Maternal Obesity and Infant Heart Defects
Dept. Ob/Gyne, University Hospital, SE-581 85 Linköping, Sweden. E-mail: email@example.com
Objective: This study determined whether obese women have an increased risk of cardiovascular defects in their offspring compared with average weight women.
Research Methods and Procedures: In a case-control study, prospectively collected information was obtained from Swedish medical health registers. The study included 6801 women who had infants with a cardiovascular defect and, as controls, all delivered women (N = 812, 457) during the study period (1992 to 2001). Infants with chromosomal anomalies or whose mothers had pre-existing diabetes were excluded. Obesity was defined as BMI >29 kg/m2, and morbid obesity was defined as BMI >35 kg/m2. Comparisons were made with average weight women (BMI = 19.8 to 26 kg/m2).
Results: In the group of obese mothers, there was an increased risk for cardiovascular defects compared with the average weight mothers [adjusted odds ratio (OR) = 1.18; 95% CI, 1.09 to 1.27], which was slightly more pronounced for the severe types of cardiovascular defects (adjusted OR = 1.23; 95% CI, 1.05 to 1.44). With morbid obesity, the OR for cardiovascular defects was 1.40 (95% CI, 1.22 to 1.64), and for severe cardiovascular defects, the OR was 1.69 (95% CI, 1.27 to 2.26). There was an increased risk for all specific defects studied among the obese women, but only ventricular septal defects and atrial septal defects reached statistical significance.
Discussion: In this sample, a positive association was found between maternal obesity in early pregnancy and congenital heart defects in the offspring. A suggested explanation is undetected type 2 diabetes in early pregnancy, but other explanations may exist.
Maternal obesity seems to be associated with the development of neural tube defects in the offspring (1, 2, 3, 4). Whether the same association exists between maternal obesity and infant congenital heart defects is still unclear (2, 5, 6, 7, 8, 9, 10). Even a weak association is important to detect because of the increasing population of obese women of childbearing age. In Sweden, the prevalence of overweight women during the fertile period doubled between 1980 and 1997 (11).
Congenital cardiovascular defects are among the most common major birth defects. The total prevalence at birth of cardiovascular defects is estimated at 7 to 9 per 1000 births (12, 13).
A few agents and maternal conditions are confirmed as cardiac teratogens, e.g., maternal rubella infection, maternal diabetes, alcoholism, anticonvulsants and/or maternal epileptic disease, thalidomide, and retinoids (14). Only a small proportion of congenital heart defects is caused by known teratogenic agents (2%); the remaining, except those of chromosomal origin (8%), are still unrelated to specific exposures (15).
The objective of this study was to thoroughly assess, in a large prospective data set from Swedish medical health registries, whether maternal obesity, defined as BMI >29 kg/m2, was associated with an increased risk of overall and specific infant cardiovascular defects.
Research Methods and Procedures
Study Population and Exposure Assessment
The study population consisted of 982, 867 infants born in Sweden January 1, 1992 through December 31, 2001. In 812, 457 (83%), information on maternal height and maternal weight in early pregnancy was available, and there was no known pre-existing maternal diabetes. Pre-existing maternal diabetes was identified from diagnoses given to the woman at delivery. Women with gestational diabetes in the current or previous pregnancies were not excluded.
The infants were identified using the Swedish Medical Birth Registry (16). Medical data on almost all (98% to 99%) infants born in Sweden are listed in the register, which also includes stillbirths after 28 complete weeks of gestation. The register contains a large number of items concerning pregnancy, delivery, and pediatric neonatal examination. It is based on copies of the standardized medical record forms completed at the maternity health care centers, at the delivery units, and at the pediatric examination of the newborn. The system is identical throughout the country.
Maternal weight and height were ascertained at the first visit at the maternity health care center and directly recorded by the midwife in the standardized form. Ninety percent of the women attended the antenatal clinic for the first time during the first trimester.
BMI (kg/m2) was calculated from these data. Women were grouped in five categories of BMI: underweight (<19.8 kg/m2), average weight (19.8 to 26 kg/m2), overweight (26.1 to 29 kg/m2), obese (>29 kg/m2), and morbidly obese (>35 kg/m2). The fifth category is thus part of the fourth.
