Paper 274 Disc

234

Maternal height and prepregnancy body mass index as risk factors for selected congenital anomalies
Gary M. Shaw,a Karen Todoroff,a Donna M. Schafferb and Steve Selvinc
a

March of Dimes Birth Defects Foundation, California Birth Defects Monitoring Program, Emeryville, bDivision of Research, Kaiser Permanente

Medical Care Program, Oakland, and cDivisions of Epidemiology and Biostatistics, University of California, Berkeley, Berkeley, CA, USA

Summary
Correspondence: Gary M. Shaw, California Birth Defects Monitoring Program, 1900 Powell St., Suite 1050, Emeryville, CA 94608, USA. E-mail: gsh@cbdmp.org

Previous studies have observed an increased risk of approximately twofold or more for neural tube defects (NTD) associated with maternal obesity before pregnancy based on a body mass index (BMI) of 4 29 kg/m2. No additional maternal factor appeared substantially to influence this association. Here, we explore further the association between BMI and NTD risk by considering the separate contributions of maternal prepregnant BMI and height. We also explore whether selected congenital anomalies, in addition to NTDs, were associated with maternal height or prepregnant BMI. Data were derived from two California population-based casecontrol studies. One study comprised 538 NTD cases and 539 non-malformed control infants. The other study included an additional 265 NTD cases, as well as 207 conotruncal cases, 165 limb anomaly cases, 662 orofacial cleft cases and 734 non-malformed controls. Maternal interviews in both studies elicited information on maternal height and prepregnant weight. Anomaly risk was described using additive linear logistic regression models. Results revealed increasing NTD risk with increasing maternal prepregnant BMI, controlling for maternal height. These patterns were observed overall as well as for most race/ethnic groups. Increasing NTD risk for decreasing height controlling for maternal BMI was also observed in one NTD study, but was not as evident in the other. Elevated risks for increasing maternal BMI and decreasing maternal height were not observed consistently for the other studied anomalies. The mechanisms underlying the association between maternal weight, or possibly maternal height, and NTD-affected pregnancy risk are unknown. Exploration of other data sets will be needed to determine whether similar patterns of NTD risk or lack of risk for other anomalies are associated with the two maternal anthropometric variables, height and prepregnant weight.

Ahed Bhed Ched Dhed Ref marker Fig marker Table marker Ref end
Ref start

Introduction
Recently, five studies have observed an increased risk of approximately twofold or more for neural tube defects (NTD) associated with maternal obesity before pregnancy.15 Four of these studies1,2,4,5 defined obesity consistent with the Institute of Medicine's definition6 of a body mass index (BMI) of 4 29 kg weight/m height2, and the other study defined obesity based on prepregnant weight only.3 Although these studies explored a variety of covariates, no additional maternal factor, including maternal periconceptional

vitamin use, appeared substantially to influence the observed association between maternal obesity and risk for NTDs. The first goal of this paper was to explore further our previously reported association between obesity and NTD risk2 by considering the separate contributions of prepregnant BMI and height. If the NTD association was stronger with BMI than with height, or vice versa, one might hypothesise differing aetiologies. For example, elevated BMI could be a reflection of a woman's recent dietary intake, whereas shorter

# Blackwell Science Ltd. Paediatric and Perinatal Epidemiology 2000, 14, 234239

Paper 274 Disc

BMI and height as anomaly risk factors stature could be the result of nutritional compromise at early stages of her development (e.g. in early childhood or adolescence). Further, although BMI attempts to standardise the influence of height for weight, some potential residual confounding associated with height may exist and, therefore, both BMI and height need to be incorporated to assess adequately the impact of both.7 As a second goal, we explored whether other congenital anomalies, in addition to NTDs, were associated with maternal height or prepregnant BMI. Only one other study1 has explored such risks for other anomalies. 235

