Abstract
Objective
Certain racial and ethnic groups have historically been labeled “high-risk” for the development of gestational diabetes mellitus (GDM). Our objective was to identify the psychosocial, behavioral, and medical factors associated with GDM and determine if they differ by race/ethnicity.
Study Design
Secondary analysis of a multicenter, prospective cohort study of pregnant nulliparous individuals with singleton gestations (2010–2013). The primary outcome was GDM. Psychosocial, behavioral, and medical characteristics were compared by self-reported race/ethnicity. Multivariable logistic regression with backward selection identified factors associated with GDM. Interaction terms between race/ethnicity and risk factors were tested.
Results
Of 8,672 pregnant individuals, 61% were non-Hispanic White, 13% non-Hispanic Black, 17% Hispanic, 4% Asian, and 5% other. The incidence of GDM differed by race/ethnicity with 4% non-Hispanic White, 3% non-Hispanic Black, 5% Hispanic, 11% Asian, and 5% other (p < 0.001). Of 34 psychosocial, behavioral, and medical factors, those associated with GDM were parent with history of diabetes (adjusted odds ratio [aOR]: 1.72; 95% confidence interval [CI]: 1.33–2.23), non-English language (aOR: 2.57; 95% CI: 1.14–5.79), daily calorie intake (aOR: 1.18; 95% CI: 1.08–1.29), daily fiber intake (aOR: 0.84; 95% CI: 0.75–0.94), maternal age (aOR: 1.53; 95% CI: 1.37–1.70), prepregnancy BMI (aOR: 1.21; 95% CI: 1.02–1.44), and waist circumference (aOR: 1.21; 95% CI: 1.03–1.43). These associations did not differ based on race/ethnicity (interaction p-values > 0.1).
Conclusion
Replacing race/ethnicity as a risk factor for GDM with significant upstream psychosocial, behavioral, and medical factors should be considered.
Keywords: race, ethnicity, gestational diabetes mellitus, psychosocial, behavioral
Gestational diabetes mellitus (GDM) complicates 18 million live births globally each year, corresponding to one in six live births, with racial and ethnic disparities in GDM development.1 Diligent management of GDM is crucial in mitigating serious adverse maternal and neonatal outcomes including preeclampsia, cesarean delivery, macrosomia, neonatal hypoglycemia, shoulder dystocia, and birth trauma, but disparities are also seen in adverse pregnancy outcomes.2–5 A history of GDM also increases the risk of long-term morbidities such as poor neuropsychiatric development and language delays in offspring, as well as obesity, type 2 diabetes mellitus, and cardiovascular disease for both the mother and her child.6–9 Due to the increased risks of both short- and long-term complications, GDM increases the costs of obstetric and lifelong healthcare costs.10
Several studies have identified risk factors for GDM. In addition to poor diet, overweight or obese body mass index (BMI), personal history of GDM, family history of diabetes, and certain minoritized racial or ethnic backgrounds have been associated with developing GDM.11–16 Specifically, individuals who identify as Asian or Pacific Islander race or Hispanic ethnicity have a higher risk of developing GDM. However, it is recognized that race is a social construct, and health disparities such as the development of GDM are the result of social determinants of health. More nuanced factors such as perceived stress and psychosocial well-being also contribute to GDM.17–19 Despite mounting evidence that race is a social construct as opposed to a biological risk factor, multiple professional societies, including the American College of Obstetricians and Gynecologists (ACOG) and the American Diabetes Association (ADA), continue to consider high-risk race or ethnicity as a risk factor for the development of diabetes.20–22
Using race and ethnicity to identify individuals at increased risk of GDM not only overly simplifies the risk factors for GDM and hinders our ability to identify interventions to address the root causes of the disease, but it also has the potential to perpetuate bias. For example, ignoring the effects of structural racism and attributing higher rates of disease to race alone can result in the propagation of institutional discrimination. In contrast, understanding how a diverse array of psychosocial, behavioral, and medical risk factors influence GDM risk within racial-ethnic groups may instead direct attention to the development of person-centered, individualized interventions to improve health equity.
Thus, our objective was to evaluate the psychosocial, behavioral, and medical factors associated with the development of GDM and determine if these associations differ based on racial and ethnic group.
Research Design and Methods
This study was a secondary analysis of data from the “Nulliparous Pregnancy Outcomes Study: Monitoring Mothers-to-Be” (nuMoM2b) study, which was a multicenter prospective cohort study of nulliparous individuals with singleton gestations enrolled at 8 centers in the United States from October 2010 to September 2013. The details of the nuMoM2b study have been previously described by Haas et al.23 In brief, pregnant individuals enrolled in this study underwent four study visits during pregnancy, with the first between 60/7 and 136/7 weeks of gestation. Detailed questionnaires were completed at each visit and medical records were abstracted by trained study staff. All testing for GDM was completed via routine clinical care (i.e., not as a part of the study protocol).
