Childhood Factors May Predict Diabetes Later in Life

According to Charles Glueck, MD, of the Jewish Hospital of Cincinnati, in two long-term cohort studies, measures such as systolic blood pressure, body mass index, and parental diabetes were associated with significantly increased risks of diabetes at ages 19 and 39. Adding laboratory measurements, such as fasting serum insulin, glucose, and lipids, improved the predictive ability, they reported. 

"Our data have practical clinical value in assessment of pre-teenaged and teenaged children," the authors wrote.

They said children with combinations of these risk factors "could be targeted for primary prevention of Type 2 diabetes through diet, exercise, and possibly insulin-sensitizing drug intervention, with special focus on overweight children with a positive family history of diabetes mellitus."

Glueck and his colleagues examined data from two long-term, prospective cohort studies -- the Princeton Follow-up Study (PFS), which included 822 schoolchildren followed for 22 to 30 years, and the National Growth and Health Study (NGHS), which followed 1,067 black and white girls for nine years.

In the PFS, the rate of Type 2 diabetes at the end of follow-up -- mean age 39 -- was 4.9%.

Significant childhood office-based predictors of developing diabetes were systolic blood pressure in the top fifth percentile, BMI in the top fifth percentile, and black race.

In a multiple logistic regression analysis, these three factors together significantly predicted Type 2 diabetes with an area under the receiver-operator curve (AUC) of 0.698).

Adding a childhood glucose level of 100 mg/dL or higher and triglyceride concentration in the top fifth percentile each improved the strength of the model (AUCs of 0.717 and 0.709, respectively).

If childhood BMI, systolic blood pressure, and diastolic blood pressure were all lower than the 75th percentile, the likelihood of diabetes at age 39 was just 2%. The rate was only 1.4% if the child's parents did not have diabetes.

In the NGHS, 1.2% of black women and 0.2% of white women had Type 2 diabetes at age 19.

Having Type 2 diabetes was significantly predicted by childhood systolic blood pressure in the top fifth percentile (OR 7.72, 95% CI 1.47 to 40.55) and parental diabetes (OR 5.22, 95% CI 1.00 to 27.23). Taken together, the two variables significantly predicted the condition with an AUC of 0.699 in a multiple logistic regression analysis.

Adding insulin in the top fifth percentile increased the AUC to 0.764. If childhood BMI, systolic blood pressure, and diastolic blood pressure were all lower than the 75th percentile, the chance of having Type 2 diabetes at age 19 was only 0.2%.

Practice Pearl: Explain to interested patients that this study did not establish a causal relationship between any of the childhood factors and adult Type 2 diabetes.

Morrison J, et al "Childhood predictors of adult Type 2 diabetes at nine- and 26-year follow-ups" Arch Pediatr Adolesc Med 2010; 164: 53-60.