Ketoacidosis during a low-carbohydrate diet.

to the editor: It is believed that low-carbohydrate diets work best in reducing weight when producing ketosis.1 We report on a 51-year-old white woman who does not have diabetes but had ketoacidosis while consuming a “no-carbohydrate” diet. There was no family history of diabetes, and she was not currently taking any medications. While adhering to a regimen of carbohydrate restriction, she reached a stable weight of 59.1 kg, a decrease from 72.7 kg. After several months of stable weight, she was admitted to the hospital four times with vomiting but without abdominal pain. On each occasion, she reported no alcohol use. Her body-mass index (the weight in kilograms divided by the square of the height in meters) was 26.7 before the weight loss and 21.7 afterward. Laboratory evaluation showed anion-gap acidosis, ketonuria, and elevated plasma glucose concentrations on three of the four occasions (Table 1). She had normal concentrations of plasma lactate and glycosylated hemoglobin. Screening for drugs, including ethyl alcohol and ethylene glycol, was negative. Abdominal ultrasonography showed hepatic steatosis. On each occasion, the patient recovered after administration of intravenous fluids and insulin, was prescribed insulin injections on discharge, and gradually reduced the use of insulin and then discontinued it while remaining euglycemic for six months or more between episodes. Testing for antibodies against glutamic acid decarboxylase and antinuclear antibodies was negative. Values on lipid studies were as follows: serum triglycerides, 102 mg per deciliter; high-density lipoprotein (HDL) cholesterol, 50 mg per deciliter; and calculated low-density lipoprotein (LDL) cholesterol, 189 mg per deciliter. The patient strictly adhered to a low-carbohydrate diet for four years, with an estimated carbohydrate intake that was often less than 20 g per day. When she was put on a diet containing normal amounts of carbohydrates, her fasting plasma glucose concentration and the results of oral glucose-tolerance tests were normal. With a normal carbohydrate intake, she had no more episodes of ketoacidosis. Citrate generated by glycolysis inhibits carnitine palmitoyltransferase complex I, limiting the beta-oxidation of fatty acids and thereby reducing ketogenesis. Lactate, which increases during starvation, can induce hepatic ketogenesis.2 Lowcarbohydrate, fat-rich meals can enhance alphacell secretion of glucagon and lower insulin concentrations.3,4 Plasma fatty acid concentrations can be twice as high during low-carbohydrate diets as compared with the usual carbohydrate intake in the postabsorptive period.5 Increased concentrations of free fatty acids in the absence of carbohydrate-induced inhibition of beta-oxidation of fatty acids and in the presence of an abnormally high ratio of glucagon to insulin and elevated concentrations of lactate may have caused ketoacidosis in our patient, who was trying to avoid all dietary carbohydrates. Low-carbohydrate, high-fat diets are generally associated with higher