We agree with Drs. Pinkerton and Holtgrave that using conventional-criteria postexposure prophylaxis is cost-effective only for sexual exposures that put persons at highest risk for seroconversion. We also agree that other forms of prevention are likely to be more cost-effective than postexposure prophylaxis. However, clinicians must always treat the patients in front of them. A patient with a recent sexual exposure to HIV has failed primary prevention. Our hope is that the intensive counseling that should accompany postexposure treatment will prevent future exposures. From this perspective, postexposure prophylaxis may be viewed as a method of motivating persons at highest risk for HIV seroconversion to initiate care; it may then be possible to engage them in more cost-effective methods of HIV prevention.

Dr. Behrman well characterizes the strengths and weaknesses of providing postexposure care in emergency departments. In the case–control study demonstrating the efficacy of zidovudine after occupational exposure to HIV, patients were treated within a mean of 4 hours of exposure [1]. Although it is unknown whether postexposure prophylaxis ceases to be effective at some point, the longer the delay after exposure, the less likely postexposure prophylaxis is to work. Given this and the fact that young, sexually active persons without HIV infection do not necessarily have regular medical providers, many exposed persons will seek care in emergency departments. However, as Dr. Behrman points out, emergency departments are challenging settings for providing counseling and follow-up care. We support emergency departments having "starter packs" with defined protocols for providing HIV testing and initiation of treatment. Ideally, patients would then be referred for counseling and follow-up care to a primary care or public health setting.

Dr. Frothingham reminds us that more is not always better when it comes to medications. We do not favor the routine use of protease inhibitors because of the additional side effects and costs (which would decrease the cost-effectiveness of postexposure prophylaxis). Adding a protease inhibitor increases the complexity of the regimen. Regimens with protease inhibitors require more pills, involve more complicated dosing schedules (for example, indinivir requires an empty stomach), and have more potential drug interactions [2].

Until more empirical research is available, clinicians must base their decisions about the indications and regimens for postexposure treatment of sexual exposures to HIV on available data from the treatment of occupational exposures to HIV [3-5] and on clinical judgment.