*Correspondence to Dr. Allen A. Mitchell, Slone Epidemiology Center, 1010 Commonwealth Avenue, Boston, MA 02215 (e-mail: ude.ub@timnella).
Contributed by
Abbreviations: BDS, Slone Epidemiology Center Birth Defects Study; CI, confidence interval; OR, odds ratio; OTC, over-the-counter.
Received 2012 Jun 8; Accepted 2012 Oct 19.
Copyright © The Author 2013. Published by Oxford University Press on behalf of the Johns Hopkins Bloomberg School of Public Health. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com.
Associated Data
Supplementary MaterialsWeb Material
supp_178_2_198__index.html (816 bytes)
GUID: B0585850-00B9-43AF-8A69-381FE7AD8653
supp_kws427_kws427_Yau_Web_Material.Final.pdf (149K)
GUID: 90477032-8919-4111-89CB-27001E4550AD
Abstract
Previous studies suggested that early pregnancy exposure to specific oral decongestants increases the risks of several birth defects. Using January 1993–January 2010 data from the Slone Epidemiology Center Birth Defects Study, we tested those hypotheses among 12,734 infants with malformations (cases) and 7,606 nonmalformed control infants in the United States and Canada. Adjusted odds ratios and 95% confidence intervals were estimated for specific birth defects, with controlling for potential confounders. Findings did not replicate several hypotheses but did support 3 previously reported associations: phenylephrine and endocardial cushion defect (odds ratio = 8.0; 95% confidence interval: 2.5, 25.3; 4 exposed cases), phenylpropanolamine and ear defects (odds ratio = 7.8; 95% confidence interval: 2.2, 27.2; 4 exposed cases), and phenylpropanolamine and pyloric stenosis (odds ratio = 3.2; 95% confidence interval: 1.1, 8.8; 6 exposed cases). Hypothesis-generating analyses involving multiple comparisons identified a small number of associations with oral and intranasal decongestants. Accumulating evidence supports associations between first-trimester use of specific oral and possibly intranasal decongestants and the risk of some infrequent specific birth defects.
Keywords: birth defects, decongestants, maternal exposure, pregnancy
Decongestants, particularly pseudoephedrine, are among the over-the-counter (OTC) medications most commonly used during pregnancy (1). Because they are available without prescription, they are widely perceived as safe by healthcare providers and pregnant women, yet the safety of prenatal exposure with regard to specific birth defects remains unclear. Given their extensive use, even a small increase in the risk of birth defects would have considerable public health implications.
All decongestants are vasoconstrictive and could share class-related effects. On the other hand, teratogenicity can vary within a given class (2), which makes it necessary that consideration also be given to specific decongestants. Epidemiologic studies of specific decongestants have identified elevated risks of specific birth defects, including defects of the heart, eyes and ears, gut, abdominal wall, and feet (Web Table 1, available at //aje.oxfordjournals.org/) (3–16).
Using data from the Slone Epidemiology Center Birth Defects Study (BDS), we tested the previously reported associations between specific birth defects and first-trimester exposure to oral decongestants and, in hypothesis-generation analyses, explored the risks of other common major congenital malformations. In addition, we explored the risk of specific birth defects in relation to first-trimester use of intranasal decongestants.
MATERIALS AND METHODS
Study population
The BDS was initiated in 1976 as a case-control surveillance program of birth defects in relation to medications. Details of the methods have been described (17, 18). The present analysis was based on data from subjects interviewed between January 1993 and January 2010. During those years, study nurses interviewed women within 6 months of delivery who had malformed infants or fetuses identified at regional centers around Boston, Massachusetts; Philadelphia, Pennsylvania; Toronto, Ontario; and San Diego, California; as well as from birth defects registries in Massachusetts and New York State. Mothers of nonmalformed (control) infants were identified at participating birth hospitals and from a population-based sample of nonmalformed infants in Massachusetts. Between January 1993 and June 1998, mothers of 3,723 cases and 781 controls were interviewed in person (“phase 1”), and between August 1998 and January 2010, mothers of 12,746 cases and 6,826 controls were interviewed by telephone (“phase 2”). Overall, among mothers of eligible cases and controls who could be contacted, the proportions who completed the interview were 69% and 67%, respectively (Web Figure). The study was approved by the institutional review boards at relevant institutions, and is compliant with requirements of the Health Insurance Portability and Accountability Act.
Cases and controls
Cases eligible for final analyses consisted of 12,734 infants and fetuses with a confirmed diagnosis of major congenital malformation(s), identified up to a minimum of 3 and a maximum of 5 months after delivery. Infants with chromosomal defects, known Mendelian inherited disorders, syndromes, DiGeorge sequence (associated with 22q deletion), and defects associated with prematurity were excluded. Cases with multiple anomalies were considered in each appropriate defect category. Controls eligible for final analyses consisted of 7,606 nonmalformed infants.
Exposure ascertainment
Subjects who provided informed consent for interview were asked to provide a release for their infant's medical record. Using highly structured interview procedures and questionnaires, trained study nurses interviewed mothers within 6 months of delivery. Information was sought about demographic characteristics; reproductive, medical, and lifestyle factors; diet; and details of exposures to all medications (prescription and OTC) at any time from 56 days before the date of the last menstrual period through the end of pregnancy. Mothers and study nurses were unaware of the various hypotheses under consideration.
To maximize recall of medication exposures and minimize errors in reported exposures, highly structured questionnaires included a series of increasingly detailed questions (19). Interviewers first asked about the occurrence of any of a list of illnesses (e.g., infections) during pregnancy and the medications taken for those illnesses, then about use of categories of medications (e.g., antibiotics, nasal sprays), and finally about use of specific products, including brand and generic names. Mothers who reported taking a medication were asked to identify the dates when use began and ended; recall was enhanced by a calendar highlighting key dates and events (e.g., last menstrual period, Christmas, delivery date). Subjects were also asked about the medication's dose and form and the frequency and duration of use. A Medication Identification Booklet with photographs of specific OTC medications was used to enhance recall of OTC products. All medications were coded via the Slone Drug Dictionary, which links each product to its individual active ingredients. Changes in product formulations over time are incorporated in the Drug Dictionary, and new formulations are coded accordingly.
Exposure definition and classification
The date of the last menstrual period was based on early ultrasound examination or maternal recall. We defined the estimated date of conception as 14 days after the date of the last menstrual period, the first trimester as 90 days after the estimated date of conception (encompassing the etiologically important period of structural development for most organs), the second trimester as 90 days after the first trimester, and the third trimester as the period after the second trimester through the end of pregnancy. We considered exposure in any given trimester to include maternal use of a decongestant on at least 1 day.
