KDOQI Clinical Practice Guideline for Diabetes and CKD ...€¦ · Minnesota Center for Obesity, Metabolism and Endocrinology, PA (MNCOME) Eagan, MN, USA ... Table 1. Grade for Strength

KDOQI CLINICAL PRACTICE GUIDELINE FOR DIABETES

AND CKD: 2012 UPDATE

Abstract

The 2012 update of the Kidney Disease Outcomes Quality Initiative (KDOQI) Clinical Practice Guideline forDiabetes and Chronic Kidney Disease (CKD) is intended to assist the practitioner caring for patients withdiabetes and CKD. Substantial high-quality new evidence has emerged since the original 2007 KDOQI guidelinethat could significantly change recommendations for clinical practice. As such, revisions of prior guidelines areoffered that specifically address hemoglobin A1c (HbA1c) targets, treatments to lower low-density lipoproteincholesterol (LDL-C) levels, and use of angiotensin-converting enzyme inhibitor (ACE-I) and angiotensinreceptor blocker (ARB) treatment in diabetic patients with and without albuminuria. Treatment approaches areaddressed in each section and the stated guideline recommendations are based on systematic reviews of relevanttrials. Appraisal of the quality of the evidence and the strength of recommendations followed the Grading ofRecommendation Assessment, Development, and Evaluation (GRADE) approach. Limitations of the evidenceare discussed and specific suggestions are provided for future research.

Keywords: Albuminuria; chronic kidney disease; Clinical Practice Guideline; diabetes; dyslipidemia; evidence-based recommendation; KDOQI.

In citing this document, the following format should be used: National Kidney Foundation. KDOQI ClinicalPractice Guideline for Diabetes and CKD: 2012 update. Am J Kidney Dis. 2012;60(5):850-886.

Am J Kidney Dis. 2012;60(5):850-886850

NOTICE

SECTION I: USE OF THE CLINICAL PRACTICE GUIDELINE

This Clinical Practice Guideline is based upon a systematic literature search that included articles publishedthrough October 2010 and upon the best information available from relevant newer publications and scientificpresentations through April 2012. It is designed to provide information and assist decision making. It is notintended to define a standard of care, and should not be construed as one, nor should it be interpreted asprescribing an exclusive course of management. Variations in practice will inevitably and appropriately occurwhen clinicians take into account the needs of individual patients, available resources, and limitations unique toan institution or type of practice. Every health-care professional making use of these recommendations isresponsible for evaluating the appropriateness of applying them in any particular clinical situation. Therecommendations for research contained within this document are general and do not imply a specific protocol.

SECTION II: DISCLOSURE

Kidney Disease Outcomes Quality Initiative (KDOQI) makes every effort to avoid any actual or reasonablyperceived conflicts of interest that may arise as a result of an outside relationship or a personal, professional, orbusiness interest of a member of the Work Group. All members of the Work Group are required to complete, sign,and submit a disclosure and attestation form showing all such relationships that might be perceived or actualconflicts of interest. This document is updated annually and information is adjusted accordingly. All reportedinformation is on file at the National Kidney Foundation (NKF).

Am J Kidney Dis. 2012;60(5):850-886 851

Work Group Membership

Work Group Co-Chairs

Robert G. Nelson, MD, PhDNational Institutes of Health

Phoenix, AZ, USA

Katherine R. Tuttle, MD, FASN, FACPProvidence Medical Research Center

University of Washington School of MedicineSpokane, WA, USA

Work Group

Rudolph W. Bilous, MDThe James Cook University Hospital

Middlesbrough, UK

J. Michael Gonzalez-Campoy, MD, PhD, FACEMinnesota Center for Obesity, Metabolism and

Endocrinology, PA (MNCOME)Eagan, MN, USA

Michael Mauer, MDUniversity of Minnesota Medical School

Minneapolis, MN, USA

Mark E. Molitch, MDNorthwestern University

Chicago, IL, USA

Kumar Sharma, MD, FAHAUniversity of California San Diego

La Jolla, CA, USA

Liasion Members

Judith E. Fradkin, MDNational Institutes of Health

Bethesda, MD, USA

Andrew S. Narva, MDNational Institutes of Health

Bethesda, MD, USA

KDOQI Evidence Review TeamUniversity of Minnesota Department of Medicine

Minneapolis VA Center for Chronic Disease Outcomes Research. Minneapolis, MN, USA:Timothy J. Wilt, MD, MPH, Professor of Medicine and Project Director

Areef Ishani, MD, MS, Chief, Section of Nephrology, Associate Professor of MedicineThomas S. Rector, PhD, PharmD, Professor of MedicineYelena Slinin, MD, MS, Assistant Professor of Medicine

Patrick Fitzgerald, MPH, Project ManagerMaureen Carlyle, MPH, PIVOT Coordinator

Am J Kidney Dis. 2012;60(5):850-886852

KDOQI Leadership

Michael V. Rocco, MD, MSCEKDOQI Chair

Jeffrey S. Berns, MDVice Chair, Guidelines and Commentary

Joseph V. Nally, Jr, MDVice Chair, Public Policy

Holly Kramer, MDVice Chair, Research

Michael J. Choi, MDVice Chair, Education

NKF-KDOQI Guideline Development StaffKerry Willis, PhD, Senior Vice-President for Scientific Activities

Emily Howell, MA, Communications DirectorMichael Cheung, MA, Guideline Development Director

Sean Slifer, BA, Guideline Development Manager

Am J Kidney Dis. 2012;60(5):850-886 853

Table of ContentsAbbreviations and Acronyms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 855Reference Keys . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 856Foreword . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 857Executive Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 858Summary of Recommendations. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 861Guideline 2: Management of Hypergylcemia and General Diabetes Care in CKD . . . . . . . . . . . . . . . . . . . 862Guideline 4: Management of Dyslipidemia in Diabetes and CKD . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 868Guideline 6: Management of Albuminuria in Normotensive Patients with Diabetes . . . . . . . . . . . . . . . . . . 873Research Recommendations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 875Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 877Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 878Biographic and Disclosure Information. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 879References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 883

TablesTable 1. Grade for Strength of Recommendation in the Diabetes and CKD Guideline . . . . . . . . . . . . . . 859Table 2. Grade for Quality of Evidence in the Diabetes and CKD Guideline . . . . . . . . . . . . . . . . . . . . . 860Table 3. Target and Achieved HbA1c Levels in the Intensively and Conventionally Treated Groups of

Three Recent Clinical Trials that Examined Different Levels of Glycemic Control in Patientswith Type 2 Diabetes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 863

Table 4. Dose Adjustment for Insulin Compounds and Oral Medicines for Diabetes in CKD . . . . . . . . 865Table 5. Summary of Four Post Hoc Analyses Reports of Lipid Lowering in People with Diabetes

Mellitus (DM) and CKD. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 869Table 6. Fibrate Treatment in Patients with Diabetes Mellitus (DM) and CKD . . . . . . . . . . . . . . . . . . . 870Table 7. Dose Adjustment for Lipid Lowering Medicines in CKD . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 871

FigureFigure 1. Key Questions (KQ) to Be Addressed By the Evidence Review . . . . . . . . . . . . . . . . . . . . . . . . 858

Am J Kidney Dis. 2012;60(5):850-886854

Abbreviations and Acronyms

4D Deutsche Diabetes Dialyse Studie4S Scandinavian Simvastatin Survival StudyACCORD Action to Control Cardiovascular Risk in DiabetesACE Angiotensin-converting enzymeACE-I Angiotensin-converting enzyme inhibitorAdDIT Adolescent type 1 Diabetes cardio-renal Intervention TrialADVANCE Action in Diabetes and Vascular Disease: Preterax and Diamicron Modified Release Controlled

EvaluationALERT Assessment of Lescol in Renal TransplantALTITUDE Aliskiren Trial in Type 2 Diabetes Using Cardiovascular and Renal EndpointsARB Angiotensin receptor blockerAURORA A study to evaluate the Use of Rosuvastatin in subjects On Regular hemodialysis: an Assessment of

survival and cardiovascular eventsAVOID Aliskiren in the Evaluation of Proteinuria in DiabetesCARDS Collaborative Atorvastatin Diabetes StudyCARE Cholesterol and Recurrent EventsCI Confidence intervalCKD Chronic kidney diseaseCVD Cardiovascular diseaseDAIS Diabetes Atherosclerosis Intervention StudyDCCT The Diabetes Control and Complications TrialDKD Diabetic kidney diseaseDM Diabetes mellitusDPP-4 Dipeptidyl peptidase-4EDIC Epidemiology of Diabetes Interventions and ComplicationseGFR Estimated glomerular filtration rateESRD End-stage renal diseaseFDA Food and Drug AdministrationFIELD Fenofibrate Intervention and Event Lowering in DiabetesGFR Glomerular filtration rateGLP-1 Glucagon-like peptide-1GRADE Grading of Recommendation Assessment, Development, and EvaluationHbA1c Hemoglobin A1cHDL-C High-density lipoprotein cholesterolHPS Heart Protection StudyHR Hazard ratioJNC Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood

PressureKDIGO Kidney Disease: Improving Global OutcomesKDOQI Kidney Disease Outcomes Quality InitiativeKQ Key questionLDL-C Low-density lipoprotein cholesterolLIPID Long-term Intervention with Pravastatin in Ischemic DiseaseMI Myocardial infarctionMICROHOPE Microalbuminuria, Cardiovascular, and Renal Outcomes in Heart Outcomes Prevention EvaluationNHANES National Health and Nutrition Examination SurveyNKF National Kidney FoundationRAS Renin-angiotensin systemRR Relative riskSCr Serum creatinineSHARP Study of Heart and Renal ProtectionTNT Treating to New TargetsUKPDS UK Prospective Diabetes StudyUSRDS United States Renal Data SystemVADT Veterans Affairs Diabetes TrialVA-HIT Veterans Affairs High-density lipoprotein Intervention Trialvs. VersusWOSCOPS West of Scotland Coronary Prevention Study

Am J Kidney Dis. 2012;60(5):850-886 855

Reference Keys

CKD NOMENCLATURE USED BY KDOQI

CKD Categories Definition

CKD CKD of any stage (1-5), with or without a kidney transplant, including bothnon–dialysis dependent CKD (CKD 1–5ND) and dialysis-dependent CKD(CKD 5D)

