2.3.0.0 A1C

Rec­om­men­da­tions

2.1 To avoid misdi­ag­no­sis or missed di­ag­no­sis, the A1C test should be per­formed using a method that is certified by the NGSP and stan­dard­ized to the Di­a­betes Con­trol and Com­pli­ca­tions Trial (DCCT) assay. B

2.2 Marked dis­cor­dance be­tween mea­sured A1C and plas­ma glu­cose lev­els should raise the pos­si­bil­i­ty of A1C assay in­ter­fer­ence due to hemoglobin vari­ants (i.e., hemoglobinopathies) and con­sid­er­a­tion of using an assay with­out in­ter­fer­ence or plas­ma blood glu­cose cri­te­ria to di­ag­nose di­a­betes. B

2.3 In con­di­tions as­so­ci­at­ed with an al­tered re­la­tion­ship be­tween A1C and glycemia, such as sick­le cell dis­ease, preg­nan­cy (sec­ond and third trimesters and the post­par­tum pe­ri­od), glu­cose-‍6-‍phos­phate de­hy­dro­ge­nase deficien­cy, HIV, hemodial­y­sis, re­cent blood loss or trans­fu­sion, or ery­thro­poi­etin ther­a­py, only plas­ma blood glu­cose cri­te­ria should be used to di­ag­nose di­a­betes. B

The A1C test should be per­formed using a method that is certified by the NGSP (www.ngsp.org) and stan­dard­ized or trace­able to the Di­a­betes Con­trol and Com­pli­ca­tions Trial (DCCT) ref­er­ence assay. Al­though point-‍of-‍care A1C as­says may be NGSP certified or U.S. Food and Drug Ad­min­is­tra­tion ap­proved for di­ag­no­sis, proficien­cy test­ing is not al­ways man­dat­ed for per­form­ing the test. There­fore, point-‍of-‍care as­says ap­proved for di­ag­nos­tic pur­pos­es should only be con­sid­ered in set­tings li­censed to per­form mod­er­ate-‍to-‍high com­plex­i­ty tests. As dis­cussed in Sec­tion 6 “Glycemic Tar­gets,” point-‍of-‍care A1C as­says may be more gen­er­ally ap­plied for glu­cose mon­i­tor­ing.

The A1C has sev­er­al ad­van­tages com­pared with the FPG and OGTT, in­clud­ing greater con­ve­nience (fast­ing not re­quired), greater pre­an­a­lyt­i­cal sta­bil­i­ty, and less day-‍to-‍day per­tur­ba­tions dur­ing stress and ill­ness. How­ev­er, these ad­van­tages may be off­set by the lower sen­si­tiv­i­ty of A1C at the des­ig­nat­ed cut point, greater cost, lim­it­ed avail­abil­i­ty of A1C test­ing in cer­tain re­gions of the de­vel­op­ing world, and the im­per­fect cor­re­la­tion be­tween A1C and av­er­age glu­cose in cer­tain in­di­vid­u­als. The A1C test, with a di­ag­nos­tic thresh­old of ≥6.5% (48 mmol/‍mol), di­ag­noses only 30% of the di­a­betes cases iden­tified col­lec­tive­ly using A1C, FPG, or 2-h PG, ac­cord­ing to Na­tion­al Health and Nu­tri­tion Ex­am­i­na­tion Sur­vey (NHA­NES) data (10).

When using A1C to di­ag­nose di­a­betes, it is im­por­tant to rec­og­nize that A1C is an in­di­rect mea­sure of av­er­age blood glu­cose lev­els and to take other fac­tors into con­sid­er­a­tion that may im­pact hemoglobin gly­ca­tion inde­pen­dently of glycemia in­clud­ing HIV treat­ment (11,12), age, race/ eth­nicity, preg­nan­cy sta­tus, ge­net­ic back­ground, and ane­mia/‍hemoglobinopathies.

