This accords with other research into breast and ovarian cancers suggesting that the constant triggering of hormones for ovulation without a rest period - such as during pregnancy - eventually sets the cancer in motion. Certainly, the American study found that pregnancy and breastfeeding had a protective effect against ovarian cancer.

Nevertheless, a number of earlier observational studies have shown a threefold risk of ovarian cancer with IVF. For instance, 12 studies analysed by the American Collaborative Ovarian Cancer Group at Stan-ford University in California discovered that the risk of invasive ovarian cancer among infertile women who’d taken fertility drugs was almost three times that of fertile women and higher than that of infertile women who had not taken the drugs (Am J Epidemiol, 1992; 136: 1175-83).

Studies of clomiphene citrate have shown that women who take the drug are increasing their chances of developing an invasive ovarian tumour by two and a half times. Those who’d taken the drug for 12 or more cycles were at an even higher risk (N Engl J Med, 1994; 331: 771-6).

There is also evidence that follicle-stimulating hormone activates a certain enzyme that has a growth-stimulatory effect on the surface cells of the ovaries (J Clin Endocrinol Metabol, 2002; 87: 2245-53).

However, the problem with the studies thus far is that they are ‘observational’ - that is, they assemble a group of women who’d taken fertility drugs, and another group who hadn’t, and then compare the rates of cancer. This kind of study doesn’t control for other factors, such as healthy lifestyle or other hormone use, known to affect cancer risk.

So, we could be comparing two groups of women who have used hormones - one with IVF and the other from the Pill. Furthermore, we could be comparing IVF cancers against the normal breast and ovarian cancer statistics - which, of course, are inflated because of the cancers caused by other hormone use in the general population. In the case of HRT, earlier case-control studies were ultimately proved false when a prospective, randomised, double-blind placebo-controlled study (where a group of similar women are randomly put into groups, with one group given infertility drugs and the other, a placebo) was finally carried out.

Children at risk Cancer is not the only concern about IVF. For a long time now, there have been worries over the possible risks of rare birth defects or other health problems in the children themselves, particularly those that develop over the longer term.

This led the Johns Hopkins University in Baltimore, Maryland, to join forces with the American Society for Reproductive Medicine and the American Academy of Pediatrics earlier this year to convene a panel of experts, who are currently analysing the available research. In the UK, the Human Fertilisation and Embryology Authority is working along the same lines.

The key risks involve multiple births, which are 27 times more likely among IVF babies than non-IVF babies, with a 5.4 per cent chance of abnormality. Babies are also five times more likely to be born prematurely and to be underweight at birth (Lancet, 1999; 354: 1579-85).

IVF babies also have three times the rate of cerebral palsy compared with children in the general population, and this applies whether they are twins, triplets or singletons of normal birthweight, according to a Swedish study (Clin Dev Med, 2000; 151: 67-83; Lancet, 2002; 359: 459-60, 461-5).