Hernandez.qxd

Copyright 2001 by The Johns Hopkins University School of Hygiene and Public Health NTDs and Folic Acid Antagonists during Pregnancy Hernández-Díaz et al.
Neural Tube Defects in Relation to Use of Folic Acid Antagonists during
Pregnancy
Sonia Hernández-Díaz,1,2 Martha M. Werler,1 Alexander M. Walker,2 and Allen A. Mitchell1 Periconceptional folic acid supplementation reduces the risk of neural tube defects (NTDs). To determine whether periconceptional exposure to folic acid antagonists (FAAs) might therefore increase the risk of NTDs,the authors examined data from an ongoing case-control study of birth defects (1979–1998) in the United Statesand Canada. They compared data on 1,242 infants with NTDs (spina bifida, anencephaly, and encephalocele)with data from a control group of 6,660 infants with malformations not related to vitamin supplementation.
Mothers were interviewed within 6 months of delivery about demographic, reproductive, medical, and behavioralfactors and about medication use. The adjusted odds ratios of NTDs related to exposure to FAAs (includingcarbamazepine, phenobarbital, phenytoin, primidone, sulfasalazine, triamterene, and trimethoprim) during thefirst or second months after the last menstrual period, compared with no use in either month, were 2.8 (95%confidence interval: 1.7, 4.6) for FAAs as a group, 4.8 (95% confidence interval: 1.5, 16.1) for trimethoprim(based on five exposed cases), and 6.9 (95% confidence interval: 1.9, 25.7) for carbamazepine (six exposedcases). These results are adjusted for region, interview year, periconceptional folic acid supplementation,maternal age, weight, education, and infections early in pregnancy. These findings suggest that a number ofFAAs may increase NTD risk, and they provide estimates of risk for selected drugs. Am J Epidemiol2001;153:961–8.
carbamazepine; case-control studies; folic acid antagonists; neural tube defects; pregnancy; teratogens;trimethoprim Neural tube defects (NTDs), characterized by a failure of as a group nor for specific FAAs such as trimethoprim. For the neural tube to close properly after conception, affect carbamazepine, the magnitude of risk is unclear. The pur- about one per 1,000 livebirths in the United States (1).
pose of the present study was to evaluate the risk of having Observational and experimental studies have shown that an infant with an NTD after periconceptional exposure to periconceptional folic acid supplementation can significantly FAAs in general and, where numbers permitted, to specific reduce the risk of NTDs (2, 3). These findings, together with the observation that low maternal red cell folate in earlypregnancy is related to NTDs (4, 5), suggest that folic acid MATERIALS AND METHODS
antagonists (FAAs) may increase the risk of NTDs.
FAAs include aminopterin, carbamazepine, methotrexate, Study population
phenobarbital, phenytoin, primidone, sulfasalazine, tri-amterene, trimethoprim, and valproic acid. Although it has Since 1976, the Slone Epidemiology Unit Birth Defects been documented or suggested that NTD risk is associated Study has been interviewing mothers of malformed children with use in early pregnancy of certain specific FAAs born in the greater metropolitan areas of Boston, (aminopterin (6), methotrexate (7), valproic acid (8, 9), and Massachusetts, Philadelphia, Pennsylvania, Toronto, carbamazepine (8, 10)), risk has not been studied for FAAs Canada, and, between 1983 and 1985, part of the state ofIowa (11). Until 1988, study subjects included only liveborn Received for publication April 21, 2000, and accepted for publica- infants ascertained at birth or tertiary care hospitals. Since then, stillborn infants and fetuses therapeutically aborted Abbreviations: CI, confidence interval; FAA, folic acid antagonist; because of a malformation have also been included. Study subjects are identified through review of admissions and dis- 1 Slone Epidemiology Unit, Boston University School of Public charges at major referral hospitals and clinics and through 2 Department of Epidemiology, Harvard School of Public Health, regular contact with newborn nurseries in community hos- pitals (the latter to identify infants born with defects who Reprint requests to Sonia Hernández-Díaz, Slone Epidemiology might not have been referred to major centers). Since 1993, Unit, Boston University School of Public Health, 1371 Beacon Street, a random sample of nonmalformed infants has been identi- Brookline, MA 02446 (e-mail: [email protected]).
