Introduction

Arterial hypertension (AH) is a major risk factor for cardiovascular disease and is associated with functional and structural changes in the heart and the vasculature. AH promotes myocardial hypertrophy, cardiomyocyte apoptosis and myocardial fibrosis through various mechanical (wall tension, volume load, blood viscosity) and neurohormonal pathophysiological pathways (activation of sympathetic and renin angiotensin aldosterone system, activation or inhibition of several humoral and genetic factors and other intracellular signals) [1]. These alterations may lead to impaired diastolic filling and systolic contraction of the left ventricle (LV) associated with clinical or subclinical cardiac dysfunction.

AH also affects the peripheral vasculature causing arterial endothelial dysfunction, accelerated vascular stiffening, increased progression of interstitial and perivascular fibrosis [2], vascular remodeling and reduction in the number of small arterioles (i.e. microangiopathy) [3]. Coronary microcirculation may be damaged either directly from vascular changes [2] or through the elevation of the LV diastolic pressure causing increased extravascular compression leading to impaired coronary flow reserve (CFR). Previous clinical studies have shown that impaired CFR may lead to exercise-induced myocardial ischaemia and to exercise-induced LV diastolic [4] and systolic dysfunction [5] in patients with AH.

Echocardiography plays a pivotal role in the assessment of the cardiac effects of AH. Conventional echocardiographic evaluation of hypertensive patients focuses on: (a) LV hypertrophy and geometry, (b) severity of diastolic dysfunction, (c) global systolic function, (d) left atrial volume and function and (e) the thoracic aorta [6]. The development of two-dimensional myocardial strain imaging with echocardiography (2DSTE) during the last decade has facilitated the simple and angle-independent measurement of LV myocardial deformation in the longitudinal, circumferential, and radial directions [7] as well as the assessment of LV rotational function [8]. Stress echocardiography with a vasodilator, such as dipyridamole (DIPSE), is used either to induce myocardial ischemia or to measure CFR in the left anterior descending territory; decreased CFR has been associated with the presence of significant microvascular and epicardial vascular dysfunction [9,10,11].

The aim of the current study was to investigate whether early (relatively healthy) hypertensive patients with normal LV ejection fraction may differ in 2DSTE indices at baseline and in response to a stress test such as dipyridamole, compared to healthy controls. Identification of indices that could differentiate early hypertensive patients would be helpful in risk stratification as well as monitoring of medical management of these patients. In the current study, the associations of 2DSTE indices and their changes following DIPSE with other studied parameters were also assessed.

Methods

Population

We prospectively evaluated 47 consecutive male patients with a history of AH between January 2012 and October 2013. Patients were followed-up at the Hypertension Clinic of the Nephrology Department of the University Hospital of Ioannina, Ioannina, Greece and had been on stable treatment for at least 3 months. Furthermore, we enrolled 20 healthy male subjects as a control group; participants in the control group belonged mainly to the hospital personnel. Subjects who were > 80 years old, had any history of coronary artery disease, left ventricular ejection fraction (LVEF) < 50%, any conduction disorder, poor echocardiographic window, any moderate-severe valvular heart disease were excluded from the study. Patients underwent a thorough physical examination and detailed echocardiographic analysis. The study protocol was approved by the local Ethics Committee. The study complied with the Declaration of Helsinki and all participants provided written informed consent.

Echocardiographic evaluation

Echocardiographic studies were performed by a single operator (DE). The echocardiographic evaluation was performed using a commercially available system (Vivid 7; GE Vingmed Ultrasound AS, Horten, Norway) as previously described [12]. Images were acquired from cine loops with high frame rates in standard parasternal and apical views and data were stored digitally in still images of high analysis (i.e. 1024 × 768 pixels) and in cine loop format for 3 consecutive analyzable beats.

Initially, a detailed basic echocardiogram was performed and all classic systolic and diastolic indices of LV function were assessed. Left atrial volume and LV mass were indexed to patients’ body surface area. All indices were assessed on the basis of the European Society of Cardiology and European Association of Cardiovascular Imaging guidelines [13]. Secondly, 2DSTE acquisition was performed.