Materials and methods

Details of the population-based casecontrol data used in this analysis have been described.810 The data come from two distinct casecontrol studies. One study,8 subsequently referred to as study 1, included 653 infants or fetuses with a NTD (anencephaly, spina bifida cystica, craniorachischisis or iniencephaly) that was ascertained by reviewing medical records, including ultrasonography, at all hospitals and clinics for those infants/fetuses delivered in selected California counties, and whose mothers gave their residence as California. Singleton liveborn infants and fetuses (including those prenatally diagnosed with a NTD and electively terminated between February 1989 and January 1991) among the cohort of 708 129 births and fetal deaths between June 1989 and May 1991 comprised the case study population. Controls were selected randomly from each area hospital in proportion to the hospital's estimated contribution to the total population of singleton infants born alive in a given month from June 1989 to May 1991. Thus, 644 singleton infants without a reportable congenital anomaly11 and whose mother was a California resident were selected. Women who only spoke languages other than English or Spanish, or who had had a previous NTD-affected pregnancy, were excluded, leaving 613 cases and 611 controls. In-person interviews were completed with mothers of 538 (87.8%) cases and 539 (88.2%) controls, an average of 4.9 months for cases and 4.6 months for controls after the actual or projected date of term delivery. This data set was the basis of an earlier report of an association between maternal BMI of 4 29 kg/m2 and increased NTD risk.2 The second casecontrol study (study 2) included NTDs (that were not included in the above-described

study), conotruncal heart defects, limb anomalies and orofacial clefts diagnosed within 1 year after birth among infants and fetal deaths delivered to women residing in most Californian counties.9,10 Included were all deliveries of infants or fetal deaths (5 20 weeks gestation) that occurred between January 1987 and December 1988 (n = 344 214), except for ascertainment of orofacial cleft cases, which included deliveries until December 1989 (an additional 208 387 infants or fetal deaths considered eligible). Case eligibility was determined by medical geneticists using detailed diagnostic information collected from medical records from all hospitals and genetics centres in the surveillance area. Eligible as NTD cases were infants or fetuses (including elective terminations) with diagnoses of anencephaly, spina bifida cystica, craniorachischisis and iniencephaly confirmed by autopsy, surgery report, ultrasound or X-ray scan. Eligible as conotruncal heart cases were all infants and fetuses with anomalies affecting aorticopulmonary septation, including tetralogy of Fallot, d-transposition of the great arteries, truncus arteriosus communis, doubleoutlet right ventricle, pulmonary valve atresia with ventricular septal defect, subaortic ventricular septal defect type I and aortico-pulmonary window, confirmed by echocardiography, cardiac catheterisation, surgery or autopsy. Eligible as limb anomaly cases were all infants and fetuses with longitudinal, transverse or amniotic band limb deficiency defects of the upper or lower limbs that were confirmed by radiology, surgery or autopsy reports. Eligible as orofacial cleft cases were those infants or fetuses with cleft palate without cleft lip (CP) and cleft lip with or without cleft palate (CLP) confirmed by surgical or autopsy report. These cases were further phenotypically classified as `isolated' CP, `isolated' CLP, `multiple' CP or `multiple' CLP based on the nature of accompanying congenital anomalies. CP and CLP cases with no other anomaly or with anomalies considered minor were classified as isolated. CP and CLP cases with at least one accompanying major anomaly were classified as multiple. Cases with monogenic conditions were excluded. Infants diagnosed with single gene disorders, trisomies, Turner syndrome (45,X) or NTD cases with amniotic band pathogenesis were also excluded. A total of 972 control infants (652 of whom were selected from the 198788 birth cohort) were randomly selected from all infants born alive (n = 548 844) in the same geographical area and time period (198789) as

