Antifungal Drugs and the Risk of Selected Birth Defects.91.E1- Am J Obstet Gynecol 2008;198:1 191.E7

Home , Butoconazole, Ciclopirox, Terconazole, Tolnaftate

Research www.AJOG.org

OBSTETRICS Antifungal drugs and the risk of selected birth defects Tonia C. Carter, PhD; Charlotte M. Druschel, MD, MPH; Paul A. Romitti, PhD; Erin M. Bell, PhD; Martha M. Werler, PhD; Allen A. Mitchell, MD; for the National Birth Defects Prevention Study

OBJECTIVE: This study examined whether first-trimester antifungaldiovascular defects. An increased risk of 1.88 was observed for dia- drug use was associated with the risk of selected birthphragmatic defects. hernia but was not statistically significant. Estimates ap- proximated unity for neural tube defects, oral clefts, anorectal atresia, STUDY DESIGN:Subjects were participants in a case-control study, hypospadias, and craniosynostosis. the National Birth Defects Prevention Study, with singleton deliveries from 1997 to 2003. Based on maternal interviews, first-trimesterCONCLUSION: anti- First-trimester antifungal drug exposure was not fungal drug use was compared between 7047 cases withstrongly isolated associated de- with the risk of most birth defects, but further stud- fects and 4774 nonmalformed controls using unconditionalies logisticshould examine the preliminary results of an association with hy- regression. poplastic left heart syndrome. RESULTS: Risk was elevated for hypoplastic left heart syndromeKey words:(oddsantifungal agents, congenital abnormalities, pregnancy, ratio, 2.30; 95% confidence interval, 1.04, 5.06) but notteratogens for other car-

Cite this article as: Carter TC, Druschel CM, Romitti PA, et al. Antifungal drugs and the risk of selected birth defects.91.e1- Am J Obstet Gynecol 2008;198:1 191.e7.

he most common clinical indica-associated with hypospadias,6 and top- data from a large, population-based T tion for antifungal drug useical ineconazole exposure was associatedstudy. women is vulvovaginal candidiasis,with cardiovascular defects but not and the Centers for Disease Controlwhen exposure was restricted to Mdrug ATERIALS ANDM ETHODS and Prevention (CDC) have recom-prescriptions documented in medical 7 Data were obtained from the National mended topical azole antifungal med-records. Birth Defects Prevention Study, an on- ication (butoconazole, clotrimazole, Although previous studies have going,not multisite, case-control study of miconazole, tioconazole, terconazole)established whether antifungal drugsthe causes of birth 8defects. The study has for the treatment of this conditioncan causein birth defects, their terato- 1 been approved by the institutional re- pregnancy. In animal studies, birthgenic potential should be evaluatedview boards of the study sites and the defects have been associated withbecause ex- of their use in pregnancyCDC. toCases and controls were identified posure to the antifungal drugs 5-fluo-treat a common, and sometimes recur- 2,3 by the birth defects surveillance systems rocytosine and fluconazole.In hu- ring, vaginal infection. The aim inof 10this states (Arkansas, California, Geor- mans, case-control studies found studyno was to determine whether gia,there Iowa, Massachusetts, New Jersey, increased risk of birth defects withis antop- association between first-trimes-New York, Texas, North Carolina, ical clotrimazole or tolnaftate expo-ter exposure to antifungal drugs and 4,5 Utah). Controls were live births without sure ; however, oral nystatin usethe was risk of selected birth defectsbirth using defects that were randomly selected from birth certificates or birth hospitals From the Department of Epidemiology, School of Public Health, State Universityin the of geographicNew regions monitored by York at Albany, Albany (Drs Carter, Druschel, and Bell); the Congenital Malformations the state surveillance systems. Cases in- Registry, Center for Environmental Health, New York State Department of Health, Troy (Dr Druschel), NY; the Department of Epidemiology, University of Iowa College of cluded live births, stillbirths 20 weeks or Medicine, Iowa City, IA (Dr Romitti); and the Slone Epidemiology Center at Boston longer or greater than 500 g, or elective University, Boston, MA (Drs Werler and Mitchell). terminations. Received Feb. 23, 2007; revised May 24, 2007; accepted Aug. 21, 2007. Reprints: Tonia Carter, Epidemiology Branch, Division of Epidemiology, Statistics, and Case classification Prevention Research, National Institute of Child Health and Human Development, 6100 Medical records, including data on phys- Executive Blvd, Room 7B03C MSC 7510, Bethesda, MD 20892-7510. [email protected] exams, clinical tests, surgeries, and This study was supported by a cooperative agreement with the Centers for Disease Control and autopsies, were obtained for all cases to Prevention, Grant U50/CCU223184. The findings and conclusions in this report are those of the confirm the presence of birth defects. authors and do not necessarily represent the views of the Centers for Disease Control and Prevention. Clinical geneticists reviewed this infor- 0002-9378/$34.00 • © 2008 Mosby, Inc. All rights reserved. • doi: 10.1016/j.ajog.2007.08.044 mation for all cases to classify them as isolated (if all birth defects were confined

