Evaluation of US State–Level Variation in Hypertensive Disorders of Pregnancy | Hypertension | JAMA Network Open | JAMA Network
[Skip to Navigation]
Sign In
Figure 1.  Prevalence of Chronic Hypertension
Prevalence of Chronic Hypertension

The circles indicate the mean prevalence. The whisker bars indicate 95% CIs. Dashed horizontal line indicates the overall mean prevalence for all states.

Figure 2.  Prevalence of Hypertensive Disorders of Pregnancy
Prevalence of Hypertensive Disorders of Pregnancy

The circles indicate the mean prevalence. The whisker bars indicate 95% CIs. Dashed horizontal line indicates the overall mean prevalence for all states.

Figure 3.  Prevalence of Eclampsia
Prevalence of Eclampsia

The circles indicate the mean prevalence. The whisker bars indicate 95% CIs. Dashed horizontal line indicates the overall mean prevalence for all states. South Carolina and Tennessee did not report eclampsia; therefore, births in these states were excluded from the analytic sample.

Table 1.  Characteristics of Women With Chronic Hypertension, Hypertensive Disorders of Pregnancy, and Eclampsia
Characteristics of Women With Chronic Hypertension, Hypertensive Disorders of Pregnancy, and Eclampsia
Table 2.  State-Level Variation in Chronic Hypertension, Hypertensive Disorders of Pregnancy, and Eclampsia
State-Level Variation in Chronic Hypertension, Hypertensive Disorders of Pregnancy, and Eclampsia
1.
Petersen  EE, Davis  NL, Goodman  D,  et al.  Vital signs: pregnancy-related deaths, United States, 2011-2015, and strategies for prevention, 13 states, 2013-2017.   MMWR Morb Mortal Wkly Rep. 2019;68(18):423-429. doi:10.15585/mmwr.mm6818e1 PubMedGoogle ScholarCrossref
2.
Sibai  B, Dekker  G, Kupferminc  M.  Pre-eclampsia.   Lancet. 2005;365(9461):785-799. doi:10.1016/S0140-6736(05)17987-2 PubMedGoogle ScholarCrossref
3.
Ying  W, Catov  JM, Ouyang  P.  Hypertensive disorders of pregnancy and future maternal cardiovascular risk.   J Am Heart Assoc. 2018;7(17):e009382. doi:10.1161/JAHA.118.009382 PubMedGoogle Scholar
4.
Ananth  CV, Keyes  KM, Wapner  RJ.  Pre-eclampsia rates in the United States, 1980-2010: age-period-cohort analysis.   BMJ. 2013;347:f6564. doi:10.1136/bmj.f6564 PubMedGoogle ScholarCrossref
5.
Bateman  BT, Bansil  P, Hernandez-Diaz  S, Mhyre  JM, Callaghan  WM, Kuklina  EV.  Prevalence, trends, and outcomes of chronic hypertension: a nationwide sample of delivery admissions.   Am J Obstet Gynecol. 2012;206(2):134.e1-134.e8. doi:10.1016/j.ajog.2011.10.878 PubMedGoogle ScholarCrossref
6.
Fingar  KR, Mabry-Hernandez  I, Ngo-Metzger  Q, Wolff  T, Steiner  CA, Elixhauser  A. Delivery hospitalizations involving preeclampsia and eclampsia, 2005-2014: statistical brief #222. In:  Healthcare Cost and Utilization Project (HCUP) Statistical Briefs. Agency for Healthcare Research and Quality; 2006.
7.
Lee  JH, Zhang  G, Harvey  S, Nakagawa  K.  Temporal trends of hospitalization, mortality, and financial impact related to preeclampsia with severe features in Hawai’i and the United States.   Hawaii J Health Soc Welf. 2019;78(8):252-257.PubMedGoogle Scholar
8.
Wallis  AB, Saftlas  AF, Hsia  J, Atrash  HK.  Secular trends in the rates of preeclampsia, eclampsia, and gestational hypertension, United States, 1987-2004.   Am J Hypertens. 2008;21(5):521-526. doi:10.1038/ajh.2008.20 PubMedGoogle ScholarCrossref
9.
Bibbins-Domingo  K, Grossman  DC, Curry  SJ,  et al; US Preventive Services Task Force.  Screening for preeclampsia: US Preventive Services Task Force recommendation statement.   JAMA. 2017;317(16):1661-1667. doi:10.1001/jama.2017.3439 PubMedGoogle ScholarCrossref
10.
Bernstein  PS, Martin  JN  Jr, Barton  JR,  et al.  National Partnership for Maternal Safety: consensus bundle on severe hypertension during pregnancy and the postpartum period.   Obstet Gynecol. 2017;130(2):347-357. doi:10.1097/AOG.0000000000002115 PubMedGoogle ScholarCrossref
11.
American College of Obstetricians and Gynecologists.  ACOG practice bulletin No. 202: gestational hypertension and preeclampsia.   Obstet Gynecol. 2019;133(1):e1-e25.PubMedGoogle ScholarCrossref
12.