CKD ND Non–dialysis-dependent CKD of any stage (1-5), with or without a kidneytransplant (i.e., CKD excluding CKD 5D)

CKD T Non–dialysis-dependent CKD of any stage (1-5) with a kidney transplant

Specific CKD StagesCKD 1, 2, 3, 4 Specific stages of CKD, CKD ND, or CKD T

CKD 3-4, etc. Range of specific stages (e.g., both CKD 3 and CKD 4)

CKD 5D Dialysis-dependent CKD 5

CKD 5HD Hemodialysis-dependent CKD 5

CKD 5PD Peritoneal dialysis–dependent CKD 5

Am J Kidney Dis. 2012;60(5):850-886856

Foreword

This publication of the Kidney Diseases Out-comes Quality Initiative (KDOQI) updatesseveral areas of the 2007 KDOQI Clinical PracticeGuidelines and Clinical Practice Recommendationsfor Diabetes and Chronic Kidney Disease. The needfor this update was the result of increasing recogni-tion that substantial high-quality new evidence hadbecome available since 2007 that could signifi-cantly change recommendations for clinical prac-tice. Using the usual rigorous analytical methods ofthe KDOQI process, an outstanding Work Group,under the leadership of Robert Nelson and Kather-ine Tuttle, working with the Minneapolis VeteransAdministration Center for Chronic Disease Out-comes Research, reviewed new studies addressingmanagement of hyperglycemia, hyperlipidemia, andalbuminuria in individuals with diabetes mellitusand chronic kidney disease (CKD). Their analysisfocuses on important outcomes such as all-causemortality, CKD progression and development ofend-stage renal disease (ESRD), fatal and non-fatalcardiovascular events, among others. Revisions ofprior guidelines are offered that specifically address

Am J Kidney Dis. 2012;60(5):850-886

HbA1c targets, treatments to lower LDL-C levels,and use of ACE-I and ARB treatment in diabeticpatients with and without albuminuria. The newguideline updates published here are each accompa-nied by an indication of the strength and quality ofsupporting evidence. Five of seven of these recom-mendations carry 1A or 1B grades indicative of thestrength of these new recommendations and thequality of evidence supporting them. Finally, impor-tant research recommendations are proposed.

As with prior KDOQI efforts, Drs Tuttle and Nel-son and members of the Work Group devoted count-less hours, all voluntarily, to the development of thisimportant document. To each of them, and to all theothers involved in this effort, we offer our mostsincere thanks for their dedication and commitment tohelping us all provide the very best care possible tothe many patients with diabetes mellitus and CKD.

Michael V. Rocco, MD, MSCEKDOQI Chair

Jeffrey S. Berns, MDVice Chair, Guidelines and Commentary

© 2012 by the National Kidney Foundation, Inc.0272-6386/$36.00

http://dx.doi.org/10.1053/j.ajkd.2012.07.005

857

http://dx.doi.org/10.1053/j.ajkd.2012.07.005

Executive Summary

INTRODUCTION

Chronic kidney disease (CKD) is a worldwide publichealth problem that affects millions of people from allracial and ethnic groups. Diabetes mellitus (henceforthreferred to as diabetes) is the leading cause of CKD, andthe rapidly increasing prevalence of diabetes worldwidevirtually assures that the proportion of CKD attributableto diabetes will continue to rise. Indeed, a recent reportfrom the National Health and Nutrition Education Sur-vey (NHANES) found that prevalence of diabetic kid-ney disease (DKD) increased steadily from 1988 through2008, and the latest United States Renal Data System(USRDS) report indicates a �30% increase in incidenceof ESRD in persons with diabetes in the USA between1992 and 2008.1,2

In 1997, as part of an effort to address the growingproblem of CKD, the National Kidney Foundation (NKF)established the Kidney Disease Outcomes Quality Initia-tive (KDOQI) to develop clinical practice guidelines forthe management of all stages of CKD.3 By 2007, withthe publication of the KDOQI Clinical Practice Guide-lines and Clinical Practice Recommendations for Diabe-tes and Chronic Kidney Disease,4 the KDOQI reachedits primary goal of producing evidence-based guidelineson the aspects of CKD most likely to improve care forpatients.5 To ensure that practitioners and patients ben-efit from the latest knowledge, an essential part of

KQ 1: In patients with diabetes (type 1 or 2control (as defined by lower target goutcomes compared to controls?

KQ 2: What harms result from more intens

(type 1 or 2)? KQ 3: In patients with diabetes (type 1 or 2

lipid management targets (defined atriglycerides) that improve health ou

KQ 4: Is there evidence for specific lipid al

or 2) and CKD? KQ 5: What harms result from more intens

altering agents in individuals with dia KQ 6: What interventions prevent incident

patients with diabetes in whom furthtreatment objective?

KQ 7: Is albuminuria a valid surrogate for h

Figure 1. Key questions (KQ) to be addressed by the evid

high-density lipoprotein cholesterol; LDL-C; low-density lipoprotein ch

858

KDOQI activities is to provide regular updates of theseguidelines.

Since publication of the diabetes guidelines in 2007,several large well-designed clinical trials have addressedmanagement issues relevant to patients with diabetesand CKD. Findings from these trials suggest that theexisting guideline recommendations for the manage-ment of hyperglycemia, hypertension, hyperlipidemia,and albuminuria may no longer accurately reflect currentmedical knowledge. To properly incorporate the newfindings from these clinical trials and other recent studiesinto a guideline update, a systematic review of the newevidence was warranted to formally determine theirapplicability and methodologic quality.

This report describes updates of guidelines for themanagement of hyperglycemia, hyperlipidemia, and al-buminuria in patients with diabetes and kidney diseaseas a result of this systematic review. An update of theguideline for management of blood pressure is presentlyunderway by Kidney Disease: Improving Global Out-comes (KDIGO), an independent not-for-profit founda-tion governed by its own international board of directors.KDIGO was established to improve international coop-eration in the development, dissemination, and implemen-tation of clinical practice guidelines.6 KDOQI andKDIGO work in concert to expand the scope of guide-lines relevant to the care of patients with CKD andimprove the care of these patients worldwide.5

or without CKD, does intensive glycemic ylated hemoglobin) improve health

emic control in individuals with diabetes

CKD, what evidence is there for specific ls for total cholesterol, LDL-C, HDL-C, s?

agent use for patients with diabetes (type 1

management or use of specific lipid s (type 1 or 2) and CKD?

inuria and/or progression of albuminuria in uction in blood pressure is not the specific

outcomes in diabetes?

review. Abbreviations: CKD, chronic kidney disease; HDL-C,

), withlycos

e glyc

) ands goatcome

tering

e lipidbete

albumer red

ealth

ence

olesterol.

Am J Kidney Dis. 2012;60(5):850-886

KDOQI Diabetes Guideline: 2012 Update

METHODS

The guideline update effort was a voluntary andmultidisciplinary undertaking that included input fromNKF scientific staff, an evidence review team from theMinneapolis VeteransAdministration Center for ChronicDisease Outcomes Research, and a Work Group ofexperts in relevant disciplines. The approach to thesystematic literature review and the comprehensive find-ings prepared for this update are reported in detailelsewhere.7 Briefly, MEDLINE was searched to identifyrandomized controlled trials published between January2003 and October 2010 that related to albuminuria,glycemic and lipid management in patients with diabe-tes. All titles and abstracts were assessed for their appro-priateness to address key questions that were developedby the multidisciplinary team and outlined in Fig 1.Study reference lists, reviews, and meta-analyses wereevaluated and references to other clinical trials wereelicited from members of the Work Group. Data fromeach study that pertained to study quality, trial character-istics, population characteristics, efficacy, outcomes, with-drawals, and adverse events were extracted. Evidencetables were created to address the key questions. Studyquality was rated as good, fair, or poor according tocriteria suggested by the Cochrane Collaboration, andincluded information on adequate allocation conceal-ment, method of blinding, use of the intention-to-treatprinciple for data analysis, reporting of dropouts, andreasons for attrition.

In formulating the guideline statements, separaterecommendation levels (1 or 2) were assigned foreach specific recommendation based on the overallstrength of the recommendation and separate lettergrades (A, B, C, or D) were assigned based on the

Table 1. Grade for Strength of Recomm

Grade* Patients

Level 1“We recommend”

Most people in your situationwould want therecommended course ofaction and only a smallproportion would not.

Mosttheof a

Level 2“We suggest”

The majority of people inyour situation would wantthe recommended courseof action, but many wouldnot.

Differapppatneemaconval

*The additional category “Not Graded” is used, typically, to proallow adequate application of evidence. The most common ecounseling, and referral to other clinical specialists. The ungr

statements, but are not meant to be interpreted as being stronger reco

Am J Kidney Dis. 2012;60(5):850-886

overall quality of the evidence for a particular interven-tion and outcome (Tables 1 and 2).8Strength of guide-line recommendations was determined by the GRADEapproach used by KDIGO. The overall quality andstrength of evidence was assessed using methodologydeveloped by the Agency for Healthcare Research andQuality and the Effective Health Care Program. Qual-ity of evidence ratings included four categories: A)high confidence, which indicated that further researchwas unlikely to change the confidence in the estimateof effect; B) moderate confidence, which indicatedthat further research may change the confidence in theestimate of effect; C) low confidence, which indicatedthat further research would likely have an importantimpact on the confidence in the estimate of effect; andD) insufficient, which indicated that the evidence wasunavailable or did not permit a conclusion.

Outcomes

The primary health outcome examined in this re-view was all-cause mortality. Secondary health out-comes included ESRD and cardiovascular death, non-fatal cardiovascular events, clinically significantretinopathy including vision loss, amputations, andsymptomatic hypoglycemia of sufficient severity torequire the assistance of another person. Intermediateoutcomes examined included changes in the level ofalbuminuria and glomerular filtration rate, doublingof serum creatinine (SCr) concentration, and progres-sion to CKD stage 4 or higher.7 The impact of treat-ments described in the recent clinical trials on thesehealth and intermediate outcomes was assessed informulating the guideline statements.

tion in the Diabetes and CKD Guideline

Implications

linicians Policy

nts should receivemmended course.

The recommendation can be evaluatedas a candidate for developing apolicy or a performance measure.

hoices will beate for different. Each patientelp to arrive at ament decisionnt with her or hisnd preferences.

