Chorionic villus sampling (CVS) is not only a dangerous test in fetuses older than nine weeks when it was assumed to be the safest period for testing but carries a high risk of inaccuracy.

Doctors should warn parents of the risks of CVS causing limb defects to their babies at least up to 76 days' (nearly 11 weeks) gestation, the US Centers for Disease Control and Prevention has recommended.

Their views go against the recent World Health Organization pronouncement that there is no risk of limb defects if CVS is carried out after the ninth week.

The Centers, based in Atlanta, Georgia, quote the research work of Firth et al, published in The Lancet on 30 April, 1994, which shows that later testing at between nine and 10 weeks have caused defects, or transverse limb deficiency (TLD).

One theory for the discrepancy in the two findings may be due to the inexperience of obstetricians at some centres, who may unwittingly be causing damage during the test (The Lancet, 13 August, 1994).

But if CVS is more dangerous than first thought, it also is more inaccurate. Doctors at hospitals in Brest, France report that CVS carried out on a fetus indicated the presence of the chromosome which causes Down's syndrome.

Despite the test results, the parents decided to proceed with the pregnancy and the mother gave birth to a healthy baby girl.

Dr M J Le Bris, writing in The Lancet, 20 August 1994, estimates the chances of such a false/positive are one in six.

But mothers who might prefer early amniocentesis (EA) instead should know that this test greatly increases the risk of miscarriage. It is also slightly more likely to cause cases of club foot than does CVS, according to research by Professor Kypros Nicolaides at the King's College Medical School in London.

!AThe Lancet, 13 August 1994.