Offline analysis was performed using a dedicated software for echocardiographic quantification (EchoPAC version 113, GE Healthcare Vingmed Ultrasound AS, Horten, Norway). The exact endocardial borders of the LV were manually traced and the region of interest was thus acquired. When necessary, the myocardial borders were manually corrected through this dedicated software. The timings of mitral and aortic valve opening and closure were defined by pulse wave Doppler of mitral and LV outflow tract. Tracking was accepted only if both visual estimation and EchoPAC related software analysis were adequate. If more than two myocardial segments at each view were not clearly visualized, the patient was excluded. However, there was not any exclusion by means of poor 2DSTE analysis. Global longitudinal and circumferential strain (GLS, GCS) were assessed for all patients. LV rotation was also assessed through acquisition of short axis views (apical, mid cavity and basal). LV TWIST was computed as the difference between basal and apical rotations. UNTWIST rate was then calculated as the peak negative derivative of time concerning twist, during diastole. Time to maximal UNTWIST was calculated as the interval of time from the peak of the R wave in the electrocardiogram of the patient to the maximal UNTWIST. Images were considered as inadequate for measurement of rotation and cardiac twist, when at least > 2 segments in each view were not adequately tracked.

Following the baseline echocardiographic examination, intravenous infusion of dipyridamole was started (0.84 mg/kg over 6 min). The assessment of CFR in the area of LAD coronary artery was performed before and after intravenous infusion. CFR was calculated as the ratio between hyperaemic and basal coronary flow (the highest three measurements were averaged for each of the parameters) (Fig. 1). A cut-off value of < 2 was used to discriminate significant LAD stenosis and/or microvascular dysfunction in the LAD area [14]. Two minutes after the end of dipyridamole infusion, a new echocardiographic evaluation was performed, which focused mainly on LV systolic and diastolic function indices. At the end of the dipyridamole infusion, 125–250 mg of aminophylline, depending on the patient, were administered to counteract the effects of dipyridamole. During dipyridamole infusion and recovery, all patients were monitored with electrocardiogram and blood pressure (BP) measurements. Beverages containing methylxanthines (e.g. coffee, tea, chocolate) were not allowed at least 24 h before the study.

Fig. 1
figure 1

The flow in the distal portion of the left anterior descending coronary artery is visualized by 2D colour Doppler in an apical modified foreshortened 2- or 3-chambers view. Coronary flow velocities are measured by pulsed-wave Doppler as a laminar flow signal directed towards the transducer pre (top) and post-dipyridamole infusion (bottom). Coronary flow reserve (CFR) is calculated as the ratio of pre- and post-dipyridamole coronary flow velocities

Full size image

In studies performed on two separate days (5–10 days apart) in 10 subjects by a single operator, the within-subject coefficient of variation of E/E′, GLS, TWIST, UNTWIST and CFR were 9.6%, 3.4%, 8.6%, 7.8% and 6.5% respectively.

Statistical analysis

Shapiro–Wilk test was used to identify continuous variables that were not normally distributed. Continuous data are presented as median (interquartile range) for not normally distributed variables and mean ± SD for normally distributed variables. Student’s t-test and Mann–Whitney U test were used to compare various parameters between hypertensive patients and healthy controls. Differences in various studied parameters between the two groups were adjusted for confounders using the General Linear Model. Paired t-test and Wilcoxon test were used to compare various parameters at baseline vs. after dipyridamole. Repeated measures ANOVA analysis was used to assess any differences in DIPSE-induced changes between hypertensive patients and healthy controls. The sphericity assumption regarding the interaction of time-point vs. group was not violated in any of the studied variables. Association analysis included the assessment of Spearman and Pearson correlation coefficients. p values were always two-sided and a value of p < 0.05 was considered significant. The SPSS statistical software package (IBM SPSS Statistics, Version 23) was used.

