Left ventricular twist mechanics during incremental cycling and knee extension exercise in healthy men

During visit 2, participants undertook a submaximal bout of cycling for echocardiographic assessment. Following a 10 min rest on the semi-supine ergometer (eBike-L, ergoline GmbH, GE Healthcare), participants performed incremental exercise of 2 × 5 min stages at 30 and 60% W _max, where the last 3 min of each stage was used for echocardiographic data collection (Stohr et al. 2012). During the data collection period, a cadence of 60 rpm was maintained to limit upper body oscillation, with HR using a pulse oximeter (9590, Nonin Medical, Netherlands), and manual blood pressures recorded at the end of each stage.

Following 10 min seated rest, participants then lay supine onto an isokinetic dynamometer (Kin-Com 125E Plus, Chattecx Corporation, Chattanooga, USA) to conduct all isometric contractions. Participants undertook an isometric contraction of the quadriceps with the knee extension angle fixed at 130°. The participant’s dominant leg was used with the joint centre aligned with the axis rotation point of the crank arm (Kong and van Haselern 2010), an ankle pad connected to the load cell strapped proximal to the malleoli with a thigh and hip strap to prevent unnecessary limb movement while avoiding restricted blood flow (Marginson and Eston 2001). Participants were instructed to keep their back flat and generate force from the quadriceps. A warm-up of 5–10 submaximal contractions was performed prior to the MVC assessment, determined as the greatest of three maximal contractions (Adler et al. 2008). Each maximal effort was 5 s in duration with a 1 s transition period from rest and each separated by 2 min passive recovery (Bojsen-Moller et al. 2005). Following a 5 min rest, participants completed an incremental protocol consisting 2 × 15 s isometric contractions, separated by 2 min passive recovery, at intensities corresponding to 40 and 75% MVC to obtain a short-axis apical and basal images at both intensities. Pilot work found 15 s was the shortest contraction time possible, whilst still able to obtain SBP and DBP measurements. An upper intensity of 75% MVC was chosen based on previous findings of an unavoidable Valsalva manoeuvre at intensities ≥80% MVC (MacDougall et al. 1992). Participants were verbally instructed to breathe normally throughout the contraction, and refrain from holding their breath to avoid the performance of a Valsalva manoeuvre and participants were also reminded throughout the exercise to breath normally. A visual digital display was provided for participants to maintain the given intensities. Five cardiac cycles were recorded at the end of each contraction, whilst HR and manual blood pressures were obtained at the termination of contraction.

Standard parameters of LV structure and function were assessed only at baseline (rest) from previously saved images. The parasternal long axis view was used to determine LV internal measures, then used in the calculation of relative wall thickness and LV mass, in accordance with the American Society of Echocardiography and scaled to body surface area to obtain LV mass index (Lang et al. 2015). The apical four and two-chamber views enabled the generation of LV end-diastolic and end-systolic volumes using Simpson’s Biplane method and conventional systolic function, stroke volume and ejection fraction. Cardiac output was determined as the product of HR and stroke volume. Standard diastolic function was assessed using the apical four-chamber view for mitral inflow velocities during early and late diastolic filling, their ratio and deceleration time. Further, using pulsed-wave Doppler, lateral and septal tissue velocities were obtained at the mitral annulus during systole, early diastole and atrial systole from a spectral trace.

Twist mechanics were measured at rest and during both submaximal cycling and isometric knee extension protocols. To account for tachycardia and prevent potential under sampling, image depth and sector width were adjusted to alter frame rate (range 70–80 frames per second), so temporal resolution was maximised but the ensuing reductions in spatial resolution were minimised. The frame rates concur with past work having used STE during submaximal dynamic cycling (Donal et al. 2011; Doucende et al. 2010; Lee et al. 2012) and IHG (Balmain et al. 2016; Stefani et al. 2008; Weiner et al. 2012). Frame rates remained constant within subjects across all conditions. The parasternal short-axis views were used to acquire basal and apical imaging planes. The basal level was determined as the highest imaging plane at which full myocardial thickness was present with the observation of surrounding mitral valve at end-systole and positioned as circular as possible with no visible papillary muscles (Weiner et al. 2010a). Given that apical rotation shows proportional increases with progressive caudal transducer movement (Stewart et al. 2016), apical images were captured proximal to the end-systolic luminal obliteration of the LV cavity with as much accuracy as possible (van Dalen et al. 2008). When an insufficient apical image was found following distal transducer movement, a second method of obtaining a ‘true’ apical view was utilised. Beginning in the apical four-chamber view, the transducer was then tilted upwards with attention paid to ensuring the lumen was as circular as possible before image capture (Stohr et al. 2016).