Individuals with pregestational diabetes, early pregnancy loss (induced or spontaneous abortions occurring prior to the opportunity for GDM diagnosis), or missing information about GDM diagnosis or maternal race or ethnicity were excluded from this analysis. We a priori chose to evaluate potential risk factors for GDM in three categories: psychosocial risk factors (defined as demographic or patient-reported factors that reflect the unique nature of each individual’s lived experience); behavioral factors (defined as factors related to health activities or choices); and medical factors (defined as factors related to medical history). Psychosocial factors included marital status, income, insurance, education, employment, primary language, perceived stress, experience of racism, resilience, perceived social support, health literacy, depression, and parents with a history of diabetes.24–29 Behavioral factors included planned pregnancy, receipt of prepregnancy counseling on appropriate weight gain, physical activity, tobacco use, daily glycemic load, daily glycemic index, total calories, total and percentage of carbohydrates, fiber, protein, fat, and the Healthy Eating Index 2010 score.30,31 Medical factors included age, BMI, waist circumference, chronic hypertension, asthma, and polycystic ovarian syndrome. In addition, the nuMoM2b study directly queried participants and their self-identified racial and ethnic identity; for this analysis, we categorized these responses as non-Hispanic White, non-Hispanic Black, Hispanic (all racial groups), Asian or Pacific Islander, and other.
The primary outcome was GDM, defined by the nuMoM2b study using clinical documentation and abnormal screening test results by 3-hour glucose tolerance test (GTT) as follows:1) fasting 3-hour 100 g GTT with two abnormal values: fasting ≥95 mg/dL, 1-hour ≥180 mg/dL, 2 hours≥155 mg/dL, 3-hours ≥140 mg/dL; 2) fasting 2-hours 75 g GTT with one abnormal value: fasting ≥92 mg/dL, 1-hour ≥180 mg/dL, 2-hours ≥153 mg/dL; 3) nonfasting 50 g GTT ≥200 mg/dL if no fasting 3- or 2-hours GTT was performed.22 In addition to GTT data, nuMoM2b chart abstractors recorded if a diagnosis of GDM was made during the course of clinical care. If no GTT data were available, the information from chart abstraction was used for GDM classification.
Baseline characteristics were compared by race/ethnicity using Chi-squared and ANOVA tests. Multivariable logistic regression was used to determine which factors were independently associated with GDM. First, we modeled the association with GDM in 3 separate groups: 1) psychosocial, 2) behavioral, and 3) medical. For each initial model, we included all factors within that group and used stepwise, backward selection eliminating those with p < 0.05 to generate the most parsimonious model and identify which factors in that group were significantly associated with GDM. We then tested for interaction between each of the factors in the parsimonious model and race to see if the association between these factors and GDM was different in certain racial/ethnic subgroups. Next, we created a combined model with all the significant factors from the first three models and again performed backward selection to identify the factors independently associated with GDM. We tested for interaction with race in a similar fashion and presented the final model overall and stratified by race.
Significance was determined using a two-tailed test with α set to 0.05. There was no imputation for missing data or adjustment for multiple comparisons. Stata version 17.0 (College Station, TX; StataCorp, LLC.) was used for all statistical analyses. Approval for this analysis of publicly available data was granted by the University of Alabama at Birmingham Institutional Review Board (approval no.: IRB-300004917). All nuMoM2b participants signed written, informed consent for participation and the original study was approved by each site.
Results
Of 9,289 pregnant individuals in the publicly available nuMoM2b dataset, 138 were excluded due to pregestational diabetes, 431 for missing information about GDM diagnosis, 3 for missing information about race or ethnicity, and 46 for early pregnancy loss or termination. Of the 8,672 pregnant individuals included in this analysis, 5,296 (61%) were non-Hispanic White, 1,122 (13%) non-Hispanic Black, 1,477 (17%) Hispanic, 342 (4%) Asian, and 435 (5%) other. Every psychosocial, behavioral, and medical characteristic differed by race and ethnicity (Table 1). For example, individuals who self-identified as non-Hispanic White or Asian were more likely to be married, have private insurance, and higher level of education whereas those who self-identified as non-Hispanic Black or Hispanic were more likely to live below the poverty level and have lower health literacy. Individuals who identified as non-Hispanic Black had the highest reported daily caloric intake and lowest Health Eating Index score, and those who identified as non-Hispanic White or Asian were more likely to have a planned pregnancy and greater physical activity.
Table 1.
