brand logo

Am Fam Physician. 2006;74(1):143-144

Clinical Question

Can A1C measurements be used to diagnose diabetes mellitus?

Evidence-Based Answer

A1C testing is highly specific compared with a two-hour oral glucose tolerance test (OGTT) or a fasting plasma glucose test. However, because A1C testing is not sensitive enough to rule out diabetes if levels are normal, the test should not be used for diagnosing diabetes.1 (Strength of recommendation: C).

Evidence Summary

The accuracy of A1C testing for the diagnosis of diabetes was first evaluated using pre-1997 World Health Organization (WHO) criteria for a diabetes diagnosis (i.e., OGTT result more than 140 mg per dL [7.8 mmol per L] fasting or more than 200 mg per dL [11.1 mmol per L] at two hours). A meta-analysis1 (including 10 studies that evaluated A1C levels of 8,984 patients) that used a cutoff (i.e., upper limit of normal) A1C level of 6.3 percent (two standard deviations [SD] above the mean) showed that A1C testing was 66 percent sensitive and 98 percent specific for diagnosing diabetes (Table 11,2). Based on an assumed diabetes prevalence rate of 6 percent in the general population (i.e., persons without known diabetes), 63 percent of patients with an A1C level more than 6.3 percent would have diabetes (positive predictive value [PPV] = 63 percent).

Cutoff A1C level (%) (standard deviations above the mean)Sensitivity (%)Specificity (%)Positive predictive value*
Peters, et al., 19961
6.3 (2)669863
6.8 (3)4810090
7.3 (4)3610097
Rohlfing, et al., 20002
5.6 (1)83.484.425
6.1 (2)63.297.457
6.5 (3)42.899.686
7.0 (4)28.399.994

A subsequent study2 that used current WHO criteria for defining diabetes (i.e., fasting plasma glucose of 126 mg per dL [7.0 mmol per L] or more) showed similar results. This cross-sectional study3 of 6,559 patients that used a cutoff A1C level of 6.1 percent and assumed a diabetes prevalence rate of 6 percent showed A1C testing was 63.2 percent sensitive and 97.4 percent specific for diagnosing diabetes (PPV = 57 percent). A subgroup analysis2 of these results showed that the sensitivity of A1C testing was considerably higher in Mexican Americans and blacks (83.6 and 75.8 percent, respectively) compared with non-Hispanic whites (58.6 percent). The reason for these differences is unclear; however, these results may suggest physiologic variations among ethnic groups. If this is verified in future, larger studies, A1C testing may have a role in diagnosing diabetes in some patient populations.

Recommendations from Others

The American Diabetes Association (ADA) does not recommend A1C testing for the diagnosis of diabetes.3 Instead, it recommends fasting plasma glucose testing (126 mg per dL or more is diagnostic) as the test of choice because it is easier and faster to perform and less expensive than other screening tests. Table 23 includes the ADA criteria for diagnosing diabetes. Although the ADA guideline points out that A1C testing historically had not been standardized among laboratories, the test has become widely standardized in the United States since the guideline’s publication.4 The European Diabetes Policy Group guideline5 also does not recommend A1C testing for diagnosing diabetes and recommends confirming the diagnosis in persons with elevated A1C levels using fasting plasma glucose testing (126 mg per dL or more is diagnostic).

Diabetes mellitus can be diagnosed if at least one of the following clinical findings is present:
Symptoms of diabetes (e.g., polyuria, polydipsia, unexplained weight loss) plus casual† plasma glucose ≥ 200 mg per dL (11.1 mmol per L)
Fasting‡ plasma glucose ≥ 126 mg per dL (7.0 mmol per L)
Two-hour postload glucose ≥ 200 mg per dL according to an oral glucose tolerance test§

Clinical Commentary

Although most of the literature on diabetes includes the two-hour OGTT as the standard test for diagnosing diabetes, in daily clinical practice, this approach is not always feasible. Fasting plasma glucose testing has been shown to have slightly lower sensitivity for predicting microvascular complications from diabetes compared with the two-hour OGTT, but it is easier to perform compared with the two-hour OGTT. Therefore, fasting plasma glucose testing is widely used as the preferred diagnostic test.

Some physicians believe that because glycemic treatment for diabetes is not aggressively initiated until A1C levels are 7 percent or more, A1C testing can be used for diagnosis despite its low sensitivity. However, there are other important interventions (e.g., improving blood pressure, lowering low-density lipoprotein cholesterol, aspirin therapy, influenza vaccination) that are recommended for patients with diabetes, even those with A1C levels less than 7 percent.6 Therefore, fasting plasma glucose testing remains the preferred diagnostic test for diabetes.

Continue Reading


More in AFP

More in Pubmed

Copyright © 2006 by the American Academy of Family Physicians.

This content is owned by the AAFP. A person viewing it online may make one printout of the material and may use that printout only for his or her personal, non-commercial reference. This material may not otherwise be downloaded, copied, printed, stored, transmitted or reproduced in any medium, whether now known or later invented, except as authorized in writing by the AAFP.  See https://www.aafp.org/about/this-site/permissions.html for copyright questions and/or permission requests.