Abstract
Background:
Chronic hypertension during pregnancy is associated with increased risk of adverse birth outcomes. In 2017, the American College of Cardiology and American Heart Association (ACC/AHA) lowered thresholds to classify hypertension in nonpregnant adults to SBP ≥ 130 mmHg and DBP ≥ 80 mmHg (i.e. stage I hypertension), resulting in an additional 4.5 million reproductive-aged women meeting criteria for hypertension. Little is known about effects of pre-pregnancy blood pressure (BP) in this range.
Objectives:
To examine the effect of pre-pregnancy maternal BP on preterm delivery.
Methods:
We analysed data from two waves of the National Longitudinal Study of Adolescent to Adult Health, including participants that had measured BP at Wave IV (2008-09) and a pregnancy that resulted in a singleton live birth between Waves IV and V (2016-18; n=2038). We categorised BP using ACC/AHA cutoffs: normal (SBP <120 mmHg and DBP < 80 mmHg), elevated (SBP 120-129 mmHg and DBP <80 mmHg), hypertension stage I (SBP 130-139 mmHg or DBP 80-89 mmHg), hypertension stage II (SBP ≥140 mmHg or DBP ≥90 mmHg). We estimated risk ratios (RR) with log-binomial regression adjusting for maternal demographics, anthropometrics, and medication use.
Results:
The prevalence of preterm delivery was 12.6%. A standard deviation (SD) increment in SBP (SD = 12.2 mmHg) and DBP (SD = 9.3 mmHg) was associated with a 14% (95% confidence interval [CI] 2, 27) and 20% (95% CI 4, 37) higher risk of preterm delivery. Compared to normotensive controls, stage I (RR 1.33, 95% CI 1.01, 1.74) and stage II (RR 1.34, 95% CI 0.89, 2.00) hypertension were associated with increased risk.
Conclusions:
We observed greater risk of preterm delivery among women with higher pre-pregnancy BP. Women with stage I hypertension during pregnancy may benefit from increased BP monitoring. Additional studies on the utility of fetal surveillance in this group are warranted.
Keywords: pre-pregnancy, chronic hypertension, blood pressure, preterm delivery
BACKGROUND
Approximately 10% of all US births are delivered preterm (i.e. <37 weeks’ gestation).1 Preterm newborns have increased risk of mortality due to inadequate in utero growth and immaturity of organs needed for survival. Also, they spend more time in the hospital resulting in greater financial and psychological stress for families.2,3 Consequences of preterm delivery extend beyond the neonatal period as babies born preterm experience poorer neurodevelopmental, socioemotional, and behavioral outcomes across development.2 The specific etiology of preterm delivery is unclear, but likely includes a complex interplay of obstetric, nutritional, environmental, and maternal factors.2
Pre-pregnancy maternal blood pressure plays a role in the development and vascular modeling of the placenta and is related to risk of preterm delivery. Chronic hypertension, defined as systolic blood pressure (SBP) ≥140 mmHg or diastolic blood pressure (DBP) ≥90 mmHg diagnosed before pregnancy or <20 weeks, is associated with two to three times increased risk of preterm delivery.4–9 Placental pathology studies suggest this may be due to altered placentation, leading to reduced functional capacity of the placenta and impaired delivery of nutrients and oxygen to the fetus.10,11 Thus, clinical recommendations suggest pregnant people with chronic hypertension receive increased surveillance of fetal growth and, potentially, antihypertensive medication.12,13
In 2017, the American College of Cardiology and American Heart Association (ACC/AHA) lowered thresholds to classify hypertension in nonpregnant adults to SBP ≥ 130 mmHg and DBP ≥80 mmHg (i.e. stage I hypertension).14 Lowering these thresholds resulted in a doubling of the prevalence of hypertension among US reproductive-aged women with approximately 4.5 million additional women meeting criteria.15–17 However, there is no clear clinical guidance on how to address stage I hypertension during pregnancy due to limited data.6,12,17 The epidemiologic studies that have examined pre-pregnancy blood pressure in this range report mixed results, with risk estimates for the association between stage I hypertension and preterm delivery ranging from 1.05 to 3.8.18–26 More research is needed to inform clinical decision making for this group.
In this paper, we leveraged data from a national cohort study of young adults to examine the effect of pre-pregnancy maternal BP on preterm delivery. First, we examined blood pressure according to the 2017 ACC/AHA categories and hypothesized that people with stage I and stage II hypertension prior to pregnancy would have greater risk of preterm delivery compared to normotensive controls. Second, we examined linear associations between blood pressure and preterm delivery and hypothesized that greater SBP and greater DBP would be associated with increased risk of preterm delivery.
METHODS
Cohort selection
The National Longitudinal Study of Adolescent to Adult Health (Add Health) is an observational cohort study designed to examine health and risk behaviors across the transition from adolescence to young adulthood.27 In 1994-1995, a sample of 80 high schools and 52 middle schools from across the US was selected using systematic random sampling. The core baseline sample included 20,745 adolescents from grades 7-12 who completed an in-school questionnaire. The study design is described in detail elsewhere.27
Including the baseline examination there have been five Add Health survey waves. Our analyses include data from Waves IV (2008-2009) and V (2016-2018) when participants were ages 24-32 and 32-42. We included non-pregnant women that participated in an in-home exam at Wave IV and reported a live singleton birth between Waves IV and V (n=2195). We excluded participants that reported conflicting or no information for gestational age (3.4%), were taking antihypertensive medication at Wave IV (3.1%), or had missing blood pressure data (3.0%), resulting in an analytic sample of 2038 women (Fig 1). A comparison between the women included in the analytic sample to all women enrolled in Add Health at the initial study baseline (Wave I) showed a slightly greater proportion of those in the analytic sample were white (66% vs 60%), were born in the US (95% vs 91%) and had parents with higher levels of education completed (34% vs 29% with college degree) (Supplemental table S1). A list of therapeutic classifications used to identify and exclude women with antihypertensive medication use at Wave IV is provided in supplemental table S2.
