12 June 2003 17:00 GMT

by Henry Nicholls

The US system of indicating risk of prescription drugs to unborn babies - the pregnancy label - is inadequate, says a senior gynecologist who urges a shake-up of current labeling.

The current five-letter system that classifies prescription drugs on the basis of their known effects on pregnancy is unsatisfactory, says Anthony Scialli, professor of obstetrics and gynecology at Georgetown University in Washington DC, a former president of the Teratology Society, which studies the causes and biological processes leading to abnormal development and birth defects.

"Distilling all available information about a drug into one of five letters is too simplistic," Scialli told BioMedNet News.

Human data are simply not available or are not of sufficient quality to provide useful information on the risks of drugs to mother and child. However, Scialli says that experimental data on animals could usefully be incorporated into the pregnancy label.

"Experimental animal data are required for the evaluation of the reproductive effects of nearly all drugs," he said. "Although there are limitations to what can be predicted from experimental animal data, recognition of these limitations can be incorporated into a useful label."

The US labeling system was introduced by the Food and Drugs Administration (FDA) in 1979, and creates artificial categories that have little clinical relevance, says Scialli. "For example, a drug that rarely causes an increase in malformations might be classed with a drug that commonly causes a malformation," he said.

Detailed pregnancy labeling is crucial for weighing the costs of prescribing a drug to a pregnant woman against the benefits it will bring, says Patricia McElhatton, head of the UK's National Teratology Information Service, which advises on all aspects of toxicity of drugs and chemicals in pregnancy.

Most epilepsy drugs, for example, are associated with an increased risk of malformation to the unborn baby, says McElhatton, with some, like sodium valproate, causing a greater incidence of spina bifida. "It's a two- to threefold increased risk of malformation," she said, but it's a risk that is usually taken. "The effects of withholding medication if it's needed are far worse than the effects [of sodium valproate] on the fetus," she said.

Although the UK has more descriptive pregnancy labeling than does the US, drugs still tend to fall into one of two categories, says McElhatton - those that are "contraindicated in pregnancy" and those "not recommended in pregnancy". Even though there may be no risk associated with drugs that are "not recommended in pregnancy", drug companies tag this onto the packaging to cover themselves against being sued should something go wrong.

This means that drugs actually carry very little information about their real risks to pregnancy. "The clinicians have absolutely no guidance whatsoever as to whether or not they can use that drug in pregnancy," she said. "What we would like to see is a clearer label as to what animal tests have been done and what has been found," she told BioMedNet News.

Later this month, Scialli will act as moderator of a workshop that aims to build on the efforts of the FDA to revise the content and format of the pregnancy label in the US. "The workshop will convene clinicians with experience and expertise in interpreting reproductive toxicity data in order to provide the FDA with guidance concerning what information should appear in the pregnancy label and how the information should be presented," he said.

"I expect that by the end of the workshop, such guidance will be available to present to FDA," he said, refusing to speculate on when changes to the current labeling scheme might be realized. "I do not know FDA's timetable for implementing the recommendations should the agency choose to do so," he concluded.