Age

The epi­demi­o­log­i­cal stud­ies that formed the basis for rec­om­mend­ing A1C to di­ag­nose di­a­betes in­clud­ed only adult pop­u­la­tions (10). How­ev­er, a re­cent ADA clin­i­cal guid­ance con­clud­ed that A1C, FPG, or 2-h PG can be used to test for predi­a­betes or type 2 di­a­betes in chil­dren and ado­les­cents. (see p. S20 SCREEN­ING AND TEST­ING FOR PREDI­A­BETES AND TYPE 2 DI­A­BETES IN CHIL­DREN AND ADO­LES­CENTS for ad­di­tion­al in­for­ma­tion) (13).

Race/‍Eth­nic­i­ty/‍Hemoglobinopathies

Hemoglobin vari­ants can in­ter­fere with the mea­surement of A1C, al­though most as­says in use in the U.S. are un­af­fect­ed by the most com­mon vari­ants. Marked dis­crep­an­cies be­tween mea­sured A1C and plas­ma glu­cose lev­els should prompt con­sid­er­a­tion that the A1C assay may not be re­li­able for that in­di­vid­u­al. For pa­tients with a hemoglobin vari­ant but nor­mal red blood cell turnover, such as those with the sick­le cell trait, an A1C assay with­out in­ter­fer­ence from hemoglobin vari­ants should be used. An up­dat­ed list of A1C as­says with in­ter­fer­ences is avail­able at www.ngsp.org/‍interf.asp.

African Amer­i­cans het­erozy­gous for the com­mon hemoglobin vari­ant HbS may have, for any given level of mean glycemia, lower A1C by about 0.3% than those with­out the trait (14). An­oth­er ge­net­ic vari­ant, X-‍linked glu­cose-‍6-‍phos­phate de­hy­dro­ge­nase G202A, car­ried by 11% of African Amer­i­cans, was as­so­ci­at­ed with a de­crease in A1C of about 0.8% in ho­mozy­gous men and 0.7% in ho­mozy­gous women com­pared with those with­out the vari­ant (15).

Even in the ab­sence of hemoglobin vari­ants, A1C lev­els may vary with race/eth­nicity inde­pen­dently of glycemia (16-18). For ex­am­ple, African Amer­i­cans may have high­er A1C lev­els than non-‍His­pan­ic whites with sim­i­lar fast­ing and postglu­cose load glu­cose lev­els (19), and A1C lev­els may be high­er for a given mean glu­cose con­cen­tra­tion when mea­sured with con­tin­u­ous glu­cose mon­i­tor­ing (20). Though conflict­ing data ex­ists, African Amer­i­cans may also have high­er lev­els of fruc­tosamine and gly­cat­ed al­bu­min and lower lev­els of 1,5-‍an­hy­droglu­ci­tol, sug­gest­ing that their glycemic bur­den (par­tic­u­lar­ly post­pran­di­al­ly) may be high­er (21,22). The as­so­ci­a­tion of A1C with risk for com­pli­ca­tions ap­pears to be sim­i­lar in African Amer­i­cans and non-‍His­pan­ic whites (23,24).

Other Con­di­tions Al­ter­ing the Re­la­tion­ship of A1C and Glycemia

In con­di­tions as­so­ci­at­ed with in­creased red blood cell turnover, such as sick­le cell dis­ease, preg­nan­cy (sec­ond and third trimesters), glu­cose-‍6-‍phos­phate de­hy­dro­ge­nase deficien­cy (25,26), hemodial­y­sis, re­cent blood loss or trans­fu­sion, or ery­thro­poi­etin ther­a­py, only plas­ma blood glu­cose cri­te­ria should be used to di­ag­nose di­a­betes (27). A1C is less re­li­able than blood glu­cose mea­surement in other con­di­tions such as post­par­tum (28-30), HIV treat­ed with cer­tain drugs (11), and iron deficient ane­mia (31).