Materials and Methods
The BWIS population consisted of infants born to residents of Maryland, the District of Columbia, and 6 adjacent counties of northern Virginia from April 1981 through December 1989. The methods of this study have been previously described in detail. All data used in our study were deidentified and analyses were performed with an exemption from the Institutional Review Board of the Centers for Disease Control and Prevention.
Cases
Cases were infants with any type of CVM ascertained from searches of community hospitals, 6 pediatric cardiology centers serving the study region, and the medical examiner’s logbooks from Maryland. CVM noted at registration were confirmed by echocardiography, cardiac catheterization, surgery, or autopsy. CVM were coded by pediatric cardiologists. Updated information about CVM diagnoses at 1 year of age obtained for all registered cases resulted in a change in only 7.8% of the initial diagnoses. Infants of gestational age <38 weeks with patent ductus arteriosus as the only CVM were not included. Also, because of improvements in diagnostic capability over the study period and the resultant rapid rise in the population prevalence among young infants, only a random sample of the infants with small ventricular septal defects (VSD) were included in BWIS. Infants with >1 cardiac defect were assigned 1 anatomic diagnosis using a hierarchical classification approach developed for BWIS based on the presumed embryonic timing of the defects. These diagnoses were then placed into categories based on their developmental mechanism. Cases were further classified based on the presence of other anomalies as isolated (ie, no noncardiac defects); chromosomal disorders (eg, Down syndrome, other trisomies); recognizable syndromes (eg, Ivemark, DiGeorge, Noonan, Williams, fetal alcohol, congenital rubella); or multiple defects (ie, with noncardiac anomalies of unknown cause).
From all identified CVM cases (n = 4390), we excluded all cases with ≥1 of the following factors: maternal reports of pregestational diabetes since this condition is a known risk factor for CVM (n = 87); recognized syndromes or chromosomal abnormalities with the exception that we included infants with Down syndrome who had atrioventricular septal defect (AVSD) (n = 947 excluded); infants who were 1 of a set of twins, triplets, or other multiple births (n = 156); and those for whom no maternal interview was obtained (n = 1013). We then evaluated singleton infants with isolated CVM whose mothers did not have pregestational diabetes and did complete interviews (final n = 2525).
Controls
Controls (n = 3572) were a random sample of all liveborn infants without CVM from the same birth cohort who were delivered in participating hospitals, stratified by month, year, and hospital of birth. Controls were similar to all area births during the study period by infant sex, race, birth weight, plurality, season of birth, and maternal age. For this analysis, we included interviewed, singleton controls with no CVM, chromosomal anomalies, syndromes, or maternal reports of pregestational diabetes (final n = 3435).
Data collection
Home interviews with the parents of case and control infants were conducted within 18 months of birth of the study subjects. A structured, standardized questionnaire was administered by trained interviewers to obtain information on sociodemographic factors, family history, maternal medical conditions, and environmental factors. The latter included reports on medication use during the periconceptional period (3 months before the last menstrual period through the first trimester of pregnancy).
Analgesic use
For this analysis, we defined exposure as maternal use of an analgesic-containing medication at any time during the periconceptional period to ensure that all relevant exposures were included regardless of errors in recall of the last menstrual period or in recall of the exact timing of medication use. Maternal reports of use of prescription and nonprescription analgesics during the periconceptional period were grouped into pharmacologic classes: salicylates, acetaminophen, other NSAIDs, and opioids.
Statistical analysis
First, we compared the frequency of selected maternal and infant demographic and clinical characteristics among cases and controls using the χ 2 statistic. A χ 2 statistic was not calculated if the proportion of subjects with missing values was >5% of the total. Then, because the presence of maternal fever or flu symptoms has been associated with an increased risk of CVM in the infant in previous analyses of BWIS data, we examined the frequency of maternal analgesic use among case and control infants by pharmacologic class stratified by the presence of fever or maternal flu symptoms during the periconceptional period. Finally, we used multiple logistic regression models to evaluate possible associations of selected specific CVM diagnostic groups with maternal periconceptional use of analgesics by pharmacologic class using adjusted odds ratios (adjORs) and 95% confidence intervals (CIs). For this part of the analysis, we excluded infants of mothers who reported use of >1 class of analgesic drug, including use of single preparations that contained >1 class of analgesic, during the periconceptional period (28% of cases and 27% of controls). However, infants of mothers who reported use of nonanalgesic drugs only, or use of single preparations that contained both an analgesic and a nonanalgesic drug, during the periconceptional period were included. All models were adjusted for the covariates of infant sex, infant race, maternal age, family history of CVM, family history of other birth defects, maternal fever or flu symptoms during the periconceptional period, maternal prepregnancy body mass index (weight in kilograms/height in m ), and maternal smoking during the periconceptional period, with inclusion of quadratic terms for maternal age and prepregnancy body mass index because of their potential nonlinear relationship with the risk for birth defects. Only subjects with nonmissing values for all covariates were included in the models. We considered only associations based on at least 3 exposed cases to be stable.
