spina bifida phenotypes in infants or fetuses of obese mothers

Spina Bifida Phenotypes in Infants orFetuses of Obese MothersGARY M. SHAW,1* KAREN TODOROFF,1 RICHARD H. FINNELL,2 AND

EDWARD J. LAMMER3

1March of Dimes Birth Defects Foundation, California Birth Defects Monitoring Program, Emeryville,California 946082Center for Human Molecular Genetics. Nebraska Medical Center, Omaha, Nebraska 681983Division of Medical Genetics, Children’s Hospital, Oakland, California 94609

ABSTRACT

Background: A twofold or greater risk of neural tubedefect (NTD)-affected pregnancy has been associatedwith prepregnant obesity, where obesity was definedas body mass index (BMI) of .29 kg/m2. Risks havebeen more elevated for spina bifida than for anenceph-aly.Methods: We investigated whether finer phenotypicclassifications of spina bifida, in combination with otherfactors, were associated with a BMI of .29 kg/m2.Data were derived from a case-control study of fetusesand infants with NTDs among 1989–1991 Californiabirths. Interviews were conducted with mothers of 277spina bifida cases and 517 nonmalformed controls.Results: Women with a BMI of .29 kg/m2 comparedwith those #29 kg/m2 revealed an odds ratio (OR) of2.2 (95% confidence interval [95% CI] 5 1.4–3.3) forspina bifida in their infants and fetuses. Elevated riskswere observed for each spina bifida subphenotype,and risks varied by subphenotype: open spina bifida,OR 5 2.0 (1.2–3.1); closed (skin-covered), 3.3 (1.4–7.5); isolated, 2.2 (1.4–3.4); nonisolated, 1.9 (0.9–4.2); high, 4.5 (2.1–9.6); low, 1.9 (1.2–2.9); open/isolated/high, 7.1 (2.8–18.1); and open/isolated/low,1.8 (1.1–3.1). Risks were higher among female in-fants/fetuses and foreign-born Latinas, and for somephenotypes the risks were quite large, e.g., OR 5 8.3(2.9–23.6) for “closed” spina bifida among femaleinfants/fetuses whose mothers were .29 kg/m2 com-pared with males whose mothers were #29 kg/m2.Maternal periconceptional vitamin use was not ob-served to influence risk as greatly across phenotypes.Conclusions: The observed pathogenetic heterogene-ity of prepregnant obesity and spina bifida risks sug-gests that there are likely to be several biologic mech-anisms underlying the association.Teratology 61:376–381, 2000. © 2000 Wiley-Liss, Inc.

INTRODUCTION

Five recent studies have observed an increased riskof approximately twofold or greater for neural tubedefects (NTDs) associated with maternal obesity before

pregnancy (Waller et al., ’94; Shaw et al., ’96; Watkinset al., ’96; Werler et al., ’96; Kallen, ’98). Four of thesestudies defined obesity consistent with that establishedby the Institute of Medicine (’90), i.e., of a body massindex (BMI) of .29 kg weight/m height2 (Waller et al.,’94; Shaw et al., ’96; Watkins et al., ’96; Kallen, ’98).Although these studies explored a variety of covariates,no additional maternal factor, including maternalpericonceptional vitamin use, appeared to influencesubstantially the observed association between mater-nal obesity and increased risk of NTDs. The risk didappear to be more elevated for spina bifida than foranencephaly, and in two studies (Waller et al., ’94;Shaw et al., ’96) the risk was greater than threefold forinfants to have lipomyelomeningocele among womenwith an elevated BMI.

In analyzing one of the datasets (Shaw et al., ’96)that previously revealed an association between mater-nal prepregnant obesity and spina bifida, our goal inthis analysis was to determine whether maternal obe-sity (defined as a BMI .29 kg/m2) was associated withany specific phenotype of spina bifida. We also exploredwhether risks by specific phenotypes were further in-fluenced by other factors including infant sex, maternalrace/ethnicity, or maternal periconceptional use ofmultivitamin supplements.

