Am. J. Epidemiol. 2002 Werler 26 31

8/11/2019 Am. J. Epidemiol. 2002 Werler 26 31

1/6

American Journal of EpidemiologyCopyright 2002 by the Johns Hopkins Bloomberg School of Public HealthAll rights reserved

26

Vol. 155, No. 1Printed in U.S.A.

Maternal Medication Use and Birth Defects Werler et al.

Maternal Medication Use and Risks of Gastroschisis and Small IntestinalAtresia

Martha M. Werler, Jane E. Sheehan, and Allen A. Mitchell

Gastroschisis and small intestinal atresia (SIA) are birth defects that are thought to arise from vasculardisruption of fetal mesenteric vessels. Previous studies of gastroschisis have suggested that risk is increasedfor maternal use of vasoactive over-the-counter medications, including specific analgesics and decongestants.This retrospective study evaluated the relation between maternal use of cough/cold/analgesic medications andrisks of gastroschisis and SIA. From 1995 to 1999, the mothers of 206 gastroschisis cases, 126 SIA cases, and798 controls in the United States and Canada were interviewed about medication use and illnesses. Risks ofgastroschisis were elevated for use of aspirin (odds ratio = 2.7, 95% confidence interval: 1.2, 5.9),pseudoephedrine (odds ratio = 1.8, 95% confidence interval: 1.0, 3.2), acetaminophen (odds ratio = 1.5, 95%

confidence interval: 1.1, 2.2), and pseudoephedrine combined with acetaminophen (odds ratio = 4.2, 95%confidence interval: 1.9, 9.2). Risks of SIA were increased for any use of pseudoephedrine (odds ratio = 2.0,95% confidence interval: 1.0, 4.0) and for use of pseudoephedrine in combination with acetaminophen (oddsratio = 3.0, 95% confidence interval: 1.1, 8.0). Reported fever, upper respiratory infection, and allergy were notassociated with risks of either defect. These findings add more evidence that aspirin use in early pregnancyincreases risk of gastroschisis. Although pseudoephedrine has previously been shown to increase gastroschisisrisk, findings of this study raise questions about interactions between medications and possible confounding byunderlying illness. Am J Epidemiol 2002;155:2631.

gastroschisis; intestinal atresia, medicine; pregnancy

Received for publication February 20, 2001, and accepted forpublication July 9, 2001.

Abbreviation: SIA, small intestinal atresia.From the Slone Epidemiology Unit, Boston University School of

Public Health, Boston, MA.Reprint requests to Dr. Martha M. Werler, Slone Epidemiology

Unit, 1371 Beacon Street, Brookline, MA 02446 (e-mail: [email protected]).

Gastroschisis and defects of the small intestine are clini-cally distinct congenital malformations, but it is thought thatthey share a common etiologic mechanism. Gastroschisis, adefect of the abdominal wall through which the intestinesprotrude at birth, putatively results from disruption of theomphalomesenteric artery by the eighth week of gestation(14). Small intestinal atresia (SIA) has been induced byocclusion of the superior mesenteric artery in dogs (5).

Previous epidemiologic studies of gastroschisis haveshown increased risks among mothers who have reported tak-ing various vasoactive over-the-counter medications, includ-ing pseudoephedrine (6, 7), phenylpropanolamine (6, 7),aspirin (69), ibuprofen (6, 7), and acetaminophen (6). Theseagents vary in their pharmacologic actions, in the prevalenceof their use in pregnancy, and in the amount of evidenceshowing increased risks of gastroschisis. However, thesemedications share common indications for their use, particu-larly upper respiratory infections. In fact, many over-the-counter products combine decongestants, antipyretics/ analgesics, and other agents and are marketed according to

the symptoms they relieve. At the same time, gastroschisishas been observed to occur in clusters, and some studies, butnot all, have reported greater frequencies of births in certainseasons (1012). These observations raise the question aboutwhether an infectious agent might increase risk. Althoughthere have been few epidemiologic studies of SIA, it has beenpostulated that risk factors for gastroschisis might be the samefor SIA, since both defects are thought to result from abnor-mal fetal mesenteric vasculature (25). Our study wasdesigned to examine gastroschisis and SIA risks in relation tocold and antipyretic/analgesic medications, with specificattention to rigorous ascertainment of relevant medicationstaken and illnesses experienced in early pregnancy.