Maternal psychosocial, behavioral, and medical characteristics compared by race/ethnicity
| non-Hispanic White (n = 5,296) | non-Hispanic Black (n = 1,122) | Hispanic (n = 1,477) | Asian (n = 342) | Other (n = 435) | |
|---|---|---|---|---|---|
| Psychosocial characteristics | |||||
| Married | 4,022 (76%) | 151 (13%) | 549 (37%) | 297 (87%) | 188 (43%) |
| Income | |||||
| > 200% federal poverty level | 3,810 (79%) | 214 (31%) | 455 (50%) | 269 (87%) | 197 (57%) |
| 100–200% federal poverty level | 575 (12%) | 144 (21%) | 201 (22%) | 26 (8%) | 61 (18%) |
| < 100% federal poverty level | 429 (9%) | 324 (48%) | 262 (29%) | 15 (5%) | 85 (25%) |
| Private insurance | 4,385 (83%) | 382 (35%) | 588 (40%) | 291 (85%) | 237 (55%) |
| Education | |||||
| Less than high school | 206 (4%) | 207 (18%) | 203 (14%) | 4 (1%) | 65 (15%) |
| High school or GED | 387 (7%) | 286 (25%) | 269 (18%) | 9 (3%) | 58 (13%) |
| Some college, associate/technical degree | 1,322 (25%) | 437 (39%) | 609 (41%) | 59 (17%) | 133 (31%) |
| College or higher degree | 3,381 (64%) | 192 (17%) | 391 (27%) | 269 (79%) | 178 (41%) |
| Employed | 3,849 (87%) | 433 (57%) | 698 (64%) | 215 (77%) | 236 (70%) |
| Non-English primary language | 4 (0.1%) | 0 | 121 (8%) | 3 (1%) | 0 |
| Cohen perceived stress | |||||
| Low | 3,431 (65%) | 470 (42%) | 751 (52%) | 198 (58%) | 210 (49%) |
| Moderate | 1,723 (33%) | 559 (50%) | 632 (44%) | 140 (41%) | 200 (46%) |
| High | 121 (2%) | 84 (8%) | 66 (5%) | 3 (1%) | 22 (5%) |
| Krieger experience of racism | |||||
| None | 4,508 (87%) | 615 (58%) | 900 (66%) | 174 (52%) | 240 (57%) |
| Moderate | 577 (11%) | 264 (25%) | 320 (24%) | 111 (33%) | 125 (29%) |
| High | 102 (2%) | 183 (17%) | 141 (10%) | 48 (14%) | 59 (14%) |
| Conner–Davidson resilience score | 80 ± 11 | 80 ± 13 | 78 ± 13 | 77 ± 11 | 80 ± 11 |
| Multidimensional scale of perceived social support score | 6.3 ± 1.1 | 5.8 ± 1.4 | 6.0 ± 1.2 | 6.3 ± 1.0 | 6.1 ± 1.2 |
| REALM-SF score | 6.9 ± 0.7 | 6.1 ±1.5 | 6.0 ± 2.0 | 6.7 ± 1.2 | 6.5 ± 1.1 |
| EPDS score | 5.4 ± 4.0 | 6.6 ± 4.8 | 6.3 ± 4.5 | 5.7 ± 3.8 | 6.0 ± 4.2 |
| Parents with a history of diabetes | 880 (17%) | 237 (23%) | 331 (25%) | 92 (29%) | 103 (25%) |
| Behavioral characteristics | |||||
| Planned pregnancy | 3,747 (71%) | 274 (24%) | 635 (43%) | 267 (78%) | 195 (45%) |
| Prepregnancy counseling about gestational weight gain | 2,054 (39%) | 266 (24%) | 492 (33%) | 132 (39%) | 110 (25%) |
| Physical activity | 4,081 (77%) | 654 (58%) | 800 (55%) | 243 (71%) | 312 (72%) |
| Tobacco use | 324 (6%) | 98 (9%) | 37 (3%) | 1 (0.3%) | 42 (10%) |
| Daily glycemic load (g) | 90 ± 44 | 145 ± 101 | 104 ± 61 | 79 ± 46 | 109 ± 68 |
| Daily glycemic index | 49 ± 4 | 52 ± 3 | 50 ± 4 | 49 ± 3 | 50 ± 3 |
| Total daily calories (kcal) | 1,497 (1,182, 1,882) | 1,965 (1,309, 2,845) | 1,558 (1,138, 2,096) | 1,308 (1,008, 1,698) | 1,564 (1,180, 2,093) |
| Total daily carbohydrates (g) | 184 (143, 235) | 250 (158, 376) | 197 (143, 274) | 160 (114, 201) | 204 (147, 267) |
| Total daily protein (g) | 56 (43, 72) | 63 (43, 96) | 56 (40, 78) | 48 (36, 66) | 53 (40, 75) |
| Total daily fat (g) | 57 (43, 74) | 74 (51, 113) | 58 (41, 81) | 54 (41, 70) | 61 (43, 82) |
| Total daily fiber (g) | 15 (11, 20) | 14 (9, 20) | 13 (9, 19) | 13 (9, 18) | 14 (9, 20) |
| % calories from carbohydrates | 50 ± 7 | 51 ± 8 | 51 ± 8 | 48 ± 7 | 51 ± 7 |
| % calories from protein | 15 ± 3 | 14 ± 3 | 15 ± 3 | 15 ± 3 | 14 ± 3 |
| % calories from fat | 34 ± 6 | 35 ± 6 | 34 ± 6 | 37 ± 6 | 34 ± 6 |
| Healthy Eating Index 2010 | 65 ± 12 | 54 ± 11 | 61 ± 12 | 69 ± 10 | 60 ± 14 |
| Medical characteristics | |||||
| Maternal age (y) | 28 ± 5 | 23 ± 5 | 25 ± 5 | 31 ± 5 | 26 ± 6 |
| Body mass index (kg/m2) | 24 (22, 28) | 27 (22, 34) | 25 (22, 29) | 22 (21, 25) | 25 (22, 31) |
| Waist circumference (cm) | 79 (74, 88) | 84 (75, 97) | 82 (75, 91) | 75 (70, 81) | 81 (73, 92) |
| Chronic hypertension | 104 (2%) | 50 (4%) | 28 (2%) | 3 (1%) | 13 (3%) |
| Asthma | 635 (13%) | 205 (20%) | 171 (14%) | 23 (7%) | 69 (17%) |
| Polycystic ovarian syndrome | 172 (3%) | 16 (1%) | 47 (3%) | 14 (4%) | 9 (2%) |
Abbreviation: GED, general educational development.