The recommendation is likely to requiresubstantial debate and involvementof stakeholders before policy can bedetermined.

guidance based on common sense or where the topic does notles include recommendations regarding monitoring intervals,recommendations are generally written as simple declarative

enda

C

patierecoction

ent cropriientsds h

nagesiste

ues a

videxamp

aded

mmendations than Level 1 or 2 recommendations.

859

e of e


Nomenclature

Guideline statements have evolved since the pub-lication of the original diabetes guideline. Themoral imperative that clinicians “should” imple-ment a particular treatment was replaced by “Werecommend” if the strength of the recommendationwas strong or moderately strong and “We suggest”if the strength of the recommendation was weak.8

This change was made to reflect the uncertaintiesinherent to all research findings and the need toadjust any recommendations to the needs of theindividual patient.

GUIDELINE STATEMENTS

The customary practice of the NKF when theoriginal diabetes guideline was published was todivide the statements into clinical practice guide-

Table 2. Grade for Quality of Evide

Grade Quality of Evidence

A High We are confi

B Moderate The true effepossibility

C Low The true effe

D Very low The estimat

lines and clinical practice recommendations. The

860

guideline statements were based on a consensuswith the Work Group that the strength of the evi-dence was sufficient to make definitive statementsabout appropriate clinical practice. When thestrength of the evidence was not sufficient to makesuch statements, the Work Group offered recommen-dations based on the best available evidence andexpert opinion. The original document containedfive clinical practice guidelines and four clinicalpractice recommendations; updates for two clinicalpractice guidelines and one clinical practice recom-mendation are reported herein. The NKF now com-bines these statements and refers to them all as aclinical practice guideline, while specifying thestrength of each recommendation and its underlyingquality of evidence. Hence, Clinical Practice Recom-mendation 1 in the original document is now referred

in the Diabetes and CKD Guideline

Meaning

that the true effect lies close to that of the estimate of the effect.

likely to be close to the estimate of the effect, but there is ait is substantially different.

ay be substantially different from the estimate of the effect.

ffect is very uncertain, and often will be far from the truth.

nce

dent

ct isthat

ct m

to as Clinical Practice Guideline 6 in this update.

Am J Kidney Dis. 2012;60(5):850-886


Summary of Recommendations

Guideline 2: Management of Hyperglycemia and General Diabetes Care in CKDHyperglycemia, the defining feature of diabetes, is a fundamental cause of vascular target organ

complications, including diabetic kidney disease (DKD). Intensive treatment of hyperglycemiaprevents elevated albuminuria or delays its progression, but patients treated by approachesdesigned to achieve near normal glycemia may be at increased risk of severe hypoglycemia.Evidence that intensive treatment has an effect on loss of glomerular filtration rate (GFR) is sparse.

2.1: We recommend a target hemoglobin A1c (HbA1c) of �7.0% to prevent or delay progression ofthe microvascular complications of diabetes, including DKD. (1A)

2.2: We recommend not treating to an HbA1c target of 30 mg/g who are at high risk of DKD or its progression. (2C)

Am J Kidney Dis. 2012;60(5):850-886 861


Guideline 2: Management of Hyperglycemia and General

Diabetes Care in CKD

Hyperglycemia, the defining feature of diabetes,is a fundamental cause of vascular target organcomplications, including diabetic kidney disease(DKD). Intensive treatment of hyperglycemia pre-vents elevated albuminuria or delays its progres-sion, but patients treated by approaches designedto achieve near normal glycemia may be at in-creased risk of severe hypoglycemia. Evidence thatintensive treatment has an effect on loss of glomer-ular filtratin rate (GFR) is sparse.

2.1: We recommend a target HbA1c of �7.0% toprevent or delay progression of the micro-vascular complications of diabetes, includ-ing DKD. (1A)

The evidence that achieving an HbA1c level of�7.0% is able to prevent the microvascular complica-tions of diabetes was presented in detail in the originalKDOQI diabetes guideline.4 For type 1 diabetes,evidence from the Diabetes Control and Complica-tions Trial (DCCT),9,10 as well as from a meta-analysis of a number of smaller studies that precededthe DCCT,11 established that this level of glycemiccontrol decreases the risk of microalbuminuria andretinopathy compared to less stringent control. Thebeneficial effects of intensive therapy on these out-comes persisted during the long-term follow-up studyof the DCCT subjects, called the Epidemiology ofDiabetes Interventions and Complications (EDIC)Study. Despite the gradual narrowing of the differencein HbA1c levels between the two DCCT groups overthe first two years in the follow-up period, and levelsremaining near 8% for both groups for the subsequent12 years, the reduction in risk of microvascular com-plications of diabetes persisted.12 Similar benefits ofglycemic control on the development of microalbu-minuria in patients with type 2 diabetes were origi-nally observed in three studies; the KumamotoStudy,13,14 the United Kingdom Prospective Diabe-tes Study (UKPDS),15,16 and the Veterans AffairsCooperative Study on Glycemic Control and Compli-cations in Type 2 Diabetes Feasibility Trial.17 Inten-sive glycemic control also significantly reduced thedevelopment of macroalbuminuria in patients withtype 1 diabetes, as shown in the DCCT/EDICStudy9,10,12 as well as the similarly designed butsmaller Stockholm study,18 and in those with type 2diabetes, as shown in the Kumamoto study13,14 andthe VA Cooperative Study.17 The UKPDS showed a

trend toward decreased development of macroalbumin-

862

uria, but this result did not achieve statistical signifi-cance.15,16

Three new studies have added to the evidence thateven more intensive glycemic control reduces thedevelopment of elevated albuminuria in patients withtype 2 diabetes. In the Action in Diabetes and VascularDisease: Preterax and Diamicron Modified ReleaseControlled Evaluation (ADVANCE) trial, more inten-sive control that achieved an HbA1c of 6.5%, com-pared with standard control (HbA1c 7.3%), was asso-ciated with a 21% reduction in new onset or worseningnephropathy defined by new onset macroalbuminuria,doubling of SCr, need for kidney replacement therapy,or death due to kidney disease (4.1% vs. 5.2%).Additionally, intensive glycemic control reduced de-velopment of macroalbuminuria by 30% (2.9% vs.4.1%), and development of new onset microalbumin-uria by 9% (23.7% vs. 25.7%).19 The Action toControl Cardiovascular Risk in Diabetes (ACCORD)study similarly showed that more intensive control,achieving an HbA1c of 6.4%, compared with standardcontrol (HbA1c 7.6%), resulted in a 32% reduction inthe development of incident macroalbuminuria (2.7%vs. 3.9%) and a 21% reduction in the development ofincident microalbuminuria (12.5% vs. 15.3%).20 Inthe Veterans Affairs Diabetes Trial (VADT), moreintensive glycemic control that achieved an HbA1c of6.9% compared with standard control (HbA1c 8.4%)resulted in a 37% reduction in macroalbuminuria(7.6% vs.12.1%) and a 32% reduction in microalbu-minuria (10.0% vs.14.7%).21

A few long-term observational cohort studies andsecondary or post hoc analyses of interventional stud-ies using ACE-Is or ARBs found that poorer glycemiccontrol is associated with a greater rate of fall of GFRin patients with type 1 diabetes.22-26 Most of theprospective, randomized studies used as evidence forthe effect of glycemic control on kidney function arelimited by the small numbers of patients reaching thisintermediate outcome. However, the EDIC/DCCT fol-low-up study recently reported that 2.0% (1.6/1000person-years) of participants in the previously inten-sive treatment group and 5.5% (3.0/1000 person-years) of those in the previously conventional treat-ment group developed sustained estimated glomerularfiltration rate (eGFR) measurements �60 mL/min/1.73 m2 with a relative risk (RR) reduction of 50%(p�0.006); there were similar RR reductions for singleeGFR measurements �45 mL/min/1.73 m2 (50%,

1.6/1000 person-years vs. 2.5/1000 person-years,

Am J Kidney Dis. 2012;60(5):850-886

VAD


p�0.045) and �30 mL/min/1.73 m2 (44%, 0.8/1000person-years vs. 1.5/1000 person-years, p�0.088) andfor ESRD (51%, 0.5/1000 person-years vs. 1.1/1000person-years, p�0.098).27 For patients with type 2diabetes, intensive treatment in the UKPDS was asso-ciated with a 67% risk reduction for a doubling ofplasma creatinine levels at 9 years (0.71% of theintensive group and 1.76% of the conventional group,p�0.027).15 None of the three more recent studiesmentioned above (ADVANCE, ACCORD, VADT)showed significant benefits of more intensive glyce-mic control on creatinine-based estimates of GFR.19-21

Accordingly, the evidence that intensive glycemiccontrol reduces the microvascular complications ofdiabetes is based almost exclusively on prevention ofmicroalbuminuria (a predictor of actual complica-tions), reduced progression to macroalbuminuria, andon prevention of retinopathy. Evidence for the preven-tion of other intermediate microvascular outcomes,including declining eGFR and doubling of SCr, issparse. Although there is no evidence that intensiveglycemic control slows progression to the clinicalendpoint of ESRD, it is likely that if the earliermanifestations of kidney disease are reduced (i.e.,albuminuria and earlier-stage CKD), then the eventualoutcome of ESRD will also be reduced. However,such assumption presumes that benefits of intensiveglycemic control are not outweighed by harms andthat patients survive to reach ESRD.

2.2: We recommend not treating to an HbA1ctarget of


CVD mortality, non-fatal CVD events, and loss ofkidney function or ESRD) were similar to thosetreated more intensively. The achieved HbA1c valuesamong the conventional treatment groups in thesestudies were 7.3-8.4%.

Years of intensive glycemic control (HbA1c � 7%)are required before a reduction in the incidence ofcomplications, such as kidney failure or blindness,becomes evident.9,10,15,16 Therefore, when institutingintensive therapy for hyperglycemia in patients withlimited life expectancy, the potential benefits must bebalanced against risks. With intensified insulin treat-ment, there is an increased risk of hypoglycemia andweight gain. In individuals 70-79 years of age who aretaking insulin, the probability of falls begins to in-crease with HbA1c �7%.