US Centers for Disease Control and Prevention. Revisions of the US standard certificates and reports. Accessed July 3, 2020. https://www.cdc.gov/nchs/nvss/revisions-of-the-us-standard-certificates-and-reports.htm
13.
National Center for Health Statistics. Guide for completing the facility worksheets for the certificate of live birth and report of fetal death. Published 2019. Updated June 9, 2020. Accessed July 3, 2020. https://www.cdc.gov/nchs/nvss/facility-worksheets-guide.htm?Sort=URL%3A%3Aasc
14.
Dietz  P, Bombard  J, Mulready-Ward  C,  et al.  Validation of selected items on the 2003 US standard certificate of live birth: New York City and Vermont.   Public Health Rep. 2015;130(1):60-70. doi:10.1177/003335491513000108 PubMedGoogle ScholarCrossref
15.
World Health Organization. Obesity: preventing and managing the global epidemic. Report of a WHO consultation. World Health Organization; 2000.
16.
Tanaka  M, Jaamaa  G, Kaiser  M,  et al.  Racial disparity in hypertensive disorders of pregnancy in New York State: a 10-year longitudinal population-based study.   Am J Public Health. 2007;97(1):163-170. doi:10.2105/AJPH.2005.068577Google ScholarCrossref
17.
Guzman  GG. Household income: 2017. US Census Bureau website. Accessed July 3, 2020. https://www.census.gov/library/publications/2018/acs/acsbr17-01.html
18.
US Census Bureau. Accessed July 3, 2020. https://data.census.gov/cedsci/
19.
US Bureau of Labor Statistics. Occupational employment statistics query system. Accessed July 3, 2020. https://www.bls.gov/oes/home.htm
20.
Larsen  K, Merlo  J.  Appropriate assessment of neighborhood effects on individual health: integrating random and fixed effects in multilevel logistic regression.   Am J Epidemiol. 2005;161(1):81-88. doi:10.1093/aje/kwi017 PubMedGoogle ScholarCrossref
21.
Larsen  K, Petersen  JH, Budtz-Jørgensen  E, Endahl  L.  Interpreting parameters in the logistic regression model with random effects.   Biometrics. 2000;56(3):909-914. doi:10.1111/j.0006-341X.2000.00909.x PubMedGoogle ScholarCrossref
22.
Merlo  J, Chaix  B, Ohlsson  H,  et al.  A brief conceptual tutorial of multilevel analysis in social epidemiology: using measures of clustering in multilevel logistic regression to investigate contextual phenomena.   J Epidemiol Community Health. 2006;60(4):290-297. doi:10.1136/jech.2004.029454 PubMedGoogle ScholarCrossref
23.
Park  S, Gillespie  C, Baumgardner  J,  et al.  Modeled state-level estimates of hypertension prevalence and undiagnosed hypertension among US adults during 2013-2015.   J Clin Hypertens (Greenwich). 2018;20(10):1395-1410. doi:10.1111/jch.13388 PubMedGoogle ScholarCrossref
24.
Sibai  BM.  Diagnosis, prevention, and management of eclampsia.   Obstet Gynecol. 2005;105(2):402-410. doi:10.1097/01.AOG.0000152351.13671.99 PubMedGoogle ScholarCrossref
25.
American College of Obstetricians and Gynecologists.  Committee opinion No. 652: magnesium sulfate use in obstetrics.   Obstet Gynecol. 2016;127(1):e52-e53. doi:10.1097/AOG.0000000000001267 PubMedGoogle ScholarCrossref
26.
Thornton  C, Dahlen  H, Korda  A, Hennessy  A.  The incidence of preeclampsia and eclampsia and associated maternal mortality in Australia from population-linked datasets: 2000-2008.   Am J Obstet Gynecol. 2013;208(6):476.e1-476.e5. doi:10.1016/j.ajog.2013.02.042 PubMedGoogle ScholarCrossref
27.
Seely  EW, Ecker  J.  Chronic hypertension in pregnancy.   Circulation. 2014;129(11):1254-1261. doi:10.1161/CIRCULATIONAHA.113.003904 PubMedGoogle ScholarCrossref
Limit 200 characters
Limit 25 characters
Conflicts of Interest Disclosure

Identify all potential conflicts of interest that might be relevant to your comment.

Conflicts of interest comprise financial interests, activities, and relationships within the past 3 years including but not limited to employment, affiliation, grants or funding, consultancies, honoraria or payment, speaker's bureaus, stock ownership or options, expert testimony, royalties, donation of medical equipment, or patents planned, pending, or issued.

Err on the side of full disclosure.

If you have no conflicts of interest, check "No potential conflicts of interest" in the box below. The information will be posted with your response.

Not all submitted comments are published. Please see our commenting policy for details.