Notes: Data presented as n (%), mean ± standard deviation, or median (interquartile range), as appropriate.
p < 0.001 for all except PCOS (p < 0.01).
Cohen Perceived Stress: 10-question survey using 5-point Likert scale ranging from 0 to 4 with total scores ranging from 0 to 13 considered low stress, 14 to 26 moderate stress, and 27 to 40 high stress.23
Krieger Experience of Racism: 9-item self-report about lifetime experiences of discrimination attributed to race, ethnicity, or skin color with moderately defined as an affirmative answer to one or two experiences and high as an affirmative answer to three or more experiences of racism.24
Conner–Davidson Resilience: 25-item questionnaire using a Likert scale ranging from 0 to 4 for a maximum score of 100 with higher scores representing higher resilience.25
The multidimensional scale of perceived support: 12-question survey using a 7-point Likert scale ranging from 1 to 7 with total score calculated as mean score across all questions with higher scores representing higher levels of support.26
REALM-SF: 7-item word recognition test with a score of 0 representing third-grade literacy and below with inability to read most low-literacy materials and needing repeated oral instructions, 1 to 3 representing 4th to 6th-grade literacy needing low-literacy materials and not able to read prescription labels, 4 to 6 representing 7th to 8th-grade literacy with some struggle to read most patient education materials but will not be offended by low literacy materials, and 7 representing high school or more, able to read most education materials.27
In a multivariable analysis of the psychosocial factors, we found that having a high experience of discrimination, a parent with diabetes, and primarily speaking a language other than English were associated with higher odds of GDM (Table 2). There was no evidence of interaction between any of these factors and maternal race or ethnicity (p > 0.1), suggesting that these relationships did not differ among racial and ethnic subgroups.
Table 2.
Psychosocial, behavioral, and medical factors associated with GDM
| aOR (95% CI)a | |
|---|---|
| Psychosocial factors (Model 1)b | |
| Experience of discrimination | |
| Low | Ref |
| Moderate | 1.19 (0.90–1.57) |
| High | 1.64 (1.14–2.36) |
| Parents with a history of diabetes | 2.20 (1.76–2.76) |
| Non-English primary language | 2.41 (1.24–4.69) |
| Behavioral factors (Model 2)c | |
| Prepregnancy counseling about gestational weight gain | 1.41 (1.17–1.86) |
| Daily calorie intake (per 500 kcal) | 1.13 (1.05–1.22) |
| Daily fiber intake (per 5 g) | 0.89 (0.80–0.99) |
| Proportion of calories from fat (per 10%) | 1.25 (1.03–1.52) |
| Medical factors (Model 3)d | |
| Maternal age (per 5 y) | 1.55 (1.41–1.70) |
| Prepregnancy BMI (per 5 kg/m2) | 1.24 (1.06–1.45) |
| Waist circumference (per 10 cm) | 1.21 (1.04–1.40) |
| Chronic hypertension | 1.65 (1.04–2.63) |
Abbreviations: aOR, adjusted odds ratio; BMI, body mass index; CI, confidence interval; GDM, gestational diabetes mellitus.
Estimates adjusted for all factors that remained in final parsimonious models after backward selection shown above.
Psychosocial factors considered: married, poverty, insurance, education, employment, language, Cohen’s perceived stress scale, Krieger racism scale, Conner–Davidson resilience scale, multidimensional scale of perceived social support, health literacy, EPDS, family history of diabetes. No evidence of interaction with race.