Presented in part at the 15th International Conference on fied at the birth hospitals as potential controls. Physicians of Pharmacoepidemiology, Boston, MA, August 1999.
study subjects confirm the diagnosis. Among eligible sub- jects, the mothers of 84 percent of cases, 83 percent of mal- group were hypertrophic pyloric stenosis (n ϭ 906), inde- formed controls, and 80 percent of nonmalformed controls terminate sex or pseudohermaphroditism (n ϭ 509), muscu- contacted agreed to an interview and to provide informed loskeletal anomalies of the skull or face (n ϭ 387), feet consent. The present study is based on subjects ascertained deformities (n ϭ 296), anomalies of the diaphragm (n ϭ 282), gastroschisis/omphalocele (n ϭ 257), esophagealstenosis (n ϭ 219), stenosis of the large intestine or anus (n ϭ 216), congenital dislocation of the hip (n ϭ 179),hypospadias (n ϭ 150), and others. We also selected two Cases consisted of 1,242 infants and fetuses with a diag- secondary series of controls, one of nonmalformed infants nosis of anencephaly, spina bifida, encephalocele, or other (n ϭ 1,626) and one of infants with chromosomal or NTDs (table 1). The relative prevalence of anencephaly and Mendelian anomalies (n ϭ 2,138).
other NTDs is not representative, mainly because stillbirthsand abortions were included later in the study. Infants with Exposure assessment
chromosomal or Mendelian-inherited anomalies (n ϭ 44) orwith amniotic bands, caudal regression, or twin disruption Within 6 months of the subject’s delivery, trained study (n ϭ 15) were excluded under the assumption that the eti- nurses unaware of the hypothesis interviewed mothers of ologies of their NTDs were different from those of the cases and controls. More than 90 percent of interviews were remaining cases. NTDs complicated with other defects (but conducted in the subjects’ homes, a setting selected to min- not as part of an identified chromosomal or Mendelian- imize anxiety, to encourage cooperation, and to maximize inherited syndrome) were included in the general analysis data collection; the remaining interviews took place at an alternative site of the mother’s choice (8 percent) or by tele-phone (1 percent). The interview was detailed and structured Controls
and included questions on demographic characteristics,patients’ medical history, an obstetric history, parents’ habits Because mothers of malformed and mothers of nonmal- and occupations, and a detailed history of the use of med- formed infants may differ in their recall of antenatal drug ication (both prescription and over-the-counter) from 2 exposure (12), our primary control group consisted of months before conception through the entire pregnancy.
infants with malformations other than NTDs. Among the Recall of medication exposures was enhanced by questions 13,666 infants with structural malformations, we excluded regarding indications for use. Recall of the timing of use 7,006 subjects with oral clefts, urinary tract defects, limb was enhanced by using a calendar that highlighted the date reduction, heart defects, and conditions related to NTD of the woman’s last menstrual period. Completed question- (hydrocephalus, microcephalus, and other anomalies of the naires were submitted to a variety of quality control proce- brain, spinal cord, or nervous system), because folic acid dures and entered into a computer file.
may play a role in the genesis of these malformations (13).
Mothers were considered exposed if they reported using The final control group consisted of 6,600 subjects. The an FAA any time during the 2 months after the last men- major types of primary diagnosis included in the control strual period; these 2 months encompass the period of neuraltube development (14).
