The teratogenic effect of maternal psychologic stress on reproductive outcomes is a subject of persistent if uneven scrutiny.1 Studies dating back to the 1960s1 have linked psychologic stress to various reproductive outcomes, including preterm delivery,2 low birth weight,3 and congenital malformations,4–10 including oral clefts6,7,9,10 and neural tube defects.4,6 In studies of congenital malformations, maternal stress was based either on a self-appraisal of “psychologic or emotional stress” or on having experienced “stressors”—life events, such as a family death, job loss, accident, or divorce. Alternative avenues of research focus on whether exogenous exposures to steroid hormones, normally produced in response to stress, are harmful to the developing fetus.11–13 In animal models, glucocorticoids are well-known teratogens, inducing cleft palate in mice.14–16
If stress can elevate the risk of an adverse health outcome, then strong emotional support may diminish this risk by “buffering” the effects of stress on health.17 That is, a strong social or emotional support system may strengthen a person’s adaptive capacities for dealing with stress.18 Numerous studies have described a relation between measures of social support and health outcomes, such as death, symptoms, and chronic diseases,19 but none has examined the analogous effect of maternal social support on reproductive risk for congenital malformations.
The etiology of neural tube defects (NTDs) among Mexican Americans living in Texas counties bordering on Mexico has been a critical issue ever since the early 1990s, when a high prevalence of NTDs was identified in this population.20 The population is poor, the consumption of folic acid-containing vitamins is relatively rare, and obesity and diabetes (two strong risk factors for NTDs) are common. Many believe that the long-observed social class gradient in NTD prevalence is tied to nutrition (namely folic acid),21–23 and we presume that inadequate intake of folic acid contributes substantially to the high prevalence of NTDs in the Texas–Mexico border population.24 However, we also conjecture that this impoverished population with poor nutritional status would be subject to more chronic stressors than other populations, which might, in turn, exacerbate risk. Therefore, we pursued the complementary questions of whether maternal stress increases and social support reduces NTD risk in this population along the Texas–Mexico border.
We identified study subjects through the Texas Department of Health’s Neural Tube Defect Project conducted along the Texas–Mexico border. The project included multisource active surveillance and a case–control study. Surveillance involved prospective case finding through the 21 hospitals, 39 birthing centers, 4 genetic clinics, 74 ultrasound centers, 4 licensed abortion centers, and approximately 150 midwives in the region. Cases were any infant or fetus with a diagnosis of anencephalus (ICD-9-CM code 740), spina bifida (741), or encephalocele (742.0) identified at birth or prenatally between March 1995 and May 2000. Of cases identified through surveillance, 53% were liveborn; 13%, stillborn; 31%, electively aborted; and 2%, spontaneously aborted. Recruitment into the case–control study was restricted to mothers who self-identified as Mexican American, were US residents, and had delivered in one of the 14 Texas counties bordering on Mexico. We identified control mothers from women delivering normal live births, that is, births without an apparent congenital malformation. Controls were randomly selected annually in proportion to the number of live births that occurred in hospitals and midwife-attended birthing centers.
The Texas Department of Health Institutional Review Board for the protection of human subjects approved the study protocol, English–Spanish consent forms, English–Spanish interview instruments, and data-collection procedures. In cooperation with hospital staff, field teams contacted women at the time of delivery or termination of pregnancy to inform them of the study and obtain consent. Women were interviewed in person about 5 to 6 weeks after their pregnancy ended. Before each interview, the staff obstetrician and interviewer estimated the date of conception for the index pregnancy using all gestational-age estimates from the medical record. We modeled the interview instrument after the 1993 Centers for Disease Control and Prevention mother questionnaire for birth defects risk factor surveillance. This included questions on maternal health history, demographics, use of medications and nutritional supplements, tobacco and alcohol use, and environmental and occupational exposures. We calculated body mass index (kg/m2) from self-reported prepregnancy height and weight. Dietary intake was assessed with a 98-item food frequency questionnaire specifically developed for this Texas–Mexico border population. For each food item, women estimated their usual intake frequency (average number of times per month, week or day) during the periconceptional period. The average daily dietary folate intake was computed with a food frequency software program (Food Frequency Data Entry and Analysis Program, University of Texas Health Science Center). Women were paid $20 for the 2-hour interview.
To measure maternal life-event stress, we used the detailed residential and occupational histories for the year before conception, as well as responses to the questions about any accidents or injuries experienced during the periconceptional period. Injury information included the type of injury, circumstances, and month of occurrence. Reported injuries during the 3 months before conception included those resulting from car accidents, falls, and physical assaults. We calculated a life-event score for each woman by summing her number of residence moves, job changes, and major injuries.
