A baby in America tested HIV positive after birth but found to be clear of the virus a year later may provide proof that babies born to HIV mothers can rid themselves of the virus.
The boy, now five and still HIV negative, is not the first case to show this phenomenon, but his case is the best documented. Others have been dismissed as laboratory mistakes.
The researchers from the UCLA School of Medicine, Los Angeles, say in the New England Journal of Medicine that their results indicate that HIV clearance can happen and may have been under-recognised.
Dr Marie-Louise Newell, co-ordinator of the European collaborative study of children born to HIV mothers, which is currently monitoring 1,800 cases, said: "This might help in pointing the way for vaccine research or confirm that research is moving in the right direction.
"What needs to happen now is investigation of the child's immune system to see if anything can be found to explain what happened.
"There have been a number of cases like this in children from Europe but people have been a bit sceptical of the results. A great deal of work was done on this case and it is nicely documented, reliable data."
Dr Newell, an epidemiologist at the Institute of Child Health in London, pointed out that levels of the virus in the baby had been very low.
The boy's mother was tested positive for HIV during pregnancy. The baby was found to have the virus when he was tested 19 days after birth and again at 51 days old. At a year old, he had a routine test and was found to be negative. Numerous, sophisticated tests since to find the virus have failed.
Reports exist of virus resistance in adults. A group of Gambian prostitutes have remained uninfected despite long exposure to HIV and sexual partners of known HIV cases have also remained clear.
The American findings raise more questions than they answer. The researchers say it is still possible that the boy has a "hidden'' virus which could re-emerge at a later date.
In an accompanying commentary in the journal, Dr Kenneth McKintosh, of Boston Children's hospital, says the findings are hard to explain.
He says: "Does this mean that in a proportion of infants HIV in some form enters the foetus or the newborn and is then cleared?"
He said it would be surprising if it was found that there was an effective immunological response in babies as the immune systems in newborns are considered to be immature.
Dr McKintosh says that in the light of the new case it seems that previous, less well studied cases which had been dismissed, perhaps "were not errors, at least not all of them".
Gaz Daly, spokesman for the charity Aids Care Education and Training, said: "Obviously it is very good to hear this news, but one has to treat it with caution. The difficulty here is that many people who have HIV and want children may be tempted to go ahead.
"They must accept that a baby could be infected and that they only have five or six years to live."