Using our exposure classification algorithm (20), which considers recall uncertainty in timing of medication exposure on the basis of the subject's self-reported precision of the start/stop dates and duration of use, we classified decongestant exposure into the following categories: “unexposed at any time from 56 days before the last menstrual period date through the end of pregnancy”; “likely exposed” and “possibly exposed” in a given trimester; and “exposed only outside a given trimester.” For each specific decongestant, our exposure group of primary interest was subjects classified as “likely exposed” to only the specific decongestant, without exposure to other decongestants in the same trimester. The reference group for all analyses was subjects unexposed to any decongestant.
Statistical analyses
We assessed specific defects previously reported to be associated with 1 or more decongestants: eye defects, ear defects, ventricular septal defect, coarctation of the aorta, endocardial cushion defect, pyloric stenosis, small-intestinal atresia/stenosis, clubfoot, gastroschisis, and hemifacial microsomia. In addition, for 21 case groups that included at least 100 subjects each, we explored possible risks for other common specific defects.
We estimated separate odds ratios and 95% confidence intervals for each specific birth defect in relation to each specific decongestant by multivariate conditional logistic regression, with calendar year of interview and geographic region of birth hospital as the matching factors to control for the effects of secular trends and regional variations in recruitment of study subjects and exposures. Potential confounders included maternal race/ethnicity; age; prepregnancy weight and body mass index; educational level; smoking; alcohol consumption; diabetes; hypertension; asthma; allergy; first-trimester episodes of convulsions, respiratory infection, and fever; first-trimester use of acetaminophen, aspirin, ibuprofen, nonaspirin nonsteroidal antiinflammatory drugs overall, antihistamines, cough medications, and antibiotics; periconceptional folic acid supplementation; family income; family history of birth defects; number of fetuses; planned pregnancy; and number of live births. We adjusted for the potential confounders associated with each specific decongestant and birth defect in our population. For our exploratory analyses, we did not estimate odds ratios for exposures with fewer than 5 exposed cases or controls because of the instability of such estimates.
Because pseudoephedrine is often combined with acetaminophen, and because a previous study that had hypothesized an association with gastroschisis and small-intestinal atresia/stenosis (7) found the highest risk for the combination of pseudoephedrine and acetaminophen, we further stratified pseudoephedrine exposure according to whether it was taken as a single-component product or was combined with acetaminophen. To assess the sensitivity of results to timing of exposure, we carried out separate analyses for second- and third-trimester exposures, restricted to subjects who had not been exposed to the specific decongestant in the first trimester. All analyses were performed in SAS for Microsoft Windows, version 9.2 (SAS Institute Inc., Cary, North Carolina).
RESULTS
Among mothers of nonmalformed infants (n = 7,606), 8.6% of subjects (n = 656) reported taking 1 or more oral or intranasal decongestants during the first trimester. Use of oral products was reported by 7.8% of subjects (n = 592); 7.0% (n = 531) took pseudoephedrine only, 0.4% (n = 28) phenylephrine only, and 0.4% (n = 30) phenylpropanolamine only. Use of intranasal decongestants was reported by 1.2% of subjects (n = 92); oxymetazoline was the most common (0.9%, n = 66). Factors associated with first-trimester decongestant use are presented in Table 1.
Table 1.
Selected Maternal Characteristics in Relation to First-Trimester Decongestant Use Among Mothers of Nonmalformed Infants, Slone Epidemiology Center Birth Defects Study, 1993–2010
CharacteristicNonusersa
(n = 5,594)Usersb
(n = 656)Crude Matched ORc95% CINo.%No.%Maternal race/ethnicity Non-Hispanic white3,85468.956085.41.0Referent Hispanic82214.7497.50.40.3, 0.6 Non-Hispanic black4738.5253.80.30.2, 0.5 Non-Hispanic Asian, Pacific Islander3185.7132.00.30.2, 0.5 Others1272.391.40.50.3, 1.1Maternal age, years <251,24922.39414.31.0Referent 25–291,48026.517526.71.71.3, 2.2 30–331,50026.821432.62.01.5, 2.5 ≥341,34624.117326.41.71.3, 2.3 Not availabled190.30Maternal prepregnancy weight, kg <551,43025.613921.21.0Referent 55–601,44025.714522.11.00.8, 1.3 61–691,27222.716925.81.41.1, 1.7 ≥701,39424.919730.01.41.1, 1.8 Not availabled581.060.90.90.4, 2.2Maternal educational level, years <131,67830.013420.40.60.5, 0.8 13–151,31723.515623.81.0Referent ≥162,59646.436655.81.21.0, 1.5 Not availabled30.10Maternal smoking Never3,29158.833951.71.0Referent Before pregnancy only1,27022.718828.71.41.2, 1.7 During pregnancy94316.912719.41.31.0, 1.6 Not availabled901.620.3Maternal alcohol consumption Never2,26040.418327.91.0Referent Before pregnancy only95117.011717.81.51.2, 2.0 During pregnancy2,38342.635654.31.91.6, 2.3Maternal conditions and medicationse Preexisting diabetes280.550.81.60.6, 4.1 Preexisting hypertension721.381.21.00.5, 2.1 Asthma during pregnancy3746.77912.01.81.4, 2.4 Allergies during pregnancy1,20021.536555.64.74.0, 5.6 Respiratory infection in first trimester1,48026.543766.67.45.9, 9.2 Fever in first trimester55810.013420.42.21.8, 2.8 Acetaminophen use in first trimester3,05354.654683.24.43.4, 5.7 Aspirin use in first trimester3756.76610.11.51.2, 2.0 Ibuprofen use in first trimester1,13420.320731.61.91.6, 2.3 Oral antihistamine use in first trimester4768.527541.99.67.9, 11.7 Cough medication use in first trimesterf1302.313520.69.97.6, 12.9 Systemic antibiotic use in first trimester60810.913620.72.11.7, 2.7 Periconceptional use of folic acidg3,53463.246971.51.81.1, 2.9
Open in a separate window
Abbreviations: CI, confidence interval; OR, odds ratio.
a Not exposed to any decongestant at any time from 56 days before the date of the last menstrual period through the end of pregnancy.
b Likely exposed to any decongestant during the first trimester of pregnancy.
c Matched on calendar year of interview and geographic region of birth hospital. Odds ratios not estimated for fewer than 5 decongestant users.
d Not available because of missing data.
e Reference category for each variable is “none during pregnancy,” except for diabetes and hypertension, where reference is “never”.
f Cough medications included antitussives (dextromethorphan, benzonatate), expectorants (guaifenesin, guaiacolsulfonate potassium, ammonium carbonate, iodinated glycerol), and mucolytics (acetylcysteine, carbocysteine).
g Any exposure from 1 month before to 1 month after the estimated date of conception.
