Geographic differences in pre-pregnancy cardiometabolic health in the United States, 2016-2019

Pre-pregnancy cardiometabolic health is independently associated with severe maternal morbidity and mortality, adverse pregnancy outcomes and long-term cardiovascular disease risk.^1–3 Given known state-level differences in pregnancy-related outcomes,² this analysis sought to describe the geographic distribution of pre-pregnancy cardiometabolic health in the US from 2016–2019.

This was a nationwide, serial cross-sectional analysis of maternal birth records from all live births in the Center for Disease Control (CDC) Natality Database from 2016–2019 when all states adopted a revised birth certificate containing information on body mass index (BMI) and pre-pregnancy diabetes. Individuals aged 20–44 years with available data on pre-pregnancy BMI, diabetes and hypertension were included. Due to censoring of limited sample size at the state-level (<10) for individuals aged ≥44 years, older persons were not included. Favorable pre-pregnancy cardiometabolic health was defined by the following health factors available in the database: normal weight [BMI 18–24.9kg/m²], absence of diabetes, and absence of hypertension. The age-specific and age-standardized prevalence of favorable cardiometabolic health, and prevalence of each risk factor (overweight [BMI≥25kg/m²], obesity [BMI≥30kg/m²], diabetes and hypertension) were calculated. Joinpoint was used to quantify the annual average percent change (APC), overall, by census region and state. Percent of individuals with live births with a high school education or less and enrolled in Medicaid were determined for each state and correlated with cardiometabolic health. A sensitivity analysis restricted to nulliparous individuals was conducted. In a secondary analysis, non-smoking status was included in the definition of favorable cardiometabolic health. This study was exempt from IRB review due to de-identified and publicly available data. All data and materials have been made publicly available through the National Vital Statistics System at CDC Wide-ranging ONline Data for Epidemiologic Research (WONDER) and can be accessed at https://wonder.cdc.gov/natality.html.

Of 14,174,625 individuals with live births, 81.4% were 20–34 years, 22.7% were Hispanic/Latina, 14% non-Hispanic Black and 52.7% non-Hispanic White. From 2016–2019, prevalence of favorable cardiometabolic health per 100 live births decreased from 43.5 (95% confidence interval: 43.3–43.6) to 40.2 (40.1–40.2) (APC = −2.6%/year [−2.9–−2.4 %/year)]) (FIGURE). Prevalence of favorable cardiometabolic health declined similarly in each 5-year age stratum from 2016–2019, and ranged from 37.1 (36.8. 37.3) in 40–44 year olds to 42.2 (42.0, 42.2) in 30–34 year olds in 2019.

Figure: State-level differences in age-standardized prevalence of favorable pre-pregnancy cardiometabolic health in 2016 (A) and 2019 (B), and percentage of individuals with a high school education or less (C) and enrolled in Medicaid (D) among 20–44 year olds with live births.

Prevalence of favorable pre-pregnancy cardiometabolic health in the United States declined from 2016 to 2019 in each state. Geographic patterns in favorable pre-pregnancy cardiometabolic health, high school education status and Medicaid enrollment were similar at the state-level in 2019.

All regions and states experienced declines in favorable cardiometabolic health; however, there was significant geographic variation. In 2019, prevalence of favorable pre-pregnancy cardiometabolic health was lower in the South (38.2 [38.1–38.3]) and Midwest (38.8 [38.6–38.9]) compared with the West (42.2 [42.1–42.]) and Northeast (43.6 [43.5–43.7]), and ranged from 31.2 (30.7–31.7) in Mississippi to 47.2 (46.7–47.6) in Utah. State-level patterns were similar for individual risk factors. There was an inverse correlation between state-level percent of favorable cardiometabolic health and state-level percent of high school education or less (r = −0.62, p < 0.01) and enrollment in Medicaid (r = −0.52, p < 0.01) at time of live birth in 2019.

In the sensitivity analysis, prevalence and trends of favorable cardiometabolic health were similar in nulliparous individuals from 2016 (48.7 [48.6–48.8]) to 2019 (45.0 [44.9–45.1]) (APC = −2.6%/year [ −3.1– −2.2%/year]) compared to the overall population. In addition, geographic and temporal patterns were similar when favorable cardiometabolic health included non-smoking status.

These data demonstrate that less than half of individuals with live births entered pregnancy with favorable cardiometabolic health. The proportion of individuals with favorable pre-pregnancy cardiometabolic health declined significantly by 3.2% between 2016–2019. There was also substantial geographic variation in cardiometabolic health with a pattern of less favorable health in the Southern and Midwestern states which reflect state-level differences in the prevalence of overweight/obesity, hypertension, and diabetes. State-level differences in educational status and Medicaid enrollment were associated with pre-pregnancy cardiometabolic health. Future work is needed to identify additional upstream social determinants of health that drive geographic differences in pre-pregnancy cardiometabolic health, and establish effective programs to eliminate disparities such the planned NIH/NHLBI Early Intervention to Promote Cardiovascular Health in Mothers and Children program.⁵ Furthermore, similar results among nulliparous individuals highlights the need to target cardiometabolic health prior to the first pregnancy.

Limitations of these analyses include the potential miscoding of pre-pregnancy cardiometabolic health factors. However, since validation studies using birth certificate data typically report low sensitivity and high specificity for cardiometabolic disease,⁵ this analysis likely overestimates the prevalence of favorable cardiometabolic health. In addition, data on cholesterol, diet, or physical activity were not available and the database does not distinguish between type 1 and type 2 diabetes. Only individuals with live births were included, which may eliminate a potentially high-risk group; however, late pregnancy losses represent <0.3% of all pregnancies. Despite these limitations, the Natality Files contain the largest database for live births in the US and allowed for the quantification of key pre-pregnancy health factors by state.

These data reveal critical deficiencies and geographic disparities in pre-pregnancy cardiometabolic health. Future research is needed to equitably improve health prior to pregnancy and quantify the potential benefits in cardiovascular disease outcomes for birthing individuals and their offspring.