Results
Characteristics of case and control infants
Compared with control infants, case infants were significantly more likely to have a family history of CVM ( P < .001). Otherwise, case and control infants were similar with respect to maternal and infant demographic and clinical characteristics ( Table 1 ).
| Characteristic | Cases (n = 2525) | % b | Controls (n = 3435) | % b | χ 2 P value c |
|---|---|---|---|---|---|
| Family history of | |||||
| Cardiovascular malformation | |||||
| Yes | 98 | 3.9 | 40 | 1.2 | < .001 |
| No | 2427 | 96.1 | 3395 | 98.8 | |
| Noncardiac malformation | |||||
| Yes | 120 | 4.8 | 155 | 4.5 | .66 |
| No | 2405 | 95.2 | 3280 | 95.5 | |
| Maternal characteristics during periconceptional period | |||||
| Treated hypertension | |||||
| Yes | 16 | 0.6 | 23 | 0.7 | – |
| No | 1743 | 69.0 | 2672 | 77.8 | |
| Prepregnancy BMI d | |||||
| <30 | 2346 | 92.9 | 3180 | 92.6 | |
| ≥30 | 174 | 6.9 | 248 | 7.2 | .62 |
| <35 | 2444 | 96.8 | 3351 | 97.6 | |
| ≥35 | 76 | 3.0 | 77 | 2.2 | .06 |
| Smoking e | |||||
| Yes | 921 | 36.5 | 1222 | 35.6 | .47 |
| No | 1604 | 63.5 | 2213 | 64.4 | |
| Alcohol use e | |||||
| Yes | 1508 | 59.7 | 2011 | 58.5 | .34 |
| No | 1015 | 40.2 | 1424 | 41.5 | |
| Education, y | |||||
| <12 | 461 | 18.3 | 637 | 18.5 | .25 |
| 12 | 949 | 37.6 | 1221 | 35.5 | |
| >12 | 1112 | 44.0 | 1575 | 45.9 | |
| Fever e | |||||
| Yes | 132 | 5.2 | 155 | 4.5 | .20 |
| No | 2393 | 94.8 | 3280 | 95.5 | |
| Flu symptoms e | |||||
| Yes | 208 | 8.2 | 261 | 7.6 | .37 |
| No | 2317 | 91.8 | 3174 | 92.4 | |
| Age, y | |||||
| <20 | 339 | 13.4 | 485 | 14.1 | .29 |
| 20-24 | 629 | 24.9 | 852 | 24.8 | |
| 25-29 | 748 | 29.6 | 1083 | 31.5 | |
| 30-34 | 576 | 22.8 | 730 | 21.3 | |
| ≥35 | 224 | 8.9 | 277 | 8.1 | |
| Gravidity | |||||
| Primiparous | 767 | 30.4 | 1119 | 32.6 | .07 |
| Multiparous | 1758 | 69.6 | 2316 | 67.4 | |
| Infant characteristics: | |||||
| Race | |||||
| White | 1624 | 64.3 | 2279 | 66.3 | .10 |
| Other | 901 | 35.7 | 1156 | 33.7 | |
| Sex | |||||
| Male | 1265 | 50.1 | 1741 | 50.7 | .66 |
| Female | 1260 | 49.9 | 1694 | 49.3 |
a Singleton infants of mothers without pregestational diabetes who completed interviews–infants with major noncardiac organ system anomalies, recognized syndromes, or chromosomal abnormalities other than Down syndrome with atrioventricular septal defect were excluded
b Percents may not add up to 100 because of missing values
c Calculations include only subjects with nonmissing values– P values were not calculated if proportion of subjects with missing values was >5% of total
d Weight in kilograms/height in m 2
Maternal analgesic use
From April 1981 through December 1989, the BWIS enrolled and interviewed 2525 singleton infants with isolated CVM or with AVSD and Down syndrome and 3435 singleton infants with no CVM, chromosomal anomalies, or syndromes whose mothers did not have pregestational diabetes. The frequency of any analgesic use during the periconceptional period was 53% among case mothers and 52% among control mothers. The frequency of analgesic use by pharmacologic class among case and control mothers, respectively, was: any salicylate-containing medication, 13.5% and 12.1%; any acetaminophen-containing medication, 42.9% and 43.5%; any NSAID-containing medication, 8.8% and 8.6%; and any opioid-containing medication, 4.4% and 3.6%. Among mothers of case infants who reported fever or flu symptoms during the periconceptional period, 177 (67.3%) used an analgesic compared with 235 (70.1%) among mothers of control infants who reported fever or flu symptoms ( Table 2 ). Among mothers of case infants who did not report fever or flu symptoms during the periconceptional period, 1160 (51.3%) used an analgesic compared with 1560 (50.3%) among mothers of control infants who did not report fever or flu symptoms. Overall, analgesic use was similar among mothers of case and control infants for all pharmacologic categories when stratified by the presence of fever or flu symptoms.
