Objective
Gastroschisis is increasing in many countries, especially among young women. Because young women may have inadequate nutrition, we assessed the relationship between individual nutrients and the risk for gastroschisis.
Study Design
We analyzed data from the National Birth Defects Prevention Study, a population-based case-control study. Cases were ascertained from 10 birth defect surveillance systems. Controls were randomly selected from birth certificates or hospital records. Nutrient intake was estimated for the year prior to conception from maternal interviews based on a 58-item food frequency questionnaire and cereal consumption reported. A total of 694 cases and 6157 controls were available for analysis.
Results
Reported intake of individual nutrients did not substantially affect the risk for gastroschisis. Stratification by maternal age, preconception body mass index, folic acid–containing supplements, or energy intake (kilocalories) did not alter risk estimates.
Conclusion
This study does not support an increased risk for gastroschisis with decreasing tertiles of individual nutrients.
Gastroschisis is a unique birth defect, notable for its increasing prevalence in many developed countries and the greatest risk observed in women <20 years of age. This combination suggests an etiologic role of environmental factors, with or without an underlying genetic susceptibility. Adequate nutrition, both prior to and after conception, is important for the normal growth and development of the fetus. Women with a lower than normal bodyweight appear to be at an increased risk for adverse perinatal outcomes such as prematurity and intrauterine growth retardation. Dietary trends in US adolescents have changed over the past several decades and many young women are now considered to have inadequate nutrition, with dieting being a common practice. Reduced intake of micronutrients, measured by maternal self-report, has been associated with an increase in risk for several birth defects. Because the dietary habits of young women are of concern and young women are at the greatest risk for gastroschisis, assessing their nutrient intake is important to understand the extent to which these factors, individually or collectively, contribute to the risk of gastroschisis.
Data on gastroschisis and nutrition are scarce. In a study of only 55 case mothers, Torfs et al reported an increased risk with low levels of some micronutrients that were thought to be indicative of poor overall nutrition: low alpha-carotene, low total glutathione, and high nitrosamines. In a related study, women in the lower range of body mass index (BMI), used as a marker for poor nutritional status, had a 3-fold increase in risk for gastroschisis. Findings from an animal model prompted Lam and Torfs to investigate the association between the risk for gastroschisis and 3 factors indicating a malnourished state: low BMI and low intake of protein and zinc. Results from a previous analysis of data from the National Birth Defects Prevention Study (NBDPS) suggest gastroschisis was modestly associated with higher maternal intake of total energy and fat.
As a next step forward in understanding the association of gastroschisis with maternal nutritional status, this analysis examined multiple nutrients, included data on almost 700 cases, and investigated whether the previously reported association of gastroschisis with young maternal age and BMI could be explained by nutrient intake. Specifically, using data from the large population-based case-control NBDPS, we investigated whether reported average intakes for micronutrients, amino acids, fatty acids, and macronutrients during the year before pregnancy were associated with an increased risk of gastroschisis. We also examined whether associations were modified by maternal intake of folic acid–containing supplements, maternal age, or BMI.
Materials and Methods
The NBDPS is a population-based case-control study involving centers from 10 states (Arkansas, California, Georgia, Iowa, Massachusetts, New Jersey, New York, North Carolina, Texas, and Utah). The NBDPS is the largest case-control study of birth defects in the United States, ongoing since 1997, with North Carolina and Utah added in 2002. Each center’s surveillance system is responsible for the identification of eligible birth defect cases and a nonmalformed control group from a defined geographic base. Birth defect cases with a known etiology (ie, chromosomal or single gene) are excluded from the NBDPS. Eligible cases may include live births, stillbirths, and pregnancy terminations and are selected from the center’s surveillance system based on strict eligibility criteria. Controls include nonmalformed liveborn infants randomly selected from either the state birth certificate files (Arkansas 2000 through 2005, Georgia 2001 through 2005, Iowa, Massachusetts, North Carolina, New Jersey, Utah) or birth hospitals (Arkansas through 1999, California, Georgia through 2000, New York, Texas) to represent the population from which the cases arise. Controls identified with a birth defect were considered ineligible and excluded from the study. Based on eligibility criteria, case and control mothers who speak either English or Spanish are invited to participate in the computerized-assisted telephone interview (CATI) between 6 weeks-24 months after the mother’s estimated date of delivery. All CATI and clinical data are sent to the central repository at the Centers for Disease Control and Prevention. For each birth defect group, 1 clinical geneticist is then responsible to clinically review all center-compiled information to classify every case based on whether their defect is eligible or noneligible and isolated or multiple. Multiple defects involve ≥2 major malformations occurring in 1 infant. For example, an infant would be considered a multiple case if both a cleft lip and gastroschisis were diagnosed. All centers maintain their own annual institutional review board approval.