# Blackwell Science Ltd. Paediatric and Perinatal Epidemiology 2000, 14, 234239

Paper 274 Disc

236 G. M. Shaw et al. overall in study 1, but not in study 2, with respect to height. These relations were explored further by dividing the NTD grouping into its two primary subphenotypes, anencephaly and spina bifida. Results for these analyses did not differ substantially from those that are displayed in Table 1, although we did observe the coefficients associated with BMI to be somewhat larger for spina bifida than for anencephaly. Coefficients associated with height tended to be larger for anencephaly than for spina bifida based on data from study 1, but not for data from study 2 (data not shown). NTD risk patterns associated with BMI and height were studied further controlling for maternal education and multivitamin use. Overall, NTD risk patterns were similar to unadjusted analyses (data not shown). Excluding women with any type of diabetes also did not substantially alter the observed results. Elevated risks associated with increasing BMI and decreasing height were not observed consistently for the other anomalies included in Table 1 (conotruncal heart defects, limb anomalies, isolated cleft lip/palate, isolated cleft palate, multiple cleft lip/palate or multiple cleft palate). Results pertaining to specific race/ ethnic groups also did not, in general, reveal strong positive associations for BMI or strong negative associations for height. Because only limited data are available in the literature on BMI as a risk factor for congenital anomalies other than NTDs, we include in Table 1 the number of case and control mothers who were obese (defined as a BMI 4 29 kg/m2). The odds ratios (ORs) associated with the non-NTD anomalies and obesity in this study were: conotruncal defects, OR = 1.0 [95% confidence interval, 0.6,1.8]; limb anomalies, OR = 0.8 [0.4,1.6]; isolated cleft lip/palate, OR = 1.0 [0.6,1.6]; isolated cleft palate, OR = 1.1 [0.6,2.0]; multiple cleft lip/palate, OR = 1.0 [0.5,2.1]; and multiple cleft palate, OR = 1.6 [0.8,3.4]. We also analysed the two largest subgroups of conotruncal defects, d-transposition of the great vessels and tetralogy of Fallot (data not shown). The ORs associated with maternal obesity for these two subgroups were 1.4 [0.7,3.1] and 0.5 [0.2,1.4] respectively.

cases. Control infants had no major congenital anomalies identified before the first birthday. Interviews were completed with 265 (84%) NTD case mothers, 207 (87%) conotruncal case mothers, 165 (82%) limb anomaly case mothers, 662 (85%) orofacial cleft case mothers and 734 (78%, [481 or 76% for the 198788 cohort]) control mothers. Interviews were completed an average of 3.7 years after the date of delivery for cases and 3.8 years for controls. In addition to information on maternal medical and reproductive history, and activities associated with various lifestyles, interviews in both casecontrol studies elicited information on maternal height and prepregnant weight. Women were asked: `In the month before you became pregnant, how much did you weigh (in light clothing and without shoes)?' and `How tall are you without shoes?' BMI was estimated for each woman from information on her reported weight and height using the algorithm,6 kg weight/(m height)2. The binary outcome of anomaly risk was described using additive linear logistic regression models. Models describing risks associated with maternal height and prepregnant BMI were constructed. Considered as covariates were maternal race/ethnicity (Latina, foreign born; Latina, US born; white, nonHispanic; black; other), education (5 high school graduate; high school graduate; college graduate) and use of multivitamins containing folic acid (in the 4month periconceptional period for study 2 or in the 3 months before conception in study 1, yes vs. no).

Results
Risks of anomaly, as measured by the estimated regression coefficients using a generalised logistic model, were described for the linear combination of maternal prepregnant BMI and height (Table 1). Focusing first on NTDs (both study 1 and study 2), the results revealed increasing NTD risk with increasing maternal prepregnant BMI, controlling for maternal height, and increasing risk for decreasing height controlling for maternal BMI. This pattern was observed overall as well as for most race/ethnic groups in both study populations, with one exception. Increasing NTD risk was observed for increasing height among foreign-born Latinas in study 2. The precision measured by the confidence intervals indicated that random variation was an unlikely explanation for the patterns observed for NTD risk

Discussion
Our analyses explored whether maternal prepregnant BMI and height influenced risk for several congenital anomalies when adjusted for the potential confounding

