Hemoglobin A1C 6.5 or higher %

One result does not make the diagnosis. Recommendation 2.1b: "In the absence of unequivocal hyperglycemia (e.g., hyperglycemic crises), diagnosis requires confirmatory testing." That means two abnormal results, either from two different tests drawn at the same time (an A1C and a fasting glucose, for example) or from the same test on two occasions. The Standards give the worked example: an A1C of 7.0 percent followed by a repeat of 6.8 confirms diabetes. If two different tests disagree, the one above its cut point is repeated, and the diagnosis rests on that confirmatory result. The exception is someone with classic symptoms such as excessive thirst and urination or unexplained weight loss and a random glucose of 200 mg/dL or higher, where the glucose alone is sufficient.

When this number should not be trusted. Recommendation 2.4 says to diagnose on plasma glucose instead of A1C in "some hemoglobin variants, pregnancy, glucose-6-phosphate dehydrogenase deficiency, HIV, and conditions that may alter red blood cell turnover." Table 2.3 lists those: anemia, iron status, splenectomy, blood loss, transfusion, hemolysis, erythropoietin, cirrhosis, kidney failure, dialysis. Some assays misread hemoglobin variants, and the Standards say an assay without that interference should be used in people with sickle cell trait. A G6PD variant carried by 11 percent of Black individuals in the US lowers A1C by roughly 0.7 to 0.8 percent in people who carry two copies. Recommendation 2.3 adds that a consistent, substantial mismatch between glucose readings and A1C should itself be investigated. The assay must be NGSP certified and traceable to the DCCT reference (Recommendation 2.2a).

What follows a confirmed diagnosis. Recommendation 6.2 sets the rhythm: assess glycemic status "at least two times a year, and more frequently (e.g., every 3 months)" when goals are not met, treatment has changed, or hypoglycemia or hyperglycemia is frequent. The default goal for many nonpregnant adults is an A1C under 7 percent, individualized up or down by health, function and hypoglycemia risk; that band has its own page. The Standards translate A1C into an estimated average glucose from the ADAG study: 6 percent corresponds to about 126 mg/dL, 7 percent to 154, 8 percent to 183, each with a wide confidence interval, which is why the ADA calls it an estimate.