Oxendine v. Merrell Dow Pharmaceuticals
506 A.2d 1100 (D.C. Ct. App. 1986)

Terry, J.:

In this product liability case, appellant seeks reversal of an order of the trial court setting aside a jury verdict in her favor. She contends that there was sufficient evidence to send the case to the jury, and that the court therefore erred in entering a judgment notwithstanding the verdict. In addition, she argues that the court abused its discretion by ordering a new trial in the alternative because the verdict was not against the great weight of the evidence. We agree with both contentions, reverse the judgment, and remand the case with directions to reinstate the verdict and proceed to trial on appellant's claim for punitive damages, which has not yet been tried.


On January 25, 1971, appellant was born with a shortened right forearm and only three fingers on her right hand. Those fingers were fused together.

Appellant, through her parents, filed a complaint in the Superior Court on February 1, 1982, alleging that her birth defects were caused by her mother's use during pregnancy of a prescription drug manufactured by appellee. The drug, known as Bendectin, was designed to alleviate the nausea which commonly accompanies pregnancy. The complaint stated five counts against appellee, alleging negligence, breach of express warranty, breach of implied warranty, strict liability, and misbranding under the Federal Food, Drug, and Cosmetic Act. It sought ten million dollars in compensatory damages and ten million dollars in punitive damages.

After hearing the testimony of eighteen witnesses in the course of a three-week trial, the jury awarded appellant $750,000 in compensatory damages. Appellee then filed a motion for judgment notwithstanding the verdict or, in the alternative, for a new trial. The trial court granted both parts of the motion after a hearing, stating in an order the basis of its decision:

In support of her case to establish proximate cause, plaintiff relies on four principal grounds. The first is the structural activity of Bendectin which included an antihistaminic component, together with the awareness that certain antihistamines have been determined to be teratogenic in certain animals. Plaintiff also relies on the animal or in vivo studies. The third ground involves the in vitro studies performed at the National Institutes of Health. Finally, plaintiff relies on human epidemiological data.

It is clear to the Court from review of the evidence adduced at the trial of this action that no conclusion one way or another can be drawn from any of the above relied upon bases, respecting whether Bendectin is a human teratogen. And it is also clear from the evidence that plaintiff has failed to prove that use of Bendectin by her mother proximately caused her birth defect.

This appeal followed.


A judgment notwithstanding the verdict is proper only in "extreme" cases, in which no reasonable person, viewing the evidence in the light most favorable to the prevailing party, could reach a verdict in favor of that party. . . . We must be particularly cautious about setting aside jury verdicts in cases, such as this one, which present difficult medical issues of causation, with expert testimony going both ways.

Judges, both trial and appellate, have no special competence to resolve the complex and refractory causal issues raised by the attempt to link low level exposure to toxic chemicals with human disease. On questions such as these, which stand at the frontier of current medical and epidemiological inquiry, if experts are willing to testify that such a link exists, it is for the jury to decide whether to credit such testimony.

Ferebee v. Chevron Chemical Co., 237 U.S. App. D.C. 164, 169, 736 F.2d 1529, 1534, cert. denied,—U.S.—, 105 S. Ct. 545 (1984).

The transcript of the hearing on the motion for judgment n.o.v. helps to clarify what the trial court meant when it said in its order that "no conclusion" could be drawn from the evidence presented at trial. Dr. Alan Done, appellant's sole causation witness, had testified that no conclusion about Bendectin's effect on humans could be drawn from any of the four types of scientific data upon which he had principally relied when each type was considered separately from the others. The court focused on this fact and concluded that if each type of data, viewed in isolation, was not sufficient to prove that Bendectin caused birth defects, then all of them taken together could not prove it either. In so ruling, the trial court erred.


