BACKGROUND Higher mortality rates among morbidly obese (BMI of > or =40 or > or =35 kg/m2 with weight-related comorbidities) subjects are mainly explained by comorbidities such as type 2 diabetes. As bariatric surgery ameliorates diabetes, obese diabetic subjects will receive great benefits from bariatric surgery. Screening for diabetes prior to surgical referral is therefore crucial. METHODS We studied 1,253 consecutively recruited (2005-2008) morbidly obese subjects (67% women). Among subjects without known diabetes, 70% (670/961) performed an oral glucose tolerance test (OGTT). Screen-detected diabetes was defined as fasting plasma glucose (fPG) > or =7.0 mmol/l and/or 2-h glucose concentration (2hPG) > or = 11.1 mmol/l. RESULTS Within the study population, 31% had diabetes, of which 8% were screen-detected. Eighty percent of those with screen-detected diabetes were diagnosed by fPG. In subjects with nondiabetic fPG concentrations, elevating the fPG cutoff value from 5.2 mmol/l to the World Health Organization's (WHO's) recommended value of 6.1 mmol/l reduced the percentage of the population needing an OGTT considerably (78-23%), but only slightly reduced the sensitivity of fPG in detecting a diabetic 2hPG concentration (100-77%). Only 7% of the patients with fPG between 6.1 and 6.9 mmol/l had a diabetic 2hPG concentration. Following the WHO's recommendations, we found that 95% of all subjects with unknown diabetes were identified. CONCLUSIONS Fasting glucose identified four out of five morbidly obese subjects with unknown diabetes. A supplemental OGTT in selected persons identified the majority of the remaining diabetic cases.