ADA: For Diabetes Screening, Simpler and Cheaper Is Just as Good

CHICAGO, June 28 -- For diabetes and prediabetes screening, random plasma glucose testing alone appears to be at least as good as the full gamut of metabolic syndrome tests, and is easier and cheaper to boot, researchers found.

Nonfasting plasma glucose was as predictive of "gold-standard" oral glucose tolerance test results as metabolic syndrome testing with fasting plasma glucose as a component (0.714 versus 0.733), said Lawrence S. Phillips, M.D., of Emory University in Atlanta, at the American Diabetes Association meeting here.

And it was more predictive than metabolic syndrome testing without the fasting glucose component (0.714 versus 0.675), Dr. Phillips and colleagues found.

Although there has been interest in using the metabolic syndrome as a marker of unrecognized glucose intolerance, "clinicians ought to pay more attention to random plasma glucose," Dr. Phillips said.

The researchers' study included 1,155 adults not known to have diabetes and without acute illness. The group of volunteers was 63% female and 54% black, with an average age of 48 and an average body mass index of 30.3 kg/m².

Participants came for an initial visit at which a plasma sample was taken for glucose testing. This was considered "random" because there were no dietary restrictions before the test and the visits could have occurred at any time in the morning or afternoon.

At the second visit, two or three weeks later, participants underwent a 75g oral glucose tolerance test in the morning after an overnight fast. Their lipids, blood pressure, and waist circumference were also measured.

Among the volunteers, 5.1% had unrecognized diabetes and 20% had pre-diabetes, which was defined as impaired glucose tolerance or a fasting plasma glucose level of 110 to 125 mg/dL or both.

The metabolic syndrome was present among 27% of the group overall. Metabolic syndrome criteria were a waist circumference over 102 cm for men or 88 cm for women, blood pressure at least 130/85 mm Hg, triglycerides at least 150 mg/dL, and HDL over 40 mg/dL for men or 50 for women.

The researchers compared accuracy of the metabolic syndrome and plasma glucose tests in identifying glucose intolerance in an area under the curve in receiver-operator-characteristic analysis.

This analysis showed that random plasma glucose was as good or better at screening as the metabolic syndrome testing parameters.

The area under the curve findings were better for diabetes screening with random plasma glucose than the metabolic syndrome testing whether it included or excluded fasting plasma glucose (0.815 versus 0.796 and 0.694, respectively).

For diabetes and prediabetes screening, accuracy was similar between random plasma glucose and metabolic syndrome when it included fasting plasma glucose (0.714 versus 0.733). But it was better with random plasma glucose than the metabolic syndrome when fasting plasma glucose was not included (0.714 versus 0.675).

The same pattern was seen when screening for all three criteria-diabetes and prediabetes defined both by impaired glucose tolerance and impaired fasting glucose (0.725 for fasting plasma glucose versus 0.76 and 0.685 for the metabolic syndrome with and without fasting plasma glucose).

"The contribution of metabolic syndrome testing to diabetes screening is mainly fasting plasma glucose," Dr. Phillips said. "People don't realize how much that contributes."

Interestingly, the pattern of predictive ability seen in the area under the curve analysis held true for all subgroups, except for younger, slimmer participants.

For patients with a body mass index less than 25 kg/m² and those younger than 40, metabolic syndrome testing was significantly better than random plasma glucose at predicting diabetes regardless of whether fasting plasma glucose was included or excluded as a component.

"We don't know what it means biologically," Dr. Phillips said. But, it may be important clinically because this is a population that is often overlooked for screening because they are assumed to be at lower risk, he said.

"There's something going on in these people's bodies that even though they are younger and even though they are slimmer, they may be more insulin deficient than you think," he said.

For most patients, though, the researchers concluded that both random plasma glucose and metabolic syndrome testing with fasting plasma glucose are good predictors of diabetes.

But, random plasma glucose testing is more convenient for patients and costs less, Dr. Phillips said. He did note, however, that metabolic syndrome testing has added usefulness in that it can be used to predict cardiovascular risk as well.