Metformin in an HIV-infected patient with protease inhibitor-induced diabetic ketoacidosis.

OBJECTIVE

To describe a case of diabetes mellitus and diabetic ketoacidosis in a patient receiving protease inhibitor therapy and to describe the patient's response to treatment with metformin.

METHODS

A 49-year-old HIV-positive white man who was receiving indinavir, stavudine, and lamivudine for more than two years presented with shortness of breath and significant weight loss over the previous month. On admission, he had a pH of 7.11 and PaCO2 of 12.9 mm Hg. Laboratory investigations revealed glucose 420 mg/dL, a total carbon dioxide 5 mEq/L, and anion gap of 32. Beta-hydroxybutyrate was 5.9 mmol/L (normal value <0.4 mmol/L). Urine was highly positive for glucose and ketones. The patient was given intravenous fluids and an insulin infusion was started. Five days later, he was discharged on 60 units of insulin per day. Following discharge, efavirenz was substituted for indinavir. Metformin was added and six months following discharge the patient's blood glucose was well controlled with 36 units of insulin per day.

CONCLUSIONS

New-onset diabetes mellitus has been reported in HIV-infected patients receiving protease inhibitors. To date, diabetic ketoacidosis has been an infrequent acute complication. The mechanism by which protease inhibitors cause diabetes is unclear; however, studies have noted insulin resistance and increased proinsulin. Metformin increases the sensitivity of peripheral tissues to insulin and appeared to be useful in this patient. However, further clinical research is needed.

CONCLUSIONS

Monitoring glucose concentrations in HIV-positive patients receiving protease inhibitors is important to prevent the development of acute complications, including diabetic ketoacidosis. We recommend that these patients have their fasting serum glucose concentration measured at baseline, with follow-up every three months. The role of metformin and the thiazolidinedione antidiabetic agents in the management of protease inhibitor-induced diabetes requires further study.