Differences in disease severity and delivery gestational age between Black and White patients with hypertensive disorders of pregnancy

Abstract

OBJECTIVES:

We sought to quantify racial differences in disease severity and delivery gestational age among Black and White patients with a diagnosis of a hypertensive disorder of pregnancy.

STUDY DESIGN:

This was a retrospective cohort of all Black and White pregnant patients carrying non-anomalous singleton or twin gestations at a single tertiary healthcare system who were diagnosed with a hypertensive disorder of pregnancy, 2014–2020.

MAIN OUTCOME MEASURES:

The primary outcome was delivery <34 weeks’ gestation. Secondary outcomes were delivery <28 weeks’, preeclampsia with severe features, acute renal insufficiency, HELLP syndrome, cesarean delivery, classical cesarean delivery, small for gestational age, severe maternal morbidity, and severe composite neonatal morbidity. Outcomes were compared by race. Data were analyzed using chi square, t-test, and logistic regression.

RESULTS:

3,522 patients (29.8% Black) met inclusion criteria. Black patients had a higher odds of delivery <34 weeks’ [adjusted odds ratio (aOR) 2.22, 95% CI 1.7–2.89] and <28 weeks’ (aOR 2.39, 95% CI 1.43–3.99) and developing preeclampsia with severe features (aOR 1.92, 95% CI 1.62–2.29) than White patients. Black patients also had higher aOR of classical cesarean, severe maternal morbidity, and a small for gestational age neonate.

CONCLUSIONS:

Black patients are more likely to experience severe hypertensive disorders of pregnancy and preterm delivery compared to White patients. These findings suggest that Black-White disparities in preterm birth may be partially attributable to disparities in onset and severity of hypertensive disorders of pregnancy.

INTRODUCTION:

Racial disparities in adverse pregnancy and birth outcomes in the United States have existed for decades, but recently, there has been increased attention to the magnitude of the problem. In 2019, the rate of preterm birth among Black patients was 50% higher than White patients.(1) Disparities in prematurity are most profound at the earliest gestational ages.(2) Furthermore, as compared to patients of other races, Black patients are more likely to have pre-existing chronic hypertension and to be diagnosed with a hypertensive disorder of pregnancy; Black patients also carry a greater risk of severe maternal morbidity and mortality during pregnancy.(3–7)

Hypertensive disorders of pregnancy [including gestational hypertension, preeclampsia, preeclampsia with severe features, eclampsia, and hemolysis, elevated liver enzymes, and low platelets (HELLP) syndrome] are the leading underlying cause of medically-indicated preterm birth.(8) In accordance with national obstetric guidelines, severity of hypertensive disease dictates delivery timing.(9)

Racial differences in the rates of hypertensive disorders of pregnancy and resultant short- and long-term complications are well-established.(10) It has been repeatedly shown that Black patients who are diagnosed with hypertensive disorders of pregnancy carry an increased risk of neonatal complications (e.g., intrauterine fetal demise, fetal growth restriction) and maternal morbidity (cerebrovascular accident, pulmonary edema, and end-organ damage such as heart or renal failure) and mortality relative to White patients.(11–13) It is unknown, however, if racial differences in severity of preeclampsia may account for some of the observed racial disparities in preterm birth. (14–16)

The objective of this study was to quantify racial differences in disease severity and delivery gestational age among Black and White patients with a diagnosis of a hypertensive disorder of pregnancy.

MATERIALS AND METHODS:

This was a retrospective cohort study of all pregnant non-Hispanic Black and White patients with a diagnosis of any hypertensive disorder of pregnancy who delivered non-anomalous singleton or twin gestations at the University of North Carolina (UNC) Hospital System from 4/2014 to 3/2020. At our center, small numbers precluded inclusion and comparison of patients self-identifying with other racial groups. We identified individuals who were diagnosed with hypertensive disorders of pregnancy, defined as one or more of the following: gestational hypertension, preeclampsia, superimposed preeclampsia, hemolysis elevated liver enzymes low platelets (HELLP) syndrome, and/or eclampsia during the antepartum period, labor, or the first two weeks following delivery. Patients were identified using ICD-9 and ICD-10 codes from the pregnancy episode (main electronic medical record; includes both outpatient and inpatient encounters) and ultrasound examinations (obstetric ultrasound coding) and free text data mining of sonographer and physician comments from obstetric ultrasound reports. Patients with a diagnosis of chronic hypertension but no pregnancy-specific exacerbation of their hypertensive disease and those who did not deliver within the University of North Carolina system were excluded. Race and ethnicity were self-identified. Each individual’s home address was geocoded and assigned a census tract to evaluate the geographic distribution and representation of the cohort.