Oral decongestants
Assessment of hypothesized associations
For pseudoephedrine and ventricular septal defect (Table 2), the risk for single-component pseudoephedrine approximated the null (odds ratio (OR) = 1.1, 95% confidence interval (CI): 0.8, 1.5), but use of pseudoephedrine/acetaminophen was associated with an increased risk (OR = 1.8, 95% CI: 1.2, 2.7). Acetaminophen itself did not affect the risk (OR = 1.0, 95% CI: 0.9, 1.2). For phenylephrine, we observed an elevated risk for endocardial cushion defect (OR = 8.0, 95% CI: 2.5, 25.3; 4 exposed cases). All 4 exposed cases had multiple anomalies, including other cardiovascular defects, and 1 had a limb defect. For phenylpropanolamine, we found increased risks of ear defects (OR = 7.8, 95% CI: 2.2, 27.2; 4 exposed cases) and pyloric stenosis (OR = 3.2, 95% CI: 1.1, 8.8; 6 exposed cases). None of the aforementioned associations was observed for second- or third-trimester exposure (Web Table 2).
Table 2.
Odds Ratios and 95% Confidence Intervals for First-Trimester Exposurea to Oral Decongestants in Relation to Specific Birth Defects Previously Reported to Be Associated With Oral Decongestant Use, Slone Epidemiology Center Birth Defects Study, 1993–2010
Outcome (Isolated and Multiple Defects Combined)bTotal No.Oral Decongestants Only
(n = 1,537)Pseudoephedrine Onlyc
(n = 1,283)No.%Adjusted ORd95% CINo.%Adjusted ORd95% CINo malformations7,6065527.31.0Referent4796.31.0ReferentEye defects207167.71.0e0.5, 1.7146.81.1f0.6, 2.0Ear defects178169.01.5g0.9, 2.7116.21.4g0.7, 2.7Ventricular septal defect1,7871548.61.31.1, 1.61246.91.31.0, 1.6Coarctation of the aorta369297.91.3h0.8, 2.0205.41.1h0.7, 1.8Endocardial cushion defect1621811.12.0i1.1, 3.6106.21.1j0.6, 2.3Pyloric stenosis757557.31.1k0.8, 1.5455.91.1k0.8, 1.5Small-intestinal atresia/stenosisl224177.61.4m0.8, 2.4156.71.6m0.9, 2.8Clubfootn543397.21.10.8, 1.6346.31.20.8, 1.7Gastroschisis258207.81.7o1.0, 2.9155.81.5o0.8, 2.8Phenylephrine Onlyc
(n = 82)Phenylpropanolamine Onlyc
(n = 57)No.%Adjusted ORd95% CINo.%Adjusted ORd95% CINo malformations260.31.0Referent160.21.0ReferentEye defects10.50Ear defects10.642.27.82.2, 27.2Ventricular septal defect120.71.40.7, 2.970.41.90.7, 4.9Coarctation of the aorta30.810.3Endocardial cushion defect42.58.0f2.5, 25.310.6Pyloric stenosis20.360.83.2j1.1, 8.8Small-intestinal atresia/stenosisl10.40Clubfootn20.40Gastroschisis31.20
Open in a separate window
Abbreviations: CI, confidence interval; OR, odds ratio.
a Refers to subjects likely exposed to decongestant in first trimester (n = 1,833) among a total of 20,340 subjects.
b Hemifacial microsomia was excluded from this table because the total number of cases (n = 47) was too small to allow evaluation.
c Not exposed to other decongestants in the first trimester.
d All odds ratios and 95% confidence intervals were estimated on the basis of conditional logistic regression matched on calendar year of interview and geographic region of birth hospital. In addition, for each exposure-outcome association, we adjusted for factors associated with the specific decongestant and the specific birth defect in our population that also materially changed the odds ratio estimate, as specified in the following footnotes. Odds ratios were not estimated for fewer than 4 exposed cases.
e Adjusted for first-trimester episodes of respiratory infection and first-trimester use of cough medications.
f Adjusted for first-trimester episodes of respiratory infection.
g Adjusted for maternal race/ethnicity.
h Adjusted for maternal race/ethnicity and first-trimester use of cough medications.
i Adjusted for alcohol consumption, first-trimester episodes of respiratory infection, and first-trimester use of antihistamines.
j Adjusted for first-trimester use of cough medications.
k Adjusted for maternal race/ethnicity and educational level.
l Restricted to non-gastroschisis cases, so as to exclude cases of small-intestinal atresia/stenosis, which can be secondary to gastroschisis.
m Adjusted for first-trimester use of antihistamines.
n Restricted to non-neural tube defect cases, so as to exclude cases of clubfoot, which can be secondary to a neural tube defect.
o Adjusted for maternal age, prepregnancy weight, educational level, and smoking.
For pseudoephedrine and gastroschisis (Table 2), use of single-component pseudoephedrine was associated with an increased risk (OR = 3.0, 95% CI: 1.4, 6.2), whereas only 1 gastroschisis case was exposed to pseudoephedrine/acetaminophen; conversely, for small-intestinal atresia/stenosis, single-component pseudoephedrine was associated with a smaller risk than the combination product (odds ratios of 1.4 and 2.3, respectively), but their lower confidence bounds included 1.0. Acetaminophen itself was not associated with gastroschisis (OR = 0.9, 95% CI: 0.7, 1.3) or small-intestinal atresia/stenosis (OR = 1.1, 95% CI: 0.8, 1.5).
For hemifacial microsomia, the total number of cases (n = 47) was too small to allow evaluation.
Our findings did not support most other previously hypothesized associations between first-trimester exposures to specific oral decongestants and specific defects; these include pseudoephedrine and coarctation of the aorta (OR = 1.1, 95% CI: 0.7, 1.8), phenylephrine and eye and ear defects (only 1 exposed case for each defect group) and clubfoot (only 2 exposed cases), and phenylpropanolamine and eye defects and gastroschisis (no exposed cases for either defect).
