Beat-to-Beat Blood Pressure Variability in the First Trimester is Associated with the Development of Preeclampsia in a Prospective Cohort: Relation with Aortic Stiffness

Abstract

Women with preeclampsia, a hypertensive disorder of pregnancy, exhibit greater beat-to-beat blood pressure variability (BPV) in the third trimester after clinical onset of the disorder. However, it remains unknown whether elevated BPV precedes the development of preeclampsia. A prospective study cohort of 139 women (age 30.2±4.0 years) were enrolled in early pregnancy (<14 weeks gestation). BPV was quantified by time domain analyses of 10-minute continuous beat-to-beat blood pressure (BP) recordings via finger photoplethysmography in the first, second and third trimesters. Aortic stiffness (carotid-femoral pulse wave velocity, cfPWV) and spontaneous cardiovagal baroreflex sensitivity (BRS) were also measured each trimester. Eighteen women (13%) developed preeclampsia. Systolic BPV was higher in all trimesters among women who developed vs. did not develop preeclampsia (1^st: 4.8±1.3 versus 3.7±1.2, P=0.001; 2^nd: 5.1±1.8 versus 3.8±1.1, P=0.02; 3^rd: 5.2±0.8 versus 4.0±1.1 mmHg, P=0.002). Elevated first trimester systolic BPV was associated with preeclampsia (OR: 1.94, 95%CI 1.27–2.99), even after adjusting for risk factors (age, body mass index, systolic BP, history of preeclampsia and diabetes mellitus) and was a significant predictor of preeclampsia (receiver operating characteristic analysis, AUC=0.75±0.07; P=0.002). cfPWV was elevated in the first trimester among women who developed preeclampsia (5.9±0.8 versus 5.2±0.8 m/s; P=0.002) and was associated with BPV after adjustment for mean BP (r=0.26; P=0.005). First trimester BRS did not differ between groups (P=0.23) and was not related to BPV (P=0.36). Elevated systolic BPV is independently associated with the development of preeclampsia as early as the first trimester, possibly mediated in part by higher aortic stiffness.

Graphical Abstract

Summary

Early pregnancy elevations in beat-to-beat blood pressure variability is independently associated with late pregnancy onset of overt preeclampsia and may be modulated in part by higher aortic stiffness.

INTRODUCTION:

Preeclampsia is a pregnancy disorder characterized by hypertension and multi-system organ dysfunction in the second half of gestation.¹ Preeclampsia affects 2–8% of pregnancies worldwide² and is associated with immediate and long-term maternal morbidity and mortality.^{3, 4} At one year postpartum, women who had preeclampsia have a 2–4 fold elevated risk for cardiovascular disease.^{5, 6} Over their lifespan, women with a history of preeclampsia are at a four-fold higher risk to develop hypertension³ and this occurs earlier than is observed with normal aging.⁴ Thus, the cardiovascular sequalae of preeclampsia emphasize the need for timely intervention in pregnancy. Women who subsequently develop preeclampsia exhibit abnormal maternal hemodynamic adaptations to pregnancy in mid-gestation prior to the onset of overt clinical signs.^7–9 Consequently, early pregnancy alterations in hemodynamic and autonomic parameters may identify women at risk to develop preeclampsia.

Fluctuations in blood pressure (BP) occur over very short-term (beat-to-beat), short-term (24-hour and day-to-day) and long-term (visit-to-visit, seasonal) time intervals.¹⁰ Greater beat-to-beat BP lability is associated with the progression and severity of target-organ damage independent of elevated mean BP.^11–13 Beat-to-beat BPV primarily reflects cardiovascular regulatory mechanisms¹⁴ and is modulated in part by autonomic reflexes such as the baroreflex¹⁵ and large elastic artery stiffness in middle-aged/older adults.^{11, 16, 17} Beat-to-beat BPV is notably exaggerated in individuals with hypertension,^{11, 12} a clinical hallmark of preeclampsia. Moreover, variation in visit-to-visit BP, rather than absolute BP values, predicts hypertensive pregnancy disorders in mid-gestation.¹⁸ Thus, BPV may represent a novel hemodynamic/autonomic biomarker for the early prediction of preeclampsia. In contrast to 24-hour ambulatory and visit-to-visit BPV, beat-to-beat BPV can be characterized over a short duration (10 minutes) in a single visit without sleep disruption. Therefore, beat-to-beat BPV may have greater clinical utility in pregnancy than other BPV methods. However, whether beat-to-beat BPV is elevated in early gestation and associated with the development of preeclampsia has not been previously assessed.

