A study found that the damage done by a mother's poor diet during pregnancy or while breastfeeding could be irreversible.

Scientists suggest that the research proves that "you are what your mother ate".

The finding comes from a study involving rats at the Royal Veterinary College in London but scientists believe it also applies to humans.

Researchers studied the health patterns of a "junk food generation" of rats whose mothers were fed a diet of doughnuts, muffins, biscuits, crisps and sweets during pregnancy and lactation.

They found that the second generation had unusually high levels of cholesterol an increased risk of heart disease because of fats in the bloodstream.

The animals also had higher than average levels of glucose and insulin, making them susceptible to type two diabetes.

"It seems that a mother's diet whilst pregnant and breastfeeding is very important for the long-term health of her child," Dr Stephanie Bayol, one of the researchers, said.

"We always say 'you are what you eat'. In fact it may also be true that 'you are what your mother ate'.

"This does not mean that obesity and poor health is inevitable and it is important that we take care of ourselves and lead a healthy lifestyle. But it does mean that mothers must eat responsibly whilst pregnant."

In previous research the same team fed pregnant rats a diet rich in fat, sugar and salt and found that many gave birth to offspring which over-ate and had a preference for less healthy food.

The new study, reported in The Journal of Physiology, found that even after the second generation were weaned off junk food, their metabolism remained altered, making them overweight and unhealthy.

Their health problems persisted beyond adolescence even after they switched to healthier diets and they stayed fatter than rats whose mothers had not eaten junk food.

Crucially, much of the extra fat surrounded the kidneys - another diabetes risk factor.

Professor Neil Stickland, a fellow researcher, said: "Humans share a number of fundamental biological systems with rats, so there is good reason to assume that the effects we see in rats may be repeated in humans.

"Our research certainly tallies with epidemiological studies linking children's weight to that of their parents."