescents aged 13-18 years. 60

Children requiri ng RRT appear to need supplementa-tion beyond this. 61

There are no pharmacotherapies available for theprevention or treatment ofAKI in adults, so recommen-

dations to avoid these are reasonable. Specically, weagreed that although there is theoretical benet to theuse of loop diuretics to prevent AKI, there are no datato support the use of diuretics to prevent AKI. Recentstudies in patients undergoing cardiac surgery or con-trast studies suggest that the use of diuretics does notprevent AKI 62 and, in the case of contrast administra-tion, increases the risk of AKI. 63,64 The KDOQI Work Group thought that it was important to comment onthe use of diuretics in patients with AKI and rhabdo-myolysis. With regard to osmotic diuretics, one retro-spective study has suggested that the use of mannitoladministration may be of benet only in patients withmarked elevations in creatinine kinase level ( 30,000U/L). 65 However, even in these patients with severerhabdomyolysis, the true benet associated with man-nitol administration remains undened. Furthermore,mannitol should be administered carefully and iscontraindicated in patients with oligoanuria. Diureticsmay be used for the treatment of volume overload inAKI. 66 We agree that low-dose dopamine, fenoldo-pam, IGF-1, and N -acetylcysteine (NAC) should notbe used for the treatment or prevention of AKI.

We concur with the recommendation to administera single dose of theophylline to neonates with severeperinatal asphyxia who are at high risk of AKI.Several clinical trials in the neonatal population sug-gest that although mortality outcomes are not af-fected, improved uid control and higher GFR areassocia ted with theophylline administration in thissetting. 67-69 A similar renal-selective response hasbeen demonstrated in asphyxiated term newbornsreceiving a single theophylline dose in the rst 60minutes of life. 70 Given the generally physiologichyperactive state of the renal autoregulatory systemapparent at transition from in utero to ex utero environ-ments and the role that adenosine plays in these

Am J Kidney Dis. 2013;61(5):649-672 659

KDOQI Commentary


12/24

processes, the use of theophylline in the aforemen-tioned settings is reasonable.

A substantial number of recommendations weremade with regard to the use of aminoglycosides andantifungal agents. Our Work Group agreed with theserecommendations but noted that there are situations in

which aminoglycosides or multidose regimens arepreferable (eg, multiple dose regimens remain thestandard of care for enterococcal endocarditis). Wealso noted that while local instillation of aminoglyco-sides is reasonable, it is unknown what the incidenceof nephrotoxicity is with local administration of ami-noglycosides; for example, there are case reports of nephrotoxicity associated with inhaled tobramycintherapy. 71-73

With regard to the use of lipid formulations of amphotericin B rather than conventional formulationsof amphotericin B, we note that it has been suggested

that the decreased incidence of AKI and the associateddecrease in hospitalization costs makes the lipid for-mulations cost-effective compared to conventionalformulations of amphotericin B. 74,75 Furthermore,other simple strategies that may reduce the risk of nephrotoxicity include discontinuation of diuretics,sodium loading and volume repletion, and potassiumand magnesium supplementation, particularly wh enusing non-liposomal preparations of amphotericin B. 76

The 2 most signicant risk factors for the develop-ment of AKI following cardiac surgery are cardi opul-monary bypass time and a history of CKD. 77,78 As aresult, there has been considerable interest in thepotential for off-pump bypass to decrease the risk of AKI. The largest retrospective study focused on pa-tients with CKD was published after the KDIGO AKIguideline. 79 This study examined nonemergent iso-lated coronary artery bypass grafting (CABG) casesin the Society of Thoracic Surgery Database from2004 through 2009 (N 742,909). Among patientswith eGFR of 15-29 mL/min/1.73 m 2 , off-pumpCABG was associated with a decreased risk of postop-erative RRT (risk difference, 2.79; 95% condenceinterval, 1.37-4.20) in the propensity-adjusted analy-sis (P 0.01). There was no difference in those with

less advanced CKD, although those with eGFR of 30-59 mL/min/1.73 m 2 were examined in aggregaterather than as separately as CKD stages 3a and 3b.This potential benet needs to be balanced against theconcern of high er gr aft occlusion rates using off-pump techniques. 80,81

