Diabetes mellitus is defined by chronic elevation of blood glucose linked to insulin resistance and/or insulinopaenia. Its diagnosis is based on a fasting blood-glucose level of ≥ 1.26 g/L or, in some countries, a blood glycated haemoglobin (HbA1c) level of > 6.5%. Of the several forms of diabetes, type-2 diabetes (T2D) is the most common and is found in patients with other risk factors. In contrast, type-1 diabetes (T1D) is linked to the autoimmune destruction of β-pancreatic cells, leading to insulinopaenia. Insulin deficiency results in diabetic ketoacidosis within a few hours. ‘Pancreatic’ diabetes develops from certain pancreatic diseases and may culminate in insulinopaenia. Treatments for T2D include non-insulin based therapies and insulin when other therapies are no longer able to control glycaemic levels. For T1D, treatment depends on long (slow)-acting insulin and ultra-rapid analogues of insulin administered according to a ‘basal-bolus’ scheme or by continuous subcutaneous delivery of insulin using a pump. For patients presenting with previously undiagnosed dysglycaemia, investigations should determine whether the condition corresponds to pre-existing dysglycaemia or to stress hyperglycaemia. The latter is defined as transient hyperglycaemia in a previously non-diabetic patient that presents with an acute illness or undergoes an invasive procedure. Its severity depends on the type of surgery, the aggressiveness of the procedure and its duration. Stress hyperglycaemia may lead to peripheral insulin resistance and is an independent prognostic factor for morbidity and mortality.