Exploratory analyses
In our exploratory analyses of exposures and defects not previously hypothesized to be associated with decongestants, there were 84 comparisons of first-trimester exposure to oral decongestants overall, pseudoephedrine, phenylephrine, and phenylpropanolamine in relation to 21 other specific birth defect groups with at least 100 cases each. These analyses yielded adjusted odds ratio estimates ranging from 0.7 to 2.4 (Table 3).
Table 3.
Exploratory Analyses of First-Trimester Exposurea to Pseudoephedrineb in Relation to Specific Birth Defects Not Previously Reported to Be Associated With Decongestant Use, Slone Epidemiology Center Birth Defects Study, 1993–2010
Outcome (Isolated and Multiple Defects Combined)Total No.Pseudoephedrine Onlyc (n = 1,283)No.%Adjusted ORd95% CINo malformations7,6064796.31.0ReferentAny neural tube defect531336.21.10.7, 1.7 Spina bifida364195.20.9e0.5, 1.5 Anencephaly11087.31.2f0.5, 3.3Agenesis/dysgenesis/other anomalies of the corpus callosum11497.91.3g0.6, 2.7Cleft palate alone526315.91.0h0.6, 1.4Cleft lip with or without cleft palate979646.51.31.0, 1.8Tracheo-esophageal fistula257145.41.20.7, 2.1Anal atresia/stenosis274165.81.30.8, 2.3Hirschsprung's disease10532.9Intestinal malrotation21194.30.7h0.4, 1.5Cryptorchidismi535234.31.10.6, 1.8Hypospadiasi669416.11.2j0.8, 1.9Renal agenesis and dysgenesis241114.60.80.4, 1.5Cystic kidney disease240145.81.10.6, 2.0Renal collecting system anomalies1,092837.61.31.0, 1.8Extra or horseshoe kidney179105.61.00.5, 2.0Limb reduction defects2312711.72.41.5, 3.7Craniosynostosis166137.81.5k0.8, 2.8Diaphragmatic hernia248166.51.40.8, 2.4Omphalocele14264.21.2l0.5, 2.9Inguinal hernia14264.20.9m0.4, 2.1
Open in a separate window
Abbreviations: CI, confidence interval; OR, odds ratio.
a Refers to subjects likely exposed in first trimester.
b Results for first-trimester exposure to phenylephrine only (n = 82) and phenylpropanolamine only (n = 57) are not presented in this table because there were fewer than 5 exposed cases for each specific birth defect, except for phenylephrine and cleft palate alone (adjusted OR = 1.7, 95% CI: 0.6, 4.8; 5 exposed cases; adjusted for first-trimester use of cough medications).
c Not exposed to other decongestants in the first trimester.
d All odds ratios and 95% confidence intervals were estimated on the basis of conditional logistic regression matched on calendar year of interview and geographic region of birth hospital. In addition, for each exposure-outcome association, we adjusted for factors associated with the specific decongestant and the specific birth defect in our population that also materially changed the odds ratio estimate, as specified in the following footnotes. Odds ratios were not estimated for fewer than 5 exposed cases for exploratory analyses.
e Adjusted for prepregnancy weight, educational level, and first-trimester use of cough medications.
f Adjusted for alcohol consumption, allergy, and first-trimester episodes of respiratory infection.
g Adjusted for maternal educational level and allergy.
h Adjusted for first-trimester use of cough medications.
i Restricted to males only. Nonmalformed controls (restricted to males only): n = 3,768.
j Adjusted for first-trimester use of acetaminophen and cough medications.
k Adjusted for maternal race/ethnicity and allergy.
l Adjusted for maternal race/ethnicity and first-trimester episodes of respiratory infection.
m Adjusted for first-trimester use of antihistamines and cough medications.
Only 1 of these comparisons had lower 95% confidence bounds that exceeded 1.0: pseudoephedrine and limb reduction defects (OR = 2.4, 95% CI: 1.5, 3.7) (Table 3); this association was not observed for second- or third-trimester exposures (Web Table 2). The risk was elevated both for first-trimester use of pseudoephedrine alone (OR = 2.5, 95% CI: 1.4, 4.5) and for pseudoephedrine/acetaminophen (OR = 3.5, 95% CI: 1.7, 7.2). Acetaminophen itself did not affect the risk (OR = 1.1, 95% CI: 0.8, 1.6).
Intranasal decongestants
In our exploratory analyses of first-trimester exposure to intranasal decongestants and defects not previously reported to be associated with intranasal decongestant use, we conducted 150 comparisons that included intranasal decongestants overall, imidazoline derivatives overall, oxymetazoline, xylometazoline, and non-imidazoline derivatives in relation to 30 common specific birth defects with at least 100 cases each. The adjusted odds ratio estimates ranged between 0.7 and 3.3 (Table 4), with 95% confidence intervals for 2 specific malformations with first-trimester exposure excluding the null.
Table 4.
Exploratory Analyses of First-Trimester Exposurea to Specific Intranasal Decongestantsb in Relation to Specific Birth Defects, Slone Epidemiology Center Birth Defects Study, 1993–2010c
Outcome (Isolated and Multiple Defects Combined)Total No.Intranasal Decongestants Onlyd
(n = 162)Imidazoline Derivatives Onlyd
(n = 130)No.%Adjusted ORe95% CINo.%Adjusted ORe95% CINo malformations7,606540.71.0Referent420.61.0ReferentSpecific birth defects previously reported to be associated with oral decongestant use Ventricular septal defect1,787160.91.20.7, 2.1130.71.40.7, 2.8 Pyloric stenosis757152.01.9f1.0, 3.5121.62.2f1.1, 4.5 Clubfooth54350.91.10.4, 2.740.7Specific birth defects not previously reported to be associated with decongestant use Any neural tube defect53150.90.9i0.3, 2.640.8 Spina bifida36451.41.3j0.5, 3.641.1 Cleft lip with or without cleft palate979141.41.40.7, 2.8111.11.70.8, 3.7 Tracheo-esophageal fistula25751.92.1g0.8, 5.551.93.3g1.2, 8.9 Hypospadiask66950.71.10.4, 3.140.6 Renal collecting system anomalies1,09250.50.70.3, 1.940.4Imidazoline Derivative: Oxymetazoline Onlyd (n = 82)Imidazoline Derivative: Xylometazoline Onlyd (n = 48)No.%Adjusted ORe95% CINo.%Adjusted ORe95% CINo malformations330.41.0Referent90.11.0ReferentSpecific birth defects previously reported to be associated with oral decongestant use Ventricular septal defect80.41.40.6, 3.150.31.60.5, 4.7 Pyloric stenosis50.71.8g0.7, 4.870.92.91.0, 7.9 Clubfooth10.230.6Specific birth defects not previously reported to be associated with decongestant use Any neural tube defect30.610.2 Spina bifida30.810.3 Cleft lip with or without cleft palate60.61.40.5, 3.950.52.20.7, 7.0 Tracheo-esophageal fistula31.220.8 Hypospadiask30.410.1 Renal collecting system anomalies20.220.2
Open in a separate window
Abbreviations: CI, confidence interval; OR, odds ratio.