Data analysis
Selection of cases and controls, Slone
Epidemiology Unit Birth Defects Study, United States and
Prevalence odds ratios of having an infant with an NTD Canada, 1976–1998
and 95 percent confidence intervals were estimated for women exposed versus women not exposed to an FAA, using unconditional logistic regression. The associations presented below were adjusted, using multivariate models, for geographic region, interview year (1976–1982, 1983–1987, 1988–1992, 1993–1998), maternal age (≤24, 25–29, 30–34, ≥35 years), periconceptional folic acid sup- plementation (during the 2 months after the last menstrualperiod—never, occasionally, or daily), prepregnancy weight, years of education, and infections during the first trimester. Folic acid supplementation was also considered as * Neural tube defects can be isolated or complicated with other The distributions of selected characteristics are shown for † Malformed controls included defects other than neural tube defects, oral clefts, urinary tract defects, limb reductions, heart cases and malformed controls in table 2. Prevalences of any defects, and conditions related to neural tube defects.
FAA use and specific FAA use are presented for cases and NTDs and Folic Acid Antagonists during Pregnancy Distribution of selected characteristics, Slone
Continued
Epidemiology Unit Birth Defects Study, United States and
Canada, 1976–1998
* p value < 0.05; ** p value ≤ 0.001.
† One pound (lb) = 0.45 kg.
‡ UTI, urinary trace infection; STD, sexually transmitted disease; § Anytime during first 2 months after the last menstrual period.
women exposed to FAAs compared with women not exposed was 2.8 (95 percent confidence interval (CI): 1.7, Table continues
4.6) (table 3). When controlled for potential confounders,the findings did not differ appreciably from the crude analy-sis; therefore, only adjusted data are presented. Control for malformed controls in table 3. Among the 1,242 infants with other factors (e.g., maternal smoking during pregnancy, NTDs, 27 (2.2 percent) were exposed to FAAs early in preg- alcohol intake, previous affected pregnancy, family history, nancy, while 67 (1.0 percent) were exposed among the 6,660 birth order, or pregnancy planning) did not materially influ- malformed controls. The estimated adjusted odds ratio for Use of FAAs* during early pregnancy in 1,242 cases and in 6,660 malformed controls, along
with prevalence odds ratios of neural tube defects and their 95% confidence intervals, associated with
use of FAAs in early pregnancy, Slone Epidemiology Unit Birth Defects Study, United States and
Canada, 1976–1998
* FAAs, folic acid antagonists; CI, confidence interval.
† Used alone or in combination with other agents; categories are not mutually exclusive.
‡ Adjusted for interview year, region, periconceptional folic acid supplementation, and maternal age, educa- tion, weight before pregnancy, and urinary tract infections or other infections early in pregnancy.
§ Two cases and one control were treated with a combination of carbamazepine and primidone.
¶ —, fewer than five cases or five controls exposed.
The odds ratio for NTDs complicated with other defects after the last menstrual period, and 0.9 (95 percent CI: 0.1, (n ϭ 506) was 3.1 (95 percent CI: 1.5, 6.2). For NTD type, 10.7) for the third month after the last menstrual period odds ratios were 6.4 (95 percent CI: 2.7, 15.3) for anen- (two, five, four, and one exposed cases, respectively).
cephaly, 2.7 (95 percent CI: 1.6, 4.7) for spina bifida, and To assess whether folic acid supplementation modifies 0.0 (95 percent CI: 0.0, 2.5) for encephalocele (there were the risk associated with carbamazepine and trimethoprim no exposed cases). Restriction of the analysis to infantswithout a family history of NTDs yielded an odds ratio of5.8 (95 percent CI: 2.5, 13.7).
Odds ratios for specific FAAs varied considerably; to minimize statistical instability, we limited estimating risk tothose specific FAAs for which there were at least fiveexposed cases: carbamazepine and trimethoprim.
In our population, 0.5 percent (n ϭ 6) of cases and 0.1 percent (n ϭ 5) of controls used carbamazepine (asmonotherapy in four cases and three controls) during thefirst 2 months after the last menstrual period, for an adjustedodds ratio of 6.9 (95 percent CI: 1.9, 25.7) (odds ratio wassimilar for monotherapy). Because carbamazepine was usedthroughout pregnancy, we were unable to consider effectsaccording to month of exposure.
Overall, 0.4 percent (n ϭ 5) of cases and 0.1 percent (n ϭ 8) of controls used trimethoprim around the time of concep-tion, yielding an odds ratio of 4.8 (95 percent CI: 1.5, 16.1).