Figure 1.
Process of selecting the analytic sample from participants enrolled in the National Longitudinal Study of Adolescent to Adult Health (Add Health)
Exposure
At the Wave IV visit, field interviewers measured BP with a Microlife BP3MC1-PC-1B oscillometric BP monitor (MicroLife USA, Inc.; Dunedin, FL). Participants were in a seated position with both feet on the ground. Measurement of the right brachial blood pressure was repeated three times in 30-second intervals. SBP and DBP were computed as the average of the second and third readings.28 We calculated mean arterial pressure (MAP, mmHg) as , which reflects the average arterial flow through one cardiac cycle. We created a four-level BP variable according to ACC/AHA guidelines: normal (SBP <120 mmHg and DBP <80 mmHg), elevated (SBP 120-129 mmHg and DBP <80 mmHg), hypertension stage I (SBP 130-139 mmHg or DBP 80-89 mmHg), or hypertension stage II (SBP ≥140 mmHg or DBP ≥90 mmHg).14 Lastly, we also considered self-reported hypertension, which was measured with the questionnaire item, “has a doctor, nurse, or other health care provider ever told you that you have or had high BP or hypertension when you were not pregnant?”
Outcomes
At Wave V, information about live births were self-reported by participants. Preterm delivery was a binary variable determined from a questionnaire item, “A preterm delivery is one that occurs before 37 weeks in pregnancy (more than 3 weeks early). Was this baby born preterm?” For women reporting more than one live birth, only the first birth (most proximal to BP measurement) was considered because we were not able to obtain interpregnancy measurements of BP.
Covariates
Sociodemographic variables that may confound the relationship between pre-pregnancy BP and preterm delivery were selected based on literature review. These included Wave IV self-reported health insurance status (private, public, uninsured), highest education level completed (less than high school, high school to some college or vocational/technical training, bachelor’s degree or more), US region (Northeast, Midwest, South, West), gravidity, and self-reported race (White, Black, or African American, American Indian or Alaskan Native, Asian or Pacific Islander, other race). Responses for several items were collapsed into the categories indicated above due to sparse data. Using data collected at Wave V, we computed time between BP measurement and birth (in months) and maternal age at birth.
We also considered whether the associations under study were independent of pre-pregnancy maternal height and waist circumference, which were measured by trained team members at the Wave IV visit,28 and maternal smoking status (self-reported at Wave IV: never, current or former). Waist circumference, rather than body mass index, was included as a covariate because the variables were strongly correlated (r=0.85), it was more strongly associated with SBP (r=0.36) and DBP (r=0.27), and is a better discriminator of cardiovascular risk.29 We included several other variables to characterize the sample which were self-reported at Wave IV (language spoken in the home, US born, Hispanic or Latina/x ethnicity, family income (below $25k, $25k-$49k, $50k-$99k, $100k-$150k, or over $150k), engagement in health care (last routine health care visit ≤12 months ago, >12 months ago).
Statistical Analysis
We used multivariable log-binomial regression to estimate risk ratios (RR) for the associations between each BP variable and preterm delivery.30 Models were adjusted for maternal age at delivery, months between maternal BP measurement and birth, self-reported race, US region, gravidity, smoking status, education, insurance status, pre-pregnancy waist circumference, and maternal height. We first estimated associations between ACC/AHA BP categories and preterm delivery. Then, we considered SBP, DBP and MAP as continuous variables and estimated risk associated per 10 mmHg unit increase in BP. We also present standardized estimates (interpreted as risk associated with one standard deviation increase in BP variable). In exploratory analyses, we examined the potential for non-linear relationships between BP and risk of preterm delivery using an unadjusted restricted cubic spline model with knots placed at the 25th, 50th, and 75th percentiles.31 We created plots of SBP and DBP (x-axis) and individual predicted risks of preterm delivery with 95% confidence intervals (y-axis) (Supplemental Figure S1. Lastly, we considered categorical variables for SBP (<120 mmHg [reference], 120-129 mmHg, 130-139 mmHg, ≥140mmHg) and DBP (<80 mmHg [reference], 80-89 mmHg, ≥90 mmHg).
We estimated associations between ACC/AHA BP categories and risk of preterm delivery stratified by maternal age (<30 years versus ≥30 years, cutoffs decided a priori) to improve comparability with other US cohorts19,22,24,32,33 that have included samples of women younger than those included in our sample. We hypothesised that effect measure modification by maternal age may help explain heterogeneity in findings across studies. In addition to presenting stratified analyses, we estimated the relative excess risk due to interaction (RERI) using a linear relative risk model which included a product term between age category and each level of the chronic hypertension variable.34
Sensitivity analysis
First, we re-fit the multivariable regression models restricting the sample to births occurring within 3 years of the Wave IV blood pressure measurement to address potential misclassification of the exposure among participants with several years between blood pressure measurement and conception. Second, we re-fit the multivariable regression models including an additional 69 women who were taking antihypertensive medication at Wave IV. Third, we identified women who reported a pregnancy loss (self-reported miscarriage, ectopic (tubal) pregnancy, stillbirth, molar pregnancy or other pregnancy loss) between Wave IV and Wave V and were excluded from the analytic sample due to no live birth. We first assessed whether these women differed from those included and estimated inverse probability (IP) weights by modeling probability of inclusion into the analytic sample as a function of age, education, health care utilization, waist circumference, SBP, and DBP.35 We repeated the multivariable regression models using the estimated IP weights as censoring weights to evaluate potential selection bias.
All data management and analyses were conducted using SAS version 9.4 (SAS Institute Inc, Cary, NC).
Missing Data
We used multiple imputation by chained equations to address missingness.36 We used a fully conditional method to identify categorical variables, which uses a discriminant function to specify prior probabilities of group membership such that missing data are imputed as full integers.36 Estimates for regression analyses are pooled regression results generated from 50 imputations. We included all independent variables from regression models in the multiple imputation process along with preterm delivery and a few auxiliary variables that may be associated with missingness or strongly correlated with values for a missing variable. The percentage of missingness for each variable ranged from 0% to 5% (Supplemental table S3).