ImplicationsWithinUSHealthCare1. In the United States, starch-containing intrave-

nous uids should be avoided. The use of albumin forresuscitation should be limited to specic situations inwhich it is clear that albumin is of benet, given the

increased costs of these solutions compared to crystal-loid and the intermittent nationwide shortages of thesesolutions. Recent studies have suggested that theremay be harm with hyperchloremic intravenous solu-tions (eg, isotonic saline), but this needs to be furthertested prior to the widespread use of balanced electro-

lyte solutions with lower chloride content (eg, Plasma-Lyte, Hartmanns and lactated Ringers) in the UnitedStates.

2. With regard to vasopressor support in patientswith shock, dopamine is associated with an increasedrate of complications in patients with septic andcardiogenic shock compared to norepinephrine. Al-though this needs to be further tested and a number of guidelines still recommend the use of dopamine as arst-line agent, norepinephrine should be used as arst-line agent over dopamine unless there are spe-cic contraindications to the use of norepinephrine.

3. Protocol-based management of perioperative pa-tients and patients with sepsis is a reasonable ap-proach, but the protocols for such management needto be further developed and tested. We note thatseveral large randomized clinical trials will test proto-col-based managementof sepsis ( ClinicalTrials.gov identi-ers NCT00975793 and NCT00510835 and Interna-tional Standard Randomised Controlled Trial NumberISRCTN36307479). Until then, care should not bebenchmarked based on the implementation of suchprotocols within a given time frame; it is reasonable touse time to antibiotics and time to uid administrationas benchmarks for care in sepsis.

4. A target blood glucose level of 110-149 mg/dL incritically ill patients may be associated with decreasedrisk of AKI and other morbidity and mortality. How-ever, emerging evidence suggests that rapid and sus-tained correction of hyperglycemia in diabetic pa-tients with previous poor glycemic control may haveworse outcomes.

5. Nutritional support in patients with AKI shouldbe provided by the enteral route when possible. Pro-tein supplementation should not be withheld to delaythe initiation of RRT. Patients with AKI are oftenhighly catabolic and protein requirements typicallyincrease with RRT due to amino acid loss, so nutri-tional prescriptions for patients with AKI need toreect these considerations.

6. No agents are recommended for the treatment orprevention ofAKI in adults; a single dose of theophyl-line can be provided to neonates at risk of AKI.

7. The cost-effectiveness of the widespread use of liposomal amphotericin B should be studied. Therapeu-tic drug monitoring for aminoglycosides should re-main standard of care.

8. Off-pump CABG may be of benet to reduce therisk of RRT in those with advanced CKD, but further

Am J Kidney Dis. 2013;61(5):649-672660

Palevsky et al
http://www.clinicaltrials.gov/http://www.clinicaltrials.gov/


13/24

studies are warranted. Off-pump CABG should not beconsidered standard of care in those with CKD.

Contrast-InducedAKICommentary The KDIGO guideline contains 12 recommenda-

tions on prevention of contrast-induced AKI; 5 are notgraded, 4 are level 1 recommendation s, and t he remain-ing 3 are level 2 recommendations ( Box 3 ). The rstof these recommendations is that contrast-induced

AKI be dened and staged using the KDIGO deni-tion and staging criteria. This recommendation makessense from the standpoint of consistency and is gener-ally applicable in the United States, albeit with thelimitations noted in the previous comments on deni-tion and staging. However, we note that most clinicalstudies related to contrast-induced AKI have usedalternative denitions based on increments in serumcreatinine level of 25% or 50% relative to base-line and/or an absolute change in serum creatininelevel 0.5 mg/dL within 2-5 days following theadministration of iodinated contrast media. Althougha multitude of largely observational studies have dem-onstrated associations of contrast-induced AKI, de-ned by these changes in serum creatinine level, withserious adverse events, the causal nature of suchassociations remains unproved, leaving unansweredthe question of how to best dene clinically signi-cant contrast-induced AKI. Thus, the recommenda-tion to operationalize the denition of contrast-inducedAKI within the KDIGO framework is justied.Furthermore, applicability of the urine output criteriato the diagnosis and staging for this form of AKI isuncertain. Most episodes of contrast-induced AKI arenonoliguric. In addition, most contrast-enhanced radio-