a Refers to subjects likely exposed to decongestant in first trimester (n = 1,833) among a total of 20,340 subjects.
b Results for first-trimester exposure to non-imidazoline derivatives only (phenylephrine, levmetamfetamine, propylhexedrine) (n = 31) are not presented in this table because there were fewer than 5 exposed cases for each specific birth defect.
c Results presented in this table are only for specific birth defects with at least 5 exposed cases.
d Not exposed to other decongestants in the first trimester.
e All odds ratios and 95% confidence intervals were estimated on the basis of conditional logistic regression matched on calendar year of interview and geographic region of birth hospital. In addition, for each exposure-outcome association, we adjusted for factors associated with the specific decongestant and the specific birth defect in our population that also materially changed the odds ratio estimate, as specified in the following footnotes. Odds ratios were not estimated for fewer than 5 exposed cases for exploratory analyses.
f Adjusted for maternal race/ethnicity and educational level.
g Adjusted for maternal race/ethnicity.
h Restricted to non-neural tube defect cases, so as to exclude cases of clubfoot, which can be secondary to a neural tube defect.
i Adjusted for maternal age; educational level; first-trimester episodes of respiratory infection; and first-trimester use of acetaminophen, antihistamines, cough medications, and antibiotics.
j Adjusted for maternal educational level and first-trimester use of cough medications.
k Restricted to males only. Nonmalformed controls (restricted to males only): n = 3,768.
Pyloric stenosis was associated with intranasal decongestants overall (OR = 1.9, 95% CI: 1.0, 3.5), imidazoline derivatives (oxymetazoline or xylometazoline) (OR = 2.2, 95% CI: 1.1, 4.5), and specifically xylometazoline (OR = 2.9, 95% CI: 1.0, 7.9) (Table 4). None of these associations was observed for second- or third-trimester exposure (Web Table 3).
The risk of tracheo-esophageal fistula was associated with first-trimester exposure to imidazoline derivatives (OR = 3.3, 95% CI: 1.2, 8.9) (Table 4). This association was not observed for second- or third-trimester exposure (Web Table 3).
We also found associations between renal collecting system anomalies and second-trimester exposure to oxymetazoline (OR = 3.1, 95% CI: 1.3, 6.9) (Web Table 3).
DISCUSSION
Findings from this case-control study support previously hypothesized associations between first-trimester exposure to phenylephrine and endocardial cushion defect (10), between first-trimester exposure to phenylpropanolamine and ear defects (9, 11), and between first-trimester exposure to phenylpropanolamine and pyloric stenosis (11). In exploratory analyses designed to generate hypotheses, multiple comparisons identified associations between first-trimester use of pseudoephedrine and limb reduction defects, between first-trimester use of intranasal decongestants and pyloric stenosis, and between first-trimester use of imidazoline derivatives and tracheo-esophageal fistula, as well as between second-trimester use of oxymetazoline and renal collecting system anomalies.
For phenylephrine, the increased risk of endocardial cushion defect identified in a previous case-control study (OR = 6.7, 90% CI: 1.3, 26; 2 exposed cases) (10) was replicated in our study (OR = 8.0, 95% CI: 2.5, 25.3; 4 exposed cases). Of note, phenylephrine belongs to the same class of substituted phenethylamines as fenfluramine and dexfenfluramine, which were associated with cardiac valvulopathy in adult users (21). All 3 compounds share the 2-phenylethylamine chemical structure.
For phenylpropanolamine, the US Collaborative Perinatal Project found an elevated risk of eye and ear defects (as a single group) (relative risk = 4.0, 95% CI: 1.6, 8.3) (9), and a more recent Swedish Register study identified a modest association for ear defects (relative risk = 1.7, 95% CI: 0.8, 3.2) but not for eye defects (relative risk = 0.8, 95% CI: 0.1, 2.7; 2 exposed cases) after exposure to “oral decongestants” (99% exposed to phenylpropanolamine and 1% pseudoephedrine) (11). We also observed an increased risk of ear defects (OR = 7.8, 95% CI: 2.2, 27.2; 4 exposed cases) but not eye defects (no exposed cases) after first-trimester phenylpropanolamine exposure. We also replicated the association between pyloric stenosis and phenylpropanolamine (OR = 3.2, 95% CI: 1.1, 8.8; 6 exposed cases) that was reported in the same Swedish study (relative risk = 2.4, 95% CI: 0.9, 6.4; 4 exposed cases) (11).
Although the elevated risks that we observed on the basis of relatively small numbers for endocardial cushion defect and first-trimester exposure to phenylephrine, as well as ear defects and pyloric stenosis in relation to phenylpropanolamine, could be due to chance, these findings were observed in the context of testing specific hypotheses raised by others (9–11). Furthermore, our assessment of confounding included illnesses that might represent confounding by indication. Controlling for factors other than those that we included in our multivariate models did not substantially influence our results. As with all observational studies, retrospective reporting of maternal conditions in the first trimester (e.g., respiratory infection, fever) might be inaccurate, and residual confounding by uncontrolled factors cannot be ruled out.
Previous investigations of gastroschisis risks by our group, including 1 study that included a subset of the cases reported here (7), identified 1.8-fold (7) and 3.2-fold (5) odds ratios associated with any pseudoephedrine use in early gestation. Our present analysis of women who were most likely to be exposed in early gestation provided weak support for such an association, with a 1.5-fold odds ratio and a lower 95% confidence interval of 0.8. However, when we separated pseudoephedrine exposures into single-component pseudoephedrine versus pseudoephedrine/acetaminophen, the pattern of odds ratios was opposite that in the earlier study (7): We found an elevated risk for gastroschisis with the use of single-component pseudoephedrine (OR = 3.0, 95% CI: 1.4, 6.2), but only 1 gastroschisis case was exposed to the combination product. Our findings for first-trimester pseudoephedrine and small-intestinal atresia/stenosis were compatible with an earlier report by our group (7); in the present analyses, we similarly observed a moderately increased risk for overall use (OR = 1.6, 95% CI: 0.9, 2.8) that was higher for pseudoephedrine/acetaminophen (OR = 2.3, 95% CI: 0.9, 6.1) than for single-component pseudoephedrine (OR = 1.4, 95% CI: 0.6, 3.1).