Maternal demographic, prior pregnancy history, medical and surgical history, antenatal course, and pregnancy outcomes were obtained through electronic medical record data abstraction which included a combination of ICD-9 and ICD-10 codes, direct electronic data abstraction from the medical records of included patients of discrete fields (e.g., parity, neonatal birthweight, laboratory values), and additional free text data mining of sonographer and physician comments from obstetric ultrasound reports. For those patients who received prenatal care at UNC, blood pressure readings from all prenatal encounters were obtained. When multiple blood pressure and/or laboratory values were available in the prenatal period, the first (earliest) value was considered the ‘baseline.’ The gestational age of the pregnancy was determined using a combination of sonographic and menstrual data, per standard American College of Obstetricians and Gynecologists criteria.(17) A portion of medical records (n=215) were manually verified by trained research assistants.

The primary outcome was preterm delivery <34^0/7 weeks’ gestation. Secondary outcomes included delivery <28^0/7 weeks’ gestation, preeclampsia with severe features [defined as the presence of one or more of the following: more than one severe range blood pressure (systolic ≥ 160 mmHg or diastolic ≥ 110 mmHg), acute renal insufficiency (serum creatinine ≥ 1.1 mg/dL), HELLP [hemolysis, elevated liver enzymes (specifically aspartate aminotransferase or alanine aminotransferase to more than twice the upper limit normal concentrations), and low platelet (platelet count <100,000/L)] syndrome, severe maternal morbidity, and severe composite neonatal morbidity.(9) Though each pregnancy was managed at the discretion of the patient’s provider, our institution has standardized order sets and nursing and provider protocols and education for the detection of and management of hypertensive disorders of pregnancy.

Though we included only individuals diagnosed with hypertensive disorders of pregnancy prior to discharge from the delivery hospitalization, we considered blood pressure and laboratory values for all patients up to two weeks postpartum to determine the worst severity of each individual’s hypertensive disorder. It is our routine practice for every patient with a hypertensive disorder of pregnancy to be scheduled for a blood pressure check within 2 weeks of delivery prior to hospital discharge. Further, our labor and delivery triage unit functions as an ‘obstetric emergency room’ through 14 days postpartum; any with signs or symptoms of hypertensive disorders of pregnancy who needed evaluation during this time frame would be seen on labor and delivery and not in the main emergency room.

Additional secondary outcomes were the development of at least one laboratory finding consistent with HELLP syndrome in the absence of other etiologies to explain laboratory abnormalities (e.g., patients with moderately decreased platelet count and a diagnosis of gestational thrombocytopenia were not considered to meet criteria for HELLP), cesarean delivery, classical cesarean delivery, birthweight <10% or <3% based upon sex- and gestational-age specific contemporary national norms, and severe composite neonatal morbidity (diagnosis of intraventricular hemorrhage, periventricular leukomalacia, bronchopulmonary dysplasia, necrotizing enterocolitis, retinopathy of prematurity, or death prior to initial hospital discharge).(18) Neonatal outcomes were assessed at the pregnancy level; thus, for twin gestations, if either neonate met criteria for composite morbidity, the pregnancy was considered to have met this adverse outcome.

Severe maternal morbidity, as defined by the Centers for Disease Control (CDC), was the final secondary outcome. Specifically, individuals who developed acute myocardial infarction, aneurysm, acute renal failure, adult respiratory distress syndrome, amniotic fluid embolism, cardiac arrest / ventricular fibrillation, conversion of cardiac rhythm, disseminated intravascular coagulation, eclampsia, heart failure / arrest during surgery or procedure, puerperal cerebrovascular disorders, pulmonary edema / acute heart failure, severe anesthesia complications, sepsis, shock, sickle cell disease with crisis, air and thrombotic embolism, hysterectomy, or temporary tracheostomy were considered to have severe maternal morbidity.(19) Two CDC severe maternal morbidity indicators were not considered when defining severe maternal morbidity: blood transfusion because it has low specificity for hemorrhage and mechanical ventilation because it was not reliably available in our database.