To measure social support, we asked women about the number of friends and the number of relatives they “can talk to about private matters or can call on for help.” The separate responses on the number of friends and relatives were summed to create a network size variable. We measured social interactions by asking, “how often do you participate in or attend group activities” and by asking about attendance at “religious worship services during the last year.” To measure emotional support we used seven questions on the degree of emotional ties to family and friends, taken from a modified version of the Duke Social Support Index.25 The scale included items such as “do family and friends understand you?” and “do you feel listened to?” Women with five or fewer positive responses to the seven items were categorized as having low emotional support (11%), women with six positive responses were categorized as having medium emotional support (15%), and women with seven positive responses were categorized as having high emotional support (74%). Network satisfaction was based on the question “how satisfied are you with relationships with family and friends?”
Of the 225 case mothers and 378 control mothers identified for study, 184 (82%) case mothers and 225 (60%) control mothers completed interviews. We did not obtain information on 26 (12%) case mothers and 101 (27%) control mothers who refused to be interviewed or on 15 (7%) case mothers and 52 (14%) control mothers who had moved out of the study area. We have previously noted that the control mothers mirrored the demographic characteristics of all border Mexican-American women who gave birth during the study years, minimizing the likelihood that controls were not representative of the border population.24
To estimate possible effects on the risk of NTD from maternal stress and social support, we calculated crude odds ratios (ORs) and 95% confidence intervals (CIs) for measures of network size, emotional support, satisfaction, and life events. We considered maternal age in years (<20, 20–24, 25–29, and ≥30), maternal birthplace (Mexico or US), years of education (0–6, 7–11, ≥12), annual household income (<$10,000, $10,000–$14,999, $15,000–$24,999, and ≥$25,000), dietary folate intake (quintiles based on the distribution of control mothers), folic acid-containing vitamin supplementation during preconception (yes or no), preconceptional cigarette smoking (nonsmoker, less than half pack per day, and half pack or more per day), preconceptional alcohol consumption (yes or no), and obesity (body mass index <30 kg/m2 or ≥30 kg/m2) as potential modifying or confounding variables. We obtained adjusted risk effect estimates from logistic regression models.
Table 1 shows the demographic characteristics of case and control mothers. Similar to the border population as a whole, about half of study subjects were born in Mexico, a substantial proportion had annual household incomes of less than $10,000, and only half had completed 12 years of education. Preconceptional use of folic acid containing vitamins was negligible for both case (2%) and control mothers (4%).
Each life-event stressor (residence moves, job changes, or major injury) was related to NTD risk (Table 2). The life-event score was strongly related to risk. Mothers experiencing one event before conception had 2.7 times the NTD risk of mothers with no such life events, whereas mothers with two or more events had 8.3 times the risk.
Table 3 shows effect estimates for categories of social network size (number of friends and relatives), social interaction (group interactions and church attendance), and social support (emotional support and network satisfaction). Other than emotional support, none of the measures was strongly related to risk. Estimated associations were weak with broad confidence intervals. Mothers who scored low on the emotional support scale had much higher risks of NTD than those who scored high (OR = 4.6; CI = 2.2–9.7).
Stratifying risk-effect estimates by selected covariates showed that effects of emotional support and life-event stress were consistent across demographic and other covariate categories (data not shown). Moreover, the effect of stressful life events (OR for one or more life events = 2.9; CI = 1.8–4.7) was neither confounded by emotional support (adjusted OR = 2.8; CI = 1.8–4.5) nor modified by such support (OR among medium/low support group = 3.9; OR among high support group = 2.5). As the stress and social support measures were likely to be correlated, we entered all variables (network size, group interactions, emotional support, satisfaction, and life events) into a logistic regression to examine their independent effects on NTD risk. As in the crude analysis, network size, group interactions, and satisfaction showed little relation with NTD risk when adjusted for the other measures. When analysis of the life-event score was adjusted for age, education, country of birth, income, obesity, vitamin supplementation, dietary folate intake, cigarette smoking, and alcohol consumption, the odds ratios for one (OR = 2.7; CI = 1.6–4.5) and two or more (OR = 5.0; CI = 1.0–26) life events were little changed from the crude estimates. Similar adjustment also had little effect on the results for medium (adjusted OR = 1.6; CI = 0.9–2.9) and low (OR = 4.8; CI = 2.1–11) emotional support.
