This issue contains the latest recommendations for prophylaxis of HIV-related infections from the U.S. Public Health Service (USPHS) and the Infectious Diseases Society of America (IDSA) [2]. These differ from previous recommendations in several ways, notably for the prevention of P. carinii pneumonia, M. avium complex disease, and toxoplasmosis. The guidelines are useful because, for the first time, they include recommendations on ways to avoid exposure to opportunistic pathogens (an area often neglected by clinicians) and a review of standard recommendations for chemoprophylaxis. They reinforce the status of trimethoprim-sulfamethoxazole as first-line prophylaxis for P. carinii pneumonia, and, for the first time, they recommend that trimethoprim-sulfamethoxazole be used for prophylaxis of toxoplasmosis in patients with antibody evidence of previous exposure to Toxoplasma gondii. The recommendations also address intolerance to trimethoprim-sulfamethoxazole, suggesting that persons with a history of non-life-threatening reactions be rechallenged with the drug [3]. For truly intolerant patients, dapsone or dapsone-pyrimethamine is the preferred second-line regimen.

The results of the largest prospective clinical trial done to date [4] indicate that the superiority of trimethoprim-sulfamethoxazole to alternative regimens is maximized in patients with CD4 lymphocyte counts of less than 100 cells/µL. This has led to the suggestion that aerosolized pentamidine be used until the CD4 lymphocyte count decreases to less than 100 cells/µL, at which time trimethoprim-sulfamethoxazole therapy could be initiated [5]. Thus, trimethoprim-sulfamethoxazole could be reserved for those persons most likely to benefit from it. It should be noted that this strategy has never been prospectively studied, and it is unclear whether waiting for lower CD4 lymphocyte counts would, in fact, increase the tolerability of trimethoprim-sulfamethoxazole. Indeed, I would argue that it is more important to direct further efforts in preventing P. carinii pneumonia toward patients with the lowest CD4 lymphocyte counts; the strongest predictor of failure of this prophylaxis is a decreasing CD4 lymphocyte count. In a recent Multicenter AIDS Cohort study [6], 19% of patients receiving prophylaxis developed P. carinii pneumonia, and 76% of failures occurred in patients with CD4 lymphocyte counts less than 50 cells/µL [6]. A similar failure rate was seen in the recent large AIDS Clinical Trials Group study [4]. The risk for P. carinii pneumonia increases as the CD4 lymphocyte count decreases. A prophylactic regimen that is effective for a patient with a CD4 lymphocyte count of 150 cells/µL may be inadequate for a patient with a count of 15 cells/µL. Alternative, more aggressive strategies may be needed for this population and should be investigated.

The USPHS/IDSA recommendations are also interesting for what they do not include. Prophylaxis for M. avium complex disease was recommended (not strongly encouraged) for consideration, despite emerging evidence from one clinical trial showing that such prophylaxis can prolong survival [7]. Increasing recognition of the importance of M. avium complex as a cause of illness and death in patients with late-stage AIDS may change the strength of that recommendation. I personally would offer M. avium complex prophylaxis (rifabutin, clarithromycin, or azithromycin) to all susceptible patients. The guidelines for M. avium complex prophylaxis were changed slightly; the recommended CD4 lymphocyte threshold for the initiation of prophylaxis was reduced to 75 cells/µL on the basis of a reanalysis of the original trials that suggested that rifabutin is effective [8]. Despite clinical trial evidence showing that systemic fungal infections can be prevented with fluconazole [9], the working group cited concerns about resistance, drug-drug interactions, and cost as factors leading it not to make a standard recommendation about antifungal prophylaxis. Additionally, clinical research has not stood still since these recommendations were put together. The Food and Drug Administration recently approved oral ganciclovir as prophylaxis for cytomegalovirus infections in patients with advanced HIV disease. However, no clear consensus exists as to how ganciclovir should be used, and the data from trials do, in fact, conflict somewhat [10, 11].

These guidelines should therefore be regarded as only the first step in managing this aspect of advanced HIV disease. It is now clear that many cases of fungal, M. avium complex, and cytomegalovirus infections can be prevented. However, the routine use of oral ganciclovir and fluconazole, rifabutin, or clarithromycin is probably impractical for most patients. These regimens are associated with some risk for inducing drug-resistant pathogens; because these agents are often pivotal in treating infection, substantial levels of resistance could become a formidable problem in several years. Already, azole-resistant candidal infections are increasingly seen in patients with advanced HIV disease [12]. Such prophylactic regimens are also likely to be costly, and they pose increased risks for drug toxicity and negative effects on quality of life. Potentially important drug-drug interactions among these agents or with newer antiretroviral agents leading to new toxicities or problems with adherence have already been identified [13]. Thus, it is important to consider alternative strategies for preventing AIDS-related opportunistic infections. Agents with activity against more than one pathogen would be useful but are unlikely to arrive in the foreseeable future. The most promising approach is that of identifying subsets of patients with advanced HIV disease who are at increased risk for infection, so that early intervention can be targeted to those at greatest risk for a specific opportunistic infection. A better understanding of the pathogenesis and natural history of opportunistic infections would greatly help. For example, if larger prospective trials verify preliminary data suggesting that cytomegalovirus retinitis can be predicted using intensive virologic monitoring [14, 15], then it may be possible to use regular viral screening and give prophylaxis (or early intervention) only to patients with the greatest risk for disease. Recommendations about prophylaxis should be tailored to individual patients on the basis of an assessment of risks and benefits. Research efforts need to focus on methods that will allow clinicians and patients to make those decisions.

It should be remembered that, ultimately, prophylaxis is merely patchwork medicine. Truly successful preventive strategies require a reversal of immunodeficiency and the elimination of the risk for opportunistic infections, either with antiretroviral chemotherapy or immunotherapy, or both. This needs to remain the primary goal of clinical research in AIDS.