FEBRUARY 2008 American Journal of Obstetrics & Gynecology 191.e1 Research Obstetrics www.AJOG.org to the same organ system or body part) Exposure neural tube defects (anencephaly, or nonisolated9; those with known Maternalsin- reports of medication use were craniorachischisis, spina bifida, enceph- gle-gene disorders or chromosomal ab- matched to the active ingredient in the alocele); cleft lip with or without cleft normalities were excluded. Classifica- Boston University Slone Epidemiology palate; anorectal atresia; hypospadias tion was intended to define case groups Center Drug Dictionary (a computer- (second or third degree); craniosynosto- that were more likely to be etiologically ized database of prescription and non- sis; gastroschisis; diaphragmatic hernia; homogenous; for example, isolated prescription drugs that links drug prod- conotruncal heart defects (truncus arte- craniosynostosis cases are probably ucts to their generic ingredients) to riosus, interrupted aortic arch type B, d- etiologically distinct from cases that identify subjects who used antifungal transposition of the great arteries, dou- have other types of isolated defects or drugs in the first trimester. The date of ble outlet right ventricle, tetralogy of cases with multiple defects including conception was considered to be 266 Fallot, pulmonary valve atresia with ven- craniosynostosis. days before the EDD reported by the tricular septal defect–tetralogy of Fallot Based on clinical and pathological cri- mother or obtained from the medical anatomy, ventricular septal defect– teria, cardiovascular defects were addi- record if no date was provided by the conoventricular); left ventricular out- tionally classified as simple, associations, mother. Subjects were classified as ex- flow tract obstructive defects (aortic or complex. The defects were character- posed if they used at least 1 antifungal stenosis, coarctation of the aorta, inter- ized as simple if no other cardiovascular drug in the first trimester (the 90-day pe- rupted aortic arch type A, hypoplastic defects were present, and they were con- riod that started with the date of concep- left heart syndrome); right ventricular sidered to be either specific, single de- tion). Five subjects reported their only outflow tract obstructive defects (pul- fects (eg, atrial septal defect) or a well- fungal infection to be a vaginal yeast in- monary valve stenosis/atresia, tricuspid defined pattern of defects recognized as a fection in the first trimester and also used atresia, Ebstein anomaly); ventricular single diagnostic entity (eg, hypoplastic an antifungal drug but did not recall the septal defects (excluding conoventricu- left heart syndrome, which is composed date of use; their antifungal drug expo- lar type); and atrial septal defects–secun- of a hypoplastic left ventricle and anom- sure was assigned as first trimester. dum type or not otherwise specified. alies of the mitral valve, aortic valve, and ascending aorta). Defects were described as associations if there were 2 or 3 simple Inclusion and exclusion criteria Statistical analysis cardiovascular defects that are known to The study included 12,733 