Ethics Approval
We obtained Add Health data through a restricted contract with the University of North Carolina at Chapel Hill, Carolina Population Center. Secondary analyses of these data were deemed exempt from Institutional Review Board review by the UC San Diego Human Research Protections Program. Cross-tabulations resulting in fewer than five cases per cell were omitted from the manuscript to avoid inadvertent disclosure of persons, families, or households.
RESULTS
The 2038 women included in the sample were, on average, 28.6 years at the Wave IV exam, and were living in all four US regions. Most were white or Black/African American, were US born citizens, and spoke English in the home. The median household income was between $50,000 and $100,000, almost all had at least a high school education, and the majority were privately insured (Table 1).
Table 1.
Summary statistics of demographic and socioeconomic characteristics at the Wave IV assessment described in the full sample and stratified by ACC/AHA blood pressure category, Add Health 2008-2009 (N = 2038)
| Full Sample | ACC / AHA Blood Pressure Category | ||||
|---|---|---|---|---|---|
| N = 2038 | Normal N = 1055 | Elevated N = 295 | Stage I HT N = 522 | Stage II HT N = 166 | |
| Wave IV age, years | 28.6 ± 1.7 | 28.5 ± 1.7 | 28.5 ± 1.7 | 28.6 ± 1.7 | 28.9 ± 1.6 |
| Maternal age at birth, years | 32.2 ± 2.7 | 32.1 ± 2.7 | 32.1 ± 2.6 | 32.3 ± 2.7 | 32.8 ± 3.0 |
| Time from Wave IV to birth, years | 3.1 (1.7, 5.2) | 3.1 (1.7, 5.1) | 3.4 (1.8, 5.2) | 3.1 (1.8, 5.2) | 3.5 (1.9, 5.7) |
| US Region | |||||
| Northeast | 329 (16.1) | 173 (16.4) | 47 (15.9) | 84 (16.1) | 25 (15.1) |
| Midwest | 479 (23.5) | 231 (21.9) | 79 (26.8) | 127 (24.3) | 42 (25.3) |
| South | 723 (35.5) | 370 (35.1) | 90 (30.5) | 195 (37.4) | 68 (41.0) |
| West | 507 (24.9) | 281 (26.6) | 79 (26.8) | 116 (22.2) | 31 (18.7) |
| Race | |||||
| White | 1341 (65.8) | 696 (66.0) | 197 (66.8) | 339 (64.9) | 109 (65.7) |
| Black or African American | 373 (18.3) | 163 (15.5) | 63 (21.4) | 108 (20.7) | 39 (23.5) |
| American Indian or Alaskan Nat. | 44 (2.2) | 24 (2.3) | 5 (1.7) | 12 (2.3) | ** |
| Asian or Pacific Islander | 130 (6.4) | 75 (7.1) | 16 (5.4) | 33 (6.3) | 6 (3.6) |
| Not listed above | 147 (7.2) | 94 (8.9) | 14 (4.8) | 30 (5.8) | 9 (5.4) |
| Missing | 3 (0.2) | ** | ** | ** | ** |
| Ethnicity | |||||
| Hispanic or Latina/x | 293 (14.4) | 178 (16.9) | 35 (11.9) | 63 (12.1) | 17 (10.2) |
| Missing | 4 (0.2) | ** | ** | ** | ** |
| Language spoken in home | |||||
| English | 1867 (91.6) | 949 (90.0) | 276 (93.6) | 486 (93.1) | 156 (94.0) |
| Spanish | 130 (6.4) | 81 (7.7) | 5.8 (17) | 24 (4.5) | 4.8 (8) |
| Other | 41 (2.0) | 25 (2.4) | ** | 12 (2.3) | ** |
| US Born Citizen | 1933 (94.9) | 988 (93.7) | 283 (95.9) | 299 (95.6) | 163 (98.2) |
| Household Annual income, $ | |||||
| <25,000 | 255 (12.5) | 118 (11.2) | 28 (9.5) | 79 (15.1) | 30 (18.1) |
| 25,000 – 50,000 | 464 (22.8) | 254 (24.1) | 71 (24.1) | 99 (19.0) | 40 (24.1) |
| 50,000 – 100,000 | 840 (41.2) | 436 (41.3) | 118 (40.0) | 222 (42.5) | 64 (38.6) |
| 100,000 – 150,000 | 255 (12.5) | 140 (13.3) | 35 (11.9) | 63 (12.1) | 17 (10.2) |
| >150,000 | 123 (6.0) | 61 (5.8) | 24 (8.1) | 31 (5.9) | 7 (4.2) |
| Missing | 101 (5.0) | 46 (4.4) | 19 (6.4) | 28 (5.4) | 8 (4.8) |
| Highest education completed | |||||
| less than high school | 67 (3.3) | 35 (3.3) | 8 (2.7) | 16 (3.1) | 8 (4.8) |
| high school to some college | 975 (47.8) | 476 (45.1) | 166 (56.3) | 254(48.7) | 79 (47.6) |
| college or more | 937 (46.0) | 515 (48.8) | 115 (39.0) | 234 (44.8) | 73 (44.0) |
| Missing | 59 (2.9) | 29 (2.8) | 6 (2.0) | 18 (3.5) | 6 (3.6) |
| Insurance status | |||||
| Uninsured | 280 (13.7) | 142 (13.5) | 44 (14.9) | 67 (12.8) | 27 (16.3) |
| Public Insurance | 159 (7.8) | 82 (7.8) | 16 (5.4) | 43 (8.2) | 18 (10.8) |
| Private Insurance | 1593 (78.2) | 829 (78.6) | 233 (79.0) | 412 (78.9) | 119 (71.7) |
| Missing | 6 (0.3) | ** | ** | ** | ** |
| Last routine health care visit | |||||
| ≤12 months ago | 1494 (73.3) | 774 (73.4) | 226 (76.6) | 383 (73.4) | 111 (66.9) |
| Over 12 months ago | 542 (26.6) | 281 (26.6) | 69 (23.4) | 137 (26.3) | 55 (33.1) |
| Missing | 2 (0.1) | ** | ** | ** | ** |
| Never pregnant | 1022 (50.2) | 522 (49.5) | 144 (48.8) | 261 (50.0) | 95 (57.2) |
| Never smoker | 1193 (58.5) | 627 (59.4) | 171 (58.0) | 313 (60.0) | 82 (49.4) |
| Waist circumference, cm | 88.3 ± 14.0 | 88.0 ± 13.6 | 98.3 ± 18.0 | 97.9 ± 17.3 | 100.5 ± 17.9 |
| Maternal height, cm | 163.5 ± 6.9 | 163.5 ± 6.9 | 165.1 ± 6.6 | 164.0 ± 6.6 | 164.3 ± 7.2 |
Alaskan Nat. = Alaskan native; ACC = American College of Cardiology; AHA = American Heart Association
For all categorical variables, we display column percentages as n (%). For years between Wave IV measurement and birth, we present median (Q1, Q3). For all other continuous variables, we present mean ± standard deviation.