Results

Hypertensive patients were older, had higher BMI, systolic and diastolic BP (p < 0.05 for all) compared to healthy controls (Table 1). After adjustment for age and BMI, MAPSE septal and lateral, IVRT, Sl were lower while TvSv was higher in hypertensive patients vs. controls (p < 0.05 for all). Regarding 2DSTE indices, GLS [− 17.8 (− 19.4, − 16.4) vs. − 20.1 (− 22.4, − 18.3), p < 0.001] and TWIST [14.3 (9.5, 18.8) vs. 8.3 (7.2, 10.6), p = 0.002] were higher while UNTWIST rate [− 80.6 (− 101.3, − 62.5) vs. − 60.2 (− 75.0, − 50.8), p = 0.018] was lower in hypertensives compared to controls respectively; however, after adjustment for age and BMI only TWIST remained significantly higher in hypertensives (p = 0.021) (Fig. 2).

Fig. 2
figure 2

Myocardial strain indices, including Global longitudinal strain (GLS), TWIST and UNTWIST, in controls (left) and hypertensive patients (right)

Full size image
Table 1 Demographic, clinical and classic echocardiographic indices in hypertensives (n = 47) and controls (n = 20)
Full size table

Statistically significant associations of 2DSTE indices with other clinical and echocardiographic parameters at baseline in hypertensive patients and healthy controls are shown in Table 2. In hypertensive patients, lower GCS was significantly associated with lower UNTWIST rate (r 0.403, p = 0.005), higher TWIST was associated with lower UNTWIST rates (r − 0.303, p = 0.043) and higher time to peak TWIST was related to higher time to UNTWIST (r 0.349, p = 0.025) (Fig. 3a–c). In healthy controls, lower CCS was significantly associated with higher TWIST (r − 0.627, p = 0.003), while time to peak TWIST was positively associated with time to UNTWIST (r 0.596, p = 0.006) (Fig. 3d, e).

Fig. 3
figure 3

Scatterplots showing the individual data for the statistically significant associations of (A) two-dimension myocardial strain indices at baseline in hypertensive patients (a–c) and healthy controls (d, e) and (B) coronary flow reserve with dipyridamole-induced changes in two-dimension myocardial strain indices (f) in healthy controls

Full size image
Table 2 Correlations of 2D myocardial strain indices of left ventricle with other clinical and echocardiographic parameters at baseline in hypertensive patients and healthy controls at significance level p < 0.1
Full size table

DIPSE-induced changes

No wall motion abnormalities were seen with DIPSE in either healthy controls or hypertensive patients. Heart rate increase with dipyridamole was similar in both groups (35% in controls vs. 26% in hypertensives, p = 0.189). CFR values were significantly higher in healthy controls compared to hypertensive patients even after adjustment for age and BMI [4.14 (IQ 3.39, 4.54) vs. 2.53 (IQ 2.03, 2.95) respectively, p = 0.001]. All controls and 87% of the hypertensive patients had a normal CFR value i.e. >2.0. Only in controls was higher CFR associated with lower UNTWIST rate (r − 0.453, p = 0.045) (Fig. 3f); no other association of CFR with 2DSTE indices was observed in either hypertensives or controls.

DIPSE-induced changes in various studied echocardiographic parameters in hypertensive patients and healthy controls are shown in Table 3. DIPSE caused a significant increase in LVOT-VTI, LVEF, MAPSE septal and lateral, Sm and Sl and a decrease in IVRT, GLS and GCS (p < 0.05 for all) in both groups. Furthermore, DIPSE caused a significant increase in E/E′ and TWIST and a decrease in MVDT, UNTWIST rate and time to UNTWIST (p < 0.05 for all) only in hypertensive patients. On the other hand, DIPSE resulted in a significant decrease in time to peak TWIST (p = 0.02) only in healthy controls. DIPSE-induced changes in various studied parameters did not differ significantly between the two groups after adjustment for age and BMI (RMANOVA, p > 0.05 for all). The associations of changes among various studied echocardiographic parameters following DIPSE in hypertensive patients are presented in Table 4. No significant associations of DIPSE-induced changes among various echocardiographic parameters was found in the control group (data not shown).