The prevalence of trimethoprim exposure according to lunarmonth of pregnancy is presented in figure 1. Trimethoprimexposure among cases was significantly higher during thefirst 2 months after the last menstrual period but not duringthe month preceding or following that period. The NTD riskfor women exposed to trimethoprim during each lunar FIGURE 1.
Trimethoprim exposure (%), according to lunar months, month, compared with women not exposed to trimethoprim in neural tube defect cases and malformed controls, SloneEpidemiology Unit Birth Defects Study, United States and Canada, during these months, was 1.3 (95 percent CI: 0.3, 5.6) for 1976–1998. The adjusted odds ratio estimate and its 95% confi- the month before the last menstrual period, 7.8 (95 percent dence interval are presented for each month around the last men- CI: 2.2, 27.0) for the first month after the last menstrual strual period. NTD, neural tube defect; LMP, last menstrual period; period, 6.4 (95 percent CI: 1.5, 26.3) for the second month OR, odds ratio; CI, confidence interval.
NTDs and Folic Acid Antagonists during Pregnancy during the 2 months after the last menstrual period, we com- If the associations were causal, for drugs with odds ratios pared use and nonuse of these drugs for users and nonusers as high as 5 or 7 the excess number of NTDs would be four of folic acid supplementation (table 4). In the absence of or six per 1,000 pregnant women exposed to these drugs carbamazepine and trimethoprim exposure, and compared early in pregnancy (assuming a prevalence of one NTD per with women who took neither these drugs nor folic acid dur- 1,000 births among women not exposed) (1).
ing the 2 months after the last menstrual period, daily folicacid supplementation was protective (odds ratio ϭ 0.7; 95 DISCUSSION
percent CI: 0.5, 0.8). Women exposed to either trimethoprimor carbamazepine and not taking folic acid had an odds ratio Exposure to FAAs as a group during the first or second of 13.3 (95 percent CI: 2.9, 61.4), whereas women taking months after the last menstrual period more than doubled one of these medications and daily folic acid had an odds the risk of NTDs. NTD risk increased more than sixfold ratio of 1.2 (95 percent CI: 0.1, 12.7).
after carbamazepine exposure and more than fourfold after All trimethoprim-exposed subjects took a combination of trimethoprim exposure. For trimethoprim, the effect was trimethoprim and a sulfonamide, either sulfamethoxazole maximal during or shortly before the time of neural tube (n ϭ 11) or sulfadiazine (n ϭ 2). There were no exposures to closure; in contrast, exposure before or after this sensitive sulfamethoxazole or sulfadiazine alone, so we could not period did not increase the risk of NTDs. As would be assess their separate effects. However, for 11 cases and 42 expected, folic acid supplementation in the absence of car- controls exposed to sulfonamides (other than the FAA sul- bamazepine and trimethoprim reduced the risk of NTDs; fasalazine), the odds ratio was 1.4 (95 percent CI: 0.7, 2.9).
though based on small numbers, such supplementation also Because trimethoprim was used primarily for urinary appeared to reduce the magnitude of NTD risk associated tract infections, we considered whether the observed associ- with exposure to carbamazepine and trimethoprim.
ation might be due to urinary tract infections rather than to Human studies have shown that exposure to FAAs such as trimethoprim. We did not find an appreciable association aminopterin (15), methotrexate (16), and valproic acid (8, 9, between urinary tract infections and NTDs (odds ratio ϭ 17) can cause NTDs. An association between carbamazepine 1.3; 95 percent CI: 0.8, 2.2). For other antibiotics used for and NTDs has also been suggested in a number of studies (8, urinary tract infections during the first 2 months after the 10, 17), but they have tended to be small, leaving the magni- last menstrual period, the odds ratio was 1.2 (95 percent CI: tude of risk unclear. Little has been described for trimetho- 0.7, 2.0) for amoxicillin/ampicillin and 1.6 (95 percent CI: prim risks, however. Clinical trials did not indicate an 0.6, 4.3) for cephalosporins. (In one subgroup analysis, increased rate of birth defects, but those studies contained cephalexin had an odds ratio of 4.0 (95 percent CI: 1.2, too few pregnancies to allow detection of a teratogenic effect (18–20). Epidemiologic studies have only rarely considered When cases were compared with nonmalformed controls, this drug; a decade ago, two case-control studies found a sig- the odds ratios were 2.8 (95 percent CI: 1.3, 6.3) for FAAs, nificant increase in the risk of oral clefts and hypospadias 2.0 (95 percent CI: 0.4, 10.1) for carbamazepine, and 7.7 (95 among trimethoprim users (21, 22). The present findings are percent CI: 0.7, 80.2) for trimethoprim (based on 13, four, the first to link trimethoprim to NTDs in humans.