graphic procedures are performed in the outpatientsetting, where monitoring of urine volume is notperformed and is impractical. Furthermore, recogniz-ing that the majority of clinicians performing contrast-enhanced procedures are not nephrologists, it is impor-tant that future efforts to rene the denition of contrast-induced AKI include input from national andinternational radiology and cardiology societies.

Underlying impairment in renal function is theprincipal risk factor for contrast-induced AKI. There-fore, identication of patients with decreased kidneyfunction is essential in order to optimize the benet of preventive care. However, measuring serum creati-nine prior to all contrast-enhanced procedures is nei-ther practical nor feasible. Simple questionnaires havebeen shown to be effective for the identication of patients at higher risk of abnormal underlying renalfunction. 82 In patients without a recent serum creati-nine measurement, it is reasonable in the UnitedStates to use such questionnaires to identify patientswho should have serum creatinine measured prior tocontrast administration. The use of dipstick testing of urine for protein is also suggested as a means of identifying pre-existing kidney disease. Our Work Group had reservations regarding this approach andsuggests that further validation of this approach isneeded before widespread adoption of this strategy isconsidered. Additional risk factors for contrast-induced AKI, including diabetes in the setting of renalimpairment, heart failure, repeated contrast exposureover short periods, and concomitant nephrotoxin ad-ministration (eg, nonsteroidal anti-inammatory drugsand aminoglycosides) are important to recognize.Equipoise on the impact of discontinuing angiotensin-converting enzymeinhibitor and angiotensin recep-

Box 3. Summary of KDIGO Recommendation Statements: Contrast-Induced AKI

4.1: Dene and stage AKI after administration of intravascular contrast media as per Recommendations 2.1.1-2.1.2. ( Not Graded )4.1.1: In individuals who develop changes in kidney function after administration of intravascular contrast media, evaluate for CI-AKI

as well as for other possible causes of AKI. ( Not Graded )4.2.1: Assess the risk for CI-AKI and, in particular, screen for pre-existing impairment of kidney function in all patients who are

considered for a procedure that requires intravascular (i.v. or i.a.) administration of iodinated contrast medium. ( Not Graded )

4.2.2: Consider alternative imaging methods in patients at increased risk for CI-AKI. ( Not Graded )4.3.1: Use the lowest possible dose of contrast medium in patients at risk for CI-AKI. (Not Graded )4.3.2: We recommend using either iso-osmolar or low-osmolar iodinated contrast media, rather than high-osmolar iodinated contrast

media in patients at increased risk of CI-AKI. ( 1B )4.4.1: We recommend i.v. volume expansion with either isotonic sodium chloride or sodium bicarbonate solutions, rather than no i.v.

volume expansion, in patients at increased risk for CI-AKI. ( 1A)4.4.2: We recommend not using oral uids alone in patients at increased risk of CI-AKI. ( 1C )4.4.3: We suggest using oral NAC, together with i.v. isotonic crystalloids, in patients at increased risk of CI-AKI. ( 2D )4.4.4: We suggest not using theophylline to prevent CI-AKI. ( 2C )4.4.5: We recommend not using fenoldopam to prevent CI-AKI. ( 1B )4.5.1: We suggest not using prophylactic intermittent hemodialysis (IHD)or hemoltration (HF) for contrast-media removal in patients

at increased risk for CI-AKI. ( 2C )

Abbreviations: AKI, acute kidney injury; CI-AKI, contrast-induced acute kidney injury; i.a., intra-arterial; i.v., intravenous; NAC,N -acetylcysteine.