Dr. Done relied on four kinds of evidence in reaching his conclusion that Bendectin caused appellant's birth defects: (a) structure-activity information, (b) in vivo studies, (c) in vitro studies, and (d) epidemiological studies. We shall discuss each in turn.3

A. Structure-Activity Information

The first category of data upon which Dr. Done based his opinion involved what he termed the "structure-activity" of Bendectin. He explained that pharmacologists are "frequently able . . . to look at the structure [of a chemical compound] and predict what kind of activity that compound will have. And one of the jobs of pharmacology is to try to find that out, because that is how we design drugs to do certain jobs." Dr. Done said that the type of "activity" or function that a drug performs usually "has a bearing on the issue of . . . teratogenicity"; some antihistamines, for example, have been found to have teratogenic effects. One of the three components of Bendectin, doxylamine succinate, is an antihistamine. Dr. Done testified that it is a known teratogen for some animals and that it is suspected of being a human teratogen as well. . . .

b. in vivo studies

Dr. Done next turned to the in vivo (animal) studies relating to Bendectin. He first discussed a study entitled "Reproductive Study of

3. Dr. Done testified as an expert in the field of teratology, which is, broadly speaking, the study of birth defects and malformations. The admissibility of Dr. Done's testimony is not at issue in this case. Appellee contests its sufficiency to prove causation, but concedes that it was admissible.

Offspring of Bendectin-Treated Dams," which was conducted by appellee and reported by a Dr. Staples. The study was designed to discover whether Bendectin would cause birth defects in the offspring of female rabbits. Dr. Done quoted the pertinent findings and conclusions of Dr. Staples:

The limb alterations observed following Bendectin administration to female rabbits during gestation are considered to have been of no consequence in view of the mildness of the change and in view of its presence among the kits of controlled females.

The sternal changes noted involving shifting ossification [bone formation] centers by past experience could point to the possibility of severe alterations should the dosage be increased above the levels employed here.

This is of particular importance in view of the fact that that type of change was noted only at the highest dose administered.

This possibility can be answered only if further experimentation employing increased dosage is conducted. Such experimentation would at least provide additional information concerning the significance of the abortions seen in this study following Bendectin administration to the rabbit.

Dr. Done said that this study raised a "suspicion" of the teratogenicity of Bendectin which could be resolved only by further studies. Because appellee conducted no further studies, Dr. Done concluded that the Staples study must therefore be viewed as "positive." With regard to the "sternal changes" which Dr. Staples found, Dr. Done testified that the study could not be viewed as negative because the changes were found only at the highest dose used, which was not the highest dose possible. Teratologists, he said, use doses which fall just short of the amount which would start making the mothers ill. They then look for a dose-response relationship in which the number of effects increases with the dosage. Such a relationship helps to demonstrate that the effect is drug-related. The Staples study, however, was lacking the proper dose levels. Dr. Done also noted that the "abortions" mentioned by Dr. Staples (i.e., spontaneous abortions or miscarriages) could have occurred because "the babies were malformed and therefore couldn't survive. . . . "

Although the principal in vivo study on which Dr. Done relied was the one reported by Dr. Staples, he testified that two other in vivo studies strengthened his opinion that Bendectin is a teratogen. . . .

As with the other types of scientific data on which he relied, Dr. Done readily acknowledged that he could not conclude that Bendectin was a teratogen on the basis of the in vivo studies alone. In this regard he testified on cross-examination: . . .

Q. Now, in studies with human beings as distinguished from studies with animals, it is much more difficult to get perfect controls; is that correct?

A. Yes.

Q. And that is because animals can be kept in the same environment, given the same food and observed very closely; is that correct?

A. That is part of the reason, yes.

Q. You can control their breeding?

A. That's right.

Q. Whereas in human beings, they are much more different, much more difficult to control even if they are on a study; is that correct?

A. That's right.

Q. Doctor, isn't it a fact that you cannot determine teratogenicity from animal studies?

A. You can determine teratogenicity in the human being from animal studies.

Q. But there is no direct correlation between the two?

A. Well, there is a direct correlation . . . in the sense that no known human teratogen to my knowledge has failed to be teratogenic in at least one species of experimental animal.