and one exposed controls, respectively). When they were Most members of the FAA “class” seem to induce NTDs.
compared with chromosomal disorders, the odds ratios were Two exceptions are phenytoin and phenobarbital, which 3.2 (95 percent CI: 1.7, 6.1) for FAAs, 10.3 (95 percent CI: have been associated with various congenital defects in ear- 1.2, 91.4) for carbamazepine, and 3.7 (95 percent CI: 0.8, lier studies (9, 17, 23, 24). However, neither those studies 16.5) for trimethoprim (based on 18, one, and four exposed nor the current one has linked these drugs to NTDs. Drug- specific differences in teratogenic effects among different Prevalence odds ratios of neural tube defects for women exposed
periconceptionally to trimethoprim or carbamazepine, folic acid supplements, or both,
Slone Epidemiology Unit Birth Defects Study, United States and Canada, 1976–1998
* The odds ratio was estimated for 1,242 cases and 6,660 malformed controls and adjusted for interview year, region, and maternal age, education, weight before pregnancy, and urinary tract infections or other infections earlyin pregnancy.
FAAs may be due to variations in folate antagonism among persons would tend to dilute the effects of folic potency or to actions at different steps of folate metabolism (25). Those that are folic acid analogs (including Known confounding factors were taken into account in aminopterin, methotrexate, valproic acid, triamterene, and the statistical analysis. Trimethoprim, typically combined trimethoprim) displace folate from enzymes and block the with sulfamethoxazole, is used primarily for urinary tract enzymatic reactions in which folate participates. For infections. Neither urinary tract infections nor other infec- instance, the antimicrobial activity of trimethoprim results tions during the first trimester were significantly associated from a selective inhibition of dihydrofolate reductase in with NTDs, and they did not materially change the odds unicellular organisms (26). Although the concentration ratio estimate for trimethoprim. Use of other drugs com- needed to inhibit human dihydrofolate reductase is monly prescribed for urinary tract infections in pregnancy, 100,000-fold greater than that required to inhibit the bacte- notably ampicillin and amoxicillin, was also not associated rial enzyme, the inhibition in humans is enough to be asso- with an increased risk nor was the use of cephalosporins as ciated with megaloblastic anemia (26), and enzymes in the a group; in the context of multiple testing, the specific find- human embryo may behave differently from those in adults.
ing for cephalexin may be due to chance. We cannot sepa- Further, folic acid supplementation reduces the clinical tox- rate an effect of trimethoprim alone from an effect of icity of these FAAs (27–29) and, in one study of rats, mal- trimethoprim combined with sulfonamides, but we found no formations induced by trimethoprim were prevented by association with sulfonamides other than the FAA sul- administration of folinic acid or dietary supplementation fasalazine. It is of note that sulfamethoxazole inhibits de novo production of folic acid only in bacteria, but trimetho- Our study has a number of limitations. Although we prim affects transformation of folic acid into required co- included stillborn infants and therapeutically aborted fetuses enzymes in both bacteria and humans (26). Thus, biologic since 1988, we still are likely to be missing affected preg- coherence would favor the effect of trimethoprim over sul- nancies, particularly early spontaneous abortions. However, famethoxazole. For carbamazepine, as for other antiepilep- if the NTDs produced by FAAs were more often lethal or tic drugs, it is always difficult to disentangle a direct effect more severe (i.e., more easily detected prenatally), our sub- of the medication from a potential effect of epilepsy itself; jects would represent a survivor cohort and would lead to an however, NTDs are not consistently associated with most underestimate of the NTD risk due to FAAs.