All outcomes were compared between women of self-reported Black and White race. The analysis was limited to patients who self-identified as a member of one of these two groups because they, respectively, have the worst and best obstetric outcomes in the United States.

Data were analyzed using chi-square, t-test, Fisher’s exact, Wilcoxon rank-sum, and backwards stepwise logistic regression as appropriate. Two sets of regression models were created; one that included the entire cohort and one that included the subset with prenatal care at UNC, enabling us to include laboratory and blood pressure data from routine outpatient obstetric visits. All regression models included a priori maternal self-reported race (Black or White), pregestational maternal body mass index (considered continuously), a diagnosis of either pregestational or gestational diabetes (dichotomized), smoking during pregnancy (dichotomized), male fetal sex (dichotomized) and twin gestation (dichotomized). Models evaluating cesarean delivery, classical cesarean delivery, and neonatal outcomes also included gestational age at delivery (continuous) in all initial models. In addition, models that included the entire cohort included the diagnosis of pregestational chronic hypertension (dichotomized), whereas UNC prenatal models included maternal blood pressure (dichotomized, with those who had an initial systolic blood pressure ≥ 140 mmHg or diastolic blood pressure ≥ 90 mmHg considered to have ‘high’ blood pressure) and aberrant laboratory studies (baseline hypertensive disorders of pregnancy labs prior to 20 weeks’ gestation with ≥1 of the following, considered dichotomously: serum creatinine ≥ 1.1 mg/dL, alanine transaminase or aspartate aminotransferase ≥ twice upper limit of normal, and/or platelets <100,000 per dL). Patient race and factors with p<0.20 remained in final models. Individuals with missing data for dichotomous variables were considered to not have the diagnosis or condition (biasing towards the null). There were no missing data for continuous variables included in regression models.

In addition, we planned an a priori sensitivity analysis to exclude patients with concomitant spontaneous preterm labor diagnoses. Those with preterm prolonged rupture of membranes (spontaneous rupture of membranes <37 weeks’ gestation for ≥ 72 hours prior to delivery) or advanced cervical dilation in the absence of active labor (< 28 weeks’ gestation, defined as ≥ 2 centimeters dilated by digital exam; 28.0–33.9 weeks’ gestation, defined as ≥ 4 centimeters dilated by digital exam) were considered to have concomitant spontaneous preterm labor diagnoses.

All data were analyzed using STATA version 16 (College Station, TX). This study was approved by the Institutional Review Board at the UNC-Chapel Hill under a waiver of informed consent.

RESULTS:

A total of 3,522 patients (29.8% Black and 70.2% White) met inclusion criteria. Patient demographic characteristics and past medical history are shown in Table 1. Geocoded patient residences were located in 586 different census tracts across the state of North Carolina, with a median of 4 individuals (range, 1–37) living in each census tract. Table 3 summarizes absolute rates of primary and secondary outcomes for both races. A total of 300 patients (8.5%) delivered prior to 34 weeks’, and 60 (1.7%) prior to 28 weeks’ gestation. Black patients were more likely than White patients to deliver prior to 34 (13.7% vs. 6.3%, p<0.001) and 28 (2.9% vs. 1.2%, p=0.001) weeks’ gestation. Black patients were significantly more likely than White patients to meet criteria for preeclampsia with severe features (33.0% vs 20.0%, p<0.001), experience severe maternal morbidity (5.3% vs. 2.5%, p<0.001), and deliver a baby who was diagnosed with major composite neonatal morbidity (4.5% vs. 1.9%, p<0.001). With regards to severe maternal morbidity, the most common morbidity was acute renal failure (n=34), followed by pulmonary edema (n=21), eclampsia (n=19), and peripartum cardiomyopathy (n=17). Each of these was more common among Black compared to White patients, though the incidence of peripartum cardiomyopathy was not statistically significant (0.8% vs. 0.4%, p=0.120).

