The Body: The Complete HIV/AIDS Resource
Follow Us Follow Us on Facebook Follow Us on Twitter Download Our App 
Professionals >> Visit The Body PROThe Body en Espanol
  • Email Email
  • Printable Single-Page Print-Friendly
  • Glossary Glossary

Public Health Service Guidelines for the Management of Possible Sexual, Injecting-Drug-Use, or Other Nonoccupational Exposure to HIV, Including Considerations Related to Antiretroviral Therapy

September 25, 1998

Considerations for Using Antiretroviral Agents

Decisions to provide antiretroviral agents to persons after possible nonoccupational HIV exposure to prevent the establishment of HIV infection must balance the potential benefits and risks. Factors influencing the potential efficacy of this intervention include the probability that the source contact is HIV-infected, the likelihood of transmission by the particular exposure, the interval between exposure and initiation of therapy, the efficacy of the drug(s) used to prevent infection, and the patient's adherence to the drug(s) prescribed.

Possible Benefits

The major potential benefit of antiretroviral postexposure prophylaxis is reducing a person's risk for acquiring HIV infection after exposure. Estimates differ for the transmission risk (if the person is not treated) after specific HIV exposures and the possible effect of early treatment after such exposures.


Probability of Transmission from One HIV Exposure

HIV can be transmitted efficiently through blood transfusions: an estimated 95% of recipients become infected from transfusion of a single unit of infected whole blood. The per-contact probability of transmission from an HIV-infected source is much lower for injecting-drug-use and sexual exposures. The risk for HIV transmission per episode of intravenous needle or syringe exposure is estimated at 0.67%.(5) The risk per episode of percutaneous exposure (e.g., a needlestick) to HIV-infected blood is estimated at 0.4% (upper limit of 95% confidence interval [CI] = 0.8%).(6) The risk for HIV transmission per episode of receptive penile-anal sexual exposure is estimated at 0.1%-3%; the risk per episode of receptive vaginal exposure is estimated at 0.1%-0.2%.(7) No published estimates of the risk for transmission from receptive oral exposure exist, but instances of transmission have been reported.(8,9)

Pathogenesis of Early HIV Infection

Information about the initial physiologic events after HIV exposure suggests that it can take several days for infection to become established in the lymphoid and other tissues. During this time, interventions to interrupt viral replication could represent an opportunity to prevent an exposure from becoming an established infection.(10,11)

Studies of Antiretroviral Agent Use to Prevent HIV Infection in Animals

Attempts to protect animals with antiretroviral monotherapy after experimental mucosal and intravenous (IV) exposures have produced various results.(12) In studies to assess the efficacy of zidovudine (ZDV) administered after IV exposure to simian immunodeficiency virus (SIV), the suppression or delay of viral replication was common but the prevention of infection was rare.(13,14) Treatment initiated within 24 hours of exposure and continued for 28 days appeared to have a greater effect than treatment initiated 72 hours after exposure. However, ZDV might not be the optimal agent to demonstrate proof-of-concept because it has not demonstrated potent inhibitory activity against SIV infection in macaques, even when treatment is initiated before viral exposure.(12,15) In another study in which a licensed antiretroviral drug was administered to macaques, initiating stavudine (d4T) treatment at the time of IV exposure to human immunodeficiency virus type 2 (HIV-2) resulted in a delay in the onset of viremia and a reduction in viral load.(16) Although protection from infection was not observed, most of the treated animals exhibited sustained control of viral replication and normal CD4+ cell levels for >1 year after receiving 16 weeks of drug treatment. More compelling evidence of the efficacy of antiretroviral drugs for postexposure prophylaxis in animal models has been generated by using unlicensed compounds. SIV infection was prevented in 100% of macaques when treatment in phase I/II clinical trials with (R)-9-(2-phosphonylmethoxypropyl)adenine (PMPA), a nucleotide analogue, was started either 4 or 24 hours after IV inoculation and was continued for 28 days.(17) Protection was diminished if treatment was delayed >24 hours or if the treatment duration was reduced.(18) In another recent study, mucosal or IV infection of macaques with SIV was blocked when a 3-day treatment with the nucleoside analogue 2',3'-dideoxy-3'-hydroxymethyl cytidine (BEA-005) was initiated within 8 hours of viral exposure.(19) Further corroborating the PMPA results, the BEA-005 study demonstrated that increasing the time between exposure and treatment initiation or decreasing the duration of treatment reduced protection.

