Dear Editor, Yeow et al. 1 describe a case of non-diabetic euglycaemic acidosis resulting from post op dysphagia and poor intake of approximately six weeks duration. We have seen a similar case of 'starvation ketoacidosis' in a patient undergoing percutaneous endoscopic gastrostomy feeding tube replacement. We think that intravenous (IV) glucose should be the initial treatment, with the addition of insulin only if required. The correspondence from Frise and Mackillop 2 states this strategy is effective for treating ketoacidosis in pregnancy; however, there are also some other pitfalls in treatment of starvation ketoacidosis which must be considered, and some overlap with alcoholic ketoacidosis. Unlike patients with diabetic ketoacidosis, patients with starvation ketosis release insulin when carbohydrate is administered. They are also producing high levels of glucose elevating hormones such as glucagon and have depleted glycogen stores. These hormones cause the lipolysis which helps generate ketones for fuel. The addition of exogenous insulin in this state risks hypoglycaemia. Once provided with adequate carbohydrate the insulin levels will rise and counterregulatory hormone levels will fall, resolving the ketosis. Alcoholics are another group prone to ketosis (alcoholic ketoacidosis (AKA)) and are particularly prone to hypoglycaemia; administration of insulin to those patients would have to be with caution and literature 3 on AKA reports resolution without insulin administration, although there is little evidence outside of case reports.