Our findings did not confirm other reported associations. Although a previous study found a 2.4-fold risk of ventricular septal defect with first-trimester exposure to oral decongestants (95% CI: 0.9, 6.3) (3), we found an odds ratio of only 1.3 (with a narrow 95% CI: 1.1, 1.6) for exposure to oral decongestants overall (mainly pseudoephedrine). Of note, we found the risk to be elevated only with pseudoephedrine/acetaminophen (OR = 1.8, 95% CI: 1.2, 2.7) and not with single-component pseudoephedrine (OR = 1.1, 95% CI: 0.8, 1.5). This finding is remarkably consistent with differences previously observed by our group for both gastroschisis and small-intestinal atresia in relation to use of pseudoephedrine/acetaminophen compared with either pseudoephedrine or acetaminophen alone (7). In other comparisons, we found little support for associations between first-trimester exposure to sympathomimetics (predominantly pseudoephedrine) and coarctation of the aorta (4), first-trimester exposure to phenylpropanolamine and gastroschisis (6), and first-trimester exposure to phenylephrine and eye and ear defects (9) and clubfoot (9).
Results from our exploratory analyses corroborated previous reports describing no associations between first-trimester exposure to oral decongestants and neural tube defects (12), hypospadias (9), or inguinal hernia (9). We also did not find any appreciable association between first-trimester oral decongestant exposure and other specific birth defects, with the exception of a 2.4-fold increased risk of limb reduction defects with pseudoephedrine.
Understanding the risks associated with phenylephrine is important given the recent trends in use (Figure 1). Although pseudoephedrine is still the most commonly used decongestant, the relative use of phenylephrine increased after federal restrictions of access to pseudoephedrine (22), which also led to the replacement of pseudoephedrine with phenylephrine in some cough/cold preparations. Phenylpropanolamine use has declined as a result of its removal from the market in 2000–2001, which was due to its reported risk of hemorrhagic stroke in adults (23).
Open in a separate window
Figure 1.
Secular trends of first-trimester use of any, as well as specific, decongestants among mothers of nonmalformed infants from Boston, Massachusetts, and Philadelphia, Pennsylvania, which throughout study period contributed data to the Slone Epidemiology Center Birth Defects Study, 1993–2010. The period 2009–2010 is essentially only 1 year because this study included data up to January 2010.
We believe ours is the first study to explore the risks and relative safety of specific intranasal decongestants used during pregnancy with regard to a wide range of specific birth defects. Inasmuch as the analyses on intranasal decongestants were exploratory in nature, the associations we found warrant cautious interpretation because they were identified in the context of no preexisting hypotheses and multiple comparisons; the latter might result in confidence intervals that could be wider than we estimated. However, we believe that the observed elevated risks of 2 malformations warrant further exploration: pyloric stenosis in relation to first-trimester exposure to intranasal decongestants, and tracheo-esophageal fistula in relation to first-trimester exposure to imidazoline derivatives. In addition, renal collecting system anomalies potentially could be associated with second-trimester exposure to oxymetazoline.
It is not known whether any of the oral decongestants (pseudoephedrine, phenylephrine, and phenylpropanolamine) cross the placenta (24) to exert any direct effect on the fetus. Nonetheless, their vasoconstrictive properties as a result of α-adrenergic stimulation could cause constriction of uterine vessels that appear to have only α-adrenergic receptors (25), potentially contributing to the risk of teratogenicity via vascular disruption or hypoxia-induced oxidative stress (26). Although administered intranasally for an intended localized effect and generally thought to carry less risk than systemically administered agents, intranasal decongestants, mainly oxymetazoline and xylometazoline, are absorbed well from the nasal mucosa to produce systemic effects (24). Though it is not known whether these agents cross the placenta (24), a report of fetal heart rate changes after maternal use of oxymetazoline nasal spray (27) suggests that oxymetazoline and possibly other intranasal decongestants could affect the fetus. In addition to their vasoconstrictive α-adrenoceptor agonistic activity (28), oxymetazoline and xylometazoline are also agonists of some 5-hydroxytryptamine (serotonin) receptor subtypes (29, 30). Increased stimulation of 5-hydroxytryptamine receptors could potentially affect fetal development and cause birth defects via 5-hydroxytryptamine receptor–mediated teratogenesis (26).
In its collection of medication histories, the BDS specifically asks about the use of OTC products, providing the opportunity to study this type of drug exposure. Such analyses are not typically possible with other types of data sources, such as claims or pharmacy dispensing databases, which do not routinely capture OTC exposures. Moreover, by relying on maternal report, the BDS focuses on actual use of medications (rather than prescription or dispensing). However, given that information about medication use is obtained retrospectively, the potential for imperfect recall is a concern, particularly the recall of OTC medications (such as decongestants) that tend to be used infrequently and for short durations. In the BDS, the use of a multilevel, highly structured questionnaire (19) and a Medication Identification Booklet with OTC medication photographs likely enhanced recall of medication exposures. In addition, the detailed medication exposure information collected by the BDS allowed us to distinguish subjects who were considered likely exposed (i.e., with high certainty of exposure) (20) from those who were possibly exposed. By creating separate categories instead of combining them into a single group of any potentially exposed subjects, we sought to minimize exposure misclassification. We further focused on subjects who likely had been exposed to only a specific decongestant, without other decongestants in the same trimester; this allowed us to clearly distinguish the specific decongestants that might be associated with specific birth defects. Although this approach could have less power than others that aggregate all potentially exposed subjects, we believe it offers greater validity by reducing exposure misclassification.
In our retrospective interview-based study, recall bias might be a concern if mothers of malformed infants recalled or reported prenatal medication use differently from mothers of nonmalformed controls. That we identified different effect estimates for different decongestants suggests that recall bias was of little relevance, because there is no reason to postulate that women would be biased in their recall of different decongestants in relation to different defects.
Our findings should be kept in perspective. As one example, the baseline prevalence of endocardial cushion defect is about 0.34 per 1,000 live births (31); thus, even if phenylephrine exposure increased the risk 8-fold, the absolute risk of having an affected child still would be small (about 2.7 per 1,000 live births; i.e., 0.27%).
In summary, our data support some previously reported associations and raise a small number of new hypotheses. Given the widespread use of decongestants by pregnant women, there is continuing need to obtain further data on the risks and relative safety of specific decongestants in relation to the wide range of specific birth defects.