Identification and Classification of Cases with Congenital Heart Defects
Infants with a congenital cardiovascular defect were identified from three sources: International Classification of Disease (ICD)1 codes recorded in the Swedish Medical Birth Registry, ICD codes in the Swedish Registry of Congenital Malformations (17), and International Society of Cardiology (ISC) codes in the Swedish Child Cardiology Register (12). Thus, all three registers gave cardiac diagnoses. The Medical Birth Register and the Register of Congenital Malformations both refer to the neonatal period, and diagnoses are often unspecific or preliminary, whereas the Cardiology Register catches cases up to the age of 1 year and contains detailed cardiac diagnoses from echocardiography, catheterization, operation, or autopsy. In most instances, the diagnoses agreed among the three registers (but cases were often not recorded in all three registers). When discrepancies existed, the Cardiology Register diagnoses were regarded as the most reliable ones.
The various sources were linked using the personal identification numbers of mothers and/or infants. These numbers are given to each person living in Sweden and are extensively used in society, including for all health care. A file was built up, cleaning the available information with a final, best possible diagnosis of cardiovascular defects for each infant, according to the register data quality mentioned above.
Primary outcome was any congenital cardiovascular defect reported to any one of the above-described registers, except patent ductus arteriosus in preterm infants (<2500 g) and single umbilical artery. All infants who also had a major noncardiovascular malformation were identified from the ICD codes in the above-described registers. Minor and variable conditions like preauricular tags, undescended testicle, unstable hip, and nevus were not included.
The congenital heart defects were classified as follows:
- 1. Isolated, if the cardiac diagnosis was the only major congenital defect present.
- 2. Nonisolated, if there were major noncardiovascular malformations present but no known chromosome anomaly.
- 3. Chromosomal, if the cardiac defect was part of a known chromosomal abnormality.
The cardiac defects were divided into groups of severe and mild forms according to a recent study (18, 19). Table 1 lists the types of cardiac defects included in the two groups. Mild cardiac defects were counted only when no severe defect was present. A third group consisted of unspecified cardiac defects (Table 1), which were identified by the ICD codes 746.9 or Q24.9 in the absence of any other cardiac defect diagnosis.
Table 1. Cardiovascular defect diagnoses (isolated or nonisolated) with number of infants with each diagnosis
|Severe defects|| |
| Hypoplastic left heart syndrome||166|
| Single ventricle||78|
| Tricuspidal atresia||36|
| Common truncus||226|
| Pulmonary atresia||26|
| Double outlet right ventricle||17|
| Endocardial cushion defect||905|
| Total abnormal pulmonary vein return||18|
| Tetralogy of Fallot||223|
| Ebstein's anomaly||26|
|Mild defects|| |
| Ventricular septal defect||2676|
| Atrial septal defect||639|
| Coarctation of aorta||117|
| Corrected transposition||3|
| Pulmonary valve stenosis||68|
| Patent ductus (term infant)||107|
| Other specified defect, mainly valve anomalies||562|
When two or more different severe cardiac defects occurred at the same time, the defect was classified as complex.
Maternal age, parity, smoking, and year of birth were thought to be potential confounding factors and were included as covariates in the adjusted analyses.
Frequency distributions were compared using χ2 statistics. Odds ratios (ORs) were determined using Mantel-Haenszel technique for adjusted rates (20). Estimates of 95% CI were made with a test-based method (21). When two stratified ORs were compared, z tests were performed based on the variances obtained from the stratified analyses.
Maternal BMI was calculated for 83% of all births registered during the study period. A total of 8947 infants with cardiovascular defects were identified during the observation period (9.1 per 1000 births). Among them, mothers of 7379 infants (82%) had information that enabled the calculation of BMI. A chromosomal abnormality (n = 491) or maternal pre-existing diabetes (mainly type 1; n = 87) was present in 578 cases (7.8%). Of 6801 remaining infants, 5714 (84%) had an isolated cardiovascular defect. Complex types occurred in 5.5% (N = 371).
The case mothers were compared with all delivered women with respect to maternal age, parity, and maternal smoking in early pregnancy (Table 2). The distribution of maternal age, parity, and smoking in early pregnancy was similar for both groups. Adjustment for these factors and the year of birth did not change the ORs noticeably; only adjusted ORs will be presented.