Dietary assessment
A modified Willett 58-item food frequency questionnaire (FFQ) was used to assess the usual dietary intake of micronutrients and macronutrients, amino acids, and fatty acids of each mother during the year prior to the pregnancy of interest. During the telephone interview, the interviewer reads each food item and queries the participant to determine her average frequency of consumption based on the food item’s specific serving size. Additional questions that are separate from the FFQ are used to query a participant on the type and amount of cereal consumed, as well as her average soda, coffee, and tea consumption. The combined reported intake of food, cereal, and beverage consumption was used to estimate the average daily intake of the micronutrients and macronutrients, amino acids, and fatty acids. Data for these nutrient calculations were derived from the SR20 version of the US Department of Agriculture (USDA) National Nutrient Database for Standard Reference. The USDA SR20 version is relatively complete for all of the studied nutrients, with the exception of choline. For choline, we used the supplemented R20 of the USDA Database of Choline Content of Common Foods. We examined the reported intake of total energy intake (kilocalories), 22 micronutrients (vitamins and minerals), 2 fatty acids (linoleic acid and oleic acid), 4 amino acids (cysteine, glycine, glutamic acid, and methionine), and 3 macronutrients (carbohydrates, protein, and total fat). We grouped these values by tertiles and used the lowest tertile as reference. Cutoff points for the tertiles were based on the distribution of each nutrient among control mothers.
Study subjects and inclusion-exclusion criteria
Case records of all gastroschisis cases were reviewed by a clinical geneticist (C.A.M.) and classified as either isolated or multiple. We excluded gastroschisis cases in which clinical presentation was suggestive of a limb-body wall complex or amniotic band sequence. This review yielded a case group of 742 eligible gastroschisis cases, together with 6807 controls, with estimated due dates from Oct. 1, 1997, through Dec. 31, 2005, and for which maternal interview data were available. We then excluded women who reported pregestational or gestational diabetes (18 cases; 472 controls), had an average daily energy intake <500 or >5000 kcal (12 cases; 82 controls), or for whom ≥3 food items in the FFQ were missing or unknown (18 cases; 96 controls). After these exclusions, 694 cases and 6157 controls were available for analysis.
Statistical analysis
We examined the following variables as potential confounders (reference group is underlined): maternal age (<20, 20-24, 25-29 , ≥30 years), race/ethnicity ( non-Hispanic white , non-Hispanic black, Hispanic, Native American/Alaskan Native, Asian/Pacific Islander, other), education ( <12 , 12, >12 years), preconception BMI (<18.5, 18.5-24.9 , 25.0-29.9, ≥30.0 kg/m 2 ), first-trimester cigarette smoking (yes, no ), alcohol ( none , 1-4, ≥5 drinks/occasion), substance abuse (yes, no ), supplements containing folic acid with or without other vitamins or minerals from 1 month before conception through the first trimester ( none , began taking either in the month before conception or during the first month after conception, began taking either in the second or third month after conception), energy intake (categorical based on controls <1223.4 , 1223.5-1722.1, ≥1722.2 kcal/d), gravidity (no previous pregnancy, at least 1 previous pregnancy ), time to interview ( <12 , ≥12 months), and center ( Arkansas , California, Georgia, Iowa, Massachusetts, New Jersey, New York, North Carolina, Texas, and Utah). We used the Student t test to test differences in means for the reported intake of energy between case and control mothers. Associations of each covariate with gastroschisis were assessed using χ 2 analysis. Any covariate with a P value < .25 was included in the initial full model. To assess effect modification we also stratified each micronutrient and macronutrient by folic acid–containing supplements, maternal age, and BMI. We used a backward selection approach in conjunction with the likelihood ratio test to determine the most parsimonious logistic regression model: the variable with the highest and nonsignificant P value in the full model was removed in the nested model and the 2 models were compared. Separate backward selection models were evaluated using 2 different micronutrients (beta-carotene and folate) to evaluate which covariates would be included in the standard set of covariates for the final logistic regression models for all micronutrients and macronutrients. For both beta-carotene and folate, the same 3 covariates were removed in the same order from the logistic regression model based on the criteria described above (ie, substance abuse, folic acid–containing supplements, and gravidity). Although energy intake (kilocalories) did not meet our inclusion criteria as a covariate in the final logistic regression model, it remained in the standard set because diet composition may vary by the level of total caloric intake. The final logistic regression model for each micronutrient and macronutrient contained the following covariates: maternal age, race/ethnicity, education, preconception BMI, smoking, alcohol, energy intake (kilocalories), time to interview, and center. Gastroschisis risk was estimated by tertiles for each micronutrient and macronutrient calculating crude odds ratios (data not shown) and adjusted odds ratios (aORs), with 95% confidence intervals (CIs) for all, isolated, and multiple gastroschisis cases. All analyses were conducted using SAS 9.1.3 (SAS Institute Inc, Cary, NC [2002–2003] or STATA/SE 10; StataCorp LP, College Station, TX [1996–2010]).
