Study protocol for the randomized controlled EVA (early vascular adjustments) trial: tailored treatment of mild hypertension in pregnancy to prevent severe hypertension and preeclampsia

The haemodynamic profile of gestational hypertension differs between individuals. On the one hand, gestational hypertension can originate from hyperdynamic circulation characterized by high cardiac output and low vascular resistance; this profile is often accompanied by late onset preeclampsia and normal foetal growth (Fig. 1). On the other hand, a hypertensive profile with high vascular resistance is associated with early onset maternal complications and impaired foetal growth [16‐19]. In the clinical phase of preeclampsia, most women exhibit a high-resistance profile with either low cardiac output (in 58% of women) or normal cardiac output (in 36% of women) [20]. Altered pre-pregnant haemodynamic phenotypes, inadequate cardiovascular adaptation to pregnancy, or crossover from hyperdynamic circulation to the more unfavourable hypertensive circulation with high vascular resistance as a consequence of endothelial derangement and loss of intravascular fluid, might underlie the divergent haemodynamic profiles [21, 22]. These heterogeneous circulatory profiles may explain the variable results of trials on antihypertensive therapy in mild to moderate hypertension during pregnancy. Beta-blockers are thought to be suitable for treating hyperdynamic hypertension, but reducing cardiac output may be detrimental in hypertensive women with high vascular resistance who already have low-cardiac-output. These women may instead benefit from vasodilation by dihydropyridine calcium channel blockers, or from central-acting alpha agonists that reduce peripheral resistance.

In non-pregnant individuals with uncontrolled hypertension, haemodynamically tailored antihypertensive treatment almost doubled the chance of reaching the target blood pressure within 3 months of initiating or adjusting the therapy compared with conventional standard treatment based on the specialist’s preference [23, 24]. During pregnancy, with respect to underlying haemodynamic profile, adequate blood pressure response to beta-blockage can be predicted by the mean arterial pressure (MAP), HR, and (if available) stroke volume index [25]. About 20–25% of gestational hypertensive women were refractory to labetalol and needed additional vasodilatory therapy. In these women, cardiac output was lowest and vascular resistance was highest before treatment. The highest response rate to labetalol primarily relates to alteration of HR and peripheral vascular resistance, and consequently blood pressure, but not stroke volume. This suggests that a simple algorithm weighing blood pressure and HR may grossly indicate the underlying haemodynamic profile. Therefore, it may be better to view antihypertensive drugs as adjusters of the haemodynamic state rather than as reducers of blood pressure [26]. If we do not take the maternal haemodynamic profile and the pharmacological mode of action of the drug into account, generic antihypertensive treatment will continue to result in disappointing, ineffective or even paradoxical outcomes for at least part of treated mothers and their offspring. To address this problem, we propose a randomized trial to compare tailored antihypertensive treatment based on the maternal circulatory profile with generally practiced active surveillance of hypertension in pregnant women with mild to moderate hypertension.

The aim of this study is to determine whether haemodynamically tailored antihypertensive therapy in pregnant women with mild to moderate hypertension reduces the incidence of severe hypertension and preeclampsia and improves offspring outcome compared with standard care.

Our primary objective is to determine whether haemodynamically tailored treatment of mild to moderate hypertension reduces disease progression to severe hypertension and preeclampsia compared with standard care. Our secondary objectives are to assess the effects of this treatment approach on maternal and offspring outcomes, including concomitant HELLP syndrome and eclampsia, gestational age at delivery, and neonatal birth weight and centile.

The Early Vascular Adjustments (EVA) trial is a randomized controlled open-label superiority trial, with two parallel groups (intervention and control). Progression to severe hypertension or preeclampsia is the primary endpoint. A schematic overview of patient enrolment and follow-up, and the SPIRIT timetable for the study are presented in Figs. 2 and 3. Participants will be randomized 1:1 to the intervention or control group using premade, non-opaque sealed envelopes, in two blocks of 104 participants to achieve equal sample sizes in the end. Generalizability of the findings will be assessed in a third group of women who do not consent to randomization, but agree to follow-up of their pregnancy outcomes. In this group, women will receive standard care, meaning that their physician will discuss treatment with them, and decide which medication to prescribe.

Participants will be recruited from the outpatient clinic and department of obstetrics in Maastricht University Medical Centre, which is both a secondary and tertiary referring hospital. During every routine pregnancy check-up, blood pressure is measured. Women with mild to moderate hypertension (defined as systolic blood pressure measures ≥140 and < 160 mmHg and/or diastolic blood pressure ≥ 90 and < 110 mmHg) without symptoms or signs of preeclampsia will be given the study patient information form, and follow-up on blood pressure is planned at least 4 h later and within a few days. Eligible women who give consent will be randomized to the intervention or control group, after which they will continue regular pregnancy check-ups with blood pressure measurements; additional laboratory analysis will be performed as indicated. After 37 weeks’ gestation, induction of labour will be recommended to all participants, as this is associated with improved maternal outcome [27]. Study endpoints are an uneventful pregnancy and 6-week postpartum period, or diagnosis of severe hypertension or preeclampsia. In the latter case, participants will receive standard care, which usually means hospitalization and intensifying or initializing antihypertensive medication for the intervention group and control group, respectively.

Women must provide written informed consent before any study procedure occurs.

