Birth outcomes have been shown to vary by maternal nativity status (Collins et al., 2013; Elo et al., 2014; Gagnon et al., 2009; Wingate and Alexander, 2006). Specifically, rates of adverse

pregnancy outcomes such as infant mortality, low birth weight, and preterm birth are generally higher among US-born mothers compared to foreign-born counterparts. Research has also demonstrated that the relative health advantage of migrants for birth outcomes is moderated by race/ethnicity. The differences in pregnancy outcomes by nativity among non-Hispanic whites and Asians are not as striking as the differences among blacks or Hispanics (Acevedo-Garcia et al., 2005; Cabral et al., 1990; Fuentes-Afflick et al., 1998: Howard et al., 2006; Janevic et al.,2011; Kramer et al., 2006; Lim et al., 2010; Madan et al., 2006; Rosenberg et al., 2005). However, information regarding the contribution of nativity and race/ethnicity to infant morbidities, specifically birth defects, is limited. Birth defects collectively are common pregnancy outcomes, and are a leading cause of infant mortality, accounting for one in five infant deaths in the United States and contributing to disability and health care costs (Parker et al., 2010; Petrini et al., 2002). Thus, birth defects are an important outcome and understudied in relation to our understanding of migration and infant morbidity.

The mechanisms by which differences in birth outcomes by nativity occur are not well understood; however, several explanations have been proposed. The healthy migrant theory posits that due to selection processes that enable healthier persons to migrate, migrants arrive in the United States with a health advantage (Palloni and Ewbank, 2004; Wingate and Alexander, 2006).

A contrasting theory proposes a potentially adverse role of acculturation and the incorporation of unhealthy norms and behaviors adopted after immigration. As time in the United States lengthens, the changes and deterioration of health status in immigrants and in subsequent generations is related to the adaptation of the populations to the behaviors and attitudes that are not in line with those of the country of origin (Collins et al., 2013; Elo et al., 2014). Identifying and understanding variations of birth defects by nativity and race/ethnicity may provide new insights into cultural, environmental, and other mechanisms influencing child health.

Previous population-based analyses, conducted through the National Birth Defects Prevention Network (NBDPN), have examined the relationship between race/ethnicity and major birth defects (Canfield et al., 2014), as well as mortality and survival with major birth defects and a detailed examination of trisomy 13 and 18 (Meyer et al., 2016; Wang et al., 2015). Smaller studies have examined the additional role of maternal nativity for specific birth defects (Canfield et al., 2014; Hoyt et al., 2014; Khodr et al., 2013; Ramadhani et al., 2009; Salemi et al., 2009), but no study has examined comprehensive NBDPN data for maternal nativity, race/ethnicity, and country of birth. Using pooled data from 11 U.S. population-based birth defects surveillance programs, this study examines the prevalence of selected major birth defects overall, and by race/ethnicity within maternal nativity status categories. We hypothesize that the prevalence for those infants born to foreign-born mothers will vary from US-born mothers within the same racial categories, but that variability will also be present across racial/ethnic classifications.

The NBDPN Data Committee issued a call to population-based birth defects surveillance programs in the United States for a multi-state collaborative project to examine major birth defects, maternal race/ethnicity, and nativity status. A previous article using this dataset focused on the prevalence of major birth defects by maternal race/ethnicity (Canfield et al., 2014). This article extends that work by examining maternal nativity status for 27 birth defects (anencephalus; spina bifida without anencephalus; encephalocele; anotia/microtia; common truncus; transposition of great arteries; tetralogy of Fallot; atrioventricular septal defect with and without Down syndrome; aortic valve stenosis; hypoplastic left heart syndrome; coarctation of the aorta; cleft palate without cleft lip; cleft lip alone; esophageal atresia; congenital hypertrophic pyloric stenosis; rectal and large intestinal atresia; hypospadias; upper, lower, and any limb deficiency; diaphragmatic hernia; gastroschisis; omphalocele; Down syndrome; trisomy 13; and trisomy 18) for years 1999–2007.

Eleven population-based surveillance programs participated: Arizona, Colorado, Florida, Illinois, Massachusetts, Michigan, North Carolina, Nebraska, New Jersey, New York, and Texas. The Texas Program served as the deferring IRB, and CDC was the central repository for the anonymized datasets. Programs were provided a data dictionary outlining the format for the selected birth defects and requested data elements. Infants with multiple defects were counted for each applicable condition except for anencephaly and spina bifida, which were mutually exclusive. All participating programs conduct population-based birth defects surveillance. Four states employed active case-finding (Arizona, Massachusetts, North Carolina, and Texas), where cases are identified by review of medical charts. The remaining states (Colorado, Florida, Illinois, Michigan, Nebraska, New Jersey, and New York) used passive case-finding methodologies, relying on administrative datasets or hospital reporting for case identification. Because some states were unable to provide data on all conditions for all study years, this is noted as relevant in the tables accompanying this study (Canfield et al., 2014). Each program provided data on cases among live births, and where available, fetal deaths or pregnancy terminations, as well as provided files with all resident state births for use in analyses. Sociodemographic data were taken from birth certificates.