MATERIALS AND METHODS

This case-control study was conducted from June 1995through March 1999 in 15 cities across the United Statesand Canada. Study subjects were ascertained from 29 pedi-atric tertiary care hospitals within 5 months of birth.Institutional review board approval was obtained from eachparticipating hospital. Gastroschisis cases were defined asnewborn infants with intestines protruding through the wallof the abdomen lateral to, but not involving, the umbilicus.SIA cases were defined as atresia, stenosis, or webbing of the duodenum, jejunum, or ileum, without gastroschisis.Gastroschisis and SIA cases with known chromosomalabnormalities or Mendelian-inherited disorders were
http://aje.oxfordjournals.org/http://aje.oxfordjournals.org/http://aje.oxfordjournals.org/http://aje.oxfordjournals.org/http://aje.oxfordjournals.org/http://aje.oxfordjournals.org/http://aje.oxfordjournals.org/http://aje.oxfordjournals.org/http://aje.oxfordjournals.org/http://aje.oxfordjournals.org/http://aje.oxfordjournals.org/http://aje.oxfordjournals.org/http://aje.oxfordjournals.org/http://aje.oxfordjournals.org/http://aje.oxfordjournals.org/http://aje.oxfordjournals.org/http://aje.oxfordjournals.org/http://aje.oxfordjournals.org/http://aje.oxfordjournals.org/http://aje.oxfordjournals.org/http://aje.oxfordjournals.org/http://aje.oxfordjournals.org/http://aje.oxfordjournals.org/


2/6

Maternal Medication Use and Birth Defects 27

Am J Epidemiol Vol. 155, No. 1, 2002

excluded. Two control groups were used for both casegroups. Malformed controls were infants with major struc-tural malformations other than gastroschisis, SIA, or othergastrointestinal defects. Nonmalformed controls wereinfants with medical conditions requiring hospital admission(e.g., infections, prematurity, hyperbilirubinemia, transienttachypnea, meconium aspiration, seizures, and gastrointesti-nal problems). Because gastroschisis, but not SIA, isstrongly associated with young maternal age, mothers of malformed and nonmalformed controls were matched to theage (+2/4 years) of gastroschisis mothers, but not SIAmothers. Eligibility for all cases and controls required per-mission from the infants physician for the mother to be con-tacted about the study. Mothers were sent a letter describingthe study and a medication identification booklet containingcolor pictures of cold and antipyretic/analgesic products tohelp women differentiate among similarly named products(e.g., Tylenol vs. Tylenol Flu (McNeil PPC, Inc., FortWashington, Pennsylvania)). After informed consent wasobtained, a nurse-interviewer administered the standardizedquestionnaire by telephone within 6 months after delivery.The interviewer was not blinded to case/control status toavoid her appearing to be insensitive to the families situa-tions, which might have undermined her rapport.Information was collected on demographic factors; repro-ductive, illness, and medication histories; and cigarette andbeverage consumption. Detailed questions were asked aboutoccurrences of allergies, colds, influenza, bronchitis, sinusproblems, headaches, and asthma and whether medicationswere taken for those illnesses. Medication use was alsoelicited by prompts of specific drugs. Details on medicationuse were obtained, including the specific product, start andstop dates of use, frequency, duration, form (liquid, pill),

amount taken per day, and reason for use. For each medica-tion reported, women were asked to retrieve the bottle, if available, to accurately identify the product. If the bottle orpackage was not available for cold or antipyretic/analgesicmedications, the woman was asked to identify the product inthe Medication Identification Booklet (Slone EpidemiologyUnit, Boston University School of Public Health, Boston,Massachusetts). Each reported product was linked to theSlone Epidemiology Unit Drug Dictionary, which identifiesindividual ingredients in multiple component products.