Dr. Done plainly did not state that in vivo data are useless, as the trial court apparently believed. Dr. Done acknowledged that, in theory, human studies may be better than animal studies for predicting a drug's effect on humans, but he also pointed out that animal studies have many advantages over human studies because better controls are possible. Moreover, he specifically said that "[y]ou can determine teratogenicity in the human being from animal studies." Dr. Done did concede that "[t]he fact that [a substance] is teratogenic in a species of animal does not necessarily mean that it will be in humans as far as we know." The extent of that concession, however, was simply that one cannot conclude that a drug is a human teratogen on the basis of animal studies alone. Dr. Done did not attempt to draw such a conclusion. . . .

c. in vitro studies

Next, Dr. Done turned to the in vitro data on the teratogenicity of Bendectin. He explained that in vitro means "in glass," and that in vitro studies are those which are performed in a test tube. Such studies are particularly valuable, he said, because they allow scientists to focus on a specific effect. For example, to determine whether a drug can cause limb-shortening birth defects, limb bud cells of animal embryos (which eventually form the limbs of animals) can be separated from the rest of the embryo, placed in a test tube, and watched to see if their development is thwarted by the drug.

Dr. Done relied principally on an in vitro study performed by a Dr. Hassell. In that study Dr. Hassell observed that Bendectin interfered with the growth and development of limb bud cells. This observation led Dr. Done to conclude that the study "clearly indicates the potential for teratogenicity is there with regard to limb defects because the focusing that was done in this instance was to focus down on limb bud cells individually and grow them as pure things in culture. When it could be shown that Bendectin interfered with their growth and development, that, of course, indicates that the potential is there."

Considering the in vitro evidence in combination with the other data, Dr. Done testified that in his opinion Bendectin caused appellant's birth defects. . . .

d. epidemiological studies

Finally, Dr. Done turned to the epidemiological data. Epidemiology is the study of the frequency of disease in human populations and the way in which that frequency varies from one population to another. Like animal studies, epidemiological data are used to determine whether there is a positive correlation between the use of a substance and the appearance of any subsequent illness.

Dr. Done relied principally on a study conducted by two employees of appellee, Drs. Bunde and Bowles. Their study analyzed certain data provided by obstetricians in different parts of the United States and Canada concerning 2218 pairs of their women patients. Drs. Bunde and Bowles concluded that these data did not show a statistically significant association between Bendectin and birth defects.

Dr. Done reexamined the same data and came to a different conclusion. Initially he found fault with the methodology used by Drs. Bunde and Bowles, stating that it did not comport with accepted epidemiological standards. He also pointed out that there were inadequate controls because many women in the control group could have been exposed to Bendectin. Since Bendectin can be purchased without a prescription in Canada, the Canadian obstetricians reporting their findings may not have known whether some of the women in the control group had taken it. Dr. Done therefore excluded the Canadian pairs from the data. He also found numerous errors in the raw data and excluded more pairs to deal with those problems. In the end he calculated that the "relative risk" of the study was between 1.3 and 1.8, depending on whether certain data were Canadian or American. He explained the significance of these figures:

Relative risk just means the likelihood of getting a defective child in this case as a result of an association with the taking of Bendectin as opposed to not taking Bendectin.

1.8 means the woman is 80% more likely or almost twice as likely, in other words, to have a malformed baby if she is in the Bendectin group than in the control group.

Or if the figure is correct, 1.3, that means 30% greater likelihood. Whereas before, the original results suggested that she was 40% less likely to have an abnormal baby if she received Bendectin than if she didn't.

