AIDS numbers rise around world

WASHINGTON -- The numbers of people infected with HIV and those dying from AIDS appear to be rising in every region of the world, and nearly a quarter of the new infections are in Asia, according to a biennial United Nations report released yesterday.

The report, which acknowledges that the UN's statistical model for years had been overstating the pandemic's reach, revises some countries' estimates of the number of people infected dramatically downward, especially in sub-Saharan Africa. The report was released in advance of the 15th International AIDS Conference, to begin in Bangkok Sunday.

UNAIDS and the World Health Organization, which are responsible for the estimates, found that roughly 38 million people worldwide were infected with HIV, the virus that causes AIDS, at the end of 2003. That represents a 13 percent reduction from its estimate two years ago, UN officials said. Many AIDS specialists outside the UN have been sharply critical of the past estimates, suggesting that the figures have been inflated by 25 to 50 percent.

UNAIDS also retroactively lowered estimates from its last report, two years ago. Using those revised figures, officials said that the numbers of people infected continue to grow.

Peter Piot, UNAIDS executive director, cautioned in a telephone interview from Geneva that the downward revisions of the estimates do not mean the epidemic is slowing. Rather, he said, AIDS is continuing to expand, reaching alarming heights in Asia and Eastern Europe.

''The number of people who became infected is higher than ever before, the number who died is higher than ever before," he said. ''We want to especially highlight Asia, where there has been an acceleration."

In Asia, the UN estimates 1.1 million people were infected in 2003, citing ''sharp" rises in HIV infections in China, Indonesia, Vietnam, and India. In 2001, roughly one-fifth of the new infections around the world were in Asia; two years later, it was roughly one-quarter, Piot said.

In India alone, the number of people living with HIV was estimated at 5.1 million, up from roughly 4 million two years earlier.

But the estimates for India were far from firm and show the difficulties in arriving at a number in the world's second most populous country. The report uses ranges in each country on numbers of people infected and then selects a number in between. In India, according to tables in the report, the number of those infected is somewhere between 2.2 million to 7.6 million people, signaling the huge degree of uncertainty. Piot acknowledged: ''India is a blank."

One positive development recorded in the report was that of the 38 countries providing statistics in sub-Saharan Africa, the region of the world hardest hit by the pandemic, 15 registered a decline in the percentage of adults infected with HIV.

Elsewhere, the report highlighted mostly daunting trends and a growing need for assistance:

The epidemic grew in Eastern Europe and Central Asia, fueled by injecting drug users and their non-sterilized needles, and now 1.3 million are believed to be infected. More than 80 percent of those infected are under the age of 30, compared to the United States and Western Europe, where only 30 percent are under the age of 30.

In the United States, an estimated 950,000 are living with HIV, up from 900,000 in 2001. Half of new infections in recent years have been among African Americans.

Global AIDS funding has increased 15-fold from 1996 to 2003, to just under $5 billion. But the report estimates that $12 billion will be needed by 2005, and $20 billion by 2007. This represents a substantial increase from earlier estimates, based somewhat on a greater need to safeguard blood supplies from HIV contamination.

Just 7 percent of HIV-infected people in developing countries have access to life-extending antiretroviral treatment, or roughly 400,000 people at the end of 2003. The number has increased to nearly 500,000 now, Piot said. That falls far short of WHO's initiative to treat 3 million people in poor countries by the end of next year.

But the report's figures did highlight some positive trends in sub-Saharan Africa, even if the numbers of infections continue at a sobering rate. Africa, with 10 percent of the world's population, has 70 percent of the people in the world living with HIV. The disease is largely transmitted through heterosexual sex.

The report found that in sub-Saharan Africa, the adult HIV prevalence rate -- the percentage of people aged 15 to 49 who are infected with the virus -- apparently is stabilizing. Piot said this is a result of both good and bad trends, citing reports of increased death rates from AIDS but also evidence that prevention messages are reaching young people, notably in cities in eastern Africa.

Of the 38 sub-Saharan African countries with estimates of HIV prevalence, 25 of them either registered a decrease or stayed the same from 2001 to 2003, according to the report's tables. In addition, researchers have found in several East African cities -- including Addis Ababa, Ethiopia; Kampala, Uganda; Lilongwe, Malawi; and Lusaka, Zambia -- that HIV prevalence has decreased among those aged 15 to 24. ''We feel these are true declines," Piot said.

The UN's revisions in some country estimates were substantial. For instance, the old estimate in 2001 for Kenya was 15 percent; the new 2001 estimate was 8 percent, and the UN revised it even further downward in 2003 to 6.7 percent. Rwanda's old estimate in 2001 was 8.9 percent; its new estimate in 2001 was 5.1 percent. And Zambia's old estimate of 21.5 percent in 2001 was revised to 16.7 percent.

Jim Chin, an epidemiologist who helped devise WHO's original models for estimating HIV prevalence, criticized UNAIDS and WHO for not emphasizing more positive trends in the report.

''The words 'peak' and 'leveling off' are not in their vocabulary," Chin said by telephone from California. ''They are still doing this very fancy soft-shoe dance around the numbers. Outside of Africa, it seems they are not admitting to any serious reduction to the estimates anyplace else. It's just not consistent with the epidemiology. Maybe it doesn't suit them to say it before Bangkok, where they are beating the drums of a gathering storm."

Piot defended his organization's new estimates, saying they reflected an openness to reexamining the trends of the epidemic as more information becomes available. ''The epidemiological surveillance is getting better, at least in Africa," he said. But he said evidence gathering on HIV prevalence remains weak in Asia. ''We still have a big job to do with Asia," he said.

John Donnelly can be reached at donnelly@globe.com.