cases and Initially, potential confounders were occur frequently together (eg, atrial sep- 4856 controls that were singleton deliv- identified based on their association with tal defect–secundum type with a ventric- eries between October 1997 and Decem- the exposure or outcomes in the pub- ular septal defect–perimembranous), ber 2003. Mothers who reported a diag- lished literature. Those associated with but no other cardiovascular defects were nosis of type 1 or 2 diabetes before the exposure included urinary tract in- present. Complex defects were those that conception (278 cases, 25 controls), had fections in the first trimester (Yes/No) occurred in multiple cardiac structures. diabetes of unknown type (18 cases, 0 and use 1 month before conception Associations and complex defects are controls), and did not provide informa- through the first trimester of the follow- considered likely to have a different eti- tion on ever being diagnosed with diabe- ing: antibiotics (Yes/No), hormonal ology from simple defects. tes (23 cases, 10 controls) were excluded. contraceptives (Yes/No), and an intra- Also excluded were 93 cases and 32 con- uterine device (Yes/No). Those related trols who did not provide information to the outcomes were maternal age-years Data collection on medication use in pregnancy as well (Ͻ20, 20-34, Ͼ34); maternal education- Structured maternal interviews were as 47 cases and 15 controls who reported years (Ͻ12, 12, Ͼ12); maternal race conducted mainly by telephone in En- use of unknown vaginal creams in the (white, African American, Hispanic, glish or Spanish no later than 24 months first trimester. other); body mass index-kg/m2 (Ͻ18.5, after the expected date of delivery (EDD) To conduct analyses on case groups 18.5-Ͻ25, 25-Ͻ30, Ն30); folic acid sup- to obtain data on maternal exposures that were more likely to be etiologically plement use from one month before during pregnancy. The proportions of homogenous, only cases with isolated conception through the first month of eligible case and control mothers who defects were included in multivariable pregnancy; fever in the first trimester not participated in the interview were 71% analysis (1654 cases with multiple de- caused by a vaginal yeast infection (Yes/ and 68%, respectively. Interviews were fects excluded); for cardiovascular de- No); respiratory illness in the first tri- completed within 6, 12, 18, and 24 fects, only those isolated defects that mester (Yes/No); maternal cocaine use months for 19.9%, 42.7%, 24.1%, and were also defined as simple were in- in the first trimester (Yes/No); maternal 12.7% of participating cases and for cluded (925 cases with isolated, non- occupational exposure to solvents in the 35.9%, 42.0, 15.1, and 6.4% of partici- simple cardiovascular defects excluded). first trimester (Yes/No); a history of mis- pating controls. Time of interview com- Case groups were required to have a carriage or stillbirth (Yes/No); gesta- pletion was missing for 0.6% of cases and minimum of 5 cases with first-trimester tional diabetes affecting the index preg- controls. antifungal drug exposure; these included nancy (Yes/No); a family history of birth