Use of pseudoephedrine, phenylpropanolamine, aspirin,ibuprofen, and acetaminophen was examined. In addition,we also examined use of antihistamines, guaifenesin, and

dextromethorphan because they are often taken with decon-gestants and antipyretics/analgesics. Exposure was definedas any reported use of a medication from the first day of thelast menstrual period through 84 days after the last men-strual period (lunar months 13), the developmentally rele-vant time period for both defects. For each drug, use wascompared with no use during lunar months 13. Use of com-bination products was also examined. Potential confoundingby reported indication for medication use and by reportedpresence of fever was assessed.

Logistic regression was used to estimate odds ratios and95 percent confidence intervals. Models for gastroschisisrisk were conditional on the match and included terms for

pseudoephedrine, phenylpropanolamine, aspirin, ibuprofen,acetaminophen, antihistamines, guaifenesin, dextromethor-phan, fever, upper respiratory infection, and allergy.Adjustments were also made for the potential confoundingeffects of family income (


3/6

28 Werler et al.


a threefold increase in risk of SIA. Use of other pseu-doephedrine- or acetaminophen-containing combinationproducts was less common; risks were estimated for gas-troschisis only, and odds ratios were increased, but were notstatistically significant. Users of multiple-component prod-ucts did not ingest higher doses of pseudoephedrine or aceta-minophen than did single-component users.

Although we did not find maternal reports of upper respi-ratory infection, allergy, or fever to be associated with gas-troschisis or SIArisk, the increased risks of gastroschisis andSIA observed for pseudoephedrine and acetaminophen takenin combination form raised the question of confounding byunderlying infectious illness, particularly by a specificmicroorganism, that could not be accurately measured frommaternal report. Because gastroschisis can occur in clusters(perhaps reflecting an infectious etiology), we used an alter-native approach to examine the possibility of confounding byinfectious disease. If we assume that cases in the clustergroup would more likely have an infectious etiology than

TABLE 1. Medication use and illnesses during lunar months 13 among cases and controls, UnitedStates and Canada, 19951999

Cases Controls

Gastroschisis SIA * Malformed NonmalformedExposure

No. % No.No.No. %%%

PseudoephedrinePhenylpropanolamineAspirinIbuprofenAcetaminophenAntihistaminesGuaifenesinDextromethorphanFeverUpper respiratory infectionAllergy

358

1330

120261217165631

1746

155813

688

2715

2725

157218

912124422

2124

125714

710103518

45141348

18448192324

10166

1243

134813

566

2617

43131247

202562433289564

1033

114914

687

2315

* SIA, small intestinal atresia.

TABLE 2. Odds ratios for medication use and illnesses,United States and Canada, 19951999

Gastroschisis SIA*Exposure

OR * 95% CI *, OR 95% CI

PseudoephedrinePhenylpropanolamineAspirinIbuprofenAcetaminophenAntihistaminesGuaifenesinDextromethorphanFeverUpper respiratory

infectionAllergy

1.81.22.71.11.50.60.71.11.0

0.90.8

1.0, 3.20.5, 3.11.2, 5.90.7, 1.81.1, 2.20.3, 1.20.3, 1.50.3, 4.50.5, 2.0

0.6, 1.50.5, 1.3

2.00.50.50.91.00.90.70.81.4

1.30.9

1.0, 4.00.1, 2.70.2, 1.80.5, 1.70.6, 1.60.4, 1.80.3, 2.00.1, 9.70.6, 3.3

0.7, 2.30.5, 1.7

* SIA, small intestinal atresia; OR, odds ratio; CI, confidenceinterval.

Odds ratios adjusted for maternal age (SIA only), education,income, medication use, illness, illicit drug use, and cigarette smoking.