Am J Kidney Dis. 2013;61(5):649-672 661

KDOQI Commentary


14/24

tor blocker therapy prior to contrast administrationpersists and the KDIGO AKI guideline acknowledgesthe inadequacy of the evidence to support the discon-tinuation of these agents prior to contrast-enhancedprocedures. While the administration of loop diureticsfor the purpose of preventing contrast-induced AKI is

ineffective and potentially harmful, there are no datato support the recommendation to routinely discon-tinue these agents in all patients prior to contrastadministration.

In patients at increased risk of contrast-inducedAKI, it is appropriate to discuss the risk to benetratio of the planned procedure with the practitionerperforming the study (ie, radiologist or interventional-ist), as well as with the patient. In considering alterna-tive procedures that do not involve the administrationof iodinated contrast, the risk of nephrogenic systemicbrosis associated with the administration of gadolin-

ium-based contrast agents is important given the po-tential morbidity associated with this condition. Pa-tients with eGFR 30 mL/min/1.73 m 2 and patientswith AKI are at risk of nephrogenic systemic brosisand appropriate cautionary measures should be takenregarding gadolinium-based contrast agents in thisgroup. While the KDOQI Work Group agrees with theimportance of avoiding unnecessary radiocontrast ad-ministration in high-risk patients, a cautionary note isrequired because several studies have suggested thatconcern over contrast-induced AKI leads to underuti-lization of imaging techniques, particularly coronaryangiography, in high-risk patients with CKD. 83

Studies of the association of dose of contrast me-dium, considered as overall volume and grams of iodine per eGFR ratio, with risk of contrast-inducedAKI sugges t that higher doses are associated withgreater risk. 84-86 Recommendations to consider notonly the total volume of contrast, but also the grams of iodine per eGFR ratio, are reasonable based on avail-able evidence, yet are most likely to be applicable toradiologists and interventionalists performing contrast-enhanced procedures. Furthermore, routine calcula-tion of the grams of iodine per eGFR ratio adds a layerof complexity to the conduct of contrast-enhancedprocedures that may not be readily acceptable to allclinicians.

The specic contrast agent used also has an effecton the risk of contrast-induced AKI. High-osmolalcontrast media are associated with higher rates of contrast-induced AKI compared with low-osmolalagents in at-risk patients and should not be used in thispatient group. 87 Clinical trials and meta-analyses com-paring the nephrotoxicity of low-osmolal contrastagents with iso-osmolal contrast have yielded conict-ing results. 88,89 Guidelines issued by the AmericanHeart Association/American College of Cardiology in

2009 recommended the use of iso-osmolal or low-osmolal contrast exclusive of iohexol and ioxaglate inpatients with nondialysis-dependent CKD undergo-ing angiography based on data demonstrating higherrates of contrast-induced AKI among patients whoreceived these 2 agents compared with iso-osmolal

contrast and other low-osmolal agents.90

The recom-mendation of our Work Group is to follow the guide-lines outlined by the American Heart Association/ American College of Cardiology in 2009. It isimportant to note that iodixanol (Visipaque) is theonly iso-osmolal contrast agent clinically available inthe United States; there is a substantially higher costfor brand-name Visipaque than the low-osmolal con-trast agents, which may be important to consider indecisions related to choice of contrast.

The provision of intravenous uids prior to andafter the administration of iodinated contrast media is

the primary intervention with demonstrated effective-ness for the prevention of contrast-induced AKI inhigh-risk patients. We concur with the recommenda-tion to not use oral uids alone in patients at increasedrisk of contrast-induced AKI. Controversy exists re-garding the benet of isotonic sodium bicarbonateadministration compared to isotonic saline solution,with divergent results reported fro m the multiplesmall- to medium-sized clinical trials 91-99 and meta-analyses. 100-110 Given the divergent results of studiesto date and the recognition that none of the random-ized clinical trials to date have demonstrated thatisotonic bicarbonate is less effective than isotonicsaline solution, the KDIGO guideline recommenda-tion to use either isotonic uid in high-risk patients isappropriate. The guideline makes an important men-tion of the potential for compounding errors withsodium bicarbonate administration because no com-mercially available isotonic bicarbonate solutions areavailable, and the attendant risk for the administrationof hypertonic solutions. This risk is not associatedwith the provision of premixed saline solution. It isalso important to note that the additional time neededto prepare isotonic sodium bicarbonate solution is notrequired when administering standard saline solu-tions. Therefore, the administration of isotonic salinesolution may be preferable in situations in whichemergent contrast procedures are indicated.

