maternal obesity and morbid obesity: the risk for birth defects in the offspring

Maternal Obesity and Morbid Obesity:the Risk for Birth Defects in the Offspring

Marie I. Blomberg1*y and Bengt Kallen2

1Division of Obstetrics and Gynaecology, Department of Clinical and Experimental Medicine, Faculty of Health Sciences,University of Linkoping, Linkoping, Sweden

2Tornblad Institute, University of Lund, Lund, Sweden

Received 13 May 2009; Revised 22 June 2009; Accepted 24 June 2009

BACKGROUND: The objective of this study was to assess, in a large data set from Swedish Medical HealthRegistries, whether maternal obesity and maternal morbid obesity were associated with an increased risk forvarious structural birth defects. METHODS: The study population consisted of 1,049,582 infants born in Swe-den from January 1, 1995, through December 31, 2007, with known maternal weight and height data.Women were grouped in six categories of body mass index (BMI) according to World Health Organizationclassification. Infants with congenital birth defects were identified from three sources: the Swedish MedicalBirth Registry, the Register of Birth Defects, and the National Patient Register. Maternal age, parity, smoking,and year of birth were thought to be potential confounders and were included as covariates in the adjustedodds ratio analyses. RESULTS: Ten percent of the study population was obese. Morbid obesity (BMI � 40)occurred in 0.7%. The prevalence of congenital malformations was 4.7%, and the prevalence of relativelysevere malformations was 3.2%. Maternal prepregnancy morbid obesity was associated with neural tubedefects OR 4.08 (95% CI 1.87–7.75), cardiac defects OR 1.49 (95% CI 1.24–1.80), and orofacial clefts OR 1.90(95% CI 1.27–2.86). Maternal obesity (BMI � 30) significantly increased the risk of hydrocephaly, anal atresia,hypospadias, cystic kidney, pes equinovarus, omphalocele, and diaphragmatic hernia. CONCLUSION: Therisk for a morbidly obese pregnant woman to have an infant with a congenital birth defect is small, but forsociety the association is important in the light of the ongoing obesity epidemic. Birth Defects Research (PartA) 88:35–40, 2010. � 2009 Wiley-Liss, Inc.

Key words: obesity; birth defects; cardiac defects; neural tube defects; orofacial clefts

INTRODUCTION

Maternal obesity is associated with an increased risk ofa range of structural birth defects. The association ismost pronounced for neural tube defects (Waller et al.,1994; Shaw et al., 1996; Werler et al., 1996; Kallen, 1998;Anderson et al., 2005; Ray et al., 2005; Waller et al., 2007;Rasmussen et al., 2008), cardiac defects (Moore et al.,2000; Shaw et al., 2000; Watkins and Botto, 2001; Mikhailet al., 2002; Cedergren and Kallen, 2003; Watkins et al.,2003; Waller et al., 2007), and orofacial clefts (Watkinset al., 2003; Cedergren and Kallen, 2005; Waller et al.,2007; Villamor et al., 2008). In a recently performed meta-analysis, obese mothers compared to mothers of recom-mended body mass index (BMI) were also at increasedrisk of having infants with anorectal atresia, hydroce-phaly, or limb reduction defects (Stothard et al., 2009)The risks are likely to be small for the individual preg-nant woman, but of general concern in the light of theongoing obesity epidemic.

There is still lack of evidence whether the associationbetween maternal prepregnancy overweight and obesityis associated with birth defects with low prevalence rates,such as kidney malformations, diaphragmatic hernia, andsevere eye and ear malformations. Insufficient numbersof cases in each study make data difficult to interpret,results from available studies are inconclusive, and thepossible association has not been investigated at all.When studying obstetric outcome, for example, pree-

clampsia, cesarean section, large for gestational ageinfants, and stillbirths, in relation to maternal pregreg-nancy BMI, there is convincing evidence that the risk

*Correspondence to: Marie Blomberg, Division of Obstetrics andGynaecology, Department of Clinical and Experimental Medicine, Facultyof Health Sciences, University of Linkoping, SE-581 85 Linkoping, Sweden.E-mail: [email protected] I. Blomberg was formerly known as Marie Cedergren.Published online 26 August 2009 in Wiley InterScience (www.interscience.wiley.com).DOI: 10.1002/bdra.20620

Birth Defects Research (Part A): Clinical and Molecular Teratology 88:35�40 (2010)

� 2009 Wiley-Liss, Inc. Birth Defects Research (Part A) 88:35�40 (2010)

increases with increasing maternal BMI, reaching thehighest odds ratios among morbidly obese women(Cedergren, 2004). There are a few studies evaluatingthe risk of congenital birth defects over the obesity class I–III strata, suggested by the WHO (World Health Organi-zation, 2000). A recent meta-analysis concerning maternalobesity and risk of neural tube defects included five stud-ies describing the risk associated with severe obesity,odds ratio 3.11 (95% CI 1.75–5.46). For all obese mothersthe odds ratio was 1.70 (95% CI 1.34–2.15). The risk formothers with severe obesity was thus nearly twice ashigh as that of obese mothers, although severe obesitywas defined differently in the five included studies(Rasmussen et al., 2008).

