Is Antiretroviral Treatment after Percutaneous HIV Exposure Justified?

Placebo-controlled Randomized Trial

In 1987, Burroughs Wellcome designed a large double-blind, placebo-controlled, randomized trial and consulted with experts in occupational HIV risk. In 1988, after establishing the national referral system [2], the trial was instituted. Because it failed to enroll enough participants, it was discontinued a year later [3]. Critics blamed this failure on clinicians who prematurely advocated routine use of zidovudine. However, this is an oversimplification and does not consider many other issues that may have a bearing on the conduct of similar studies.

First, when the trial was initiated, the impact of HIV infection risk on health care workers was not fully appreciated and motivation to offer zidovudine may have been lower. Second, because few employee health physicians had managed occupational HIV exposures, efficient systems to facilitate timely evaluation, referral, and enrollment were often lacking. Third, study participation necessitated detailed and burdensome case reporting and follow-up documentation by providers, many of whom had no previous involvement or experience in clinical trials research. Furthermore, the initially alarming reports of zidovudine toxicity in patients with advanced HIV infection may have dissuaded some physicians from advocating its use.

Despite more than 2000 inquiries, only 122 participants were randomized, and only 84 received at least one dose of study drug. Those participants who perceived a low risk for infection may not have been willing to risk the potential toxicity of zidovudine (1200 mg/d for 6 weeks). Those who perceived a high risk for infection and potential benefit from treatment may have obtained the drug elsewhere and not risked the 50% chance of randomization to the placebo group. The importance of participating in the randomized trial was not publicly emphasized by opinion leaders. The inability to do this important study motivated infectious disease experts who were concerned with occupational exposure to HIV to lobby for research related to HIV chemoprophylaxis for health care personnel. Discussions focused on the difficulties inherent in conducting a clinical trial with sufficient power to evaluate drug efficacy and on the requirements for labeling a study drug for chemoprophylaxis. The feasibility of using existing AIDS clinical trial sites where retroviral research expertise and exposure incidence were high was also considered.

Multicenter Open Study

Ultimately, the clinical trial was not funded and investigators obtained funding from the manufacturer to support a multicentered open study of zidovudine toxicity. In 1991, a consensus treatment protocol was implemented in 21 participating medical institutions, which is now used by more than 300 other medical centers [4]. The trial was designed to provide useful information about safety, tolerance, cost, and the feasibility of timely administration after exposure. Moreover, it was hoped that the experience gained in developing collaborative relationships with employee health clinicians might facilitate implementation of efficacy trials in the future.

Preliminary study data suggest that prompt initiation of postexposure treatment is possible. More than 80% of the first 180 participants enrolled received zidovudine within 3 hours after percutaneous exposure [5]. To date, no hematologic or other objective toxicity severe enough to warrant dose reduction has been observed. The study has shown that a successful clinical trial can be done by clinicians with little previous research experience.

Additional Studies Are Needed To Determine Drug Efficacy

To make intelligent decisions about the necessity of postexposure treatment, we need more scientific information. A coordinated effort involving experts in occupational infection, study design, retrovirology, animal models, pharmacology, and medical ethics will be required. Suitable animal models are necessary to compare the relative efficacy of candidate prophylactic drugs or drug combinations. A murine model, using mice with severe combined immunodeficiency that have been immunologically reconstituted with human lymphoid tissue (SCID-hu mice), may serve as an appropriate screening model for at least some categories of antiretroviral drugs [6]. The efficacy and toxicity of candidate regimens could then be studied in primates exposed to simian immunodeficiency virus or HIV-2. Studies in infant rhesus macaques have provided preliminary data supporting the efficacy of oral zidovudine when administered 2 hours before exposure to a small inoculum of simian immunodeficiency virus [7]. This model more closely approximates the conditions of occupational exposure (low inoculum) than do earlier primate models of prophylaxis using reverse transcriptase inhibitors that employed high titers of infecting virus. This simian model may be useful in determining the efficacy of prophylactic treatment after low inoculum, transcutaneous, or transmucosal exposures. Even if effective prophylaxis were documented in a primate model, doing a clinical efficacy trial will still be difficult given the need for a sample size of more than 2000 in each arm. The study size could become more manageable if a subset of persons could be defined who had an infection risk statistically greater than the average reported risk (0.2% to 0.4%). Use of a placebo arm could still deter some potential participants, especially after greater risk exposures. Alternate study designs include trials comparing zidovudine to a new agent or zidovudine alone compared with zidovudine plus a new agent as well as trials using unequal randomization schemes to increase the probability that a participant would be assigned to the treatment group.

Counseling Exposed Health Care Workers

The decision to use zidovudine after exposure is very personal. Health care workers should be educated about the pros and cons before receiving treatment. Those considering treatment with zidovudine should be fully informed about its unproven value in this setting. The limited efficacy data from animal models should be interpreted objectively without assuming their relevance to occupational HIV exposures.

Counseling should include information about the relatively high frequency of nausea, vomiting, headache, and fatigue associated with zidovudine treatment. Patients should also be told about the possibility of hematologic toxicity and about the need for clinical and laboratory monitoring at least every 2 weeks while receiving therapy. Additional theoretical concerns include the potential for zidovudine treatment to result in delayed seroconversion, in selection for resistant virus, and in late toxic reactions. Patients should also be reassured that the decision to start treatment does not mean it must be continued.

Until a scientific basis for treatment is established, the use of zidovudine prophylaxis (or any other antiretroviral agent) should not be considered a standard of care. Treatment is justified only as an experimental option for informed employees when resources and expertise to monitor therapy are available [8]. Fortunately, most exposed health care providers can be reassured that the statistical risk for infection is low, regardless of their decision about receiving antiretroviral agents.