Behavioral factors considered: planned pregnancy, prepregnancy weight counseling, physical activity, tobacco use, daily glycemic load, daily glycemic index, total daily calories, total carbohydrates, total protein, total fat, total dietary fiber, % calories from carbohydrates, % calories from protein, % calories from fat, Healthy Eating Index 2010. Evidence of interaction between % daily fat and race only (p = 0.03).
Medical factors considered: maternal age, prepregnancy BMI, waist circumference at the first visit, chronic hypertension, asthma, PCOS. No evidence of interaction with race (p > 0.1 for all).
With regards to behavioral factors, receipt of prepregnancy counseling about appropriate gestational weight gain, higher daily calorie intake, and a higher proportion of calories from fat was associated with higher odds of GDM (Table 2). Higher daily dietary fiber intake was associated with lower odds of GDM. There was evidence of interaction between the proportion of calories from fat and race and ethnicity, but the association among all other behavioral factors and GDM did not differ by race and ethnicity (p > 0.1).
With regards to medical factors, being older, having higher BMI, larger waist circumference, and chronic hypertension were associated with higher odds of GDM (Table 2). There was no evidence of interaction between race and ethnicity with any of these factors (p > 0.1).
When combining the factors from these three models, we found that having a parent with a history of diabetes (adjusted odds ratio [aOR]: 1.72; 95% confidence interval [CI]: 1.33–2.23), primarily speaking a language other than English (aOR:2.57; 95% CI: 1.14–5.79), higher daily caloric intake (aOR:1.18; 95% CI: 1.08–1.29), older age (aOR: 1.53; 95% CI: 1.37–1.70), higher BMI (aOR: 1.21; 95% CI: 1.02–1.44), and larger waist circumference (aOR: 1.21; 95% CI: 1.03–1.43) were independently associated with higher odds of GDM (Table 3). Having higher daily dietary fiber intake was associated with lower odds of GDM (aOR: 0.84; 95% CI:0.75–0.94). Again, there was no evidence of interaction between any of these factors and maternal race and ethnicity (p > 0.1). Stratified models by race and ethnicity similarly demonstrate that the association between the psychosocial, behavioral, and medical factors and GDM did not significantly differ by race and ethnicity. (Table 4).
Table 3.
Combined models of factors associated with GDM
| All factors combined, aOR (95% CI) | All factors with stepwise backward selection (Model 4), aOR (95% CI)a | |
|---|---|---|
| Experience of discrimination | ||
| Low | Ref | – |
| Moderate | 1.17 (0.86–1.59) | |
| High | 1.13 (0.73–1.74) | |
| Parents with a history of diabetes | 1.71 (1.32–2.21) | 1.72 (1.33–2.23) |
| Non-English primary language | 2.50 (1.11–5.62) | 2.57 (1.14–5.79) |
| Prepregnancy counseling about gestational weight gain | 1.27 (1.00–1.29) | – |
| Daily calorie intake (per 500 kcal) | 1.18 (1.08–1.29) | 1.18 (1.08–1.29) |
| Daily fiber intake (per 5 g) | 0.84 (0.74–0.94) | 0.84 (0.75–0.94) |
| Proportion of calories from fat (per 10%) | 1.00 (0.98–1.02) | – |
| Maternal age (per 5 y) | 1.49 (1.33–1.67) | 1.53 (1.37–1.70) |
| Prepregnancy BMI (per 5 kg/m2) | 1.19 (1.00–1.42) | 1.21 (1.02–1.44) |
| Waist circumference (per 10 cm) | 1.20 (0.80–2.38) | 1.21 (1.03–1.43) |
| Chronic hypertension | 1.38 (0.80–2.38) | – |
Abbreviations: aOR, adjusted odds ratio; BMI, body mass index; CI, confidence interval; GDM, gestational diabetes mellitus.
Estimates adjusted for all factors that remained in the final parsimonious model after backward selection are shown above. No evidence of interaction with race (p > 0.1 for all).
Table 4.