The objective of this study was to assess, in a largedata set from Swedish medical health registries, whethermaternal obesity and maternal morbid obesity wereassociated with an increased risk for various structuralbirth defects.

MATERIALS AND METHODS

The study population consisted of infants born in Swe-den January 1, 1995, through December 31, 2007. Theywere identified using the Swedish Medical Birth Registry(National Board of Health and Welfare, 2003). Medicaland other data on almost all (98–99%) deliveries inSweden are listed in the register, which also includesstillbirths after 28 weeks of gestation. It is based on cop-ies of the standardized medical record forms completedat the maternity health care centers at the start of prena-tal care, usually in gestational week 10–12, records fromthe delivery units, and the pediatric examination of thenewborn. The system is identical throughout the country.A description and validation of the register content isavailable (National Board of Health and Welfare, 2003).

Midwives recorded pregregnancy maternal weight andheight on a standardized form at the first visit to the ma-ternity health care center. Ninety percent of women whowill give birth present themselves to the antenatal clinicduring the first trimester of their pregnancy.

Body mass index (BMI, kg/m2) was calculated frommaternal weight and height data. Data were missing forabout 15%. Women with known BMI were grouped insix categories of BMI: underweight (<18.5), averageweight (18.5–24.9), overweight (25–29.9), obese class I(30–34.9), obese class II (35–39.9), and obese class III(�40) according to the WHO (World Health Organiza-tion, 2000).

Infants with congenital birth defects were identifiedfrom three sources: the Swedish Medical Birth Registry,the Swedish Register of Birth Defects (previously calledthe Registry of Congenital Malformations), and theNational Patient Register (previously called the HospitalDischarge Registry) (National Board of Health and Wel-fare, 2004). The Medical Birth Register and the Registerof Birth Defects both contain data from the neonatalperiod. The National Patient Register contains informationon discharge diagnoses of all patients admitted to Swedishhospitals and was used up to and including 2007.

Any congenital malformations were defined as ICD-9codes 740.0–759.9 or ICD-10 codes beginning with Q.Relatively severe malformations exclude infants who hadonly one or more of the following diagnoses: preauricularappendix, tongue tie, patent ductus arteriosus in preterm

infants, single umbilical artery (SUA), undescended tes-ticle, unstable hip, and nevus. These conditions are com-mon, of minor clinical significance, and with a markedlyvariable registration rate between different hospitals. Thisrestriction was made to obtain a better defined and morerelevant definition of malformations, but it should berealized that some minor anomalies are still included inthe concept.Any cardiovascular malformation excludes patent

ductus arteriosus in preterm infants and SUA. Seriouscardiac defects were defined as tetralogy of Fallot, hypo-plastic left heart syndrome, transposition of great arteriesand common truncus, coarctation of aorta, Ebstein anom-aly, and endocardial cushion defect.The analysis was performed for each malformation

type regardless of the possible presence of other malfor-mations but with the exclusion of infants with knownchromosome anomalies.All three information sources were linked using the

personal identification numbers of mothers and/orinfants. These numbers are given to each person living inSweden and are extensively used in society, including allhealth care. In most instances, the diagnoses agreedbetween the three registers (but cases were not alwaysrecorded in all three registers).In this study, the rate of infants with a congenital mal-

formation reported to any one of the above-describedregisters and with maternal data in the Medical BirthRegister was estimated according to maternal BMI class.The malformation rates for infants of obese or morbidobese women were compared to those for infants ofwomen with a normal BMI.We did not remove from the analysis women with a

diagnosis of diabetes. Obesity is mainly associated withdiabetes type 2, which will act as an intermediary in theeffects of obesity. The strongest effect on congenitalmalformations is from diabetes type 1.Maternal age (seven five-year classes), parity (1–41),

smoking (unknown, no smoking, <10 cigarettes/day,�10 cigarettes/day), and year of delivery were thoughtto be potential confounding factors and were included ascovariates in the adjusted analyses. The register informa-tion on these variables was obtained from the antenatalcare center records. Odds ratios (OR) were determinedusing Mantel-Haenszel technique for adjusted rates(Mantel and Haenszel, 1959). Estimates of 95% confi-dence intervals (95% CI) were made with a test-basedmethod, based on the Mantel-Haenszel chi-square (Miet-tinen, 1974). When the expected number of cases waslow (<10), the ratio between observed and expectednumbers were used, and 95% confidence intervals werebased on exact Poisson distributions.