Acute administration of both mannitol and furo-semide has been ass ociat ed with increased risk of contrast-inducedAKI. 63,64 However, a series of recenttrials examined the benet of generating high urineow rates using furosemide and saline solution infu-sions combined with a device that matches intrave-nous uid administration with urine output. 111,112

These trials, which randomly assigned patients to thisstrategy or to conventional periprocedural intrave-

Am J Kidney Dis. 2013;61(5):649-672662

Palevsky et al


15/24

nous uids, reported lower rates of contrast-inducedAKI among patients randomly assigned to this device.However, these studies were relatively small and didnot conrm whether sodium intake in the form of intravenous uids was matched to urinary sodiumexcretion in patients randomly assigned to the device.

Recognizing that urine composition following furo-semide administration is likely to be hypotonic rela-tive to saline, patients randomly assigned to thedevice-based protocol may have experienced net positivesodium balance, which could confound the interpreta-tion of the benets of this device.

A multitude of clinical trials have investigatedNAC for the prevention of contrast-induced AKI andgenerated, much like the studies comparing isotonicbicarbonate with isotonic saline solution, highly con-icting ndings. 113-138 These studies have been lim-ited by small study populations, clinically implausible

effect sizes, and a paucity of data for the effect of NAC on hard patient outcomes. Meta-analysesbasedon these trials also documented conicting ndingson the efcacy of NAC and underscore the limitationsin the primary clinical trials. 139-142 In 2011, a largerandomized controlled trial of 2,308 patients reportedno reduction in the rate of contrast-induced AKI with1,200 mg of oral NA C dosed daily for 2 days ascompared to placebo. 143 While this study is largerthan all prior published trials of NAC, several limita-tions to its design have been raised, including the factthat only 15.7% of study participants had a serumcreatinine level 1.5 mg/dL and the baseline serumcreatinine level may have been obtained up to 90 daysprior to the angiographic procedure. Intravenous NACis associated with potentially serious adverse effectsand, in the absence of sound data supporting itseffectiveness, should not be routinely administeredfor the prevention of contrast-induced AKI. However,oral NAC is inexpensive and largely devoid of ad-verse side effects in the doses employed to preventcontrast-induced AKI. Therefore, the recommenda-tion to administer oral NAC together with intravenousisotonic crystalloid is not inappropriate despite thequestions of its efcacy. However, oral NAC shouldnot be used in lieu of intravenous isotonic crystalloidin high-risk patients. While the KDIGO guidelinedoes not recommend a specic dose or duration of therapy, data suggesting a possible dose-dependenteffect justify the administration of 1,200 mg by mouthtwice daily for 2 days.

Our WorkGroup concurred with theKDIGO recom-mendations against the use of fenoldopam and theoph-ylline for prevention of contrast-induced AKI. Al-though preliminary uncontrolled trials of fenoldopamsuggested a benet with regard to prevention of con-trast-inducedAKI, this was not conrmed in a random-

ized controlled trial. 144 While studies have suggesteda potential benet to the adenosine antagonist theoph-ylline for the prevention of contrast-induced AKI, thedata to date are not conclusive. 145 Moreover, unlikeNAC, theophylline is associated with potential cardio-vascular side effects and has numerous drug interac-

tions. Our KDOQI Work Group also concurred withthe recommendations not to use prophylactic hemodi-alysis or hemoltration for contrast media removaland prevention of contrast-induced AKI. Althoughextracorporeal therapies are able to remove contrastmedia from the circulation, the rate of removal isunlikely to prevent the kidney damage, which devel-ops within minutes of contrast administration. Meta-analysis of clinical trials of hemodialysis and hemol-tration have demonstrated an absence of benet andpotential risk of harm with prophylactic hemodialy-sis. 146

ImplicationsWithinUSHealthCare1. All patients undergoing contrast-enhanced imag-

ing procedures should be evaluated for risk of AKI.Screening can most appropriately be performed usingstandardized questionnaires. Routine measurement of serum creatinine in patients identied as low risk byquestionnaire is not indicated; however, if there is anyquestion regarding risk of kidney disease, serum creat-inine level should be obtained prior to contrast admin-istration.