Limit 140 characters
Limit 3600 characters or approximately 600 words
    Original Investigation
    Obstetrics and Gynecology
    October 1, 2020

    Evaluation of US State–Level Variation in Hypertensive Disorders of Pregnancy

    Author Affiliations
    • 1Department of Anesthesiology, Perioperative and Pain Medicine, Stanford University School of Medicine, Stanford, California
    • 2Department of Obstetrics and Gynecology, Stanford University School of Medicine, Stanford, California
    • 3Division of Neonatal and Developmental Medicine, Department of Pediatrics, Stanford University School of Medicine, Stanford, California
    JAMA Netw Open. 2020;3(10):e2018741. doi:10.1001/jamanetworkopen.2020.18741
    Key Points

    Question  Does the prevalence of chronic hypertension, hypertensive disorders of pregnancy, and eclampsia vary across the US by state?

    Findings  In this cross-sectional study of 3 659 553 women who had live births in the US in 2017, the median odds of eclampsia were 2.4-fold higher if the same woman delivered in a state with a higher vs lower prevalence of eclampsia. The median odds ratios were substantially lower for chronic hypertension and hypertensive disorders of pregnancy.

    Meaning  The findings of this study suggest that substantial variation among states exists in the prevalence of eclampsia across the US, despite controlling for multiple patient characteristics.

    Abstract

    Importance  Hypertensive disorders of pregnancy are important causes of maternal and perinatal morbidity in the US. However, the extent of statewide variation in the prevalence of chronic hypertension, pregnancy-induced hypertension or preeclampsia, and eclampsia in the US remains unknown.

    Objective  To examine the extent of statewide variation in the prevalence of chronic hypertension, hypertensive disorders of pregnancy (including pregnancy-induced hypertension or preeclampsia), and eclampsia in the US.

    Design, Setting, and Participants  A cross-sectional study using 2017 US birth certificate data was conducted from September 1, 2019, to February 1, 2020. A population-based sample of 3 659 553 women with a live birth delivery was included.

    Main Outcomes and Measures  State-specific prevalence of chronic hypertension, hypertensive disorders of pregnancy, and eclampsia was assessed using multilevel multivariable logistic regression, with the median odds ratio (MOR) to evaluate statewide variation.

    Results  Of the 3 659 553 women, 185 932 women (5.1%) were younger than 20 years, 727 573 women (19.9%) were aged between 20 and 24 years, 1 069 647 women (29.2%) were aged between 25 and 29 years, 1 037 307 women (28.3%) were aged between 30 and 34 years, 523 607 women (14.3%) were aged between 35 and 39 years, and 115 487 women (3.2%) were 40 years or older. Most women had Medicaid (42.8%) or private insurance (49.4%). Hawaii had the lowest adjusted prevalence of chronic hypertension (1.0%; 95% CI, 0.9%-1.2%), and Alaska had the highest (3.4%; 95% CI, 3.0%-3.9%). Massachusetts had the lowest adjusted prevalence of hypertensive disorders of pregnancy (4.3%; 95% CI, 4.1%-4.6%), and Louisiana had the highest (9.3%; 95% CI, 8.9%-9.8%). Delaware had the lowest adjusted prevalence of eclampsia (0.03%; 95% CI, 0.01%-0.09%), and Hawaii had the highest (2.8%; 95% CI, 2.2%-3.4%). The degree of statewide variation was high for eclampsia (MOR, 2.36; 95% CI, 1.88-2.82), indicating that the median odds of eclampsia were 2.4-fold higher if the same woman delivered in a US state with a higher vs lower prevalence of eclampsia. Modest variation between states was observed for chronic hypertension (MOR, 1.27; 95% CI, 1.20-1.33) and hypertensive disorders of pregnancy (MOR, 1.17; 95% CI, 1.13-1.21).

    Conclusions and Relevance  The findings of this study suggest that after accounting for patient-level and state-level variables, substantial state-level variation exists in the prevalence of eclampsia. These data can inform future public-health inquiries to identify reasons for the eclampsia variability.

    Introduction

    Hypertensive disorders that impact pregnant women, which include chronic hypertension, pregnancy-induced hypertension or preeclampsia, and eclampsia, are important causes of maternal death. Between 2011 and 2014, hypertensive disorders accounted for 7.1% of all US maternal deaths.1 Hypertensive disorders, especially severe preeclampsia and hemolysis, elevated liver enzyme levels, and low platelet count syndrome, are associated with adverse perinatal and maternal outcomes, including preterm delivery, intrauterine growth retardation, placental abruption, stroke, renal failure, and long-term cardiovascular morbidity.2,3

    In population-based studies using administrative data, the reported US prevalence of hypertensive disorders was 1.7% to 1.8% for chronic hypertension, 3.0% to 3.8% for pregnancy-induced hypertension, 3.0% to 3.4% for preeclampsia, and 0.08% for eclampsia.4-8 However, the extent of regional variation for these disorders is unknown. Identifying such variation has important implications for state and national health policy,9 national guidelines,10,11 and etiologic research.

    The main objective of this study was to examine the extent of statewide variation in the prevalence of chronic hypertension, pregnancy-induced hypertension or preeclampsia, and eclampsia in the US. We hypothesized that, after adjustment for patient-level and state-level factors, the prevalence of chronic hypertension, pregnancy-induced hypertension or preeclampsia, and eclampsia varies among states across the US. To test this hypothesis, we conducted a retrospective cross-sectional analysis of approximately 3.6 million women who had a live birth in the US in 2017.