31 Moreover, in patients withtype 2 diabetes, one study showed that the presence ofco-morbidities abrogates benefits of lower HbA1clevels on CVD events.32 Therefore, a target HbA1c of�7.0% is suggested for patients with diabetes who areat risk of hypoglycemia and have clinically-signifi-cant co-morbidities or limited life expectancy.

LIMITATIONS

Recommendations regarding glycemic control inpatients with diabetes and CKD are based primarilyon reductions in the appearance and progression ofalbuminuria, yet the relationship between elevatedalbuminuria and clinical endpoints is often discor-dant. Less is known about appropriate glycemic con-trol in patients with diabetes and more advancedCKD, because no prospective, randomized clinicaltrials evaluating the level of glycemic control onhealth outcomes have been carried out in patients withCKD stages 3-5. Extended follow-up of patients withtype 1 diabetes in DCCT/EDIC showed a beneficialeffect of prior intensive therapy on later CKD end-points, but the numbers of patients were small. Arecent observational, claims-based study in peoplewith type 1 or type 2 diabetes and CKD33 reported aU-shaped relationship between HbA1c level and riskof death, with deaths increasing significantly for HbA1clevels below 6.5% and above 8% over nearly 4 yearsof follow-up. Risks of doubling of SCr, ESRD, CVDevents, and hospitalization increased in a graded man-ner with higher levels of HbA1c.

HbA1c levels of �7-9% are associated with betteroutcomes for survival, hospitalization, and CVD inpatients on hemodialysis in some34-38 but not allobservational studies;39,40 however, this relationshiphas not been tested in prospective, randomized stud-ies. Nevertheless, patients with diabetes who are treatedby dialysis or kidney transplant may continue tobenefit from good glycemic control because of reduc-

tions in eye and neurologic outcomes. Complicating

864

glycemic management in patients with diabetes andadvanced CKD, however, are the many new medi-cines now available for glycemic control; some whichare potentially useful and others which are harmful ormust be used with care due to reduced clearance of thedrug or its metabolites by the kidneys.

IMPLEMENTATION ISSUES

Management of hyperglycemia involves a multifac-torial approach that includes medicines, proper nutri-tion and meal planning, and physical activity. Each ofthese approaches may need to be modified in thesetting of CKD. Nutritional management in diabetesand CKD is addressed in Guideline 5 and physicalactivity is addressed in Clinical Practice Recommen-dation 4 of the previously published guideline.4

Special Considerations inAdvancedCKD

The risk of hypoglycemia is increased in patientswith substantial decreases in eGFR (CKD stages 4and 5) for two reasons: (1) decreased clearance ofinsulin and of some of the oral agents used to treatdiabetes and (2) impaired renal gluconeogenesis withreduced kidney mass.41 The contribution of reducedrenal function to the risk of hypoglycemia is difficultto quantify. About one-third of insulin degradation iscarried out by the kidneys and impairment of kidneyfunction is associated with a prolonged half-life ofinsulin. Patients with type 1 diabetes receiving insulinwho have significant creatinine elevations (mean 2.2mg/dL) have a 5-fold increase in the frequency ofsevere hypoglycemia.42,43 Therefore, it is imperativethat patients being treated intensively monitor theirglucose levels closely and reduce their doses of medi-cine as needed to avoid hypoglycemia.

Progressive falls in kidney function result in de-creased clearances of the sulfonylureas or their activemetabolites,44-46 necessitating a decrease in drug dos-ing to avoid hypoglycemia. Table 4 provides recom-mendations for drug dosing of medicines used to treathyperglycemia in patients with CKD. First generationsulfonylureas (e.g., chlorpropamide, tolazamide, andtolbutamide) should be avoided altogether in patientswith CKD. These agents rely on the kidneys to elimi-nate both the parent drug and its active metabolites,resulting in increased half-lives and the risk of hypo-glycemia. Of the second-generation sulfonylureas (e.g.,glipizide, glyburide, and glimepiride), glipizide is thepreferred agent as it does not have active metabolitesand does not increase the risk of hypoglycemia inpatients with CKD. An increase in the levels of theactive metabolite of nateglinide occurs with decreasedkidney function,47,48 but this increase does notoccur with the similar drug, repaglinide.49 On the

other hand, repaglinide can accumulate when the

Am J Kidney Dis. 2012;60(5):850-886


GFR �30 mL/min/1.73 m2.49 Although hypoglyce-

Table 4. Dose Adjustment for Insulin Com

Medication Class and Agents

InsulinGlargine No advised dose adjustmDetemir No advised dose adjustmNeutral Protamine Hagedorn (NPH) No advised dose adjustmRegular No advised dose adjustmAspart No advised dose adjustmLispro No advised dose adjustmGlulisine No advised dose adjustm

First-generation sulfonylureasAcetohexamide** Avoid useChlorpropamide GFR 50-80 mL/min/1.73Tolazamide Avoid useTolbutamide Avoid use

Second-generation sulfonylureasGlipizide No dose adjustmentGlimepiride Start conservatively at 1Glyburide Avoid useGliclazide** No dose adjustment

MeglitinidesRepaglinide If GFR �30 mL/min/1.73Nateglinide If GFR �30 mL/min/1.73

BiguanidesMetformin*** United States FDA label

British National FormulareGFR �30 mL/min/1.73

ThiazolidinedionesPioglitazone No dose adjustmentRosiglitazone No dose adjustment

Alpha-glucosidase inhibitorsAcarbose Avoid if GFR �30 mL/miMiglitol Avoid if GFR �25 mL/mi

DPP-4 inhibitorSitagliptin GFR �50 mL/min/1.73 m

GFR 30-50 mL/min/1.73GFR �30 mL/min/1.73 m

Saxagliptin GFR �50 mL/min/1.73 mGFR �50 mL/min/1.73 m

Linagliptin No dose adjustmentVildagliptin** GFR �50 mL/min/1.73 m

GFR �50 mL/min/1.73 m

Incretin mimeticExenatide Not recommended in GFLiraglutide Not recommended in GF

Amylin analogPramlintide No dose adjustment and

Dopamine receptor agonistBromocriptine mesylate* Not studied in patients w

*Adjust dose based on patient response.**Not currently licensed for use in the U.S.***These levels are controversial (see text).

mia has not been demonstrated to increase substan-

Am J Kidney Dis. 2012;60(5):850-886

tially with progressive falls in GFR,49,50 it would

ds and Oral Medicines for Diabetes in CKD

CKD stages 3, 4, and 5 ND

educe dose 50%, GFR �50 mL/min/1.73 m2: avoid use

aily

tart conservatively at 0.5 mg with mealstart conservatively at 60 mg with meals

s, “do not use if SCr �1.5 mg/dL in men, �1.4 mg/dL in women”the Japanese Society of Nephrology recommend cessation if

3 m2

3 m2

0 mg daily0 mg dailymg dailyg dailymg daily

mg twice dailymg daily

0 mL/min/1.73 m2

0 mL/min/1.73 m2

ecommended for patients with CKD stage 4 or greater

duced GFR

poun

ent*ent*ent*ent*ent*ent*ent*

m2: r

mg d

m2 sm2 s

statey andm2

n/1.7n/1.7

2: 10m2: 52: 252: 5 m2: 2.5

2: 502: 50

R �3R �6

not r

ith re

seem prudent to start treatment with a 0.5 mg dose

865


of repaglinide with each meal and titrate upwardscautiously when the GFR is �30 mL/min/1.73 m2.Similarly, nateglinide should be used with cautionwhen the GFR is �30 mL/min/1.73 m2, startingwith 60 mg at meals and cautiously titrating up-wards.

Metformin does not cause hypoglycemia. Lacticacidosis, however, is a rare and serious side effect ofmetformin use, which can occur when toxic levels ofmetformin accumulate. Metformin is cleared by thekidneys, thus its use in CKD is restricted. A UnitedStates Food and Drug Administration (FDA) man-dated black-box warning exists regarding the risk oflactic acidosis with metformin use. The label indicatesthat metformin should not be used in men with a SCrof �1.5 mg/dL or in women with a SCr of �1.4mg/dL. It is also reasonable to consider a GFR cutofffor metformin use as well, since SCr can translate intodifferent eGFR levels depending on weight, race orage. The clearance of metformin decreases by about75% when the GFR is �60 mL/min/1.73 m2 withoutfurther change when the GFR declines to 30 mL/min/1.73 m2;51 however, serum concentrations of met-formin at both of these lower GFR levels are onlyabout two-fold higher than in normal kidney functionand these levels are still only about 3% of those foundin patients with true metformin-associated lactic aci-dosis.51,52 In studies of patients continuing to receivemetformin with GFR levels in the 30-60 mL/min/1.73m2 range, lactic acidosis is still exceedingly rare evenin the presence of comorbid conditions like conges-tive heart failure, chronic obstructive pulmonary dis-ease, and liver disease.53,54 Given its marked clinicalbenefit, restriction of metformin use based on thecreatinine cutoffs provided by the FDA, or a GFRcutoff of �60 mL/min/1.73 m2, has been called intoquestion.55,56 At present the exact GFR cutoff formetformin use to avoid lactic acidosis is controver-sial. A recent review proposed that metformin use bereevaluated when GFR is �45 mL/min/1.73 m2 andstopped when �30 mL/min/1.73 m2; this advice wasadopted by the British National Formulary and theJapanese Society of Nephrology.57,58,58a

The thiazolidinediones, pioglitazone and rosiglita-zone, do not lead to hypoglycemia, are metabolizedby the liver, and thus can be used in CKD. However,fluid retention is a major limiting side effect and theyshould not be used in advanced heart failure andCKD. They have been linked with increased fracturerates and bone loss;59 thus the appropriate use inpatients with underlying bone disease (such as renalosteodystrophy) needs to be considered. The FDA hasrestricted use of rosiglitazone based on informationlinking the medicine with increased cardiovascular

events.60 Currently, rosiglitazone has to be dispensed

866

by the manufacturer and may no longer be prescribed,except by physicians registered to do so.