MATERIALS AND METHODS

The data used for this analysis were described else-where (Shaw et al., ’95, ’96). Diagnostic informationfrom medical records, autopsies, and surgical reports ofall infants and fetuses with NTDs or similar centralnervous system (CNS) malformations was reviewed byone of us (E.J.L.) to restrict eligibility to those withanencephaly, spina bifida cystica, craniorachischisis,or iniencephaly. Eligible cases were singleton fetuses,including those electively terminated between Febru-ary 1989 and January 1991, and liveborn infants diag-

*Correspondence to: Gary M. Shaw, March of Dimes Birth DefectsFoundation, California Birth Defects Monitoring Program, 1900 Pow-ell St. Suite 1050, Emeryville, CA 94608.

Received 16 September 1999; Accepted 18 December 1999

TERATOLOGY 61:376–381 (2000)

© 2000 WILEY-LISS, INC.

nosed with a NTD among the cohort of 708,129 Cali-fornia births (includes fetal deaths) between June 1989and May 1991. Ascertained were 653 singleton infantsor fetuses with a NTD. Excluded were women who onlyspoke languages other than English or Spanish, as wellas 11 women who had a previous NTD-affected preg-nancy. Controls were randomly selected from each areahospital in proportion to the hospital’s estimated con-tribution to the total population of singleton liveborninfants in a given month from June 1989 to May 1991.Selected were 611 singleton infants without a report-able congenital anomaly, whose mother was a Califor-nia resident, spoke either English or Spanish, and didnot have a previous NTD-affected pregnancy. In-per-son interviews were completed with mothers of 538(87.8%) cases and of 539 (88.2%) controls an average of4.9 months for cases and 4.6 months for controls afterthe actual or projected date of term delivery. The cur-rent analyses were restricted to the 296 case motherswho delivered an infant or fetus with spina bifida, 277of whom provided information on height (m) andprepregnant weight (kg). 517 control mothers providedheight and weight information enabling the calculationof BMI for 794 case and control mothers.

All cases were further classified by a medical genet-icist (E.J.L), initially as either isolated or nonisolatedbased on the presence and nature of accompanyingcongenital anomalies. An infant or fetus with spinabifida and an accompanying major anomaly that wasnot a dysplasia, deformation, disruption, or a conse-quence of the NTD was considered nonisolated; theremainder were classified as isolated. Cases were fur-ther subgrouped on the basis of the anatomic locationof the spinal defect and on whether the defect was openor covered with full-thickness skin. Medical records,radiographic or magnetic resonance imaging (MRI) re-ports, surgical notes, and autopsy findings were used tomake these distinctions. “High” spina bifida defectswere those in which the upper bony extent of defectwas determined to be above the 12th thoracic vertebra.“Low” spina bifida defects were those that were at orbelow the 12th thoracic vertebra. Spinal defects wereconsidered “closed” only if full-thickness skin waspresent over the entire area of the defect and wereotherwise considered “open.” Of the 32 infants/fetusesconsidered to have closed spinal defects, 12 were deter-mined (n 5 10) or suspected (n 5 2) to be lipomyelo-meningoceles.

Effects were estimated by odds ratios (OR) and their95% confidence intervals (95% CI). Effects were esti-mated for women whose prepregnant BMI was .29kg/m2 (defined as obesity by the Institute of Medicine,’90) compared with those whose BMI was #29 kg/m2.These analyses were further stratified by infant sex,maternal race/ethnicity (non-Hispanic whites, U.S.-born Latina, foreign-born Latina, and black; women of“other” groups were excluded, 11 case and 3 controlmothers), and periconceptional use of multivitaminscontaining folic acid (women who began using vitamins

in the period 3 months before conception versus womenwho began after conception or did not use vitamins atall during pregnancy).

