Author's Reply: Thank you for the opportunity to address the letters from Drs. Polifka and Shepard and Dr. Brent. My responses to the issues are summarized below.

First, regarding the contention of our critics that the chicken embryo is not a good model to carry out teratologic screening: I quite agree. Our study in Pediatric Research(1) was designed to test a unifying hypothesis for a biologic mechanism to explain the relationships among a broad set of well described teratogens as stated in the introductory section of the article (pp. 1-2), and as pointed out by Drs. Polifka and Shepard. Our results support the hypothesis very well, and show a new mechanism for the induction of dysmorphogenesis. I agree with Dr. Brent that the chicken embryo is "an elegant tool" for this purpose. Based upon our results, and in light of the many other peer-reviewed experimental and epidemiologic studies that were used to derive our hypothesis. I feel that the intelligent use of our results by the teratology community and the drug companies could result in better prediction of the teratogenicity of compounds that may act as NMDA receptor antagonists, and also could result ultimately in the development of effective antitussives that do not pose a potential threat via their NMDA receptor antagonism. Teratologists need to understand, embrace and apply the unifying concepts of modern molecular developmental biology.

Second, regarding their contention that no epidemiologic study has found a risk to babies with maternal use of dextromethorphan: this is simply not true. The Baltimore-Washington Infant Study(2) reported that maternal use of antitussives containing dextromethorphan (p. 50) had a highly significant relationship with three kinds of congenital heart defects whose standardized relative risk(SRR) were 8.9, p < 0.01 (p.120), 6.3, p < 0.01(p. 52) and 4.7, p < 0.01 (p. 162), respectively. That is, maternal use of dextromethorphan was associated with an apparent increase in the risk of some congenital heart defects by up to 890%. These defects occur at a low rate, and this increased risk may result in only a small number of additional heart defects. Nevertheless it is a risk that is eminently avoidable, and it is impossible to adduce these data in support of the argument that dextromethorphan poses no risk.

Drs. Polifka and Shepard report that they got their information from TERIS (TERatology Information System), which assigns a risk assessment of "None" to the maternal use of dextromethorphan, based upon two epidemiologic studies. Dr. Brent refers to the same two studies. In the first of these two studies, 1 baby had a congenital defect out of 59 babies born to mothers who used dextromethorphan during pregnancy(3). In the second reference(4), it is reported that the use of dextromethorphan by 300 mothers during mo 1-4 of pregnancy increased the risk of first trimester defects by 25%. This result did not achieve statistical significance; however, it does not support a risk assessment of "None," as discussed in the same book. Furthermore, the same reference implies a significant risk for defects whose induction is not limited to the first trimester ["anytime" defects(4)]: "For antitussives, the standardized relative risk (SRR) overall (for "anytime" defects) was 1.59 and the result was statistically significant." The major contributions to this elevated SRR were from caramiphen (SRR1.65) and dextromethorphan hydrobromide(SRR1.39)" (4, p 438). It is noteworthy that dextromethorphan users represented 2/3 of this sample population (580/867). The expected rate of occurrence of "anytime" defects was 18/1000 but the observed rate with dextromethorphan use was 25/1000. This increase of 7/1000, again, is not very high but it is avoidable, and does not support the argument than dextromethorphan poses no risk. Page 496 of reference 4 contains the information that inguinal hernia occurred 30% more often than expected among 580 children who were exposed to dextromethorphan in utero; this datum is used inexplicably by our critics to support their contention that dextromethorphan is safe. Based upon the results of these two large epidemiologic studies, it appears to be at least imprudent for TERatology Information System (TERIS) to continue to report that there is no risk from maternal use of dextromethorphan.

None of the three references used for the above discussion was included in our report(1), as pointed out by our critics. The use of data that has not itself been peer-reviewed(2,4) should have limited use in peer-reviewed manuscripts; however, I am pleased to have been given this alternative venue for their discussion. The other reference(3) reported the occurrence of a single event in a small population and should not be given the status it has been given by our critics.

The contention by our critics that our article unjustifiably extrapolates avian results to humans is moot because the epidemiologic studies cited above showed an increased risk to babies when their mothers used dextromethorphan while they were pregnant, especially for certain heart defects(2,4). Although the point is moot, our report in fact made only qualified comments about the application of our data to humans(e.g. see the quote in Dr. Brent's letter). However, I believe that the "extrapolation from chicken to human" that has most upset our critics did not occur in the pages of Pediatric Research.

The final contention by our critics is that we discussed our report with the popular press in an irresponsible or careless way. A statement about our study came from our university's institutional relations division as a routine part of their function, because our report was made the cover story for the January edition of Pediatric Research. After its release, we were asked by the media to answer questions about the study. In providing information to the press, I referred to some of the studies of human use of dextromethorphan cited above and stated that the defects were known to occur at low frequency; I concluded that further studies should be carried out. These comments by me were ignored in the story written by the Associated Press that contained the attributions that have so inflamed Dr. Brent, but some balance was achieved in the story by caveats provided by others including a senior representative of the March of Dimes.

Although the quotes in the article were not necessarily accurate, their general meaning does not appear to have been completely fallacious and inflammatory as implied by our critics: dextromethorphan may indeed act via the same teratogenic mechanism in man and in chickens, but further testing of dextromethorphan should be carried out; and women probably should try to avoid dextromethorphan if they plan to become pregnant. Given the experimental and epidemiologic data cited above, what responsible individual would recommend otherwise? This advice would be guided by a sincere desire to give children their best opportunity for normal development, not "misguided scientific advocacy" as Dr. Brent suggests.

I am well aware that TERIS and other information services are telling physicians, and some physicians are repeating to their patients, that the use of dextromethorphan during pregnancy is perfectly safe. Unfortunately, I can think of no more irresponsible or careless claim than a claim of "no risk to your baby" when a risk exists. I urge that more epidemiologic and experimental studies be carried out. Pending their results, I ask Drs. Polifka, Shepard, and Brent to reconsider their reluctance to provide a well considered warning instead of a virtual endorsement for the use of this drug. If the absolute chance for the occurrence of a defect is increased by only 7/1000 or even 1/1000, women still should be given the opportunity to choose to avoid this risk.

Sincerely,

Thomas H. Rosenquist

College of Medicine; University of Nebraska Medical Center; Omaha, NE 68198