Evidence from cross-sectional studies suggests that beat-to-beat BPV is elevated in late pregnancy in women with preeclampsia compared with healthy pregnancy controls.^19–22 These assessments of BPV were limited to the third trimester and the confounding effect of elevated absolute BP among women with preeclampsia was not considered. Furthermore, the potential mechanisms that modulate elevated beat-to-beat BPV in preeclampsia have not been explored. Therefore, the primary objective of this prospective study was to determine if beat-to-beat BPV in the first trimester is associated with the development of preeclampsia independent of elevations in BP alone and assess the diagnostic performance of first trimester BPV. Secondly, this study aimed to identify potential hemodynamic and/or autonomic mechanisms, including aortic stiffness and cardiovagal BRS, that may contribute to higher beat-to-beat BPV in early pregnancy among women who develop preeclampsia.

METHODS:

All procedures were approved by the University of Iowa Institutional Review Board and all participants reviewed and signed an Institutional Review Board-approved informed consent document before participating in the research study. No human subjects research was performed at the Medical College of Wisconsin. The data that support the findings of this study are available from the corresponding author upon reasonable request.

Participants:

Two-hundred and nine pregnant women between 18 and 40 years of age were recruited and prospectively enrolled in the first trimester (defined as <14 weeks gestation) via the University of Iowa Hospitals and Clinics Department of Obstetrics and Gynecology Maternal Fetal Tissue Bank (MFTB).²³ Participants were co-enrolled in the Early Vascular Dysfunction and Elevated Copeptin in Human Preeclampsia (EVECO) study between March 2015 and January 2020. Additional details on participant eligibility is provided in the Supplemental Methods. Those enrolled in the EVECO study prior to the introduction of these measurements are excluded from the present analysis. Subject characteristics of the EVECO study cohort (N=209) and the study sample in which BPV was measured (N=139) are presented in Table S1. Continuous beat-to-beat BP measurements were initiated in November 2015 and collected in one-hundred and thirty-nine women through June 2019 (Table S1). The diagnosis of preeclampsia and further classification as with or without severe features was determined according to the American College of Obstetrics and Gynecology definitions.¹ Assessments of BPV were conducted prospectively in the first (6–13 weeks), second (18–26 weeks), and third trimester (27–37 weeks) and included beat-to-beat systolic and diastolic BPV, spontaneous cardiovagal BRS measured by the sequence method, and aortic stiffness measured by carotid-femoral pulse wave velocity (cfPWV).

Beat-to-beat blood pressure variability:

Participants were instrumented with a 3-lead electrocardiogram (ECG) (AccuSync, Mildford, CT) and finger photoplethysmography (Nexfin; BMEYE, Amsterdam, The Netherlands) to synchronously record heart rate and continuous arterial blood pressure, respectively, while supine.²⁴ BPV was calculated in the time domain by standard deviation and frequency domain via fast Fourier transform spectral analysis (Cardioseries v2.7; Sao Paulo, Brazil). Spectral decomposition of systolic and diastolic BP oscillations was conducted to quantify variability in frequency bands corresponding to mechanisms of cardiovascular control.^{25, 26} Oscillations in BP were quantified in low frequency (0.04–0.15 Hz; BPV_LF) and high frequency (0.15–0.40 Hz; BPV_HF) bands (mmHg²).²⁷ Refer to data supplement for additional details.

Spontaneous baroreflex sensitivity:

Cardiovagal BRS was assessed in the time domain via the sequence technique (Cardioseries v2.7; Sao Paulo, Brazil) as detailed in the data supplement.

Aortic stiffness:

Aortic stiffness was measured using applanation tonometry and expressed as the reference-standard cfPWV as previously described²⁸ and detailed in the data supplement.

