Hypertension (HTN) is a major cause of heart failure but the precise pathways by which HTN leads to heart failure are not resolved. Several pathways have been proposed that lead to heart failure with reduced ejection fraction (HFrEF) or heart failure with preserved ejection fraction (HFpEF) [1]. HFpEF comprises nearly half of all patients with heart failure and is growing in prevalence [2]. HTN is frequently associated with HFpEF [3] implicating HTN as an important potential cause of HFpEF.

One advancement in the field of assessment of cardiac function is the introduction of the concept of cardiac strain and the techniques to assess it. Myocardial strain is a dimensionless index of length change between two given points, which reflects the degree of myocardial deformation [4]. Cardiac contraction is due to the shortening of myocardial fibers that have different orientations at various levels of the heart [5–7]. Myocardial fibers have a longitudinal arrangement on the oblique parts of the heart and a circumferential arrangement on other parts of the heart [6]. Contraction of myocardial fibers that have different orientations produces deformation in different directions that translates into left ventricular (LV) shortening that can be measured as percentage longitudinal, circumferential and radial strain [8, 9]. Longitudinal strain evaluates the apex-base deformation, circumferential, strain evaluates circumferential deformation while radial strain represents radial thickening of the myocardium [8, 10]. The distribution and angulation of myofibers in all layers of the heart contribute to each of these three kinds of strain [8]. It has only been recently possible to readily assess changes in myocardial contractility in the different orientations in the heart. Speckle tracking echocardiography permits a quantitative assessment of myocardial motion that is reflective of different layers of the heart [11]. This technique provides accurate and angle-independent measurements of LV dimensions [12]. There is evidence that assessment of myocardial strain may be superior to the LV ejection fraction (LVEF) as a predictor of major adverse cardiac events [13, 14].

Assessment of myocardial strain has advanced the assessment of cardiac functional changes during chemotherapy for various malignancies which involves the use of potentially cardiac toxic drugs [15]. The question is the extent to which myocardial strain can identify abnormalities in cardiac contractile function in patients with HTN and HFpEF. The objective of this study is to determine whether LV strain is significantly different in patients with HTN and HFpEF compared to patients with HTN or individuals without HTN. In addition, myocardial strain will be compared to the standard assessment of cardiac contractility namely LVEF.

The initial search produced 103 studies and with one identified through other sources, 104 studies were screened and then relevant studies were assessed and evaluated for inclusion (Figure 1). Eight studies were included for quantitative synthesis (Table 1) [17, 19–26]. Excluded were studies that focused on HFpEF in animal models [23].

LVEF was not significantly different in patients with HFpEF compared to individuals without HTN or other conditions (Figure 2) as well as compared to those with HTN (Figure 3).

GLS was significantly (P < 0.0001) different in patients with HFpEF and HTN compared to control individuals without HTN or other conditions (Figure 4). The odds ratio (OR), the ratio of the odds in the HFpEF with HTN vs. the HTN group, was 8.9 (z = 16.0, P < 0.001). Although there was significant heterogeneity between studies, all but one study showed a significant difference and it was only the magnitude of the difference which varied between studies. Evaluation for publication bias using the classic fail-safe N showed a relative lack of publication balance as it would take 518 studies to reverse the significant findings. The values for trim and fill were for the fixed model 1.205 (1.046–1.365), for random model 1.588 (0.874–2.301) and the values were unchanged using trim and fill.

GLS was also significantly (P < 0.0001) different in patients with HFpEF and HTN compared to patients HTN without HFpEF (Figure 5). Although there was significant heterogeneity between studies, most studies showed a significant difference. OR in the HFpEF with HTN vs. the HTN group, was 3.03 (z = 8.9, P < 0.001). Evaluation for publication bias the classic fail-safe N was 126 indicating that it would take 126 null studies to reverse the significant findings. The values for trim and fill were for the fixed model 0.60631 (0.457–0.756), for random model 0.715 (0.259–1.172) and the values were unchanged using trim and fill analysis.

GCS was significantly (P < 0.0001) different in patients with HFpEF and HTN compared to individuals without HTN or other conditions (Figure 6). OR was 6.1, (P < 0.001). Evaluation for publication bias the classic fail-safe N was 221, indicating that it would take 221 null studies to reverse the significant findings. The values for trim and fill were for the fixed model 0.997 (0.815–1.178), for random model 1.449 (0.630–2.268) and the values were unchanged using trim and fill analysis.

GCS was also significantly (P < 0.0001) different in individuals with HFpEF and HTN compared to individuals with HTN (Figure 7). There was only one exception to the overall findings [17]. There was significant heterogeneity between studies. OR in the HFpEF with HTN vs. the HTN group, was 4.7 (P < 0.001). However, fail-safe N was 121 indicating that it would take 121 null studies to reverse the significant findings. The values for trim and fill were for the fixed model 0.850 (0.676–1.024), for random model 0.921 (0.429–1.41) and the values were unchanged using trim and fill analysis.

There were few studies that evaluated GRS. GRS was not quite significantly (P < 0.054) different in patients with HFpEF and HTN compared to individuals without HTN or other conditions (Figure 8).

GRS was significantly (P < 0.01) different in individuals with HFpEF and HTN compared to individuals with HTN (Figure 9).