RESULTS

A total of 1,235,877 infants were born during the studyperiod. Maternal prepregnancy weight and height couldbe retrieved for 84.9%. The risk for any congenital mal-formation and the risk for a relatively severe malforma-tion among infants of women with unknown BMI wasOR 0.98 (95% CI 0.95–1.01) and OR 0.99 (95% CI 0.96–1.03) when compared to infants of women with a knownBMI. This indicates that the data loss is not skewed.Thirty-five percent of the study population was over-

weight, and 10.2% were obese. Morbid obesity (BMI �

36 BLOMBERG AND KALLEN

Birth Defects Research (Part A) 88:35�40 (2010)

40) occurred in 0.7%. The prevalence of any congenitalmalformation among mothers with known BMI was 4.7%(49,630/1,049,582), and the prevalence of relatively severemalformations was 3.2% (33,821/1,049,582). The oddsratios for any congenital malformation and the oddsratios for relatively severe malformations according tomaternal BMI are presented in Figure 1.

Maternal characteristics are presented in Table 1. Theobese mothers were slightly older, more often multipa-rous, and more often smokers than normal weightwomen.

Table 2 shows the distribution of congenital malforma-tions among infants of women who were underweight,normal weight, overweight, or obese before pregnancy.

No statistically significant association between under-weight women and infant birth defects was seen,although the risks for abdominal wall defects and hydro-cephaly were increased.Maternal overweight was associated with an increased

risk for neural tube defects, hydrocephaly, severe earmalformations, orofacial clefts, cardiac defects, pes equi-novarus, and omphalocele.Obese women had an increased risk for neural tube

defects, hydrocephaly, cardiac defects, orofacial clefts,anal atresia, hypospadias, cystic kidney, pes equinovarus,omphalocele, and diaphragmatic hernia.There was a decreased risk for gastroschsis among

infants of obese women of marginal statistical signifi-cance and based on only five exposed cases. None of themothers of infants with gastroschisis were of obesityclass II–III (3.3 expected). In 33 infants with abdominalwall defects, the defect was of another nature thanomphalocele or gastroschisis or was unclassifiable (if cod-ing was made with ICD-9, the specific subtypes couldnot be differentiated).Four major congenital heart defects were also analyzed.

There were 444 infants with tetralogy of Fallot, 514 withtransposition of great vessels, 274 with hypoplastic leftheart syndrome (HLHS), and 156 with common truncus.These four major defects were together significantly asso-ciated with maternal obesity, OR 1.31 (95% CI 1.08–1.58).Each heart defect was also evaluated separately in rela-tion to maternal obesity. The risk for tetralogy of Fallotamong obese women was OR 1.37 (95% CI 0.99–1.88), fortransposition OR 1.14 (95% CI 0.83–1l.58), for HLHS 1.67(95% CI 1.13–2.46), and for common truncus was 1.26(95% CI 0.72–2.21). Thus, only for HLHS the increasedOR is statistically significant, but all ORs may well esti-mate the same common OR.In Table 3 data for obese women were divided into

three classes (I–III) of obesity. The risk for neural tubedefects in infants of women with obesity class I wasalmost twofold increased compared to infants of womenwith a normal BMI. Infants of women with morbid obe-

Figure 1. The odds ratios for any congenital malformation (solidline) and the odds ratios for relatively severe malformations(dashed line) according to maternal BMI. Vertical lines indicateconfidence intervals.

