The study was published in the Archives of Disease in Childhood alongside other articles examining the links between cot death and mothers who drink alcohol and coffee.

It found no evidence that duvets could contribute to cot deaths.

This contradicts an earlier UK study which found that the use of duvets doubled the risk of sudden death before the age of one.

But the Foundation for the Study of Infant Deaths says the new research is misleading, and could confuse parents.

The problem stems from the definition of duvet, which in the UK means a soft quilted cover.

Blanket confusion

However, in New Zealand, where the new study was carried out, the word "duvet" could also mean a thin crochet cover.

Joyce Epstein FSID's Secretary General, said: "We are not comparing like with like in the New Zealand and UK research.

"It's a great pity that parents will be confused by these results. Parents in the UK should continue to avoid the use of duvets for babies under one year."

The research into alcohol consumption found that babies of mothers who drank heavily were more likely die suddenly.

However, this research was based on only an extremely small sample of women.

A larger sample revealed that high levels of coffee consumption by the mother after birth appeared to have no effect on the likelihood of cot death.

Ms Epstein commented: "Mothers drinking alcohol during pregnancy do not appear to increase the risk of cot death, however, it is well known that it can be harmful to babies for a number of reasons.

"Binge drinking once the baby is born, however, does seem to increase the risk of cot death and mothers are recommended to keep their intake low or refrain altogether.

"Babies should not sleep with their parents if either parent has been drinking, smokes, or is excessively tired."

Cot deaths decreasing

Recent figures have shown cot deaths decreasing in the UK.

There here were 284 cot deaths in England and Wales in 1998 compared with 393 in 1997. This is a fall of 28%.