Severe Diabetes Associated with Protease Inhibitor Therapy

TO THE EDITOR:

Protease inhibitors block the action of aspartate protease, which is required to produce new virions. We describe a patient in whom symptomatic hyperglycemia was associated with the protease inhibitor nelfinavir, perhaps as a result of inhibition of the protease that converts proinsulin to insulin.

A 46-year-old man with advanced-stage AIDS was receiving didanosine and lamuvidine. His viral load was 380 000 RNA copies/mL, and his CD4⁺ count was 20 cells/mm³. Treatment was changed to nelfinavir, zidovudine, and lamivudine. Other medications were trimethoprim-sulfamethoxazole, ciprofloxacin, clarithromycin, ethambutol, itraconazole, megestrol acetate, phenytoin, amitryptyline, and carbamazepine. The patient's mother had type 2 diabetes, but the patient was repeatedly euglycemic (Figure 1).

View larger version:

• In this window
• In a new window

• Download as PowerPoint Slide

Figure 1. Each symbol indicates a serum glucose evaluation. Duration of megestrol acetate therapy is shown. Insulin therapy (40 U/d) was initiated on 19 June 1997 and was maintained at 45 U/d. Serum glucose levels in a patient with AIDS before and after therapy with the protease inhibitor nelfinavir began.

Two weeks after starting nelfinavir therapy, the patient noticed polyuria, blurred vision, oral thrush, dysphagia, and fever. On physical examination, he was thin, weak, and febrile with marked oral thrush. His lung, heart, and abdomen were normal. His leukocyte count was 2200 cells/mm³. The serum glucose level was 657 mg/dL, and the sodium level was 123 mEq/L; results of other laboratory tests were normal. No enzymatic or radiologic evidence showed pancreatitis. All blood cultures were negative, as was the result of an ophthalmoscopic examination for cytomegalovirus. Glucose levels were controlled with neutral protamine Hagedorn insulin (40 U/d), vision improved, and oral thrush responded to fluconazole.

In this case, treatment with a protease inhibitor was rapidly followed by new-onset, symptomatic diabetes mellitus. Megestrol acetate [1] was not thought to be responsible because hyperglycemia began with initiation of nelfinavir therapy and continued for 2 months after megestrol acetate therapy had been discontinued. No pancreatitis was present. Before initiation of nelfinavir therapy, the patient was euglycemic; however, a familial predisposition may have played a role.

It is estimated that more than 100 000 patients are taking protease inhibitors. The U.S. Food and Drug Administration has received 83 reports of new-onset or worsening diabetes (ranging from mild hyperglycemia to diabetic ketoacidosis) associated with each of the four protease inhibitors used to treat HIV infection [2].

The mechanism by which protease inhibitors cause diabetes is unknown. The proteolytic processing of prohormones within neuroendocrine cells is needed to generate biologically active peptides [3-5]. Three families of proteases, including an aspartate protease, have been localized to the chromaffin granules that process prohormones [3]. Human processing of proinsulin to cleave C-peptide requires serine endopeptidases PC1/PC3 and PC2 [3, 4]. Genetic deficiency of endopeptidases has recently been implicated as a cause of hyperglycemia [4]. In patients who are functionally deficient in these endopeptidases, the mammalian homologue of yeast aspartic protease (YAP3p) may play a more critical role in processing proinsulin and other prohormones [5]; as a result, it may be subject to inhibition by nelfinavir, an aspartic protease, inhibitor. The mechanism of this effect is under further study. Since this case, we have documented new-onset diabetes in two more patients, both of whom developed diabetes while receiving indinavir.

• Copyright ©2004 by the American College of Physicians