Use of malformed controls would have introduced bias in Maternal periconceptional exposure to FAAs such as this study if FAAs were related to malformations other than trimethoprim and carbamazepine appears to increase the NTDs. However, subjects with conditions potentially asso- risk of NTDs. The strength, time-specific effects, and bio- ciated with folate supplementation were excluded. Further, logic plausibility of the findings, together with previous since the controls include a variety of malformations, a pre- studies in animals, in humans, and with other FAAs, offer viously undocumented effect of FAA on any one malforma- support for a causal association. Fortunately, among women tion would have little impact on these findings. In fact, the who use carbamazepine or trimethoprim, the absolute risk use of alternative control groups did not materially change appears to be modest, because more than 99 percent of the results. Moreover, under most scenarios of biased con- women exposed to these drugs early in pregnancy will trol selection, the reported risk would underestimate the true deliver an infant unaffected by an NTD. Moreover, theprevalence of carbamazepine and trimethoprim use in our population was low, so that the number of cases potentially Because we relied on women’s ability to recall and report attributable to these drugs was small. It is unclear whether information, there may have been differential misclassifica- supplemental folic acid will reduce this excess risk or tion of past exposure between case and control mothers. Use whether it will affect the therapeutic efficacy of these drugs.
of malformed controls should reduce this possibility. In addi- Finally, though these findings extend the number of FAAs tion, the carefully designed questionnaire, administered that may cause NTDs, they are based on relatively few sub- relatively soon after delivery by interviewers blind to the jects and need to be evaluated in further studies.
hypothesis, is likely to have substantially reduced informa-tion errors. On average, mothers of NTD cases were inter-viewed sooner after the last menstrual period than weremothers of malformed controls (53.4 and 58.8 weeks, respec-tively), mainly because of a higher number of stillbirths and ACKNOWLEDGMENTS
therapeutic abortions among cases. However, these differ-ences did not affect maternal recall, as reflected by the simi- Supported in part by the Pharmacoepidemiology lar time intervals between the last menstrual period and Teaching and Research Fund of the Harvard School of interview found for exposed and unexposed controls.
Public Health: the National Center for Environmental Nondifferential underreporting, or misclassification of Health, Centers for Disease Control and Prevention, through the dates of the last menstrual period, or of exposure a grant to the Massachusetts Center for Birth Defects would underestimate the true association. We did not con- Research and Prevention; and by the Massachusetts sider food sources of folate, which have greatest impact on Department of Public Health: the National Institute of Child folate status among nonusers of folic acid supplements Health and Human Development grant HD27697 and the (31). However, unaccounted variations in folate intake National Heart, Lung, and Blood Institute grant HL50763.
NTDs and Folic Acid Antagonists during Pregnancy Additional support for the Slone Epidemiology Unit Birth REFERENCES
Defects Study was provided by Hoechst Marion Roussel,Inc. (Kansas City, Missouri); Pfizer, Inc. (New York, New 1. Elwood M, Little J, Elwood H. Conclusions. In: Elwood JM, ed. Epidemiology and control of neural tube defects. New York); the Glaxo-Wellcome Company (Research Triangle York, NY: Oxford University Press, 1992:756–77.
Park, North Carolina); and Rhone Pouleuc Rorer 2. Prevention of neural tube defects: results of the Medical Research Council Vitamin Study. MRC Vitamin Study The authors thank Rachel Wilson, project coordinator; Research Group. Lancet 1991;338:131–7.