191.e2 American Journal of Obstetrics & Gynecology FEBRUARY 2008 www.AJOG.org Obstetrics Research defects (Yes/No); consanguineous pa- 1-2 times/day. Mothers reported use for an estimated prevalence of 1-4 cases per rental relationship (Yes/No); household more than 30 days when infections lasted 10,000 births10,11; therefore, if our find- income (less than $20,000, $20,000 to for multiple months during gestation, ings reflect a real association, an approx- less than $50,000, or $50,000 or more). suggesting the presence of recurring imate 2-fold increase in risk will result in In addition, time to interview comple- infections. 2-8 cases per 10,000 births. It is unknown tion (within 1 year vs more than 1 year Apart from antifungal drugs, mothers whether the observed increase in risk of after the EDD) and study center were also reported use of other medications. these birth defects was directly related to considered as potential confounders to Among exposed subjects, 83.3% of con- the use of antifungal drugs or whether adjust for differences in interview time trols (70 of 84) and 84.1% of cases (111 the drugs might be a marker for other and geographic location, respectively. of 132) reported using multiple medica- exposures or diseases related to the risk Odds ratios and 95% confidence in- tions in the first trimester (excluding vi- of birth defects. Another explanation is tervals were estimated using uncondi- tamins). Among all study subjects, that the estimates were due to chance be- tional logistic regression. Any poten- 42.1% of controls (2012 of 4774) and cause of the small number of exposed tial confounders that were associated 45.4% of cases (3198 of 7047) reported subjects and should be conservatively in- with both exposure and outcome in bi- use of more than 1 medication in the terpreted as suggesting the possibility of variate analysis with a P Ͻ .2 based on same period. The most frequently used an effect until these results can be con- Fisher’s exact test were included in the medications were acetaminophen, ibu- firmed in another study. initial regression model. Potential con- profen, pseudoephedrine, amoxicillin, Mothers in this study most often used founders were retained in the final naproxen, and aspirin. topical antifungal agents in pregnancy, model if they were related to the out- but some were exposed to ketoconazole Ͻ come at a P .1 and also caused a Birth defect risk and fluconazole, which are available for change of more than 10% in the coef- systemic use. Topical antifungal drugs Estimates of the risk of selected birth de- ficient estimate for antifungal drug use applied at usual therapeutic doses have fects associated with first-trimester ex- when they were dropped from the low systemic absorption (0-15%) with a posure to any antifungal drugs are model. The data were too sparse to re- concentration in serum of 1-100 ng/mL shown in Table 2. The crude and ad- liably determine effect modification. at 12-72 hours after 12-14dose. In con- justed estimates were similar; therefore, Analyses were performed using SAS 9.1 trast, oral antifungal drugs administered only the adjusted estimates are pre- statistical software (SAS Institute, at usual therapeutic doses have a concen- sented. A statistically significant in- Cary, NC). tration in plasma of 1-100 ␮g/mL (10- creased risk was observed for hypoplastic 1000 times greater than for topical left heart syndrome, based on 7 exposed drugs) at 2-48 hours after15,16 anddose, cases: 3 used miconazole, 1 used tercon- RESULTS their higher systemic concentration azole, 1 used ketoconazole, and 2 used an Antifungal drugs raises concern if they are potential unknown antifungal drug. Risk in- The antifungal drugs used in the first tri- teratogens. creased almost 2-fold for diaphragmatic mester are listed in Table 1. Mothers re- We could not perform analyses based hernia, although the effect was not statis- ported using these drugs for vulvovagi- on categories of duration and frequency tically significant. The 9 exposed dia- nal candidiasis except for 3 case mothers of use, topical vs systemic drugs, or use of phragmatic hernia cases used micon- who reported ringworm, an unspecified over-the-counter vs prescription drugs. azole, fluconazole, and unknown fungal skin infection, or oral thrush. The Most drugs were used for no more than antifungal drugs (Table 1). No increase most commonly used drugs were micon- 14 days at a frequency of 1-2 times/day in crude or adjusted risk was observed azole, terconazole, and clotrimazole. and the variation in duration and fre- for the other case groups. There was little Among specifically named antifungal quency of use was not sufficient to form difference in the results obtained when drugs, these 3 drugs accounted for 52 of categories for comparison. In addition, the analyses were repeated for live births 59 exposures to controls (88.1%) and 62 topical/systemic status could not be de- only or for subjects with no family his- of 84 exposures to cases (73.8%). Use of termined for unknown antifungal drugs, tory of birth defects. an unknown antifungal drug was re- and mothers were not asked in the inter- ported by the mothers of 29 controls and view whether drugs were obtained over 55 cases. Most drugs were used 1-2 COMMENT the counter or by prescription. These times/day for periods between 1 and 30 First-trimester exposure to antifungal limitations did not allow us to be more days, but 10 controls and 20 cases re- drugs was associated with a modestly in- specific about the type of antifungal drug ported the duration of use to be between creased risk of isolated hypoplastic left use in pregnancy that could be most rel- 31 and 90 days for miconazole, tercon- heart syndrome and diaphragmatic her- evant to the risk of birth defects. azole, clotrimazole, and unknown anti- nia, although only the increase in risk for The antifungal drugs used by exposed fungal drugs. All 5 subjects who used ke- hypoplastic left heart syndrome was sta- cases of hypoplastic left heart syndrome toconazole reported that exposure was tistically significant. Hypoplastic left and diaphragmatic hernia were micon- between 31 and 90 days at a frequency of heart syndrome is a rare birth defect with azole, terconazole, fluconazole, and ke-