Dr. Done testified that the Bunde-Bowles study, taken together with the rest of the available data, supported his conclusion that Bendectin caused appellant's birth defects. He identified more than twenty other epidemiological studies which he had reviewed and analyzed in his own separate study of the literature on Bendectin and its components. He also referred to an internal memorandum of appellee entitled "Bendectin and Congenital Malformations," which reported that 48 percent of all Bendectin-related defects observed in a study were limb defects. Finally, Dr. Done noted that in a memorandum written by an employee of appellee, which was admitted into evidence over appellee's objection, there appears the statement that "Bendectin is not an obligatory teratogen as Thalid[omide] (for example)." Dr. Done explained that an obligatory teratogen is one which almost always causes birth defects, and that the author of this statement would not have made it unless he or she believed that Bendectin was a teratogen, but not as "flagrant" or severe as Thalidomide.

Dr. Done was on the witness stand for three and a half days. His testimony fills almost 600 pages of transcript. Although he was vigorously and exhaustively cross-examined by very able counsel, he did not waver from his opinion that Bendectin had caused appellant's birth defects. Throughout his testimony he repeatedly stated that this opinion was based not on any single study or type of evidence, but on four different types of scientific data viewed in combination. He conceded his inability to conclude that Bendectin was a teratogen on the basis of any of the individual studies which he discussed, but he also made clear that all of these studies must be viewed together and that, so viewed, they supported his conclusion. To argue on the basis of his concession, as appellee now does, that there was no factual support for his opinion simply ignores the rest of his testimony.

e. appellee's evidence

Dr. Brian MacMahon was appellee's principal witness on epidemiology and statistics. He testified that in evaluating the data on which an epidemiological study is based, one must first calculate the "relative risk" of the substance tested. Next, one must determine the "confidence interval" for the study, which sets the range within which the relative risk would fall if the study were repeated many times. Finally, one must look to see if the number "1" falls within the confidence interval. If so, the results of the study are essentially meaningless; if not, the study shows a statistically significant relationship between the substance tested and the effects observed.

Dr. MacMahon then reviewed a number of epidemiological studies on Bendectin. Many of them had a relative risk higher than 1, which meant that the risk of having a baby with a birth defect was higher for those mothers who took Bendectin than those who did not. Nevertheless, Dr. MacMahon believed that these studies were meaningless because the number "1" fell within the confidence interval of each of the studies. He therefore concluded that Bendectin was not responsible for the birth defects observed in the studies and, moreover, that it did not cause appellant's birth defects. Appellee relies principally on Dr. MacMahon's testimony and on similar statements by its other expert witnesses in arguing that causation cannot be established without evidence of a statistically significant association between Bendectin and birth defects. There are at least three reasons why this argument lacks merit.

First, Dr. MacMahon conceded that there were in fact several studies showing statistically significant associations between Bendectin and various types of birth defects. In a study by a Dr. Cordero, for example, Dr. MacMahon noted that the relative risk for one type of limb reduction effect was statistically significant. The relative risk was 3.88, and the confidence interval did not include the number "1." Other defense witnesses also testified to such statistically significant associations. Therefore, even assuming that a study must be deemed meaningless whenever the number "1" falls within its confidence interval, the jury was made aware of several meaningful studies on which it could base a finding of causation.

Second, appellant's rebuttal witness, Dr. Shanna Swan, an expert in biostatistics and epidemiology, testified that the mere fact that the number "1" falls within the confidence interval does not invalidate the study. She cited as an example an epidemiological study by a Dr. Smithells, in which the confidence interval was from 0.71 to 2.61, and the relative risk was 1.36. Dr. Swan testified that the relative risk calculation is the "best estimate of the true situation based on [these] data. . . . " A relative risk of 1.36 means that women who take Bendectin have a 36 percent higher risk of having a baby with a birth defect than those who do not. Confidence intervals, she said, are used to predict the range in which the relative risk would fall if the study were repeated several times. Thus, in the Smithells study, "the data [are] consistent with the relative risk being as high as 2.61," which means that women who took Bendectin would have a 161 percent higher risk of having a baby with a birth defect than those who did not take Bendectin. Dr. Swan's testimony thus contradicted that of Dr. MacMahon on the question of whether many of the studies which he rejected could be relied upon as evidence that women exposed to Bendectin had a higher risk of giving birth to babies with birth defects. That contradiction was properly left to the jury to resolve.