3. Lewis DP, Van Dyke DC, Stumbo PJ, et al. Drug and environ- Fiona Rice, research coordinator; Rita Krolak, quality con- mental factors associated with adverse pregnancy outcomes.
trol/coder; Sally Perkins, Mary Krieger, Kathleen Sheehan, Part II. Improvement with folic acid. Ann Pharmacother 1998; Karen Bennett Mark, Deborah Kasindorf, and Clare Coughlin, interviewers; Joan Shander, Diane Gallagher, and 4. Yates JRW, Ferguson-Smith MA, Shenkin A, et al. Is disor- dered folate metabolism the basis for the genetic predisposi- Valerie Hillis, research assistants; Thomas Kelley, research tion to neural tube defects? Clin Genet 1987;31:279–87.
pharmacist; and Nastia Dynkin and John Farrell, program- 5. Daly LE, Kirke PN, Molloy A, et al. Folate levels and neural tube defects. Implications for prevention. JAMA 1995;274: The authors also are indebted to the following institutions 6. Thiersch JB, Seattle W. Therapeutic abortions with a folic acid that generously provided them access to patients: Boston antagonist, 4-aminopteroylglutamic acid (4-amino P.G.A.) area: Baystate Medical Center, Beth Israel Deaconess administered by the oral route. Am J Obstet Gynecol 1952;63: Medical Center, Boston Medical Center, Brigham & Women’s Hospital, Brockton Hospital, Cambridge Hospital, 7. Milunsky A, Graef JW, Gaynor MF. Methotrexate-induced congenital malformations. J Pediatr 1968;72:790–5.
Children’s Hospital, Charlton Memorial Hospital, 8. Samren EB, Van Duijn CM, Koch S, et al. Maternal use of Deaconess Waltham Hospital, Emerson Hospital, Falmouth antiepileptic drugs and the risk of major congenital malforma- Hospital, Franciscan Children’s Hospital, Good Samaritan tions: a joint European prospective study of human teratogenesis Medical Center, Haverhill Municipal-Hale Hospital, Holy associated with maternal epilepsy. Epilepsia 1997;38:981–90.
9. Dravet C, Julian C, Legras C, et al. Epilepsy, antiepileptic Family Hospital, Jordan Hospital, Kent County Memorial drugs, and malformations in children of women with epilepsy: Hospital, Lawrence General Hospital, Lowell General a French prospective cohort study. Neurology 1992;42(suppl 5): Hospital, Melrose-Wakefield Hospital, UMASS Memorial Health Care (Memorial Campus), Metro West Medical 10. Rosa F. Spina bifida in infants of women treated with carba- mazepine during pregnancy. N Engl J Med 1991;324:674–7.
Center (Framingham Campus), Mt. Auburn Hospital, New 11. Mitchell AA, Rosenberg L, Shapiro S, et al. Birth defects related England Medical Center, Newton-Wellesley Hospital, North to Benedectin use in pregnancy. JAMA 1981;245:2311–14.
Shore Medical Center, Rhode Island Hospital, Saints 12. Mitchell AA. Special considerations in studies of drug-induced Memorial Medical Center, South Shore Hospital, Southern birth defects. In: Strom BL, ed. Pharmacoepidemiology. 3rd New Hampshire Regional Medical Center, St. Elizabeth’s ed. Chichester, UK: John Wiley & Sons, Inc, 2000:595–608.
13. Werler MM, Hayes C, Louik C, et al. Multivitamin supplemen- Medical Center, St. Luke’s Hospital, St. Vincent Hospital, tation and birth defect risks. Am J Epidemiol 1999;150:675–82.
and Women & Infants’ Hospital; Philadelphia area: 14. Lewis DP, Van Dyke DC, Stumbo PJ, et al. Drug and environ- Abington Memorial Hospital, Albert Einstein Medical mental factors associated with adverse pregnancy outcomes.
Center, Alfred I. duPont Institute, Bryn Mawr Hospital, Part I. Antiepileptic drugs, contraceptives, smoking, andfolate. Ann Pharmacother 1998;32:802–17.
Chester County Hospital, Children’s Hospital of 15. Carter CO. Spina bifida and anencephaly: a problem in Philadelphia, Christiana Care Health Services, City Avenue genetic-environmental interaction. J Biosoc Sci 1969;1:71–83.
