Left ventricular strain and peak systolic velocity: responses to controlled changes in load and contractility, explored in a porcine model

Since its introduction tissue velocity echocardiography (TVE) has opened new possibilities for non-invasive quantification of myocardial function [1]. Tissue Doppler echocardiography is an established cardiac diagnostic method for measurement and characterization of both the systolic and diastolic cardiac dysfunction [2‐5]. Myocardial velocity is the most widely used and validated parameter from this method [6‐8]. Through further derivation and integration of tissue velocities, additional parameters like displacement, strain rate and strain can be calculated. Strain is a dimensionless quantity, which is the product of stress. It represents the fractional change (or percentage) from the original or unstressed dimension. Positive strain indicates expansion or lengthening and negative strain compression or shortening.

An ideal index of myocardial contractility should be independent of loading conditions (preload and afterload), heart size and mass, and sensitive to changes in inotropy [9]. In this study, we have focused only on the issue of systolic velocities and load, since the load-dependence of diastolic tissue velocities has been well demonstrated [10, 11].

The issue of load-dependency and tissue velocities during systole has been raised earlier, and clinical findings have been published that support the idea that load reduction can lead to decrase in systolic velocities [12] or alternatively increase in systolic velocity [13‐15] , and in these cases, the authors have suggested mechanism other than load change (decrease) that lead to higher tissue velocities. Some preliminary findings for controlled load reduction in an experimental setting have shown that systolic velocities may decrease in conjunction with load decrease [16]. In the pediatric population with congenital heart disease, acute load reduction with inferior vena cava occlusion caused significant decrease in peak systolic velocities in the right ventricle [17].

We hypothesized that TVE systolic measures of the left ventricular function, peak systolic velocity (PSV) and systolic strain, would change in relation to loading changes. We further hypothesized that PSV and strain would behave in the same fashion during pharmacologically altered contractile status. We aimed to test these hypotheses in an experimental large animal model, which included exposure to positive and negative inotropic drugs and carefully controlled transient load changes. We also aimed to compare tissue velocities and strain measured from different directions with respect to myocardial axis.

After approval from the Umeå Regional Animal Experimental Ethics Committee, and in conformation with the Guide for the Care and Use of Laboratory Animals (US National Academy of Sciences, 1996, USA), 13 juvenile Yorkshire/Hampshire pigs with a mean weight of 36.7 kg (SD = 4.1) were anaesthetized and instrumented, using methods that have been well described previously [18].

The anaesthetized pigs all demonstrated central circulatory parameters consistent with health during rest before the start of the protocol. All 13 animals completed the protocol, with simultaneous tissue velocities echocardiography measurements together with LV pressure volume results for each sequence. Load changes were achieved by the VCBO, where the first eight beats in the sequences (Figure 1) are analyzed and a clear load alteration (reduction in end-diastolic volume, as well as reduction in stroke work) is shown, and this was consistent with all 3 inotropic conditions and in both projections (radial/short axis and longitudinal/long axis) (Figure 2).

Hemodynamic results show that the animals had their VCBO sequences with start and end ventricular pressures and volumes not in extreme ranges (Figures 1 and 2). Inotropic interventions, both positive and negative, achieved significant changes in contractile status as exemplified by change in PRSW from baseline 74.1 ± 10.2 mmHg to 104.0 ± 14.8 mmHg in positive inotropic status and to 51.3 ± 12.9 mmHg in negative inotropic status (Table 1).

The PSV results during the unloading sequence of 8 consecutive beats remained relatively unchanged in both short and long axis at baseline (Figure 3), and behaved in the same manner in the pharmacologically-induced positive inotropic condition. In negative inotropic state, in the short axis results, no significant change in PSV was observed, though in the long axis there was a significant decrease in PSV during the course of the preload reduction sequence.

Systolic strain findings (Figure 4) showed that during the unloading sequence of 8 consecutive beats, peak systolic strain had a tendency for increase in short axis in baseline and positive inotropic condition, but reached significance only in negative inotropic condition. For the long axis results, a reduction in systolic strain was observed which reached statistical significance in baseline and negative inotropic conditions.

The inotropic interventions led to differences in the tissue\ velocity parameters (Table 2), where PSV increased with the positive inotropic intervention, for both the longitudinal and radial axis measurements. PSV decreased, at least for the radial measurement, with the negative inotropic intervention. There was no clear relation observed for systolic strain and inotropic status. There was no difference between the absolute strain levels for the negative and positive inotropic interventions (Figure 4).

When the radial and longitudinal axes for PSV were compared, there were no differences between the longitudinal axis and radial axis measurements for the resting control and negative inotropic measurements. For the positive inotropic measurement, the PSV average for the longitudinal axis measurement was greater than for the radial axis measurement.

