OBJECTIVES: To compare rates of obesity-associated comorbidities between the General Electric (GE) Centricity EMR and Thomson Reuters MedStat MarketScan commercial claims databases in patients with either a recorded BMI ≥25 (GE EMR) or ICD-9 code for overweight/obesity (MarketScan). METHODS: From the GE EMR, subjects aged 20-64 with at least one BMI value ≥25 and ≥2 y of EMR activity prior to BMI index date were included. From MarketScan, subjects ≥20 y (mostly <65 y), with a claim (ICD-9 code) for overweight, obesity, or morbid obesity, and ≥365 days of claims activity after selected comorbidity was first recorded, were included. Selected comorbidities were identified by ICD-9 codes in both databases and stratified by BMI or overweight/obesity claims. RESULTS: In the GE EMR, 109,885 subjects (36.7%) were overweight or obese (BMI ≥25). Of these, 51% had BMI 25-29.9, 28% had BMI 30-34.9, and 22% had BMI ≥35. In the MarketScan database, 246,261 subjects (0.7%) had overweight, obesity, or morbid obesity claims. Of these, 4% were overweight, 66% were obese, and 30% were morbidly obese. The three most prevalent comorbidities in both databases, across all weight categories, were hypertension, hyperlipidemia, and chronic back pain, except diabetes replaced back pain in the morbid obesity group in MarketScan. For all comorbidities, prevalence was higher in MarketScan, except polycystic ovary syndrome prevalence was higher in GE EMR BMI ≥35 group compared to morbid obesity group in MarketScan. CONCLUSIONS: Despite different methods (BMI vs. ICD-9 codes) for capturing obesity between two databases, prevalence rankings of comorbidities were similar. Obesity is therapeutically classified by BMI; however, most large claims databases capture obesity by ICD-9 codes. This may result in variation in weight group distributions and significant under-reporting of obesity claims in MarketScan. Our findings highlight the need for further research in optimal sources for data in obese patients.