TABLE 3. Medication use according to the number of components, United States and Canada,19951999

Cases Controls

Gastroschisis SIA* Malformed NonmalformedMedication (type)

No. % No.No.No. %%%

PseudoephedrineSingle

CombinationWith acetaminophenWithout acetaminophen

AspirinSingleCombination

AcetaminophenSingleCombination

With pseudoephedrineWithout pseudoephedrine

6

2923

6

121

903023

7

3

1411

3

60.5

441511

3

10

1714

3

41

551714

3

8

1411

2

31

441411

2

20

2521

4

103

1552921

8

5

761

31

41862

16

2721

6

111

167352114

4

751

30.2

40853

* SIA, small intestinal atresia.


4/6



would cases in the noncluster group, then confounding byinfectious disease would be minimized in the nonclustergroup. For gastroschisis, we arbitrarily defined a cluster asascertainment of at least three cases from the same geo-graphic area born within a 30-day interval. There were 71cases in the cluster group and 134 cases in the noncluster

group (the ascertainment date was unclear for one case).Odds ratios for medication use and illnesses are presented foreach group in table 5. The prevalences of medication use andillnesses were generally similar among cases in the clusterand the noncluster groups, with the exception of aspirin use,which was nearly twice as common among cluster cases. Inthe noncluster group, the risk estimates for any pseu-doephedrine use and any acetaminophen use approached thenull, but estimates for pseudoephedrine in combination with

acetaminophen and for any aspirin use remained elevated.No clear pattern of risk was observed for reported illnesseswithin each group.

Although SIA is not known to occur in clusters, we iden-tified 37 cases that fit our criteria as having occurred in acluster. Among the noncluster group, there were nine casesexposed to pseudoephedrine in combination with aceta-minophen, for a risk of estimate of 2.7 (95 percent confi-dence interval: 1.2, 5.8).

DISCUSSION

Findings on risks of gastroschisis from this study confirm,to a certain extent, those from previous studies. Specifically,aspirin use has consistently been shown to increase gas-troschisis risk in both animals (13) and humans (69). Ourstudy confirmed that association with an observed 2.7-foldincreased risk for any use and 3.2-fold risk for single-component use, which were not altered by control for othermedication use or reported illnesses. The effects of aspirin on

fetal vasculature are not known, but both vasoconstrictionand vasodilation can result from aspirin intake at varyingdoses in adults (14). Findings on other medications in thisstudy are less consistent, however. For example, ibuprofen,which shares some pharmacologic actions with aspirin, wasnot observed to increase gastroschisis risk, in contrast to anearlier report (7). On the other hand, acetaminophen is notknown to affect adult vasculature (14), but overall use wasassociated with a slightly increased risk of gastroschisis, asreported in one other study (6). While decongestants such aspseudoephedrine and phenylpropanolamine are vasocon-strictive (14) and have been observed in previous studies toincrease gastroschisis risk (6, 7), we found that pseu-

doephedrine use overall was only modestly associated withgastroschisis risk (not statistically significant) and thatphenylpropanolamine did not increase risk. However, thisstudy was the first to examine use of combination products.We found that single-component pseudoephedrine use wasnot associated with an increased risk (although this findingwas based on few users) and that multiple-component pseu-doephedrine use was associated with such a risk. The major-ity of multiple-component products contained both pseu-

TABLE 4. Odds ratios for medication use by type ofproduct, United States and Canada, 19951999

Gastroschisis SIA*Medication (type)

OR * 95% CI *, OR 95% CI

PseudoephedrineSingleCombination

Without acetaminophen

AspirinSingleCombination

AcetaminophenSingleCombination

Without pseudoephedrine

Pseudoephedrine andacetaminophen (withor without others)

1.80.72.5

3.4

3.2

1.51.51.72.0

4.2

1.0, 3.20.2, 2.11.0, 6.2

1.0, 11.3

1.4, 7.1

1.1, 2.21.0, 2.10.8, 4.00.8, 5.2

1.9, 9.2

2.01.14.2

1.01.00.7

3.0

1.0, 4.00.5, 2.91.3, 14.1

0.6, 1.60.6, 1.60.2, 2.3

1.1, 8.0

* SIA, small intestinal atresia; OR, odds ratio; CI, confidenceinterval. Odds ratios adjusted for maternal age (SIA only), education,

income, medication use, illness, illicit drug use, and cigarettesmoking.

