Assessment of Undiagnosed Diabetes Puts the Brakes on Climbing Prevalence Rates

SANTA MONICA, Calif., Aug. 13 -- Better diabetes detection over the past 25 years has actually reduced estimates of skyrocketing disease and shifted disparities from race and ethnicity to education.

Over the past 25 years, diabetes prevalence among men doubled from 3.1% to 7.1% using the conventional measure of diagnosed cases only, said James P. Smith, Ph.D., of the nonprofit Rand Corporation here, in an analysis of National Health and Nutrition Examination Surveys (NHANES).

But a more accurate prevalence measure, including both diagnosed and undiagnosed cases, for the same period showed only a 50% increase from 6.0% to 8.9%, Dr. Smith reported in the August issue of the Proceedings of the National Academy of Sciences.

"Although race and ethnic differentials in undiagnosed diabetes were eliminated over the last 25 years, the disparities became larger across other measures of disadvantage such as education," he noted.

Dr. Smith analyzed nationally representative data from NHANES II (1976 to 1980), III (1988 to 1994), and IV (1999 to 2002).

The surveys included data on prevalence of physician-diagnosed diabetes self-reported during personal interviews without distinction between type 1 and type 2 disease.

Participants also underwent a physical exam and a random subset had laboratory tests. Undiagnosed diabetes was determined by absence of self-reported diabetes but presence of clinical diabetes on the laboratory tests, defined as fasting plasma glucose of 140 mg/dl or greater or a glycosylated hemoglobin A1c of at least 6.5%.

The study included adults ages 25 through 70, the range for which full data was available, but excluded women because gestational diabetes reporting was inconsistent across the surveys.

Dr. Smith found a dramatic increase in diagnosed diabetes, from 3.1% in the period starting in 1976 to 4.6% in the third wave of the survey to 7.1% for the period ending in 2002.

Undiagnosed diabetes, though, declined. In the period starting in 1976 it accounted for 48.2% of all diabetes cases; in the period ending in 2002, that figure was just 21.6%.

Therefore, the trend in total diabetes prevalence among men was less dramatic, "although still real and significant," Dr. Smith said. It rose from 6.0% to 6.8% to 8.9% over the respective survey periods across all races and ethnicities.

Notably, a quarter of the increase in diagnosed diabetes during the study actually represented improved detection, he said, making the drop in undiagnosed diabetes one of the top factors accounting for trends.

Another of the most important factors in prevalence trends over the past 25 years has been the increase in the number of men who had a diabetic parent, which predicted a 1.39% increase in prevalence of the 2.9% increase seen in the study.

Even more important, though, was the increase in obesity. Increasing incidence of overweight and obesity among men over the study period was predicted to raise diabetes prevalence 2.15%.

Race and ethnicity factored little in the overall increase, Dr. Smith said. Together these factors predicted only a 0.15% increase compared with the 2.9% that prevalence rose, he estimated.

Total prevalence was significantly higher among both Hispanic and African-American men than among non-Hispanic white men even after controlling for other factors.

But, the disparity in the proportion of undiagnosed cases had "for all practical purposes been eliminated" in the most recent NHANES (21.4%, 24.3%, and 21.2%, respectively) compared with NHANES II (65.4%, 40.3%, and 46.0%).

These results suggest "a declining significance of race" in the likelihood of being diagnosed, Dr. Smith said.

But as these disparities disappeared, others-most notably, education-arose.

In NHANES II, there was essentially no difference in the prevalence of undiagnosed diabetes between men with less than a high school degree (6.4%), only a high school diploma (6.0%), and more than a high school diploma (7.2%).

But by the most recent survey, the difference was pronounced between men who had more than a high school education and those without a high school diploma (6.0% versus 9.8% prevalence).

The reason education became a factor may be that lower educational level was linked to Hispanic and African-American race or ethnicity, less vigorous physical exercise, more smoking, and more obesity.

Age, race, ethnicity, and body mass index accounted for 60% to 75% of the effect of schooling on diabetes prevalence.

"Those in lower education groups face a triple diabetes threat" with higher risk of developing the disease, of going undiagnosed, and of having difficulty managing their disease with complex treatments, Dr. Smith said.

However, "partially counteracting these disturbing trends in diabetes prevalence, several recent studies have shown that health consequences of diabetes have declined over time," he concluded.