Background A quarter of adults in the UK are obese, and 2 to 3% are morbidly obese. Evidence from small clinical trials and cohort studies based in centres of excellence shows that bariatric surgery for obesity may contribute to weight loss and better health outcomes in obese patients. However, there is scant follow-up on the majority of patients who undergo bariatric surgery. The aim of this study was to evaluate health outcomes in a population-based cohort using primary care electronic health records.

Methods A cohort of patients aged 20–100 years with a record for gastric banding, gastric bypass or sleeve gastrectomy was selected from the UK Clinical Practice Research Database (CPRD) from 2002 to 2012. Changes in weight, body mass index (BMI) and cardiovascular risk factors after surgery were evaluated. A control group of patients with a similar BMI profile who did not have surgery were also selected for comparison.

Results Of the 2839 participants identified, 2265 (80%) were women. There were 229 (8%) followed up for weight at 12 months, 305 (11%) at 2 years and 335 (12%) at 5 years after operation. The mean BMI before surgery was 45.3 kg/m² in men and 43.5 kg/m² in women. The mean reduction in body weight at one year was 33.5 (28.1 to 38.8) kg in men and 25.2 (22.1 to 28.3) kg in women, with a reduction in BMI of 10.9 and 9.4 kg/m² respectively. Systolic blood pressure was reduced by 7.9 (6.0 to 9.9) mm Hg in men and 6.2 (5.1 to 7.2) mm Hg in women. In the control patients a significant reduction in weight of -1.9 (-1.1 to -2.7) kg in men and -0.7 (-0.5 to -1.0) kg in women was observed. Reductions in body weight and blood pressure were maintained in bariatric surgery patients up to 5 years after surgery.

Conclusion In a population-based cohort, bariatric surgery was found to be associated with substantial reductions in body weight, body mass index and systolic blood pressure that were maintained for up to five years after operation. Only a minority of participants had measurements during follow-up contributing to increased risk of bias.