Table 1Maternal Characteristics

<18.5 18.5–24.9 25–29.9 30–34.9 35–39.9 �40Maternal Body Mass Index N 5 26,460 N 5 660,493 N 5 256,017 N 5 77,052 N 5 22,236 N 5 7324

(%) (%) (%) (%) (%) (%)

Maternal age<20 1382 (5.2) 12,462 (1.9) 3609 (1.4) 1057 (1.4) 261 (1.2) 72 (1.0)20–24 6780 (25.6) 91,610 (13.9) 33,551 (13.1) 11,112 (14.4) 3235 (14.5) 1024 (14.0)25–29 9153 (34.6) 221,959 (33.6) 81,812 (32.0) 24,692 (32.4) 7284 (32.8) 2349 (32.1)30–34 6700 (25.3) 226,898 (34.4) 86,803 (33.9) 25,016 (32.5) 7134 (32.1) 2442 (33.3)35–39 2125 (8.0) 91,951 (13.9) 41,520 (16.2) 12,494 (16.2) 3552 (16.0) 1207 (16.5)40–45 304 (1.1) 15,085 (2.3) 8367 (3.3) 2558 (3.3) 741 (3.3) 224 (3.1)145 16 528 355 123 29 6

Parity1 13,757 (52.0) 301,002 (45.6) 98,266 (38.4) 27,151 (35.2) 7698 (34.6) 2516 (34.4)2 9177 (34.7) 240,658 (36.4) 94,386 (36.7) 27,809 (36.1) 7882 (35.4) 2588 (35.3)3 2865 (10.8) 87,551 (13.3) 41,329 (16.1) 13,142 (17.1) 3790 (17.0) 1216 (16.6)4 961 (3.6) 31,282 (4.7) 22,036 (8.6) 8950 (11.6) 2866 (12.9) 1004 (13.7)

Maternal smokingUnknown 590 (2.2) 13,551 (2.1) 5472 (2.1) 1589 (2.1) 495 (2.2) 170 (2.3)No smoking 21,884 (82.7) 581,208 (88.0) 220,421 (86.1) 64,237 (83.4) 18,126 (81.5) 5946 (81.2)<10 cig/day 2805 (10.6) 46,429 (7.0) 20,513 (8.0) 7398 (9.6) 2287 (10.3) 706 (9.6)�10 cig/day 1181 (4.5) 19,305 (2.9) 9611 (3.8) 3828 (5.0) 1328 (6.0) 502 (6.9)

37MATERNAL MORBID OBESITY AND BIRTH DEFECTS


sity (class III) had a fourfold increased risk of neural tubedefects. The same pattern could be seen concerning therisk of orofacial clefts, a slightly increased risk amongwomen in obesity class I, and a twofold increased riskamong morbidly obese women. The risk for cardiacdefects also increases with increasing degree of maternalobesity. Anal atresia was significantly associated withmorbid maternal obesity (class III). The risk for hypospa-dias was almost the same over the obesity class strata

(class I OR 1.30, class II OR 1.29, class III OR 1.40), butdid not reach statistical significance among infants of themost obese women.

DISCUSSION

This large population-based study showed that mater-nal prepregnancy obesity was associated with a numberof severe congenital birth defects: neural tube defects,

Table 2Adjusted Odds Ratios for the Association between Maternal Body Mass Index and Congenital Malformations

BMI < 18.5 BMI 18.5–24.9 BMI 25–29.9 BMI � 30

N OR * (95% CI) N OR * (95% CI) N OR * (95% CI) N OR * (95% CI)