Acarbose, a disaccharidase inhibitor, is only mini-mally absorbed, but with reduced kidney function,serum levels of the drug and its metabolites increasesignificantly. Although no adverse effects have beenreported, its use in patients with a GFR�26 mL/min/1.73 m2 is not recommended.61 Miglitol has greatersystemic absorption and undergoes kidney excretion,and it should not be used in patients with GFR �25mL/min/1.73 m2.61

The dipeptidyl peptidase (DPP-4) inhibitors, sita-gliptin, saxagliptin, linagliptin, and vildagliptin de-crease the breakdown of the incretin hormones, suchas glucagon-like peptide 1 (GLP-1), and improve bothfasting and post-prandial glucose levels. All can beused in CKD patients but sitagliptin, saxagliptin, andvildagliptin need downward dose adjustments as de-tailed in Table 4.

Exenatide and liraglutide are injectable incretinmimetics that facilitate insulin secretion, decreaseglucagon secretion, delay gastric emptying and causeearly satiety. Although their use is associated withpancreatitis in some patients, the overall frequency ofpancreatitis with their use is not greater than inpatients with diabetes using other agents. Exenatide isexcreted by the kidneys, and its clearance is reducedby 36% with a GFR of 45 mL/min/1.73 m2 and by64% with a GFR of �30 mL/min/1.73 m2.62 There-fore, exenatide is not recommended for use with aGFR �30 mL/min/1.73 m2. Furthermore, exenatidehas been associated with acute kidney injury or accel-eration of CKD progression in case reports.63,64 Lira-glutide is fully degraded elsewhere in the body, andthe kidneys are not a major organ of elimination.65 Insingle dosing, there is no effect on the area under thecurve in subjects with stages 4 and 5 CKD.65 How-ever, there are few data on long term use and themanufacturer recommends avoiding this medicinewhen GFR is �60 mL/min/1.73 m2 .57

Pramlintide is an injectable amylin analog availableas a complement to insulin therapy and normally it isgiven with each meal. Although pramlintide is metabo-lized and eliminated predominantly by the kidneys, ithas a wide therapeutic index and dosage adjustmentsare not usually required in the presence of mild-to-moderate decreases in GFR. However, use of pramlint-ide is not recommended for patients with CKD stage 4or greater.

Bromocriptine mesylate is a dopamine agonist thatis predominantly metabolized in the liver and only2-6% appears in the urine. No studies evaluating thesafety of this medicine in patients with reduced GFRhave been performed; therefore it should be used with

caution in patients with CKD.

Am J Kidney Dis. 2012;60(5):850-886


Assessment ofGlycemic Control

Inaccuracy of the HbA1c measurement in reflectingambient glucose concentrations must be considered inthe assessment of glycemic control in patients withprogressive kidney disease. Factors that may contrib-ute to falsely decreased values include a reduced redblood cell lifespan, transfusions, and hemolysis. Onthe other hand, falsely increased values may occurdue to carbamylation of the hemoglobin and acidosis.However, Morgan et al found that the relationshipbetween HbA1c and glucose levels was not differentbetween patients with normal kidney function andthose with kidney failure (creatinine mean of 6.6mg/dL), but some hemodialysis patients had lowerthan expected HbA1c levels relative to the ambientglucose concentrations.66 Opposite findings for dialy-sis patients were reported by Joy et al;67 an HbA1cincrease of 1% correlated with a change in meanglucose of 20 mg/dL in hemodialysis patients and 30mg/dL in those with normal kidney function. Studiespublished since the release of the previous KDOQIdiabetes guidelines contributed further to our under-standing of the relationship between HbA1c and glu-cose in advanced CKD. Inaba et al68 found lowercorrelation of plasma glucose levels with HbA1c lev-els in patients with diabetes on hemodialysis (r �0.520) compared to those with normal kidney func-tion (r � 0.630), and they also had shallower regres-sion slopes. Riveline et al69 also found a shallowerregression slope for hemodialysis patients comparedto those without DKD. At lower levels of glucose(�160 mg/dL and HbA1c �7.5%), hemodialysis pa-

tients tended to have higher glucose levels for a given

Am J Kidney Dis. 2012;60(5):850-886

HbA1c, whereas at higher levels the correlations weresimilar. When patients with CKD stages 3 and 4 wereevaluated, glucose levels were also found to be slightlyhigher than expected for given HbA1c levels.

70 Ironsupplementation or erythropoietin administration leadto a modest fall of 0.5-0.7% in HbA1c along with therise in total hemoglobin in patients with advancedCKD. These effects are likely due to the formationof new red cells and to alterations in hemoglobinglycation rates.68,71 Importantly, all of these studiesshow a very wide variability in the glucose-HbA1crelationship.66-71 The modest changes with decreas-ing eGFR from 75 to15 mL/min/1.73 m2, and evenwith hemodialysis, do not appear to be of clinicalsignificance compared to the wide inter-individualvariability. Neither peritoneal nor hemodialysisacutely change HbA1c levels.

72 Fructosamine orglycated albumin correlate either more poorly66,67,69

or better68,70 with blood glucose than HbA1c inpatients with stages 4 and 5 CKD. Nevertheless, arecent prospective study found that glycated albu-min, which reflects glycemic control over a 2-weekperiod, is a better predictor of mortality and hospi-talizations than HbA1c in dialysis patients withdiabetes.35 In summary, HbA1c remains the bestclinical marker of long-term glycemic control, par-ticularly if combined with self-monitoring of bloodglucose, in patients with diabetes and CKD. Othermarkers such as glycated albumin that reflect glyce-mic control over a shorter period may be of greatervalue for predicting clinical outcomes in patients

with advanced CKD.

867


Guideline 4: Management of Dyslipidemia in Diabetes and CKD

Dyslipidemia is common in people with diabetesand CKD. Cardiovascular events are a frequentcause of morbidity and mortality in this popula-tion. Lowering low-density lipoprotein cholesterol(LDL-C) with statin-based therapies reduces riskof major atherosclerotic events, but not all-causemortality, in patients with CKD including thosewith diabetes.

4.1: We recommend using LDL-C loweringmedicines, such as statins or statin/ezetimibecombination, to reduce risk of major athero-sclerotic events in patients with diabetesand CKD, including those who have re-ceived a kidney transplant. (1B)

The evidence that lowering the LDL-C concentra-tion reduces the risk of major atherosclerotic eventsin patients with diabetes and CKD (other than stage5) was presented in detail in the original KDOQIdiabetes guideline.4 Recommendations were basedlargely on four post hoc analyses73-76 that reportedresults of lipid lowering therapy for a subpopula-tion of patients with CKD and diabetes comparedwith placebo (Table 5).

A new clinical trial has added to the evidence thatlowering LDL-C reduces cardiovascular events in awide range of patients with diabetes and CKD. TheStudy of Heart and Renal Protection (SHARP) trial77

randomized 9438 participants �40 years old withCKD (mean eGFR of 27 mL/min/1.73 m2) to receivesimvastatin 20 mg plus ezetimibe 10 mg daily orplacebo, and followed them for 5 years. Thirty-threepercent of the patients (n�3023) were receiving main-tenance dialysis at randomization and 23% (n�2094)of the participants had diabetes, with equal propor-tions in the simvastatin plus ezetimibe and placebogroups. Statin plus ezetimibe therapy was associatedwith a significant 17% relative reduction in the risk ofthe primary outcome of major atherosclerotic events(coronary death, myocardial infarction [MI], non-hemorrhagic stroke, or any revascularization) com-pared with placebo (hazard ratio [HR], 0.83; 95%confidence interval [CI], 0.74-0.94). This finding wasattributable in large part to significant reductions innon-hemorrhagic stroke and arterial revascularizationprocedures. There was no reduction in the risk ofall-cause mortality, and among the patients with CKDnot treated by dialysis at randomization (n�6247),treatment with simvastatin plus ezetimibe did notreduce the frequency of doubling of the baseline SCr

concentration or progression to ESRD. Although the

868

study was not powered to reliably estimate the effectof treatment on primary outcomes among clinicalsubgroups, the proportional effect on major atheroscle-rotic events did not appear to differ between thosewith or without diabetes.

The Assessment of Lescol in Renal Transplant(ALERT) trial78 examined the effect of statin therapyon cardiovascular risk reduction in 2102 patients withfunctioning kidney transplants who were followed for5-6 years. Fluvastatin therapy (40-80 mg/day), com-pared with placebo, was associated with a significant35% relative reduction in the risk of cardiac death ordefinite nonfatal MI (HR, 0.65; 95% CI, 0.48-0.88).The study included a pre-specified analysis for asubset of 396 patients with diabetes, of whom 197were randomized to fluvastatin and 199 to placebo. Inthis subset, the benefit was similar in magnitude as inthe overall cohort, but was not statistically significant(HR, 0.71; 95% CI, 0.41-1.21), suggesting limitationsof under-powering due to small sample size. Giventhese limitations and the lack of a significant interac-tion between diabetes and treatment assignment forthe primary outcome, the Work Group based its recom-mendation for statin treatment in kidney transplantpatients on the overall results from the ALERT study.

Accordingly, the evidence that treatment with statinor statin/ezetimibe combination improves health out-comes is based primarily on prevention of CVDevents. There is no evidence from these trials thatsuch treatment improves kidney disease outcomes,including doubling of SCr or progression to ESRD, orall-cause mortality.

4.2: We recommend not initiating statin therapyin patients with diabetes who are treated bydialysis. (1B)

Results of the Die Deutsche Diabetes Dialyse Studie(4D)79 motivated the recommendation regarding sta-tin treatment in patients with type 2 diabetes onmaintenance hemodialysis in the original KDOQIdiabetes guideline.4 Concerns that the results of 4Dwere attributable to the futility of a single interventionin such high-risk patients inspired A Study to Evaluatethe Use of Rosuvastatin in Subjects on Regular Hemo-dialysis (AURORA)80, a clinical trial that randomized2776 patients on hemodialysis to rosuvastatin 10 mg aday and placebo. Only 26% of the patients inAURORA had diabetes. As found in 4D, AURORAreported no significant effect of statin therapy on theprimary cardiovascular outcome that included car-

diac death or non-fatal MI and fatal or non-fatal

Am J Kidney Dis. 2012;60(5):850-886

Table 5. Summary of Four Post Hoc Analyses Reports of Lipid Lowering in People with Diabetes Mellitus (DM) and CKD

Study# treated/# with

DM and CKDRandomized

statin CVD outcome vs. placebo Definition of kidney impairment Kidney outcome vs. placebo

OSCOPS, CARE,LIPID – Tonelli76

290/571 Pravastatin,40 mg/day

All cause mortality 18.0% onpravastatin vs. 19.2%.(Absolute reductiondecreased from 6.4 to3.5% comparing peoplewith DM and CKD tothose with neither). HR forCABG or PTCA 0.69,(95% CI 0.47-1.01). HRfor stroke, 1.12 (95% CI0.63-1.97).