Finally, Dr. MacMahon conceded that statisticians must be careful to weigh other evidence and not to apply their rules rigidly and blindly. On cross-examination, he agreed with Dr. Byron Brown, whom he recognized as an authority on the subject:

Q. Do you agree with [Dr. Brown], if the scientist is to go further than reporting his set of data, discussing it as fully as possible and relating it to the body of current knowledge, if he is to regard himself as a decision maker, then he must consider more than error rates? Do you agree with that?

A. Yes.

Q. He must weigh past evidence in with the data he has accumulated and he must weigh the penalties or costs involved in making the various types of error. Do you agree with that?

A. Yes.

Q. And Doctor, one of the things you have to consider, then, when you come up with your statistics is what happens if you are wrong. You have to consider that risk, don't you, as a biostatistician?

A. Yes.

Q. And if the studies that show a relative risk as Smithells did at 36 percent greater chance of a defect, Mitchell, 50 percent, and the other ones that we referred to were right with that relative risk and you are wrong saying they are statistically insignificant, and Bendectin is in fact a teratogen, then the risk is that children are going to be born with limb defects of mothers who take Bendectin, isn't that right?

A. That's very hypothetical, but yes.

F. Conclusion

Although the trial in this case was long and the evidence complex, the issue before the jury was a straightforward one: did Bendectin cause appellant's birth defects? Expert witnesses testified at length on both sides of that issue. Not surprisingly, their testimony revealed a disagreement as to how the epidemiological and other data should be interpreted. "The case was thus a classic battle of the experts, a battle in which the jury must decide the victor." Ferebee v. Chevron Chemical Co., supra, 237 U.S. App. D.C. at 170, 736 F.2d at 1535 (citation omitted). The trial court, however, granted appellee's motion for judgment notwithstanding the verdict on the ground that appellant's causation expert admitted in his testimony that each of the studies on which he relied could not, by itself, support a finding of causation. Where the court erred was in failing to consider the same expert's testimony that all of the studies, taken in combination, did support such a finding, as he carefully and repeatedly explained.

In ruling on a motion for judgment n.o.v., the court must view the evidence as a whole, not in fragments. Like the pieces of a mosaic, the individual studies showed little or nothing when viewed separately from one another, but they combined to produce a whole that was greater than the sum of its parts: a foundation for Dr. Done's opinion that Bendectin caused appellant's birth defects. The evidence also established that Dr. Done's methodology was generally accepted in the field of teratology, and his qualifications as an expert have not been challenged. Because he fully explained to the jury that his opinion was based on all of the studies, we cannot say that no reasonable juror could reach a verdict in favor of appellant. Hence we must reverse the order granting a judgment n.o.v.


Appellant also contends that the court abused its discretion by ordering a new trial in the alternative on the ground that the verdict was against the weight of the evidence. We agree. . . . Although the court did not explain its ruling further, it appears to have been based on the same reasoning that led to the court's flawed ruling on the motion for judgment n.o.v., namely, that Dr. Done's testimony must be discounted because he could not base a finding of teratogenicity on any of the individual studies on which he relied. Having given the court's new trial order the close scrutiny which the case law commands, we are convinced that the court erred in concluding that the verdict was against the weight of the evidence. . . .

It is clear from a review of the record that the verdict was not against the weight of the evidence. There was evidence on both sides of nearly every issue, and it was fairly evenly weighted. Taking all the testimony into account, we hold that the trial court abused its discretion in granting a new trial on the ground that the verdict was against the weight of the evidence. . . .


div1.gif (1531 bytes)
Home | Contents | Topical Index | Syllabi | Search | Contact Us | Professors' Pages
Cases | Problems | Rules | Statutes | Articles | Commentary