Neural tube defect 9 1.03 (0.47–1.95) 207 1.00 109 1.33 (1.06–1.66) 64 2.04 (1.53–2.71)Hydrocephaly 14 1.83 (1.00–3.08) 178 1.00 82 1.31 (1.04–1.66) 43 1.59 (1.14–2.23)Microcephaly 4 1.09 (0.30–2.79) 84 1.00 42 1.29 (0.91–1.80) 17 1.27 (0.75–2.15)An- or microphthalmia 1 0.75 (0.02–41.8) 39 1.00 14 0.87 (0.48–1.59) 6 0.92 (0.34–2.00)Severe ear malformation 4 0.93 (0.25–2.38) 107 1.00 61 1.45 (1.07–1.95) 25 1.53 (0.99–2.27)Cardiac defects 265 0.96 (0.65–2.38) 6,823 1.00 2796 1.06 (1.01–1.11) 1279 1.17 (1.10–1.24)Choanal atresia 0 57 1.00 31 1.24 (0.81–1.92) 11 1.11 (0.59–2.13)Orofacial clefts 46 1.02 (0.76–1.37) 1,132 1.00 508 1.15 (1.04–1.28) 232 1.26 (1.09–1.95)Esophageal atresia 9 1.30 (0.59–2.47) 169 1.00 64 1.05 (0.80–1.37) 34 1.34 (0.92–1.95)Small gut atresia 5 0.79 (0.26–1.84) 152 1.00 75 1.23 (0.94–1.60) 28 1.16 (0.78–1.75)Anal atresia 12 1.22 (0.63–2.13) 222 1.00 102 1.21 (0.98–1.51) 65 1.87 (1.42–2.47)Hypospadias 85 1.15 (0.92–1.43) 1,781 1.00 735 1.08 (0.99–1.18) 364 1.31 (1.17–1.46)Kidney agenesis/dysgenesis 8 1.09 (0.43–2.15) 174 1.00 77 1.15 (0.89–1.48) 32 1.14 (0.77–1.68)Cystic kidney 13 1.37 (0.73–2.74) 230 1.00 96 1.11 (0.89–1.38) 52 1.40 (1.03–1.90)Pes equinovarus 35 0.99 (0.71–1.40) 826 1.00 438 1.36 (1.25–1.53) 205 1.50 (1.29–1.75)Limb reduction defects 14 0.87 (0.51–1.49) 377 1.00 150 1.04 (0.86–1.27) 69 1.22 (0.94–1.58)Carniostenosis 18 1.13 (0.70–1.83) 398 1.00 163 1.06 (0.88–1.28) 68 1.09 (0.84–1.42)Diaphragmatic hernia 7 1.22 (0.49–2.51) 145 1.00 60 1.12 (0.84–1.47) 41 1.81 (1.29–2.55)Abdominal wall defecta 14 1.75 (0.91–2.93) 155 1.00 64 1.23 (0.95–1.59) 24 1.02 (0.67–1.57)Gastroschisis 6 1.28 (0.47–2.78) 79 1.00 31 1.14 (0.79–1.66) 5 0.42 (0.18–1.00)Omphalocele 5 2.30 (0.75–5.37) 55 1.00 25 1.28 (0.84–1.96) 18 2.03 (1.42–2.90)

Adjustments were made for maternal age, parity, smoking in early pregnancy, and year of birth using the Mantel-Haenszel method.aA total number of 33 cases had another abdominal wall defect than gastroschisis or omphalocele, or could not be definitively

classified.

Table 3Adjusted Odds Ratios for the Association between Maternal Obesity Class I–III and Congenital Malformations

BMI 30–34.9 BMI 35–39.9 BMI � 40

N OR * (95% CI) N OR * (95% CI) N OR * (95% CI)

Neural tube defect 41 1.80 (1.28–2.54) 14 2.07 (1.13–3.48) 9 4.08 (1.87–7.75)Hydrocephaly 28 1.40 (0.93–2.10) 10 1.82 (0.87–3.35) 5 2.74 (0.89–6–39)Microcephaly 14 1.41 (0.79–2.52) 3 1.09 (0.22–3.18) 0An- or microphthalmia 3 0.68 (0.14–1.98) 3 2.11 (0.44–6.16) 0Severe ear malformation 18 1.52 (0.92–2.43) 6 1.74 (0.64–3.78) 1 0.83 (0.02–4.62)Cardiac defects 877 1.11 (1.03–1.19) 290 1.26 (1.12–1.42) 112 1.49 (1.24–1.80)Choanal atresia 9 1.31 (0.60–2.49) 0 2 3.36 (0.41–12.1)Orofacial clefts 146 1.09 (0.91–1.30) 62 1.62 (1.26–2.09) 24 1.90 (1.27–2.86)Esophageal atresia 64 1.05 (0.80–1.37) 20 1.08 (0.68–1.73) 9 1.69 (0.77–3.21)Small gut atresia 23 1.33 (0.85–2.06) 4 0.78 (0.21–2.01) 1 0.60 (0.02–3.32)Anal atresia 45 1.77 (1.29–2.44) 11 1.48 (0.74–2.64) 9 3.72 (1.70–7.07)Hypospadias 262 1.30 (1.14–1.48) 75 1.29 (1.02–1.63) 27 1.40 (0.95–2.06)Kidney agenesis/dysgenesis 18 0.86 (0.52–1.47) 11 1.87 (0.93–3.34) 3 1.63 (0.34–4.76)Cystic kidney 36 1.33 (0.93–1.89) 11 1.40 (0.70–2.59) 5 2.02 (0.65–4.70)Pes equinovarus 158 1.59 (1.34–1.88) 32 1.14 (0.80–1.63) 15 1.60 (0.90–2.65)Limb reduction defects 50 1.21 (0.90–1.63) 13 1.09 (0.62–1.90) 6 1.61 (0.59–3.51)Carniostenosis 43 0.95 (0.70–1.31) 20 1.53 (0.97–2.42) 5 1.23 (0.40–2.88)Diaphragmatic hernia 32 1.94 (1.33–2.83) 7 1.42 (0.57–2.93) 2 1.29 (0.16–4.67)Abdominal wall defect 19 1.11 (0.69–1.79) 3 0.62 (0.13–1.83) 2 1.25 (0.15–4.51)Chromsome anomaly 176 1.18 (1.00–1.38) 55 1.29 (0.98–1.69) 20 1.48 (0.93–2.31)