The animal-model data demonstrate that antiretroviral agents administered after SIV or HIV-2 exposure can prevent infection. However, extrapolating these results to humans is problematic because of several factors, including differences between a) the laboratory-adapted strains of SIV and HIV used in animal studies and the HIV strains that circulate among persons; b) inoculum size; c) routes of inoculation or exposure; d) time of treatment initiation; e) drug(s) used; f) treatment duration; and g) host metabolism, host immunology, and other biological parameters. Animal studies offer proof-of-concept and demonstrate the challenges to understanding the requirements for effective use of antiretrovirals to prevent HIV transmission in humans.

Studies of Antiretroviral Agent Use to Prevent HIV Infection in Humans

In 1995, investigators used case-control surveillance data from health-care workers in France, Italy, the United Kingdom, and the United States to document that ZDV use was associated with an 81% (95% CI = 48%-94%) decrease in the risk for HIV infection after percutaneous exposure to HIV-infected blood.(1,2,20,21) This study was a retrospective case-control study, rather than a prospective trial, which is the preferred method of assessing clinical drug efficacy. Additional limitations were that a) the number of case-patients was small, b) the case-patients and controls came from separate populations, c) some case-patients were reported anecdotally before formal surveillance was established, and d) some details of exposures in case-patients were obtained retrospectively, whereas information for controls was collected prospectively. Although the health-care worker study demonstrated antiretroviral effectiveness following percutaneous HIV exposure, some researchers have suggested that the magnitude of the effect might be overestimated because of the methodologic questions raised.(22) ZDV has failed to prevent HIV infection in health-care workers in 13 reported instances.(23)

In a prospective, randomized controlled trial of ZDV administered to HIV-infected women during pregnancy and labor and to their infants for 6 weeks postpartum, perinatal transmission was reduced 67% among those randomly assigned to the treatment group compared with those in the control group, who received no antiretroviral therapy. Results of multivariate analyses suggested that a prophylactic effect on the fetus during antenatal, intrapartum, or postpartum exposure(24) could account for some reduction in perinatal transmission. In a prospective trial of ZDV in Thailand, perinatal HIV transmission was reduced 51% for women treated from 36 weeks' gestation until delivery.(25) Perinatal transmission despite use of ZDV prophylaxis in pregnancy also has been reported.(26)

Although these studies suggest that antiretroviral agents are potentially valuable for treating HIV exposures in these settings, the data might not be directly relevant to nonoccupational exposures. Health-care workers often are exposed to HIV in settings where antiretroviral therapy can begin within 1-2 hours of exposure and where the HIV status of the source patient usually can be determined quickly. These circumstances are unlikely for many nonoccupational exposures. The perinatal transmission model also might not be directly relevant to nonoccupational exposures. If most perinatal infections occur at the time of delivery, the observed effectiveness of ZDV therapy could represent a preexposure not a postexposure effect. Despite the apparent usefulness of antiretroviral agents in perinatal and occupational settings, it is unclear whether these findings can be extrapolated to other settings. Further studies are needed before one can conclude whether using antiretroviral agents to prevent HIV infection after nonoccupational exposures is effective.

Possible Risks

Potential risks of antiretroviral postexposure prophylaxis include drug toxicity, reduced effectiveness of behavioral HIV-prevention measures, and the acquisition of antiretroviral-resistant HIV strains. Also, the cost of medications could tax already scarce public funds for antiretroviral agents for HIV-infected persons, which offer cost-effectiveness and therapeutic benefit. Many insurers will not cover the cost of this unproven therapy, so any possible benefit will be limited based on the patient's ability to pay.