Table 1.

Patient demographics and past medical history stratified by race. Data are n(%) unless specified.

Characteristic	Black patients n=1051	White patients n=2471	p-value
Patient age at estimated date of confinement, mean years ± SD a	31.3 ± 5.4	29.0 ± 5.9	<0.001
Nulliparity	379 (36.1)	1,104 (44.7)	<0.001
Multifetal gestation	31 (3.0)	118 (4.8)	0.010
Patient age greater than 35 years at estimated date of confinement	195 (18.6)	623 (25.2)	<0.001
Prepregnancy body mass index (BMI) class			<0.001
- Underweight/normal (BMI <25.0 kg/m2)	113 (10.8)	337 (13.6)
- Overweight (BMI 25.0 – 29.9 kg/m2)	181 (17.2)	541 (21.9)
- Obese (BMI ≥ 30 kg/m2)	757 (72.0)	1,593 (64.5)
Excessive gestational weight gain b	308 (29.3)	598 (24.2)	0.002
Chronic hypertension	47 (15.2)	59 (11.1)	0.090
Any diabetes mellitus	325 (13.2)	164 (15.6)	0.054
Chronic kidney disease	5 (0.48)	0 (0)	0.002
Autoimmune disease c	18 (1.7)	26 (1.1)	0.106
Tobacco use during pregnancy	128 (5.2)	65 (6.2)	0.230

^aSD, standard deviation

^bGestational weight gain was calculated adjusting for gestational age at delivery and prepregnancy body mass index. Gestational weight gain recommendations are per the Institute of Medicine which states pregnant patients should gain 1.1–4.4lb in first trimester then 2nd and 3rd trimester weight gain should depend on prepregnancy BMI as follows:

underweight = 1–1.3/week

normal weight = 0.8–1.0/week

overweight = 0.5–0.7/week

obese = 0.4–0.6/week

^cIncludes systemic lupus erythematosus, antiphospholipid antibody syndrome, rheumatoid arthritis, sarcoidosis, and autoimmune disease not otherwise specified.

Table 3.

Pregnancy outcomes, including severity of hypertensive disorder of pregnancy, stratified by patient race. Data are n(%) unless specified.

Characteristic	Black patients n=1051	White patients n=2471	p-value
Preterm birth <34 weeks’ gestation	144 (13.7)	156 (6.3)	<0.001
Preterm birth <28 weeks’ gestation	30 (2.9)	30 (1.2)	0.001
Preeclampsia with severe features	347 (33.0)	493 (20.0)	<0.001
Intrapartum systolic blood pressure(s) ≥ 160 mmHg
≥ 1 value	550 (53.4)	958 (40.7)	<0.001
≥ 10 values	226 (21.9)	291 (12.4)	<0.001
Intrapartum diastolic blood pressure(s) ≥ 110 mmHg			<0.001
≥ 1 value	334 (32.4)	445 (18.9)
≥ 10 values	35 (3.4)	28 (1.2)
Hemolysis, elevated liver enzymes, low platelets syndrome	48 (4.6)	102 (4.1)	0.555
Acute renal insufficiency (serum creatinine ≥ 1.1 mg/dL)	47 (4.7)	46 (2.0)	<0.001
Aspartate aminotransferase and/or alanine transaminase elevated ≥ twice upper limit of normal	101 (9.6)	190 (7.7)	0.058
Platelets <100,000 per dL	56 (5.3)	103 (4.2)	0.129
Eclampsia	10 (1.0)	9 (0.4)	0.03
Male fetal sex	512 (50.0)	1262 (52.7)	0.150
Underwent induction or augmentation of labor	746 (71.0)	1,753 (70.9)	0.982
Cesarean delivery	457 (43.5)	977 (39.5)	0.029
Classical cesarean delivery	43 (4.1)	29 (1.2)	<0.001
Small for gestational age at birth, <10%	135 (12.8)	204 (8.3)	<0.001
Small for gestational age at birth, <3%	35 (3.3)	40 (1.6)	0.001
Severe maternal morbidity a	56 (5.3)	61 (2.5)	<0.001
Severe composite neonatal morbidity b	47 (4.5)	47 (1.9)	<0.001

^aSevere maternal morbidity, as defined by Centers for Disease Control indicators, excluding transfusion and maternal intubation.