Combined models of factors associated with GDM stratified by race
| non-Hispanic White | non-Hispanic Black | Hispanic | Asian | Other | |
|---|---|---|---|---|---|
| Parents with a history of diabetes | 1.97 (1.39–2.78) | 1.14 (0.49–2.63) | 1.30 (0.71–2.38) | 0.91 (0.36–2.33) | 2.91 (1.04–8.14) |
| Non-English primary language | – | – | 2.03 (0.81–5.09) | 6.21 (0.51–76.0) | – |
| Daily calorie intake (per 500 kcal) | 1.25 (1.09–1.45) | 1.23 (1.00–1.52) | 1.04 (0.83–1.30) | 1.05 (0.70–1.57) | 0.92 (0.57–1.48) |
| Daily fiber intake (per 5 g) | 0.78 (0.66–0.92) | 0.85 (0.60–1.20) | 0.90 (0.70–1.17) | 1.13 (0.74–1.73) | 0.92 (0.59–1.42) |
| Maternal age (per 5 y) | 1.56 (1.33–1.83) | 1.67 (1.22–2.28) | 1.44 (1.13–1.82) | 0.97 (0.62–1.54) | 1.46 (0.92–2.32) |
| Prepregnancy BMI (per 5 kg/m2) | 1.28 (1.03–1.59) | 1.39 (0.86–2.24) | 1.10 (0.72–1.69) | 1.07 (0.55–2.09) | 1.02 (0.44–2.35) |
| Waist circumference (per 10 cm) | 1.20 (0.96–1.49) | 1.07 (0.66–1.73) | 1.43 (0.96–2.14) | 1.96 (1.01–3.79) | 1.24 (0.56–2.75) |
Abbreviations: BMI, body mass index; GDM, gestational diabetes mellitus.
Note: Estimates adjusted for all factors that remained in the final parsimonious model after backward selection shown above. No evidence of interaction with race (p > 0.1 for all).
Discussion
GDM is a global public health challenge that carries significant health implications both during pregnancy and lifelong.1–10 In this evaluation of data from a study with comprehensive, prospective ascertainment of social and behavioral determinants of health, we found that risk factors for developing GDM are multifactorial and span the psychosocial, behavioral, and medical domains. Specifically, having a parent with diabetes, non-English primary language, higher daily caloric intake, older age, higher BMI, and larger waist circumference was associated with higher odds of GDM whereas having higher daily fiber intake was associated with lower odds of GDM. Although the prevalence of these factors differed significantly by racial and ethnic identity, we identified that the relationship between these factors and GDM did not differ by race and ethnicity, suggesting these underlying factors may be more important markers of increased risk of GDM rather than race itself.
Prior studies have highlighted the racial and ethnic disparities in GDM diagnosis.11–16 However, there are limited data evaluating the underlying social determinants of health beyond race and ethnicity, also a social determinant, which may contribute to GDM. For example, in a study of 24,195 pregnant individuals using California birth certificate records, individuals from multiple Asian subgroups, including Vietnamese, Korean, Filipino, Chinese, and Asian Indian, were found to have a higher risk of GDM compared with non-Hispanic White individuals; in contrast, there was no association between Hispanic ethnicity and GDM, and non-Hispanic Black individuals had a lower risk for GDM compared with non-Hispanic White.15 Other risk factors including education, insurance, and nativity were considered along with maternal age and BMI and found to account for almost two-thirds of the risk of GDM. The authors concluded that the observed variation in population attributable fraction of GDM was due to race and ethnicity and differences in the prevalence of the other risk factors among individuals of different racial and ethnic groups; however, this study lacked the nuanced collection of patient-reported social determinants that is available in nuMoM2b. Similarly in a retrospective cohort study of 16,258 pregnant individuals, Asian, Hispanic, and Arab Americans were noted to have a higher risk of GDM compared with White individuals.14 In adjusted analyses the authors accounted for differences in BMI between racial and ethnic groups and concluded that the racial and ethnic disparities in GDM incidence persisted despite this adjustment. While recognizing that there are health disparities between racial and ethnic groups is an important first step in identifying inequities, it does not tell us anything about why the disparities exist.
An umbrella review of 30 meta-analyses demonstrated that the majority of GDM risk factors evaluated in prior studies have primarily been biomarkers, medical conditions, and BMI, with little attention to behaviors such as physical activity and sleep, and no evaluation of nonracial social determinants such as income, education, social support, and racism.13 In contrast, a more recent prospective case-control study of 100 pregnant individuals with GDM and 273 matched control individuals identified maternal stress, education, occupation, and socioeconomic status as risk factors for GDM in addition to age, BMI, exercise, and family history.17 Our study builds on these findings by including a wider array of social determinants such as experiences of discrimination and primary language, and by testing their relative importance across race and ethnicity. Awareness of these social determinants with conscious efforts to address them is needed to mitigate these disparities.32 Beyond recognizing minoritized groups with a higher incidence of GDM, screening for and intervening upon social determinants of health as recommended by ACOG to identify underlying adverse environmental conditions may reduce the risk of adverse health outcomes and correspondingly reduce disparities in those outcomes.20,22
Strengths and Limitations
Our study has several strengths. We utilized a large cohort study of nearly 9,000 pregnant individuals across multiple centers in geographically diverse regions of the United States with rigorously collected prospective data on psychosocial, behavioral, and medical factors that allowed for the evaluation of upstream social determinants of health. Multivariable modeling with a careful selection of factors facilitated the simultaneous evaluation of a diverse group of risk factors to determine which were independently associated with GDM. Furthermore, testing for effect modification by self-identified race/ethnicity allowed us to assess if the association between any of these risk factors and GDM differed by race and ethnicity. Nevertheless, this study has limitations. As in any observational study, we are unable to determine causation and cannot rule out the possibility of unmeasured confounding. The study only enrolled nulliparous individuals, which precludes our ability to assess parity, interpregnancy weight patterns, or history of GDM as risk factors. Additionally, given that the overall prevalence of GDM was low and there were fewer individuals who self-identified as Asian, our results may not be generalizable to other populations. Moreover, “Asian” and “Hispanic” are broad categories combining diverse subgroups that have been shown to have differing risks of GDM, but we did not have sufficient data to examine subgroups. Finally, this study enrolled individuals who were engaged in the healthcare system in the first trimester; further study is needed to understand the risk for GDM in individuals who enter care later, who are known to have a greater risk for adverse pregnancy outcomes yet may also have less opportunity for diagnosis and treatment.