Side Effects and Toxicity of Antiretroviral Agents

The frequency, severity, duration, and reversibility of side effects must be weighed against the usefulness of antiretroviral agents for any patient. All antiretroviral agents have been associated with side effects. Adverse events have been reported for persons with advanced HIV disease (and longer treatment courses), but persons with less advanced disease or those who are uninfected might have different experiences.(27) Many side effects can be managed symptomatically, but when the probability of transmission is low, one must weigh this probability against the risk of a severe side effect. Although the most common side effects are mild, studies have demonstrated that 50%-75% of health-care workers receiving ZDV alone for possible HIV exposure reported one or more subjective complaints, and as many as 35% did not complete the full course of therapy because of side effects.(6,28,29) Preliminary information on health-care workers receiving combination therapy for postexposure prophylaxis demonstrated that 50%-90% reported subjective side effects and 24%-36% reported side effects severe enough to discontinue therapy.(30-33)

Many antiretroviral agents are associated with gastrointestinal side effects (e.g., nausea, vomiting, and diarrhea). In general, using combinations of agents has not caused more instances of adverse effects, but serious drug interactions have occurred when antiretrovirals were used with certain other medications. Current medications must be evaluated before patients are prescribed any antiretrovirals, and health-care providers must monitor patients closely for toxicities. Protease inhibitors recently have been associated with the occurrence of lipid abnormalities,(34-36) as well as the development of diabetes mellitus, hyperglycemia, and diabetic ketoacidosis, and they can exacerbate preexisting diabetes mellitus.(27,37) Some health-care workers using combination drugs for postexposure prophylaxis of occupational HIV exposure have developed serious side effects -- including nephrolithiasis, hepatitis, and pancytopenia -- sometimes within 3 days of initiating therapy.(31,32,38)

Behavior Changes Potentially Related to Prophylactic Antiretroviral Therapy

Some persons actively seek and repeatedly participate in high-risk behaviors (e.g., unprotected sex or needle-sharing injecting-drug use). The widespread availability of antiretroviral agents for treating possible nonoccupational HIV exposure could undermine public health efforts aimed at increasing and maintaining sexual and injecting-drug-use behaviors that prevent HIV exposure. If persons perceive that postexposure antiretroviral prophylaxis prevents HIV infection, they could increase the frequency of risk behaviors or shift from lower-risk to higher-risk activities. If many persons increase higher-risk behaviors, the widespread availability of antiretroviral agents for treating HIV exposure paradoxically could increase the number of new infections because the treatment's effectiveness will be <100%. One study of 54 men who had sex with men (recruited as part of an intervention counseling study) documented that 15% already had taken "a chance of getting infected when having sex" because of the availability of new treatments. It is unclear whether this reported behavior was a response to the existence of antiretroviral postexposure prophylaxis or a decreased fear of HIV disease because of the effectiveness of combination antiretroviral therapy.(39)

Acquiring Antiretroviral-Resistant Virus

The use of antiretroviral agents after possible nonoccupational HIV exposure, particularly if a patient does not adhere to the prescribed drug treatment, poses the theoretical risk that the patient could become infected with an antiretroviral-resistant strain of HIV if postexposure prophylaxis fails to prevent infection. In nonoccupational exposures, information regarding the antiretroviral-susceptibility patterns of the source virus likely will not be known, making it difficult to tailor antiretroviral therapy appropriately.

Cost of Antiretroviral Postexposure Prophylaxis

A 28-day course of antiretroviral agents for a single possible exposure to HIV costs an estimated $800 (range: $600-$1,000), depending on the agents used.(4,40) This cost generally is more per client than the cost of enrollment in intensive, behavioral, HIV primary prevention programs designed to reduce the likelihood of future exposures.(41) If postexposure prophylaxis proves effective in reducing HIV transmission after nonoccupational exposure, its cost dictates that use be restricted to high-risk exposures to avoid compromising funds for more cost-effective behavior intervention programs.

Uncertainties about key factors make it difficult to estimate the cost-effectiveness of treating nonoccupational HIV exposure with antiretroviral drugs. However, recent studies have used mathematical modeling to estimate cost-effectiveness ratios for this treatment.(42,43) These studies demonstrate that antiretroviral drugs could be cost-effective for persons who engage in behaviors with high per-act infectivity (e.g., receptive anal intercourse) with persons known or likely to be HIV-seropositive. However, the drugs might not be cost-effective for treating exposures with low per-act infectivity or involving partners at low risk for HIV infection.

Back | Next
Table of Contents

  • Email Email
  • Printable Single-Page Print-Friendly
  • Glossary Glossary

This article was provided by U.S. Centers for Disease Control and Prevention. It is a part of the publication Morbidity and Mortality Weekly Report. Visit the CDC's website to find out more about their activities, publications and services.