In this Mexican-American border population, we found a strong association between the occurrence of stressful life events and the risk of NTDs. Although low emotional support was independently associated with NTD risk, we found no indication that strong emotional support modified or “buffered” the impact of stressful life events on risk. Two previous studies of maternal stress and NTDs have also shown effects from stress. A 1969 study from South Wales reported that the frequency of “accidents in pregnancy” during the first trimester was twice as high in mothers of offspring with anencephaly than in mothers of normal offspring.4 A recent case–control study in California reported a modest association between NTD risk and having experienced a close death, divorce, or job loss during the periconceptional period (OR = 1.5; CI = 1.1–2.1).6 Risk effects among these California women were slightly more pronounced among the lowest educated women (OR = 2.6) than among the more educated (OR = 1.2), and among foreign-born Hispanics (1.4) than among US-born women (1.1). These findings suggest that poor populations might be particularly susceptible to the damaging effects of life stressors.
Biased recall or reporting is a concern because our measures of stress and social support were constructed from self-reported responses solicited after the NTD occurrence. The experience of an NTD-affected pregnancy has the potential for influencing responses to queries about emotional support. The mothers’ questionnaire was administered an average of 5 to 6 weeks after the pregnancy ended and referenced the period surrounding conception. Mothers with affected babies may be more apt to perceive low emotional support subsequent to experiencing an NTD-affected pregnancy, as it is likely that the limits of this support would have been severely tested. However, reporting bias would not readily explain the large effect we observed for stressful life events. We obtained detailed medical, residential, and occupational histories from both case and control mothers. This included the types of injuries sustained and the periconceptional month of occurrence, the dates and address locations of residences before conception, and company names, job titles, and dates of employment. Although we did not verify self-reported responses pertaining to life event occurrences, the detailed probing on these events reduces the likelihood of reporting bias.
Another concern is selection bias caused by the unequal participation of case and control mothers. More control mothers declined to be interviewed (27%) or had moved out of the study area (14%) than case mothers (of whom 12% declined and 7% moved). Although control mothers who participated in the study were similar demographically to the overall population on the border, we cannot be sure that participation was independent of life events related to mobility. However, assuming that control mothers lost to follow-up had the same mobility and stressful-event profile as case mothers, their inclusion would have reduced the estimated odds ratio for stressful life events only to 2.5.
How stress might cause NTDs is unclear. Stress may indirectly affect neural tube development by inducing high-risk behaviors.18 In response to stress, women might smoke more, drink more alcohol, or eat poorly, behaviors that contribute to an increased risk of NTDs. We adjusted for diet, cigarette smoking, and alcohol consumption, but the observed effects may still reflect unmeasured environmental or personal factors. It is also possible that factors, such as personality, lifestyle, or environment, predispose a woman to both stress and NTDs.
An underlying biologic mechanism whereby maternal stress affects neural development of the fetus has been postulated.26 Wadhwa et al27 has demonstrated that psychologic indices of maternal stress are correlated with increased levels of circulating adrenocorticotropin and cortisol. Maternal cortisol levels are highly correlated with fetal concentrations, suggesting a direct effect of maternal stress hormones on fetal development.26 Elevated cortisol levels during pregnancy are known to induce oral clefts in a variety of animal species.14–16
As in previous studies of NTDs,4,6 our measure of life event stress was limited to a few items available from our questionnaire. Standard approaches to measuring life event stress are based on more comprehensive checklists of 30–50 items,28 including job and residence changes as well as personal injuries.29 Because our stress measure was created ad hoc, we are unable to attest to its reliability or validity in comparison to the standard measures. Better-designed studies using comprehensive measures of stress and social support are needed to confirm the apparent relationship between risk of neural tube defects and maternal stress.
In summary, maternal stress experienced preconceptionally and resulting from certain life events may affect the normal development of the neural tube in offspring. Furthermore, these stress effects may be exacerbated in populations with poor nutritional status and meager economic resources.
We thank the following NTD Project team members for their crucial role in interviewing case and control mothers: El Paso—Hilda Chavarria, Maria Torres, Carmen Ramos, Donna Brom, and Patricia Velazquez; Harlingen—Oralia Villafranca, San Juana Thompson, Graciela Rubio, Manuela Flores, Rene Rodriguez, Sara Mungia, and Jorge Trevino; Laredo—Ricardo Treviñno, Miguel Madrigal, Olivia Macias Gutierrez, Cordelia Bernal, Cynthia Medina de Llano, Jackie Bassini, and Armandina Ortiz. We also acknowledge Rich Ann Baetz, Kelly Johnson, Hermia Brooks, Billie Woullard, Jennifer Tisch, John Dunn, and Jackie Stroupe for their careful work that insured the accuracy of the data. Finally, we are especially grateful to Scott Simpson, who helped write protocols and identify resources in the early years of the study, and to Russell D. Larsen, who provided many years of supportive supervision to the entire team before his retirement.
