Katz and Gerberding [1] address the timely subject of prophylaxis after sexual exposure to HIV. Their standard protocol consists of two nucleoside analogues administered for 28 days. A protease inhibitor is optional in some circumstances, although the authors note that some experts routinely use three drugs. Katz and Gerberding discuss short-term but not long-term adverse drug effects. Nucleoside analogues have been used for more than a decade, but few data are available on the long-term safety of protease inhibitors.

In June 1997, the U.S. Food and Drug Administration (FDA) reported 83 cases of diabetes or hyperglycemia associated with protease inhibitors, 27 of which required hospitalization and 6 of which were life-threatening [2]. The average onset was 76 days after initiation of therapy, but cases occurred as early as 4 days. When protease inhibitor therapy was discontinued, many patients had resolution of hyperglycemia, but diabetes persisted in some patients. For patients with HIV infection, the FDA considered the benefits of protease inhibitors to outweigh the risks.

There is a plausible rationale for adding protease inhibitors to postexposure regimens, but no current evidence supports a benefit [1]. The rate of HIV infection per exposure is low under most circumstances. Even a low rate of diabetes could dramatically influence the risk–benefit ratio for protease inhibitors in this setting. Given the degree of uncertainty, physicians may choose to "do no harm." Patients who receive protease inhibitors for any indication should be informed of the possible association with diabetes. They should know the warning signs of hyperglycemia: increased thirst and hunger, fatigue, unexplained weight loss, and increased urination [2].