RESULTS

Women whose prepregnant BMI was .29 kg/m2 re-vealed an OR 5 2.2 (95% CI 5 1.4–3.3) for spina bifidain their offspring compared with women whose BMIwas #29 kg/m2. An elevated risk was observed for eachspina bifida subphenotype, but the magnitude of riskvaried somewhat by subphenotype: open spina bifida,OR 5 2.0 (1.2–3.1); closed, 3.3 (1.4–7.5); isolated, 2.2(1.4–3.4); nonisolated, 1.9 (0.9–4.2); high, 4.5 (2.1–9.6);low, 1.9 (1.2–2.9); open/isolated/high, 7.1 (2.8–18.1);and open/isolated/low, 1.8 (1.1–3.1).

We investigated whether additional factors in com-bination with elevated BMI influenced risks of spinabifida phenotypes. The data presented in Table 1 showthat female infants or fetuses whose mothers were .29kg/m2 were at elevated risk, compared with male in-fants or fetuses whose mothers were #29 kg/m2, par-ticularly for closed, high, or open/isolated/high pheno-types. In general, risks were also elevated, but to alesser extent, for female infants whose mothers werenot .29 kg/m2, and for males whose mothers were .29kg/m2.

Comparisons were also made for maternal race/eth-nic backgrounds (Table 2). Using as referents womenwho were non-Hispanic white and #29 kg/m2, women.29 kg/m2 in most race/ethnic groups were at elevatedrisks of delivering infants or fetuses with most spinabifida subphenotypes. Effect estimates were generallyhighest among foreign-born Latinas whose BMI was.29 kg/m2.

Because maternal diabetes has been associated withNTD risk, the infant sex and race/ethnic group analy-ses were repeated restricting analyses to those womenwithout diabetes. This restriction did not substantiallyalter observed results.

We further explored the elevated risks among fe-males and foreign-born Latinas. The OR associatedwith overall spina bifida (irrespective of subphenotype)and BMI .29 kg/m2 for child sex 5 female and mater-nal race/ethnicity 5 foreign-born Latina was 6.5 (1.9–21.9), relative to child sex 5 male, race/ethnicity 5non-Hispanic white, and BMI of #29kg/m2. Effects ofthe same magnitude or higher, however, were observedfor females whose mothers were non-Hispanic whitewith BMI of .29 kg/m2 and for males whose motherswere foreign-born Latina with BMI of .29 kg/m2 (datanot shown). Sample sizes were insufficient to ade-quately estimate effects for the combination of childsex, maternal race/ethnicity, and maternal BMI .29kg/m2 on risks of subphenotypes.

We also explored whether periconceptional use ornonuse of vitamins containing folic acid altered BMIassociations by spina bifida phenotypes. We observedelevated risks of all subphenotypes in infants/fetuses

SPINA BIFIDA PHENOTYPES IN INFANTS/FETUSES OF OBESE MOTHERS 377

delivered to women .29 kg/m2 irrespective of whetherthey used or did not use vitamins periconceptionally(data not shown).

DISCUSSIONThese data extend previous observations regarding

prepregnant obesity in women and their risks of havingNTD-affected pregnancies. The nature of the biologicmechanism underlying this risk is not obvious. It istempting to speculate that the embryonic environmentis metabolically different for obese women. Obesity

might affect nutritional and metabolic aspects of theembryoplacental unit. For example, perhaps glucosemetabolism or placental hemodynamics are altered.Such metabolic differences might play a role in increas-ing risk for NTDs. However, depending on the specificbehavioral and genetic factors that contribute to obe-sity, the metabolic changes experienced by the fetusmay substantially vary between different obesewomen.