    Methods

    We performed a population-based cross-sectional study using 2017 US birth certificate data. The study was conducted from September 1, 2019, to February 1, 2020. Because the birth certificate data we used were publicly available and deidentified, we obtained an exemption of review from the Stanford University institutional review board. We followed Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guideline for cross-sectional studies. We sourced birth certificate data for 100% of US live births in all 50 states and the District of Columbia. Information on US birth certificates follows the 2003 revised US Standard Certificate of Live Birth format and comprises demographic, medical, obstetric, and neonatal data.12 On the revised birth certificate, the National Center for Health Statistics describes hypertensive disorders as prepregnancy (or chronic), gestational (which includes pregnancy-induced hypertension or preeclampsia), and eclampsia; definitions and key words for these birth certificate variables are published by the US National Center for Health Statistics.13 To avoid ambiguity in terminology, we have relabeled gestational hypertension as hypertensive disorders of pregnancy, which is in line with definitions used by the American College of Obstetricians and Gynecologists.11 Based on specifications for reporting items on the birth certificate, birth coders can report chronic hypertension or hypertensive disorders of pregnancy individually but not together.12 Therefore, this mutual exclusivity precluded us from identifying women with chronic hypertension and preeclampsia (ie, superimposed preeclampsia). Based on data from 3 states, the positive predictive value for hypertension disorders reported on the revised birth certificate is high (99%).14

    We initially identified women who underwent a live birth in 2017. We excluded women who had missing data for chronic hypertension, hypertensive disorders of pregnancy, and eclampsia; gestational age at delivery; and any patient-level factor. Frequencies of missingness for each type of hypertensive disorder and patient factors are presented in eTable 1 in the Supplement. Given the low rates of missingness for these variables, we performed complete case analyses. Other exclusion criteria included women with gestational age at delivery reported as less than 20 weeks or more than 45 weeks.

    To examine state-level variation in the prevalence of chronic hypertension, hypertensive disorders of pregnancy, and eclampsia, we created 3 analytic samples for each type of hypertensive disorder. Because South Carolina and Tennessee did not report eclampsia, births in these states were excluded from the analytic sample for eclampsia.

    Patient-Level Factors

    The following patient-level factors were selected for inclusion in multivariable models: maternal age in years (<20, 20-24, 25-29, 30-34, 35-39, and >40 years) race/ethnicity (non-Hispanic White, non-Hispanic Black, non-Hispanic other, non-Hispanic Asian, and Hispanic), educational level (high school or less, college or associated bachelor’s degree, and master’s or doctorate degree), insurance (private insurance, Medicaid, self-pay, or other), prepregnancy body mass index using the World Health Organization categories,15 smoking history before pregnancy (presence or absence), prepregnancy diabetes (presence or absence), and number of prior live births (0 or ≥1). Race and ethnicity were examined as disparities exist in all hypertensive disorders16 and population distributions in states vary by race and ethnicity. Race and ethnicity are categorized by the National Center for Health Statistics in the birth certificate data files. In the models for hypertensive disorders of pregnancy and eclampsia, we also included gestational diabetes (presence or absence), plurality (single, twin, or triplet, or higher-order pregnancy) and trimester when women initiated prenatal care (first, second, or third trimester) as additional covariates. Because these variables are pregnancy related and cannot precede a prepregnancy diagnosis of chronic hypertension, they were not included in the chronic hypertension models. Women with missing details for prenatal care were also excluded from the analytic samples for pregnancy-related disorders and eclampsia.

    Statistical Analysis

    We performed multilevel logistic regression using the GLIMMIX procedure in SAS version 9.4 (SAS Institute), with maximum likelihood estimation based on the Laplace approximation. We fit 3 regression models as follows: a null model that included state as a random effect (model 1), a model that included patient-level factors (model 2), and a model that included patient-level and state-level factors (model 3).

    State-level factors in model 3 included 2017 median household income taken from the American Community Survey17 and percentage of families with a household income below the poverty level obtained from the American Community Survey for 2013 to 2017.18 We also included data for the number of general practitioners per 1000 deliveries (included only in the model for chronic hypertension) and number of obstetrician-gynecologists per 1000 deliveries (included in models for hypertensive disorders of pregnancy and eclampsia) using 2018 data from the US Bureau of Labor Statistics.19 We did not include data for both physician specialties in models for all hypertensive disorders because we assumed that general practitioners may manage care for women with chronic hypertension and obstetricians may manage care for women with hypertensive disorders of pregnancy.

    In each model, we quantified the between-state variation using random-effects variance by computing the median odds ratio (MOR).20-22 In this study, the MOR indicates the extent to which an individual’s probability of hypertensive disease is determined by the state and is comparable to an OR used for patient-level factors. An MOR of 1.0 implies that there are no differences between states in the odds of a woman developing a hypertensive disorder. An MOR greater than 1 implies state-level variation in the individual odds of developing a hypertensive disorder: the larger the OR, the stronger the variation. We calculated 95% CIs for each MOR value. Detailed information about the quantification of state-level adjusted prevalence and the MOR are presented in the eMethods in the Supplement. Prevalence data for individual states are presented as caterpillar plots and heat maps.