^bSevere composite neonatal morbidity included intraventricular hemorrhage, periventricular leukomalacia, bronchopulmonary dysplasia, necrotizing enterocolitis, retinopathy of prematurity, and death prior to initial hospital discharge.

After adjusting for confounders, Black patients had higher adjusted odds of delivery <34 weeks’ (2.22, 95% CI 1.7–2.89) and <28 weeks’ (2.39, 95% CI 1.43–3.99), and developing preeclampsia with severe features (1.92, 95% CI 1.62–2.29) when compared to White patients (Table 4). Black patients also had higher aOR of delivering by classical cesarean section, having a small for gestational age neonate, and developing severe maternal morbidity. Severe composite neonatal morbidity did not vary by race in adjusted models. These results were consistent regardless of whether prenatal care characteristics were considered (Table 4).

Table 4.

Logistic regression results. Shown are unadjusted odds ratios (OR) and adjusted odds ratios (aOR) for association between Black race and each primary and secondary outcome. Statistically significant values are bolded.

Outcome	Model 1: All patients N=3,422		Model 2: Patients with prenatal care datab n=1,314
	Unadjusted analyses	Adjusted analysesa	Unadjusted analyses	Adjusted analysesb
	OR (95% CI)	aOR (95% CI)	aOR (95% CI)	aOR (95% CI)
Preterm birth <34 weeks’ gestation	2.36 (1.85–2.99)	2.22 (1.70–2.89)	2.46 (1.67–3.64)	2.67 (1.78–4.00)
Preterm birth <28 weeks’ gestation	2.39 (1.43–3.99)	2.39 (1.43–3.99)	2.00 (0.72–5.54)	1.64 (0.61–4.44)
Preeclampsia with severe features	1.98 (1.68–2.33)	1.92 (1.62–2.29)	1.54 (1.16–2.05)	1.54 (1.16–2.06)
Hemolysis, elevated liver enzymes, and low platelets (HELLP) syndrome	1.11 (0.78–1.58)	1.30 (0.88–1.95)	1.22 (0.64–2.31)	1.40 (0.73–2.68)
Cesarean delivery	1.18 (1.02–1.36)	1.02 (0.86–1.20)	1.23 (0.97–1.56)	1.02 (0.80–0.86)
Classical cesarean delivery	3.59 (2.23–5.78)	2.11 (1.19–3.73)	3.23 (1.68–6.22)	2.10 (1.19–3.72)
Small for gestational age at birth, <10%	1.64 (1.30–2.06)	1.63 (1.27–2.10)	1.45 (1.00–2.09)	1.62 (1.26–2.09)
Small for gestational age at birth, <3%	2.09 (1.32–3.31)	1.98 (1.19–3.29)	1.75 (0.84–3.64)	2.08 (0.97–4.45)
Severe maternal morbidity c	2.22 (1.54–3.22)	2.19 (1.51–3.17)	2.44 (1.36–4.39)	2.27 (1.24–4.17)
Severe composite neonatal morbidity d	2.41 (1.60–3.61)	0.99 (0.54–1.79)	1.49 (0.73–3.02)	0.45 (0.16–1.28)

^aall initial model 1 logistic regression models included the full cohort and independent variables included chronic hypertension, pregestational body mass index, diabetes, smoking during pregnancy, male fetus, and twin gestation; the outcomes of cesarean delivery, classical cesarean delivery, and neonatal morbidity also included gestational age in the initial models. Independent variables with p<0.20 remained in final models.

^ball initial model 2 logistic regression models included the subset of patients who received prenatal care at University of North Carolina and had prenatal blood pressure and laboratory data available; independent variables included initial blood pressure ≥ 140 mmHg systolic or ≥ 90 mmHg diastolic, first lab with at least one abnormality (serum creatinine ≥ 1.1 mg/dL, alanine transaminase or aspartate aminotransferase ≥ twice upper limit of normal, or platelets <100,000 per dL), pregestational body mass index, diabetes, smoking during pregnancy, male fetus, and twin gestation; the outcomes of cesarean delivery, classical cesarean delivery, and neonatal morbidity also included gestational age in the initial models. Independent variables with p<0.20 remained in final models.