In this analysis, we explored potential factors thatmight contribute to the observation that risk is more

TABLE 1. Effect estimates (odds ratios) for spina bifida phenotypes by maternal prepregnant body massindex category and infant sex

NTD phenotypeMaternal BMI

category (kg/m2)Infant

sex OR 95% CI Cases Controls

All NTDs.29 F 3.0 1.7–5.3 47 23#29 F 1.2 0.9–1.6 184 224.29 M 1.6 0.9–3.0 28 25#29 M Reference — 168 245

Spina bifida.29 F 3.5 1.9–6.2 31 23#29 F 1.3 0.9–1.8 114 224.29 M 2.0 1.0–3.7 19 25#29 M Reference — 95 245

Open.29 F 2.8 1.5–5.3 20 23#29 F 1.3 0.9–1.8 90 224.29 M 1.9 1.0–3.8 15 25#29 M Reference — 76 245

Closed.29 F 8.3 2.9–23.6 7 23#29 F 1.8 0.8–4.2 15 224.29 M 1.1 0.5–9.0 1 25#29 M Reference — 9 245

Isolated.29 F 3.6 1.9–6.6 26 23#29 F 1.3 0.9–1.9 93 224.29 M 2.0 1.0–4.0 16 25#29 M Reference — 77 245

Nonisolated.29 F 3.0 1.0–8.4 5 23#29 F 1.3 0.7–2.4 21 224.29 M 1.6 0.5–5.6 3 25#29 M Reference — 18 245

High.29 F 6.4 2.2–18.6 6 23#29 F 1.5 0.7–3.4 14 224.29 M 4.9 1.6–14.9 5 25#29 M Reference — 10 245

Low.29 F 3.2 1.8–6.0 25 23#29 F 1.3 0.9–1.8 96 224.29 M 1.4 0.7–2.9 12 25#29 M Reference — 82 245

Open/isolated/high.29 F 14.2 3.3–60.4 4 23#29 F 2.9 0.8–10.2 8 224.29 M 13.1 3.1–55.3 4 25#29 M Reference — 3 245

Open/isolated/low.29 F 2.8 1.4–5.7 15 23#29 F 1.3 0.9–1.9 67 224.29 M 1.7 0.8–3.7 10 25#29 M Reference — 57 245

BMI, body mass index; OR, odds ratio; CI, confidence interval; NTD, neural tube defect; Reference, group designated forcomparison to compute OR.

378 SHAW ET AL.

TABLE 2. Effect estimates (odds ratios) for spina bifida phenotypes by maternal prepregnant body massindex category and maternal race/ethnicity


(kg/m2) Race/ethnicity OR 95% CI Cases Controls

All NTDs.29 White 2.9 1.6–5.2 37 18#29 White Reference — 178 251.29 US-born Latina 1.5 0.7–3.1 16 15#29 US-born Latina 1.2 0.8–1.9 57 65.29 Foreign-born Latina 4.7 1.9–11.6 20 6#29 Foreign-born Latina 2.1 1.5–2.9 137 91.29 Black 2.1 0.8–5.8 9 6#29 Black 0.9 0.5–1.8 16 24

Spina bifida.29 White 2.2 1.1–4.3 17 18#29 White Reference — 110 251.29 US-born Latina 1.8 0.8–4.0 12 15#29 US-born Latina 1.1 0.6–1.7 30 65.29 Foreign-born Latina 5.3 2.1–13.8 14 6#29 Foreign-born Latina 1.6 1.1–2.4 65 91.29 Black 2.3 0.8–6.9 6 6#29 Black 1.0 0.4–2.0 10 24

Open.29 White 2.6 1.2–5.2 15 18#29 White Reference — 82 251.29 US-born Latina 1.6 0.7–3.9 8 15#29 US-born Latina 1.0 0.6–1.8 22 65.29 Foreign-born Latina 5.1 1.9–13.9 10 6#29 Foreign-born Latina 1.7 1.1–2.7 52 91.29 Black 1.0 0–4.5 2 6#29 Black 1.3 0.6–2.7 10 24

Closed.29 White 1.4 0–9.1 1 18#29 White Reference — 10 251.29 US-born Latina 5.0 1.4–19.0 3 15#29 US-born Latina 1.9 0.7–5.6 5 65.29 Foreign-born Latina 8.4 0–42.0 2 6#29 Foreign-born Latina 2.5 1.0–6.1 9 91.29 Black 4.2 0–30.0 1 6#29 Black — — 0 24