    We performed several prespecified analyses. First, we conducted stratified analyses for the prevalence of eclampsia for women with and without hypertensive disorders of pregnancy. Because chronic hypertension and hypertensive disorders of pregnancy cannot both be checked on the 2003 revised birth certificate, we could not perform stratified analyses for the prevalence of preeclampsia for women with and without chronic hypertension. To obtain finer resolution of prevalence, we also calculated crude county-level prevalence of chronic hypertension, hypertensive disorders of pregnancy, and eclampsia for counties that had equal to or more than 100 deliveries in 2017.

    Results

    We identified 3 855 500 US live births in 2017. Complete data were available on 3 659 553 women (Table 1); 185 932 women (5.1%) were younger than 20 years, 727 573 women (19.9%) were aged between 20 and 24 years, 1 069 647 women (29.2%) were aged between 25 and 29 years, 1 037 307 women (28.3%) were aged between 30 and 34 years, 523 607 women (14.3%) were aged between 35 and 39 years, and 115 487 women (3.2%) were 40 years or older. Most women had Medicaid (42.8%) or private insurance (49.4%). A flow diagram depicting women who met exclusion criteria and numbers in our final analytic samples is presented in eFigure 1 in the Supplement. The final analytic samples for assessing state-level variation included 3 659 553 deliveries for chronic hypertension, 3 588 122 deliveries for hypertensive disorders of pregnancy, and 3 461 192 deliveries for eclampsia. In these samples, the prevalence was 1.9% (95% CI, 1.9%-1.9%) for chronic hypertension, 6.5% (95% CI, 6.4%-6.5%) for hypertensive disorders of pregnancy, and 0.3% (95% CI, 0.3%-0.3%) for eclampsia; the prevalence for any hypertensive disorder was 8.6% (95% CI, 8.5%-8.6%).

    Table 1 presents the demographic characteristics of women in each analytic sample. eTable 2 in the Supplement presents adjusted ORs from multilevel models adjusting for patient-level factors. Maternal factors associated with all 3 hypertensive disorders were age 35 years or older, non-Hispanic Black race, overweight and all obesity classes, and prepregnancy diabetes. Medicaid was associated with chronic hypertension and eclampsia, while self-pay was inversely associated with chronic hypertension and hypertensive disorders of pregnancy. Gestational diabetes, twin pregnancy, and triplet or higher-order pregnancy were associated with hypertensive disorders of pregnancy and eclampsia. No prenatal care was associated with eclampsia.

    Prevalence of Hypertensive Disorders

    Caterpillar plots are presented for the unadjusted and adjusted prevalence of chronic hypertension (Figure 1), hypertensive disorders of pregnancy (Figure 2), and eclampsia (Figure 3). eFigures 2, 3, and 4 in the Supplement show heat maps for the adjusted state-level prevalence. The adjusted chronic hypertension prevalence ranged from 1.0% (95% CI, 0.9%-1.2%) in Hawaii to 3.4% (95% CI, 3.0%-3.9%) in Alaska. The adjusted hypertensive disorders of pregnancy prevalence ranged from 4.3% (95% CI, 4.1%-4.6%) in Massachusetts to 9.3% (95% CI, 8.9%-9.8%) in Louisiana. The adjusted eclampsia prevalence ranged from 0.03% (95% CI, 0.01%-0.09%) in Delaware to 2.8% (95% CI, 2.2%-3.4%) in Hawaii. Three other states (Alabama, Virginia, and Alaska) had an eclampsia prevalence greater than 1% (eTable 3 in the Supplement). Alaska and Missouri were among the 10 states with the highest adjusted prevalence for each hypertensive disorder (eTable 3 in the Supplement).

    Table 2 presents the MORs of the unadjusted and adjusted models for the 3 hypertensive disorders. In model 1, the point estimates for the MORs were greater than 1 in all analytic samples. After sequential adjustment for patient-level factors (model 2) and then state-level factors (model 3), there was a slight decrease in the MORs in all analytic samples. The degree of statewide variation was high for eclampsia (MOR, 2.54; 95% CI, 1.88-2.82), indicating that the median odds of eclampsia were 2.4-fold higher if the same woman delivered in a US state with a higher vs lower prevalence of eclampsia. Modest variation between states was observed for chronic hypertension (MOR, 1.27; 95% CI, 1.20-1.33) and hypertensive disorders of pregnancy (MOR, 1.17; 95% CI, 1.13-1.21).

    Secondary Analyses

    We performed prespecified stratified analyses to examine the crude prevalence of eclampsia for women with and without hypertensive disorders of pregnancy. eFigure 5 and eFigure 6 in the Supplement present caterpillar plots for the unadjusted statewide prevalence of eclampsia among women with and without hypertensive disorders of pregnancy. Among women with hypertensive disorders of pregnancy, 3 states had a high prevalence of eclampsia: Alaska (5.5%), Missouri (5.4%), and Alabama (5.4%). Among women without hypertensive disorders of pregnancy, the prevalence of eclampsia in Hawaii (3.2%) was substantially higher than all other states.