^cSevere maternal morbidity, as defined by Centers for Disease Control indicators, excluding transfusion and maternal intubation.

^dSevere composite neonatal morbidity included intraventricular hemorrhage, periventricular leukomalacia, bronchopulmonary dysplasia, necrotizing enterocolitis, retinopathy of prematurity, and death prior to initial hospital discharge.

A subset of patients [1,314/3,522 (37.3%)], received prenatal care at UNC and had serial blood pressure and laboratory data available (Table 2). The percentage of Black patients who received prenatal care at our facility (30.7%) and the percentage of Black patients who received prenatal care at our facility and delivered preterm <37 weeks’ gestation (25.3%) did not differ from the percentages in the overall cohort. Of women receiving prenatal care at our facility, 456/1,314 (34.7%) had ‘baseline’ liver and renal function laboratory studies prior to 20 weeks’ gestation; these studies were obtained at median 9.6 weeks’ gestation [Interquartile Range (IQR) 7.6, 13.5], which did not vary by race. However, Black patients were more likely than White patients to have baseline laboratory studies obtained (45.9% vs 29.7%, p<0.001). Though absolute laboratory and blood pressure values were similar by race, Black patients were more likely to have baseline laboratory abnormalities. However, the number of individuals with these perturbations was small (Table 2).

Table 2.

Prenatal care laboratory and blood pressure results, by patient race. Data are n(%) unless specified.

Characteristic		Black patients N=1051	White patients N=2471	p-value
First lab value <20 weeks gestation a	Serum creatinine, median mg/dL (IQR c )	0.62 (0.55, 0.70)	0.56 (0.51, 0.63)	<0.001
	Serum creatinine ≥ 1.1 mg/dL	6 (1.49)	0 (0)	0.001
	Alanine transaminase, median IU/L (IQR)	22 (17, 29)	24 (19, 32)	0.008
	Alanine transaminase ≥ twice upper limit of normal	7 (1.7)	8 (0.9)	0.182
	Aspartate aminotransferase, median IU/L (IQR)	19/185 (15, 24)	19 (16, 24)	0.842
	Aspartate aminotransferase ≥ twice upper limit of normal	5 (1.2)	5 (0.6)	0.188
	Platelets, median per dL (IQR)	272 (233, 337)	275 (238, 337)	0.678
	Platelets <100,000 per dL	1/185 (0.5)	0/271 (0)	>0.999
	First lab with either serum creatinine ≥ 1.1 mg/dL, alanine transaminase or aspartate aminotransferase ≥ twice upper limit of normal, or platelets <100,000 per dL	15 (3.7)	9 (1.0)	0.001
Antenatal blood pressure measurements b	Blood pressure classification at first prenatal visit			0.650
	■ Normal: systolic <120 and diastolic <80 mmHg	112 (27.8)	277 (30.1)
	■ Borderline: systolic 120–130 and diastolic <80 mmHg	98 (24.3)	202 (22.2)
	■ Grade 1: systolic 130–139 ± diastolic 90–99 mmHg	141 (35.0)	306 (33.6)
	■ Grade 2: systolic 140–159 ± diastolic 100–109 mmHg	52 (12.9)	126 (13.8)
	■ Grade 3: systolic ≥160 ± diastolic ≥ 110	9 (2.2)	10 (1.1)
	Median number of blood pressure readings obtained during prenatal care period (IQR)	9 (5, 13)	11 (6, 14)	<0.001
	Difference between initial and lowest systolic blood pressure, median mmHg (IQR)	8 (0, 16)	7 (0, 16)	0.892
	Difference between initial and lowest diastolic blood pressure, median mmHg (IQR)	7 (0, 13)	6 (0, 12)	0.400

^aFirst lab values <20 weeks’ are based on 185 Black patients and 271 White patients

^bBlood pressure parameters during pregnancy are based on 403 Black and 911 White patients

^cIQR, interquartile range

We also considered patients within the cohort who had complications during pregnancy suggestive of preterm birth due to indications unrelated to hypertension. Overall, rates of suspected spontaneous preterm birth were low; 31 (0.88%) had preterm prelabor rupture of membranes and 4 (0.11%) were diagnosed with advanced cervical dilation. Due the small number of individuals (<1%) who had overlap with potential spontaneous mechanisms of preterm birth, we did not perform a sensitivity analysis excluding patients with features consistent with concomitant spontaneous preterm birth and hypertensive disorders of pregnancy.