# Blackwell Science Ltd. Paediatric and Perinatal Epidemiology 2000, 14, 234239

# Blackwell Science Ltd. Paediatric and Perinatal Epidemiology 2000, 14, 234239

Table 1. Congenital anomaly risk relative to women's prepregnant body mass index (BMI) and height, stratified by maternal race/ethnic group No. of casesa Study 1 Neural tube defects Overall White non-Hispanic Foreignborn Latina US-born Latina Black Study 2 Neural tube defects Overall White non-Hispanic Foreign-born Latina US-born Latina Black Conotruncal defects Overall White non-Hispanic Foreignborn Latina US-born Latina Black Limb anomalies Overall White non-Hispanic Foreignborn Latina US-born Latina Black Isolated cleft lip/palate Overall White non-Hispanic Foreign-born Latina US-born Latina Black Isolated cleft palate Overall White non-Hispanic Foreignborn Latina US-born Latina Black No. of controlsa bb BMI (kg/m2) b Height (m) No. of case mothers 4 29 kg/m2 No. of control mothers 4 29 kg/m2

[95% CIc]

[95% CI]

500 215 157 73 25

517 269 97 80 30

0.53 0.59 0.79 0.40 0.34

[0.26, 0.79] [0.21, 0.98] [0.09, 1.49] [0.23, 1.02] [0.32, 1.01]

2.44 1.82 0.81 0.82 2.81

[4.08,0.80] [4.38, 0.74] [4.66, 3.03] [5.90, 4.25] [11.0, 5.39]

83 37 20 16 9

48 18 6 15 6

247 152 40 32 6 202 136 14 30 7 156 100 16 20 9 333 218 46 37 7 140 96 14 13 6

461 276 55 69 13 461 276 55 69 13 461 276 55 69 13 706 426 82 103 23 706 426 82 103 23

0.47 0.61 0.61 0.53 0.83 0.09 0.29 1.47 0.02 1.12 0.02 0.18 0.29 1.14 0.72 0.08 0.06 0.23 0.34 0.92 0.25 0.06 0.59 1.83 2.04

[0.14, 0.80] [0.16, 1.06] [0.24, 1.45] [0.31, 1.37] [1.78, 3.45] [0.31, [0.23, [3.36, [0.92, [3.97, [0.42, [0.39, [1.82, [2.63, [1.14, [0.40, [0.36, [1.21, [1.19, [3.72, [0.72, [0.64, [2.16, [3.73, [0.17, 0.49] 0.81] 0.42] 0.88] 1.74] 0.46] 0.75] 1.24] 0.34] 2.57] 0.25] 0.48] 0.76] 0.51] 1.88] 0.21] 0.53] 0.99] 0.07] 4.24]

0.34 1.87 4.47 0.23 1.89 0.69 0.89 7.69 3.37 7.90 1.77 0.06 8.75 1.67 4.36 0.35 0.04 1.87 3.03 1.39 1.14 1.41 4.40 6.50 4.84

[2.41, [4.82, [1.36, [6.18, [15.1, [1.49, [3.88, [3.24, [9.12, [20.0,

1.73] 1.07] 10.3] 5.71] 11.4] 2.87] 2.10] 18.6] 2.37] 4.23]

36 21 5 8 1 18 12 0 4 0 11 8 1 0 1 29 20 3 4 1 13 7 1 2 2

40 15 9 12 1 40 15 9 12 1

BMI and height as anomaly risk factors

[0.56,4.10] [3.32, 3.20] [0.97, 18.5] [8.66, 5.31] [4.62, 13.3] [1.39, [2.44, [3.60, [8.13, [12.3, 2.09] 2.36] 7.33] 2.07] 15.0]

40 15 9 12 1 61 26 11 18 1 61 26 11 18 1

Paper 274 Disc

[-3.53, 1.26] [4.63, 1.80] [12.0, 3.19] [15.4, 2.36] [9.11, 18.8]

237

Paper 274 Disc

238 G. M. Shaw et al.
No. of control mothers 4 29 kg/m2

Multiple cleft lip/palate Overall White non-Hispanic Foreignborn Latina US-born Latina Black Multiple cleft palate Overall White non-Hispanic Foreignborn Latina US-born Latina Black