    After excluding counties with less than 100 deliveries in 2017, we performed prespecified analyses to examine the crude prevalence of chronic hypertension and hypertensive disorders of pregnancy in 1501 counties and of eclampsia in 1430 counties. eFigures 7, 8, and 9 in the Supplement are heat maps for the unadjusted prevalence in counties of chronic hypertension, hypertensive disorders of pregnancy, and eclampsia.

    Discussion

    Our analysis of 3 855 500 US live births in 2017 provides what is, to our knowledge, the most recent prevalence estimates of chronic hypertension, hypertensive disorders of pregnancy, and eclampsia in the US. We observed that the median odds of eclampsia were 2.4-fold higher if the same woman delivered in a US state with a higher vs lower prevalence of eclampsia. There was substantially less state-level variation in the prevalence of chronic hypertension and hypertensive disorders of pregnancy. These data suggest that public health efforts are needed to understand and reduce the degree of statewide variation in the prevalence of eclampsia.

    Sparse US data exist on regional variation in the prevalence of hypertensive disorders of pregnancy. A population-based study examining hospital discharge data reported that the prevalence of preeclampsia and eclampsia were highest in the South (3.4%) and Northeast (1.2%).8 However, no state-specific prevalence was reported. Our modeled estimates may guide national efforts for hypertension prevention and treatment, especially in states with the highest prevalence of hypertensive disorders of pregnancy and eclampsia.

    Hawaii had a very low prevalence of chronic hypertension and hypertensive disorders of pregnancy but the highest prevalence of eclampsia. Possible explanations for these discordant findings are that in women with eclampsia, hypertension may have been underreported or underdiagnosed. A study reporting state-level estimates of hypertension prevalence in US adults between 2013 and 2015 using National Health and Examination Survey data and Behavioral Risk Factor Surveillance System data reported that the prevalence of undiagnosed hypertension was highest in Hawaii (6.5%).23 Population-wide studies of blood pressure data are needed to ascertain whether underreporting or underdiagnosis explain why the prevalence of eclampsia was discordant in select states, such as Hawaii.

    To our knowledge, this analysis reveals new evidence that variability exists in the prevalence of each hypertensive disorder, especially eclampsia. Patient-level and state-level variables only modestly attenuated the point estimates for the MORs for each disorder. Our post hoc analyses of crude county-level prevalence for each hypertensive disorder did not suggest obvious geographic areas of clustering. Given the substantial statewide variation in eclampsia prevalence and the association of eclampsia with maternal morbidity (eg, disseminated intravascular coagulation, pulmonary edema, cardiac arrest, and perinatal mortality from placental abruption, prematurity, and severe fetal growth restriction),24 studies are needed to explore reasons for this variation. Potential state or local measures that may explain this variation include differences in the screening and use of magnesium sulfate in women with impending eclampsia or severe preeclampsia,25 state-level economic factors, quality and access to adequate and equitable antenatal and intrapartum care, and distributions of physical or environmental etiologic factors. Detailed analyses of prehospital and hospital-based care in the highest and lowest prevalence counties and states would be an important first step in reducing regional differences in eclampsia. Linkage of existing data sets, such as electronic health records and registries containing blood pressure data with other data on health and nonhealth exposures, may also inform how to improve hypertensive disease surveillance and treatment before and during pregnancy.

    Strengths and Limitations

    Strengths of our study include a large sample of approximately 4 million births representing all US live births in 2017, information about the states and counties where delivery occurred, and a large selection of individual-level covariates. This information allowed robust analyses of statewide variation for the 3 hypertensive disorders.

    This study has limitations. First, the observational design limits our ability to identify what state attributes influence the prevalence of each disorder. Second, owing to the classification of hypertensive disorders on the birth certificate, we could not ascertain hypertension severity or the timing of the diagnosis. However, in an observational study of preeclampsia and eclampsia in Australia, nearly half (44%) of all eclamptic seizures occurred during labor.26 Third, 17% to 25% of women with chronic hypertension develop superimposed preeclampsia27; therefore, it is unclear how these diagnoses are coded on the birth certificate. Fourth, coding variability may exist for each hypertensive disorder among states. In a study examining the accuracy of variables on birth certificates in New York City and Vermont, the sensitivity for any hypertensive disorder varied from poor in New York City (39%) to good in Vermont (76%).14 However, the specificity (New York City: 99%, Vermont: 99%), positive predictive values (New York City: 86%, Vermont: 95%), and negative predictive values (New York City: 95%, Vermont: 97%) were high. To enable assessment of future changes to hypertension screening and treatment, there is a public health need to assess state-level coding accuracy of hypertension disorders recorded on US birth certificates. Because of either low birth volume or lack of reporting, we could not estimate the degree of county-level variation in the prevalence for each hypertensive disorder. Obtaining accurate data for all US counties may help guide public health officials to narrow the geographic areas most in need of improved surveillance and treatment, especially in high prevalence states. Fifth, because the 95% CIs were relatively wide in states with a high prevalence of eclampsia, there is uncertainty about the accuracy of the prevalence estimates for these states. Nonetheless, the lower bounds of the prevalence estimate in the highest prevalence states for eclampsia (Hawaii, Alaska, Virginia, and Alabama) were above the upper bounds for most other states.