TABLE 5. Medication use and illnesses in relation to risk of gastroschisis according to cluster group, United States and Canada,19951999

Cluster ( n = 71) Noncluster ( n = 134)

OR * 95% CI *, OR 95% CINo. % No. %Exposure

PseudoephedrineAcetaminophenPseudoephedrine and

acetaminophen (with orwithout others)

AspirinFeverUpper respiratory infectionAllergy

1445

967

1710

2063

139

102414

1.31.9

2.13.91.70.70.8

0.5, 3.11.0, 3.6

0.7, 6.31.0, 15.50.5, 5.70.3, 1.60.3, 1.8

2175

1479

3921

1656

1057

2916

1.31.3

2.43.60.80.81.2

0.7, 2.40.9, 2.0

1.0, 5.71.3, 10.10.3, 1.90.5, 1.40.7, 2.0

* OR, odds ratio; CI, confidence interval. Odds ratios (95% confidence intervals) adjusted for education, income, medication use, illness, illicit drug use, and cigarette smoking.


5/6

30 Werler et al.


doephedrine and acetaminophen, and this combination prod-uct was associated with an approximate 4.2-fold increasedrisk of gastroschisis.

It is possible that an interaction between pseudoephedrineand acetaminophen might affect the development of gas-troschisis and SIA, but differences in women who take thesecombination products, such as their underlying illnesses,might also affect risk. We hypothesized that use of pseu-doephedrine/acetaminophen combination products was amarker for an infectious agent that itself increased gas-troschisis risk. Although we adjusted for reported upper res-piratory illnesses, measurements of specific infectiousagents are difficult to obtain in observational studies, leav-ing open the opportunity for uncontrolled confounding. Wetherefore used an alternative approach to examine the possi-bility of confounding by underlying infection by identifyingclusters of gastroschisis cases, under the hypothesis thatcases that occur in clusters more likely have an infectiousetiology. If medication use is confounded by infectious dis-ease, then risk estimates for medications would approach thenull among the noncluster group. For gastroschisis, esti-mates for any pseudoephedrine use and any acetaminophenuse did, indeed, approach the null within the nonclustergroup. However, increased risks of gastroschisis and SIAwere observed for use of pseudoephedrine in combinationwith acetaminophen within both the cluster and the non-cluster groups. A similar pattern was observed for aspirinuse and gastroschisis risk. The persistence of an elevatedrisk among the subgroup of cases least likely to be causedby an infectious illness suggests that the observed associa-tion for specific fever/pain and cold medications may not beconfounded by maternal illness. However, since reportedfever and upper respiratory illness were not associated with

cluster status among cases, it is possible that such reportingmight be inaccurate and/or that the cluster grouping may notbe indicative of an upper respiratory infectious illness.

To our knowledge, this was the first study to examinemedication use in early pregnancy in relation to SIA risk. Asdiscussed for gastroschisis, it is biologically plausible thatvasoactive effects of pseudoephedrine, phenylpropano-lamine, aspirin, and ibuprofen might affect the developmentof SIA by fetal vascular disruption, but only multiple-component pseudoephedrine (primarily combined withacetaminophen) was associated with an increased SIA risk.While the combination of acetaminophen and pseu-doephedrine was associated with a threefold increase in risk,

acetaminophen use overall did not appear to affect risk.Our findings have limitations. Selection bias might be a