Data not shown because cell count < 5
The prevalence of preterm delivery was 13%. The distributions of SBP, DBP, and MAP were normally distributed (supplemental figure S2). Of the 2038 participants, 8% met criteria for stage II hypertension at Wave IV, 26% met criteria for stage I hypertension, 15% had elevated blood pressure, and 52% were normotensive (Table 2).
Table 2.
Multivariable adjusted associations between categorical and continuous measures of pre-pregnancy maternal blood pressure with preterm delivery, Add Health (N=2038)
| Total Births Number (col %) | Preterm Delivery Number (row %) | Unadjusted Risk Ratio (95% CI) | Adjusted Risk Ratio (95% CI) | |
|---|---|---|---|---|
| Chronic Hypertension | ||||
| Normal | 1055 (51.8) | 114 (10.8) | 1.00 (Reference) | 1.00 (Reference) |
| Elevated | 295 (14.5) | 37 (12.5) | 1.16 (0.82, 1.64) | 1.08 (0.76, 1.54) |
| Hypertension – Stage I | 522 (25.6) | 80 (15.3) | 1.42 (1.09, 1.85) | 1.33 (1.01, 1.74) |
| Hypertension – Stage II | 166 (8.1) | 26 (15.7) | 1.45 (0.99, 2.15) | 1.34 (0.89, 2.00) |
| Self-reported hypertension diagnosis | ||||
| No | 1940 (95.2) | 239 (12.3) | 1.00 (Reference) | 1.00 (Reference) |
| Yes | 98 (4.8) | 18 (18.4) | 1.49 (0.97, 2.30) | 1.32 (0.85, 2.03) |
|
| ||||
| Systolic Blood Pressure | ||||
| Continuous, per 1 SD (12.2 mmHg) | --- | --- | 1.19 (1.07, 1.31) | 1.14 (1.02, 1.27) |
| Continuous, per 10 mmHg | --- | --- | 1.15 (1.06, 1.26) | 1.11 (1.02, 1.22) |
| Categorical | ||||
| < 120 mmHg | 1154 (56.6) | 129 (11.2) | 1.00 (Reference) | 1.00 (Reference) |
| 120 - 129 mmHg | 561 (27.5) | 79 (14.1) | 1.26 (0.97, 1.64) | 1.19 (0.91, 1.56) |
| 130 - 139 mmHg | 231 (11.3) | 35 (15.2) | 1.36 (0.96, 1.92) | 1.22 (0.85, 1.74) |
| ≥ 140 mmHg | 92 (4.5) | 14 (15.2) | 1.36 (0.82, 2.26) | 1.28 (0.76, 2.15) |
|
| ||||
| Diastolic Blood Pressure | ||||
| Continuous, per 1SD (9.3 mmHg) | --- | --- | 1.25 (1.13, 1.39) | 1.20 (1.04, 1.37) |
| Continuous, per 10 mmHg | --- | --- | 1.27 (1.14, 1.42) | 1.21 (1.05, 1.41) |
| Categorical | ||||
| < 80 mmHg | 1446 (71.0) | 166 (11.5) | 1.00 (Reference) | 1.00 (Reference) |
| 80-89 mmHg | 460 (22.6) | 68 (14.8) | 1.29 (0.99, 1.67) | 1.26 (0.96, 1.64) |
| ≥ 90 mmHg | 132 (6.5) | 23 (17.4) | 1.52 (1.02, 2.26) | 1.43 (0.96 2.13) |
| Mean Arterial Pressure | ||||
| Continuous, per 1SD (10 mmHg) | --- | --- | 1.25 (1.12, 1.39) | 1.10 (0.93, 1.30) |
SD = standard deviation
Adjusted risk ratios were estimated from models that included months between exam and birth, maternal age, race, US region, gravidity, Wave IV smoking status, insurance status, pre-pregnancy waist circumference, and height
Greater SBP, DBP, and MAP were associated with increased risk of preterm delivery with similar point estimates across variables (Table 2). A one standard deviation increase in SBP was associated with a 14% increased risk (95% CI 2, 27), a one standard deviation increase in DBP was associated with a 20% increased risk (95% CI 4, 37), and a one standard deviation increase in MAP was associated with a 10% (95% CI −7, 30) increased risk of preterm delivery (Table 2). For ACC/AHA BP category, stage I hypertension and stage II hypertension were associated with 1.33 (95% CI 1.01, 1.74) and 1.34 (95% CI 0.89, 2.00) times greater risk of preterm delivery compared to normal BP. Limited sample size in the stage II hypertension category resulted in wide confidence intervals. We found slightly larger risk ratios for stage I and stage II hypertension among participants aged 30 years or more compared to those with maternal age <30 years. However, there was no evidence of an additive interaction (Table 3).