Table 1. continued

# Blackwell Science Ltd. Paediatric and Perinatal Epidemiology 2000, 14, 234239

Race/ethnic groups may not add to overall because `other' race/ethnic groups and those with missing information were excluded. Regression coefficient. Coefficient for BMI has been multiplied by 10. c95% confidence interval. Interval for BMI has been multiplied by 10.

influence of the other. Our results pertaining to NTDs extend previous findings involving obesity defined by BMI as a risk factor for NTD. We observed both prepregnant BMI and maternal height to contribute to NTD risk. However, the relation with decreasing height was less evident in study 2. The reason for this discrepancy in results concerning height is unknown, but appeared to be associated with one race/ethnic group in study 2. Thus, the relation between decreasing maternal height and NTD risk should be interpreted with more caution than the relation between maternal prepregnant BMI and NTD risk. Maternal height,12 not controlled for weight, has been observed previously, as has increasing maternal weight,3 not controlled for height, to influence NTD risk. Further, our results suggest that decreasing height, increasing BMI or elevated BMI are not important risk factors for conotruncal heart defects, limb anomalies and most orofacial clefts. These results contribute substantially to a relatively small literature on risks for anomalies other than NTDs associated with maternal prepregnant height or weight. Waller et al.1 provided data on other congenital anomalies and observed an elevated risk (from elevated BMI) for a subphenotype (transposition of great vessels) of the heart anomalies studied here. We observed a modest (OR = 1.4) elevated risk for this subphenotype. For decades, BMI as an index of obesity has been used as a surrogate for body size and body composition to predict risk for a number of health outcomes. Generally, when BMI is used as a measure of body size, it is considered to be independent of height. This assumption has been demonstrated to be inaccurate.7 In the current data (e.g. study 2), the correlation between BMI and height was 0.22, indicating that there may be a substantial residual association with height in a model that considers risk from BMI only. Approaches have been suggested for removing this residual,7 such as including both weight and height in models, or including both BMI and height. We chose to explore risk based on the linear combination of BMI and height. Our results revealed evidence for a positive association between NTD risk and maternal prepregnant BMI and a negative association between NTD risk and maternal height. The mechanism underlying the association between maternal obesity and NTDaffected pregnancy risk is unknown. It may be the consequence of a metabolic abnormality associated with being obese or a risk factor more common among

61 26 11 18 1 8 4 0 2 1 1.51] 5.52] 7.89] 2.57] 9.37] [4.07, [3.47, [6.31, [14.2, [12.8, 0.46] 0.99] 0.46] 0.82] 2.76] 0.08 0.21 1.10 0.51 0.41 706 426 82 103 23 93 44 19 14 7 [0.61, [0.57, [2.65, [1.84, [1.94, 1.28 1.03 0.79 5.81 1.72

No. of case mothers 4 29 kg/m2

[95% CI]

b Height (m)

bb BMI (kg/m2)

No. of controlsa

No. of casesa

67 43 5 13 2

706 426 82 103 23

0.61 0.85 0.74 0.08 1.89

[0.08, 1.14] [0.18, 1.53] [1.52, 2.99] [1.38, 1.21] [2.58, 6.35]

[95% CIc]

0.13 0.20 2.52 5.70 15.9

[3.11, [4.42, [9.88, [14.2, [13.8,

3.37] 4.82] 14.9] 2.83] 45.5]