2. The risks and benets of contrast administration

need to be carefully evaluated in patients at high risk of contrast-induced AKI. The risk of contrast-inducedAKI should not preclude the performance of neededdiagnostic imaging and therapeutic procedures in high-risk patients.

3. Iso-osmolal or selected low-osmolal contrast inthe lowest possible dose should be used in high-risk patients.

4. The only intervention that has consistently beendemonstrated to decrease the risk of contrast-inducedAKI is periprocedural intravenous volume administra-tion using isotonic crystalloid. Thus, all patients at

increased risk of contrast-induced AKI should receiveperiprocedural intravenous isotonic crystalloid. Theoptimal rate of uid administration remains uncertain,and whether bicarbonate administration is associatedwith greater benet at risk reduction than saline re-mains unresolved.

5. The benet associated with NAC administrationremains uncertain. Given the minimal risk and costassociated with this agent, we do not recommendagainst its use, although it should not be used in lieuof more effective interventions, particularly periproce-dural administration of isotonic crystalloid.

Am J Kidney Dis. 2013;61(5):649-672 663

KDOQI Commentary


16/24

Box 4. Summary of KDIGO Recommendation Statements: Dialysis Interventions for Treatment of AKI

5.1.1: Initiate RRT emergently when life-threatening changes in uid, electrolyte, and acid-base balance exist. ( Not Graded )5.1.2: Consider the broader clinical context, the presence of conditions that can be modied with RRT, and trends of laboratory

testsrather than single BUN and creatinine thresholds alonewhen making the decision to start RRT. ( Not Graded )5.2.1: Discontinue RRT when it is no longer required, either because intrinsic kidney function has recovered to the point that it is

adequate to meet patient needs, or because RRT is no longer consistent with the goals of care. ( Not Graded )

5.2.2: We suggest not using diuretics to enhance kidney function recovery, or to reduce the duration or frequency of RRT. ( 2B )5.3.1: In a patient with AKI requiring RRT, base the decision to use anticoagulation for RRT on assessment of the patients potential

risks and benets from anticoagulation (see Figure 17). ( Not Graded )5.3.1.1: We recommend using anticoagulation during RRT in AKI if a patient does not have an increased bleeding risk or

impaired coagulation and is not already receiving systemic anticoagulation. (1B) 5.3.2: For patients without an increased bleeding risk or impaired coagulation and not already receiving effective systemic

anticoagulation, we suggest the following:5.3.2.1: For anticoagulation in intermittent RRT, we recommend using either unfractionated or low-molecular-weight heparin,

rather than other anticoagulants. (1C) 5.3.2.2: Foranticoagulation in CRRT, we suggest using regional citrate anticoagulationrather than heparin in patients who do

not have contraindications for citrate. (2B)5.3.2.3: For anticoagulation during CRRT in patients who have contraindications for citrate, we suggest using either

unfractionated or low-molecular-weight heparin, rather than other anticoagulants. (2C) 5.3.3: For patients with increased bleeding risk who are not receiving anticoagulation, we suggest the following for anticoagulation

during RRT:

5.3.3.1: We suggest using regional citrate anticoagulation, rather than no anticoagulation, during CRRT in a patient withoutcontraindications for citrate. (2C)