Table 3.
Risk ratios examining association between chronic hypertension and preterm delivery stratified by maternal age at birth, Add Health (N=2038)
| Full sample (n=2038) | Maternal age < 30 (n=462) | Maternal age ≥ 30 (n=1576) | ||||
|---|---|---|---|---|---|---|
|
| ||||||
| Adj. Risk Ratio (95%CI) | PTD/n | Adj. Risk Ratio (95% CI) | PTD/n | Adj. Risk Ratio (95%CI) | RERI (95% CI) | |
| Normal | 1.00 (Reference) | 29/256 | 1.00 (Reference) | 85/799 | 1.00 (Reference) | |
| Elevated | 1.08 (0.76, 1.54) | 11/68 | 1.29 (0.61, 2.70) | 26/227 | 1.01 (0.66, 1.55) | −0.40 (−1.62, 0.82) |
| Hypertension Stage I | 1.33 (1.01, 1.74) | 16/111 | 1.21 (0.63, 2.33) | 64/411 | 1.38 (1.00, 1.89) | 0.18 (−0.62, 0.98) |
| Hypertension Stage II | 1.34 (0.89, 2.00) | 4/27 | 1.24 (0.41, 3.73) | 22/139 | 1.40 (0.90, 2.20) | 0.17 (−1.35, 1.69) |
PTD = Preterm Delivery; RERI = Relative Excess Risk Due to Interaction (reference group = maternal age <30); Adj. = Adjusted
Stratified adjusted risk ratios were estimated from models that included months between exam and birth, race, US region, gravidity, Wave IV smoking status, insurance status, pre-pregnancy waist circumference, and height
Our study results did not meaningfully change after including women that were taking antihypertensive medications at Wave IV (Supplemental table S4) or after restricting to births that occurred within 3 years of blood pressure measurement (Supplemental table S5). We report small differences between women included in the analytic sample and those excluded from the analysis due to a pregnancy loss between Waves IV and V (Supplemental table S6). We found no meaningful differences in effect estimates after weighting the sample based on propensity to be included (Supplemental table S7).
COMMENT
Principal Findings
Using data from women enrolled in Add Health, we observed greater risk of preterm delivery among women with higher pre-pregnancy BP. The magnitude of associations between BP and preterm delivery were slightly stronger when comparing pre-pregnancy DBP (20% increase per SD) to MAP (10% increase per SD) and SBP (14% increase per SD). Additionally, we observed that women with pre-pregnancy stage I hypertension and stage II hypertension had similarly increased risk of preterm delivery compared to those with normal blood pressure. These associations were independent of several sociodemographic factors and pre-pregnancy waist circumference and were robust to sensitivity analyses.
Strengths of the study
The study leveraged existing data from a well-characterized sample of women living in all four US regions. Importantly, the data included both participant self-report of hypertension diagnosis and measured BP prior to pregnancy. SBP and DBP were measured using standardized protocols by trained research staff in the participant’s home.28 This allowed for examination of SBP, DBP, and MAP as continuous variables, rather than limiting our analyses to the study of diagnosed hypertension or blood pressure categories. Lastly, the prospective design allowed for investigation of selection bias due to restriction on live birth because we were able to identify a subgroup of women that had pregnancies which did not end with a live birth during the study period.
Limitations of the data
Several limitations with the study design and variable measurement should be considered. Because blood pressure was measured at Wave IV (ages 24-32), we only analyzed data from births occurring after that assessment. As a result, the average maternal age at birth in our sample (32 years) is higher than the national average (27 years).1 Also, we did not use Add Health sampling weights due to small cell sizes and unstable effect estimates. Thus, the descriptive statistics and prevalence estimates presented in this paper are not considered to be nationally representative estimates. The sample had a small proportion of Hispanic or Latina/x participants, was predominantly English speaking, US born, highly educated and with adequate prenatal care. These analyses must be replicated in samples with varying demographics to improve generalisability.
Additionally, the time between Wave IV blood pressure measurement and birth was about 3 years (interquartile range: 1.7 to 5.2 years). This could have resulted in misclassification of our independent variable if, for example, women moved from stage I to stage II hypertension by the time of conception. However, we conducted a sensitivity analysis restricting our sample to people with a blood pressure measurement within 3 years of birth and found little change in estimates. Additionally, prior literature found no change in SBP and only 1-2 unit increases in DBP in the five years prior to childbearing,33 thus we expect the amount of misclassification in our sample to be low.
Although we ran sensitivity analyses assessing potential selection bias due to restriction on live births,37,38 we were not able to assess potential selection bias due to women having difficulty getting pregnant because this information was not measured. If women with stage I or II hypertension had more difficulty getting pregnant, and thus were not included in our analytic sample, we would expect the presented effect estimates to be biased towards the null.
Finally, spontaneous preterm delivery versus clinician-initiated preterm deliveries have distinct aetiologies that we were not able to evaluate or disentangle due to no data on subtype. Research shows that increases in preterm delivery associated with chronic hypertension are ascribed to clinician-initiated, and not spontaneous, preterm deliveries.39 This appears to hold when examining associations between stage I hypertension and preterm delivery as well.22 Further, preterm delivery was self-reported, so misclassification of the study outcome is possible, though maternal report of preterm delivery has good specificity and sensitivity.40
Interpretation
The findings for preterm delivery are mostly consistent with a growing body of literature examining the same associations. In the Young Finns study, standard deviation increases in SBP and DBP were associated with 28% and 19% increased risk of preterm delivery, respectively.41 In a US cohort of women in Louisiana, Mississippi and Texas, the associations between pre-pregnancy BP and preterm delivery were similar in magnitude but imprecise.32 A study from Shanghai, China analysed MAP continuously and reported a 11% increase per 10-mmHg unit change in MAP, which is like the estimate presented in our study.