concern because our study was not population based. Toinclude sufficient numbers of cases in less than 4 years, studysubjects were ascertained at large tertiary pediatric care cen-ters across the United States and Canada. However, the find-ings on pseudoephedrine and aspirin in this hospital-basedstudy confirm those observed by Torfs et al. (7) in their pop-ulation-based study. In addition, higher proportions of ascer-tained controls than of cases were ineligible for the studybecause physicians were either not able to be contacted orthey denied permission for the family to be approached.However, it is unlikely that medication use in pregnancy

would be related to successful contact of and approval byphysicians. In addition, proportions of mothers whom wewere unable to contact or who refused to participate weresimilar among cases and controls, but represented approxi-mately 2030 percent of the women approached. Although itis unlikely that medication use is related to successful contactand interview of mothers, if it were, these findings might notbe generalizable to all women.

Information was retrospectively collected in this study,raising the possibilities of both random and differential mis-classification of exposure. Several steps were taken to min-imize these possibilities. The interval between birth andinterview was less than 6 months for all study subjects, andthe interview used standardized, detailed questions to iden-tify medications and illnesses during early pregnancy. If there were nondifferential reporting of medication use, theobserved odds ratios would underestimate true effects.Concern about differential classification (e.g., recall bias)led us to use two separate control groups: One comprisedinfants with malformations and the other infants withoutmalformations. While the nonmalformed group includedinfants with medical conditions requiring hospitalization inthe first few months of life and, thus, was not normal,maternal recall of early pregnancy events and exposuresmay be closer to that of mothers of healthy infants than tothat of mothers of malformed infants. It is noteworthy thatthe rigorous, systematic inquiry of our standardized inter-view resulted in strikingly similar reported prevalences of medication use and other vasoconstrictive exposures in thetwo control groups, suggesting similar recall accuracy.

In addition, exposure to specific medications was deter-mined by examining the components of combination prod-ucts. In other words, we did not rely on the womens recallof specific ingredients of cough, cold, and antipyretic/anal-gesic products, but rather we asked them to identify in themedication identification booklet the products they took.Therefore, the higher prevalences of pseudoephedrine andacetaminophen use among cases relative to controls, butsimilar rates of other cough and cold components, suggestthat recall bias is unlikely to account for our findings.

Finally, sociodemographic factors, medications, illnesses,and behaviors were controlled as potential confounders,resulting in little change in risk estimates. There remains thepossibility of confounding by other factors, particularlyinfectious agents, that were not accurately ascertained in thisstudy.

Medications for colds, coughs, and pain are commonlyused in early pregnancy. Our findings suggest that many of those medications do not increase risks of gastroschisis andSIA, but suggest that others might affect risks. Specifically,aspirin use appears to increase risk of gastroschisis, aceta-minophen is modestly associated with an increased risk, andibuprofen has no effect. Use of pseudoephedrine alone didnot increase risk, but use of multiple-component pseu-doephedrine, particularly pseudoephedrine in combinationwith acetaminophen, was associated with increases in risk of both gastroschisis and SIA. We were not able to delineate thenature of observed increased risks for multiple-componentmedications further.


6/6



Several questions are raised by these findings, includingwhether the less common use of single-component pseu-doephedrine is without risk, whether medications taken incombination form interact to affect fetal development, andwhether an infectious agent (independently or synergisti-cally with pseudoephedrine and acetaminophen) affects risk.Given the widespread use of many these medications, thesequestions deserve attention in laboratory research settings.

ACKNOWLEDGMENTS

Support for this study was provided by The Procter &Gamble Company, Hoechst Marion Roussel, Inc., andWarner-Lambert Company.

The authors thank the members of our AdvisoryCommittee whose guidance throughout the study wasinvaluable. These members were Dr. Lowell Sever (chair),Dr. Ann Aschengrau, Dr. Michael Shannon, Dr. ClaudineTorfs, and Dr. Linda Van Marter. The authors are thankfulfor our dedicated staff members: Deborah Kasindorf, nurse-interviewer; Joanne Lightbown and Kathy OBrien,research assistants; John Farrell, systems analyst; ThomasKelley, research pharmacist; and Dorothy Roach, wordprocessor.