GFR �60 mL/min/1.73 m2 or GFR 60-89mL/min/1.73 m2 concomitant with traceor greater proteinuria on dipstickurinalysis

Not reported

4S – Chonchol73 105/200 Simvastatin,20 mg/day

All cause mortality 13.5% onsimvastatin vs. 27.9%

GFR �75 mL/min/1.73 m2 Not reported

CARDS – Colhoun74 482/970 Atorvastin,10 mg/day

All cause mortality 5.6% onatorvastatin vs. 6.2%.Stroke 1.2% onatorvastatin vs. 3.1% onplacebo. Coronaryrevascularization 1.0% onatorvastatin vs. 2.5% onplacebo.

GFR �60 mL/min/1.73 m2 20.5% regression from micro- tonormoalbuminuria vs. 19.4%.

HPS – Collins75 142/310 Simvastatin,40 mg/day

All cause mortality notreported.

Creatinine �110 �mol/L (1.24 mg/dL) forwomen, and �130 �mol/L (1.47 mg/dL)for men

Significantly smaller fall in the GFRduring follow-up (5.9 [0.1] vs. 6.7[0.1] mL/min, difference �0.8[0.2] mL/min; p�0·0003). Thisdifference appeared to beslightly larger among those whohad diabetes than among thosewho did not (�1.4 [0.4] mL/minvs. �0.5 [0.2] mL/min;heterogeneity p�0.08).

Abbreviations: 4S, Scandinavian Simvastatin Survival Study; CABG, coronary artery bypass graft surgery; CARDS, Collaborative Atorvastatin Diabetes Study; CARE, Cholesterol andRecurrent Events; CI, confidence interval; CKD, chronic kidney disease; GFR, glomerular filtration rate; HPS, Heart Protection Study; HR, hazard ratio; LIPID, Long-Term Intervention withPravastatin in Ischaemic Disease; PTCA, percutaneous coronary angiography WOSCOPS, West of Scotland Coronary Prevention Study.

Am

JK

idneyD

is.2012;60(5):850-886869

KD

OQ

IDiabetes

Guideline:2012

Update

prote


stroke, either in the overall study population or inthe subgroup of patients with diabetes. A subsequentpost hoc analysis of the participants with diabetes inAURORA81 found that treatment with rosuvastatinsignificantly reduced the risk of a redefined end pointof cardiac death or non-fatal MI by 32%, but signifi-cantly increased the risk of hemorrhagic stroke bymore than 5-fold. Although the number of hemor-rhagic strokes among participants with diabetes inAURORA was small and the overall stroke rate didnot differ by treatment group, the finding is concern-ing, since the original report from the 4D trial foundthat treatment with atorvastatin was associated with2-fold increase in fatal stroke.79 A recent post hocanalysis of the 4D trial82 found that fatal and nonfatalcardiac events were significantly reduced if the pre-treatment LDL-C was �145 mg/dL. Although thesepost hoc analyses provide a different look at the datafrom the previous studies, they must be viewed ashypothesis-generating, and therefore do not alter themain message of the guideline update, which is basedon the primary pre-specified outcomes from theseclinical trials.

The SHARP trial77 indicated that risk for the pri-mary outcome of major atherosclerotic events otherthan death was reduced by simvastatin/ezitimibe com-bination among a wide range of patients with CKD.Yet, the “subgroup” of over 3000 patients on dialysisdid not show a statistically significant reduction inrisk of the primary outcome. The SHARP investiga-tors advocate that this group is still likely to benefitbecause of lack of statistical heterogeneity. However,even as a subgroup, this is still the largest trial of LDLcholesterol-lowering conducted to-date in patients ondialysis. Taking into account the 4D and AURORAtrials along with the SHARP data, overall evidence tosupport a favorable effect of initiating LDL-choles-terol lowering treatment on atherosclerotic events in

Table 6. Fibrate Treatment in Patie

Study# treated/# with

DM and CKD Randomized fibrate

VA-HIT – Tonelli86 136/297 Gemfibrozil,600 mg BID

Comth41

FIELD – Keech85,87 4895/9795 Fenofibrate,200 mg/day(mean dose)

Not

DAIS – Ansquer84 155/314 Fenofibrate,200 mg/day

Not

Abbreviations: BID, twice a day; DAIS, Diabetes AtherosclerLowering in Diabetes; VA-HIT, Veterans Affairs High-density lipo

dialysis patients is lacking. Moreover, since most of

870

the clinical CVD events experienced by hemodialysispatients with diabetes are deaths, for which statinsprovide little or no benefit as illustrated in the SHARPtrial,77 the Work Group concluded that the availableevidence continues to support the recommendationthat statin therapy not be initiated in dialysis patientswith diabetes. Whether previously treated patientsshould be continued on statin therapy once they com-mence dialysis, or not, has not been studied, and assuch, data are insufficient to provide guidance for thisgroup.

LIMITATIONS

With the exception of SHARP, data to supportrecommendations for LDL cholesterol-lowering comefrom post hoc subgroup analyses of clinical trials thatincluded CKD patients with and without diabetes.Nevertheless, a growing body of evidence demon-strates a clear benefit of statin therapy on clinicalCVD events among patients with diabetes. This ben-efit holds true across a wide range of CKD stages,perhaps with the exception of those on dialysis. Ofnote, the Adolescent type 1 diabetes mellitus cardio-renal Intervention Trial (AdDIT)83 is under way, andwill provide data on the effectiveness of atorvastatinand quinapril to prevent cardiovascular and kidneycomplications in adolescents with type 1 diabetes.

Other lipid altering medicines may also be of valuein the management of patients with diabetes andCKD, but the Work Group concluded that the avail-able evidence for these medicines was not yet suffi-cient to make specific management recommendations.Randomized treatment trials that examined the effectof fibrates relative to placebo in patients with diabetesand CKD are summarized in Table 6.84-86 The Veter-ans Affairs High-density lipoprotein Intervention Trial(VA-HIT)86 found evidence that gemfibrozil reducesrisk of major cardiovascular events (fatal coronary

ith Diabetes Mellitus (DM) and CKD

CVD OutcomeRegression from micro- to

normoalbuminuria

te outcome 26.5% (36/136) infibrozil treated group vs.

(66/161) in the placebo

Not reported

ted 47.0% (462/983) in thefenofibrate group vs. 39.3%(400/1017) in the placebogroup

ted 37.7% (20/53) in the fenofibrategroup vs. 34.1% (15/44) inthe placebo group

Intervention Study; FIELD, Fenofibrate Intervention and Eventin Intervention Trial.

nts w

posie gem.0%

repor

repor

osis

heart disease, nonfatal MI, and stroke) by 42% com-

Am J Kidney Dis. 2012;60(5):850-886


pared with placebo (RR, 0.58; 95% CI, 0.38-0.89) in apost hoc analysis of 297 individuals with low eGFR(GFR �75 mL/min/1.73 m2) and diabetes. The Diabe-tes Atherosclerosis Intervention Study (DAIS)84 andthe Fenofibrate Intervention and Event Lowering inDiabetes (FIELD) study85,87 reported that fenofibratetreatment significantly lowered the risk of developingnew onset microalbuminuria compared with placebo(RR, 0.87 in patients with type 2 diabetes; 95% CI,0.77-0.97). Fenofibrate also promoted regression frommicroalbuminuria to normoalbuminuria (RR, 1.15;95% CI, 1.04-1.28, n�2260, 2 trials), but did notchange the risk of progression from microalbuminuriato macroalbuminuria (RR, 1.07; 95% CI, 0.43-2.63, 1trial, n�97). There is moderate evidence that fenofi-brate decreases risk of progression from normoalbu-minuria to microalbuminuria and leads to regressionof microalbuminuria to normoalbuminuria comparedwith placebo in patients with type 2 diabetes. None ofthe trials of fibrate therapy in diabetes published sincethe original guideline reported CVD or kidney diseaseoutcomes for the subgroup of patients with CKD.


Management of dyslipidemia involves a multifacto-rial approach that includes medicines, proper nutri-tion, and physical activity. Each of these approachesmay need to be modified in the setting of CKD.Nutritional management in diabetes and CKD is ad-

Table 7. Dose Adjustment fo

Medication Class and Agents No CKD or stages 1-2

Statins (mg/day)Atorvastatin 10-80Fluvastatin 20-80Lovastatin 10-80Pravastatin 10-40Rosuvastatin 5-40Simvastatin 5-40

Bile acid sequestrants (g/day)Cholestipol 5-30Cholestyramine 4-16Colesevelam 2.6-3.8

Fibric acid derivatives (mg/day)Bezafibrate* 400-600Clofibrate 1000-2000Ciprofibrate* 200Fenofibrate 96Gemfibrozil 1200

Other (mg/day)Ezetimibe 10Niacin 2000

*Not currently licensed for use in the U.S.

dressed in Guideline 5 and physical activity is ad-

Am J Kidney Dis. 2012;60(5):850-886

dressed in Clinical Practice Recommendation 4 of thepreviously published guideline.4

Higher doses of statins may be beneficial in somepatients with diabetes and mild-to-moderate CKD(stages 1-3). The Treating to New Targets trial(TNT)88 reported a benefit for secondary preven-tion of major cardiovascular events from treatmentwith atorvastatin, 80 mg/day compared with atorva-statin, 10 mg/day, in 546 patients with diabetes andCKD and pre-existing coronary artery disease over5 years of follow-up. The risk of stroke was 4.8%(13/273) for the higher dose, compared with 7.3%(20/271) for the lower dose. There was no reductionin all-cause mortality.

