Foreign-born status and risk of gestational diabetes mellitus by years of residence in the United States

To explore the association between acculturation among foreign-born (FB) women, gestational diabetes (GDM) and GDM-associated adverse birth outcomes, we conducted a retrospective cohort study of 34,696 singleton pregnancies from Houston, TX, between 2011 and 2022. FB women (n = 18,472) were categorized based on years of residence in US (0–5, 6–10, and > 10 years), while US-born women (n = 16,224) were the reference group. A modified Poisson regression model determined the association between acculturative level and GDM within the entire cohort and stratified by race/ethnicity. Compared to US-born women, FB women with 0–5 years [adjusted relative risk (RRadj.) 1.27, 95% confidence interval [CI] 1.14–1.42)], 6–10 years (RRadj. 1.89, 95%CI 1.68–2.11) and > 10 years in the US (RRadj. 1.85, 95%CI 1.69–2.03) had higher risk of GDM. Results were consistent for all racial/ethnic groups, although associations were not significant at 0–5 years. FB women had lower risk of other adverse pregnancy outcomes, except for preeclampsia with severe features at higher levels of acculturation. Results were similar among those with and without GDM. In conclusion, FB status increases risk of GDM among all racial/ethnic groups but is elevated with higher acculturation levels.

Gestational diabetes mellitus (GDM) is defined as the first onset of insulin intolerance during pregnancy and is one of the most frequently occurring metabolic disorders of pregnancy 1 . Recent estimates place the annual prevalence of GDM in the United States at approximately 7.6% 2 . GDM is associated with increased risk of adverse birth outcomes such as preeclampsia, Cesarean section, macrosomia, and preterm delivery 3 . In addition, those diagnosed with GDM have an increased risk of developing Type 2 diabetes within two years postpartum, and infants born to GDM pregnancies have an elevated risk of metabolic disorders later in life 4 . Thus, GDM is a significant public health concern, and there is a need to understand population-level factors that increase the risk of GDM and GDM-associated complications.
Foreign-born immigrants have higher rates of GDM than their native-born counterparts, despite typically experiencing better health outcomes in a phenomenon commonly called the healthy migrant effect 5 . This health advantage appears to wane with increased acculturation, which can be defined as gradual socio-cultural assimilation to the dominant culture 6 . A recent study among foreign-born Asian women showed a negative relationship between increased acculturation and GDM risk 7 . Although GDM was not examined, other studies among Hispanic, African and White immigrants show an increase in diabetes risk [8][9][10] . The relationship between acculturation and GDM is unclear across different race/ethnic groups.
Our primary objective was to examine the association between foreign-born status (FBS) and GDM across different race/ethnicity groups. In addition, we investigated the effect of acculturation, using years in the US as our proxy, on GDM risk among foreign-born women. As a secondary analysis, we examined the association between different acculturation levels in foreign-born women on GDM-associated adverse birth outcomes.

Methods
Study design and population. We conducted a retrospective cohort study of singleton pregnancies in the Peribank database 11 . Briefly, Peribank was developed by Baylor College of Medicine and recruits women from four hospitals in Houston, Texas, at the time of admission to labor and delivery. Demographic and clinical variables were obtained through interviews and electronic health records. A board-certified obstetrician performs regular audits of the data to maintain high reliability. All participants provided informed consent. The Governance Board of Peribank provided approval for the current study. The University of Texas Medical Branch Institutional Review Board reviewed this study and determined it to be exempt. Study population and exclusion criteria. We identified 49,841 singleton pregnancies between August 2011 and April 2022. Only the first recorded delivery was used in cases where multiple births occurred within that timeframe, and any subsequent birth was excluded (n = 6044). Patients with missing self-reported country of birth (n = 1593) and time spent in the US (n = 7508) were also excluded from our study. After all exclusions, we had a final analytic sample of 34,696 singleton deliveries.