The authors are indebted to the following hospitals, whichprovided us access to their patients: Boston, MA :Massachusetts General Hospital, Tufts New EnglandMedical Center, and Childrens Hospital; Philadelphia, PA :Thomas Jefferson University Hospital, Childrens Hospitalof Philadelphia, and St. Christophers Hospital ForChildren; Pittsburgh, PA : Childrens Hospital; Buffalo, NY :

Childrens Hospital; Cincinnati, OH : Childrens HospitalMedical Center; Denver, CO : Childrens Hospital andUniversity Hospital; Minneapolis, MN : Childrens MedicalCenter; Chicago, IL : Childrens Memorial Hospital; St.

Louis, MO : Cardinal Glennon Childrens Hospital, St.Johns Mercy Medical Center, and Childrens Hospital;Toronto, Ontario, Canada : Hospital For Sick Children;Falls Church, VA : Fairfax Hospital; Washington, DC :Childrens National Medical Center and GeorgetownUniversity Hospital; Wilmington, DE : A. I. Dupont Institute;

St. Paul, MN : Childrens Hospital; Robinsdale, MN : NorthMemorial Medical Center; Detroit, MI : Childrens Hospital;

Ann Arbor, MI : Mott Childrens Hospital; Lebanon, NH :Dartmouth/Hitchcock Medical Center; Milwaukee, WI :Childrens Hospital; Birmingham, AL : Childrens Hospital;and Indianapolis, IN : James Whitcomb Riley Hospital ForChildren.

REFERENCES

1. Hoyme HE, Higginbottom MC, Jones KL. The vascular patho-genesis of gastroschisis: intrauterine interruption of theomphalomesenteric artery. J Pediatr 1981;98:22831.

2. Van Allen MI. Fetal vascular disruptions: mechanisms andsome resulting birth defects. Pediatr Ann 1981;10:21933.

3. Hoyme HE, Jones KL, Dixon SD, et al. Prenatal cocaine expo-sure and fetal vascular disruption. Pediatrics 1990;85:7437.

4. Martin ML, Khoury MJ, Cordero JF, et al. Trends in rates of multiple vascular disruption defects, Atlanta 19681989: is

there evidence of a cocaine teratogenic epidemic? Teratology1992;45:64753.5. Louw JH, Barnard CN. Congenital intestinal atresia: observa-

tions on its origin. Lancet 1955;2:10657.6. Werler MM, Mitchell AA, Shapiro S. First trimester maternal

medication use in relation to gastroschisis. Teratology 1992;45:3617.

7. Torfs CP, Katz EA, Bateson TF, et al. Maternal medicationsand environmental exposures as risk factors for gastroschisis.Teratology 1996;54:8492.

8. Drongowski RA, Smith RK, Coran AG, et al. Contribution of demographic and environmental factors to the etiology of gas-troschisis: a hypothesis. Fetal Diagn Ther 1991;6:1427.

9. Martinez-Frias ML, Rodriguez-Pinilla E, Prieto L. Prenatalexposure to salicylates and gastroschisis: a case-control study.Teratology 1997;56:2413.

10. Lynberg MC, Cordero JF, Khoury MJ. Increasing prevalenceat birth of gastroschisis in metropolitan Atlanta, 19681990.(Abstract). Teratology 1992;45:453.

11. Penman DG, Fisher RM, Noblett HR, et al. Increase in inci-dence of gastroschisis in the South West of England in 1995.Br J Obstet Gynaecol 1998;105:32831.

12. Goldbaum G, Daling J, Milham S. Risk factors for gastroschi-sis. Teratology 1990;42:397403.

13. Szabo KT. Congenital malformations in laboratory and farmanimals. Berkeley, CA: Academic Press, Inc, 1989.

14. Goodman GA. The pharmacologic basis of therapeutics. Ninthed. New York, NY: McGraw-Hill Book Co, 1996.

Documents

Am. J. Epidemiol. 2002 Werler 26 31