Higher doses of lipid lowering medicines, however,are associated with increased risk of myopathy,89

particularly among patients with reduced kidney func-tion.90 Therefore, doses of some lipid-lowering medi-cines should be modified in moderate–to-advancedCKD (stages 3-5). Additionally, reliance less on higherdosing of statins and more on combination therapy toreduce LDL-C is an attractive strategy. The SHARPtrial77 addressed this issue by using lower dose simva-statin (20 mg/day) and adding the cholesterol-absorp-tion inhibitor ezetimibe (10 mg/day) to achieve anaverage LDL-C reduction of about 1 mmol/L. Table 7provides guidance for drug dosing of lipid-loweringmedicines in patients with CKD. Updated recommen-dations on management of dyslipidemia in CKD (in-

Lowering Medicines in CKD

CKD stage 3 CKD stages 4-5 Kidney transplant

10-80 10-80 10-2020-80 10-80 10-8010-80 10-40 10-4010-40 10-20 10-205-20 5-10 55-40 5-20 5-20

5-30 5-30 5-304-16 4-16 4-16

2.6-3.8 2.6-3.8 2.6-3.8

200 Avoid Avoid500 500 Avoid

Unknown Avoid Unknown48 Avoid Avoid

1200 600 600

10 10 Unknown2000 1000 Unknown

r Lipid

cluding diabetes) are expected from KDIGO in 2013.

871


Of note, the U.S. FDA issued a Safety Announce-ment in June 2011 that recommends limited use of thehighest approved dose of simvastatin (80 mg) becauseof increased risk of myopathy. Simvastatin 80 mgshould be used only in patients who have been takingthis dose for 12 months or more without evidence ofmuscle injury. Simvastatin 80 mg should not be startedin new patients, including patients already takinglower doses of the drug. In addition to these newlimitations, the FDA is requiring changes to the simva-statin label to add new contraindications (concurrentcyclosporine or gemfibrozil use) and dose limitations

for use with other medicines such as calcium channel

872

blockers or amiodarone. The lovastatin label has alsobeen updated extensively with new contraindicationsand dose limitations when it is taken with certainmedicines that can increase the risk of myopathy; andhuman immunodeficiency virus and hepatitis C virusprotease inhibitors are now contraindicated with sim-vastatin and lovastatin because of increased risk ofmyopathy. The FDA has also added information tostatin labels about the potential for generally non-serious and reversible cognitive side effects and re-ports of increased HbA1c levels. Further informationcan be obtained at the FDA website (fda.gov/Drugs/

DrugSafety).

Am J Kidney Dis. 2012;60(5):850-886

http://fda.gov/Drugs/DrugSafetyhttp://fda.gov/Drugs/DrugSafety


Guideline 6: Management of Albuminuria in Normotensive

Patients with Diabetes

Treatments that produce a lasting decrease inurinary albumin excretion may slow the progres-sion of DKD even in the absence of hypertension.However, most people with diabetes and albumin-uria have hypertension. Assessment of albumin-uria is addressed in Guideline 1.4 Management ofhypertension is addressed in Guideline 34 and theKDIGO Clinical Practice Guideline for the Man-agement of Blood Pressure in CKD.

6.1: We recommend not using an angiotensin-converting enzyme inhibitor (ACE-I) or anangiotensin receptor blocker (ARB) for theprimary prevention of DKD in normoten-sive normoalbuminuric patients with diabe-tes. (1A)

There is currently strong evidence that use of agentsthat block the renin-angiotensin system (RAS) does notprevent the development of microalbuminuria or slowthe rate of biopsy-assessed diabetic renal lesions innormoalbuminuric normotensive patients with type 1diabetes, at least over a study duration of 4-5 years.91-93

Studies in normoalbuminuric normotensive pa-tients with type 2 diabetes are fewer but also show nobenefit.91,94 For normoalbuminuric patients with type2 diabetes and hypertension, or pre-existing CVD, useof ACE-Is with or without diuretics reduces the abso-lute risk of developing microalbuminuria by 2-4%over 4-5 years.95-97 However, these studies used avariety of definitions of microalbuminuria, were oftenbased on single urine samples taken at 1-2 yearintervals, and were not tested for a durable effect vs. atransient hemodynamic effect of RAS blockade. Morestringent definitions of microalbuminuria using mul-tiple timed collections at more frequent intervalsrevealed incidence rates of 3-4% per annum,91,97 lessthan half the incidence reported by studies usingless stringent definitions of microalbuminuria suchas MICROHOPE96 and ADVANCE.95 Thus, the evi-dence does not support a clinical benefit of interven-tion to prevent the intermediate outcome of persistentmicroalbuminuria or the changes in kidney structureassociated with DKD.

6.2: We suggest using an ACE-I or an ARB innormotensive patients with diabetes andalbuminuria levels >30 mg/g who are athigh risk of DKD or its progression. (2C)

There are no long-term studies that show a benefit

of treatment with RAS blocking agents on CKD

Am J Kidney Dis. 2012;60(5):850-886

progression or health outcomes in normotensive pa-tients with diabetes and increased levels of albumin-uria. Some of these patients, perhaps especially thosewith additional risk factors for DKD, may benefitfrom such treatment, although there is no strongevidence to support this belief. Those more likely todevelop or progress to more serious DKD includepatients with increasing levels of albuminuria in themicroalbuminuria range, macroalbuminuria, declin-ing glomerular filtration rate, increasing blood pres-sure, presence of retinopathy, elevated lipids and/oruric acid concentrations, or a family history of hyper-tension, macrovascular disease, or DKD. Patientswith microalbuminuria and none of these additionalrisk factors may be at relatively low risk of DKD or itsprogression and could be followed without treatmentwith RAS blocking agents to see whether some ofthese additional risk factors subsequently develop.The presence of macroalbuminuria without retinopa-thy, especially if present within 10 years of diabetesonset, suggests a need for investigations to rule outnondiabetic kidney diseases.

In hypertensive patients with type 1 and type 2diabetes, RAS blocking agents prevent developmentof macroalbuminuria, but even after two or moreyears of treatment, albuminuria increases soon afterthe withdrawal of these drugs.98 This observationcalls into question the durability of the treatmenteffect on underlying disease processes.

For patients with macroalbuminuria and moder-ately reduced eGFR, there is strong evidence showingthat angiotensin-converting enzyme (ACE) inhibitionin type 199 and ARBs in type 2100,101 diabetes conferbenefit in terms of loss of GFR (both rates of declinein eGFR and doubling of baseline SCr) and ESRD,but few of these patients are normotensive. Further-more, there are no long-term studies that examine therenoprotective efficacy of ACE-Is in type 2 or ARBsin type 1 diabetes.

Fundamental to the interpretation of all of these stud-ies is the definition of hypertension, which is currentlyunder review by KDIGO and the Joint National Commit-tee on Prevention, Detection, Evaluation, and Treatmentof High Blood Pressure (JNC). Early studies in “normo-tensive” patients with type 1 diabetes used definitions ofthe upper limit of normal for blood pressure that wouldnow be considered hypertension. The meta-analysis pub-lished by the ACE Inhibitors in Diabetic NephropathyTrialist Group102 included studies in patients with type 1diabetes whose upper limit of baseline blood pressures at

inclusion ranged from 145-160 mm Hg systolic and

873


90-120 mm Hg diastolic. These investigators also didnot report correction for systolic pressure, only for dia-stolic. Moreover, preferred statistical approaches to con-trolling for the effects of blood pressure are debatable,and as such, are not established methodologies.

LIMITATIONS

Past recommendations regarding management ofalbuminuria in patients with diabetes assumed that theappearance of microalbuminuria signaled an inevi-table progression to macroalbuminuria. There is in-creasing evidence, however, of spontaneous remis-sion of microalbuminuria in up to 40% of patientswith type 1 diabetes,103,104 while about 30-40% re-main microalbuminuric and do not progress to mac-roalbuminuria over 5-10 years of follow-up.105

Whether this observation reflects a better understand-ing of the natural course of albuminuria in diabetes or is,in part, a response to treatment is uncertain, but theadoption of a blanket policy for the use of RAS blockingdrugs in these patients has made it difficult to exploreother potential therapies and hindered studies of thenatural history of early diabetic nephropathy.

The use of albuminuria as a surrogate marker ofbenefit of intervention in DKD was the subject of anFDA/NKF symposium in 2008.106 Major questions wereraised about how to define abnormal albuminuria; atwhat level should intervention take place; how manytests over what period of time would be required beforeintervention should commence; what would be regardedas an adequate response to intervention and how wouldthis be defined; and how would long term benefit bemeasured? The conference produced a consensus reportthat concluded that the evidence was not strong enoughto use changes in albuminuria as an adequate surrogateendpoint of long term kidney benefit in people withdiabetes or other kidney disease.

A major confounding problem with interpretingstudies of intervention in the course of DKD is chang-ing natural history. In the 1970s, the median time toESRD from the development of overt (dipstick posi-tive) proteinuria in type 1 diabetes was �7 years;107 itis now �14 years.105 The incidence of ESRD in type1 diabetes from Finland is now only 7.8% at 30 yearsduration.108 Recent data from the DCCT/EDIC cohortshows a 10 year incidence of 3% for ESRD in 325patients with incident microalbuminuria during thecourse of the study.105 Thus, intervention studies withbenefit in terms of health outcomes related to kidneydisease as their primary end point could require manyyears of observation, and be costly in terms of re-source. For these reasons, such studies are unlikely tobe performed. In clinical practice, changes in eGFRand albuminuria are suggested to be used together to

monitor kidney status, even in the absence of conclu-

874

sive evidence that they predict precisely long-termreduction in risk of actual health outcomes such asESRD.


Dosages forACE-Is orARBs

In normotensive persons with diabetes and albumin-uria the target dose of ACE-Is or ARBs is unknown. Inthe absence of side effects or adverse events (e.g.,hyperkalemia or acute kidney injury) the Work Groupsuggests titration up to the maximum approved dosefor the treatment of hypertension.

CautionsAboutUsageofACE-Is andARBs

The use of a combination of ACE-Is and ARBs as adual blockade of the RAS cannot be recommended atpresent. At least 1 clinical trial has shown an increase inadverse events, particularly impaired kidney functionand hyperkalemia, compared to either agent alone, de-spite a reduction in albuminuria using combinationtherapy.109,110

The use of ACE-Is and ARBs in early pregnancy isreportedly associated with harm to the fetus (neonatalacute kidney injury; lung toxicity; skull hypoplasia;congenital malformations of the cardiovascular sys-tem, central nervous system, and kidney),111 althoughmore recent studies have not confirmed theserisks.112,113 The FDA is currently reviewing its adviceon the use of these agents in the first trimester.

