Giving antibiotics to some pregnant women in premature labour may increase their chance of having a child with cerebral palsy, according to a study.

The seven-year study found babies had a greater risk of developing the disability if their mothers' waters had not broken and they were given antibiotics despite showing no signs of infection.

The Medical Research Council (MRC) report found the risk was not greater where mothers had gone into premature labour and their waters had broken.

The new study followed up the Oracle trial completed seven years ago to see how children developed if their mothers had been given antibiotics during premature labour.

Women who showed obvious signs of infection were not included in the study as the risk to mother and child was considered too great for antibiotics not to be administered. But some women who showed no obvious symptoms were given antibiotics in case an underlying infection had caused them to go into premature labour.

Sara Kenyon, of the University of Leicester who led the Oracle Children Study, said: "It is unclear why the follow-up showed this unexpected increase in the number of cases of cerebral palsy in babies born to the group of women whose waters had not broken and not in the other group.

"Before the Oracle trial, there was some evidence of short-term benefits of antibiotics in premature labour, but we did not know what the long-term outcomes would be, which is why we conducted the follow-up."

The study, published in medical journal The Lancet, found the risk of cerebral palsy was increased by the antibiotics erythromycin and co-amoxiclav but "the overall risk of this condition was low".

Children had a 3.3% chance of developing cerebral palsy if their mothers had been given the antibiotic erythromycin, compared with 1.7% if they had not received the drug. There was a 3.2% chance of developing cerebral palsy for children whose mothers received the antibiotic co-amoxiclav, compared with a 1.9% chance for those who had not taken it.

Pregnant women who were given both drugs had an even bigger chance of having a child who developed cerebral palsy, almost three times that for those who took only a placebo, although the risk was still low.