London Aug 5 : A US study has questioned the credibility of the global HIV statistics, claiming that US health officials have annually been underestimating new HIV infections by 40 percent, for almost a decade now.
The most recent global revisions, that came in November last year, included drastic downgrades of the numbers of cases, thanks mainly to revised and more optimistic figures from India.
The researchers who brought to light the discrepancy have said that the implications for global HIV figures are unclear, but the World Health Organisation has defended the accuracy of its current estimates.
This underestimate in US HIV infections came after a new blood test revealed how recently someone has been infected. And this led the US Centers for Disease Control in Atlanta, Georgia, to raise its national estimate of the annual infection rate by 40 percent, taking the figure to 56,300, from the previous calculation of 40,000.
"Our earlier estimate was based on less precise methods," New Scientist quoted Irene Hall, chief of the CDC's HIV Incidence and Case Surveillance Branch, as saying.
However, she said that the CDC had always suspected that 40,000 was an underestimate, based on data from smaller studies monitoring infection trends. But, she couldn't really say if the same applies to other countries as well, for "the methods are all different in other countries".
While most countries take the latest data on the national prevalence of HIV and estimate the number of new cases by applying pre-calculated "correction factors" to the overall numbers, the new estimate for the US is the first to be based on direct analysis of samples from patients.
The test, called the BED HIV-1 Capture Enzyme Immunoassay, tells about the proportion of a person's antibodies primed to combat HIV. As there's an increase in the anti-HIV antibodies weeks after infection subsides, it's possible to tell how recently someone was infected.
The researchers took into account the unused blood samples and medical data taken in 2006 from 6864 patients in 22 US states, which was then anonymised and sent to CDC for analysis. It was revealed that 31 percent turned out to be recent infections, allowing Hall and her colleagues to calculate the new national figure.
The data confirmed earlier studies showing that the highest rate of infection is in men who have sex with other men, accounting for 53 percent of all new infections in 2006. Some 31 percent were through high-risk heterosexual sex. And, ethnically, black Americans were worst affected, accounting for 45 percent of all new cases.
However, Kevin De Cock, director of HIV/AIDS at the WHO, doesn't agree with it.
"The big change last November was a reduction from 40 million people worldwide living with HIV to 33.2 million. There was also a reduction in estimates of new infections, from 4 to 5 million down to around 2.5 million per year," he said.
He defended the accuracy of the global figures, saying: "Do I think the current estimates are good? Yes I do," he said, adding that they were based on the best figures available in difficult setting. But he said that the US discrepancies were startling, and that "there is a risk that we might sometimes underestimate".
The US study findings were revealed at the 17th International AIDS Conference in Mexico City.