Foreign-born status (FBS) and years lived in the US.
Patients were categorized by race and ethnicity as non-Hispanic White (White), non-Hispanic Black (Black), non-Hispanic Asian (Asian), and Hispanic. Patients' self-reported country of birth was used to determine foreign-born status (FBS). Women who listed 'USA' as their country of birth were indicated as US born, while those with a non-US country were listed as foreign-born. Self-reported time spent in the US by foreign-born women was used as our proxy for acculturation. We used 5-year increments, i.e., 0-5, 6-10, and greater than 10 years (> 10), as our indicator for the level of acculturation. Prior research shows that after 5 years or more in the US, foreign-born women begin to show a decline in the immigrant health advantage 12 .
Primary outcome. Gestational diabetes mellitus (GDM) was our primary outcome of interest. Diagnoses of GDM followed standardized hospital protocols typically using a two-step screening process between 24 and 28 gestation weeks. The first step was administering a 50 g oral glucose test with a 140 mg/dl cutoff indicator, while the second was a 100 g oral glucose tolerance test 13 . Only 118 women (0.3%) were not screened for GDM. Secondary outcomes. For our secondary analysis, we examined the association between acculturation level and GDM-associated adverse birth outcomes. The outcomes examined included: Cesarean section delivery (C-section), gestational hypertension, superimposed preeclampsia, preeclampsia, preeclampsia with severe features, medically indicated preterm birth (PTB) and spontaneous PTB. Preeclampsia was diagnosed based on ACOG criteria, as the new onset of hypertension, along with proteinuria > 300 mg in a urine sample or in the absence of proteinuria evidence of multisystem failure 14 . Preeclampsia with severe features was defined as preeclampsia with one or more of the following signs: (1) multi-organ failure, (2) systolic blood pressure > 160 mmHg or diastolic blood pressure > 110 mmHg, (3) proteinuria, (4) reduced platelet count (5) intrauterine growth restriction 14 .
Superimposed preeclampsia was defined among those with chronic hypertension who went on to develop preeclampsia symptoms 14 . Gestational hypertension was defined as the presence of newly occurring hypertension after 20 gestation weeks without any other preeclampsia symptoms 14 . PTB was defined as a live birth occurring before 37 gestation weeks, where gestational age was determined using ACOG criteria using last menstrual period and correction by ultrasound when needed 15 . Medically indicated PTB was defined as defined as a preterm birth, medically induced due to maternal or fetal distress, while spontaneous PTB occurred due to preterm labor or rupture of membranes 16 . Covariates. Demographic variables obtained from Peribank included maternal age, education (> than high school, high school degree or less), insurance method (private, Medicaid/Children's health insurance program (CHIP), no insurance/other), smoking, and alcohol use. We also obtained data on maternal health history and current health indicators such as pre-pregnancy body mass index (BMI), prior GDM, mental health issues, chronic health conditions (cardiovascular disease, asthma, chronic hypertension), and gravidity. Mental health issues were defined as the presence of either anxiety or depression. Statistical analysis. We compared maternal demographics and clinical variables between foreign-born and US-born women, stratified by race/ethnicity, using the Chi-squared test (Fisher's exact test if cell size < 5) and T-test to determine the presence of significant differences between groups. Next, we examined the distribution of years in US among foreign-born women, stratified by race/ethnicity, both as a categorical (0-5, 6-10 and > 10) and continuous variable. The distribution of years in US, among the most frequently reported countries within each race/ethnic group was also examined.
To determine the association between FBS and GDM, we used modified Poisson regression models 17 to calculate the relative risk (RR) and 95% confidence intervals (CI), a method frequently used to measure relative risk 18 . A log-binomial model was used initially but there were converge issues when adding covariates to our model. To select our confounders, we used a directed acyclic graph (DAG) 19 . However, acknowledging that we do not fully know the relationship between years in the US and some variables such as development of chronic health conditions, we used the disjunctive cause criterion, which includes covariates that are causes of exposure, causes of outcome, or both 20 . Considering criterion and the DAG we created 21 (Supplementary Fig. S5), we decided to utilize 3 models for our analysis. Model 1 was unadjusted, model 2 adjusted for maternal age only, and model 3, www.nature.com/scientificreports/ our fully adjusted model, adjusted for maternal age, pre-pregnancy BMI, insurance method, and chronic health conditions. We considered prior GDM, however this variable is most likely a mediator for some pregnancies.
As a sensitivity analysis we conducted a mediation analysis, using VanderWeele's causal mediation macro 22 , to determine if prior GDM has a mediating effect between FBS and GDM in the current pregnancy. We found some evidence of mediation, therefore, prior GDM was not adjusted for in the model. We stratified our foreign-born population by time spent in the US (0-5, 6-10, > 10 years), using US-born women as our reference group. We explored this analysis among the entire cohort and stratified by race/ethnicity, utilizing the same 3 models as our prior analysis. As a sensitivity analysis we evaluated the association between years in US, as a continuous variable, and GDM among foreign-born women, stratified by race/ethnicity using a spline regression model 23 . Lastly, we examined the association between acculturation and adverse birth outcomes using the methods described above. As GDM is known to increase risk of several adverse birth outcomes, we stratified these analyses by GDM status 24 .
The missingness among the variables in our models ranged from 0.3 to 8.9%. To account for missing data, we used a fully conditional specified multiple imputation with 10 iterations. As a sensitivity analysis we also conduced our primary analysis, looking at the association between FBS and GDM, using complete case analysis. However, effect estimates were nearly identical [e.g., for association between FBS and GDM: (RRadj. 1.62, 95% CI 1.49-1.77) vs. using multiple imputation (RRadj. 1.64, 95% CI 1.54-1.83)], therefore results are shown using the multiple imputation model. All analysis was conducted using SAS version 9.4, Cary, North Carolina.