Results
The distributions of self-reported maternal characteristics for case and control mothers are presented in Table 1 . Compared to controls, case mothers were more likely to be younger (<25 years), be primigravid, be of Hispanic or of other race/ethnicity, be less educated, have a lower BMI, smoke cigarettes, use illicit drugs, drink at least 5 drinks per occasion, and infrequently consume folic acid–containing supplements.
| Maternal characteristic | Cases, n = 694 | Controls, n = 6157 | cOR (95% CI) |
|---|---|---|---|
| Age, y | |||
| <20 | 326 (47.0) | 863 (14.0) | 7.1 (5.5–9.1) |
| 20-24 | 240 (34.6) | 1469 (23.9) | 3.1 (2.4–4.0) |
| 25-29 | 88 (12.7) | 1656 (26.9) | 1.0 |
| ≥30 | 40 (5.8) | 2169 (35.2) | 0.3 (0.2–0.5) |
| Race/ethnicity | |||
| White, non-Hispanic | 377 (54.3) | 3739 (60.7) | 1.0 |
| Black, non-Hispanic | 49 (7.1) | 691 (11.2) | 0.7 (0.5–0.96) |
| Hispanic | 205 (29.5) | 1327 (21.6) | 1.5 (1.3–1.8) |
| Other | 62 (8.9) | 375 (6.1) | 1.6 (1.2–2.2) |
| Missing | 1 (0.1) | 25 (0.4) | – |
| Education, y | |||
| <12 | 206 (29.7) | 1001 (16.3) | 1.0 |
| 12 | 265 (38.2) | 1494 (24.3) | 0.9 (0.7–1.1) |
| >12 | 215 (31.0) | 3630 (59.0) | 0.3 (0.2–0.4) |
| Missing | 8 (1.1) | 32 (0.5) | – |
| Preconception BMI, kg/m 2 | |||
| <18.5 | 60 (8.7) | 340 (5.5) | 1.23 (0.9–1.6) |
| 18.5-24.9 | 485 (69.9) | 3380 (54.9) | 1.0 |
| 25.0-29.9 | 105 (15.1) | 1313 (21.3) | 0.6 (0.4–0.7) |
| ≥30.0 | 30 (4.3) | 893 (14.5) | 0.2 (0.16–0.3) |
| Missing | 14 (2.0) | 231 (3.8) | – |
| Cigarette smoking | |||
| No | 437 (63.0) | 4986 (81.0) | 1.0 |
| Yes | 255 (36.7) | 1159 (18.8) | 2.5 (2.1–3.0) |
| Missing | 2 (0.3) | 12 (0.2) | – |
| Substance abuse | |||
| No | 584 (84.2) | 5834 (94.8) | 1.0 |
| Yes | 84 (12.1) | 247 (4.0) | 3.4 (2.6–4.4) |
| Missing | 26 (3.7) | 76 (1.2) | – |
| Alcohol use (drinks/occasion) | |||
| None | 408 (58.8) | 3834 (62.3) | 1.0 |
| 1-4 | 164 (23.6) | 1763 (28.6) | 0.9 (0.7–1.1) |
| ≥5 | 110 (15.9) | 492 (8.0) | 2.1 (1.7–2.6) |
| Missing | 12 (1.7) | 68 (1.1) | – |
| Folic acid supplements | |||
| None | 113 (16.3) | 769 (12.5) | 1.0 |
| B1 or P1 | 265 (38.2) | 3238 (52.6) | 0.6 (0.4–0.7) |
| P2 or P3 | 304 (43.8) | 2067 (33.6) | 1.0 (0.8–1.3) |
| Missing | 12 (1.7) | 83 (1.3) | – |
| Gravidity | |||
| Primigravid | 351 (50.6) | 1847 (30.0) | 2.4 (2.0–2.8) |
| Multigravid | 343 (49.4) | 4310 (70.0) | 1.0 |
| Time to interview (mo) | |||
| ≤6 | 146 (21.0) | 2059 (33.4) | 1.0 |
| 7-11 | 317 (45.7) | 2656 (43.1) | 1.7 (1.4–2.1) |
| 12-18 | 153 (22.1) | 979 (15.9) | 2.2 (1.7–2.8) |
| ≥19 | 75 (10.8) | 442 (7.2) | 2.4 (1.8–3.2) |
| Missing | 3 (0.4) | 21 (0.3) | – |
