Results
We analyzed a total of 69,539,875 hospitalizations from 2002 to 2017 including 674,040 women diagnosed with PDM (1.0%) and 2,960,797 (4.3%) women diagnosed with GDM.
Figure 1 displays the temporal trends in the rates of PDM and GDM in hospitalized pregnant women by race and ethnicity from 2002 to 2017. Overall, the prevalence of PDM and GDM increased over the 15-year study period from 11.1 per 1,000 hospitalizations to 12.8 per 1,000 hospitalizations and from 42.7 per 1,000 hospitalizations to 91.6 per 1,000 hospitalizations, respectively. The average annual increase in prevalence was 5.2% (95% CI [4.2, 6.2]) for GDM and 1.0% (95% CI [-0.1, 2.0]) for PDM over the study period. Hispanic women had the highest average annual percent increase in prevalence of GDM (AAPC 5.3; 95% CI [3.6, 7.1]). The prevalence of GDM among Hispanic women increased from 50% in 2002 to 109.5% in 2017. NH-Black women had the highest average annual percent increase in the prevalence of PDM over the study period (AAPC 0.9; 95% CI [0.1, 1.7]). The prevalence of PDM among NH-Black women increased from 18.9% in 2002 to 21.7% in 2017. The lowest average annual percent increase in the prevalence of GDM (AAPC 5.1; 95% CI [3.9, 6.2]) and PDM (AAPC 0.3; 95% CI [-1.4, 2.0]) was observed among NH-Whites. The prevalence of GDM among NH-Whites increased from 37.7% in 2002 to 78.9% in 2017, while the prevalence of PDM increased from 9.2% to 9.5% over the same time.
Table 1 shows the relationship between socio-demographic factors and diagnosis of PDM and GDM among hospitalized pregnant women by race/ethnicity. The age distribution of PDM and GDM was similar across groups. The prevalence of PDM and GDM increased with age in each race/ethnic group. The prevalence of PDM was highest among NH-Black hospitalized women in each age group ranging from 1.1% for those less than 24 years of age to 4.3% for those 35 to 49 years of age. The prevalence of GDM was highest among Hispanic women across all age groups ranging from 2.9% for those less than 24 years of age to 15.8% for those 35 to 49 years of age.
Among hospitalized pregnant women for which discharge status was known, there were notable differences in the prevalence of pregnancy-associated diabetes by race/ethnicity (Table 1). The highest prevalence rates for pregnancy-associated diabetes were observed among NH-Black (9.8%) and Hispanic (9.9%) women who were discharged against medical advice. Additionally, the prevalence of PDM among NH-Black mothers who died in the hospital was 7.9%, notably higher than the prevalence in other racial/ethnic groups. The prevalence of GDM among mothers who died during hospitalization was highest for Hispanic women (5.5%).
The diagnosis of PDM was more common among hospitalized women who reported a residence in a low-income zip code for all race/ethnicity groups (Table 1). In contrast, the prevalence of GDM was highest among hospitalized Hispanic (8.1%) and NH-Black (6.4%) women with residences in high-income zip codes. There were also differences in primary payer for health care services by race/ethnicity. Most NH-White hospitalized pregnant women used private insurance to pay for care, while the majority of NH-Black and Hispanic women had Medicaid as their primary payor. Among hospitalized pregnant women for which payor source was known, the prevalence of PDM and GDM was highest among women who had Medicare as their primary payor in each race and ethnic group.
With respect to hospital characteristics, the diagnosis of PDM and GDM was similar across regions for race/ethnic groups with one exception (Table 1). The prevalence of PDM among mothers increased with hospital size in each race/ethnic group but was similar across racial/ethnic groups for GDM. The prevalence of PDM and GDM was highest in urban teaching hospitals in each race/ethnic group, where most women were admitted for care.
Figure 2 depicts the rates for stillbirth among women with PDM and GDM by race and ethnicity. Overall, the rate of stillbirth was low for both groups, but the occurrence of stillbirth among hospitalized women with PDM (2.40%) was about 4.4 times higher than that for hospitalized women with GDM (0.54%). The highest frequency of stillbirth among women with PDM and GDM was in the NH-Black (0.9% and 3.13%, respectively) group and the lowest in the NH-White (0.54% and 2.40%, respectively) group.
Figure 3 depicts the association between pregnancy-associated diabetes and stillbirth by race/ethnicity. Odds ratios for stillbirth adjusted for sociodemographic hospital characteristics are provided for hospitalized women diagnosed with PDM and GDM by race/ethnicity. Unadjusted odds ratios are provided in Table S1 for PDM and in Table S2 for GDM. Compared to hospitalized pregnant women who did not have diabetes, the adjusted odds of stillbirth more than doubled for women diagnosed with PDM (OR=2.14; 95% CI [2.07,2.22]). The increase in the adjusted odds of stillbirth observed among hospitalized mothers with PDM was consistent for each race/ethnic group. The adjusted odds of stillbirth among hospitalized women with PDM was highest for Hispanic women (OR=2.41; 95% CI [2.23,2.60]) and lowest for NH Black women (OR=1.81; 95% CI [1.71,1.94]).
In contrast, the adjusted odds of stillbirth for hospitalized mothers diagnosed with GDM was lower than that for hospitalized pregnant women without diabetes. Hospitalized pregnant woman diagnosed with GDM were 49% (OR=0.51; 95% CI [0.50,0.53]) less likely to have a stillbirth. The lower adjusted odds of stillbirth observed among hospitalized women with GDM was consistent across race/ethnic groups. The adjusted odds of stillbirth among hospitalized women with GDM were the same for NH-White (OR=0.52; 95% CI [0.49,0.55]), NH-Black (OR=0.52; 95% CI [0.48 -0.56]) and Hispanic OR=0.52; 95% CI [0.49,0.56]) women. The data show that GDM is protective for stillbirth among hospitalized pregnant women when compared to their respective non-diabetic counterparts regardless of race/ethnicity.