| Analgesic class | Cases (n = 2525) | % | Controls (n = 3435) | % |
|---|---|---|---|---|
| No fever or flu symptoms | 2262 | 100 | 3100 | 100 |
| No analgesic drugs | 1102 | 48.7 | 1540 | 49.7 |
| Any analgesic drug | 1160 | 51.3 | 1560 | 50.3 |
| Any aspirin-containing drug | 295 | 13.0 | 359 | 11.6 |
| Aspirin only | 127 | 5.6 | 171 | 5.5 |
| Any acetaminophen-containing drug | 935 | 41.3 | 1296 | 41.8 |
| Acetaminophen only | 649 | 28.7 | 905 | 29.2 |
| Any NSAID | 190 | 8.4 | 257 | 8.3 |
| NSAID only | 51 | 2.3 | 60 | 1.9 |
| Any opioid-containing drug | 94 | 4.2 | 100 | 3.2 |
| Opioid drug only | 8 | 0.4 | 7 | 0.2 |
| With fever and/or flu symptoms | 263 | 100 | 335 | 100 |
| No analgesic drugs | 86 | 32.7 | 100 | 29.9 |
| Any analgesic drug | 177 | 67.3 | 235 | 70.1 |
| Any aspirin-containing drug | 46 | 17.5 | 57 | 17.0 |
| Aspirin only | 16 | 6.1 | 23 | 6.9 |
| Any acetaminophen-containing drug | 149 | 56.7 | 199 | 59.4 |
| Acetaminophen only | 100 | 38.0 | 136 | 40.6 |
| Any NSAID | 33 | 12.5 | 37 | 11.0 |
| NSAID only | 9 | 3.4 | 7 | 2.1 |
| Any opioid-containing drug | 16 | 6.1 | 23 | 6.9 |
| Opioid drug only | 1 | 0.4 | 4 | 1.2 |
a Singleton infants of mothers without pregestational diabetes who completed interviews–infants with major noncardiac organ system anomalies, recognized syndromes, or chromosomal abnormalities other than Down syndrome with atrioventricular septal defects were excluded.
CVM diagnostic groups and maternal analgesic use
When comparing use of analgesics by pharmacologic class and case or control status, multiple logistic regression analyses showed few significant associations between analgesic use and CVM ( Table 3 ). Mothers of infants with tetralogy of Fallot were significantly more likely to have used acetaminophen during the periconceptional period than were control mothers (adjOR, 1.57; 95% CI, 1.08–2.27); mothers of infants with dextrotransposition of the great arteries (dTGA) with intact ventricular septum were significantly more likely to have used NSAIDs during the periconceptional period (adjOR, 3.24; 95% CI, 1.19–8.77). Maternal use of salicylates or opioids during the periconceptional period was not associated with CVM in the offspring.
