FDA advisers said that was not enough and recommended that the drugs not be used for children younger than 6. An expert panel said children older than 2 could keep taking the medications while studies are undertaken to settle scientific questions about safety and effectiveness.

It turns out that when the FDA set standards for cough and cold medicines some 30 years ago, no separate studies were done for kids.

Cough and cold medicines send about 7,000 children to hospital emergency rooms each year with symptoms ranging from hives and drowsiness to unsteady walking. Low doses of a medicine are not likely to cause a problem; the main risk comes from unintentional overdoses.

The same ingredients usually are found in different products. For example, giving a child a cough syrup and a decongestant could inadvertently lead to an overdose.

The Consumer Healthcare Products Association, which represents the manufacturers, says preventable errors are the problem, not the safety of the ingredients in the medicines. The industry is starting an educational campaign aimed at parents, doctors and day care providers on the importance of following directions and storing medicines in places where kids cannot get at them.

"The data clearly show a majority of adverse events are direct result of misuse of our products," said Linda Suydam, who heads the industry group.

Baltimore health commissioner Dr. Joshua Sharfstein sought to reassure FDA officials worried about unintended consequences if the government moves to restrict the medications and parents start dispensing adult drugs to their preschoolers. Sharfstein said the state of Maryland saw an immediate benefit after OTC cough and cold remedies for tots were removed from store shelves last fall. Calls to poison control about problems with the medicines involving children younger than 2 dropped by 40 percent, from 99 to 60, in the first six months of this year when compared with 2007. Calls involving children 2 to 6 also dropped, but by much less.

"The feared increases in poisonings simply did not happen," said Sharfstein. "In fact, the opposite occurred."