Statistical Analyses:

Analyses were conducted using IBM SPSS Statistics Ver. 25 (IBM, Armonk, NY). Pregnancy outcomes were obtained for all participants. Detailed descriptions of a priori sample size estimation and statistical analysis protocol is presented in the data supplement. Logistic regression models were constructed to determine if first trimester BPV was associated with the development of preeclampsia and whether clinically meaningful covariates including maternal age, body mass index, systolic BP, family or individual history of preeclampsia, parity and diabetes mellitus altered this association. Receiver operator characteristic (ROC) curves were constructed to quantify the diagnostic performance of first trimester BPV. Bivariate Pearson’s and partial correlations were performed to investigate the association of BP, cfPWV, and cardiovagal BRS to BPV in the first trimester. Data are expressed as mean ± SD or median [IQR]. An α-level of 0.05 was selected to indicate statistical significance for all analyses.

RESULTS:

Participants:

Eighteen women enrolled in the first trimester subsequently developed preeclampsia in the third trimester as verified from the medical record. All preeclampsia cases exhibited de novo hypertension and proteinuria. Eleven women developed preeclampsia with severe features including severely elevated BPs (n=7), HELLP syndrome (n=5), de novo headache (n=5) and visual disturbance (n=2). Women who did not develop preeclampsia (n=121) were characterized by uncomplicated pregnancy, specifically, pregnancies with no evidence of gestational diabetes or gestational hypertension. Subject and delivery characteristics are presented in Table 1. Women who developed preeclampsia did not differ in age, body mass index, previous pregnancies (gravida), or parity from those who did not develop preeclampsia. Participants were predominantly white and racial distribution was similar between groups (P=0.66). As expected, a greater proportion of the women who developed preeclampsia exhibited more traditional risk factors for preeclampsia²⁹ such as pre-existing hypertension (P<0.001) and a history of preeclampsia in a previous pregnancy (P<0.001). The women who developed preeclampsia exhibited significantly lower gestational age at delivery (37 [35–38] versus 39 [39–40] weeks; P<0.001) and babies born to preeclamptic women had a lower birthweight (2819±687 versus 3442±529 g; P<0.001) compared with women who did not have preeclampsia, consistent with higher rates of preterm and small for gestational age birthweight associated with preeclampsia.²⁹

Table 1:

Subject and pregnancy characteristics

Characteristic	Preeclampsia (n=18)	Without Preeclampsia (n=121)	p-value
Maternal Age, y	31.1 ± 4.0	30.1 ± 4.0	0.32
Body mass index (1st trimester), kg/m2	29 [24–31]	25 [22–30]	0.11
Preeclampsia with severe features, n (%)	11 (61.1)	-	-
Early onset preeclampsia, n (%)	3 (16.7)	-	-
Gravida, n	2 [1–3]	2 [1–3]	0.23
Parity, n	1 [0–1]	1 [0–1]	0.41
Primigravida, n (%)	8 (44.4)	32 (26.4)	0.16
Pre-existing hypertension, n (%)	5 (27.8)	0 (0)	<0.001
Pre-existing diabetes mellitus, n (%)	1 (5.55)	3 (2.5)	0.43
History of preeclampsia, n (%)	4 (22.2)	0 (0)	<0.001
Family history of preeclampsia, n (%)	1 (5.5)	4 (3.3)	0.50
Current smoker, n (%)	0 (0)	1 (0.8)	0.87
Race:
White, not Hispanic, n (%)	18 (100)	111 (91.7)	0.66
Hispanic, n (%)	0 (0)	1 (0.8)	0.66
Asian, n (%)	0 (0)	5 (4.1)	0.66
Black, n (%)	0 (0)	4 (3.3)	0.66
Gestational age at delivery, wk	37 [35–38]	39 [39–40]	<0.001
Twin gestation, n (%)	0 (0)	2 (1.7)	0.76
Birth weight, g	2819 ± 687	3442 ± 529	<0.001
Birth weight for gestational age, centile	33 [17–51]	54 [30–77]	0.027
APGAR 1 minute	8 [4.5–8]	8 [7.25–9]	0.007
APGAR 5 minute	8.5 [7–9]	9 [9–9]	0.001

Data are mean ± SD or median [IQR]. APGAR, Appearance, Pulse, Grimace, Activity, Respiration

Comparison of beat-to-beat BPV in preeclampsia and healthy pregnancy:

BPV, BRS, and cfPWV were assessed in the first (12±1 weeks gestation), second (22±2 weeks gestation), and third trimester (32±3 weeks gestation). Brachial systolic and diastolic BP were elevated among women who developed preeclampsia in every trimester compared with those who did not develop preeclampsia (all P<0.01; Table S2). Women who developed preeclampsia had elevated systolic BPV during every trimester (1^st: 4.8±1.3 vs. 3.7±1.2 mmHg, P=0.001; 2^nd: 5.1±1.8 vs. 3.8±1.1 mmHg, P=0.02; 3^rd: 5.2±0.8 vs. 4.0±1.1 mmHg; P=0.002; Table S2) compared with women who did not develop preeclampsia, paralleled by elevated diastolic BPV in every trimester (1^st: 3.1±0.9 vs. 2.5±0.7 mmHg, P=0.02; 2^nd: 3.0±0.8 vs. 2.5±0.5 mmHg, P=0.007; 3^rd 3.5±1.3 vs. 2.5±0.8 mmHg, P=0.04; Table S2).

Systolic and diastolic BPV_LF were elevated throughout pregnancy in those who developed preeclampsia compared with those who did not develop preeclampsia, whereas BPV_HF sharply increased in the preeclampsia group only in the third trimester (Table S3). Cardiovagal BRS tended to be lower in the second trimester among women who developed preeclampsia compared with women who did not (9.9 [5.2–13.2] vs. 12.3 [8.2–20.2] ms/mmHg; P=0.06), but did not differ in the first (P=0.23) or third trimesters (P=0.51; Table S2). Women who developed preeclampsia had elevated cfPWV in the first (5.9±0.8 vs. 5.2±0.8 m/s; P=0.002), second (5.5±1.0 vs. 4.8±0.8 m/s; P=0.003) and third trimester (5.9±1.1 vs. 5.0±0.8 m/s; P=0.001; Table S2) compared with women who did not develop preeclampsia.

Early pregnancy BPV and development of preeclampsia:

Logistic regression models were constructed to determine whether clinical risk factors confounded the association between first trimester BPV and preeclampsia diagnosis. Maternal age, body mass index, systolic BP, family or individual history of preeclampsia, parity and diabetes mellitus were included as covariates (Table 2). Elevated first trimester systolic BPV was significantly associated with preeclampsia (OR: 1.94, 95%CI 1.27–2.99) and remained significant when controlling for all covariates individually (Table 2). First trimester diastolic BPV was associated with the development of preeclampsia (OR: 3.14, 95%CI 1.47–6.70) and remained significant when controlling for all covariates individually (data not shown). Among normotensive women only (N=133), first trimester systolic BPV was associated with preeclampsia independent of all covariates (OR: 1.91, 95%CI 1.20–3.03; Table S4).

Table 2:

Logistic regression models for the association between first trimester systolic blood pressure variability and the development of preeclampsia

Model	β (SBPV)	SE	OR	95% Confidence interval for OR		P value
				Lower	Upper
SBPV (unadjusted)*	0.66	0.22	1.94	1.27	2.99	0.002
SBPV adjusted for maternal age	0.65	0.22	1.91	1.24	2.94	0.003
SBPV adjusted for body mass index	0.64	0.22	1.91	1.24	2.93	0.003
SBPV adjusted for SBP	0.54	0.23	1.72	1.10	2.70	0.019
SBPV adjusted for history of PE	0.71	0.23	2.04	1.30	3.20	0.002
SBPV adjusted for diabetes mellitus	0.69	0.22	2.00	1.30	3.09	0.002
SBPV adjusted for parity	0.68	0.22	1.97	1.28	3.04	0.002

SBPV indicates systolic blood pressure variability; OR, odds ratio; SBP, systolic blood pressure; PE, preeclampsia.

^*ROC curve is presented in Figure 1.