Table 2. Characteristics of mothers of infants with cardiovascular defects (isolated or nonisolated) and of all delivered women
|Maternal age|| || || || |
| <20||152||2.2||16, 509||2.0|
| 20 to 24||1, 194||17.6||139, 356||17.2|
| 25 to 29||2, 492||36.6||299, 711||36.9|
| 30 to 34||1, 930||28.4||243, 484||30.0|
| 35 to 39||871||12.8||96, 132||11.8|
| 40+||162||2.4||17, 265||2.1|
|Parity|| || || || |
| 1||2, 849||41.9||333, 678||41.1|
| 2||2, 362||34.7||295, 217||36.3|
| 3||1, 089||16.0||124, 931||15.4|
| 4||501||7.4||58, 631||7.2|
|Maternal smoking|| || || || |
| Unknown||174||2.6||19, 832||2.2|
| No smoking||5, 585||82.1||664, 066||81.7|
| Smoking||1, 042||15.3||128, 559||15.8|
The observed prevalence of obesity (BMI > 29 kg/m2) was 10.6% in the control group and 12.1% in the case group. The corresponding prevalence of overweight, if the cut-off level is BMI >26 kg/m2, was 23.3% among controls and 25.0% among cases. Morbid obesity was slightly more common in the case group (2.7%) compared with the control group (2.0%).
Table 3 shows the distribution of some groups of cardiac defects according to maternal BMI. There is no significant difference in the BMI distribution among mothers of infants with specified and with unspecified cardiovascular defects (χ2 = 2.93 at 4 df, p = 0.57).
Table 3. Distribution of BMI among all delivered women and women with infants with cardiovascular defects, according to cardiovascular defect type
|All delivered women||82, 193||540, 853||103, 378||69, 728||16, 305||812, 457|
|All cardiovascular defects||627||4474||874||640||186||6801|
|Isolated cardiovascular defects||524||3778||728||537||147||5714|
|Nonisolated cardiovascular defects||103||696||146||103||39||1087|
|Severe cardiovascular defects||132||983||179||140||48||1482|
| Among them, complex||36||251||39||33||12||371|
|Mild cardiovascular defects||362||2533||504||370||107||3876|
|All specified cardiovascular defects||494||3516||683||510||155||5358|
|Unspecified cardiovascular defects||133||958||191||130||31||1443|
|Hypoplastic left heart syndrome||13||103||27||17||6||166|
|Endocardial cushion defect||78||608||108||86||25||905|
|Tetralogy of Fallot||28||148||21||16||10||223|
|Ventricular septal defect||248||1782||326||256||64||2676|
|Atrial septal defect||54||403||89||66||27||639|
|Coarctation of aorta||10||71||19||13||4||117|
Table 4 shows the risk of cardiovascular defects by class of BMI. Average weight (BMI = 19.8 to 26 kg/m2) was used as a reference. In the overweight group (BMI = 26.1 to 29 kg/m2), the risk is slightly above one for all defects as well as for isolated and nonisolated defects separately, although statistical significance was not reached.
Table 4. Adjusted OR for each BMI class to have an infant with a cardiovascular defect using BMI class 19.8 to 26 kg/m2 as a reference (1.00)
|All defects||0.92||0.84 to 1.00||1.03||0.96 to 1.11||1.18||1.09 to 1.27||1.41||1.22 to 1.64|
|Isolated defects||0.90||0.82 to 0.99||1.02||0.94 to 1.10||1.17||1.07 to 1.27||1.34||1.14 to 1.58|
|Nonisolated defects||1.00||0.76 to 1.31||1.09||0.91 to 1.30||1.25||1.04 to 1.50||1.80||1.30 to 2.49|
|Severe defects||0.89||0.70 to 1.12||0.95||0.81 to 1.12||1.23||1.05 to 1.44||1.69||1.27 to 2.26|
|Complex severe defects||0.97||0.68 to 1.38||0.84||0.60 to 1.18||1.21||0.88 to 1.67||1.78||1.00 to 3.16|
|Mild defects||0.89||0.79 to 0.99||1.06||0.97 to 1.15||1.15||1.04 to 1.27||1.40||1.18 to 1.68|
|Unspecified defects||0.86||0.69 to 1.07||1.08||0.93 to 1.27||1.15||0.98 to 1.37||1.22||0.85 to 1.75|
|Hypoplastic left heart syndrome||0.83||0.23 to 2.95||1.34||0.87 to 2.06||1.41||0.90 to 2.21|| || |
|D-transposition||0.63||0.15 to 2.61||0.77||0.45 to 1.30||1.48||0.97 to 2.27|| || |
|Endocardial cushion defect||0.85||0.61 to 1.18||0.95||0.77 to 1.17||1.21||0.99 to 1.49|| || |
|Tetralogy of Fallot||1.26||0.42 to 3.76||0.73||0.46 to 1.15||1.07||0.70 to 1.64|| || |
|Ventricular septal defect||0.92||0.79 to 1.06||0.96||0.85 to 1.08||1.14||1.01 to 1.28|| || |
|Atrial septal defect||0.91||0.59 to 1.40||1.12||0.89 to 1.40||1.37||1.09 to 1.72|| || |
|Coarctation of aorta||0.95||0.14 to 6.27||1.39||0.84 to 2.30||1.47||0.88 to 2.46|| || |
In the group of obese (BMI > 29 kg/m2) mothers, there is an overall statistically significant increased risk for cardiovascular defects, except for complex ones, but the number is low in that group (N = 45), except for unspecified defects. In the subgroup of morbid obesity, BMI >35 kg/m2, the ORs are still higher and reach statistical significance also in the complex group. Because of the small numbers, this subgroup was not presented for specific cardiac defects.
