Risk of Adverse Pregnancy Outcomes Among Pregnant Individuals With Gestational Diabetes by Race and Ethnicity in the United States, 2014-2020

Abstract

Importance: Gestational diabetes, which increases the risk of adverse pregnancy outcomes, has been increasing in frequency across all racial and ethnic subgroups in the US.

Objective: To assess whether the frequency of adverse pregnancy outcomes among those in the US with gestational diabetes changed over time and whether the risk of these outcomes differed by maternal race and ethnicity.

Design, setting, and participants: Exploratory serial, cross-sectional, descriptive study using US National Center for Health Statistics natality data for 1 560 822 individuals with gestational diabetes aged 15 to 44 years with singleton nonanomalous live births from 2014 to 2020 in the US.

Exposures: Year of delivery and race and ethnicity, as reported on the birth certificate, stratified as non-Hispanic American Indian, non-Hispanic Asian/Pacific Islander, non-Hispanic Black, Hispanic/Latina, and non-Hispanic White (reference group).

Main outcomes and measures: Maternal outcomes of interest included cesarean delivery, primary cesarean delivery, preeclampsia or gestational hypertension, intensive care unit (ICU) admission, and transfusion; neonatal outcomes included large for gestational age (LGA), macrosomia (>4000 g at birth), small for gestational age (SGA), preterm birth, and neonatal ICU (NICU) admission, as measured by the frequency (per 1000 live births) with estimation of mean annual percentage change (APC), disparity ratios, and adjusted risk ratios.

Results: Of 1 560 822 included pregnant individuals with gestational diabetes (mean [SD] age, 31 [5.5] years), 1% were American Indian, 13% were Asian/Pacific Islander, 12% were Black, 27% were Hispanic/Latina, and 48% were White. From 2014 to 2020, there was a statistically significant increase in the overall frequency (mean APC per year) of preeclampsia or gestational hypertension (4.2% [95% CI, 3.3% to 5.2%]), transfusion (8.0% [95% CI, 3.8% to 12.4%]), preterm birth at less than 37 weeks (0.9% [95% CI, 0.3% to 1.5%]), and NICU admission (1.0% [95% CI, 0.3% to 1.7%]). There was a significant decrease in cesarean delivery (-1.4% [95% CI, -1.7% to -1.1%]), primary cesarean delivery (-1.2% [95% CI, -1.5% to -0.9%]), LGA (-2.3% [95% CI, -2.8% to -1.8%]), and macrosomia (-4.7% [95% CI, -5.3% to -4.0%]). There was no significant change in maternal ICU admission and SGA. In comparison with White individuals, Black individuals were at significantly increased risk of all assessed outcomes, except LGA and macrosomia; American Indian individuals were at significantly increased risk of all assessed outcomes except cesarean delivery and SGA; and Hispanic/Latina and Asian/Pacific Islander individuals were at significantly increased risk of maternal ICU admission, preterm birth, NICU admission, and SGA. Differences in adverse outcomes by race and ethnicity persisted through these years.

Conclusions and relevance: From 2014 through 2020, the frequency of multiple adverse pregnancy outcomes in the US increased among pregnant individuals with gestational diabetes. Differences in adverse outcomes by race and ethnicity persisted.

Figure 1.. Birth Records Included for Primary Analysis in a Study of the Trends in Adverse Pregnancy Outcomes With Gestational Diabetes by Race and Ethnicity in the US From 2014 to 2020

^aPrior to implementation of the 2003 Standard Certificate of Live Birth, state birth certificate data did not include information on gestational diabetes.

Figure 2.. Maternal Outcomes (Frequency per 1000 Live Births) From 2014 to 2020 of Pregnant Individuals Aged 15 to 44 Years With Gestational Diabetes

The numbers of pregnant individuals for each outcome are as follows: cesarean delivery (n = 1 560 511), primary cesarean delivery (n = 1 236 666), preeclampsia or gestational hypertension (n = 1 560 822), intensive care unit (ICU) admission (n = 1 559 445), and transfusion (n = 1 559 445). The y-axes vary in range. Corresponding data, including the calculated annual percentage change, are shown in eTable 2 in the Supplement.

Figure 3.. Infant Outcomes (Frequency per 1000 Live Births) From 2014 to 2020 to Pregnant Individuals Aged 15 to 44 Years With Gestational Diabetes

Preterm birth included a birth at less than 37 weeks based on the best obstetric estimate for any indication. Large for gestational age was defined as a parity-specific standardized birth weight greater than the 90th percentile, and small for gestational age as less than the 10th percentile. Macrosomia was defined as a birth weight greater than 4000 g. The y-axes vary in range. The numbers of infants for each outcome are as follows: preterm birth (n = 1 560 604), neonatal intensive care unit (NICU) admission (n = 1 559 441), large for gestational age (n = 1 556 467), macrosomia (n = 1 560 366), and small for gestational age (n = 1 556 467). Corresponding data, including the calculated annual percentage change, are shown in eTable 2 in the Supplement.

Figure 4.. Maternal Outcomes and Association Between Race and Ethnicity Among Pregnant Individuals With Gestational Diabetes, US Natality Files, 2014-2020

^aMaternal outcomes of intensive care unit admission and transfusion were ascertained at delivery hospitalization. ^bAll models adjusted for age; parity; tobacco use in pregnancy; chronic hypertension; Special Supplemental Nutrition Program for Women, Infants, and Children receipt; prepregnancy body mass index; education status; insurance status; prior preterm birth; and delivery year. All models adjusted for prior cesarean delivery except primary cesarean delivery. Model for primary cesarean delivery conducted among individuals without a history of cesarean delivery.

Figure 5.. Infant Outcomes and Association Between Race and Ethnicity Among Pregnant Individuals With Gestational Diabetes, US Natality Files, 2014-2020

^aInfant outcomes of neonatal intensive care unit admission, large for gestational age (LGA), macrosomia, and small for gestational age (SGA) also adjusted for infant sex. ^bAll models adjusted for age; tobacco use in pregnancy; chronic hypertension; Special Supplemental Nutrition Program for Women, Infants, and Children receipt; prepregnancy body mass index; education status; insurance status; prior preterm birth; and delivery year. All models, except LGA and SGA, adjusted for parity, because parity-standardized LGA and SGA definitions were used.