The main findings in this study were that PSV was load-independent in this model. The tissue velocity parameter PSV increased with inotropic status increase, and decreased with the negative inotropic intervention. It has been recognized for many decades that there is a strong relation between load and ventricular performance, though it is not yet clear how strongly related tissue velocities are on ventricular performance. These results support the idea that PSV changes, increases or decreases, reflect changes in contractile status (for example as brought about by inotropic interventions in this model).

Previous reports have been mixed concerning the relation of tissue velocity to load: some pulsed Doppler velocities in the AV plane during load alteration results have supported the idea that PSV is load-independent [16, 25‐27], others that PSV is load-dependent [28]. Some have performed these studies using human subjects, where although a load or inotropic intervention can be implemented, the actual load-altering effect of these interventions can be difficult to assess and verify [29]. In this experimental model, both loading conditions and contractile status were carefully controlled, and altered independently for serial comparisons.

The results for strain in relation to load supported the idea that strain is load-dependent, in contrast to PSV. Though not reaching statistical significance for all the inotropic conditions, strain appeared to increase generally with decreasing load, in this model, and this agrees with some previously published results [25, 30]. There was relatively more variation in the grouped measures for strain (grouped by beat in the preload alteration sequence) compared to PSV. There was relatively less variability in the strain assessments in the longitudinal axis compared to that from the radial axis, and this was expected since conducting measurements in the same tissue plane with the radial axis is more challenging. Strain did not seem to be as affected by inotropic state compared to PSV, also supporting the idea that strain is not a good method for identifying changes in contractile function [31, 32].

There are measurement-technical aspects which can affect the quality of the signals. We analyzed velocities which were measured in the basal aspect of the interventricular septum. Signal from the ventricular lateral wall was also acquired, but there was often inadequate imaging quality during the VCBO, so that we could only consistently use the septal assessment for futher analysis. It is possible that there was some minimal drift of the strain curves over time, since the (Echopac) proprietary software uses drift compensation. We examined signals both with and without drift compensation. Without this software-driven compensation the baseline was consistently drifting. We are confident that the absolute amount of strain was adequately assessed. Strain rate was examined for all VCBO sequences, thought the signal had enough ‘noise’ that clear maximum strain rates were not reliably identified. Therefore, no strain rate results are shown (Figure 5).

This study was conducted with an undisturbed thorax. In cases where the thorax and pericardium are open, it is likely that there is an effect on wall motion and tissue velocities [33]. There have been reports that suggest that transthoracic echocardiographic studies are limited in pigs [31], but imaging was definitely adequate in our experiments with possibly a more modern generation of hardware and software. Also, in a healthy heart, the axis of shortening is not only in the long or short axis [34, 35]. It is reasonable perhaps measure in one particular area and then extrapolate to other regions of the heart. In a diseased or injured heart, inhomogeneity of tissue velocities may be known or presumed. In our model, there was not injury. Still, this is not readily generalizable to the clinical setting.

The clinical relevance [35] of these findings is that PSV shows promise as a potential bedside means to obtain a quantitative assessment of contractile function during serial measures in a patient. A recent expert group [36] has suggested that pre-systolic myocardial loading must be taken in account for any clinical application of these parameters. There is a well-recognized degradation of myocardial function, for example with age or illness, which can be followed. Alterations up and down in the range of potential contractile status or function for any individual are part of normal myocardial function, and these can make single beat assessment of general myocardial condition very difficult, since despite implementing an alteration in load, it is much more difficult to absolutely control contractile status, as long as the autonomic nerve system is responsive. On tradition has been to measure single beat parameters in patients at rest, but this may not be adequate for assuring repeatable autonomic nerve system conditions for serial measures. Contractile status needs to be quantitatively assessed, taking advantage of situations where the effect of beat-to-beat loading changes on mechanical systolic function can be approximated. It is possible to do this with high precision measurement of load and mechanical effect for families of beats where load is slightly different for each beat (as has been done here), though the current methodologies for direct ventricular pressure and volume measurements are highly invasive.

This type of assessment is needed today in settings where myocardial dysfunction is suspected, or where inotropic drug intervention is contemplated. Peak systolic velocity shows promise as a relatively load-independent parameter, which reflects systolic mechanical myocardial function, and more clinical experience needs to be gathered.

In summary, these results showed that peak systolic velocities showed a strong load-independence during acute load reduction, though this was not so for strain. We conclude that PSV is a clinically robust parameter of LV regional and global performance under changing load. Peak systolic strain seems to be load-dependent, and has no clear relation to inotropic changes in serial measures. These findings support a broader use of PSV in serial measures in patients to assess changes in ventricular function.