The proportional attributable risk of cardiovascular defects caused by maternal obesity (BMI > 29 kg/m2) is [(1.18 − 1)/1.18] = 0.15. In the general population of delivered women in Sweden, 10.6% are obese; thus, the population attributable fraction of cardiac defect is 0.15 × 10.6% = 1.6.
For some specific types of cardiovascular defects where numbers were reasonably large, separate analyses were performed (Table 4). There was an increased risk for all specific cardiac defects studied, but the only ones reaching statistical significance were ventricular septal defects and atrial septal defects. There was no statistically significant difference between the highest OR for D-transposition (1.48) and the lowest OR for tetralogy of Fallot (1.07; z = 0.99, p = 0.24).
Underweight (BMI < 19.8 kg/m2) seemed to be protective for isolated cardiovascular defects (Table 4). The risk among the underweight women was slightly below unity in all subgroups, although the findings could be random. The pattern of lower ORs among underweight women was consistent across the specific defect subgroups evaluated, except for tetralogy of Fallot (OR = 1.26; 95% CI, 0.42 to 3.76). CIs become wide in the analysis of specific subgroups of cardiovascular defects and include unity.
We observed a positive association between maternal obesity (BMI > 29 kg/m2) in early pregnancy and congenital heart defects in the offspring. This association has been pointed out in other studies, although low numbers of cases limited their statistical power. A case-control study from Germany reported elevated ORs among women with BMI >30 kg/m2 for transposition of the great arteries and truncus arteriosus, but no increased risk was found for any cardiovascular defect (7). A recent study concluded that obese African-American women were more likely to have infants with a cardiac anomaly (OR = 6.5; 95% CI, 1.2 to 34.9); this finding was based on seven cases of cardiac malformations, and chromosomal anomalies were not excluded (8). Watkins and Botto found that overweight women (BMI > 26 kg/m2) were more likely to have a child with a major isolated heart defect (OR = 1.36; 95% CI, 0.95 to 1.93; N = 66) than women with BMI = 19.9 to 22.7 kg/m2 (6). In a study from California (10), no effect of BMI was seen at conotruncal defects, but numbers were low. The percentage of women with a BMI >29 kg/m2 were the same (9%) among case and control mothers.
Our study can be seen as a large case-control study based on 6801 mothers giving birth to an infant with a congenital heart defect. The completeness of ascertainment of the infants with congenital cardiovascular defects is probably high. We used three Swedish medical health registers as sources of information, and the best possible diagnosis for each infant was identified. Infants with chromosomal anomalies and infants of mothers with diabetes were excluded from the analysis because of known associations with cardiovascular malformations. The advantage of register studies is the large number of individuals, which gives high statistical power and makes it possible also to demonstrate weak effects on reproductive outcome. The drawback is the sometimes low validity of infant cardiac diagnosis, despite the use of multiple sources. Within the group of unspecified cardiovascular defects, a substantial proportion of infants probably did not have such a defect (12). A dilution of the material with healthy infants will bias the observed risk estimates toward unity. In Table 4, there is a tendency for lower ORs at high BMI for unspecified defects than for the other groups.