Our findings relating SBP and DBP to preterm delivery were dissimilar to results from the CARDIA study,33 a US cohort that reported no differences in trajectories of pre-pregnancy BP when comparing women with preterm versus term deliveries. The authors of that report suggested the null finding could relate to the relatively young age of participants included (mean age ~ 24 years), when “differences in metabolic and vascular function…might be too minimal to detect or have yet to appear.”33 Notably, mean SBP and DBP in the CARDIA study were about 12 mmHg and 10 mmHg lower compared to our Add Health sample. In response to the possibility that differing maternal age may contribute to heterogeneity across studies, we conducted post hoc analyses stratified by maternal age at birth, finding slightly larger associations among women 30 years and older compared to those under 30 years at child’s birthdate. Age-specific associations should be investigated in future research.
Our finding that stage I hypertension is associated with increased risk of preterm delivery is consistent with what has been reported in other studies18,20–23,25,26 In aggregate, these findings suggest that women falling in this category may benefit from increased BP monitoring. We note that current practice in pregnancy is to tolerate permissive mild hypertension to maintain adequate placental perfusion.13 As such, questions about potential effects of BP lowering need to be tested in further studies. Clinical trials examining treatment of hypertension at lower thresholds are needed to answer questions surrounding potential benefits for maternal and perinatal outcomes.15 Women who become normotensive by virtue of medication use may still incur identical risks of preeclampsia and indicated preterm delivery.13,42–44 Additional studies assessing the relationship between stage I hypertension and utility of fetal surveillance could be warranted. Decisions to increase antenatal testing and ultrasound surveillance must consider potential harms of increased surveillance including costs and utilization of healthcare resources, in addition to potential benefits. Lastly, a change of thresholds used to classify hypertension will result in cases of gestational hypertension reclassified as chronic hypertension, affecting clinical decisions surrounding timing of delivery and induction.45
Conclusions
Data from this well-characterized sample of women, living in all four US census regions adds to the literature investigating the effects of pre-pregnancy BP on preterm delivery. Our data suggest that elevated pre-pregnancy BP below traditionally held clinical thresholds may confer higher risk for preterm delivery, and those women with stage I hypertension may benefit from increased BP monitoring throughout pregnancy. Optimizing women’s health early in life is necessary for not only an individual’s lifelong health, but for the health of the next generation.
Supplementary Material
SYNOPSIS.
Study question:
How does pre-pregnancy blood pressure (BP) relate to risk of preterm delivery?
What’s already known:
Women with chronic hypertension, traditionally defined as systolic BP ≥140 mmHg or diastolic BP ≥90 mmHg diagnosed before pregnancy or before 20 weeks of gestation, have two to three times the risk of delivering preterm.
What this study adds:
We found associations between higher pre-pregnancy systolic and diastolic BP and greater risk of preterm delivery across the full spectrum of BP, suggesting increased risk at levels below traditional clinical cutoffs.
ACKNOWLEDGEMENTS
This research uses data from Add Health, a program project designed by J. Richard Udry, Peter S. Bearman, and Kathleen Mullan Harris, and funded by a grant P01-HD31921 from the Eunice Kennedy Shriver National Institute of Child Health and Human Development, with cooperative funding from 17 other agencies. We thank Ronald R. Rindfuss and Barbara Entwisle for assistance in the original design. No direct support was received from grant P01-HD31921 for this analysis. The research was supported by the National Heart, Lung, and Blood Institute grant 2T32HL079891 (University of California San Diego training grant support to Erin Delker).
DATA AVAILABILITY
Persons interested in obtaining Data Files from Add Health should contact Add Health, The University of North Carolina at Chapel Hill, Carolina Population Center, Carolina Square, Suite 210, 123 W. Franklin Street, Chapel Hill, NC 27516 (addhealth_contracts@unc.edu).
REFERENCES
- 1.National Vital Statistics Reports Volume 68, Number 13, November 30, 2019, Births: Final Data for 2018. :47. [PubMed] [Google Scholar]
- 2.Vogel JP, Chawanpaiboon S, Moller AB, Watananirun K, Bonet M, Lumbiganon P. The global epidemiology of preterm birth. Best Pract Res Clin Obstet Gynaecol. 2018;52:3–12. doi: 10.1016/j.bpobgyn.2018.04.003 [DOI] [PubMed] [Google Scholar]
- 3.Institute of Medicine (US) Committee on Understanding Premature Birth and Assuring Healthy Outcomes; Behrman RE, Butler AS, editors. Preterm Birth: Causes, Consequences, and Prevention. Washington (DC): National Academies Press (US); 2007. 10, Mortality and Acute Complications in Preterm Infants. Preterm Birth: Causes, Consequences, and Prevention. [PubMed] [Google Scholar]
