Updated

LONDON - The world's first malaria vaccine, developed by GlaxoSmithKline, provides some protection after three doses but its effect dwindles to almost nothing after seven years, scientists said on Wednesday.

Publishing a long-term study of the vaccine - called RTS,S or Mosquirix and designed for children in Africa where the disease claims most of its victims - researchers said the decline in its efficacy over time is fastest in children living in areas with higher than average rates of malaria.

This raises questions about whether Mosquirix can play a meaningful role in fighting malaria, they said, and suggests a four-dose schedule would be needed if it were used.

"The results suggest that the implementation of RTS,S will need to be considered carefully and in a way that takes into account different levels of malaria exposure," said Mike Turner, head of infections at the Wellcome Trust global health charity which helped fund the research.

He added that while the vaccine "isn't perfect", it still has the potential to save lives "and will provide an important springboard for improved second-generation vaccines".

Malaria, a mosquito-borne parasitic disease, kills around 400,000 people a year, the vast majority of them children and babies in sub-Saharan Africa. World Health Organization (WHO) data show there are around 200 million malaria cases a year.

Hopes that GSK's shot could wipe out malaria were dashed when trial data in 2011 and 2012 showed that Mosquirix reduced malaria episodes in babies aged six to 12 weeks by 27 percent, and by about 46 percent in children aged five to 17 months.

Despite its limited efficacy, Mosquirix last year became the first ever malaria vaccine to win regulatory approval, when the European Medicines Agency gave it a green light. The WHO, meanwhile, has said that Mosquirix is promising in its potential to reduce cases of malaria, but should be deployed only on a pilot basis before any wide-scale use.

For this latest study, published in the New England Journal of Medicine, researchers at the KEMRI-Wellcome Trust research program in Kilifi, Kenya, followed 447 children who had received three doses of either Mosquirix or a control vaccine when they were 5 to 17 months old. After seven years, there were 312 children still involved in the study.

The results showed that during the first year, the risk of getting malaria in the vaccinated children was 35.9 percent less than in the control group, but after seven years this difference fell to 4.4 percent.

After five years, in children exposed to higher than average rates of malaria, there were more cases (1,002) in the vaccinated group compared with the control group (992).

Philip Bejon, director of the KEMRI programme, said this "rebound" effect is thought to be caused by the vaccinated children developing their natural immunity against malaria more slowly than unvaccinated children.

GSK, which has been working on Mosquirix for 30 years, has promised it will make no profit from it, pricing it at the cost of manufacture plus a 5 percent margin which it will reinvest in research on malaria and other neglected tropical diseases. The shot also contains an adjuvant, or booster, made by the U.S. biotech company Agenus.

(Reporting by Kate Kelland)