Conclusion
In conclusion, we present data showing there are many psychosocial, behavioral, and medical factors that alter the risk for GDM, but the associations of these risk factors with GDM do not differ by race and ethnicity. Rather than relying on race and ethnicity when assessing risk for and intervening upon GDM, these data highlight the importance of fully understanding the social and behavioral context of illness, which extends far beyond race and ethnicity. Validation of these findings in other populations may support removing race/ethnicity from risk factor-based screening guidelines and replacing them with the underlying factors contributing to increased risk to improve outcomes and reduce disparities.
Key Points.
GDM varies in incidence based on race.
GDM was linked to a parent with various factors.
These factors are the history of diabetes, non-English language, and daily calorie intake.
These factors also include lower daily fiber intake, maternal age, prepregnancy BMI, and waist circumference.
The psychosocial, behavioral, and medical factors associated with GDM did not differ based on race/ethnicity.
Footnotes
Conflict of Interest
None declared.
References
- 1.Cho NH, Shaw JE, Karuranga S, et al. IDF diabetes atlas: global estimates of diabetes prevalence for 2017 and projections for 2045. Diabetes Res Clin Pract 2018;138:271–281 [DOI] [PubMed] [Google Scholar]
- 2.Metzger BE, Lowe LP, Dyer AR, et al. ; HAPO Study Cooperative Research Group. Hyperglycemia and adverse pregnancy outcomes. N Engl J Med 2008;358(19):1991–2002 [DOI] [PubMed] [Google Scholar]
- 3.Yogev Y, Xenakis EMJ, Langer O. The association between preeclampsia and the severity of gestational diabetes: the impact of glycemic control. Am J Obstet Gynecol 2004;191(05):1655–1660 [DOI] [PubMed] [Google Scholar]
- 4.Ehrenberg HM, Durnwald CP, Catalano P, Mercer BM. The influence of obesity and diabetes on the risk of cesarean delivery. Am J Obstet Gynecol 2004;191(03):969–974 [DOI] [PubMed] [Google Scholar]
- 5.Kc K, Shakya S, Zhang H. Gestational diabetes mellitus and macrosomia: a literature review. Ann Nutr Metab 2015;66(Suppl 2):14–20 [DOI] [PubMed] [Google Scholar]
- 6.Nahum Sacks K, Friger M, Shoham-Vardi I, et al. Prenatal exposure to gestational diabetes mellitus as an independent risk factor for long-term neuropsychiatric morbidity of the offspring. Am J Obstet Gynecol 2016;215(03):380.e1–380.e7 [DOI] [PubMed] [Google Scholar]
- 7.Dionne G, Boivin M, Séguin JR, Pérusse D, Tremblay RE. Gestational diabetes hinders language development in offspring. Pediatrics 2008;122(05):e1073–e1079 [DOI] [PubMed] [Google Scholar]
- 8.Zhu Y, Zhang C. Prevalence of gestational diabetes and risk of progression to type 2 diabetes: a global perspective. Curr Diab Rep 2016;16(01):7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Vounzoulaki E, Khunti K, Abner SC, Tan BK, Davies MJ, Gillies CL. Progression to type 2 diabetes in women with a known history of gestational diabetes: systematic review and meta-analysis. BMJ 2020;369:m1361. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Danyliv A, Gillespie P, O’Neill C, et al. Short- and long-term effects of gestational diabetes mellitus on healthcare cost: a cross-sectional comparative study in the ATLANTIC DIP cohort. Diabet Med 2015;32(04):467–476 [DOI] [PubMed] [Google Scholar]
- 11.Chen P, Wang S, Ji J, et al. Risk factors and management of gestational diabetes. Cell Biochem Biophys 2015;71(02):689–694 [DOI] [PubMed] [Google Scholar]
- 12.Plows JF, Stanley JL, Baker PN, Reynolds CM, Vickers MH. The pathophysiology of gestational diabetes mellitus. Int J Mol Sci 2018;19(11):3342. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Giannakou K, Evangelou E, Yiallouros P, et al. Risk factors for gestational diabetes: an umbrella review of meta-analyses of observational studies. PLoS One 2019;14(04):e0215372. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Liu B, Lamerato LE, Misra DP. A retrospective analysis of the relationship between race/ethnicity, age at delivery and the risk of gestational diabetes mellitus. J Matern Fetal Neonatal Med 2020; 33(17):2961–2969 [DOI] [PubMed] [Google Scholar]