Isolated.29 White 2.7 1.4–5.5 17 18#29 White Reference — 87 251.29 US-born Latina 1.7 0.7–4.0 9 15#29 US-born Latina 1.1 0.7–1.9 25 65.29 Foreign-born Latina 5.3 2.0–14.2 11 6#29 Foreign-born Latina 1.7 1.1–2.5 53 91.29 Black 1.9 0.6–6.5 4 6#29 Black 1.2 0.6–2.6 10 24

Nonisolated.29 White — 0 18#29 White Reference — 23 251.29 US-born Latina 2.2 0.6–7.6 3 15#29 US-born Latina 0.8 0.3–2.2 5 65.29 Foreign-born Latina 5.5 1.4–21.5 3 6#29 Foreign-born Latina 1.4 0.7–3.0 12 91.29 Black 3.6 0–16.9 2 6#29 Black — — 0 24

High.29 White 5.4 1.8–16.2 5 18#29 White Reference — 13 251.29 US-born Latina 2.6 0–11.3 2 15#29 US-born Latina 1.2 0.4–3.6 4 65.29 Foreign-born Latina 9.7 2.4–40.0 3 6#29 Foreign-born Latina 1.5 0.6–3.7 7 91.29 Black 3.2 0–22.4 1 6.29 Black — — 0 24

(Continued)


elevated for spina bifida relative to anencephaly (Shawet al., ’96). No particular spina bifida phenotype alonewas observed to explain the elevated risk overall.Risks, however, were higher among female infants/fetuses and among foreign-born Latinas. For some ofthese phenotypes, the risks were quite large, e.g., OR 58.3 for closed spina bifida among female infants/fetuseswhose mothers were .29 kg/m2 compared with maleswhose mothers were #29 kg/m2. These risks remainedelevated irrespective of maternal periconceptional useof vitamin supplements, an important factor for NTDsin general. Despite the magnitude of some observedrisks, however, one cannot rule out the potential role ofchance, given the number of comparisons made and thelack of knowledge for which observations should be themost biologically meaningful.

Nevertheless, we investigated whether particularspina bifida phenotypes were associated with maternalobesity because specificity for a particular phenotypemight provide clues about the pathogenesis underlyingthe association. If the risk of “closed” spina bifida waselevated, for example, then the underlying mechanismmight be similar to that of caudal dysgenesis associ-ated with infants delivered of diabetic women (i.e.,women who may also be more likely to be obese). Weobserved such a specific increased risk (OR 5 8.3), butour finding was limited to female infants/fetuses. Bycontrast, an elevated risk of open spina bifida wouldsuggest a mechanism acting upon the neurulation pro-cess of neural tube closure, elevated risk of open defects

at one location of the spinal column would suggest amore specific mechanism acting selectively on one ofthe posited five closure sites of neurulation (Van Allenet al., ’93). What we observed were elevated effect es-timates associated with obesity for essentially everytype of spina bifida, with the effects among females alsohigher for nearly every subphenotype. These findingssuggest a heterogeneous pathogenesis contributing toincreased risk for both abnormal neurulation and cau-dal dysgenesis. Similarly, associations observed forboth isolated and nonisolated spina bifida suggest thatat least some of the obesity-related pathogenetic pro-cesses are not unique to neural tube formation, butadversely affect the embryogenesis of other tissues aswell. Two observations associated with unusually highrisks may be the strongest evidence for some specificityof pathogenesis: (1) an OR of 8.3 for closed spina bifidaamong females; and (2) an OR of 14.2 for open/high/isolated spina bifida among females. Thus, our resultsshow both specificity and nonspecificity indicating thatthere may be several pathogenetic mechanisms in ac-tion.