    Conclusions

    The findings of this study suggest that substantial differences exist in state-level prevalence of eclampsia within the US. Smaller differences appear to be present for chronic hypertension and hypertensive disorders of pregnancy. These data can inform future public health inquiries to identify reasons for the state-level variability in eclampsia prevalence.

    Back to top
    Article Information

    Accepted for Publication: July 17, 2020.

    Published: October 1, 2020. doi:10.1001/jamanetworkopen.2020.18741

    Open Access: This is an open access article distributed under the terms of the CC-BY License. © 2020 Butwick AJ et al. JAMA Network Open.

    Corresponding Author: Alexander J. Butwick MBBS, MS, Department of Anesthesiology, Perioperative and Pain Medicine (MC:5640), Stanford University School of Medicine, 300 Pasteur Dr, Stanford, CA 94305 (ajbut@stanford.edu).

    Author Contributions: Drs Butwick and Guo had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

    Concept and design: Butwick, Shaw, Guo.

    Acquisition, analysis, or interpretation of data: All authors.

    Drafting of the manuscript: Butwick, Shaw, Guo.

    Critical revision of the manuscript for important intellectual content: All authors.

    Statistical analysis: Butwick, Shaw, Guo.

    Administrative, technical, or material support: Butwick.

    Supervision: Butwick.

    Conflict of Interest Disclosures: Dr Druzin is cochair of the California Maternal Quality Care Collaborative preeclampsia taskforce. No other conflicts were reported.

    Funding/Support: This study was supported and funded internally by the Department of Anesthesiology, Perioperative, and Pain Medicine and Department of Obstetrics and Gynecology, Stanford University School of Medicine. Dr Shaw is currently funded by Centers for Disease Control (grant 5U01DD001226), National Institutes of Health (grants 5R01HD092316 and R01HL13984401), and the Bill and Melinda Gates Foundation.

    Role of the Funder/Sponsor: The funding sources had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

    Additional Contributions: Jason Bentley, PhD (Children's Hospital at Westmead Clinical School, University of Sydney, Australia) provided statistical input related to the quantitative measures of state-level variation. There was no financial compensation.