1.Istvan J. Stress
, anxiety, and birth outcomes:a critical review of the evidence. Psychol Bull
2.Wadhwa PD, Sandman CA, Garite TJ. The neurobiology of stress
in human pregnancy: implications for prematurity and development of the fetal central nervous system. Prog Brain Res
3.McAnarney ER, Stevens-Simon C. Maternal psychological stress
/depression and low birth weight. Is there a relationship? Am J Dis Child
4.Richards ID. Congenital malformations and environmental influences in pregnancy. Br J Prev Soc Med
5.Adams MM, Mulinare J, Dooley K. Risk factors for conotruncal cardiac defects in Atlanta. J Am Coll Cardiol
6.Carmichael SL, Shaw GM. Maternal life event stress
and congenital anomalies. Epidemiology
7.Czeizel A, Nagy E. A recent aetiological study on facial clefting in Hungary. Acta Paediatr Hung
8.Hansen D, Lou HC, Olsen J. Serious life events and congenital malformations: a national study with complete follow-up. Lancet
9.Laumon B, Martin JL, Collet P, et al. Exposure to organic solvents during pregnancy and oral clefts: a case-control study. Reprod Toxicol
10.Saxen I. Cleft lip and palate in Finland: parental histories, course of pregnancy and selected environmental factors. Int J Epidemiol
11.Carmichael SL, Shaw GM. Maternal corticosteroid use and risk of selected congenital anomalies. Am J Med Gene
12.Rodriguez-Pinilla E, Martinez-Frias ML. Corticosteroids during pregnancy and oral clefts: a case-control study. Teratology
13.Czeizel AE, Rockenbauer M. Population-based case-control study of teratogenic potential of corticosteroids. Teratology
14.Fraser FC, Fainstat TD. Production of congenital defects in the offspring of pregnant mice treated with cortisone. Pediatrics
15.Lahti A, Antila E, Saxen L. The effect of hydrocortisone on the closure of the palatial shelves in two inbred strains of mice in vivo and in vitro. Teratology
16.Teramoto S, Hatakenaka N, Shirasu Y. Effects of the Ay
gene on susceptibility to hydrocortisone fetotoxity and teratogenicity in mice. Teratology
17.Heaney CA, Israel BA. Social networks and social support
. In: Glanz K, Lewis FM, Rimer BK, eds. Health Behavior and Eeducation: Theory, Research, and Practice
. San Francisco: Josey-Bass Publishers; 1997:179–205.
18.Cohen S, Kessler RC, Gordon LU. Strategies for measuring stress
in studies of psychiatric and physical disorders. In: Cohen S, Kessler RC, Gordon LU, eds. Measuring Stress. A Guide for Health and Social Scientists
. New York: Oxford University Press; 1997:3–26.
19.Broadhead WE, Kaplan BH, James SA, et al. The epidemiologic evidence for a relationship between social support
and health. Am J Epidemiol
20.Hendricks KA, Simpson JS, Larsen RD. Neural tube defects
along the Texas-Mexico border, 1993-1995. Am J Epidemiol
21.Little J, Elwood JH. Chapter 14. Socio-economic status and occupation. In: Elwood JM, Little J, Elwood JH, eds. Epidemiology and Control of Neural Tube Defects
. Oxford: Oxford University Press; 1992:456–520.
22.Watkins ML. Efficacy of folic acid prophylaxis for the prevention of neural tube defects
. Ment Retard Dev Disabil Res Rev
23.Willett WC. Folic acid and neural tube defects
: can’t we come to closure. Am J Public Health
24.Suarez L, Hendricks KA, Cooper SP, et al. Neural tube defects
among Mexican Americans
living on the US-Mexico border: effects of folic acid and dietary folate. Am J Epidemiol
25.Koenig HG, Westlund RE, George LK, et al. Abbreviating the Duke Social Support
Index for use in chronically ill elderly individuals. Psychosomatics
26.Gitau R, Cameron A, Fisk NM, et al. Fetal exposure to maternal cortisol. Lancet
27.Wadhwa PD, Dunkel-Schetter C, Chicz-DeMet A, et al. Prenatal psychosocial factors and the neuroendocrine axis in human pregnancy. Psychosom Med
28.TurnerRJ, Wheaton B. Checklist measurement of stressful life events. Measuring Stress: A Guide for Health and Social Scientists
. In: Cohen S, Kessler RC, Gorden LU. 1997:29;58, Oxford University Press, New York
29.Holmes TH, Rahe RH. The social readjustment rating scale. J Psychosom Res