Table 3 Comparisons of echocardiographic parameters at baseline and following dipyridamole administration in hypertensives patients and healthy controls
Full size table
Table 4 Associations of changes among various studied echocardiographic parameters following dipyridamole infusion in hypertensive patients
Full size table

Discussion

The current study systematically investigated the combination of various novel echocardiographic biomarkers and the stress response to dipyridamole including CFR, in the assessment of early hypertension. We demonstrated that relatively healthy patients with early hypertension showed subclinical deterioration in several echocardiographic indices of systolic (LVEF, MAPSE septal and lateral, S lateral and GLS) and diastolic (IVRT, E′ and E/E′) LV function compared to healthy controls. However, hypertensive patients were older (by ca. 10 years) and had a higher BMI, factors that may affect all the above mentioned indices [15,16,17,18]. Indeed, after adjustment for age and BMI, significant differences remained between hypertensives and healthy controls in MAPSE septal and lateral, S lateral and IVRT, but not in GLS. It has been previously shown that impairment of LV long-axis systolic function occurs at the very first stages in many heart diseases including AH; this has been considered as a useful tool in the evaluation of the hypertensive patient even before the appearance of LV hypertrophy [19,20,21,22]. Changes in LV longitudinal systolic function may also be affected by the presence of other factors such as increased body weight and age, as suggested in our study; therefore, these factors should be taken into account when evaluating patients with AH. Similar to our findings, no deterioration in circumferential systolic function (i.e. GCS) has been previously reported in relatively healthy hypertensive patients without overt cardiomyopathy or heart [21, 23, 24] while GCS deteriorates in the presence of LV hypertrophy [24]. Furthermore, diastolic function has been shown to deteriorate in hypertensive patients in accordance to previous studies [25].

Moreover in hypertensive patients, lower GLS (i.e. improved) was currently associated with lower LV mass, improved indices of systolic LV function and cardiac output (MAPSE and LVOT VTI) as well as ameliorated diastolic function as assessed by higher E′ velocity. These findings have been previously described in hypertensive populations [24, 26]. The close relation of systolic and diastolic function in hypertensive patients has been attributed to myocardial changes observed in the presence of AH related to increased loading conditions, increased wall stress and myocardial fibrosis [27,28,29]. Importantly, neither GLS nor GCS was related to LVEF in hypertensive patients in contrast to healthy controls; this finding may strengthen the use of myocardial strain indices as measures of subclinical deterioration of LV systolic function long before the overt decrease of measures of global systolic function such as LVEF and the appearance of clinical symptoms.

Hypertensive patients showed higher TWIST and lower UNTWIST rate compared to healthy controls in the current study; TWIST, and not UNTWIST rate, remained significantly higher in hypertensives even after adjustment for age and BMI. Although unanticipated, this finding has been previously reported in hypertensive populations with mild diastolic dysfunction without advanced heart disease or heart failure [30,31,32] while in cases of hypertensive patients with advanced LV hypertrophy and diastolic dysfunction or reduced systolic LV function, TWIST was significantly reduced [30, 31, 33]. Similarly, lower UNTWIST rate has been also described in hypertensive patients early in the course of AH [30]. It has been suggested that this “hyper-rotation” phenomenon may be a compensatory mechanism early in patients with AH to maintain a relatively normal global systolic LV function (i.e. normal LVEF). Twisting and untwisting are tightly coupled since the greater the LV twist the more potential energy is stored for subsequent higher LV untwisting rate [34, 35] leading to preserved LV filling in the early stages of hypertensive cardiomyopathy. Pathophysiologically, the early subendocardial dysfunction observed in hypertensive patients (i.e. decreased longitudinal systolic LV function) is probably compensated by subepicardial myocardial fibers function and the imbalance results in increased TWIST [36]. However, as the disease progresses, subepicardial myocardium is affected by pathological changes, and therefore LV twist may be reduced in the later stages in AH patients accompanied by decreased LVEF and advanced diastolic dysfunction.