DISCUSSION:

We found that Black patients with hypertensive disorders of pregnancy had increased unadjusted and adjusted odds of multiple adverse maternal and neonatal outcomes compared to White patients with hypertensive disorders of pregnancy – including higher odds of preterm delivery, preeclampsia with severe features, acute renal insufficiency, and severe maternal morbidity, and delivery of a small for gestational age neonate. Further, we found that Black and White patients who ultimately developed hypertensive disorders of pregnancy had similar blood pressure profiles in early pregnancy. More Black than White patients had baseline laboratory abnormalities prior to 20 weeks’ gestation (Table 2), though the absolute number of patients with any laboratory abnormality in early pregnancy (n=24) was very small. Our findings add to the growing body of evidence that individual patient level factors are unable to explain racial disparities in outcomes among patients with hypertensive disorders of pregnancy.

While it is well documented that Black patients have higher rates of hypertensive disorders of pregnancy compared to White patients,(21, 22) few investigators have evaluated racial differences in disease severity and resultant delivery gestational age.(14–16) One study found that Black patients developed preeclampsia earlier in pregnancy than White patients, but this study was limited by a small sample size (165 patients).(16) Another study demonstrated that among patients with preeclampsia with severe features, Black patients were more likely than their White counterparts to have severe hypertension at presentation and require antihypertensives, but White patients were more likely to develop HELLP syndrome.(14) Both of these studies utilized older cohorts (2006–2009 and 1993–2003 respectively).(14, 16) Since then, obstetric guidelines, including recommendations for delivery timing for patients with any hypertensive disorder of pregnancy and definitions of severe hypertensive disease have changed.(9)

Many investigators and clinicians have postulated that racial differences in pre-pregnancy chronic conditions account for the higher rates of preeclampsia among Black patients. For example, a 2001 study concluded much of the risk of preeclampsia was attributable to racial differences in baseline rates of chronic hypertension.(23) However, this study utilized an administrative database including data from 1988 to 1996 and was unable to evaluate early pregnancy blood pressures or laboratory values. One small study of 101 patients challenged the notion that chronic hypertension is the culprit for racial disparities in preeclampsia,(24) but it is still remains a popular theory. Differences in pre-pregnancy body mass index (BMI) have also been postulated as a key reason underlying racial disparities in the incidence of preeclampsia, and one California study reported racial disparities in preeclampsia risk attenuated with increasing BMI.(25)

In contrast to the aforementioned literature, our contemporaneous study found that Black patients experienced higher rates of severe disease, delivered at earlier gestational ages (likely as a result of disease severity), and had increased maternal morbidity even after controlling for patient level factors including chronic hypertension and an obese pre-pregnancy BMI. Indeed, we hypothesize that the earlier delivery gestational age may account for the small reduction in the number of blood pressure readings obtained prenatally among Black patients (Table 2). Further, we hypothesize that the higher rate of c-section among Black patients was likely related to the higher proportion of small for gestational age neonates, severely elevated blood pressure values intrapartum, and/or acute renal insufficiency among Black patients compared to White patients, which may have contributed to reduced fetal tolerance of labor and/or a lower threshold for providers to await vaginal delivery.

Additionally, we add new knowledge that parameters including blood pressure readings at the time of routine obstetric visits prior to 20 weeks’ gestation are similar between Black and White patients who eventually develop hypertensive disorders of pregnancy. Our regression models that included only women with prenatal care data available and thus were able to incorporate initial blood pressure and lab data were very similar to the overall cohort models. These results suggest that there are no clear medical factors that are present early in pregnancy to account for the observed disparities in hypertensive disorders of pregnancies.