A ROC analysis was performed to describe the diagnostic characteristics of BPV in the first trimester for preeclampsia later in pregnancy. The time domain index (standard deviation) of BPV was selected as the predictor due to its prognostic significance and clinical utility.³⁰ The area under the ROC curve (AUC) demonstrated the ability of systolic BPV in first trimester to discern pregnancy outcome (AUC=0.75 ± 0.07; P=0.002; Figure 1). The optimal systolic BPV cutoff value associated with the prediction of preeclampsia was 3.9 mmHg, with 86.7% sensitivity and 61.7% specificity (Figure 1), corresponding to a positive likelihood ratio of 2.3 and negative likelihood ratio of 0.2. Accounting for the prevalence of preeclampsia estimated between 3–10%, first trimester systolic BPV yielded a positive and negative predictive value range of 7–20% and 98–99%, respectively (Table S5). First trimester diastolic BPV predicted the development of preeclampsia (AUC=0.71 ± 0.08; P=0.009) with an optimal cutoff value of 2.5 mmHg with 73.3% sensitivity and 61.9% specificity, corresponding to a positive likelihood ratio of 1.9 and a negative likelihood ratio if 0.4. Diastolic BPV yielded a positive and negative predictive value range of 6–18% and 95–99% respectively (Table S5).

Figure 1:

ROC curve illustrating the ability of first trimester systolic BPV to identify pregnancies that will eventually develop preeclampsia with an optimal cut off value of 3.9 mmHg (grey).

These findings were compared to the discriminatory ability of systolic and diastolic BP alone among women with and without previous hypertension. Systolic BP exhibited an optimal cut-off of 113 mmHg, associated with 72.2% sensitivity and 75.8% specificity. Systolic BP identified 22% of future preeclampsia cases at a 5% false positive rate (FPR; Figure S1, Table S6) compared with 33% detected by systolic BPV (Figure 1, Table S7). Diastolic BP was associated with an optimal cut-off of 63 mmHg associated with 66.7% sensitivity and 70.5% specificity and identified 27% of future preeclampsia cases at 5% FPR (Figure S1, Table S6), compared with 33% indentified by diastolic BPV. Among normotensive women only (n=133), systolic BP was associated with 60.5% sensitivity and 70.1% specificity and diastolic BP was associated with 53.8% sensitivity and 65.6% specificity for the future diagnosis of preeclampsia.

First trimester systolic BPV_LF, which reflects sympathetic modulation of vascular tone,²⁷ was associated with the development of preeclampsia (OR: 1.30 95%CI 0.97–1.66) but was not associated with preeclampsia independent of systolic BP (OR: 1.27, 95%CI 0.97–1.66; Table S8). First trimester systolic BPV_HF, which reflects the mechanical influence of respiration,²⁷ was not associated with preeclampsia.

Logistic regression models were also constructed to determine if first trimester cfPWV was independently associated with the development of preeclampsia. Elevated cfPWV was significantly associated with preeclampsia (OR: 2.60, 95%CI 1.38–4.91; Table S9), but not independent of mean arterial pressure (OR: 1.38, 95%CI 0.65–2.97) or systolic BPV (OR: 1.96, 95%CI 0.89–4.32; Table S9).

Logistic regression models and ROC curves were constructed to determine if systolic BPV was associated with the development of preeclampsia when assessed in mid (2^nd trimester) to late (3^rd trimester) gestation. The AUC demonstrated the ability of second trimester systolic BPV (AUC=0.71; P=0.017) and third trimester systolic BPV (AUC=0.80; P=0.002) to identify preeclampsia (Figure S2). However, neither second trimester systolic BPV nor third trimester systolic BPV predicted the development of preeclampsia independent of corresponding systolic BP (Table S10)

Baroreflex, aortic stiffness as putative mechanisms:

Bivariate Pearson’s and partial correlations were performed to determine the relation of first trimester BP, cfPWV, and BRS with BPV in the entire cohort. Higher systolic BP (r=0.23; P=0.01) and cfPWV (r=0.31; P=0.001) correlated with systolic BPV, whereas cardiovagal BRS was not related to systolic BPV (r=−0.08; P=0.36). The relation between cfPWV and systolic BPV remained significant after adjusting for systolic BP (r=0.26; P=0.005; Figure 2). Similarly, higher cfPWV correlated with elevated diastolic BPV (r=0.20; P=0.03) after accounting for diastolic BP, while cardiovagal BRS was not related to diastolic BPV (r=0.11; P=0.24). Diastolic BP weakly correlated with diastolic BPV (r=0.18; P=0.045). Logistic regression models were constructed to assess whether the addition of predictors BRS and cfPWV altered the association between first trimester systolic BPV and preeclampsia outcome. The coefficient corresponding to systolic BPV (β_BPV=0.661; Table 2) was reduced by a mere 0.3% after BRS was added to the model (β_BPV=0.659; data not shown) and BRS was not associated with the development of preeclampsia (OR: 0.98, 95%CI 0.92–1.03; data not shown). The coefficient corresponding to systolic BPV was reduced by 7.9% with the addition of cfPWV (β_BPV=0.609; data not shown), but cfPWV was not associated with the development of preeclampsia independently of BPV (Table S9).

