Correspondence
Teratogenicity of High Vitamin A Intake
N Engl J Med 1996; 334:1195-1197May 2, 1996
- Article
To the Editor:
The article by Rothman et al. (Nov. 23 issue)1 followed our study, in which we identified a possible increase in the risk of certain birth defects associated with vitamin A supplementation, but information on dose was not available.2 We are pleased that their study examined this relation according to the dose of retinol, but the methods used raise several questions.
Information on pregnancy outcomes was obtained from the delivering physician or the mother; both are considered poor sources of information on birth-defect diagnoses, but we cannot evaluate the accuracy of ascertainment without knowing the range and distribution of abnormalities. The authors used a case-classification approach, like ours,2 which groups together defects in structures that had some embryologic contribution from cranial-neural-crest cells. However, the application of this approach was flawed, resulting in the misclassification of outcomes. Malformations of the central nervous system, eye, and heart were included, but experimental studies have not shown a contribution of cranial-neural-crest cells to these abnormalities, except for malformations affecting the meninges, ocular pigmentary cells, the cardiac outflow tract, and the mesenchyme surrounding branchial-arch arteries.3-5
Rather than excluding the 201 pregnancies with missing information on retinol supplementation from the analysis, the authors chose to estimate the missing doses. However, the doses were imputed for 39 percent of the women classified as having intakes of more than 10,000 IU of retinol per day, as compared with less than 1 percent of those assigned to a lower dose category. Does the disproportion arise from the assignment of a starting date for the prenatal vitamins that preceded the discontinuation of multivitamins, thereby assuming the simultaneous use of two vitamin supplements? Typically, pregnant women stop taking a multivitamin the day they start taking prenatal vitamins, suggesting that the authors overestimated the dose.
The finding of an increased risk of “cranial-neural-crest outcomes” is based on seven cases in the intake category of more than 10,000 IU of retinol daily, but there is a possibility that exposure, diagnosis, or both were misclassified for at least four cases (two with missing data on vitamin intake and four with unclear or inappropriate diagnoses). One might argue that random misclassification would bias the relative risks toward a finding of no effect, but with so few exposed cases the predictability of misclassification bias is less reliable. Furthermore, the possibility of differential misclassification of outcomes may bias risk estimates away from the null finding.
Apart from other considerations, these weaknesses alone make it inappropriate to attempt to identify a threshold dose of vitamin A that is associated with an increased frequency of birth defects. High-dose vitamin A may increase the risk of certain malformations, but we do not believe that the results of this study alone provide sufficient evidence of a dose effect for clinicians and the public to be warned of the dangers of any vitamin A dose greater than 10,000 IU.
5 ReferencesMartha M. Werler, Sc.D.
Boston University School of Medicine, Boston, MA 02146Edward J. Lammer, M.D.
Children's Hospital of Oakland, Oakland, CA 94609Allen A. Mitchell, M.D.
Boston University School of Medicine, Boston, MA 021461
Rothman KJ ,Moore LL ,Singer MR ,Nguyen UDT ,Mannino S ,Milunsky A . Teratogenicity of high vitamin A intake. N Engl J Med 1995;333:1369-1373
Full Text | Web of Science | Medline2
Werler MM ,Lammer EJ ,Rosenberg L ,Mitchell AA . Maternal vitamin A supplementation in relation to selected birth defects. Teratology 1990;42:497-503
CrossRef | Medline3
Johnston MC ,Noden DM ,Hazelton RD ,Coulombre JL ,Coulombre AJ . Origins of avian ocular and periocular tissues. Exp Eye Res 1979;29:27-43
CrossRef | Web of Science | Medline4
Kirby ML ,Waldo KL . Role of neural crest in congenital heart disease. Circulation 1990;82:332-340
CrossRef | Web of Science | Medline5
Le Douarin N. The neural crest. Cambridge, England: Cambridge University Press, 1982.