| Center | |||
| Arkansas | 100 (14.4) | 772 (12.5) | 1.0 |
| California | 155 (22.3) | 765 (12.4) | 1.6 (1.2–2.1) |
| Georgia | 60 (8.7) | 661 (10.7) | 0.7 (0.5–1.0) |
| Iowa | 76 (11.0) | 675 (11.0) | 0.9 (0.6–1.2) |
| Massachusetts | 64 (9.2) | 801 (13.0) | 0.6 (0.4–0.9) |
| New Jersey | 46 (6.6) | 547 (8.9) | 0.6 (0.5–0.9) |
| New York | 38 (5.5) | 549 (8.9) | 0.5 (0.4–0.8) |
| North Carolina | 22 (3.2) | 359 (5.8) | 0.5 (0.3–0.8) |
| Texas | 84 (12.0) | 688 (11.2) | 0.9 (0.7–1.3) |
| Utah | 50 (7.2) | 340 (5.5) | 1.1 (0.8–1.6) |
Energy and macronutrients
Overall, the reported mean caloric (energy in kilocalories) intake during the year prior to conception was 7% higher among case mothers compared to control mothers (1763 vs 1638 kcal, P < .0001). Compared to the lowest tertile for carbohydrates, protein, and fat, the middle and highest tertiles were very close to 1, with all 95% CI including 1, and no obvious trend was evident ( Table 2 ). Among the group of isolated gastroschisis there was no trend of increasing or decreasing risk with increasing macronutrient intake. Among the group of multiple gastroschisis (gastroschisis with additional major malformations), a much smaller group, the results were mixed and inconsistent for carbohydrate and protein intake, whereas higher total fat intake tended to be associated with lower risk for gastroschisis, although aOR estimates were imprecise.
| Macronutrient | Controls, n = 6157 | All cases, n = 694 | Isolated cases, n = 633 | Multiple cases, n = 61 | ||||||
|---|---|---|---|---|---|---|---|---|---|---|
| No. of controls | No. of cases | aOR a | (95% CI) | No. of cases | aOR a | (95% CI) | No. of cases | aOR a | (95% CI) | |
| Carbohydrates, g | ||||||||||
| 39.3-166.9 | 2034 | 187 | 1.00 | 170 | 1.00 | 17 | 1.00 | |||
| 167.0-246.0 | 2092 | 221 | 1.02 | (0.76–1.37) | 206 | 1.02 | (0.75–1.38) | 15 | 0.91 | (0.37–2.26) |
| 246.1-923.0 | 2031 | 286 | 0.98 | (0.65–1.46) | 257 | 0.95 | (0.63–1.45) | 29 | 1.21 | (0.35–4.20) |
| Protein, g | ||||||||||
| 10.5-54.7 | 2045 | 261 | 1.00 | 235 | 1.00 | 26 | 1.00 | |||
| 54.8-76.0 | 2081 | 195 | 0.88 | (0.68–1.14) | 184 | 0.92 | (0.71–1.21) | 11 | 0.49 | (0.21–1.13) |
| 76.1-280.8 | 2031 | 238 | 0.98 | (0.71–1.35) | 214 | 1.00 | (0.72–1.39) | 24 | 0.77 | (0.31–1.92) |
| Total fat, g | ||||||||||
| 5.5-37.6 | 2032 | 221 | 1.00 | 198 | 1.00 | 23 | 1.00 | |||
| 37.7-54.7 | 2101 | 222 | 1.02 | (0.78–1.32) | 206 | 1.07 | (0.81–1.40) | 16 | 0.59 | (0.26–1.32) |
| 54.8-214.7 | 2024 | 251 | 0.91 | (0.66–1.27) | 229 | 0.97 | (0.69–1.37) | 22 | 0.47 | (0.18–1.21) |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