Figure 2:

Bivariate Pearson’s and partial correlations between aortic stiffness and A) Systolic and B) Diastolic blood pressure variability in the first trimester of pregnancy. BPV, blood pressure variability; PWV, pulse wave velocity; CI, confidence interval. r indicates bivariate correlation, r_p indicates partial correlation adjusted for A) systolic or B) diastolic blood pressure.

DISCUSSION:

The present study investigated the association between very short-term beat-to-beat BPV in the first trimester and the subsequent development of preeclampsia. The novel findings of this study indicate a strong relation between first trimester beat-to-beat systolic BPV and the development of preeclampsia. First, beat-to-beat BPV was considerably elevated within the first trimester among women who eventually developed preeclampsia and remained elevated in mid (2^nd trimester) to late (3^rd trimester) pregnancy. Second, beat-to-beat BPV was significantly associated with the diagnosis of preeclampsia, independent of clinical risk factors, as early as the first trimester. Its association is sustained in the 2^nd and 3^rd trimesters. Additionally, the present study found that aortic stiffness, but not spontaneous cardiovagal BRS, was elevated among women who developed preeclampsia and was significantly related to greater BPV. Overall, these findings support the concept that beat-to-beat BPV is a feature of early pregnancy hemodynamic/autonomic dysregulation in preeclampsia. While BPV is a modest predictor of the disease, higher BPV is associated with cfPWV independent of average BP, suggesting a link between elevated BPV and higher aortic stiffness in pregnancy.

The most important and clinically relevant finding of the present study is that BPV is independently associated with the future development of preeclampsia by the twelfth week of pregnancy. The relatively low prevalence of preeclampsia in the general obstetric population poses significant challenges in the development of accurate diagnostic tests. Preeclampsia is characterized by a lengthy latent stage between the initiating event, possibly at placentation, and the clinical presentation in late gestation.³¹ Early identification of women at higher risk during this subclinical period may guide pharmacologic prophylactic measures, such as aspirin administration, and closer perinatal BP monitoring.³² A multitude of hemodynamic parameters implicated in the pathogenesis and progression of preeclampsia have been explored in prediction models but have proved insufficient or impractical for clinical use. Mean arterial pressure is elevated among women who develop preeclampsia in the first trimester and predicts select phenotypes of preeclampsia in combination with maternal characteristics.^{33, 34} Parameters such as cardiac output and total peripheral resistance⁷ are altered in the first trimester in women who subsequently develop preeclampsia, but these variables have limited efficacy for predicting preeclampsia without severe features and require complex measurements for screening that are not routinely performed. First trimester arterial stiffness does not notably enhance systolic BP as a predictor of hypertension in pregnancy,³⁵ consistent with the present findings. Uterine artery velocimetry has been investigated as a method for the early prediction of preeclampsia,^{36, 37} but the validity of this measure is significantly influenced by operator expertise³⁸ and indices used to describe uterine artery hemodynamics are heterogeneous across studies.³⁶ A frequency index of diastolic beat-to-beat BPV has previously been reported to be predictive of preeclampsia in the second trimester when used in conjunction with abnormal uterine artery perfusion.²¹ In contrast, the current study demonstrates that both systolic and diastolic BPV alone are independently associated with preeclampsia earlier in pregnancy (first trimester) using a simpler index of BPV that may be more practical for clinical use. In evaluating the diagnostic performance of ‘low’ and ‘high’ systolic BPV according to optimal threshold, high systolic BPV had a small to moderate impact in identifying or ruling out future preeclampsia, respectively, with slightly higher accuracy compared with systolic BP in normotensive women. Collectively, our findings identify beat-to-beat BPV as a novel hemodynamic feature in early pregnancy that may further complement maternal risk factors and circulating biomarkers of preeclampsia in improving risk stratification in early pregnancy.