The underlying reason for the elevated risks ob-served for females or for foreign-born Latinas is un-clear. Among female infants elevated risks of spinabifida are uncommon, whereas elevated risks of anen-cephaly are more characteristic (Elwood et al., ’92).Elevated NTD risks for foreign-born Latinas have beenobserved previously (Shaw et al., ’97). The risks ob-served here, however, suggest that foreign-born Lati-

TABLE 2. Effect estimates (odds ratios) for spina bifida phenotypes by maternal prepregnant body massindex category and maternal race/ethnicity (continued)


(kg/m2) Race/ethnicity OR 95% CI Cases Controls

Low.29 White 1.6 0.8–3.5 11 18#29 White Reference — 94 251.29 US-born Latina 1.8 0.8–4.0 10 15#29 US-born Latina 1.1 0.6–1.8 26 65.29 Foreign-born Latina 4.5 1.6–12.1 10 6#29 Foreign-born Latina 1.6 1.1–2.4 55 91.29 Black 2.2 0.7–7.0 5 6#29 Black 1.1 0.5–2.4 10 24

Open/isolated/high.29 White 11.6 3.4–39.8 5 18#29 White Reference — 6 251.29 US-born Latina — 0 15#29 US-born Latina — 0 65.29 Foreign-born Latina 20.9 4.7–97.4 3 6#29 Foreign-born Latina 2.3 0.7–7.0 5 91.29 Black — 0 6#29 Black — 0 24

Open/isolated/low.29 White 2.3 1.0–5.1 10 18#29 White Reference — 61 251.29 US-born Latina 2.2 0.9–5.3 8 15#29 US-born Latina 1.1 0.6–2.0 17 65.29 Foreign-born Latina 3.4 1.1–11.0 5 6#29 Foreign-born Latina 1.6 1.0–2.6 36 91.29 Black 1.4 0–6.1 2 6#29 Black 1.7 0.8–3.7 10 24

BMI, body mass index; OR, odds ratio; CI, confidence interval; NTD, neural tube defect; Reference, group designated forcomparison to compute OR.

380 SHAW ET AL.

nas who are obese before pregnancy (relative to non-Latinas who are not obese) are at substantially ele-vated risk of spina bifida. The apparent pathogeneticheterogeneity of prepregnant obesity and spina bifidasuggests that there are likely to be several biologicmechanisms underlying the association betweenprepregnant obesity and spina bifida risk in infants.

LITERATURE CITED

Elwood JM, Little J, Elwood JH. 1992. Epidemiology and control ofneural tube defects. Oxford: Oxford University Press.

Institute of Medicine, National Academy of Sciences. 1990. Nutri-tion during pregnancy. Washington, DC: National AcademyPress. p 5.

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

Shaw GM, Schaffer D, Velie EM, Morland K, Harris JA. 1995.Periconceptional vitamin use, dietary folate, and the occurrence ofneural tube defects. Epidemiology 6:219–226.

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

Shaw GM, Velie EM, Wasserman CR. 1997. Risk of neural tubedefect-affected pregnancies among women of Mexican descent andwhite women in California. Am J Public Health 87:1467–1471.

Van Allen MI, Kalousek DK, Chernoff GF, Juriloff D, Harris M,McGillivray BC, Yong S-L, Langlois S, MacLeod PM, Chitayat D,Friedman JM, Wilson RD, McFadden D, Pantzar J, Ritchie S, HallJG. 1993. Evidence for multi-site closure of the neural tube inhumans. Am J Med Genet 47:723–743.

Waller DK, Mills JL, Simpson JL, Cunningham GC, Conley MR,Lassman MR, Rhoads GG. 1994. Are obese women at higher riskfor producing malformed offspring? Am J Obstet Gynecol 170:541–548.

Watkins ML, Scanlon KS, Mulinare J, Khoury MJ. 1996. Is maternalobesity a risk factor for anencephaly and spina bifida? Epidemiology7:507–512.

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


spina bifida phenotypes in infants or fetuses of obese mothers

Documents