To the Editor:
We are concerned that the paper by Rothman et al. can be misinterpreted and therefore be more harmful than helpful. The authors conclude that a pregnant woman who ingests more than 10,000 IU of preformed vitamin A (in the form of retinol and retinyl esters) daily from supplements has an increased risk of having a malformed infant. Our main objection is to their classification of malformations related to abnormal migration of cranial-neural-crest cells, presumably caused by fetal exposure to an excessive amount of preformed vitamin A. For example, we would not have assigned all heart defects to this group. Their classification of defects could have reflected the findings in human infants and monkey infants exposed to 13-cis-retinoic acid (isotretinoin, Accutane),1,2 which other studies in animals have shown to be similar to the fetal effects of excess amounts of preformed vitamin A.3 Unfortunately, there was no indication that the authors considered the distinctive nature of the 13-cis-retinoic acid–associated malformations of the ear, heart (predominantly conotruncal defects), cerebellum (absence or hypoplasia of the vermis), and thymus. The reclassification of a small number of the infants with malformations might substantially change the observations in the study by Rothman et al.
3 ReferencesRobert L. Brent, M.D., Ph.D., D.Sc.
Alfred I. Dupont Institute, Wilmington, DE 19899Andrew G. Hendrickx, Ph.D.
University of California, Davis, Davis, CA 95616Lewis B. Holmes, M.D.
Massachusetts General Hospital, Boston, MA 02114Richard K. Miller, Ph.D.
University of Rochester Medical Center, Rochester, NY 14642-86681
Lammer EJ ,Chen DT ,Hoar RM , et al. Retinoic acid embryopathy. N Engl J Med 1985;313:837-841
Full Text | Web of Science | Medline2
Hummler H ,Korte R ,Hendrickx AG . Induction of malformations in the cynomolgus monkey with 13-cis retinoic acid. Teratology 1990;42:263-272
CrossRef | Medline3
Teratology Society position paper: recommendations for vitamin A use during pregnancy
CrossRef
To the Editor:
Rothman et al. report that daily consumption by pregnant women of supplemental vitamin A in doses exceeding 10,000 IU results in an increased risk of birth defects in sites derived from cranial-neural-crest cells. Several methodologic considerations suggest the need for cautious interpretation of these results.
First, the designation of 10,000 IU as the threshold for increased risk may be misleading. The 317 women with vitamin A intakes exceeding 10,000 IU per day had a mean daily intake of 21,675 IU. The authors do not report the actual levels consumed by the mothers of the seven infants with cranial-neural-crest abnormalities. Depending on what those levels were, the true threshold for teratogenicity may be much higher than 10,000 IU. We would like to know the exposure levels in these seven pregnancies and see them plotted on Figure 1 of the article.
Second, the study's low overall rate of birth defects (1.5 percent, excluding chromosomal defects) suggests that not all birth defects were ascertained. In fact, the birth-defect rate in the group with high vitamin A intake (3 percent) approximates the generally accepted background rate of 3 to 4 percent.1
Third, the authors' broad definition of cranial-neural-crest defects, including all defects in organ systems that have some contribution from cranial-neural-crest cells, is open to question. This classification system implies pathogenetic homogeneity, but it results in a very heterogeneous group of defects, some of which are in sites unlikely to be of true cranial-neural-crest origin. Also, the classification system used for infants with multiple defects favors cranial-neural-crest–derived defects and could have obscured an elevated risk of other defects. We would like to see a detailed phenotypic description of the infants who were included in the group with cranial-neural-crest defects, with further clarification of nonspecific terms such as “multiple heart defects,” characterization of patterns of defects in infants with multiple anomalies, and discussion of the existence of other conditions that may have contributed to the defect (e.g., prematurity and hydrocephalus).
Although women of reproductive age need to remain cautious about taking high doses of vitamin A, we believe that these methodologic issues raise doubts about the authors' conclusion that a dose of 10,000 IU per day is the threshold for vitamin A teratogenicity. This conclusion may generate unjustified fear about dietary supplements and may discourage women of childbearing age from using multivitamin supplements that can prevent neural-tube defects and perhaps other defects. In addition, some women who are concerned about their “exposure” to more than 10,000 IU of vitamin A per day may choose, perhaps needlessly, to terminate a pregnancy.
1 ReferencesMargaret Watkins, M.P.H.
Cynthia Moore, M.D.