Nativity was categorized as US-born if the maternal country of birth as reported on the birth certificate indicated United States; otherwise, if the maternal country of birth was not missing, the nativity status was categorized as foreign-born. Because categorization by nation of origin varied by state, we were unable to further classify nativity status. We excluded cases with multiple or missing maternal race/ethnicity and/or missing nativity, as well as those who were native American/Alaskan natives. We could not compare American Indians or Alaska natives by nativity status, as a substantial number of all women within this racial/ethnic group were born in the United States, and the very small numbers of women in this sub-population born outside the United States were insufficient for valid statistical comparisons. Puerto Ricans were grouped in the US-born category if they self-identified their country of origin as United States; otherwise, they were classified as foreign-born.

Data cleaning and analysis were performed using SAS 9.3 (Cary, NC). Crude prevalence rates were calculated per 10,000 live births, with 95% confidence intervals (CIs) calculated using the exact method. Poisson regression was used to create crude prevalence ratios (PR) with the reference category of US-born for each racial/ethnic group (non-Hispanic white, all Hispanic ethnicities, Mexican, non-Hispanic black, and non-Hispanic Asian/Pacific Islander). Multivariable analyses were performed to adjust the PR by maternal state of residence at delivery and maternal age category (<20, 20–34, and ≥35 years). Adjusted prevalence ratios were calculated as the ratio of the prevalence of the child’s defect among infants born to foreign-born mothers over the prevalence among infants born to US-born mothers, adjusted for maternal age and state of residence. The 95% CIs were calculated using the Wald method.

Table 1 presents the percentage of foreign-born mothers by race/ethnicity and state, both for cases and total live births. Specific demographic information is included in Supporting Information Table SS1. This study examined 172,235 cases of birth defects among 12,954,369 live births. Births to foreign-born women constituted 26% of all births, and 20% of birth defects cases. While child sex did not differ by maternal nativity, foreign-born mothers were less likely to be younger than 20 years of age, or of non-Hispanic White race/ethnicity. While the relative contribution of foreign-born mothers to all births varied by state of residence, each state contributed a sizeable number of births and cases to the study population. It is notable that for most states, approximately 90% of Asian mothers were foreign-born. Over one-half of Hispanic and specifically Mexican mothers (a subset of Hispanic mothers) were also foreign-born. The percentage of foreign-born non-Hispanic black mothers was lower, but varied greatly across states. Non-Hispanic white mothers were <10% foreign-born, with several exceptions (Massachusetts, New Jersey).

Prevalence rates and 95% CIs for each of the 27 birth defects, by race/ethnicity and nativity, are shown in Table 2. For many conditions the greatest differences in the US-born versus foreign-born prevalence were observed within the non-Hispanic white category. Crude prevalence ratios comparing prevalence of infants born to foreign-born mothers to the prevalence among infants born to US-born mothers within each racial/ethnic category and for each defect are shown in Supplemental Table 2.

Table 3 presents adjusted prevalence ratios (aPRs) with 95% CIs for each condition. (Among non-Hispanic white mothers, about one-half of the birth defects showed significantly lower prevalence among foreign-born relative to US-born mothers, as evidenced by down arrows (aPRs <1.0 and 95% CI not including the null.) We observed aPRs with magnitudes as small as 0.5 for pyloric stenosis comparing foreign-born to USborn mothers in each racial/ethnic group. There were very few significantly elevated aPRs for other racial/ethnic categories, as evidence by up arrows (aPRs >1.0 and 95% CI not including the null). Among non-Hispanic black mothers, elevated aPRs for foreign-born versus US-born were seen for trisomy 13 and 18, while elevated aPRs were observed for spina bifida, anotia/microtia, and Down syndrome among Hispanic foreign-born versus US-born. Among Asian mothers, no elevated aPRs by nativity were observed; however, significantly lower prevalence was observed for six birth defects (encephalocele, anotia/microtia, tetralogy of Fallot, pyloric stenosis, gastroschisis, and omphalocele) for foreign-born mothers. We observed aPRs below 1 and statistically significant for at least three of the five foreign-born groups for the following birth defects: pyloric stenosis, hypospadias, and gastroschisis.

This large population-based study in the United States examined the prevalence of selected birth defects by race/ethnicity and maternal nativity. Our study provides estimates of birth defect prevalence for US-born and foreign-born non-Hispanic white, non-Hispanic black, non-Hispanic Asian women, as well as for all Hispanic and a subset limited to Mexican women. The covered live birth population (almost 13 million) represents approximately a third of all births in the United States during 1999–2007. The large size of the study population enabled comparisons of birth defects prevalence by nativity status for 27 specific conditions, providing a first look at the potential role of nativity as a factor influencing prevalence.

