Skip to main content
Premium Trial:

Request an Annual Quote

Diabetes Could be Diagnosed in Single Blood Test, Hopkins Study Suggests


NEW YORK (360Dx) – Researchers at Johns Hopkins University have determined that a single blood test can be sufficient to confirm a diabetes diagnosis.

In a study published this week in Annals of Internal Medicine, the researchers found that patients with elevated levels of both blood glucose and hemoglobin A1c in a single blood sample were subsequently diagnosed with diabetes with a specificity greater than 98 percent.

This suggests that in these patients, the standard practice of performing a second confirmatory blood test is not necessary, said Elizabeth Selvin, professor of epidemiology at Hopkins and first author on the study. Making a diagnosis based on a single blood test could reduce costs and improve diabetes care, she added.

"The current clinical practice guidelines require a confirmatory test to make the diagnosis of diabetes, and they specifically state that the same test should be repeated in a second blood sample at a separate time within a short period," she said. "That's going to require a second patient visit and a second blood draw, and that whole process can be burdensome on the patient."

Waiting for a second visit can delay the start of treatment for patients who do have diabetes. Additionally, a certain proportion of patients never return for follow-up, though Selvin said there was not good data on what percentage of at-risk patients do not return for a second test, adding that this likely varies depending on the specific patient population and healthcare system.

In any case, she said, "if you can avoid a second visit and make treatment decisions right then and there, that's a benefit to the patient and can make the whole process of delivering care more streamlined."

To investigate whether testing at a single time point could effectively diagnose diabetes, Selvin and her colleagues used data from the Atherosclerosis Risk in Communities (ARIC) study, which looked at 13,346 patients (12,268 without diabetes) over 25 years, tracking their cardiovascular outcomes along with the development of diabetes and kidney disease and mortality.

Of the 12,268 subjects without diabetes, 978 had elevated levels of fasting glucose (≥7.0 mmol/L) or HbA1c (≥6.5 percent) at the study baseline. Of those 978, 39 percent had elevated levels of both markers, which the researchers considered "confirmed undiagnosed diabetes," while 61 percent had elevated levels of only one of the two markers, which the researchers considered "unconfirmed undiagnosed diabetes."

The confirmed undiagnosed diabetes definition identified patients who would receive a diagnosis of diabetes within five years with a sensitivity of 54.9 percent and a specificity of 98.1 percent, indicating that most patients with elevated levels of both glucose and HbA1c at the single time point were, in fact diabetic. Extending follow-up time to 15 years, specificity rose to 99.6 percent.

The relatively low sensitivity of the confirmed undiagnosed definition means that a large percentage of patients positive for diabetes would not be diagnosed by the single sampling time point, but the high specificity suggests that patients who do have elevated levels of both markers at a single time point can start treatment for diabetes without waiting for a second confirmatory test.

"There will be the people for whom both [markers] are negative, and in that case, you're ruling out diabetes—those individuals are low risk for anything happening in the short run," she said. "If one is elevated and the other is not, then we would follow current practice guidelines to follow up [with a second test]. But if both are elevated, we should be able to diagnose diabetes [based on the single blood draw] … and figure out based on the A1C and other factors what the appropriate treatment would be."

Selvin said that while it would be good to replicate the findings in other cohorts, she believes the evidence presented in the Annals of Internal Medicine study is sufficient to start changing clinical practice.

"The ARIC study [used] a large, well-characterized cohort of both black and white adults in the United States," she said. "Their findings are generally very generalizable."

"It's already common for physicians to measure the two blood tests in one sample, and I think it makes a lot of sense to streamline and formally recommend that the diagnosis can be made if both those tests are elevated," she said. "I think this study should have a direct impact on guidelines and that we should incorporate this into the clinical practice."

An accompanying editorial in Annals of Internal Medicine by Venkat Narayan and Ram Jagannathan, researchers at Emory University, was more circumspect, however.

The authors, who were not involved in Selvin's work, noted the promise of the findings, "especially for resource-challenged settings," but highlighted several potential concerns with the study.

One they noted was the fact that glucose measurements were done at five-year intervals in the ARIC study, meaning that the procedure for diagnosing diabetes within that study was not equivalent to that used in current clinical practice.

"Ideally, an effort to determine whether the diagnostic accuracy of a single-sample, [two]-test approach approximates that of a repeated test at a subsequent visit would use a confirmatory test within a much shorter period to replicate clinical practice," they wrote.

They also noted that as the ARIC cohort consisted of black and white Americans between 45 and 64 years of age, "the generalizability to other groups with different prevalence of diagnosed and undiagnosed diabetes is not clear."

The approach put forth by Selvin and her coauthors "has appeal," they noted, "but it needs replication in other populations before becoming accepted clinical practice."