Joseph Mulinare, M.D., M.S.P.H.
Centers for Disease Control and Prevention, Atlanta, GA 30341To the Editor:
Rothman et al. concentrated on vitamin A excesses, not deficiencies. Yet there is evidence that vitamin A deficiency also leads to birth defects in both animals and humans.1 Rothman et al. showed that the vast majority of the 22,748 women in the study consumed safe levels of vitamin A, with the implicit definition of “safe” being no more than approximately twice the recommended dietary allowance (RDA). Unfortunately, they failed to analyze how many women actually consumed less than the RDA for vitamin A. If they had done so, they might have clarified the effects of a low intake of vitamin A on teratogenicity, as well as identified the optimal dose for pregnant women.
The greater public health problem may be not a small number of cases of vitamin A overdose, but widespread vitamin A deficiency. Although the latest edition of the Recommended Dietary Allowances describes vitamin A deficiency as rare in the United States,2 some evidence suggests that the opposite may be true — that is, that potentially large numbers of Americans may consume levels of vitamin A below the RDA.3 Again, Rothman et al. might have clarified the extent of these deficiencies by including lower levels of vitamin A in their stratification categories.
3 ReferencesJack J. Challem
, Aloha, OR 970061
Bendich A ,Langseth L . Safety of vitamin A. Am J Clin Nutr 1989;49:358-371
Web of Science | Medline2
National Research Council. Recommended dietary allowances. 10th rev. ed. Washington, D.C.: National Academy Press, 1989.
3
Block G . The data support a role for antioxidants in reducing cancer risk. Nutr Rev 1992;50:207-213
CrossRef | Web of Science | Medline
To the Editor:
In the United States, there is very little risk of vitamin A deficiency. According to the latest Department of Agriculture food-consumption survey,1 women of childbearing age consume on average over 5000 IU of vitamin A daily (more than 120 percent of the RDA2) from food alone. This figure may underestimate actual consumption by 20 to 25 percent, since it is based on a reported energy consumption of only 1650 kcal. As indicated by the Food and Nutrition Board's report, supervised supplementation may be desirable for some groups of pregnant women, such as recent immigrants from countries in which vitamin A deficiency is endemic. However, with the much greater risk of harm than of benefit, the medical and nutrition community should be encouraging the removal of preformed vitamin A from over-the-counter nutrient supplements, or at least the replacement of preformed vitamin A with carotene. Although there is general scientific acceptance of the need to ensure the consumption of at least 400 μg of folic acid daily from foods, supplements, or both by women of childbearing age to help prevent neural-tube defects, this goal is not a sound basis for accepting supplements that contain up to 8000 IU of vitamin A, as suggested by Oakley and Erickson in their editorial.3
3 ReferencesJanet R. Hunt, Ph.D., R.D.
U.S. Department of Agriculture, Grand Forks, ND 58202-90341
Tippett KS, Mickle SJ, Goldman JD, et al. Food and nutrient intakes by individuals in the United States, 1 day, 1989-91. Washington, D.C.: Department of Agriculture, 1995. (NFS report no. 91-2.)
2
National Research Council. Recommended dietary allowances. 10th rev. ed. Washington, D.C.: National Academy Press, 1989.
3
Oakley GP Jr ,Erickson JD . Vitamin A and birth defects -- continuing caution is needed. N Engl J Med 1995;333:1414-1415
Full Text | Web of Science | Medline
Author/Editor Response
The authors reply:
To the Editor: Without access to our data, Werler et al. claim that 39 percent of the subjects in our highest intake category had imputed doses. They are wrong. In this category, we used imputed doses to categorize only 4 percent of subjects and none of those with affected infants; as we reported, our analyses were unchanged after we excluded subjects with imputed doses. Werler et al., as well as Brent et al. and Watkins et al., also imply that misclassification of birth defects arising in cranial-neural-crest cells may explain the nearly fivefold increase in the prevalence of these defects. Existing evidence is not nearly so precise about the classification as they imply.1 More important, misclassification of outcomes that is unrelated to exposure cannot produce a spurious effect if none exists in the first place.2 Furthermore, we reported a large increase in the prevalence of total birth defects, a finding that could not possibly result from the erroneous classification of defects arising in cranial-neural-crest cells.
