Background: Results from recent studies have raised the controversy of Obesity Paradox in patients with type 2 diabetes (T2DM). While some studies reported significantly higher mortality risk in patients with normal body weight at diagnosis of diabetes, compared to those with obesity, a few studies could not find such evidence.

Aims:  The aims of this real-world primary care based observational study were to explore the co-morbidities at diagnosis of T2DM, and the association of body mass index (BMI) at diagnosis with all-cause mortality.

Methods A cohort of 192,079 T2DM patients with diagnosis from January 2000 were extracted from UK primary care system (The Health Improvement Network database) under the conditions: complete data on age at diagnosis (18-70 years), sex, BMI at diagnosis (≥ 15 kg/m² at diagnosis), dates of first records of cardiovascular diseases (CVD), renal diseases and cancer.  

Results: The patients were mean 54 years old, 57% male, 33% current or ex-smokers, 16 / 5 / 8% had CVD / renal disease / cancer. At diagnosis, 0.2% were underweight, 7% had normal weight and 38 / 16 / 11% had Grade 1 / Grade 2 / Grade 3 obesity. During 7 years of median follow-up, 10 / 13 / 17% developed CVD / renal disease / cancer, and 8% died. Among patients without disease history at diagnosis, compared to patients with Grade 1 obesity, the adjusted mortality risk were significantly higher in underweight, normal weight and Grade 3 obese patients by 200%, 10% and 30% respectively.

Conclusion: At least 29% and 40% T2DM patients had comorbidities at diagnosis and post-diagnosis respectively. After accounting for time-varying prevalence of co-morbidities, the underweight and normal weight patients at diagnosis had significantly higher mortality risk compared to patients with BMI 30-35 kg/m² at diagnosis.