Our study findings challenge the hypothesis that disparities in hypertensive disorders of pregnancy can be attributed to medical disparities in pre-pregnancy health. One possible alternative explanation for the observed disparities is interpersonal, institutional, and/or structural racism. It is imperative that clinicians are aware of this distinct possibility and begin to act to mitigate the risk of racism to patient care. Clinics, hospitals, and health systems should evaluate how racism and bias are impacting clinical care so they can be addressed. For example, with regards to hypertensive disorders of pregnancy, rates of prescription of aspirin for at-risk patients, access to outpatient blood pressure cuffs for home monitoring, and timing of administration of acute antihypertensive treatment for patients with severe range blood pressures should be critically evaluated and compared by maternal race.

This study raises several questions that deserve further investigation. Additional research is needed to determine why Black patients with hypertensive disorders of pregnancy are more likely to deliver preterm, and why Black patients and their neonates are more likely to experience severe morbidity in association with this pregnancy complication. Additionally, future research should evaluate whether cumulative stress, weathering, individual experience with the health care system and prenatal care, and/or racism at any level may contribute to inequities in outcomes.(26) Elucidation of underlying etiologies may permit development of strategies to mitigate this increased risk at the individual, institutional, and structural level. This could include known strategies applied equitably (such as aspirin use) or novel strategies that have yet to be discovered.

This study has multiple strengths. We had a large, diverse sample size that included 29.8% Black patients. Furthermore, we considered factors such as patients’ baseline health including presence of chronic disease and were able to include early pregnancy blood pressure readings and laboratory values for a subset of the sample; these are often unavailable or excluded in other studies. Inclusion of patients from a single site may be considered a weakness by some, however, standardized order sets and nursing and provider protocols and education for the detection of and management of hypertensive disorders of pregnancy and management by a single maternal-fetal medicine team limits heterogeneity in management. In the State of North Carolina, pregnancy Medicaid provides care access upon confirmation of a positive pregnancy test. As evidenced by the large number of census tracts of primary residence that are represented by this cohort, UNC is a major referral center in addition to holding multiple different prenatal clinics covering a wide geographic area, spanning a range of socioeconomic statuses. Further, UNC is the major public tertiary care center for the State of North Carolina, and does not restrict access to care based on an individual’s healthcare insurer or ability to pay for services and offers a comprehensive patient assistance program for those with financial need who do not qualify for Medicaid. Additionally, because this was a robust institutional database, we were able to obtain granular details such as number of severe range blood pressures that are usually lacking in administrative databases.

However, these data must be interpreted within the context of study limitations. The study design did not allow us to elucidate reasons why these disparities exist. Additionally, we were unable to determine the precise timing and severity of each individual’s hypertensive disorder of pregnancy at initial diagnosis. We also were unable to reliably identify patients who met criteria for preeclampsia with severe features by symptoms and could not include aspirin use in our model as it was captured inconsistently. We also were not able to confirm whether the hypertensive disorder was the sole indication for iatrogenic preterm delivery. It is possible that the patients who did not receive prenatal care at UNC differed from those who did, biasing our results evaluating prenatal course. However, both “lower-risk” individuals (e.g., those who receive care at local health departments) and “higher-risk” individuals (e.g., those who are transferred from outlying facilities due to severe hypertensive disorders) did not receive prenatal care at UNC.

CONCLUSION:

Racial inequity in pregnancy outcomes is a critical problem threatening public health today. Hypertensive disease and preterm birth are two major contributors to obstetric morbidity and obstetric Black-White disparities. This study shows that hypertensive disorders of pregnancy are more severe and likely occur at earlier gestational ages among Black compared to White patients. These findings suggest that the well-described Black-White disparities in preterm birth may be partially attributable to disparities in onset and severity of hypertensive disorders of pregnancy. Importantly, our study demonstrates that these disparities are not caused by differences in pre-pregnancy maternal health. Future studies should aim to identify and apply strategies to decrease risk of, delay onset of, and/or decrease severity of hypertensive disorders specifically among Black patients.

HIGHLIGHTS:

• Black and White patients had similar blood pressure values and classification at their first prenatal visit <20 weeks’ gestation

• Black patients had increased adjusted odds of preeclampsia with severe features

• Black patients had an increased adjusted odds of birth <34 and <28 weeks’ gestation

• Nearly all preterm births in this cohort were iatrogenic rather than spontaneous

• Disparities in hypertension severity may contribute to disparities in preterm birth observed