- 4.Al Khalaf SY, O’Reilly ÉJ, Barrett PM, et al. Impact of Chronic Hypertension and Antihypertensive Treatment on Adverse Perinatal Outcomes: Systematic Review and Meta-Analysis. J Am Heart Assoc. 2021;10(9):e018494. doi: 10.1161/JAHA.120.018494 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Bramham K, Parnell B, Nelson-Piercy C, Seed PT, Poston L, Chappell LC. Chronic hypertension and pregnancy outcomes: systematic review and meta-analysis. BMJ. 2014;348. doi: 10.1136/bmj.g2301 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Agrawal A, Wenger NK. Hypertension During Pregnancy. Curr Hypertens Rep. 2020;22(9):64. doi: 10.1007/s11906-020-01070-0 [DOI] [PubMed] [Google Scholar]
- 7.Berger H, Melamed N, Davis BM, et al. Impact of diabetes, obesity and hypertension on preterm birth: Population-based study. PloS One. 2020;15(3):e0228743. doi: 10.1371/journal.pone.0228743 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Maducolil MK, Al-Obaidly S, Olukade T, Salama H, AlQubaisi M, Al Rifai H. Maternal characteristics and pregnancy outcomes of women with chronic hypertension: a population-based study. J Perinat Med. 2020;48(2):139–143. doi: 10.1515/jpm-2019-0293 [DOI] [PubMed] [Google Scholar]
- 9.Johnson S, Liu B, Kalafat E, Thilaganathan B, Khalil A. Maternal and Perinatal Outcomes of White Coat Hypertension During Pregnancy: A Systematic Review and Meta-Analysis. Hypertens Dallas Tex 1979. 2020;76(1):157–166. doi: 10.1161/HYPERTENSIONAHA.119.14627 [DOI] [PubMed] [Google Scholar]
- 10.Atlass J, Menke M, Parks WT, Catov JM. Pre-conception blood pressure and evidence of placental malperfusion. BMC Pregnancy Childbirth. 2020;20(1):25. doi: 10.1186/s12884-019-2699-3 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Kovo M, Schreiber L, Ben-Haroush A, Wand S, Golan A, Bar J. Placental vascular lesion differences in pregnancy-induced hypertension and normotensive fetal growth restriction. Am J Obstet Gynecol. 2010;202(6):561.e1–5. doi: 10.1016/j.ajog.2010.01.012 [DOI] [PubMed] [Google Scholar]
- 12.Chronic Hypertension in Pregnancy. Accessed July 1, 2020. https://www.acog.org/en/Clinical/ClinicalGuidance/PracticeBulletin/Articles/2019/01/ChronicHypertensioninPregnancy
- 13.Lu Y, Chen R, Cai J, Huang Z, Yuan H. The management of hypertension in women planning for pregnancy. Br Med Bull. 2018;128(1):75–84. doi: 10.1093/bmb/ldy035 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Whelton Paul K, Carey Robert M, Aronow Wilbert S, et al. 2017 ACC/AHA/AAPA/ABC/ACPM/AGS/APhA/ASH/ASPC/NMA/PCNA Guideline for the Prevention, Detection, Evaluation, and Management of High Blood Pressure in Adults: Executive Summary: A Report of the American College of Cardiology/American Heart Association Task Force on Clinical Practice Guidelines. Hypertension. 2018;71(6):1269–1324. doi: 10.1161/HYP.0000000000000066 [DOI] [PubMed] [Google Scholar]
- 15.Battarbee AN, Sinkey RG, Harper LM, Oparil S, Tita ATN. Chronic hypertension in pregnancy. Am J Obstet Gynecol. 2020;222(6):532–541. doi: 10.1016/j.ajog.2019.11.1243 [DOI] [PubMed] [Google Scholar]
- 16.Topel ML, Duncan EM, Krishna I, Badell ML, Vaccarino V, Quyyumi AA. Estimated Impact of the 2017 American College of Cardiology/American Heart Association Blood Pressure Guidelines on Reproductive-Aged Women. Hypertens Dallas Tex 1979. 2018;72(4):e39–e42. doi: 10.1161/HYPERTENSIONAHA.118.11660 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.SMFM Publications Committee. Electronic address: pubs@smfm.org. SMFM Statement: benefit of antihypertensive therapy for mild-to-moderate chronic hypertension during pregnancy remains uncertain. Am J Obstet Gynecol. 2015;213(1):3–4. doi: 10.1016/j.ajog.2015.04.013 [DOI] [PubMed] [Google Scholar]
- 18.Li Q, Zheng L, Gu Y, et al. Early pregnancy stage 1 hypertension and high mean arterial pressure increased risk of adverse pregnancy outcomes in Shanghai, China. J Hum Hypertens. Published online March 23, 2021:1–8. doi: 10.1038/s41371-021-00523-6 [DOI] [PubMed] [Google Scholar]
- 19.Hauspurg A, Sutton EF, Catov JM, Caritis SN. Aspirin Effect on Adverse Pregnancy Outcomes Associated With Stage 1 Hypertension in a High-Risk Cohort. Hypertension. 2018;72(1):202–207. doi: 10.1161/HYPERTENSIONAHA.118.11196 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Wu D dan, Gao L, Huang O, et al. Increased Adverse Pregnancy Outcomes Associated With Stage 1 Hypertension in a Low-Risk Cohort. Hypertension. 2020;75(3):772–780. doi: 10.1161/HYPERTENSIONAHA.119.14252 [DOI] [PubMed] [Google Scholar]
- 21.Liu J, Tao L, Cao Y, et al. Stage 1 hypertension defined by the 2017 American College of Cardiology/American Heart Association guideline and risk of adverse birth outcomes in Eastern China. J Hypertens. 2020;38(6):1090–1102. doi: 10.1097/HJH.0000000000002380 [DOI] [PubMed] [Google Scholar]
- 22.Sutton EF, Hauspurg A, Caritis SN, Powers RW, Catov JM. Maternal Outcomes Associated With Lower Range Stage 1 Hypertension. Obstet Gynecol. 2018;132(4):843–849. doi: 10.1097/AOG.0000000000002870 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Xiao Y, Liu J, Teng H, Ge W, Han B, Yin J. Stage 1 hypertension defined by the 2017 ACC/AHA guidelines and neonatal outcomes: Systematic review and meta-analysis. Pregnancy Hypertens. 2021;25:204–212. doi: 10.1016/j.preghy.2021.06.011 [DOI] [PubMed] [Google Scholar]
- 24.Sabol BA, Porcelli B, Diveley E, et al. Defining the risk profile of women with stage 1 hypertension: a time to event analysis. Am J Obstet Gynecol MFM. 2021;3(4):100376. doi: 10.1016/j.ajogmf.2021.100376 [DOI] [PubMed] [Google Scholar]