The results of this study indicate elevated BPV precedes the clinical manifestation of preeclampsia and may reflect underlying alterations in autonomic baroreflex control and/or large artery stiffness that modulate beat-to-beat BP fluctuations. To date, the mechanisms that underpin BPV in healthy pregnancy and preeclampsia are not fully understood. The findings in our prospective study support and extend those of previous cross-sectional studies which demonstrated greater BPV in women who later develop preeclampsia at 20²⁰ and 30–36 weeks of gestation.¹⁹ Faber et al. reported similarly elevated BPV in pregnant women with chronic hypertension, suggesting that BPV in women with preeclampsia may be attributable to higher BP in the third trimester.¹⁹ In the present prospective study, first trimester BPV was associated with the development of preeclampsia independently of BP. Taken together, these data suggest that elevated BPV in preeclampsia cannot be explained by a concomitant rise in BP alone.

The baroreflex regulates beat-to-beat increases in BP by coupling vagally-mediated reductions in heart rate with inhibition of sympathetic nerve outflow to the periphery in order to lower peripheral resistance and BP. Baroreflex function is defined by the degree of the efferent response, cardiovagal BRS and sympathetic BRS respectively, and these indices of BRS are not associated in healthy adults.³⁹ Previous studies report reduced cardiovagal BRS among women with preeclampsia in the third trimester,^{19, 40–42} but alterations in sympathetic BRS have not been characterized in human preeclamptic pregnancy in any trimester. In the present study, cardiovagal BRS was not reduced among women who developed preeclampsia in the first trimester and was not correlated with first trimester BPV. While these results indicate that elevated BPV in the first trimester in women who develop preeclampsia is likely not attributable to reduced BRS alone, it is important to note that the sequence method used to determine BRS reflects baroreflex control of cardiovagal drive to the heart rather than of sympathetic outflow to the heart and periphery.⁴³ Baroreflex control of sympathetic outflow and peripheral resistance may be a more important determinant of BPV. Indeed, low frequency fluctuations in BP are thought to reflect baroreceptor-mediated sympathetic modulation of vascular tone,²⁵ while high frequency fluctuations in BP are attributed to the mechanical influence of respiration. Therefore, the elevated BPV_LF observed throughout pregnancy in women who developed preeclampsia compared with healthy pregnancy may reflect impaired baroreflex-mediated sympathetic control of arterial tone, although we did not measure this directly.

BPV may be influenced indirectly by the effect of the large elastic arteries (e.g., aorta, carotid arteries) on the baroreceptors as well as by the effect of the large arteries on BP pulsatility. The elastic properties of the large arteries are a principal determinant of baroreceptor stretch and relaxation in response to a given change in BP. Aortic and/or carotid stiffness limits the stimulation of the mechanosensitive baroreceptors which results in attenuated activation of the baroreflex and reduced efficacy in modulating BP fluctuations.⁴⁴ In support of this, aortic stiffness is associated with reduced sympathetic BRS⁴⁵ and cardiovagal BRS⁴⁶ in non-pregnant humans. Moreover, aortic stiffness is related to greater short-term fluctuations in BP in older adults.¹¹ Importantly, aortic stiffness is elevated in early pregnancy among women who develop preeclampsia prior to the onset of clinical symptoms.³⁵ In accordance with previous studies, we report that aortic stiffness is higher in women who developed preeclampsia as early as the twelfth week of gestation and associated with reduced cardiovagal BRS. Furthermore, aortic stiffness is moderately correlated with greater BPV independently of BP in the first trimester. Taken together, these results suggest that aortic stiffness potentially reduces the responsiveness of the baroreflex, thereby attenuating both the cardiovagal response and sympathetic modulation of peripheral resistance. Because the present study was limited to the measurement of cardiovagal BRS, additional studies are required to fully elucidate the potential causal pathway between aortic stiffness, sympathetic BRS, and BPV in pregnancy. Overall, these findings suggest that aortic stiffness may be a factor related to elevated BPV in preeclampsia in early pregnancy, even in the absence of hypertension.