9 7 1 1 0

61 26 11 18 1

Paper 274 Disc

BMI and height as anomaly risk factors obese women. Our data suggest that risk increases as maternal prepregnant BMI increases and, therefore, NTD-affected pregnancy risk is not restricted only to those women who were considered to be obese based on the Institute of Medicine's definition of obesity.6 This association clearly needs to be investigated further to determine what concomitant factors may be contributing to this risk pattern. The mechanism underlying the potential inverse association between maternal height and risk for a NTD-affected pregnancy is also unknown. Some investigators have speculated that NTD risk may be associated with disturbances in earlier developmental and growth processes of the mother herself.13 Shorter maternal height may, indeed, reflect stunting from nutritional deprivation or other biological processes and, thus, provide some indirect support for this speculation. These analyses had several advantages, including large sample size, population-based ascertainment of cases and controls, reasonably high participation of eligible case and control mothers and information on relevant covariates. It is unlikely that our results associated with NTDs arise from differential reporting accuracy of height and weight by case and control mothers. When data were collected for the two NTD studies, shorter height and elevated weight had not been widely reported as risk factors for NTD-affected pregnancies. Moreover, the risk patterns observed for NTDs associated with maternal height and prepregnant BMI were not observed for the other congenital anomalies studied here, further suggesting that differential reporting accuracy between case and control mothers was unlikely. It is also unlikely that the lack of similar relations with other studied anomalies was caused by non-differential misclassification, given that self-reporting of past weight has been observed by others to be accurate when the recall period is 10 years or less.14 Despite the advantages and limitations of this study, other data sets will need to be explored to determine whether similar `linear' patterns of NTD risk or lack of risk for other anomalies are associated with the two anthropometric variables considered in this study. 239

gram, University of California, 1RT466 and 3RT0413. The opinions expressed are the views of the authors. They do not necessarily reflect the official position of the California Department of Health Services or other institutions involved.

References
1 Waller DK, Mills JL, Simpson JL, Cunningham GC, Conley MR, Lassman MR, et al. Are obese women at higher risk for producing malformed offspring? American Journal of Obstetrics and Gynecology 1994; 170:541548. 2 Shaw GM, Velie EM, Schaffer D. Risk of neural tube defectaffected pregnancies among obese women. JAMA 1996; 275:10931096. 3 Werler MM, Louik C, Shapiro S, Mitchell AA. Prepregnant weight in relation to risk of neural tube defects. JAMA 1996; 275:10891092. 4 Watkins ML, Scanlon KS, Mulinare J, Khoury MJ. Is maternal obesity a risk factor for anencephaly and spina bifida? Epidemiology 1996; 7:507512. 5 Kallen K. Maternal smoking, body mass index, and neural tube defects. American Journal of Epidemiology 1998; 147:11031111. 6 Institute of Medicine National Academy of Sciences. Nutrition During Pregnancy. Washington, DC: National Academy Press, 1990; p 5. 7 Michels KB, Greenland S, Rosner BA. Does body mass index adequately capture the relation of body composition and body size to health outcomes? American Journal of Epidemiology 1998; 147:167172. 8 Shaw GM, Schaffer D, Velie EM, Morland K, Harris JA. Periconceptional vitamin use, dietary folate, and the occurrence of neural tube defects. Epidemiology 1995; 6:219226. 9 Shaw GM, Wasserman CR, Lammer EJ, O'Malley CD, Murray JC, Basart AM, et al. Orofacial clefts, parental cigarette smoking, and transforming growth factor-alpha gene variants. American Journal of Human Genetics 1996; 58:551561. 10 Wasserman CR, Shaw GM, O'Malley CD, Tolarova MM, Lammer EJ. Parental cigarette smoking and risk for congenital anomalies of the heart, neural tube, or limb. Teratology 1996; 53:261267. 11 Croen LA, Shaw GM, Jensvold NJ, Harris JA. Birth defects monitoring in California: a resource for epidemiological research. Paediatric and Perinatal Epidemiology 1991; 5:423427. 12 Anderson WJR, Baird D, Thomson AM. Epidemiology of stillbirths and infant deaths due to congenital malformation. Lancet 1958; 1:13051306. 13 Emanuel I, Sever L. Questions concerning the possible association of potatoes and neural-tube defects, and an alternative hypothesis relating to maternal growth and development. Teratology 1973; 8: 325332. 14 Perry GS, Byers TE, Mokdad AH, Serdula MK, Williamson DF. The validity of self-reports of past body weights by US adults. Epidemiology 1995; 6:6166.

Acknowledgements
The authors are grateful to Jennifer Stiling for manuscript preparation. This work was partially supported by the Cigarette and Tobacco Surtax Fund of California Tobacco-Related Disease Research Pro-

# Blackwell Science Ltd. Paediatric and Perinatal Epidemiology 2000, 14, 234239