Distribution of years in US among foreign-born women.
We explored the distribution of years in the US among foreign-born women. Among our foreign-born study population, 7610 women (41.2%) had spent > 10 years in the US, followed by 6947 (37.6%) and 3915 (21.2%) women in the 0-5 and 6-10-year ranges respectively ( Table 2). Among foreign-born Black women, the distribution skewed towards those with less time in the US; those in the 0-5 (61.7%) and 6-10 (22.2%) year groups accounted for over 80% of the population. The majority of foreign-born Black women were from Nigeria, accounting for about 50% of the Black immigrant population.
Among Hispanic, Asian, and White foreign-born women, the distribution of years in the US followed a similar pattern where most women had lived > 10 years in US (42.4%, 46.1%, and 44.9%, respectively), followed by more recent immigrants with 0-5 years in US (35.6%, 35.0%, and 38.0%, respectively) and those with 6-10 years in US (22.1%, 19.0%, and 17.1%, respectively).
Within the Hispanic foreign-born population, three countries, Mexico (49.9%), Honduras (18.0%), and El Salvador (16.5%), accounted for over 84% of the Hispanic population. In contrast there were numerous countries of birth for Asian and White women. Foreign-born Asian women from India (20.7%) followed by China (13.6%) and Vietnam (12.4%) were the most frequently reported in our Asian population. Foreign-born White women from the United Kingdom (8.7%), Canada (6.9%) and Germany (6.5%) were the most frequently reported in our White population.

Association between FBS and GDM.
We explored the association between FBS and GDM across the entire cohort and stratified by race/ethnicity. Among the entire cohort, FBS was associated with an increased risk of GDM (RRadj. 1.64, 95% CI 1.54-1.83) relative to US-born women ( Association between FBS and GDM, stratified by years in the US. Among our entire cohort, we observed foreign-born women with 0-5 years in the US had an increased risk of GDM (RRadj. 1.27, 95% CI 1.14-1.42); however, GDM risk was higher among those with 6-10 years (RRadj. 1 and those with 0-5 years showed no association. Lastly, foreign-born Asian women with 6-10 years in the US had a higher risk of GDM (RRadj. 1.68, 95% CI 1.08-2.60), but the other groups showed no significant association with GDM compared to US-born Asians.
Association between acculturation and adverse birth outcomes. The prevalence of other adverse birth outcomes stratified by race/ethnicity can be found in Supplementary Table S1. Among our entire cohort,    www.nature.com/scientificreports/ 6-10 years (RR 0.71, 95% CI 0.59-0.85) had a lower risk but not those with > 10 years in the US. With preeclampsia with severe features, we observed the opposite trend, foreign-born women with 6-10 years (RRadj. 1.42, 95% CI 1.10-1.84) and > 10 years in the US (RRadj. 1.48, 95% CI 1.19-1.83) had an elevated risk, while the 0-5year group showed no association (RRadj. 1.05, 95% CI 0.83-1.33). Spontaneous preterm birth showed a trend towards higher risk across all acculturative levels in foreign-born women; however, this effect was attenuated in the fully adjusted model. Foreign-born women at all acculturative levels were at lower risk for gestational hypertension, superimposed preeclampsia, and Cesarean section. Results were relatively similar when we stratified by GDM status (Supplementary Table S3).

Sensitivity analyses.
We conducted several sensitivity analyses. First, we examined the association between years in the US as a continuous variable and GDM among foreign-born women. Within the entire cohort, each additional year in the US resulted in a moderate increase in GDM risk (βadj. 0.01, 95% CI 0-0.02, p = 0.01) (Supplementary Table S4). Among foreign-born Hispanic women, each additional year in the US moderately increased GDM risk (βadj. 0.01, 95% CI 0-0.02 p-value 0.01). However, among Black, Asian, and White foreign-born women there was no association. Next, we conducted a mediation analysis to determine whether prior GDM was a mediator in the association between FBS and GDM. The total causal effect of prior GDM was 2.48, 95% CI 2.20-2.77, and the proportion mediated was 0.53. Based on these results, we concluded that prior GDM was a mediator in the association between FBS and GDM.