- 25.Yang Y, He Y, Li Q, et al. Preconception blood pressure and risk of preterm birth: a large historical cohort study in a Chinese rural population. Fertil Steril. 2015;104(1):124–130. doi: 10.1016/j.fertnstert.2015.03.024 [DOI] [PubMed] [Google Scholar]
- 26.Greenberg VR, Silasi M, Lundsberg LS, et al. Perinatal outcomes in women with elevated blood pressure and stage 1 hypertension. Am J Obstet Gynecol. 2021;224(5):521.e1–521.e11. doi: 10.1016/j.ajog.2020.10.049 [DOI] [PubMed] [Google Scholar]
- 27.Harris KM, Florey F, Tabor J, Bearman PS, Jones J, Udry J. The National Longitudinal Study of Adolescent Health: Research Design; 2009. http://www.cpc.unc.edu/projects/addhealth/design
- 28.Entzel P Add Health Wave IV Documentation: Cardiovascular and Anthropometric Measures. Published online 2009. doi: 10.17615/C6HM11 [DOI] [Google Scholar]
- 29.Huxley R, Mendis S, Zheleznyakov E, Reddy S, Chan J. Body mass index, waist circumference and waist:hip ratio as predictors of cardiovascular risk--a review of the literature. Eur J Clin Nutr. 2010;64(1):16–22. doi: 10.1038/ejcn.2009.68 [DOI] [PubMed] [Google Scholar]
- 30.Spiegelman D, Hertzmark E. Easy SAS Calculations for Risk or Prevalence Ratios and Differences. Am J Epidemiol. 2005;162(3):199–200. doi: 10.1093/aje/kwi188 [DOI] [PubMed] [Google Scholar]
- 31.Desquilbet L, Mariotti F. Dose-response analyses using restricted cubic spline functions in public health research. Stat Med. 2010;29(9):1037–1057. doi: 10.1002/sim.3841 [DOI] [PubMed] [Google Scholar]
- 32.Harville EW, Myers L, Shu T, Wallace ME, Bazzano LA. Pre-pregnancy cardiovascular risk factors and racial disparities in birth outcomes: the Bogalusa Heart Study. BMC Pregnancy Childbirth. 2018;18. doi: 10.1186/s12884-018-1959-y [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Sun Baiyang, Bertolet Marnie, Brooks Maria M., et al. Life Course Changes in Cardiometabolic Risk Factors Associated With Preterm Delivery: The 30-Year CARDIA Study. J Am Heart Assoc. 2020;9(15):e015900. doi: 10.1161/JAHA.119.015900 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Richardson DB, Kaufman JS. Estimation of the Relative Excess Risk Due to Interaction and Associated Confidence Bounds. Am J Epidemiol. 2009;169(6):756–760. doi: 10.1093/aje/kwn411 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Hernán M, Robins J. Causal Inference: What If. In: Ch 12. IP Weighting and Marginal Structural Models. Chapman & Hall/CRC. [Google Scholar]
- 36.Multiple Imputation in SAS Part 1. Accessed July 22, 2020. https://stats.idre.ucla.edu/sas/seminars/multiple-imputation-in-sas/mi_new_1/
- 37.Snowden JM, Bovbjerg ML, Dissanayake M, Basso O. The Curse of the Perinatal Epidemiologist: Inferring Causation Amidst Selection. Curr Epidemiol Rep. 2018;5(4):379–387. doi: 10.1007/s40471-018-0172-x [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.Heinke D, Rich-Edwards JW, Williams PL, et al. Quantification of selection bias in studies of risk factors for birth defects among livebirths. Paediatr Perinat Epidemiol. 2020;34(6):655–664. doi: 10.1111/ppe.12650 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39.Panaitescu AM, Syngelaki A, Prodan N, Akolekar R, Nicolaides KH. Chronic hypertension and adverse pregnancy outcome: a cohort study. Ultrasound Obstet Gynecol Off J Int Soc Ultrasound Obstet Gynecol. 2017;50(2):228–235. doi: 10.1002/uog.17493 [DOI] [PubMed] [Google Scholar]
- 40.Hakim RB, Tielsch JM, See LC. Agreement between Maternal Interview- and Medical Record based Gestational Age. Am J Epidemiol. 1992;136(5):566–567. doi: 10.1093/oxfordjournals.aje.a116534 [DOI] [PubMed] [Google Scholar]
- 41.Harville EW, Viikari JSA, Raitakari OT. Preconception cardiovascular risk factors and pregnancy outcome. Epidemiol Camb Mass. 2011;22(5):724–730. doi: 10.1097/EDE.0b013e318225c960 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 42.Nobles CJ, Mendola P, Mumford SL, et al. Preconception Blood Pressure and Its Change Into Early Pregnancy: Early Risk Factors for Preeclampsia and Gestational Hypertension. Hypertens Dallas Tex 1979. 2020;76(3):922–929. doi: 10.1161/HYPERTENSIONAHA.120.14875 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 43.Bellos I, Pergialiotis V, Papapanagiotou A, Loutradis D, Daskalakis G. Comparative efficacy and safety of oral antihypertensive agents in pregnant women with chronic hypertension: a network metaanalysis. Am J Obstet Gynecol. 2020;223(4):525–537. doi: 10.1016/j.ajog.2020.03.016 [DOI] [PubMed] [Google Scholar]
- 44.Abalos E, Duley L, Steyn DW, Gialdini C. Antihypertensive drug therapy for mild to moderate hypertension during pregnancy. Cochrane Database Syst Rev. 2018;10:CD002252. doi: 10.1002/14651858.CD002252.pub4 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 45.Bello NA, Zhou H, Cheetham TC, et al. Prevalence of Hypertension Among Pregnant Women When Using the 2017 American College of Cardiology/American Heart Association Blood Pressure Guidelines and Association With Maternal and Fetal Outcomes. JAMA Netw Open. 2021;4(3):e213808. doi: 10.1001/jamanetworkopen.2021.3808 [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
Persons interested in obtaining Data Files from Add Health should contact Add Health, The University of North Carolina at Chapel Hill, Carolina Population Center, Carolina Square, Suite 210, 123 W. Franklin Street, Chapel Hill, NC 27516 (addhealth_contracts@unc.edu).