Model Predicts Halt to Africa's AIDS Epidemic

By David Brown
Washington Post Staff Writer
Wednesday, November 26, 2008

A strategy of testing adults every year for HIV and immediately treating every person found to be infected could virtually end the AIDS epidemic in Africa in about a decade, new research suggests.

While nobody is seriously espousing that approach, the "thought experiment" outlined this week in the Lancet journal emphasizes the usefulness of antiretroviral drugs as tools for preventing the spread of HIV infection as well as treating it.

The power of AIDS drugs to dramatically slow the epidemic is the consequence of two well-established facts.

The first is that the amount of virus circulating in the bloodstream is the most important factor determining whether an infected person transmits the disease to another during a high-risk encounter. The second is that AIDS drugs can lower this "viral load" in the bloodstream to one-millionth of what it is without treatment.

The researchers who constructed the mathematical model used in the new study found that if a testing-and-treatment approach were pushed aggressively enough, it could prove to be "the greatest strategy for reducing transmission" of HIV, said Reuben M. Granich, a biostatistician at the World Health Organization in Geneva.

HIV prevention today consists mostly of urging people to change their behavior: abstaining from sexual activity, having fewer sex partners, using condoms. The search for less difficult prevention strategies has been largely unsuccessful.

There is no AIDS vaccine and no prospect of one for at least a decade. Microbicide gels designed to kill or cripple the virus during sexual activity have not worked. Male circumcision reduces the risk of transmission by about 40 percent, but public health authorities have not yet tried to roll it out on a massive scale in countries where it is not routine.

The authors of the study, all WHO employees, said emphatically that neither they nor the United Nations agency are endorsing universal, annual HIV testing and treatment.

"This is a theoretical exercise based on mathematical modeling to stimulate discussion," said Kevin de Cock, head of WHO's AIDS department.

There were 33 million people living with HIV worldwide in 2007, with 22 million in sub-Saharan Africa. There were about 2.3 million new infections and 2 million deaths globally that year. About 3 million people were on lifesaving antiretroviral therapy -- taking three or more AIDS drugs simultaneously -- but another 6.7 million needed it.

The model assumes that HIV testing would be voluntary and no one would be compelled to start treatment. In developing countries, about 20 percent of HIV-infected people know they have the disease. This strategy would find most of them.

The study took into account the rate at which people on antiretroviral therapy would stop taking the medications or have to switch drugs because they stopped working, based on actual data from Malawi.

The model applies only to the type of epidemic seen in southern Africa, where nearly all transmission occurs through heterosexual intercourse. Whether the findings might also apply to epidemics in which anal intercourse between men and intravenous drug use are common modes of transmission is not known.

If adolescents and adults were tested annually and started treatment even if they had no symptoms, transmission rates would fall from 20 new cases per 1,000 people per year to 1 case per 1,000 in about a decade.

The current recommendation is that people start treatment once their CD4 cell count -- a measure of immune system health -- drops below 350 from a normal of about 900 cells per microliter of blood. Under that strategy, 8.7 million people are expected to die between now and 2050. With universal treatment, that number would be cut to 3.9 million, according to the study.

The strategy would require large up-front investment in testing and treatment but would eventually cost less than the current strategy of putting off treatment until the infection has advanced because so many new infections would be averted.

The model assumes testing, drugs and clinical management would cost about $730 a year per patient with first-line drugs, and $3,300 when a person switched to more expensive second-line drugs. By 2032, the cost of the current strategy or delayed treatment and the cost of universal treatment would be about the same: $1.7 billion a year.

De Cock said WHO will convene a meeting of experts early next year to discuss the possible implications of the model's projections.

View all comments that have been posted about this article.

© 2008 The Washington Post Company