FEBRUARY 2008 American Journal of Obstetrics & Gynecology 191.e3 Research 191.e4 mrcnJunlo Obstetrics of Journal American

TABLE 1

First-trimester antifungal drug use among cases and controls, National Birth Defects Prevention Study (1997-2003) Obstetrics Cleft lip with or Neural without tube cleft Anorectal Diaphragmatic Cardiovascular & Antifungal Controls, defects, palate, atresia, Hypospadias, Craniosynostosis, Gastroschisis, hernia, defects, a(3049 ؍ n) (275 ؍ n) (455 ؍ n) (447 ؍ n) (776 ؍ n) (209 ؍ n) (1086 ؍ n) (750 ؍ n) (4774 ؍ Gynecology drug (n Miconazole 39 10 9 3 3 2 1 2 13 ...... Clotrimazole 7 2 0 1 2 2 0 0 5 ...... ERAY2008 FEBRUARY Terconazole 6 0 2 0 1 0 0 0 4 ...... Fluconazole 0 2 2 0 2 1 0 1 4 ...... Tioconazole 4 0 1 0 0 1 0 0 0 ...... Ketoconazole 1 1 0 0 1 0 0 0 2 ...... Econazole 1 0 1 0 0 0 0 0 0 ...... Butoconazole 1 0 0 0 0 0 0 0 0 ...... Ciclopirox olamine 0 0 1 0 0 0 0 0 0 ...... Propionic acid 0 1 0 0 0 0 0 0 0 ...... Terbinafine 0 0 1 0 0 0 0 0 0 ...... Unknown antifungal 29 4 8 1 4 5 4 6 23 drug ...... Any antifungal drug 84b 18c 23d 51310e 59 49f ...... a Cardiovascular defect case groups have been combined to present data on drug use. b Four control mothers used more than 1 drug: 1 used miconazole and clotrimazole; 1 used miconazole and terconazole; 2 used miconazole and an unknown antifungal drug. c Two neural tube defect case mothers used more than 1 drug: 1 (encephalocele) used ketoconazole and an unknown antifungal drug; 1 (anencephaly) used fluconazole and miconazole. d One cleft lip with cleft palate case mother used econazole, ciclopirox olamine, and terbinafine. e One craniosynostosis case mother used fluconazole and an unknown antifungal drug. f Two cardiovascular defect case mothers used more than 1 drug: 1 (aortic stenosis) used clotrimazole and an unknown antifungal drug; 1 (perimembranous ventricular septal defect) used miconazole and an unknown antifungal drug. www.AJOG.org Carter. Antifungal drugs and the risk of selected birth defects. Am J Obstet Gynecol 2008. www.AJOG.org Obstetrics Research

TABLE 2 Odds ratios for the association between first-trimester use of antifungal drugs and the risk of selected birth defects, National Birth Defects Prevention Study (1997-2003) Adjusted odds Cases Controls ratio (95% Cases, Controls, exposed, exposed, Covariates included in the confidence Case group na na n (%) n (%) regression model interval) Neural tube defects 699 4526 16 (2.3) 81 (1.8) Periconceptional folic acid use, 1.25 (0.72, 2.15) urinary tract infections in the first