HEART scans for newborn babies could prevent thousands of cot deaths, a study says today.

Half of all such deaths appear to be caused by a heart abnormality that can be detected in the third day of life, scientists have found. The discovery raises the prospect that all babies could be screened and, if need be, treated for a year with a common beta-blocker drug until after the danger is over.

But this would be enormously expensive and would involve giving drugs to many babies who would not have died anyway. It could also cause "needless parental anxiety", said one expert in the field of infant death research.

The study of cot deaths - sudden infant death syndrome (SIDS) - was carried out by doctors in Italy over 20 years and involved giving heart scans (ECGs) to 34,000 babies.

Over that period 34 of the babies died, 24 from cot death. The Italian rate of SIDS is 0.7 per 1,000. In Britain it is 2.5 per 1,000. The team, led by Dr Peter Schwartz, of the University of Pavia, examined the ECGs of all the babies on their third or fourth day, measuring the time between the start of the main heartbeat and when it finished - the "QT interval".

Babies with a long QT interval were 40 times more likely to die of SIDS over the next year. Scientists in the team said the link was "dramatic", much greater than known risk factors such as allowing a baby to sleep on its front or smoking by the mother.

The reason why a prolonged heartbeat was linked to the deaths is not clear. The doctors think that it indicates that the heart rhythms are less stable. However, many babies with a prolonged heartbeat did not die, which indicates some unknown outside trigger.

The doctors, writing in The New England Journal of Medicine, suggest that nerves may be growing at different rates in some babies, causing a temporary and harmful imbalance in the regulation of the heart.

Dr Jeffrey Towbin, a specialist in children's heart conditions at Baylor College in Houston, Texas, said: "If these findings are right, we may save a significant number of babies by doing ECGs." But the costs would be very great, he said, hampering other areas of medicine.

While the link was "dramatic" and the trial was large, he said, more studies were needed to compensate for any bias caused by the fact that the research was all conducted by one institute.

The question of needless parental worries was raised by Dr Shireen Chantler, a scientific adviser to Britain's Foundation for the Study of Infant Deaths. She said: "Caution is recommended, as screening has implications for all children, the majority of whom will not die."

The number of cot deaths in the United Kingdon has declined rapidly in recent years, from 1,400 in 1990 to 470 in 1996. However, last year the figure rose by six per cent, the first rise in eight years, prompting the Foundation for the Study of Infant Deaths to suggest that possibly parents are becoming complacent. Cot death is the most common killer of babies aged between a month and one year.

22 May 1998: Cook Report on cot deaths 'was scaremongering'
12 March 1998: Smoking in the home 'kills babies'
5 May 1997: Study undermines fumes link to cot deaths
20 April 1997: Experts dispute cot deaths link to vaccinations
8 January 1997: Bed sharing may cut cot death risk