Spermicides and birth defects.

Spermicides and birth defects 

An epidemiologic study showing an association betweenspermicide use and birth defects was flawed, claims one of its authors. A second author author says it should never have been published. But in a battle being played out in the letters section of the Dec. 12 JOURNAL OF THE AMERICAN MEDICAL ASSOCIATION, other authors of the initial report are standing by it.

In 1981, Hershel Jick of Boston University and others studiedthe incidence of birth defects for 790 Seattle women who had filled prescriptions for spermicides fewer than 600 days before miscarrying or delivering a baby. They compared this with the birth defect incidence of women who delivered or miscarried around that time who had not purchased spermicides. They found a 2.2 percent incidence of birth defects and a 3.5 percent miscarriage rate in the spermicide purchasers, compared with 1.0 percent and 2.0 percent in the control group (SN: 4/11/81, p.229).

But subsequent studies and a Food and Drug Administrationpanel failed to confirm the relationship, and study coauthor Richard N. Watkins of Seattle's Group Health Cooperative of Puget Sound now says that the original conclusion "was based on an inaccurate presumption of exposure to spermicide near the time of conception."

Watkins reexamined the medical charts of the eight womenin the spermicide-purchasing group who had had deformed children and found that four said their pregnancies were planned, suggesting they had stopped using spermicides before conceiving.

Another coauthor, Lewis B. Holmes of Massachusetts GeneralHospital in Boston, notes that no subsequent articles have shown "unequivocal evidence" of the association and that the type of careful reservations about the strength of the study included in the original publication tend to be ignored in courts of law. "In retrospect," he says, "I believe our article should never have been published."

But Jick and two coauthors stand by their work. "A number ofauthors have found associations between spermicide use and both chromosomal anomalies and limb anomalies very close to those reported by us," they write. While an association with limb-related birth defects "remains in substantial doubt," they say, the literature on chromosomal abnormalities "is more supportive ... but it is far from definitive." They "still believe that [the] report did address plausible hypotheses, was valid in its conception and conduct, was circumspectly reported and is consonant with much material published both before and since."