AIM: To assess glycaemic control and factors associated with poor glycaemic control at initiation of second-line therapy in the DISCOVER programme. MATERIALS AND METHODS: DISCOVER (NCT02322762 and NCT02226822) comprises two similar prospective observational studies of 15 992 people with type 2 diabetes (T2D) initiating second-line glucose-lowering therapy in 38 countries across six regions (Africa, Americas, South-East Asia, Eastern Mediterranean, Europe and Western Pacific). Data were collected using a standardized case report form. Glycated haemoglobin (HbA1c) levels were measured according to standard clinical practice in each country, and factors associated with poor glycaemic control (HbA1c >8.0%) were evaluated using hierarchical regression models. RESULTS: HbA1c levels were available for 80.9% of patients (across-region range [ARR] 57.5%-97.5%); 92.2% (ARR 59.2%-99.1%) of patients had either HbA1c or fasting plasma glucose levels available. The mean HbA1c was 8.3% (ARR 7.9%-8.7%). In total, 26.7% of patients had an HbA1c level ≥9.0%, with the highest proportions in South-East Asia (35.6%). Factors associated with having HbA1c >8.0% at initiation of second-line therapy included low education level, low country income, and longer time since T2D diagnosis. CONCLUSIONS: The poor levels of glycaemic control at initiation of second-line therapy suggest that intensification of glucose-lowering treatment is delayed in many patients with T2D. In some countries, HbA1c levels are not routinely measured. These findings highlight an urgent need for interventions to improve monitoring and management of glycaemic control worldwide, particularly in lower-middle- and upper-middle-income countries.