5.3.3.2: We suggest avoiding regional heparinization during CRRT in a patient with increased risk of bleeding. (2C) 5.3.4: In a patient with heparin-induced thrombocytopenia (HIT), all heparin must be stopped and we recommend using direct

thrombin inhibitors (such as argatroban) or Factor Xa inhibitors (such as danaparoid or fondaparinux) rather than other or noanticoagulation during RRT. ( 1A)5.3.4.1: In a patient with HIT who does not have severe liver failure, we suggest using argatroban rather than other thrombin

or Factor Xa inhibitors during RRT. ( 2C )5.4.1: We suggest initiating RRTin patientswith AKIviaan uncuffednontunneleddialysis catheter, rather thana tunneledcatheter. (2D) 5.4.2: When choosing a vein for insertion of a dialysis catheter in patients with AKI, consider these preferences (Not Graded ):

First choice: right jugular vein; Second choice: femoral vein; Third choice: left jugular vein; Last choice: subclavian vein with preference for the dominant side.

5.4.3: We recommend using ultrasound guidance for dialysis catheter insertion. (1A) 5.4.4: We recommend obtaining a chest radiograph promptly after placement and before rst use of an internal jugular or subclavian

dialysis catheter. (1B) 5.4.5: We suggest not using topical antibiotics over the skin insertion site of a nontunneled dialysis catheter in ICU patients with AKI

requiring RRT. (2C) 5.4.6: We suggest not using antibiotic locks for prevention of catheter-related infections of nontunneled dialysis catheters in AKI

requiring RRT. (2C) 5.5.1: We suggest to use dialyzers with a biocompatible membrane for IHD and CRRT in patients with AKI. (2C)5.6.1: Use continuous and intermittent RRT as complementary therapies in AKI patients. (Not Graded) 5.6.2: We suggest using CRRT, rather than standard intermittent RRT, for hemodynamically unstable patients. (2B) 5.6.3: We suggest using CRRT, rather than intermittent RRT, for AKI patients with acute brain injury or other causes of increased

intracranial pressure or generalized brain edema. ( 2B )5.7.1: We suggest using bicarbonate, rather than lactate, as a buffer in dialysateandreplacementuid forRRT in patients with AKI. (2C) 5.7.2: We recommend using bicarbonate, rather than lactate, as a buffer in dialysate and replacement uid for RRT in patients with

AKI and circulatory shock. (1B) 5.7.3: We suggest using bicarbonate, rather than lactate, as a buffer in dialysate and replacement uid for RRT in patients with AKI

and liver failure and/or lactic acidemia. (2B) 5.7.4: We recommend that dialysis uids and replacement uids in patients with AKI, at a minimum, comply with American

Association of Medical Instrumentation (AAMI) standards regarding contamination with bacteria and endotoxins. (1B) 5.8.1: The dose of RRT to be delivered should be prescribed before starting each session of RRT. (Not Graded) We recommend

frequent assessment of the actual delivered dose in order to adjust the prescription. (1B) 5.8.2: Provide RRT to achieve the goals of electrolyte, acid-base, solute, and uid balance that will meet the patients needs. (Not

Graded )5.8.3: We recommend delivering a Kt/V of 3.9 per week when using intermittent or extended RRT in AKI. (1A) 5.8.4: We recommend delivering an efuent volume of 20-25 mL/kg/h for CRRT in AKI (1A). This will usually require a higher

prescription of efuent volume. (Not Graded)

Abbreviations: AKI, acute kidney injury; BUN, blood urea nitrogen; CRRT, continuous renal replacement therapy; ICU, intensivecare unit; RRT, renal replacement therapy.


Am J Kidney Dis. 2013;61(5):649-672664

Palevsky et al


17/24

6. Fenoldopam, theophylline, and prophylactic he-modialysis and hemoltration should not be used todecrease the risk of contrast-induced AKI.