Our results should be interpreted in the context of several limitations. First, zero prevalence of chronic hypertension and previous preeclampsia in the women who did not develop preeclampsia prevented the use of these clinically important covariates in logistic regression models. Systolic BP was included as a continuous covariate to mitigate the former limitation, but future work is required to confirm that BPV improves the risk stratification provided by clinical BP guidelines. Second, with only 18 preeclampsia cases observed, we were limited in the number of covariates that could be included in multivariable models without the risk of overfitting. Third, this study did not distinguish between early-onset and late-onset phenotypes of preeclampsia with or without severe features as outcomes in regression analyses because of insufficient sample size. Women with preeclampsia without severe features did not differ from those with preeclampsia with severe features in any subject characteristics or indices of BPV (Table S11). Therefore, it is unlikely that the association between BPV and pregnancy outcome was driven by the severity of preeclampsia. However, future studies are required to address whether the association between first trimester BPV and the development of preeclampsia is altered by disease onset and severity. Further, because the earliest assessment of BPV was conducted at 6–13 weeks gestation, we cannot determine whether elevated BPV preceded pregnancy in women who developed preeclampsia. Additional studies are indicated to determine whether BPV is a potential pre-pregnancy predictor for the development of preeclampsia. Finally, the participants in this study were predominantly non-Hispanic white women. In the United States, black women are at a higher risk to develop preeclampsia and are disproportionately affected by severe morbidities and mortality associated with preeclampsia.⁴⁷ Additional studies are needed to validate the utility of BPV as a predictor of preeclampsia in a more racially/ethnically diverse population.

The present study is the first to prospectively assess beat-to-beat BPV, cardiovagal BRS and aortic stiffness in pregnancy. Overall, our results demonstrate that beat-to-beat BPV is elevated as early as the first trimester in women who develop preeclampsia compared with healthy pregnancies and identify first trimester BPV as a novel predictor for the development of preeclampsia. First trimester BPV remained significantly associated with preeclampsia independent of traditional risk factors and is associated with the development of preeclampsia earlier in pregnancy than other prominent biomarkers. Furthermore, aortic stiffness is associated with greater beat-to-beat fluctuations in BP in the first trimester. The link between aortic stiffness and BPV provides novel insight into the autonomic cardiovascular dysregulation that occurs in early pregnancy, well before the clinical signs manifest among women who develop preeclampsia. Further studies are needed to establish the prediction of preeclampsia by BPV in non-white women, to elucidate the potential mechanisms contributing to elevated BPV in preeclampsia, such as sympathetic BRS.

PERSPECTIVES

Beat-to-beat BPV is elevated in late pregnancy in women with preeclampsia compared with a healthy pregnancy.^19–22 This prospective study demonstrates that elevated BPV precedes the clinical onset of preeclampsia and is independently associated with the disorder as early as the twelfth week of pregnancy. This is clinically meaningful because current diagnostic markers do not adequately predict preeclampsia until the second trimester, shortly before clinical signs are detected, which precludes focused monitoring and early intervention. Recent data suggest that other maternal hemodynamic alterations such as systemic vascular resistance and cardiac output identify different hemodynamic phenotypes that further stratifies appropriate treatment for women with preeclampsia.^{7, 48} Furthermore, our study identifies aortic stiffness, but not cardiovagal BRS, as a potential contributor in part to greater BPV in early pregnancy among women who develop preeclampsia. Certainly, understanding the altered vascular and autonomic regulation of beat-to-beat BPV and other hemodynamic factors adds significant understanding to the nature of the early pregnancy pathophysiology of preeclampsia. Moreover, these data provide additional insight into distinct hemodynamic/blood pressure phenotypes present in very early pregnancy in women prior to the manifestation of clinical signs and symptoms of preeclampsia.

Novelty and Significance.

What is new?

• In a prospective cohort, beat-to-beat blood pressure variability is associated with the development of preeclampsia as early as the twelfth week of pregnancy independent of absolute blood pressure levels and clinical covariates.

• Elevated aortic stiffness, but not cardiovagal baroreflex sensitivity, is associated with greater blood pressure variability in the first trimester of pregnancy.

What is relevant?

• Current diagnostic markers of preeclampsia are confined to mid-to-late gestation, shortly preceding the overt clinical onset of the disorder. Beat-to-beat blood pressure variability represents a novel, non-invasive biomarker of early pregnancy abnormal hemodynamics and predictor of preeclampsia.

• Early pregnancy disruptions in cardiovascular regulatory mechanisms that contribute to the progression of preeclampsia are not well characterized.