trimester, prepregnancy body mass index, maternal education ...... Spina bifida 441 4581 11 (2.5) 81 (1.8) Periconceptional folic acid use, 1.40 (0.74, 2.66) prepregnancy body mass index ...... Cleft lip with or without 1040 4415 23 (2.2) 78 (1.8) Urinary tract infections in the 1.24 (0.78, 1.99) cleft palateb first trimester, prepregnancy body mass index ...... Anorectal atresia 199 4565 5 (2.5) 81 (1.8) Prepregnancy body mass index, 1.42 (0.66, 3.06) maternal education ...... Hypospadias (second and 775 2397 13 (1.7) 44 (1.8) Maternal education 0.86 (0.46, 1.61) third degree)c ...... Craniosynostosis 440 4581 8 (1.8) 81 (1.8) Prepregnancy body mass index 1.02 (0.49, 2.13) ...... Gastroschisis 450 4733 5 (1.1) 84 (1.8) Urinary tract infections in the 0.64 (0.25, 1.62) first trimester, maternal age ...... Diaphragmatic hernia 274 4717 9 (3.3) 84 (1.8) Hormonal contraceptives use in 1.88 (0.93, 3.78) the month before conception through the first trimester ...... Conotruncal heart defects 707 4581 15 (2.1) 81 (1.8) Prepregnancy body mass index 1.20 (0.69, 2.10) ...... Tetralogy of Fallot 250 4581 5 (2.0) 81 (1.8) Prepregnancy body mass index 1.11 (0.45, 2.77) ...... Left ventricular outflow 513 4698 8 (1.6) 84 (1.8) Antibiotics use in the month 0.84 (0.40, 1.74) tract obstructive heart before conception through the defects first trimester ...... Hypoplastic left heart 176 4581 7 (4.0) 81 (1.8) Prepregnancy body mass index 2.30 (1.04, 5.06) syndrome with intact ventricular septum ...... Right ventricular outflow 507 4581 9 (1.8) 81 (1.8) Prepregnancy body mass index 1.00 (0.50, 1.99) tract obstructive heart defects ...... Ventricular septal defects 709 4698 9 (1.3) 84 (1.8) Antibiotics use in the month 0.70 (0.35, 1.39) before conception through the first trimester ...... Perimembranous 531 4698 7 (1.3) 84 (1.8) Antibiotics use in the month 0.72 (0.33, 1.57) ventricular septal defects before conception through the first trimester ...... Atrial septal defects 529 4733 5 (1.0) 84 (1.8) Urinary tract infections in the 0.52 (0.21, 1.28) (secundum type or not first trimester otherwise specified) ...... a Number after excluding subjects with missing values for covariates in unconditional logistic regression analysis. b Subjects from the Utah study center were excluded from the control group in analyses with oral cleft cases because these cases were ascertained in all study centers except Utah. c Hypospadias cases were compared with male controls. Carter. Antifungal drugs and the risk of selected birth defects. Am J Obstet Gynecol 2008.

