Rethinking AIDS Strategy After a String of Failures

Here in Durban, one of the most heavily infected cities in the world, researchers at Abdool Karim's center called each of their 53 vaccine trial participants with the grim news.

Bonga Mkhize, 24, who had received a shot in his upper arm over the past few months, said glumly, "I was expecting it to work."

Among a group of nearly 700 subjects worldwide who received two doses of the vaccine, 19 became infected with HIV, compared with 11 for a similarly sized group that received placebos. The finding alarmed some scientists and underscored the tricky ethics of using human subjects to test potential remedies for incurable diseases.

South African researchers last week began warning hundreds of volunteer test subjects that the vaccine might actually have increased their risk of contracting HIV.

Two trials for microbicides -- gels that women insert into their vaginas to prevent infections -- also ended when more women using the experimental substance became infected with HIV than those using placebos. Scientists theorize that vaginal irritation caused by these products may have made it easier, not harder, for the virus to infect women. A study of whether diaphragms might inhibit HIV found that they were also ineffective.

"It's been an appalling year for the biologists," said Francois Venter, president of the Southern African HIV Clinicians Society.

Technology vs. Reality

A technology that has worked in highly controlled settings often fails in the context of actual sexual behavior.

Hospitals routinely use antiretroviral drugs, for example, to prevent infections in doctors and nurses stuck by HIV-infected needles. But when researchers asked healthy West African women to take such medicine every day, the difference in infection rates was so small that scientists could not determine whether the medicine worked.

Condoms, meanwhile, can block HIV but are not used routinely enough to reverse the widespread epidemics in sub-Saharan Africa.

Researchers have struggled to prove the effectiveness of other popular and heavily funded strategies. For example, many scientists believe that treating sexually transmitted infections should slow HIV by healing the ulcers that encourage infection. But five of six large studies so far have ended in failure.

Theories about the ability of HIV testing and counseling to encourage safer sexual behavior also remain unproved. In some studies, those who learned they had the virus reported altering their behavior; those who discovered they were not yet infected did not. But most studies have found that making HIV testing more available does not slow the spread of the virus, and a rigorous new one published recently in the publication JAIDS found the virus spread most swiftly among those with the greatest access to testing and counseling.

For a study of people at high risk for HIV, Beauty came each month to a clinic that had muffins and hot tea at the ready. Researchers tested her for HIV and gave her free condoms, extensive AIDS counseling and a modest stipend.