The relationship between diet, exercise and fetal growth is complex, but there is no doubt that lifestyle has an enormous impact on the size of a baby at birth.

According to Jim Clapp 3rd of the Department of Reproductive Biology and the Schwartz Center for Metabolism and Nutrition at Case Western Reserve University, Cleveland, Ohio, the main determinant of fetal growth is the average daily flow of glucose into the placenta. That correlates directly with the glucose content of a woman's diet, but the effects of exercise are more subtle.

During exercise, blood is diverted away from the placenta to the muscles. But if a woman is in good shape as a result of taking regular exercise before her pregnancy, that siphoning-off effect is buffered by her increased overall blood volume and cardiac capacity.

Observational studies have shown that moderate exercise during pregnancy can boost a baby's growth and decrease the risk that it will be born prematurely. But until now, nobody has carried out a controlled, randomized trial, following through the effects of different exercise regimes on the size of the baby at birth.

Now Clapp and colleagues have done so with 600 American women. The women, who were all regular exercisers before their pregnancy, didn't smoke and ate balanced diets.

In one study, women were randomly assigned to two groups - one that took 20 minutes of moderate, weight-bearing exercise such as power-walking or aerobics three to five times a week, and a control group that took no regular sustained exercise until after the birth of their child.

The babies born to the exercise group were on average 250g heavier than those of the controls. Ultrasound analyses during the pregnancy showed that their placentas also grew faster, and they were larger at birth.

In the second study, women were assigned to three groups. The first took 20 minutes of moderate exercise five times a week in early pregnancy, increasing to an hour five times a week after the 20th week. The second reversed this pattern, while the third took 40 minutes five times a week throughout their pregnancy.

Those who cut down on exercise in late pregnancy gave birth to larger offspring with larger placentas than either of the other two groups, and those who increased their exercise in late pregnancy were delivered of the lightest babies.

Clapp says his findings demonstrate that physical activity could be used in a therapeutic way to regulate the growth of the placenta and fetus. "This may well have very positive effects in inactive women, those with gestational diabetes, and those at risk for growth restriction," he said.

Maureen Hatch of the division of cancer epidemiology and genetics at the US National Cancer Institute in Rockville, Maryland, who has studied fetal growth, says a randomized trial like this has been a long time coming.

Although very small babies tend to have health problems, it's not clear whether the babies born 250g lighter in this study will suffer any adverse consequences in later life. However, says Hatch, "We think smoking is bad, and in general smokers have babies who are only about 250g lighter than [those of] non-smokers.

She also warns that the benefits of exercise in pregnancy only hold for fit women. If a woman is not fit, but wants to take up exercise when she finds she is pregnant, she should start gently and gradually build up her exercise regime.

"The most damaging thing you can do is to launch a heavy exercise program when you've been a couch potato prior to your pregnancy," she said, adding that in late pregnancy, she advises women to switch to non-weight-bearing exercise such as cycling or swimming.

Clapp's research was presented on Friday 25 April at a meeting on controversies in obstetrics, gynecology and infertility held in Berlin, Germany.