FEBRUARY 2008 American Journal of Obstetrics & Gynecology 191.e5 Research Obstetrics www.AJOG.org toconazole. Previous reports indicate months) likely contributed to the failure nation for the results. Data were not avail- that miconazole use does not increaseof mothers to recall medication use, al- able for nonparticipants (subjects who the risk of birth 17defects, but the evi-though this period was implemented to were eligible but did not perform an inter- dence is less consistent for ketoconazole allow for attempts to contact mothers view), and it is unknown whether the ex- and fluconazole. A Hungarian study who were difficult to locate and to obtain posure-outcome association in this group found an increased risk of cardiovascular sufficient medical data to confirm case differed from that of participants; how- defects after oral ketoconazole exposure, diagnoses. Recall bias could have oc- ever, it was unlikely that the participation but this was based on only 2 exposed curred if 1 study group underreported of cases and controls was related to the use cases and was not statistically signifi- exposure, compared with another of antifungal drugs in pregnancy. cant.18 Case reports have described agroup, pat- or if case mothers were more The results indicated that first-trimes- tern of craniofacial and skeletal malfor- likely than control mothers to recall ter use of antifungal drugs did not in- mations after fluconazole exposure,19-21 medication use to explain the occurrence crease the risk for most of the birth defect but these findings were not observed in of birth defects in their offspring. There- groups studied. Antifungal drugs are of- the current study. Studies that have used fore, recall bias cannot be ruled out as a ten used by pregnant and nonpregnant computerized medical or prescription possible explanation for the results. women without known adverse conse- databases to identify women exposed Limitations in our exposure assess- quences, and there is evidence of a re- during pregnancy to fluconazole, mi- ment also included the absence of infor- duced risk of preterm birth with gesta- conazole, ketoconazole, econazole, itra- mation on dose and route of administra- tional clotrimazole 27use; therefore, conazole, clotrimazole, and nystatin tion. For example, mothers exposed to there needs to be confirmation of a real found no increased risk of birth defects ketoconazole reported the drug name as association with birth defects before cau- overall after first-trimester expo- Nizoral, which is available for topical tioning against use of these drugs in sure.22,23 No published reports relating(McNeil Consumer, Fort Washington, pregnancy. The increased risk of hypo- to topical terconazole use and the risk of PA) or oral (Janssen Pharmaceutica, Ti- plastic left heart syndrome associated birth defects in humans were found. tusville, NJ) use, and the dose to which with antifungals should be considered a One of the limitations of our study was the mother (and fetus) is exposed is ex- preliminary result and causal inferences probable underreporting of exposure. pected to vary, depending on the route of avoided, particularly because use among Approximately 10-20% of women have administration. Another limitation was exposed cases was not confined to a sin- vulvovaginal candidiasis during preg- that individual medications could not be gle antifungal drug. Further studies nancy24,25; however, in this study, identifiedonly when mothers reported use of should examine the risk of this specific 1.8% of control mothers reported first- unknown antifungal drugs or unknown defect in relation to antifungal drug ex- trimester antifungal use. Therefore, it vaginal creams; therefore, the number of posure and should attempt to separate was estimated that at least 80% of af- women actually exposed to the individ- the effect of the medication from that of fected women did not report their yeast ual antifungal drugs listed in Tablethe infection.1 is The observation that 2 of infection and associated antifungal use. likely to be higher than that presented. the 4 cases exposed to ketoconazole had Missing from the interview was an item The unknown drugs reported may have isolated cardiovascular defects (right that asked specifically about use of anti- also included antifungal medications ventricular outflow tract obstructive de- fungal drugs based on drug name or clin- other than those in Table 1. fect–pulmonary valve stenosis and hy- ical indication. Instead, reports of anti- In addition, it could not be clearly dis- poplastic left heart syndrome) warrants fungal drug use were provided in tinguished whether the observed in- that attention also be focused on this response to a question about the occur- crease in the risk of certain birth defects particular exposure. f rence of fever or to a general question was due to the maternal infection or to about the presence of diseases such as the antifungal drugs used to treat the in- ACKNOWLEDGMENTS chronic, infectious, or sexually transmit- fection. Very few subjects reported having We thank Lenore Gensburg for her assistance ted diseases during pregnancy; thus, un- a fungal infection and being untreated; with statistical analysis and providing com- derreporting of drug use may have con- therefore, the risk of birth defects in in- ments on the manuscript. We appreciate the tributed to the small number of exposed fected but untreated subjects could not be contributions made by the study centers that subjects in the study. obtained. Immunocompromised condi- include the Arkansas Reproductive Health In addition, many antifungal medications can increase the probability of devel- Monitoring System, the California Birth Defects 26 Monitoring Program, the Metropolitan Atlanta tions for the treatment of vulvovaginal oping fungal infections; however, sepa-Congenital Defects Program, the Iowa Birth De- candidiasis are available as over-the- rate analyses could not be performed for fects Registry, the Massachusetts Center for counter creams with a recommended women having these conditions. A disad- Birth Defects Research and Prevention Moni- duration of use of 1-7 days, and it is pos- vantage of the study was that these women toring Program, the Birth Defects and Special sible that case and control mothers could not be reliably identified because no Needs Registry of New Jersey, the New York State Congenital Malformations Registry, the tended to forget exposure to medications interview question asked specifically about Texas Birth Defects Monitoring Division, the used for such a short period. The long such conditions. Finally, we considered North Carolina Birth Defects Monitoring Pro- time-to-interview period (up to 24 participation bias as an alternative expla- gram, and the Utah Birth Defects Network.