Adjustments were made for maternal age, parity, smoking in early pregnancy, and year of birth using the Mantel-Haenszel method.



hydrocephaly, cardiac defects, orofacial clefts, anal atre-sia, hypospadias, cystic kidney, pes equinovarus, ompha-locele, and diaphragmatic hernia. The risk increased withincreasing degree of obesity for neural tube defects, car-diac defects, and orofacial clefts, reaching the highestlevel among the morbidly obese women.

Morbid obesity before pregnancy is strongly associatedwith adverse obstetric and neonatal outcome (Cedergren,2004). Current knowledge mainly concerning obstetricoutcomes supports a division of obesity into threeclasses, as the magnitude of risk differs significantlybetween the classes. Whether the risk of a birth defectincreases with increasing maternal pregregnancy obesityhas not been sufficiently evaluated. This is mainlybecause of low numbers of women with BMI above 35 orabove 40 in combination with low prevalence outcomes(birth defects). A recent meta-analysis examined the rela-tionship between maternal morbid obesity and the risk ofneural tube defects. The overall risk was concluded to beOR 3.11 (95% CI 1.75–5.46) based on five studies(Rasmussen et al., 2008). The definition of morbid obesityvaried substantially: two studies used maternal weight >110 kg, two studies defined morbid obesity as BMI > 38,and the fifth used BMI � 40. Anderson et al. (2005)defined morbid obesity as BMI � 40, the same definitionas we used. The number of cases evaluated in that studywas one case with anencephaly and nine cases withspina bifida in the morbid obesity group, numbers com-parable with those in our study. The risk estimates forneural tube defects in association with maternal obesitywas OR 4.08 (95% CI 1.87–7.75) in our study compared toOR 4.8 (95% CI 1.7–13) (spina bifida only) in their case-control study, which was based on retrospectivestructured telephone interviews and a relatively largedrop-out rate.

As far as we know, no earlier studies described therisk of cardiac defects among morbidly obese women. Inthe present data set, the risk of cardiac defects amonginfants of morbidly obese women was increased by 50%,based on 112 cases. We had an insufficient number ofcases in this group to evaluate the different cardiacdefects separately. This would, of course, be of greatinterest as the severity of cardiac defects differs substan-tially, ranging from insignificant septal defects to life-threatening conditions such as HLHS. An earlier studyfrom the Swedish medical health registers indicated thatmaternal obesity (BMI � 29) was more strongly associ-ated with severe cardiac defects than with less severedefects (Cedergren and Kallen, 2003). This study alsorevealed an increased risk among obese mothers of fourmajor cardiac defects: tetralogy of Fallot, transposition ofgreat vessels, HLHS, and common truncus.

We were also able to analyze the risk of orofacial cleftsover the maternal obesity class strata. The pattern wascomparable with that of cardiac defects, although theabsolute risk estimates were slightly more pronounced,reaching a twofold increased risk among the morbidlyobese women. We have found no previous studiesconcerning the risk of orofacial clefts among infants ofmorbidly obese women.

One possibility is that the increased risks of neuraltube defects, cardiac defects, and orofacial clefts, higherin relation to the higher degree of obesity the mother suf-fers from, could be explained by insufficient detectionrates antenatally potentially resulting in fewer termina-

tions of pregnancy. A study evaluated ultrasound exami-nations for singleton gestations. The rate of suboptimalultrasound visualization (SUV) of the fetal structures washigher for obese than nonobese women for both cardiac(37.3 vs. 18.7%) and craniospinal structures (42.8 vs.29.5%). Increased severity of maternal obesity was associ-ated with increased SUV rate for both cardiac and cranio-spinal structures (Hendler et al., 2004). It is possible thatinsufficient detection of neural tube defects among mor-bidly obese women could affect the birth rate. It is lesslikely that incomplete detection rates by ultrasoundexplain the increased risk of cardiac defects among obesewomen, as only four chamber views are included inmost ultrasound screening protocols. Orofacial clefts arenot screened for at all in general clinical practice.Another potential explanation for an increased risk of

congenital defects among obese mothers is undetected di-abetes type II with hyperglycemia and insulin resistance(Towner et al., 1995; Hendricks et al., 2001).Maternal obesity was significantly associated with an

