LONG BEACH, CALIF. — Identifying patients most at risk for progression to diabetes in 5-10 years may become easier with the development of several scoring systems for quantifying risk.
About 57 million Americans have prediabetes, as judged by impaired fasting glucose level, but only about 10% of them will go on to develop type 2 diabetes within 5 years. Because it's impractical to target all of those 57 million people with intensive interventions, several strategies have emerged to identify those most at risk.
According to speakers at a conference on diabetes sponsored by the Centers for Disease Control and Prevention, some of these strategies appear to get good results. Investigators at the CDC, for example, have developed two simple scoring systems—one of which doesn't even require a blood test—that quantify an individual's chance of developing diabetes within 10 years. And researchers at a private company, Tethys Bioscience of Emeryville, Calif., have developed an algorithm based on seven biomarkers that can categorize individuals as being at high, medium, or low risk for progressing to diabetes within 5 years.
“I think in the U.S. population, our user-friendly scoring systems can really help us to target individuals, at least among blacks and whites, who will benefit most from a prevention program,” said Dr. Henry S. Kahn of the CDC. “Our basic system could be used to identify those maybe who should just be escalated … to a simple blood test. You could use the basic scoring system through elementary examinations to find out which subsets of your community are in the greatest need of prevention resources.”
Dr. Kahn and his colleagues used 10-year follow-up data from the Atherosclerosis Risk in Communities (ARIC) study, which involved 12,729 U.S. adults who did not have diabetes at baseline. The investigators randomly selected 75% of that cohort to derive their scoring systems, reserving the other 25% to demonstrate that the systems work.
The basic scoring system uses only easily obtainable data, and the enhanced scoring system adds information from simple blood tests. Both systems yield a diabetes prediction score (DPS) from 0 to 100 that represents the chance that an individual would develop diabetes within 10 years.
The basic DPS assigns point values to hypertension, family history, black race, age 55-64 years, smoking history, waist circumference, height, resting pulse rate, and weight. Among the validation cohort, basic DPS scores in the lowest quintile corresponded to a 10-year incidence of diabetes of 5%, while scores in the highest quintile corresponded to a 10-year incidence of 32%.
The enhanced DPS eliminates smoking history and weight from the scoring system and adds point values for height and for levels of blood glucose, triglycerides, HDL cholesterol, and uric acid. Among the validation cohort, scores in the lowest quintile corresponded to a 10-year incidence of diabetes of 4%, while scores in the highest quintile corresponded to a 10-year incidence of 46%.
Dr. Kahn said that if incorporated within public health surveys, either scoring system would likely identify some adults with undiagnosed diabetes and some at elevated risk for cardiovascular disease.
The Tethys test employs a different strategy. After analyzing several large European cohorts, investigators identified seven biomarkers that together form a fingerprint identifying individuals likely to develop diabetes within 5 years. Those biomarkers are levels of fasting glucose, hemoglobin A1c, insulin, adiponectin (associated with obesity), ferritin (associated with cell death, especially in the liver), and C-reactive protein and interleukin-2 receptor-alpha (both associated with inflammation and cardiovascular risk).
Michael McKenna, Ph.D, chief scientific officer at Tethys, explained that physicians interested in ordering the test need only request a blood draw and have a standard 3-mL red-cap tube of serum sent to Tethys's CLIA-approved laboratory.
The test, which is called PreDx, yields a score between 0 and 10 that corresponds to the risk of incident diabetes within 5 years. Among the overall test cohort, 5.7% developed diabetes within 5 years. But that was about 2% among the 54% of individuals judged to be at low risk based on their PreDx score, 7% among the 36% judged to be at medium risk, and 20% among the 10% evaluated as high risk. Individuals with high scores were 16.7 times as likely to develop diabetes as those with low scores.
Dr. Eric Book, Tethys's chief medical officer, said that a report describing these results in Northern European populations was to be published in the June 2009 issue of the journal Diabetes Care. But in order to persuade American health plans to cover PreDx, the company will need to demonstrate that the results hold up in multiethnic populations. Those studies will be conducted during summer and fall 2009, with results being published late in 2009 and in spring 2010.