An increase in low birthweight babies born in and around New York in the months after 9/11 has been blamed on stress caused by the attacks.
The stress of the attacks may have affected pregnancies
The journal Human Reproduction reported a two-thirds increase in the risk of giving birth to a slightly underweight baby in the week after 9/11.
The University of California at Berkeley researchers found it was higher even months later.
However, a UK expert said other factors could have played a role.
While many researchers have suggested that the release of stress hormones might, in some circumstances, affect the development of the unborn child, the evidence that it leads to premature birth is not conclusive.
The New York study looked at information from more than 1.6 million birth certificates for babies born in the city between 1996 and 2002.
They divided the babies into groups depending on how close to Ground Zero their families lived.
Compared to the week leading up to 9/11, the following seven days say a higher risk of babies being born weighing less than 2kg.
Normal birthweight is based on a figure of 2.5kg.
There was a 67% increase in the risk of a baby weighing between 1.5 and 2kg, and a 44% increase in the chance of a baby weighing less than 1.5kg.
In December 2001, the risk of a baby weighing less than 1.5kg was 36% higher than normal, and in January the risk was still 22% higher.
The effect was not just confined to the immediate New York City area - in the surrounding areas, or "upstate" New York, the risk of a low birthweight baby was increased by 46% in January 2002.
The researchers said that the initial shock of the attacks may have triggered early labour in some women close to the Twin Towers - while longer-term stresses for women across the whole of New York State may also have interfered with the pregnancy.
Professor Brenda Eskenazi, who led the study, said: "We think the increased incidence in low birth weights is mainly due to stress-initiated early deliveries.
"We observed immediate effects in New York City, but long-term effects both in the city and upstate.
"This may indicate that higher levels of stress are necessary to induce acute effects on birth outcome, but that, in the longer term, women in both locations suffered stress as a result of the disaster and this is reflected in the later peaks in low birth weights."
Dr Virginia Beckett, a spokesman for the Royal College of Obstetricians and Gynaecologists, described the study as "interesting."
But she said it was hard to draw any firm conclusions without knowing the week in pregnancy each baby was born - and whether they had been born naturally, induced, or by caesarean section.
She said: "I don't think we fully understand the effects of stress on pregnancy - although we know that babies can respond to levels of the stress hormone cortisol in the mother's bloodstream."
Professor Andrew Shennan, a consultant obstetrician with the charity Tommy's , said that stress might not be having a direct effect on the foetus or the mother.
He said: "It's possible that some of these early births are due to decisions by doctors to deliver early. They may be intervening because of anxieties on the part of the mother.
"It's difficult to determine the precise effect of stress on the pregnancy."