increased risk for anal atresia (N 5 65), OR 1.87. The riskwas even more pronounced among the morbid obesewomen, OR 3.72. That is in accordance with data fromthe National Birth Defects Prevention Study, where therisk of anorectal atresia (N 5 75) associated with mater-nal obesity was OR 1.46 (95% CI 1.10–1.95) (Waller et al.,2007), mothers with preexisting diabetes excluded. Thesedata were retrospective and had a high nonresponse rate.There was an almost twofold increased risk for dia-

phragmatic hernia among obese mothers in this study.When dividing obesity into three classes, the increasedrisk persisted among infants of women in obesity class I.This result contradicts findings in a recent meta-analysiswhere no association between maternal obesity and therisk of an infant affected by diaphragmatic hernia wasfound (based on four studies) (Stothard et al., 2009).In the same meta-analysis no association between

maternal obesity and hypospadias was found, althoughin our data set the risk was slightly but significantlyincreased.Maternal obesity seems to be protective against the

development of gastroschisis: OR 0.17 (95% CI 0.10–0.30)(Stothard et al., 2009). We subanalyzed our data ofabdominal wall defects and also found a decreased riskof gastroschisis among obese mothers, OR 0.42 (95% CI0.18–1.00) in spite of adjustment for maternal age. Thereis, however, a very strong association between very lowmaternal age and gastroschisis, and the adjustment mayhave been incomplete.Cystic kidney occurred significantly more often among

infants of the group of obese women compared tonormal weight women, in the present study based on 52cases. Honein et al. (2003) investigated women with highBMI, defined as BMI � 25, compared to women withBMI < 25, and found no association between renalanomalies in the infants and high BMI (N 5 19).An advantage of register studies is the presence of

many individuals, which gives high statistical power andmakes it possible also to demonstrate weak effects onreproductive outcome. Another advantage is the accessto information on putative confounders.The completeness of ascertainment of the infants with

congenital malformations was probably high becausemultiple sources for ascertainment were used. Exposureinformation (weight and height) was recorded in early

39MATERNAL MORBID OBESITY AND BIRTH DEFECTS


pregnancy and was therefore prospective as regards theidentification of the congenital malformation. Recall biaswas thus avoided.

A problem concerning studies in this field is the defini-tion of obesity. Different thresholds or cutoff values fordefining obesity were used in different studies, whichmakes it difficult to exactly compare risk estimates. Ourdata were based on the WHO classification.

The odds ratio assesses how much more likely awoman with morbid obesity is to give birth to a childwith a relatively severe congenital birth defect comparedto an average weight woman, but it gives no indicationof the significance of the excess risk in the population.This could be measured by the attributable riskexpressed as a proportion of the total incidence rateamong the exposed. The proportional attributable risk ofa relatively severe congenital birth defect due to maternalmorbid obesity (BMI > 40) is ((1.41–1)/1.41) 5 0.29; thusmaternal morbid obesity accounts for 29% of all cases ofsevere congenital birth defect among morbidly obesemothers. The proportion of morbidly obese women is,however, small (0.7%), so the contribution to the totalnumber of relatively severe congenital malformations bymaternal morbid obesity is also small, about 3 per 1000.At present, overweight or moderate obesity may contrib-ute more to the burden of congenital malformationsbecause of the larger numbers in these groups.

In conclusion, this large register study showed thatmaternal pregregnancy morbid obesity was associatedwith neural tube defects, cardiac defects, and orofacialclefts, reaching the highest risk estimate among womenwith BMI � 40. Maternal obesity increased the risk ofseveral severe congenital birth defects: hydrocephaly,anal atresia, hypospadias, cystic kidney, pes equinovarus,omphalocele, and diaphragmatic hernia. The absoluterisk for the individual pregnant woman is still low, butin light of the ongoing obesity epidemic, these results area challenge on a population-based level.

ACKNOWLEDGMENTSWe thank the National Board of Health and Social

Welfare, Stockholm, for access to the health registers. Thestudy was supported by a grant from Evy and GunnarSandberg Foundation (B.K.).

REFERENCES

Anderson JL, Waller KD, Canfield MA, et al. 2005. Maternal obesity, ges-tational diabetes, and central nervous system birth defects. Epidemi-ology 16:87–92.

Cedergren M, Kallen B. 2005. Maternal obesity and the risk for orofacialclefts in the offspring. Cleft Palate Craniofac J 42:367–371.

Cedergren MI. 2004. Maternal morbid obesity and the risk of adversepregnancy outcome. Obstet Gynecol 103:219–224.

