Physical activity and pre-diabetes-an unacknowledged mid-life crisis: findings from NHANES 2003-2006.

The prevalence of pre-diabetes (PD) among US adults has increased substantially over the past two decades. By current estimates, over 34% of US adults fall in the PD category, 84% of whom meet the American Diabetes Association's criteria for impaired fasting glucose (IFG). Low physical activity (PA) and/or sedentary behavior are key drivers of hyperglycemia. We examined the relationship between PD and objectively measured PA in NHANES 2003-2006 of 20,470 individuals, including 7,501 individuals between 20 and 65 yrs.We excluded all participants without IFG measures or adequate accelerometry data (final N = 1,317). Participants were identified as PD if FPG was 100-125 mg/dL (5.6-6.9 mmol/L). Moderate and vigorous PA in minutes/day individuals were summed to create the exposure variable "moderate-vigorous PA" (MVPA). The analysis sample included 884 normoglycemic persons and 433 with PD. There were significantly fewer PD subjects in the middle (30.3%) and highest (24.6%) tertiles of PA compared to the lowest tertile (35.5%). After adjusting for BMI, participants were 0.77 times as likely to be PD if they were in the highest tertile compared to the lowest PA tertile (p < 0.001). However, these results were no longer significant when age and BMI were held constant. Univariate analysis revealed that physical activity was associated with decreased fasting glucose of 0.5 mg/dL per minute of MVPA, but multivariate analysis adjusting for age and BMI was not significant. Overall, our data suggest a negative association between measures of PA and the prevalence of PD in middle-aged US adults independent of adiposity, but with significant confounding influence from measures of BMI and age.