Blood From High-Risk AIDS Areas May Be Safer

Researchers at the Harvard School of Public Health (SPH) and Tufts have shown that blood taken from groups with historically higher incidences of AIDS may be safer.

A team of six researchers from Harvard and Tufts universities have proposed a mathematical explanation for how the risk of HIV contamination in the blood supply can be reduced. The group's findings appear in this month's issue of Medical Decision Making.

"We developed a model which could show the overall probability for error [in HIV testing] which takes into account both random error and window error," said Eugene Litvak, senior research associate in the Department of Health Policy and Management at the SPH.

When tests screen for HIV, they are determining the presence of antibodies to the HIV virus. Once infected, a person enters the window period in which they develop the antibodies to HIV.

So it is possible for infected individuals to test negative for HIV because their antibodies have yet to reach a detectable level.

According to Litvak, this so-called window error was discovered several years ago. Scientists had previously thought that random error was the only factor to blame for inaccuracies in test results.

Litvak worked in collaboration with Joanna E. Siegel, former assistant professor of maternal and child health at SPH; Stephen G. Pauker, vice chair of the department of medicine at Tufts University; Marc J. Lallemant, former visiting scientist in the department of cancer biology at SPH; Harvey V. Feinberg '67, provost of Harvard University and Milton C. Weinstein, Kaiser professor of health policy and management and biostatistics at SPH.

Last December scientists at the Center for Disease Control (CDC) published an article in the New England Journal of Medicine that reported that in areas where AIDS incidence was high, there were fewer cases of HIV infection by blood transfusion, a finding which Litvak described as "counter-intuitive."

Litvak's group developed a model which accounts for this seemingly paradoxical situation.

When HIV incidence is high, there are more people entering the window. However it is also important how quickly those infected leave the window or become symptomatic--progress from HIV infection to AIDS.

In order to better determine HIV prevalence, the researchers determined that the incidence of AIDS should also be taken into account. Their model focuses on the ratio of HIV over AIDS incidence as an important determinant of error.

In areas where AIDS incidence is high, the window error is minimized. This explains the CDC's findings that where there was a greater incidence of AIDS, there was a lower probability of blood contamination.

According to this model, the window error is likely to be highest in groups where AIDS incidence is low, but HIV incidence is growing. A group that fits this criteria are females in their twenties.

The study shows that the long held belief that female donors are better than male donors may no longer hold true if the incidents of HIV infection in females continues to increase.

Litvak said that he "would definitely be hesitant" to say at this point that female donors are better.

According to the findings, window error poses the greatest risk in areas or groups where the epidemic is starting. As an epidemic matures, more people pass the window.

In the United States, where the epidemic is now more mature, Litvak says "nature is working in our favor."