NewerAgents that Target theRAS

One 6-month phase 2 study of Aliskiren in theEvaluation of Proteinuria in Diabetes (AVOID) re-ported a further reduction of albuminuria with theaddition of the direct renin inhibitor aliskiren to losar-tan in patients with type 2 diabetes and microalbumin-uria.114 A phase 3 study, Aliskiren Trial in Type 2Diabetes Using Cardio-Renal Endpoints (ALTITUDE),was subsequently initiated in 3 groups of patients withtype 2 diabetes: 1) albuminuria �200 mg/g; 2) eGFR30-60 mL/min/1.73 m2 with albuminuria �20 mg/gand �200 mg/g; 3) eGFR 30-60 mL/min/1.73 m2

with CVD. However, ALTITUDE was stopped earlydue to therapeutic futility and increased risk of strokeand adverse events including hyperkalemia, hypoten-sion, and ESRD or death due to kidney disease. Giventhese findings, dual blockade of the RAS with directrenin inhibition and either ACE-I or ARB cannot berecommended. Indeed, the manufacturer (Novartis)recommends that aliskiren be stopped in diabeticpatients treated with ACE-I or ARB, and in April 2012the U.S. FDA announced a new contraindicationagainst the use of aliskiren with ACE-Is or ARBs inpatients with diabetes because of the risk of kidney

impairment, hypotension, and hyperkalemia.

Am J Kidney Dis. 2012;60(5):850-886


Research Recommendations

These guideline updates, and the clinical trials onwhich they were based, illustrate the importance ofcontinuing to conduct research that challenges orexpands established clinical practice. As stated in theoriginal guideline, uncertainty is an immutable ele-ment of all scientific research, and the establishmentof a guideline should neither preclude nor renderunethical further inquiry.4 Even as knowledge regard-ing approaches to managing diabetes to prevent ortreat DKD and related complications has advancedsubstantially since the publication of the originalKDOQI Clinical Practice Guidelines and ClinicalPractice Recommendations for Diabetes and ChronicKidney Disease, many essential questions remain tobe answered. The original guideline publication pro-vided a list of research suggestions that meaningfullyinformed the investigative agenda, one of many indi-cations of the importance of these efforts. The presentresearch recommendations arose directly from theguideline update process. The Work Group providedresearch suggestions they considered key to advanc-ing knowledge concerning clinical care. These sugges-tions are arranged by topic area in order to link back tothe specific guidelines. Some of the recommendationsare not for new research studies per se, but areproposals for how research studies might be designedto enhance their value, validity, or generalizability.

Guideline 2: Management of Hyperglycemia andGeneral Diabetes Care in CKD

1. Determine effects of glycemic control on earlyand late GFR loss and health outcomes of CKD.Evaluate different levels of glycemic control tooptimize safety as well as clinical outcomes ofsurvival, hospitalization, and CVD events inadvanced CKD and/or ESRD.

2. Perform validation studies of HbA1c, glycatedalbumin, and potentially other markers of long-term glycemic control in patients with diabetesand various stages of CKD.

3. Assess metformin safety in patients with CKDstages 4 and 5.

Guideline 4: Management of Dyslipidemia in Diabetesand CKD

1. Perform clinical trials of statins for primaryand secondary prevention of CVD in patientswith diabetes and CKD stages 1-4 and meta-analyses of completed studies in CKD stage 5.

2. Conduct studies of other lipid-lowering agents

for primary and secondary prevention of CVD,

Am J Kidney Dis. 2012;60(5):850-886

or patient level meta-analyses of completedstudies, in patients with diabetes and CKDstages 1-4.

3. Establish LDL-C levels for treatment and initia-tion of therapy as well as targets for primary andsecondary prevention in patients with diabetesby CKD stage.

4. Evaluate lipid-lowering therapy for CVD inpatients with diabetes who are treated by hemo-dialysis or peritoneal dialysis, or who havereceived a kidney transplant.

5. Examine results of previous studies, e.g.,SHARP, by CKD stage.

6. Examine data from completed clinical trials toassess effects of lipid lowering agents on out-comes such as albuminuria, eGFR, and ESRD inparticipants with diabetes.

7. Examine lipid-lowering therapy for CVD inadolescent patients with diabetes.

Guideline 6: Management of Albuminuria inNormotensive Patients with Diabetes

1. Durability of RAS inhibition for the delay inmicroalbuminuria onset should be tested by atreatment washout phase of at least two monthsduration.

2. Post hoc adjustment for blood pressure differ-ences may be fraught with faulty assumptions.Therefore, equivalent blood pressure levels arean important design element to be considered infuture clinical trials that test specificity of adrug’s mechanism of action independent ofblood pressure effects.

3. Since the “endpoint” of preventing incidentalbuminuria derives validity from predictingincreased risk of GFR loss, treatments to reducealbuminuria should not be offset by greater GFRdecline. Measurement of GFR (e.g., eGFR orother more precise methods) should be per-formed as a companion to albuminuria. Inclinical trials to demonstrate prevention of el-evated albuminuria, the demonstration of nor-moalbuminuria at baseline should follow wash-out of at least of two months duration fromprevious RAS blockade, with careful bloodpressure control by alternative antihypertensiveagents. This approach is necessary to avoidrandomization of participants in whom albumin-uria is already present, but masked by RAStreatment, an effect which may be posited in

several studies where there was rapid progres-

875


sion to microalbuminuria in the first few monthsafter randomization to placebo.

4. Given the limitations of albuminuria as anoutcome measure and the recent consensuspanel’s recommendation against acceptance ofalbuminuria as a surrogate outcome,106 studiesare needed to evaluate durability of effects onurinary albumin excretion. The categorization ofalbuminuria outcomes should be based on aminimum of two of three consecutive urinesamples being in the same category.115

5. Consider an indication for regulatory approvalbased on demonstration of a lasting reduction inurinary albumin excretion, but conditional uponfirm commitment to continue long-term studiesto determine effects on GFR loss and clinically

relevant outcomes.

876

6. Evaluate the relative roles of ACE-Is, ARBs,renin blockers, and mineralocorticoid receptorblockers on progression of DKD in patients withalbuminuria.

7. Kidney biopsy outcomes based on carefullymeasured structural variables that correlatestrongly with GFR loss may reduce the durationof primary prevention or early interventionDKD clinical trials.93 Consider enzyme replace-ment for Fabry’s Disease as an example.116

8. Clinical trials represent important opportunitiesto advance knowledge beyond addressing theprimary hypotheses themselves. Protocols shouldinclude plans for acquiring and banking blood,urine, DNA, and other samples for eventualbiomarker discovery and validation. Consider

the DCCT as an example.117

Am J Kidney Dis. 2012;60(5):850-886


Conclusion

Earlier and more aggressive therapeutic interven-tion is believed to be responsible, at least inpart, for the general decline in the incidence of ESRDattributable to diabetes among several racial and eth-nic groups in recent years.2 Encouraged by theseobservations and by the results of previous trials usingless aggressive endpoints, several large, well-de-signed clinical trials were conducted among patientswith diabetes to determine whether even earlier ormore intensive therapy might further reduce the fre-quency of CKD and important health outcomes ofCKD, including ESRD. Results from these trials sug-gest that “more is not always better,” as such interven-tions often did not improve clinical outcomes, and insome settings were actually harmful. After examiningthe new evidence, the Work Group moved each guide-line to a more conservative position than was taken inthe original guideline published in 2007.

Clinical Practice Guideline 2 was modified to rec-ommend that a target HbA1c of �7% is mainly usefulto prevent or delay microvascular complications in-cluding DKD in both type 1 and type 2 diabetes. Arecent series of three large clinical trials found nomi-nal to no benefit of more intensive glycemic control(target HbA1c levels �7%) on macrovascular compli-cations or clinical kidney disease endpoints (loss offunction or requirements for dialysis or transplanta-tion) in older people with established type 2 diabetes.Moreover, the risk of severe hypoglycemia washigh.19-21 Therefore, a target HbA1c �7% is notrecommended for patients with diabetes at risk ofhypoglycemia, a group that includes many in theCKD population. Finally, patients with diabetes andadvanced CKD often have multiple co-morbidities orlimited life expectancy that would nullify the poten-tial benefits of intensive glycemic control. In suchpatients, an extension of target HbA1c to �7% issuggested.

Clinical Practice Guideline 4 was updated to reflectthe results of recent clinical trials of lipid-loweringtherapies that included patients with diabetes and

CKD. In particular, the SHARP trial added new data

Am J Kidney Dis. 2012;60(5):850-886

to support treatment with a statin-ezitimibe combina-tion in patients with CKD.77 This treatment reducedrisks of major atherosclerotic events, but not deaths orhealth outcomes related to kidney disease. Addition-ally, SHARP and AURORA expanded the knowledgebase about initiating statins or statin-ezitimibe inpatients with diabetes treated by hemodialysis.77,80

Taken together, these data do not provide convincingevidence for benefits on overall clinical CVD eventsin this specific group. Guideline 4 no longer includesrecommendations for an LDL-C concentration atwhich statin therapy should be initiated or a therapeu-tic target concentration to be achieved because thestudies were not conducted in this manner and evi-dence is lacking to guide therapy by LDL-C concen-tration.

Clinical Practice Recommendation 1, now referredto as Clinical Practice Guideline 6, was revised torecommend that ACE-Is and ARBs not be used inpatients with diabetes and CKD who have normalblood pressure and normoalbuminuria. Clinical trialsof patients with either type 1 or 2 diabetes found thesetreatments did not reduce the development of elevatedurinary albumin excretion or the structural evidenceof DKD.91,93 The use of ACE-Is or ARBs is stillsuggested in normotensive patients with diabetes andelevated albuminuria who are at high risk of DKD orits progression, but this suggestion is based on low-level evidence.

This update to the KDOQI Clinical Practice Guide-lines and Clinical Practice Recommendations for Dia-betes and Chronic Kidney Disease reflects our under-standing of the present state of knowledge. Manyquestions about optimal management of DKD remainunanswere

Documents

KDOQI Clinical Practice Guideline for Diabetes and CKD ...€¦ · Minnesota Center for Obesity, Metabolism and Endocrinology, PA (MNCOME) Eagan, MN, USA ... Table 1. Grade for Strength