Exposure information (weight and height) was recorded in early pregnancy and, therefore, was prospective with regard to the identification of congenital cardiac defects. Recall bias was thus avoided. Some effect on the weight measurement will exist according to when during the first trimester the woman attended the antenatal clinic, but it is unlikely that this will covary with the risk for a cardiac defect. Theoretically, maternal alcoholism may result in late attendance, but quantitatively, this is a small error because maternal alcoholism is rare among pregnant women in Sweden; in addition, in order to have a confounding effect, there would need to be an association between alcoholism and obesity.
A number of potential confounding factors associated with maternal obesity, such as maternal age, parity, and smoking in early pregnancy, were adjusted for in this study, but they changed the OR estimates little. A putative confounding factor, not stratified for in this study, is socioeconomic level, which could have affected the result, but smoking during pregnancy is strongly correlated with socioeconomic level in Sweden. Maternal education can be used as a proxy for socioeconomic conditions—it does not covary with the risk for a cardiovascular defect in Sweden. Information on maternal education is available only up to 1996; for the period 1992 to 1995, we found no correlation between maternal educational level and the risk for a cardiac defect (M.I. Cedergren and B.A.J. Källén, unpublished information).
Another problem concerning studies in this field is the definition of obesity. Different thresholds or cut-off values for defining obesity were used in different studies, which makes it difficult to compare risk estimates. We found a slightly increased risk for infant cardiovascular defects of any type in the overweight group (BMI = 26.1 to 29 kg/m2), although the finding could be random, whereas maternal obesity (BMI > 29 kg/m2) was significantly associated with cardiovascular defects in the offspring. This association was even more pronounced among the morbidly obese (BMI > 35 kg/m2). These findings indicate the necessity of dividing the group of women with excess weight into not only overweight, but also obese and morbidly obese (BMI > 35 kg/m2), categories when conducting further studies. An “epidemic” of obesity is a current issue in Sweden as well as in other developed countries, especially in the United States (22). In our sample, 10.6% of the control mothers had a BMI >29 kg/m2, which is in accordance with a recent survey in Sweden (11). That study reported the prevalence of overweight (BMI ≥ 25 kg/m2) to be 41.6% in Swedish women, which is higher than in our pregnant women, where 23% had a BMI >26 kg/m2.
The observed increased risk among obese women seems to be similar for isolated and nonisolated cardiovascular defects in infants. There is no major difference in the risk estimates whether the defect is severe or mild. This finding speaks toward an overall excess risk for cardiovascular defects associated with maternal obesity.
Among specific cardiovascular defects, there appeared to be an increased risk of ventricular septal defects and atrial septal defects. Despite the large number of subjects, no statistically significant association could be found between maternal obesity and hypoplastic left heart syndrome, transposition, endocardial cushion defect, tetralogy of Fallot, or coarctation of aorta, although ORs were increased. Even larger groups of these specific cardiac defects are needed to evaluate this indicated association.
Whatever the underlying mechanism for the observed association between maternal obesity and infant cardiovascular defect is, it affects fetal heart development in early pregnancy. The critical period for most defects occurs 14 to 60 days after conception. A likely explanation is undetected type 2 diabetes, because obese women, in the absence of overt diabetes, have been found to have an impaired glucose metabolism (23), which may be associated with an increased risk of congenital heart disease (24, 25). Another possible explanation for the association between obesity and cardiovascular defects could be improper nutrition. It has been suggested that lack of folic acid could increase the risk of a cardiovascular defect (26, 27), but data on folic acid supplementation in Sweden gave no support for such an association (28).
The observed association between maternal obesity and cardiovascular defects in the offspring is of importance from various points of view. Congenital cardiovascular defects contribute to increased morbidity and mortality during childhood, and little is known about putative preventive mechanisms. Overweight or obesity among women of childbearing age is associated with a number of complications, especially later in life. Pregnancy is a life event when women may be more inclined to behavioral changes. It is possible that appropriate counseling and management during pregnancy for the obese women could reduce the gestational weight gain, improve perinatal outcome, and also contribute to persistent behavioral changes concerning nutrition and physical exercise.
Östergötland County Council (M.C.) and a grant from K. and A. Wallenberg's foundation (B.K.) supported this study. We thank the National Board of Health and Social Welfare, Stockholm, for access to the health registers.
Nonstandard abbreviations: ICD, International Classification of Disease; OR, odds ratio.