Supplementary Material
Web Material:
Click here to view.
ACKNOWLEDGMENTS
Author affiliations: Department of Epidemiology, Harvard School of Public Health, Harvard University, Boston, Massachusetts (Wai-Ping Yau, Kueiyu Joshua Lin, Sonia Hernández-Díaz); Department of Pharmacy, Faculty of Science, National University of Singapore, Singapore (Wai-Ping Yau); and Slone Epidemiology Center at Boston University, Boston, Massachusetts (Allen A. Mitchell, Martha M. Werler).
This work was supported by grant R01 HD046595 from the Eunice Kennedy Shriver National Institute of Child Health and Human Development (NICHD). The Pharmacoepidemiology Program at the Harvard School of Public Health and the Slone Epidemiology Center receive support from various pharmaceutical companies, some of which could manufacture products included in these analyses. These analyses were not supported by any pharmaceutical manufacturer. Wai-Ping Yau was supported by the National University of Singapore-Overseas Postdoctoral Fellowship.
We thank Dawn Jacobs, Fiona Rice, Rita Krolak, Kathleen Sheehan, Moira Quinn, Clare Coughlin, Nancy Rodriquez-Sheridan, Carolina Meyers, Joan Shander, and Paula Wilder for their assistance in data collection; Nastia Dynkin for computer programming; the staff of the Massachusetts Department of Public Health Center for Birth Defects Research and Prevention, Dr. Charlotte Druschel and the New York State Health Department, and Drs. Christina Chambers and Kenneth Jones of the University of California, San Diego, for assistance with case ascertainment. We also thank the medical and nursing staff at all participating hospitals: Baystate Medical Center, Springfield, Massachusetts; Beth Israel Deaconess Medical Center, Boston, Massachusetts; Boston Medical Center, Boston, Massachusetts; Brigham and Women's Hospital, Boston, Massachusetts; Brockton Hospital, Brockton, Massachusetts; Cambridge Hospital, Cambridge, Massachusetts; Caritas Good Samaritan Medical Center, Brockton, Massachusetts; Charlton Memorial Hospital, Children's Hospital, Fall River, Massachusetts; Emerson Hospital, Concord, Massachusetts; Falmouth Hospital, Falmouth, Massachusetts; Haverhill-Hale Hospital, Haverhill, Massachusetts; Jordan Hospital, Plymouth, Massachusetts; Kent Hospital, Warwick, Rhode Island; Lawrence General Hospital, Lawrence, Massachusetts; Lowell General Hospital, Lowell, Massachusetts; Melrose-Wakefield Hospital, Melrose, Massachusetts; Metro West Medical Center-Framingham, Framingham, Massachusetts; Mt. Auburn Hospital, Cambridge, Massachusetts; New England Medical Center, Boston, Massachusetts; Newton-Wellesley Hospital, Newton, Massachusetts; North Shore Medical Center, Salem, Massachusetts; Rhode Island Hospital, Providence, Rhode Island; Saints Memorial Medical Center, Lowell, Massachusetts; South Shore Hospital, Weymouth, Massachusetts; Southern New Hampshire Medical Center, Nashua, New Hampshire; St. Elizabeth's Medical Center, Brighton, Massachusetts; St. Luke's Hospital, New Bedford, Massachusetts; St. Vincent Hospital, Worcester, Massachusetts; UMASS Memorial Health Care, Worcester, Massachusetts; Women and Infants' Hospital, Providence, Rhode Island; Abington Memorial Hospital, Abington, Pennsylvania; Albert Einstein Medical Center, Philadelphia, Pennsylvania; Alfred I. duPont Hospital for Children, Wilmington, Delaware; Bryn Mawr Hospital, Bryn Mawr, Pennsylvania; Chester County Hospital, West Chester, Pennsylvania; Children's Hospital of Philadelphia and their Clinical and Translational Research Center, Philadelphia, Pennsylvania; Christiana Care Health Services, Wilmington, Delaware; Community Hospital, Chester, Pennsylvania; Crozer-Chester Medical Center, Upland, Pennsylvania; Doylestown Hospital, Doylestown, Pennsylvania; Frankford Hospital, Philadelphia, Pennsylvania; Hahnemann University Hospital, Philadelphia, Pennsylvania; the Hospital of the University of Pennsylvania, Lankenau Hospital, Philadelphia, Pennsylvania; Lancaster General Hospital, Lancaster, Pennsylvania; Lehigh Valley Hospital, Allentown, Pennsylvania; Nanticoke Memorial Hospital, Seaford, Delaware; Pennsylvania Hospital, Philadelphia, Pennsylvania; Sacred Heart Hospital, Allentown, Pennsylvania; St. Christopher's Hospital for Children, Philadelphia, Pennsylvania; St. Mary Medical Center, Langhorne, Pennsylvania; Temple University Health Sciences Center, Philadelphia, Pennsylvania; Reading Hospital and Medical Center, West Reading, Pennsylvania; Thomas Jefferson University Hospital, Philadelphia, Pennsylvania; Grand River Hospital, Kitchener, Ontario; Guelph General Hospital, Guelph, Ontario; Hamilton Health Sciences Corp, Hamilton, Ontario; the Hospital for Sick Children, Toronto, Ontario; Humber River Regional Hospital-Church Site, Toronto, Ontario; Humber River Regional Hospital-Finch Site, Toronto, Ontario; Joseph Brant Memorial Hospital, Burlington, Ontario; Lakeridge Health Corp, Durham Region, Ontario; London Health Sciences Center, London, Ontario; Mt. Sinai Hospital, Toronto, Ontario; North York General Hospital, Toronto, Ontario; Oakville Trafalgar Memorial Hospital, Oakville, Ontario; Scarborough Hospital–General Division, Scarborough, Ontario; Scarborough Hospital–Grace Division, Scarborough, Ontario; St. Joseph's Health Center-London, London, Ontario; St. Joseph's Health Center-Toronto, Toronto, Ontario; St. Joseph's Healthcare-Hamilton, Hamilton, Ontario; St. Michael's Hospital, Toronto, Ontario; Sunnybrook and Women's College Health Sciences Center, Toronto, Ontario; Toronto East General Hospital, East York, Ontario; Toronto General Hospital, Toronto, Ontario; Trillium Health Center, Mississauga, Ontario; William Osler Health Center, Etobicoke, Ontario; York Central Hospital, Richmond Hill, Ontario; York County Hospital, Newmarket, Ontario; Alvarado Hospital, San Diego, California; Balboa Naval Medical Center, San Diego, California; Camp Pendleton Naval Hospital, Camp Pendleton North, California; Children's Hospital and Health Center, San Diego, California; Kaiser Zion Medical Center, San Diego, California; Palomar Medical Center, Escondido, California; Pomerado Hospital, Poway, California; Scripps Mercy Hospital, San Diego, California; Scripps Memorial Hospital-Chula Vista, Chula Vista, California; Scripps Memorial Hospital-Encinitas, Encinitas, California; Scripps Memorial Hospital-La Jolla, La Jolla, California; Sharp Chula Vista Hospital, Chula Vista, California; Sharp Coronado Hospital, Coronado, California; Sharp Grossmont Hospital, La Mesa, California; Sharp Mary Birch Hospital, San Diego, California; Tri-City Medical Center, Oceanside, California; and University of California, San Diego Medical Center, San Diego, California.