- 15.Pu J, Zhao B, Wang EJ, et al. Racial/ethnic differences in gestational diabetes prevalence and contribution of common risk factors. Paediatr Perinat Epidemiol 2015;29(05):436–443 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Shah NS, Wang MC, Freaney PM, et al. Trends in gestational diabetes at first live birth by race and ethnicity in the US, 2011–2019. JAMA 2021;326(07):660–669 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Mishra S, Shetty A, Rao CR, Nayak S, Kamath A. Effect of maternal perceived stress during pregnancy on gestational diabetes mellitus risk: A prospective case-control study. Diabetes Metab Syndr 2020;14(05):1163–1169 [DOI] [PubMed] [Google Scholar]
- 18.Gilbert L, Gross J, Lanzi S, Quansah DY, Puder J, Horsch A. How diet, physical activity and psychosocial well-being interact in women with gestational diabetes mellitus: an integrative review. BMC Pregnancy Childbirth 2019;19(01):60. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Delanerolle G, Phiri P, Zeng Y, et al. A systematic review and meta-analysis of gestational diabetes mellitus and mental health among BAME populations. EClinicalMedicine 2021;38:101016. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.ACOG Practice Bulletin No. ACOG Practice Bulletin No. 190: Gestational Diabetes Mellitus . Obstet Gynecol 2018;131(02):e49–e64 [DOI] [PubMed] [Google Scholar]
- 21.ElSayed NA, Aleppo G, Aroda VR, et al. ; The American Diabetes Association. 2. Classification and diagnosis of diabetes: standards of medical care in diabetes-2023. Diabetes Care 2023;46(Suppl 1): S19–S40 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Gandhi M, et al. ACOG clinical practice update: screening for gestational and pregestational diabetes in pregnancy and postpartum. Obstet Gynecol 2024 [Google Scholar]
- 23.Haas DM, Parker CB, Wing DA, et al. ; NuMoM2b study. A description of the methods of the nulliparous pregnancy outcomes study: monitoring mothers-to-be (nuMoM2b). Am J Obstet Gynecol 2015;212(04):539.e1–539.e24 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Cohen S, Kamarck T, Mermelstein R. A global measure of perceived stress. J Health Soc Behav 1983;24(04):385–396 [PubMed] [Google Scholar]
- 25.Krieger N, Smith K, Naishadham D, Hartman C, Barbeau EM. Experiences of discrimination: validity and reliability of a self-report measure for population health research on racism and health. Soc Sci Med 2005;61(07):1576–1596 [DOI] [PubMed] [Google Scholar]
- 26.Connor KM, Davidson JRT. Development of a new resilience scale: the Connor-Davidson resilience scale (CD-RISC). Depress Anxiety 2003;18(02):76–82 [DOI] [PubMed] [Google Scholar]
- 27.Zimet GD, Powell SS, Farley GK, Werkman S, Berkoff KA. Psychometric characteristics of the multidimensional scale of perceived social support. J Pers Assess 1990;55(3–4):610–617 [DOI] [PubMed] [Google Scholar]
- 28.Arozullah AM, Yarnold PR, Bennett CL, et al. Development and validation of a short-form, rapid estimate of adult literacy in medicine. Med Care 2007;45(11):1026–1033 [DOI] [PubMed] [Google Scholar]
- 29.Cox JL, Holden JM, Sagovsky R. Detection of postnatal depression. Development of the 10-item Edinburgh postnatal depression scale. Br J Psychiatry 1987;150:782–786 [DOI] [PubMed] [Google Scholar]
- 30.U.S. Department of Agriculture Food and Nutrition Service. Healthy Eating Index (HEI). Accessed December 15, 2023 at: https://www.fns.usda.gov/cnpp/healthy-eating-index-hei
- 31.Guenther PM, Kirkpatrick SI, Reedy J, et al. The Healthy Eating index-2010 is a valid and reliable measure of diet quality according to the 2010 dietary guidelines for Americans. J Nutr 2014;144(03):399–407 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Grobman WA, et al. Executive summary: Workshop on social determinants of health and obstetrical outcomes, February 1–2, 2022, cosponsored by the Society for Maternal-Fetal Medicine, the Commonwealth Foundation, and the Alliance for Innovation on Maternal Health, with support from the Society for Women’s Health Research. SMFM STATEMENT. 2023. Mar;223(3): PB18–B24 [DOI] [PubMed] [Google Scholar]