191.e6 American Journal of Obstetrics & Gynecology FEBRUARY 2008 www.AJOG.org Obstetrics Research

REFERENCES plastic left heart syndrome in eastern Wisconsin treatment for birth outcomes. Congenit Anom 1. Centers for Disease Control and Prevention. for birth cohorts 1997-1999. Birth Defects Res (Kyoto) 2005;45:5-8. Sexually transmitted diseases treatment guide- A Clin Mol Teratol 2004;70:114-20. 19. Lee BE, Feinberg M, Abraham JJ, Murthy lines 2002. MMWR Recomm Rep 2002;51:1-78. 11. Ferencz C, Loffredo C, Correa-Villasenor A, AR. Congenital malformations in an infant born 2. Chaube S, Murphy ML. The teratogenic ef- Wilson P. Genetic and environmental risk fac- to a woman treated with fluconazole. Pediatr fects of 5-fluorocytosine in the rat. Cancer Res tors of major cardiovascular malformations: the Infect Dis J 1992;11:1062-4. 1969;29:554-7. Baltimore-Washington Infant Study, 1981- 20. Pursley TJ, Blomquist IK, Abraham J, 3. Menegola E, Broccia ML, Di Renzo F, Giavini 1989. Armonk (NY): Futura Publishing Co; Andersen HF, Bartley JA. Fluconazole-induced E. Pathogenic pathways in fluconazole-induced 1997. p. 178–89. congenital anomalies in three infants. Clin Infect branchial arch malformations. Birth Defects Res 12. Vukovich A, Heald A, Darragh A. Vaginal Dis 1996;22:336-40. A Clin Mol Teratol 2003;67:116-24. absorption of two imidazole antifungal agents, 21. Aleck KA, Bartley DL. Multiple malformation 4. Czeizel AE, Toth M, Rockenbauer M. No ter- econazole and miconazole. Clin Pharmacol syndrome following fluconazole use in preg- atogenic effect after clotrimazole therapy during Ther 1977;21:121. nancy: report of an additional patient. Am J Med pregnancy. Epidemiology 1999;10:437-40. 13. Ritter W, Patzschke K, Krause U, Stetten- Genet 1997;72:253-6. 5. Czeizel AE, Kazy Z, Puho E. Tolnaftate spray dorf S. Pharmacokinetic fundamentals of vagi- 22. Jick SS. Pregnancy outcomes after mater- treatment during pregnancy. Reprod Toxicol nal treatment with clotrimazole. Chemotherapy nal exposure to fluconazole. Pharmacotherapy 2004;18:443-44. 1982;28(Suppl 1):37-42. 1999;19:221-2. 6. Czeizel AE, Kazy Z, Puho E. A population- 14. Daneshmend TK. Systemic absorption of 23. Rosa FW, Baum C, Shaw M. Pregnancy based case-control teratological study of oral miconazole from the vagina. J Antimicrob Che- outcomes after first-trimester vaginitis drug nystatin treatment during pregnancy. Scand mother 1986;18:507-11. therapy. Obstet Gynecol 1987;69:751-5. J Infect Dis 2003;35:830-35. 15. Larosa E, Cauwenbergh G, Cilli P, 24. Meis PJ, Goldenberg RL, Mercer B, et al. 7. Czeizel AE, Kazy Z, Vargha P. A population- Woestenborghs R, Heykants J. Itraconazole The preterm prediction study: significance of based case-control teratological study of vaginal pharmacokinetics in the female genital tract: vaginal infections. Am J Obstet Gynecol econazole treatment during pregnancy. Eur J Ob- plasma and tissue levels in patients undergoing 1995;173:1231-5. stet Gynecol Reprod Biol 2003;111:135-40. hysterectomy after a single dose of 200 mg itra- 25. Cotch MF, Hillier SL, Gibbs RS, Eschen- 8. Yoon PW, Rasmussen SA, Lynberg MC, et conazole. Eur J Obstet Gynecol Reprod Biol bach DA. Epidemiology and outcomes associ- al. The National Birth Defects Prevention Study. 1986;23:85-9. ated with moderate to heavy Candida coloniza- Public Health Rep 2001;116:32-40. 16. Force RW. Fluconazole concentrations in tion during pregnancy. Am J Obstet Gynecol 9. Rasmussen SA, Olney RS, Holmes LB, Lin breast milk. Pediatr Infect Dis J 1995;14:235-6. 1998;178:374-80. AE, Keppler-Noreuil KM, Moore CA. Guidelines 17. Czeizel AE, Kazy Z, Puho E. Population- 26. Sobel JD. Vulvovaginitis in healthy women. for case classification for the National Birth De- based case-control teratologic study of topical Compr Ther 1999;25:335-46. fects Prevention Study. Birth Defects Res A Clin miconazole. Congenit Anom (Kyoto) 2004; 27. Czeizel AE, Fladung B, Vargha P. Preterm Mol Teratol 2003;67:193-201. 44:41-5. birth reduction after clotrimazole treatment dur- 10. Cronk CE, Pelech AN, Malloy ME, Mc- 18. Kazy Z, Puho E, Czeizel AE. Population- ing pregnancy. Eur J Obstet Gynecol Reprod Carver DG. Excess birth prevalence of hypo- based case-control study of oral ketoconazole Biol 2004;116:157-63.

FEBRUARY 2008 American Journal of Obstetrics & Gynecology 191.e7