The present study showed that individuals with HFpEF had reduced LV strain while there was no difference in LVEF. This finding was evident regardless of whether the comparator was individuals free of HTN or those with HTN. GLS and GCS provided the best separation between patients with HFpEF and HTN compared to individuals with HTN without HFpEF. GRS, however, did not differentiate patients with HFpEF and HTN from individuals without HTN.

This meta-analysis demonstrated a consistent and significant reduction in GLS in HFpEF in the absence of any change in LVEF. The complex LV architecture allows a fiber shortening of only 20% to be transformed into a 60% change in LV volume [9] so that small changes in LV strain are meaningful. GLS is the change in deformation from the apex to base so that a small percentage change is reflective of a reduced LV contractility. In animal studies, GLS correlates strongly with LV +dP/dt_max [27] so that reduced GLS is indicative of reduced ventricular +dP/dt_max. GLS has been demonstrated to identify early and subclinical LV dysfunction [28, 29]. Subendocardial and subepicardial layers are purported mainly responsible for longitudinal strain [8] and subendocardial layers are more likely to detect impaired cardiac function because they are more vulnerable to myocardial damage.

Longitudinal data support the value of GLS. In patients with cardiac disease, GLS is better than LVEF for predicting major adverse cardiac events [13]. In the general population, GLS provides independent and incremental prognostic information regarding long-term risk of cardiovascular morbidity and mortality [30]. GLS is better at predicting the composite cardiovascular outcome and heart failure incident beyond the Framingham risk score, the SCORE risk chart, and the modified American College of Cardiology/American Heart Association (ACC/AHA) pooled cohort equation [30]. Patients with preserved LVEF but impaired LV GLS have an increased risk of hospitalization for heart failure and all-cause mortality [31].

This meta-analysis demonstrates a consistent and significant lower GCS in HFpEF in the absence of any change in LVEF. GCS is the change in circumferential deformation or the change in length in systole along the circumferential axis of the LV. A smaller GCS is an indicator of a reduction in cardiac contractility. Mathematical modeling of LV contraction suggested that both longitudinal and mid-wall circumferential shortening contribute different extents to LV contractility depending on the degree of abnormality of myocardial shortening [32]. Circumferential shortening in systole compensates for impaired longitudinal shortening that occurs first when LV systolic dysfunction occurs [33]. GCS provides incremental independent prognostic value after considering LVEF, and late gadolinium enhancement [34].

This meta-analysis did not demonstrate a consistent and significant reduction in GRS in HFpEF. GRS is a result of fiber shortening in all layers, augmented by thickening and inward displacement of the myocardium [9]. Because GRS evaluates the change in radial deformation, GRS is an indicator of LV contractility in this dimension. In animal studies, the correlation between GRS and LV +dP/dt_max was weaker than the correlation with GLS [27], suggesting that GRS is not as good an indicator of cardiac contractility and explains why GLS and not GRS is highly significantly abnormal in HFpEF.

Szelényi et al. [35] studied patients with HTN with or without echocardiographic evidence of diastolic dysfunction—a component of HFpEF, but did not specifically study HFpEF. They found that GLS but not GCS or GRS was significantly lower in HTN compared to control and there was no significant difference in GCS or GRS between HTN with or without echo evidence of diastolic dysfunction [35].

Patients with HFpEF have limited exercise capacity which has been ascribed to increases in LV diastolic pressure with exercise. The demonstration of associated impairment in LV strain in HFpEF suggests another element that compromises cardiac output during exercise. A concept that is consistent with the idea that limited myocardial systolic function reserve may be underlying limited exercise capacity in HFpEF [36].

The superiority of LV strain over LVEF should not be unexpected. It is generally accepted that a major limitation of LVEF as an indicator of cardiac function is the dependence of LVEF on preload, afterload, chamber size, thickness, as well as its relative insensitivity to identify patients with early-stage heart failure [37, 38]. The question of whether LV systolic function is reduced in HFpEF has been controversial. Baicu et al. [39] compared 75 patients with HFpEF to 75 normal control subjects and concluded that HFpEF was associated with normal LV systolic performance. In contrast, other investigators have found that approximately one-quarter of patients with HFpEF have subtle systolic dysfunction. Other data suggest that the progression to HFpEF is mediated by processes that impair both myocardial contractility and increase myocardial stiffness [40].

This study raises the possibility of redefining the classification of heart failure. Currently, heart failure is divided into HFpEF, HFrEF, and heart failure with mid-range EF and heart failure with improved EF. We have previously proposed the classification of HFrEF with diastolic dysfunction (HrEFwDD) [41]. The present study highlights the presence of reduced systolic function in HFpEF. In a model of HFpEF, the Dahl salt-sensitive rat, reduction in GLS occurs early and continues to deteriorate until heart failure develops [42]. This study raises several novel perspectives. First, it suggests the possibility of a classification of heart failure based solely on the myocardial strain with heart failure severity being ranked according to the amount of reduction in LV strain. Another approach is that HFpEF could be sub-classified according to the degree of impairment in LV stain. A classification that includes HFpEF with small reductions in systolic function, identified by LV strain, may be beneficial as it may indicate the development of HFrEF. It would also stimulate clinical trials to test whether drugs that are effective in HFrEF will be beneficial in this subset of patients with HFpEF.