The

Pros

and

Cons

of

Using

HbA1cfor

Diabetes

DiagnosisDavid

B.

Sacks

MB,

ChB,

FRCPathSenior

Investigator,

NIHAdjunct

Professor

of

Medicine,

Georgetown

UniversityClinical

Professor

of

Pathology,

George

Washington

UniversityHonorary

Professor

of

Clinical

Laboratory

Sciences,

UCTHbA1c

SymposiumChinaMarch,

2017Worldwide

Diabetic

PopulationNature

Medicine

2005;

12:75IDF

Estimates

of

DiabetesPrevalence:

2015-20401.

2015

-415

million2.

2040

-project

642

millionIDF

20151Top

10

Countries

for

Diabetes

2015IDF

DiabetesAtlas

2015Prevalence

of

ObesityIncrease

over

40

y:Men

7.8-foldWomen

5.3-foldNature

2016;

532:11Estimated

Financial

Costs

of

Diabetes

in

the

US2007

CostsTotal

=$174

b2012

CostsTotal

=$245

b(41%

increase)Nature

Medicine

2005;

12:75Diabetes

Care

2013;

36:10332Investment

in

DiabetesNature

2012;

485:S2Why

Do

We

Need

to

Identify

Diabetes?1

Many

with

diabetes

are

undiagnosed:US~40%AfricaChina~7

8

%~50%2.

7-yr

gap

between

onset

and

diagnosis

of

type2diabetes3.

At

the

time

of

diagnosis,

25%-30%

of

patientswith

diabetes

have

irreversible

complications4.

Treatment

significantly

reduces

complicationsIDF

2015Diabetes

Care

2009;32:287

NEJM

1993;

329:977The

Major

Complications

of

DiabetesDiabetesAtlas:

IDF

20033Diagnosis

of

Diabetes

is

Difficult1.

Comprehension

of

pathophysiology

ofdiabetes

islimited2.

Lack

of

identified

unique

biologicalmarker

of

diabetes3.

Use

historically

associated

metabolicderangement,

i.e.,

hyperglycemiaDiagnosis

of

Diabetes

Mellitus

-

1975Diabetes

(%)Normal

(%)Suspects

(%)Fajans&

Conn45550WilkersonWHO244151481876424952330170BDAUGDPESGDE049Diabetes

1975;

24:585Diagnosis

of

Diabetes

Mellitus

–

ADA

19971.

FPG

3

126

mg/dL

(7.0

mmol/L

)OR2.

2-h

glucose

3

200

mg/dL

(11.1

mmol/L)

during

OGTTOR3.

Symptoms

of

hyperglycemia

and

casual

plasmaglucose

3

200

mg/dL

(11.1

mmol/L)Unless

unequivocal

hyperglycemia,

confirm

by

repeat

testing

on

adifferent

dayDiabetes

Care

19974Fasting

Plasma

Glucose

(FPG)Advantages1.

Easily

automated2.

Widely

available3.

InexpensiveDisadvantages

of

FPG

for

the

Diagnosisof

Diabetes1.

Patient

must

fast

3

8h2.

Large

biological

variability:Intraindividual

CVs

4.6-8.3%Interindividual

CVs

7.5-12.5%3.

Lack

of

sample

stability

-invitro

glycolysisChangesin

Glucose

Concentration

withTimeFluoride

vs.

HeparinWHOLEBLOODF5.00H060120424Time(min)(hr)Chan

Clin

Chem

1989;

35:

3155GlycolysisFluorideDisadvantages

of

FPG

for

the

Diagnosisof

Diabetes

(ctd.)6.

Numerous

other

factors

alter

glucose

concentrations

e.g.,acute

illness7.

No

harmonization

of

glucose

testing8.

Concentration

varies

with

source

of

the

sample

(venous,capillary

or

arterial

blood)9.

Concentration

in

whole

blood

is

different

to

that

in

plasma10.Guidelines

recommend

plasma,

but

many

laboratoriesmeasure

serum

glucose11.FPG

less

tightly

linked

to

diabetes

complications

(than

HbA1c)12.Reflects

glucose

homeostasis

at

a

single

point

in

timeSacks

Diabetes

Care

2011;34:518Post-Prandial

Glucose

(PPG)1.

Subject

to

same

limitations

as

FPG2.

Time-consuming

and

inconvenient3.

Expensive4.

Unpalatable5.

Influenced

by

numerous

medications6.

Extensive

patient

preparation7.

Lack

of

reproducibility6Poor

Reproducibility

of

OGTTFPG60

overweight

youths,2

OGTTs

1-25

d

apart2h

PGJCEM

2008;

93:4231People

With

Undiagnosed

Diabetes

2015193

million

undiagnosed

(47%

of

total)IDF

2015“…

the

Committee

believes

that

it

isstill

premature

to

add

[Hb]A1C

to

thegroup

of

tests

used

for

the

definitivediagnosis

of

diabetes.”Diabetes

Care

2003;

26:31607“Currently

HbA1c

is

not

considered

asuitable

diagnostic

test

for

diabetes….”WHO/IDF

2006“[Hb]A1c

assay

may

be

a

better

means

of

diagnosingdiabetes

than

measures

of

glucose”Diabetes

Care

2009;

32:1327WHO

Position

on

HbA1c

for

Diagnosis“HbA1c

can

be

used

as

adiagnostic

test

fordiabetes”WHO

Jan.,

20118Comparison

of

WHO/IDF

and

ADADiagnostic

CriteriaWHO/IDF

2011ADA

2010HbA1c≥6.5%≥6.5%ororFasting

plasmaglucose126

mg/dL(≥7.0

mmol/l)126

mg/dL(≥7.0

mmol/l)oror2-h

glucose*200

mg/dL(≥11.1

mmol/l)200

mg/dL

(≥11.1

mmol/l)*Venous

plasma

glucose

2-h

after

ingestion

of

75g

oral

glucose

loadWHO

recommendation

accepted

by

presidents

ofIDF,

ADA&

EASDHbA1c

for

Diagnosis:

ProBiology1.