Conflict of interest: Until August of 2012, Dr. Mitchell owned stock in Johnson and Johnson (<$20,000 value), which manufactures decongestant and acetaminophen products.
REFERENCES
1. Werler MM, Mitchell AA, Hernandez-Diaz S, et al. Use of over-the-counter medications during pregnancy. Am J Obstet Gynecol. 2005;193(3 Pt 1):771–777. [PubMed] [Google Scholar]
2. Mitchell AA. Studies of drug-induced birth defects. In: Strom BL, Kimmel SE, Hennessy S, editors. Pharmacoepidemiology. Chichester, UK: Wiley-Blackwell; 2012. pp. 487–504. [Google Scholar]
3. Bateman DN, McElhatton PR, Dickinson D, et al. A case control study to examine the pharmacological factors underlying ventricular septal defects in the North of England. Eur J Clin Pharmacol. 2004;60(9):635–641. [PubMed] [Google Scholar]
4. Ferencz C, Loffredo CA, Correa-Villasenor A, et al. Perspectives in Pediatric Cardiology, Volume 5. Genetic and Environmental Risk Factors of Major Cardiovascular Malformations. The Baltimore-Washington Infant Study 1981–1989. Armonk, NY: Futura Publishing Company, Inc.; 1997. Coarctation of aorta; pp. 190–201. [Google Scholar]
5. Werler MM, Mitchell AA, Shapiro S. First trimester maternal medication use in relation to gastroschisis. Teratology. 1992;45(4):361–367. [PubMed] [Google Scholar]
6. Torfs CP, Katz EA, Bateson TF, et al. Maternal medications and environmental exposures as risk factors for gastroschisis. Teratology. 1996;54(2):84–92. [PubMed] [Google Scholar]
7. Werler MM, Sheehan JE, Mitchell AA. Maternal medication use and risks of gastroschisis and small intestinal atresia. Am J Epidemiol. 2002;155(1):26–31. [PubMed] [Google Scholar]
8. Werler MM, Sheehan JE, Hayes C, et al. Vasoactive exposures, vascular events, and hemifacial microsomia. Birth Defects Res A Clin Mol Teratol. 2004;70(6):389–395. [PubMed] [Google Scholar]
9. Heinonen OP, Slone D, Shapiro S. Birth Defects and Drugs in Pregnancy. Littleton, MA: Publishing Sciences Group, Inc.; 1977. Drugs affecting the autonomic nervous system; pp. 345–356. [Google Scholar]
10. Rothman KJ, Fyler DC, Goldblatt A, et al. Exogenous hormones and other drug exposures of children with congenital heart disease. Am J Epidemiol. 1979;109(4):433–439. [PubMed] [Google Scholar]
11. Kallen BA, Olausson PO. Use of oral decongestants during pregnancy and delivery outcome. Am J Obstet Gynecol. 2006;194(2):480–485. [PubMed] [Google Scholar]
12. Shaw GM, Todoroff K, Velie EM, et al. Maternal illness, including fever and medication use as risk factors for neural tube defects. Teratology. 1998;57(1):1–7. [PubMed] [Google Scholar]
13. Gilbert-Barness E, Drut RM. Association of sympathomimetic drugs with malformations. Vet Hum Toxicol. 2000;42(3):168–171. [PubMed] [Google Scholar]
14. Schatz M, Zeiger RS, Harden K, et al. The safety of asthma and allergy medications during pregnancy. J Allergy Clin Immunol. 1997;100(3):301–306. [PubMed] [Google Scholar]
15. Jick H, Holmes LB, Hunter JR, et al. First-trimester drug use and congenital disorders. JAMA. 1981;246(4):343–346. [PubMed] [Google Scholar]
16. Aselton P, Jick H, Milunsky A, et al. First-trimester drug use and congenital disorders. Obstet Gynecol. 1985;65(4):451–455. [PubMed] [Google Scholar]
17. Werler MM, Shapiro S, Mitchell AA. Periconceptional folic acid exposure and risk of occurrent neural tube defects. JAMA. 1993;269(10):1257–1261. [PubMed] [Google Scholar]
18. Louik C, Lin AE, Werler MM, et al. First-trimester use of selective serotonin-reuptake inhibitors and the risk of birth defects. N Engl J Med. 2007;356(26):2675–2683. [PubMed] [Google Scholar]
19. Mitchell AA, Cottler LB, Shapiro S. Effect of questionnaire design on recall of drug exposure in pregnancy. Am J Epidemiol. 1986;123(4):670–676. [PubMed] [Google Scholar]
20. Yau WP, Lin KJ, Werler MM, et al. Drug certainty-response in interview-based studies. Pharmacoepidemiol Drug Saf. 2011;20(11):1210–1216. [PMC free article] [PubMed] [Google Scholar]
21. Centers for Disease Control and Prevention (CDC) Cardiac valvulopathy associated with exposure to fenfluramine or dexfenfluramine: U.S. Department of Health and Human Services interim public health recommendations, November 1997. MMWR Morb Mortal Wkly Rep. 1997;46(45):1061–1066. [PubMed] [Google Scholar]
22. Drug Enforcement Administration (DEA), Department of Justice. Retail sales of scheduled listed chemical products; self-certification of regulated sellers of scheduled listed chemical products. Interim final rule with request for comment. Fed Regist. 2006;71(186):56008–56027. [PubMed] [Google Scholar]
23. Meadows M. FDA issues public health advisory on phenylpropanolamine in drug products. FDA Consum. 2001;35(1):9. [PubMed] [Google Scholar]
24. Micromedex® Healthcare Series. Thomson Reuters (Healthcare) Inc. //www.thomsonhc.com. (Accessed February 27, 2010)