Watkins et al. worry that our estimated threshold dose of vitamin A was too low because women in the highest supplemental-dose category who had infants with malformations could have had daily doses well above 10,000 IU; this point was also raised by Oakley and Erickson in their editorial.3 Our dose–response curve, however, was fitted through the actual data points, not the midpoints of the categories. Furthermore, we reported a 2.6-fold increase in cranial-neural-crest defects among infants born to women taking 8001 to 10,000 IU of retinol supplements daily, a finding that refutes the concern that the effect occurs only at much higher doses. Table 1Table 1
Number and Percentage of Pregnancies Resulting in Birth Defects and Prevalence Ratios, According to the Daily Intake of Retinol from Supplements. gives the prevalence of defects after the intake category of 10,001 IU or more of supplements daily is split into two subcategories. Despite the small numbers, the effect appears about as strong within the range from 10,001 to 20,000 IU as it is above 20,000 IU. An extended spline regression curve also indicates that the effect plateaus above 20,000 IU daily.Watkins et al. also question an apparently low overall prevalence of birth defects in our population. All our comparisons were internal to our study, however, and nondifferential underascertainment of cases — as with errors in categorization — could not produce a spurious effect if none existed.2
Challem, Watkins et al., and Khoury et al.4 have expressed concern that our findings will undermine the beneficial role of vitamin supplementation in pregnancy. We hope not. We investigated the teratogenicity of retinol, not the costs and benefits of multivitamins. Concern with the broader issue is no reason to deny the possible teratogenic effects of retinol supplements.
4 ReferencesKenneth J. Rothman, Dr.P.H.
Lynn L. Moore, D.Sc.
Martha R. Singer, M.P.H., R.D.
Aubrey Milunsky, M.B., B.Ch., D.Sc.
Boston University School of Medicine, Boston, MA 021181
Kirby ML ,Waldo KL . Role of neural crest in congenital heart disease. Circulation 1990;82:332-340
CrossRef | Web of Science | Medline2
Rothman KJ. Modern epidemiology. Boston: Little, Brown, 1986.
3
Oakley GP Jr ,Erickson JD . Vitamin A and birth defects -- continuing caution is needed. N Engl J Med 1995;333:1414-1415
Full Text | Web of Science | Medline4
Khoury MJ ,Moore CA ,Mulinare J . Vitamin A and birth defects. Lancet 1996;347:322-322
Web of Science | Medline
- Citing Articles (6)
Citing Articles
1
Richard D. Semba. (2002) Safety of daily oral vitamin A supplementation for individuals with retinitis pigmentosa. Annals of Ophthalmology 34:3, 194-198
CrossRef2
Guy B. Mulder, Nancy Manley, John Grant, Karen Schmidt, Weiping Zeng, Christian Eckhoff, Lillian Maggio-Price. (2000) Effects of excess vitamin A on development of cranial neural crest-derived structures: A neonatal and embryologic study. Teratology 62:4, 214-226
CrossRef3
H Dolk. (1999) Dietary vitamin A and teratogenic risk: European Teratology Society discussion paper. European Journal of Obstetrics & Gynecology and Reproductive Biology 83:1, 31-36
CrossRef4
P. Mastroiacovo, T. Mazzone, A. Addis, E. Elephant, P. Carlier, T. Vial, H. Garbis, E. Robert, M. Bonati, A. Ornoy, A. Finardi, C. Schaffer, L. Caramelli, E. Rodr
guez-Pinilla, M. Clementi. (1999) High vitamin A intake in early pregnancy and major malformations: A multicenter prospective controlled study. Teratology 59:1, 7-11
CrossRef5
Lennart Dencker, Per Eriksson. (1998) Susceptibility in utero and upon neonatal exposure. Food Additives and Contaminants 15:sup001, 37-43
CrossRef6
James L. Mills, Joe Leigh Simpson, George C. Cunningham, Mary R. Conley, George G. Rhoads. (1997) Vitamin A and birth defects. American Journal of Obstetrics and Gynecology 177:1, 31-36
CrossRef
