Background: The American Diabetes Association (ADA) recommends that treatment with metformin be considered for prevention of type 2 diabetes in persons with prediabetes. However, metformin use outside the setting of diagnosed diabetes in US adults is not well characterized.

Objective: To examine trends in self-reported prediabetes and treatment with metformin. We also compared characteristics of adults self-reported prediabetes who were vs. were not taking metformin.

Design: Cross-sectional analysis.

Participants: Adults ≥ 20 years of age who participated in the 2005-2014 National Health and Nutrition and Examination Survey (NHANES), n = 28,461.

Approach: We characterized trends in self-reported prediabetes and metformin use in this population. We used multiple logistic regression models to identify predictors of metformin use among adults with self-reported prediabetes. All analyses accounted for the weighted complex survey design to generate nationally representative estimates.

Key results: The prevalence of self-reported prediabetes increased from 5.1% in 2005-2006 to 7.4% in 2013-2014 (P-for-trend < 0.001). In persons with self-reported prediabetes, metformin use increased, from 2.4 to 8.3% (P-for-trend = 0.013). Adults who were taking metformin were more likely to be obese and to report trying to lose weight in the past year.

Conclusions: Self-reported prediabetes has increased in the past decade. Metformin use in the setting of prediabetes has also increased but remains relatively uncommon at 8% in adults who self-report prediabetes, despite current clinical recommendations.