Cedergren MI, Kallen BA. 2003. Maternal obesity and infant heart defects.Obes Res 11:1065–1071.

Hendler I, Blackwell SC, Bujold E, et al. 2004. The impact of maternalobesity on midtrimester sonographic visualization of fetal cardiacand craniospinal structures. Int J Obes Relat Metab Disord 28:1607–1611.

Hendricks KA, Nuno OM, Suarez L, Larsen R. 2001. Effects of hyperinsu-linemia and obesity on risk of neural tube defects among MexicanAmericans. Epidemiology 12:630–635.

Honein MA, Moore CA, Watkins ML. 2003. Subfertility and prepregnancyoverweight/obesity: possible interaction between these risk factors inthe etiology of congenital renal anomalies. Birth Defects Res A ClinMol Teratol 67:572–577.

Kallen K. 1998. Maternal smoking, body mass index, and neural tubedefects. Am J Epidemiol 147:1103–1111.

Mantel N, Haenszel W. 1959. Statistical aspects of the analyses of datafrom retrospective studies of disease. J Nat Cancer Inst 22:719–748.

Miettinen OS. 1974. Simple interval estimation of risk ratio. Am JEpidemiol 100:515–516.

Mikhail LN, Walker CK, Mittendorf R. 2002. Association between mater-nal obesity and fetal cardiac malformations in African Americans.J Natl Med Assoc 94:695–700.

Moore LL, Singer MR, Bradlee ML, et al. 2000. A prospective study of therisk of congenital defects associated with maternal obesity anddiabetes mellitus. Epidemiology 11:689–694.

National Board of Health and Welfare. 2003. The Swedish Medical BirthRegister: a summary of content and quality. Stockholm: National Boardof Health and Welfare. Available from: http://www.socialstyrelser.se/Publicerat/2003/1697/2003-112-3.htm. Accessed April 14, 2009.

National Board of Health and Welfare. 2004. Registration of congenital mal-formations in the Swedish health registers. Stockholm: National Boardof Health and Welfare. Available from: http://www.socialstyrelsen.se/Publicerat/2006/5120/2004-112-1.htm. Accessed April 14, 2009.

Rasmussen SA, Chu SY, Kim SY, et al. 2008. Maternal obesity and risk ofneural tube defects: a metaanalysis. Am J Obstet Gynecol 198:611–619.

Ray JG, Wyatt PR, Vermeulen MJ, et al. 2005. Greater maternal weightand the ongoing risk of neural tube defects after folic acid flour forti-fication. Obstet Gynecol 105:261–265.

Shaw GM, Todoroff K, Schaffer DM, Selvin S. 2000. Maternal height andprepregnancy body mass index as risk factors for selected congenitalanomalies. Paediatr Perinat Epidemiol 14:234–239.

Shaw GM, Velie EM, Schaffer D. 1996. Risk of neural tube defect-affectedpregnancies among obese women. JAMA 275:1093–1096.

Stothard KJ, Tennant PW, Bell R, Rankin J. 2009. Maternal overweightand obesity and the risk of congenital anomalies: a systematic reviewand meta-analysis. JAMA 301:636–650.

Towner D, Kjos SL, Leung B, et al. 1995. Congenital malformations inpregnancies complicated by NIDDM. Diabetes Care 18:1446–1451.

Villamor E, Sparen P, Cnattingius S. 2008. Risk of oral clefts in relation toprepregnancy weight change and interpregnancy interval. Am J Epi-demiol 167:1305–1311.

Waller DK, Mills JL, Simpson JL, et al. 1994. Are obese women at higherrisk for producing malformed offspring? Am J Obstet Gynecol170:541–548.

Waller DK, Shaw GM, Rasmussen SA, et al. 2007. Prepregnancy obesityas a risk factor for structural birth defects. Arch Pediatr Adolesc Med161:745–750.

Watkins ML, Botto LD. 2001. Maternal prepregnancy weight and congeni-tal heart defects in offspring. Epidemiology 12:439–446.

Watkins ML, Rasmussen SA, Honein MA, et al. 2003. Maternal obesityand risk for birth defects. Pediatrics 111:1152–1158.

Werler MM. Louik C, Shapiro S, Mitchell AA. 1996. Prepregnantweight in relation to risk of neural tube defects. JAMA 275:1089–1092.

World Health Organization. 2000. Obesity: preventing and managing theglobal epidemic. WHO technical report series 894. Geneva, Switzer-land: World Health Organization.



maternal obesity and morbid obesity: the risk for birth defects in the offspring

Documents