Reflects

chronic

glycemia

(~8-12

weeks)2.

Independent

of

acute

factors

e.g.,

stress,

exercise3.

Very

low

intra-individual

variability

(CV

~1%)Analysis1.

Fasting

not

necessary2.

Blood

may

be

collected

any

time

of

the

day3.

Sample

is

stable4.

Assay

is

standardized

across

instruments5.

Accuracy

of

test

is

monitoredSacks

Diabetes

2013;

62:419HbA1c

for

Diagnosis:

Pro

(ctd.)Clinical1.

Monitor

long-term

glucose

control2.

Used

to

guide

therapy

-target

value3.

Concentration

predicts

risk

of

microvascularcomplications

of

diabetesSacks

Diabetes

2013;

62:41Prevalence

of

Diabetes-SpecificRetinopathyDETECT-2;

n=27

933Data

pooled

from

9

studiesDiabetes

Care

2011;34:145Prevalence

of

Diabetes-Specific

RetinopathyFPGOGTTHbA1cDiabetes

Care

2011;34:14510HbA1c

for

Diagnosis:

ConMay

be

altered

by

factors

other

than

glycemiaNon-glycemic

Factors

That

Alter

HbA1cNat

Rev

Endocrinol

2010;

6:589Conditions

That

Alter

HbA1c1.

Most

reports

based

on

very

small

n2.

Changes

often

small

and

unlikely

to

be

clinicallysignificant3.

Many

of

those

conditions

inold

literature

do

notalter

values

with

current

HbA1c

methods11Factors

That

Influence

HbA1c1.

Factors

that

influence

interpretation

i.e.,

alterHbA1c

value2.

Factors

that

interfere

with

HbA1c

measurementi.e.,

analyticFactors

That

May

Influence

Interpretationof

HbA1c1.

Physiological

e.g.,

age,

race2.

Chronic

kidney

disease3.

Iron

deficiency

anemia4.

Erythrocyte

lifespan5.

Glycation

“phenotypes”6.

Drugs

e.g.,

dapsone,

antiretroviralDiabetes

Care

2016;

39:1299Effect

of

Race/Ethnicity

on

HbA1c1.

Studies

inUSA

reveal

that

HbA1c

concentrationinblacks

ishigher

than

inwhites2.

Are

differences

clinically

meaningful?3.

Highly

controversial12Prevalence

of

Retinopathy

in

JapanLongitudinal

data20

433

adultsFollowed

3

yDiabetes

2012;

61:3280Prevalence

of

Retinopathy

-

Asia4cross-sectionalstudies,Singaporen=13

170

adults5

834

Chinese3

596

Malays3

740

IndiansJCEM

2015;

100:689Effect

of

Diabetes

Definition:

NCD-RisC1.

Pooled

analysis

of

96

population-based

studies46

-

Australia,

USA,

W.

Europe18

-East

&SE

Asia10

-Latin

America7-Oceana6-sub-Saharan

Africa5-South

Asia3-Middle

East

&N.

Africa2.

331

288

participantsLancet

Diab

Endo

2015;

3:62413Effect

of

Diabetes

Definition:

NCD-RisC1.

Compared

diagnosis

of

diabetes

by

FPG

only,FPG-or-2hOGTT

or

HbA1c

(≥6.5%)2.

Prevalence

based

on

FPG-or-2hOGTT

higherthan

FPG

aloneLancet

Diab

Endo

2015;

3:624Prevalence:

FPG-or-2hOGTT

vs

FPGLancetDiabEndo2015;

3:624HbA1c

vs

FPG

or

FPG-or-2hOGTTLancet

Diab

Endo

2015;

3:62414Chronic

Kidney

Disease

and

HbA1c1.

CKD

common

complication

of

diabetes2.

RBC

survival

reduced

inCKD3.

Erythropoietin

stimulates

RBC

turnover4.

Reduced

RBC

lifespan

results

inlower

HbA1c5.

Difficult

to

monitor

long

term

glycemic

controlinCKDSelvin

&

Sacks,

Clinical

Chemistry

2016,

in

the

pressFactors

That

May

Interfere

withHbA1c

Measurement1.

Uremia

(carbamylated

Hb)2.

Hemoglobin

variants3.

Drugs

e.g.,

opiates4.

Other

e.g.,

bilirubin,

triglyceride,

alcoholHemoglobin

Variants1.

1268

hemoglobin

variants

identified2.

893

(70%)

involve

the

beta

gene3.

64

mutations

infirst

13

amino

acids4.

Most

common

variants

HbS,

HbC,

HbE,

HbDHbVar:ADatabase

of

Human

Hemoglobin

Variants15Hemoglobin

Variants

and

HbA1c1.

Cannot

use

HbA1c

inindividuals

homozygousfor

HbS

or

HbC

or

with

HbSC

-no

HbA2.

Can

measure

HbA1c

accurately

in

mostheterozygous

Hb

variants,

if

appropriateassay

used

()Interference

from:MethodHbASHbACHbAEHbADImmunoassaysAbbottArchitect/AerosetBayer

(Metrika)A1cNOWYe