Dialysis InterventionsforTreatmentofAKICommentary

In the KDIGO AKI guideline, there are 33 recom-mendation statements on dialysis interventions fortreatment of AKI; 9 are not graded, 10 are level 1recom mendat ions, and 14 are level 2 recommenda-tions ( Box 4 ). The 9 recommendations that are notgraded are largelypractical suggestions with the excep-tion of the recommendation for dialysis catheter-siteselection. There are data to suggest that the risk of subclavian stenosis is greater with large, nontunneledcatheters and data for the association of catheter sitewith catheter fun ction and risk of infection from theCathedia trial. 147 Our Work Group recommends that

in addition to site selection, dialysis catheters shouldbe of an adequate length to minimize the risks of access recirculation and catheter malfunction. Theremaining recommendations are largely practicalstatements about the initiation and termination of dialysis, as well as the use of anticoagulation. TheWork Group thought that it was important to notethe paucity of evidence to support a recommenda-tion regarding the timing of initiation of RRT inAKI and believes that this is an area in need of well-designed clinical trials.

The KDOQI Work Group also agreed with the 10

level 1 recommendations. These included using anti-coagulation for patients who are receiving RRT whoare not at increased risk of bleeding, using heparin foranticoagulation during intermittent hemodialysis,avoiding heparin products in patients with heparin-induced thrombocytopenia, using ultrasound for dialy-sis catheter insertion and chest radiography to conrmline placement, and appropriate dosing of bothintermittent and continuous RRTs. With regard tothe level 1 recommendation to use bicarbonate-containing dialysate and replacement uid in pa-tients with shock, the KDOQI Work Group notedthat bicarbonate has replaced lactate and acetate asthe dialysate buffer of choice for intermittent RRT.The evidence of the benet of bicarbonate-basedover lactate -based solutions for CRRT is inconsis-tent. 148-150 Nevertheless, the increased availabilityof commercially prepared bicarbonate-based CRRTuids in the United States seems to support theiruse as the buffer of choice.

The recommendations regarding dose of RRT inAKI are based upon some of the most rigorous evi-dence for the entire guideline. The RENAL(Random-ized Evaluation of Normal Versus Augmented Level)Replacement Therapy study, conducted in Australia

and New Zealand, randomly assigned 1,508 patie-nts to continuous venovenous hemodialtration(CVVHD F) at an efuent ow rate of either 40 or 25mL/kg/h. 12 The VA/NIH Acute Renal Failure TrialNetwork (ATN) study randomly assigned 1,124 pa-tients in the United States to a strategy of more

intensive (CVVHDF at 35 mL/kg/h or intermittenthemodialysis or SLED [sustained low-efciency dialy-sis] on a 6-day-per-week schedule) or less intensive(CVVHDF at 20 mL/kg/h or intermittent hemodialy-sis or SLED on a 3-day-per-week schedule) RRT. 11 Inthe ATN study, patients moved between modality of therapy as their hemodynamic status changed. In bothstudies, there was no added benet with regard tosurvival or recovery of kidney function associatedwith more intensive therapy. Although the KDIGOGuideline bases their target doses of therapy for bothintermittent and continuous therapy on the results of

these studies, it should be recognized that neitherstudy was designed to identify the minimum adequatedose of therapy. In addition, it should also be notedthat the KDIGO recommended weekly delivered Kt/Vof 3.9 is based upon the arithmetic sum of the dose perindividual treatment. Kinetic modeling suggests thatthis is erroneous and that prescription of intermittenthemodialysis to provide a Kt/V of 1.3 three times perweek or a Kt/V of 0.65 six times per week are notequivalent. 151 Available evidence suggests that deliv-ery of RRT often falls short of the prescribed dose, 152-155

supporting the need for frequent assessment of theactual delivered dose.

With regard to the level 2 recommendations, ourWork Group agreed that diuretics should not be usedspecically for improving kidney function or reduc-ing the need for RRT. However, with regard to the useof anticoagulation for CRRT, the Work Group be-lieved that there is no consensus on which anticoagu-lant should be rst choice for CRRT. Rather, thechoice of anticoagulant for CRRT should be deter-mined by patient characteristics, local expertise, nurs-ing comfort, ease of monitoring, and pharmacy issues.Monitoring should include evaluation of anticoagu-lant effect, lter efcacy, and circuit life and compli-cations. Although regional citrate anticoagulation isgaining wide acceptance w ith the development of simpler and safer protocols, 156,157 citrate is not ap-proved by the US Food and

aki commentary 2013

Documents