| Cardiac malformation | Total no. | Salicylates | Acetaminophen | Other NSAIDs | Opioids | ||||
|---|---|---|---|---|---|---|---|---|---|
| Exposed/nonexposed d | AdjOR (95% CI) | Exposed/nonexposed d | AdjOR (95% CI) | Exposed/nonexposed d | AdjOR (95% CI) | Exposed/nonexposed d | AdjOR (95% CI) | ||
| No malformation | 2953 | 194/1640 | 1041/1640 | 67/1640 | 11/1640 | ||||
| Any cardiac malformation | 2149 | 143/1188 | 1.02 (0.81–1.28) | 749/1188 | 0.99 (0.88–1.12) | 60/1188 | 1.23 (0.86–1.77) | 9/1188 | 1.02 (0.41–2.57) |
| Lateral/looping | 36 | 3/20 | 1.47 (0.42–5.13) | 12/20 | 1.00 (0.48–2.08) | 0/20 | N/A | 1/20 | 25.57 (2.42–270.25) e |
| Conotruncal | 321 | 18/173 | 0.85 (0.51–1.43) | 119/173 | 1.08 (0.84–1.39) | 10/173 | 1.45 (0.72–2.89) | 1/173 | 0.92 (0.11–7.31) |
| dTGA | 154 | 9/85 | 0.81 (0.39–1.66) | 53/85 | 0.89 (0.62–1.29) | 6/85 | 1.74 (0.72–4.21) | 1/85 | 1.63 (0.20–13.38) |
| dTGA with IVS | 78 | 5/40 | 0.98 (0.37–2.57) | 28/40 | 0.96 (0.58–1.59) | 5/40 | 3.24 (1.19–8.77) e | 0/40 | N/C |
| dTGA with VSD | 47 | 2/26 | 0.65 (0.15–2.83) | 18/26 | 1.00 (0.54–1.87) | 0/26 | N/A | 1/26 | 8.09 (0.90–72.42) |
| dTGA with DORV | 14 | 1/7 | 0.82 (0.10–7.12) | 5/7 | 0.91 (0.28–2.98) | 1/7 | 2.92 (0.34–25.16) | 0/7 | N/C |
| Truncus arteriosus | 14 | 1/10 | 0.78 (0.10–6.36) | 3/10 | 0.47 (0.13–1.76) | 0/10 | N/A | 0/10 | N/C |
| Tetralogy of Fallot | 135 | 8/64 | 1.10 (0.51–2.36) | 59/64 | 1.57 (1.08–2.27) e | 4/64 | 1.61 (0.56–4.61) | 0/64 | N/C |
| Any AVSD | 213 | 16/107 | 1.19 (0.68–2.09) | 81/107 | 1.14 (0.83–1.55) | 9/107 | 1.97 (0.94–4.13) | 0/107 | N/C |
| AVSD with Down syndrome | 162 | 13/81 | 1.24 (0.66–2.32) | 62/81 | 1.15 (0.81–1.63) | 6/81 | 1.68 (0.69–4.10) | 0/81 | N/C |
| AVSD without Down syndrome | 51 | 3/26 | 1.07 (0.31–3.66) | 19/26 | 1.10 (0.59–2.03) | 3/26 | 3.37 (0.95–11.93) | 0/26 | N/C |
| Membranous VSD | 410 | 27/236 | 1.01 (0.66–1.56) | 133/236 | 0.92 (0.73–1.16) | 11/236 | 1.13 (0.58–2.18) | 3/236 | 2.08 (0.56–7.67) |
| Atrial septal defect secundum | 162 | 14/89 | 1.48 (0.81–2.70) | 55/89 | 1.01 (0.70–1.45) | 3/89 | 0.95 (0.29–3.11) | 1/89 | N/C |
| Left-sided obstruction | 210 | 17/116 | 1.11 (0.64–1.92) | 70/116 | 0.87 (0.64–1.20) | 7/116 | 1.42 (0.63–3.21) | 0/116 | N/C |
| Hypoplastic left heart | 95 | 10/53 | 1.62 (0.79–3.30) | 30/53 | 0.89 (0.56–1.42) | 2/53 | 0.95 (0.22–4.03) | 0/53 | N/C |
| Coarctation of aorta | 66 | 6/37 | 1.30 (0.52–3.21) | 20/37 | 0.81 (0.46–1.42) | 3/37 | 1.99 (0.58–6.82) | 0/37 | N/C |
| Aortic valve stenosis | 49 | 1/26 | 0.24 (0.03–1.82) | 20/26 | 0.88 (0.48–1.62) | 2/26 | 1.62 (0.37–7.16) | 0/26 | N/C |
| Right-sided obstruction | 231 | 16/135 | 1.05 (0.60–1.85) | 74/135 | 0.90 (0.66–1.22) | 5/135 | 0.92 (0.36–2.36) | 1/135 | 1.11 (0.14–8.94) |
| Pulmonary valve stenosis | 175 | 13/103 | 1.11 (0.59–2.10) | 54/103 | 0.87 (0.62–1.24) | 4/103 | 0.97 (0.34–2.76) | 1/103 | 1.43 (0.17–11.80) |
| Pulmonary atresia with IVS | 35 | 2/18 | 0.94 (0.21–4.17) | 14/18 | 1.17 (0.57–2.40) | 1/18 | 1.23 (0.16–9.51) | 0/18 | N/C |
| Ebstein anomaly | 27 | 3/12 | 2.26 (0.60–8.45) | 11/12 | 1.36 (0.59–3.14) | 1/12 | 1.55 (0.18–13.59) | 0/12 | N/C |
| Patent ductus arteriosus f | 42 | 5/21 | 2.63 (0.94–7.36) | 15/21 | 1.11 (0.56–3.14) | 1/21 | 1.22 (0.15–9.71) | 0/21 | N/C |
| Total anomalous pulmonary venous return | 37 | 1/20 | 0.44 (0.06–3.33) | 14/20 | 1.07 (0.53–2.17) | 2/20 | 2.74 (0.61–12.34) | 0/20 | N/C |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