In hypertensive patients, it was currently shown that various anti-hypertensive medications were associated with potential beneficial effects on myocardial deformation mechanics. Circumferential myocardial systolic LV function (i.e. GCS) was associated with the use of diuretics while UNTWIST rate was better in patients receiving renin-angiotensin system inhibitors. It has been previously reported that telmisartan (an angiotensin II receptor inhibitor) improved longitudinal and circumferential strain as well as TWIST without affecting UNTWIST rate [37]. The involvement of renin-angiotensin system in the promotion of myocardial fibrosis and the increase of afterload of LV is well established [27, 38]. The potential benefit of anti-hypertensive medications on regional myocardial function is probably related to their anti-fibrotic and vasodilating effects; whether certain medications may affect specific aspects of myocardial deformation mechanics is not known and further studies are needed to address this issue.

Dipyridamole stress echocardiography is a highly reproducible method to evaluate CFR in various populations [39]. In the absence of significant coronary stenosis, CFR is a marker of coronary microcirculatory function [11]. It has been previously shown that CFR is impaired in the early stages of hypertension [4] as well as in the pre-hypertension stage [40] before LV hypertrophy is apparent [41]. Similarly, we demonstrated that CFR is decreased in relatively healthy patients with AH even after adjustment for potential confounders such as increased body weight and age; almost half of the hypertensive cohort presented with abnormal CFR i.e. < 2.0. Decreased CFR (measured with DIPSE) has been associated with mortality in patients with known or suspected coronary artery disease [42] although in patients with AH the prognostic role of CFR has not been studied.

Furthermore, DIPSE induced several changes in systolic and diastolic function indices in patients with AH. A significant improvement of left ventricular systolic function, including classic indices of LV systolic function (i.e. LVOT VTI, LVEF, MAPSEsep, MAPSElat, Sm, Slat) and myocardial strain indices (i.e. GLS, GCS, TWIST) was observed while an improvement in UNTWIST rate was also shown. These changes did not differ in a statistically significant manner between healthy controls and hypertensives. Similar changes have been previously reported in other populations as well [12] and underline the importance of DIPSE as a stress test for systolic and diastolic myocardial reserve [12, 43, 44]. Interestingly, DIPSE-induced improvement in GLS was significantly associated with higher CFR suggesting a close pathophysiological relationship between coronary microcirculatory function and myocardial long-axis systolic function, as shown in previous studies [45]. On the other hand, DIPSE-induced increase in TWIST was associated with greater DIPSE-induced increase in E/E′ ratio and decrease in E′; this finding suggests a complex interplay between diastolic function and torsional deformation in hypertensive patients. As discussed above, increased LV twist acts as a compensatory mechanism in patients with AH while it may also be a marker for the early stage of diastolic dysfunction in these patients.

Conclusion

The current study aimed to identify novel echocardiographic indices that could potentially differentiate early hypertensive patients from healthy controls; these indices could be proven to be helpful in risk stratification and monitoring of medical management of these patients. The study showed that well-controlled hypertensive patients with normal LV mass and geometry have only mild differences in echocardiographic indices compared to controls, using both conventional and speckle tracking echocardiography. These mild differences as well their clinical significance need to be validated in larger studies due to the limitations related to the sample of the current work. The higher TWIST in AH patients compared to healthy controls as well as the association of DIPSE-induced increase in TWIST with greater DIPSE-induced increase in E/E’ ratio may suggest that a ‘hyper-rotation’ early in AH could serve as a mechanism to compensate for a stress-induced impairment of diastolic filling and to maintain a relatively normal global systolic LV function. CFR was found to be decreased in AH patients, albeit within normal range, even after adjustment for potential confounders. The association of CFR with DIPSE-induced improvement in GLS indicates a close pathophysiological relationship between coronary microcirculatory function and myocardial long-axis systolic function. Further research is needed to explore the progression of these observations with longer duration of AH, the occurrence of comorbidities and the effect of treatment with various antihypertensive agents. The prognostic role of DIPSE-induced changes in myocardial strain indices should also be further investigated.