    References
    1.
    Petersen  EE, Davis  NL, Goodman  D,  et al.  Vital signs: pregnancy-related deaths, United States, 2011-2015, and strategies for prevention, 13 states, 2013-2017.   MMWR Morb Mortal Wkly Rep. 2019;68(18):423-429. doi:10.15585/mmwr.mm6818e1 PubMedGoogle ScholarCrossref
    2.
    Sibai  B, Dekker  G, Kupferminc  M.  Pre-eclampsia.   Lancet. 2005;365(9461):785-799. doi:10.1016/S0140-6736(05)17987-2 PubMedGoogle ScholarCrossref
    3.
    Ying  W, Catov  JM, Ouyang  P.  Hypertensive disorders of pregnancy and future maternal cardiovascular risk.   J Am Heart Assoc. 2018;7(17):e009382. doi:10.1161/JAHA.118.009382 PubMedGoogle Scholar
    4.
    Ananth  CV, Keyes  KM, Wapner  RJ.  Pre-eclampsia rates in the United States, 1980-2010: age-period-cohort analysis.   BMJ. 2013;347:f6564. doi:10.1136/bmj.f6564 PubMedGoogle ScholarCrossref
    5.
    Bateman  BT, Bansil  P, Hernandez-Diaz  S, Mhyre  JM, Callaghan  WM, Kuklina  EV.  Prevalence, trends, and outcomes of chronic hypertension: a nationwide sample of delivery admissions.   Am J Obstet Gynecol. 2012;206(2):134.e1-134.e8. doi:10.1016/j.ajog.2011.10.878 PubMedGoogle ScholarCrossref
    6.
    Fingar  KR, Mabry-Hernandez  I, Ngo-Metzger  Q, Wolff  T, Steiner  CA, Elixhauser  A. Delivery hospitalizations involving preeclampsia and eclampsia, 2005-2014: statistical brief #222. In:  Healthcare Cost and Utilization Project (HCUP) Statistical Briefs. Agency for Healthcare Research and Quality; 2006.
    7.
    Lee  JH, Zhang  G, Harvey  S, Nakagawa  K.  Temporal trends of hospitalization, mortality, and financial impact related to preeclampsia with severe features in Hawai’i and the United States.   Hawaii J Health Soc Welf. 2019;78(8):252-257.PubMedGoogle Scholar
    8.
    Wallis  AB, Saftlas  AF, Hsia  J, Atrash  HK.  Secular trends in the rates of preeclampsia, eclampsia, and gestational hypertension, United States, 1987-2004.   Am J Hypertens. 2008;21(5):521-526. doi:10.1038/ajh.2008.20 PubMedGoogle ScholarCrossref
    9.
    Bibbins-Domingo  K, Grossman  DC, Curry  SJ,  et al; US Preventive Services Task Force.  Screening for preeclampsia: US Preventive Services Task Force recommendation statement.   JAMA. 2017;317(16):1661-1667. doi:10.1001/jama.2017.3439 PubMedGoogle ScholarCrossref
    10.
    Bernstein  PS, Martin  JN  Jr, Barton  JR,  et al.  National Partnership for Maternal Safety: consensus bundle on severe hypertension during pregnancy and the postpartum period.   Obstet Gynecol. 2017;130(2):347-357. doi:10.1097/AOG.0000000000002115 PubMedGoogle ScholarCrossref
    11.
    American College of Obstetricians and Gynecologists.  ACOG practice bulletin No. 202: gestational hypertension and preeclampsia.   Obstet Gynecol. 2019;133(1):e1-e25.PubMedGoogle ScholarCrossref
    12.
    US Centers for Disease Control and Prevention. Revisions of the US standard certificates and reports. Accessed July 3, 2020. https://www.cdc.gov/nchs/nvss/revisions-of-the-us-standard-certificates-and-reports.htm
    13.
    National Center for Health Statistics. Guide for completing the facility worksheets for the certificate of live birth and report of fetal death. Published 2019. Updated June 9, 2020. Accessed July 3, 2020. https://www.cdc.gov/nchs/nvss/facility-worksheets-guide.htm?Sort=URL%3A%3Aasc
    14.
    Dietz  P, Bombard  J, Mulready-Ward  C,  et al.  Validation of selected items on the 2003 US standard certificate of live birth: New York City and Vermont.   Public Health Rep. 2015;130(1):60-70. doi:10.1177/003335491513000108 PubMedGoogle ScholarCrossref
    15.
    World Health Organization. Obesity: preventing and managing the global epidemic. Report of a WHO consultation. World Health Organization; 2000.
    16.
    Tanaka  M, Jaamaa  G, Kaiser  M,  et al.  Racial disparity in hypertensive disorders of pregnancy in New York State: a 10-year longitudinal population-based study.   Am J Public Health. 2007;97(1):163-170. doi:10.2105/AJPH.2005.068577Google ScholarCrossref
    17.
    Guzman  GG. Household income: 2017. US Census Bureau website. Accessed July 3, 2020. https://www.census.gov/library/publications/2018/acs/acsbr17-01.html
    18.
    US Census Bureau. Accessed July 3, 2020. https://data.census.gov/cedsci/
    19.
    US Bureau of Labor Statistics. Occupational employment statistics query system. Accessed July 3, 2020. https://www.bls.gov/oes/home.htm
    20.
    Larsen  K, Merlo  J.  Appropriate assessment of neighborhood effects on individual health: integrating random and fixed effects in multilevel logistic regression.   Am J Epidemiol. 2005;161(1):81-88. doi:10.1093/aje/kwi017 PubMedGoogle ScholarCrossref
    21.
    Larsen  K, Petersen  JH, Budtz-Jørgensen  E, Endahl  L.  Interpreting parameters in the logistic regression model with random effects.   Biometrics. 2000;56(3):909-914. doi:10.1111/j.0006-341X.2000.00909.x PubMedGoogle ScholarCrossref
    22.
    Merlo  J, Chaix  B, Ohlsson  H,  et al.  A brief conceptual tutorial of multilevel analysis in social epidemiology: using measures of clustering in multilevel logistic regression to investigate contextual phenomena.   J Epidemiol Community Health. 2006;60(4):290-297. doi:10.1136/jech.2004.029454 PubMedGoogle ScholarCrossref
    23.
    Park  S, Gillespie  C, Baumgardner  J,  et al.  Modeled state-level estimates of hypertension prevalence and undiagnosed hypertension among US adults during 2013-2015.   J Clin Hypertens (Greenwich). 2018;20(10):1395-1410. doi:10.1111/jch.13388 PubMedGoogle ScholarCrossref
    24.
    Sibai  BM.  Diagnosis, prevention, and management of eclampsia.   Obstet Gynecol. 2005;105(2):402-410. doi:10.1097/01.AOG.0000152351.13671.99 PubMedGoogle ScholarCrossref
    25.
    American College of Obstetricians and Gynecologists.  Committee opinion No. 652: magnesium sulfate use in obstetrics.   Obstet Gynecol. 2016;127(1):e52-e53. doi:10.1097/AOG.0000000000001267 PubMedGoogle ScholarCrossref
    26.
    Thornton  C, Dahlen  H, Korda  A, Hennessy  A.  The incidence of preeclampsia and eclampsia and associated maternal mortality in Australia from population-linked datasets: 2000-2008.   Am J Obstet Gynecol. 2013;208(6):476.e1-476.e5. doi:10.1016/j.ajog.2013.02.042 PubMedGoogle ScholarCrossref
    27.
    Seely  EW, Ecker  J.  Chronic hypertension in pregnancy.   Circulation. 2014;129(11):1254-1261. doi:10.1161/CIRCULATIONAHA.113.003904 PubMedGoogle ScholarCrossref
    ×