Discussion
We observed that foreign-born women have a higher risk of GDM compared to US-born women across our entire cohort and in all race/ethnic groups. Furthermore, when examining the effect of acculturation on the risk of GDM among foreign-born women, we found a positive relationship between increasing acculturation and GDM risk. Lastly, we observed that foreign-born women across all three acculturative levels were at lower risk for adverse birth outcomes with a few exceptions, mainly preeclampsia with severe features, which was elevated at higher levels of acculturation.
Similar studies conducted within the United States 25 and other countries, such as Norway 26 and Denmark 27 , show an elevated risk of GDM among foreign-born women and a positive association between acculturation and Table 4. Association between foreign-born (FB) status and gestational diabetes (GDM), by years in US, stratified by race and ethnicity. a Model 1 was unadjusted/crude. b Model 2 adjusted for maternal age. c Model 3 adjusted for maternal age, chronic health conditions, insurance type, and pre-pregnancy body mass index.   25 found no significant association between foreign-born women with less than 10 years in the US and GDM. In contrast, we observed foreign-born women with both 0-5 and 6-10 years in US, had elevated risk of GDM compared to US-born women, although risk was highest at > 10 years in the US and we found a modest increase in GDM per yearly increase of residence in the US. One potential theory explaining this effect is the negative acculturation theory 28 , which follows a social conformity framework. This viewpoint hypothesizes that a longer duration in a new country leads to higher exposure to social norms and further integration with the native-born populace. Within the US, this could be seen with the adoption of a more sedentary lifestyle and high physical inactivity levels among immigrants 29 , which has been shown to increase GDM risk 30 . Furthermore, changes in dietary patterns, such as the consumption of sugar-sweetened food and beverages, which are prevalent in the US, could lead to increased risk of weight gain, Type 2 diabetes and pregnancy complications like GDM [31][32][33] .
There are several factors that could be contributing to our results, showing an increased risk of GDM and preeclampsia with severe features among foreign-born women, particularly at increased acculturative levels. One is the presence of significant barriers to accessing healthcare, along with a lack of familiarity with the system and anti-immigration policies, which has been shown to lead to worse long-term health outcomes [34][35][36] . Also, the lessening of the protective effect from the healthy migrant effect, along with high levels of stress levels due to high levels of prejudice and discrimination, could contribute to worse birth outcomes for our 6-10 and > 10 year foreign-born groups 37 . However, for other adverse outcomes, we observed a lower risk among foreign-born women. But in the case of preeclampsia this protective effect appeared to wane with increasing acculturation.
Our study has several strengths; over 99% of our population was screened for GDM, in contrast with birth record data, where previous studies have found evidence of underreporting of maternal risk factors and pregnancy complications, like diabetes and preeclampsia 38,39 . Peribank recruits about 85% of eligible expecting mothers within the Houston area and our foreign-born populace was highly diverse with a large enough sample size to examine various geographical regions or birth origins. Some of our study's limitations include a need for a quantitative measure of the stress and discrimination our foreign-born population experiences. Another limitation is the lack of a standard measure of acculturation, like a validated acculturation scale 40 . However, length of stay is a highly reliable way of quantifying acculturation by proxy 41 . Another limitation is our lack of data on glucose levels within our population. Due to this we were unable to evaluate the association between FBS and glucose levels. We did not have any records indicating when or where prior GDM was diagnosed. We suspect the same criteria was applied among patients diagnosed and treated in the US, however we have a diverse foreign-born population and the diagnostic criteria for GDM varies by country. Lastly, there is potential selection bias because women with uncertain legal status or high levels of mistrust may not consent to be included in the database. This group could be associated with low levels of prenatal care, leading to an underestimation of the effect in our results. Although we excluded women with missing foreign-born status and length of residence data, we did not observe much difference in GDM prevalence between the missing group and our study population (8.9% vs. 9.6%).
In conclusion, we observed a positive relationship between increased acculturation and GDM risk among foreign-born women across multiple race/ethnic groups. Future research could examine the impact of early interventions, specifically within their first 5 years of residence, in lowering the GDM risk among foreign-born women. Future research is also needed to understand social factors that may drive increased risk of GDM among foreign-born women.

Data availability
The data used in this manuscript was provided to by Peribank under request/license. Access to this data, requires permission from Peribank.