Hospital, Community General Hospital, Crozer-Chester 16. Sulik KK, Cook CS, Webster WS. Teratogens and craniofacial Medical Center, Doylestown Hospital, Frankford Hospital malformations: relationships to cell death. Development 1988;103(suppl):213–32.
(Torresdale Division), Hospital of the University of 17. Czeizel AE, Bod M, Halasz P. Evaluation of anticonvulsant Pennsylvania, Lancaster General Hospital, Lehigh Valley drugs during pregnancy in a population-based Hungarian Hospital, Nanticoke Memorial Hospital, Pennsylvania study. Eur J Epidemiol 1992;8:122–7.
Hospital, Reading Hospital & Medical Center, Sacred Heart 18. Williams JD, Brumfitt W, Condie AP, et al. The treatment of bacteriuria in pregnant women with sulphamethoxazole and Hospital, St. Mary Medical Center, Temple University trimethoprim. Postgrad Med J 1969;45:71–5.
Hospital, and Thomas Jefferson University Hospital; and 19. Brumfitt W, Pursell R. Trimethoprim-sulfamethoxazole in the Toronto area: Cambridge Memorial Hospital, Grand River treatment of bacteriuria in women. J Infect Dis 1973;128: Hospital, Guelph General Hospital, Hospital for Sick Children, Humber River Regional Hospital, Joseph Brant 20. Bailey RR. Single-dose antibacterial treatment for bacteriuria in pregnancy. Drugs 1984;27:183–6.
Memorial Hospital, London Health Sciences Center, 21. Hill L, Murphy M, McDowall M, et al. Maternal drug histories McMaster University Medical Center, Mt. Sinai Hospital, and congenital malformations: limb reduction defects and oral North York General Hospital, Oakville Trafalgar Memorial clefts. J Epidemiol Community Health 1988;42:1–7.
Hospital, Lakeridge Health Corporation, Peel Memorial 22. Czeizel A. A case-control analysis of the teratogenic effects of co-trimoxazole. Reprod Toxicol 1990;4:305–13.
Hospital, Scarborough General Hospital, St. Joseph’s Health 23. Lindhout D, Meinardi H, Meijer J, et al. Antiepileptic drugs Centre (London), St. Joseph’s Health Centre (Toronto), St.
and teratogenesis in two consecutive cohorts: changes in pre- Joseph’s Hospital (Hamilton), St. Michael’s Hospital, scription policy paralleled by changes in pattern of malforma- Toronto East General Hospital, Toronto Hospital (General tions. Neurology 1992;42(suppl 5):94–110.
24. Shaw GM, Wasserman CR, O’Malley CD, et al. Orofacial Division), Trillium Health Center, Women’s College clefts and maternal anticonvulsant use. Reprod Toxicol Hospital, York County Hospital, and York Central Hospital.
25. Picciano MF, Stokstad ER, Gregory JF III. Folic acid metabo- clefting. Br J Orthod 1999;26:115–20.
lism in health and disease. New York, NY: Wiley-Liss, 1990.
29. Wegner C, Nau H. Alteration of embryonic folate metabolism 26. Hardman JG, Limbird LE, Molinoff PB, et al. Goodman & by valproic acid during organogenesis: implications for mech- Gilman’s the pharmacological basis of therapeutics. 9th ed.
anism of teratogenesis. Neurology 1992;42(suppl 5):17–24.
30. Udall V. Toxicology of sulphonamide-trimethoprim combina- 27. Morgan SL, Baggott JE, Vaughn WH, et al. Supplementation tions. Postgrad Med J 1969;45(suppl):42–5.
with folic acid during methotrexate therapy for rheumatoid 31. Werler MM, Shapiro S, Mitchell AA. Periconceptional folic arthritis. Ann Intern Med 1994;121:833–41.
acid exposure and risk of occurrent neural tube defects. JAMA 28. Hartridge T, Illing H, Sandy J. The role of folic acid in oral

Source: http://www.idpas.org/pdf/1076-neuraltubedefects.pdf