PREGNANCY LABELING CATEGORY INTERPRETATION PROMOTES "BAD SCIENCE," ARGUS RESEARCHER TELLS FDA; RISK NARRATIVES COULD REPLACE CURRENT CATEGORIES

FDA reviewers should be more willing to reject applications based on inadequate reproductive toxicity studies, Argus Research teratologist Mildred Christian, PhD, told an agency public hearing on pregnancy labeling categories Sept. 12.

"There is now a problem in rejecting poorly designed or conducted [animal] studies, and I've heard that from many FDA reviewers," Christian maintained. "I know that there is great political pressure to put compounds through, to internationalize, to not hold things up, but I would recommend very highly that consideration be given to rejecting poorly designed and poorly conducted studies, and you do know what they are."

"There's a great reluctance of the companies to negotiate with the agency before performing [an animal] study and as a contract lab, we certainly cannot always convince our clients...to do the studies we recommend," Christian stated. "The well-designed study should identify when one approaches toxicity," she argued. "These studies are sometimes designed to not show toxicity, or to show only maternal toxicity at a particular level."

FDA called the hearing as part of its effort to develop a guidance on pregnancy labeling that will likely involve an overhaul of its current category approach. Drugs are currently placed in one of five pregnancy categories -- A, B, C, D or X -- which the agency fears "convey the impression that there is a gradation of reproductive risk from drug exposure across categories" ("The Pink Sheet" Aug. 4, T&G-8).

The current interpretation of pregnancy categories for drug labeling encourages "bad science," Argus' Christian maintained, "because if you don't look, you won't see things." On the other hand, "good science should not be penalized by labeling," she said.

Sponsors may be able to determine the outcome of studies through the design. If "you only look once a week you won't have an effect, but if you look every day you will see toxicity," Christian noted. Argus is a contract research firm that conducts reproductive and developmental toxicity studies.

"Many times I have companies...come to me and say, `Our competitor has a B category, we have a C category. Can you help us get ours changed or can we do studies of theirs that show that they should be degraded to a C, because our marketing people say there is such a difference,'" Christian said. "Unfortunately many times both companies have inadequate animal studies to support the labeling."

Christian urged FDA and the pharmaceutical industry to work with communication experts when writing pregnancy labeling to define terms and communicate the meaning of data effectively. "Certainly working with drug companies, when one writes up that [a drug] is safe or not safe or it is or is not a teratogen, one really begins to appreciate that even across companies those words are not commonly interpreted, and certainly when that type of information is in the public domain it needs to have a better level of communication," Christian added.

Speakers at the hearing agreed on the types of changes needed to pregnancy labeling, including the replacement of the category system with a more detailed narrative discussing what is known about reproductive toxicity for each drug.

Speaking on behalf of the Teratology Society, Robert Brent, MD/PhD, Jefferson Medical College (Wilmington, Del.), stated that "there seems to be an overwhelming number of reasons to get rid of classifications; it's very hard to think of reasons to keep" the system. Brent called the classification system a "rubber stamp that isn't applicable with the complexity of the knowledge that we have in 1997."

The classification system is "very confusing and very anxiety provoking," Brent maintained. In drug labeling, "FDA should utilize the same principles that are used by clinicians who counsel women with regard to reproductive risks," he said. "The present classification of drugs would never be used by an experienced counselor because they are inconsistent and do not provide a basis for a risk estimate of reproductive toxicity."

Experienced clinicians "want to know what is the stage of your pregnancy, what was the exposure, the dose, was it acute or chronic, what is the status of her pregnancy, does she have other diseases, is she a patient with lupus, does she have several other previous losses of pregnancy, what is her genetic and reproductive history -- and then we look at the human and animal reproductive toxicity data. That's a very small part of the evaluation of the individual patient's case," Brent said.

Animal study results in labeling should eventually include pharmacokinetic data, Brent recommended. "It's much different to know the dose was 10 mg/kg in a mouse; I haven't the faintest idea what the blood level was, which is what's important," Brent maintained. "Those are things in the future that the FDA may require to make animal studies more appropriate."

Animal studies should focus more on the "no effect" dose, Brent said. "When you get below a certain level even thalidomide doesn't have a reproductive effect," he said. "So when you give a drug at a hundred or a thousand times the therapeutic dose, you know that it has no meaning with regard to human reproduction."

Argus' Christian agreed that "we need to know exposures in animals relative to humans and we can tell that from the animal studies if we have appropriate species," she said. "We need to identify the relative vulnerability...not is there an effect but is there an effect on the reproductive system at any time which occurs at lower doses than other toxicities so that one can see this as a special problem."

Speakers agreed that pregnancy labeling needs to be more detailed in a number of respects, including a discussion of the quality and quantity of study data available for each drug, a "biological plausibility" statement explaining the background incidence of congenital malformation so physicians can more accurately assess the risk associated with a drug, the separation of statements dealing with risk from those dealing with risk/benefit assessments, discussions of reproductive risk at different stages, such as fertility and lactation, and a description of the specific types of malformations that each product could cause. A Lilly working group is working on a labeling format that would include much of that information in a table.

"Teratogens, even the most potent, cannot produce every congenital malformation," Brent said. "If you read the present package insert, you would have no idea that every drug can't produce every malformation."

SmithKline Beecham teratologist Patrick Wier, PhD, recommended that FDA consider "standardized core statements" instead of the categories.

"While narrative text is the best means to communicate the technical details, some care must be taken to avoid labels becoming difficult to compare by virtue of such detail," Wier said. Core statements would "facilitate comparisons between drugs in terms of the most frequently sought information....Labels would include one or more of the most appropriate core statements."

"Core statements would not compare risks to benefits," Wier said, because "it is not possible to predict with certainty the conditions or level of necessity for human use or the suitability or availability of alternative therapies prospectively."

Centers for Disease Control & Prevention Birth Defects Group Chief David Ericson supported having a "summary descriptive statement, maybe something which is quantitative" of the risk of using a particular drug, such as "the risk of fetal harm is 25%," he said. Ericson added that "it would be better to have a clear statement of risk without clouding the issue with benefit, but also to have some statement of benefit and some education of the potential of the adverse effects of not treating the condition." FDA said it is considering holding another public workshop on the issue.