In general, within the four major racial/ethnic groups, prevalence of the specific birth defects was similar between US-born and foreign-born women. Infants born to foreign-born non-Hispanic white women were significantly less likely to be affected by 11 of the 27 conditions analyzed (Table 3). However, the prevalence ratios tended to be in the range of 0.7–0.8 for most conditions, suggesting a relatively small diminution in prevalence for infants born to foreign-born non-Hispanic white women. The high number of modest but statistically significant differences in prevalence identified for this group is probably associated with the much higher number of total births among non-Hispanic whites in the dataset. Three conditions with a reduced prevalence were identified in the foreign-born Hispanic group, all of which had also been identified as conditions for which foreign-born non-Hispanic white mothers were at a reduced risk.

Prevalence was significantly elevated for spina bifida, anotia/microtia, and trisomy 21 among infants born to foreign-born Hispanic mothers. Similar patterns were found in the separate analysis of Mexican births to foreign-born and US-born mothers for these conditions except for anotia/microtia (although elevated but not statistically significant). These findings are generally concordant with those reported by Shumate et al. (2018), Padula et al. (2017), and Hoyt et al. (2014). Among infants born to non-Hispanic black women, prevalence was significantly lower for only one condition, pyloric stenosis, while higher prevalence was found for trisomy 13 and 18 among those with foreign-born mothers. The analysis for births among non-Hispanic Asian/Pacific Islander women revealed six conditions for which prevalence varied between foreign-born and US-born.

Gastroschisis and pyloric stenosis showed a persistent pattern of reduced prevalence by foreign nativity across all racial/ethnic groups (excepting only for non-Hispanic foreign-born black women for gastroschisis), generally confirming a Florida-specific analysis by Salemi et al. (2009). Similar to the present analysis, Salemi et al. were unable to explore the role of maternal pre-pregnancy body mass index. Risk factors for gastroschisis pose an explanatory conundrum yet to be unraveled—the highest prevalence as well as the greatest increase in recent prevalence is found among women of young maternal age (Jones et al., 2016; Kirby et al., 2013), while women with lean body mass tend to have the highest risk (Siega-Riz et al., 2009); however, the complex inter-relationships between maternal age, nativity, race/ethnicity, and maternal body mass and pregnancy weight gain have not been fully explored.

A large body of literature considers nativity and the migrant paradox of healthy birth outcomes (Juarez and Revuelta-Eugercios, 2016). However, because individual birth defects are rare and the data required for examination of these outcomes is beyond the scope of the typical administrative databases utilized to analyze other birth outcomes, for many of the 27 birth defects in this study there are few or no studies of the role of maternal nativity. Explanations for the paradox include US-born women’s disadvantage in regard to biomedical, nutritional, and psychosocial risk factors (de la Rosa, 2002). These mechanisms may also explain the protective effect of foreign-born status we found for many conditions. A growing body of literature, however, has recognized that immigrant women may be at greater risk of certain maternal health outcomes during pregnancy, such as gestational diabetes (Savitz et al., 2008) and severe maternal morbidity (Howell et al., 2017). It is plausible that the conditions for which we found foreign-born women to be at an increased risk share factors with these maternal conditions, such as poor access to health care. Other conditions in our study for which foreign-born women have an increased prevalence, such as spina bifida and Trisomy 21, are those for which prenatal diagnostic testing are available given optimal prenatal care. We were unable to assess access to or quality of care before or during pregnancy or explore other potential mechanisms for the observed differences in these conditions. Given there is no clear pattern between US-born and foreign-born women, it is possible that multiple protective and deleterious factors are simultaneously at play. Overall, our findings fit into the complex story that is emerging in research on immigration and maternal and infant health.

This study examined differences in birth defects prevalence by nativity status across the United States, showing elevated prevalence ratios for some specific conditions and lower prevalence for others. We hope the findings spur interest in more detailed analyses using a broader range of maternal behaviors and characteristics, including pre-pregnancy body mass index, diet and nutrition in the preconception period and during pregnancy, and use or exposure to tobacco products (Ramadhani et al., 2011). Maternally linked pregnancy outcome files would enable examination of risk for birth defects in subsequent pregnancies. Assimilation and acculturation cannot be directly measured with available vital statistics data, yet these socio-cultural factors may influence the prevalence of birth defects among immigrant women, leading potentially to significantly lower or in some cases higher prevalence of specific birth defects. Previous studies have also illustrated the disparities in access to care, among foreign-born or non-US citizens compared to US-born women (Derose et al., 2007; El-Sayed and Galea, 2012; Goldfarb et al., 2017; Korinek and Smith, 2011; Massey et al., 2017). Evidence suggests that these factors should also be explored in the context of birth defects outcomes. The limited access to services including preconception and prenatal care services could influence the overall prevalence by nativity.