Eligible patients who consent to participate will be randomized to receive tailored antihypertensive medication (intervention group) or standard care (control group). The antihypertensive medications we plan to administer to participants in the intervention group can be safely prescribed to pregnant women [6]. Total peripheral vascular resistance (calculated as: 80 × MAP / cardiac output) indicates whether the hypertension is cardiac- or vascular-driven. In a healthy pregnancy, total peripheral vascular resistance drops from approximately 1300 to 1000 dyne.s/cm⁵ during the second and third trimester [28, 29]. Considering a standard deviation (SD) of 150 dyne.s/cm⁵, the upper limit of healthy dilated total peripheral vascular resistance is 1300 dyne.s/cm⁵. Women can be considered vasoconstricted when gestational vascular resistance exceeds 1300 dyne.s/cm⁵ and vasodilated when vascular resistance is below 1300 dyne.s/cm⁵. To avoid the need for an additional tool to assess cardiac output (stroke volume × HR), the MAP/HR ratio is used as a proxy for the haemodynamic profile. The MAP/HR ratio differs from total peripheral vascular resistance assessment in that it does not account for stroke volume. During healthy pregnancy, the mean ± SD stroke volume is 75 ± 10 ml. In the estimation of which proportion of women can be assumed to be vasoconstricted or vasodilated using the MAP/HR ratio, we can estimate the range in which total peripheral vascular resistance must be by taking the mean ± 2 SD stroke volume. When the MAP/HR ratio exceeds 1.4, more than 95% of all women must be considered vasoconstrictive. Contrary, when the MAP/HR ratio is 1.1 or less, more than 50% of women are likely to have a vasodilated circulation (Fig. 4). In the absence of empirical data that support the determination of cut-off values, we assumed a likelihood of more than 95% vasoconstrictive and relatively hypodynamic, corresponding a MAP/HR ratio 1.4 as reference value for considering vasodilating medication, and opposite, a likelihood of more than 50% of women to be vasodilated and hyperdynamic, corresponding MAP/HR ratio of 1.1 as a reference value for considering HR-lowering and with it, cardiac output-lowering medication. In the latter case, a beta-blocker with dual alpha- and beta-adrenergic receptor antagonism will account for the potential overlap in profiles. Thus, we will administer labetalol in the intervention group when hyperdynamic hypertension is assumed (MAP/HR ratio ≤ 1.1), and slow-release nifedipine when hypodynamic hypertension is assumed (MAP/HR ratio ≥ 1.4). Women with normodynamic hypertension will be identified by a MAP/HR ratio between 1.1 and 1.4 and will be prescribed methyldopa (Fig. 5). Treatment will be increased if the targeted blood pressure of < 130/80 mmHg (MAP 97 mmHg) is not achieved [7, 30]. The maximum dosages for blood pressure control are 800 mg three times daily for labetalol, 30 mg three times daily for slow release nifedipine and 1000 mg three times daily for methyldopa. If the maximum dosage has been administered and the MAP/HR ratio does not indicate another medication class, the blood pressure will be accepted. Moreover, in the unlikely event that blood pressure measurements fall below 95/50 mmHg, the treatment regime will be reduced and the last added medication step will be stopped. In case of intolerable side effects, the medication class will be switched; methyldopa will substitute labetalol or nifedipine, and nifedipine will substitute methyldopa when the MAP/HR ratio is ≥1.3 and labetalol when the MAP/HR ratio is < 1.3. Antihypertensive drugs will be provided by the local pharmacy. Since this is an open-label study, the participant, researcher, and physician will know the type and dosage of the prescribed medication. Adherence to the treatment and side effects of the medication in the intervention group will be discussed and recorded during each scheduled check-up.

Data will be recorded in predesigned case record forms. Data on pregnancy and neonatal outcomes will be collected from the hospital maternity records. Additional neonatal outcomes will be collected from the discharge summary when neonates are admitted to the neonatal care department.

Severe hypertension and preeclampsia develop respectively in 20% and 15–25% of women initially diagnosed with mild to moderate gestational hypertension [6, 32]. As severe hypertension and preeclampsia require comparable in-hospital treatment, we consider an 15% reduced incidence of both as clinically relevant. We calculated the sample size based on a progression level of 10% in the intervention group and a progression level of 25% in the control group. For a desired power of 80% and a two-tailed alpha level of 0.05, we need to recruit 99 women per group. Considering an anticipated dropout rate of 5%, we will recruit 208 women for randomization. We expect to include approximately 160 women who do not consent for randomization but agree to follow-up of their pregnancy outcomes. In this observational cohort, we will evaluate general obstetrical care for mild to moderate hypertension, patient characteristics, and patient outcomes comparable to the randomized population.

The effectiveness of haemodynamically tailored treatment of mild to moderate gestational hypertension will be evaluated based on the intention-to-treat principle. Descriptive statistics will be used to compare baseline characteristics between study groups. The primary analysis will be an X² test, or Fisher’s exact test in groups with less than five cases, to compare the incidence of severe hypertension and preeclampsia in the intervention group and the control group. We planned to conduct a per protocol sub-analysis in with women with at least 80% of their medication intake assessed by self-report at each pregnancy check-up. Secondary outcomes will be assessed by using X,² Fisher’s exact test, Mann-Whitney U or T test as appropriate. For both primary and secondary outcomes, crude and adjusted odds ratio’s will be calculated (control group will be considerate as reference) and adjustments will be made for gestational age at recruitment. Separate analyses will be conducted to evaluate whether or not optimal blood pressure control was reached, if the type of medication and dosage affected the outcomes, which haemodynamic profiles were most prevalent, and if the initial haemodynamic profile affected the outcome, and to explore the effect of BMI, parity and age on the haemodynamic profile and response to medication.