Methods used to estimate the number of people infected by HIV/Aids in India are flawed and the actual number of cases may be far lower, a study says.
The study says infection rates could be 40% lower than estimated
The UN estimates that India has the highest number of HIV infections, with 5.7 million people carrying the virus.
The survey of blood samples reported by British journal BMC Medicine suggests the true figure may be 40% of that.
UN and Indian experts queried whether a study conducted in one district in the south could be extrapolated nationwide.
The study is based on research in Guntur district in the southern Indian state of Andhra Pradesh, which is worst-hit by the infection.
Investigators collected blood samples from 12,617 people aged between 15 and 49 in Guntur - one of the worst affected areas in the state - to come to their conclusions.
The method estimated that there were 45,900 people living with HIV in Guntur, compared with the estimate of 112,600 reached by the official method.
Extrapolating findings in Guntur, the study's suggests there may be between 3.2 million and 3.5 million adults with the infection in India.
"India may be overestimating its HIV burden with the currently used official estimation method," the study said.
However, study investigator Dr Lalit Dandona said that even though his team's numbers were smaller, they were "by no means suggesting that the problem is already taken care of".
Speaking to the BBC, Dr Dandona agreed that more research needed to be conducted elsewhere in India to substantiate the study's findings, but he stood by its methodology.
The official method uses data collected from ante-natal and STD clinics and public hospitals.
Dr Dandona said he believed that there was no intentional effort to inflate the numbers, but the official method gave a flawed picture.
This is because the number of people with the infection reporting to the clinics and public hospitals were not representative of their true numbers in the population, he said.
UNAids chief in India Dr Denis Broun told the Associated Press that this was a "good study and definitely useful".
But he said there were problems with the methods used for arriving at its conclusion.
"Even if it were right in Guntur, it would not mean it is right all over India," he said.
A member of India's National Aids Council, Dr Sarojit Jana, stood by the way official figures were collated. He told the BBC that it was "not right" to make conclusions on the basis of a single district.
Aids awareness campaigners at the Naz Foundation agreed and warned that the study could create a "false sense of security".
Recently, former US President Bill Clinton called India the epicentre of the global HIV/Aids epidemic.
A report in India warned that its economy would suffer if the country fails to check the spread of HIV and Aids - economic growth currently at 8% could fall by nearly 1% if the disease is not contained, it said.