The United Nations on Monday radically lowered years of estimates of the number of people worldwide infected by the AIDS virus, revealing that the growth of the AIDS pandemic is waning for the first time since HIV was discovered 26 years ago.
The revised figures, which were the result of much more sophisticated sampling techniques, indicate that the number of new infections peaked in 1998 and the number of deaths peaked in 2005.
The new analysis shows that the total number of people living with HIV has been gradually increasing, but at a slower rate than in the past.
Hints of those trends were present in the older estimates, but at much greater numbers.
UNAIDS estimated in a report to be issued today that about 2.5 million people will be infected with the AIDS virus, called HIV, this year -- a 40% drop from the 2006 estimate.
The report also says that about 33 million people worldwide are infected with the virus, compared with last year's estimate of almost 40 million.
Reports over the last decade or longer have portrayed a disease spiraling out of control, but improved methods of counting people with AIDS have unveiled a different picture.
"For the first time, we are seeing a decline in global AIDS deaths," said Dr. Kevin De Cock, director of the AIDS department at the World Health Organization.
UNAIDS Executive Director Dr. Peter Piot said the new estimates also reflected improved treatment rates and changes in sexual behavior in some affected regions of the world.
"These improved data present us with a clear picture of the AIDS epidemic," he said. "Unquestionably, we are beginning to see a return on investment."
The data represent some of the first good news in the battle to control the pandemic, particularly coming after recent reports indicating that promising HIV vaccines are ineffective and perhaps even dangerous.
The numbers have been highly politicized because they are used to govern the distribution of the billions of dollars in aid that is being poured into the problem by industrialized countries -- an estimated $10 billion this year.
Some critics viewed the changed estimates as a remarkable admission by world health authorities that they had made a mistake.
Dr. James Chin of UC Berkeley, a former WHO AIDS expert who has been tracking the disease since it first emerged in California in the 1980s, has been arguing for years that the UNAIDS figures have been inflated.
"It's getting closer to what it ought to be, but it's still high," he said. "It seemed to me that high-rise house of numbers had to crumble."
Chin estimated the total number of cases worldwide at 20 million to 30 million.
UNAIDS has "been overemphasizing and exaggerating numbers in an effort to get more and more money," Chin said. "A lot of people say the ends justify the means. It's going to come back and bite you when the real numbers appear."
Dr. Paul De Lay of UNAIDS said that he considered it "absurd" to think his agency would exaggerate the data.
"It would be technically impossible to somehow rig the numbers," he said.
De Lay said the revised numbers would only have a small effect on the budget the agency has recommended for treatment programs. With the downward revisions, the $40 billion recommended for 2010 would drop to $38 billion, he said.
Dr. Roger Detels, a UCLA epidemiologist, cautioned that the reduced numbers should not be used as an excuse to dismiss concerns about the pandemic.
"Even though the estimates are lower than we had previously thought, they're still pretty significant," Detels said. "You're still talking about prevalences in sub-Saharan Africa where you've got over 20% of adults infected with HIV. . . . I think the danger here is to say: 'Oh my Lord, you know they overestimated. It's not a very serious epidemic.' I would say 33 million is a pretty serious epidemic."
Detels noted that getting accurate numbers is difficult in any epidemic.
"You want to raise public concern enough so they'll do something," he said. "On the other hand, you don't want to overestimate because people get fatalistic about the possibility of doing something."
Health officials say that the rate of new infections has legitimately declined in several countries in both eastern and western Africa due to widespread changes in sexual behavior.
But the bulk of the apparent decrease comes from improved techniques for counting cases.
In the early days of the pandemic, researchers studied primarily women receiving prenatal care because they were more likely to encounter the medical establishment. Analysts then projected those AIDS rates to the entire population. Such women tended to be more urban, wealthier and more likely to be sexually active -- factors that appear to have skewed the data.
The new prevalence numbers, in contrast, were based on data obtained in house-to-house surveys in 30 high-prevalence countries.
The surveys included extensive questionnaires and the drawing of blood samples.
Results from one such study in India, released earlier this year, cut the estimated number of cases from 5.7 million to 2.5 million. There, researchers interviewed members of 120,000 households to arrive at the new prevalence data.
Researchers had previously said that India had the majority of the world's cases.
The revision affects past prevalence numbers as well. A 2002 UNAIDS report, for example, estimated a worldwide total of 42 million cases. The new report says the actual number that year was only 30 million.
Chin believes that the numbers have probably been wrong since at least 1997. He argues that if the number of new cases peaked in the late 1990s, the total number of cases could not keep climbing.
He said the infection would max out in the high-risk populations and then begin a natural decline as people who were untreated began to die about a decade later.
He says that prevention programs, such as condom use for sex workers, are also helping to decrease the numbers.
De Lay of UNAIDS said Monday that accurate figures have been hard to get because of the high cost of household surveys, about $2 million to $3 million per country means they cannot be done every year.
"They're big, expensive, time-consuming studies," he said.