Images were analysed offline using semiautomated software (EchoPac software, GE Healthcare, UK) by one investigator. The endocardial border was manually detected and initially traced prior to automatic software tracking. The region of interest was manually adjusted until the epicardial border was correctly aligned to encompass the entire LV wall thickness whilst avoiding the echogenic pericardium (Mor-Avi et al. 2011). Data files were imported into an Excel spreadsheet (Microsoft Corporation, Washington, USA) where a cubic spline add-in (SRS1 Software, Boston, USA) was implemented to generate 300 data points for both systole and diastole. Due to variances in HR, rotations were normalised to 5% increments of systolic and diastolic cycles with the pre-defined aortic valve closure indicating end-systole (100% systole). Peak LV twist during ejection was determined as the difference between basal and apical rotations, expressed in degrees (°) (Santoro et al. 2014a); whilst peak untwisting velocity (PUV) was considered the largest negative deflection following peak twisting velocity (PTV) (Stohr et al. 2011). Time-to-peak untwisting velocity and the percentage of diastole that PUV occurred (%diastole) were also recorded.

All participants completed the full protocol and echocardiographic data was determined in all participants at rest. However, cardiac images during all exercise conditions could not be obtained for one participant leading to exclusion from the study, resulting in a total sample of twenty-six participants available for statistical analyses. Furthermore, due to poor image quality, twisting mechanics data were not acquired during cycling at 30% W _max (n = 2) and 60% W _max (n = 3) and during knee extension at 75% MVC (n = 3). Consequently, an equal number of participants were required to determine net twist mechanics from apical and basal rotations which led to 22 and 23 participants during cycling and knee extension, respectively. A full data set was available for HR during both types of exercises (n = 26). An inability to obtain SBP and DBP during both intensities of cycling and knee extension (n = 1) and DBP at 75% MVC (n = 1), enabled 25 participants for SBP, DBP and MAP during cycling, whilst during knee extension 25 were available for SBP and 24 for DBP and MAP. Baseline participant characteristics and haemodynamic measures for the full cohort are presented in Table 1 and conventional echocardiographic structure and function shown in Table 2 (both n = 26).

To establish the quality of images gathered during cycling and knee extension exercises, a grading system was employed by the sonographer rating images 1–4 (1 = excellent—all endocardial and epicardial boarders clearly defined and successful tracking of all segments; 2 = good—fully visible wall with successful tracking of at least five segments but image considered ‘hazy’; 3 = poor—areas of image undefined, but successful tracking of at least four out of six segments; 4 = no image available or unsuccessful tracking). Images were considered excellent good quality from 65% during cycling at 30% W _max and knee extension at 40 and 75% MVC, whilst 54% of images at 60% W _max.

Findings that systolic twist initially increased from rest to submaximal exercise concur with past work (Doucende et al. 2010; Lee et al. 2012; Stohr et al. 2012). Systolic rotations then remained unchanged following a further increase in intensity, suggesting a plateau in the ‘absolute’ magnitude of net rotation at moderate intensity (between 30 and 60% W _max) and agree with our first hypothesis. In addition, these observations support the findings of the only previous investigation to use intensities greater than 40% peak power (Stohr et al. 2011). Mechanical constraints of the LV could be limited by multiple factors which may have profound implications on stroke volume (Stohr et al. 2011), including (1) pericardial constraints preventing further ventricular distension, (2) maximum systolic tissue compression, (3) increased blood pressure given the known afterload-twist interaction (Dong et al. 1999), and (4) an upper limit in titin phosphorylation. Acute phosphorylation during dynamic exercise may contribute to the Frank-Starling mechanism (Muller et al. 2014); therefore, an upper limit of phosphorylation may effectively limit twist, at least in recreationally active individuals. A combination of these factors may (in part) reflect an upper physiological limit to twist. Further study in elite endurance athletes may establish potential adaptability since athletes demonstrate reduced twist mechanics than untrained counterparts, suggestive of a functional reserve (Nottin et al. 2008; Santoro et al. 2014a).

Peak rotational mechanics were greater at the apex than base at rest and all cycling intensities. This is unsurprising since apical rotation is dominant in global LV twist (Notomi et al. 2006) and the apex demonstrates a more dynamic behaviour than the base (Doucende et al. 2010; Stohr et al. 2011) because of the LV geometric structure (Stöhr et al. 2015) and greater β-adrenergic receptor density (Mori et al. 1993).

Unlike absolute twist, PTV and PUV increased at both 30 and 60% W _max and further increases at higher intensities must be considered. Similar to this study PTV and PUV initially increased, but did not increase beyond moderate intensities indicating a potential upper limit (Stohr et al. 2011). The transition from rest to exercise diminishes the total cardiac cycle duration, most notably diastole, consequent to elevated HR (Doucende et al. 2010). As a result, faster myocardial rotation may be required during systole to accommodate these reductions (Doucende et al. 2010) and generate sufficient energy for release during early diastole to increase the intraventricular pressure gradient and augment LV filling during exertion (Burns et al. 2009). Although systolic twist and PUV are positively associated (Notomi et al. 2006), PTV may better predict PUV rather than the absolute twist as systolic and diastolic velocities were paralleled, expressing a systolic-diastolic coupling during incremental exercise in this study.

To our knowledge this is the first study examining twist mechanics during lower body isometric exercise, showing preserved LV twist at both intensities, which is discordant with prior IHG work (Balmain et al. 2016; Weiner et al. 2012) and opposes our second hypothesis. An explanation for our findings is a potential lack of afterload effect. Despite progressively increased arterial blood pressure, at the same relative intensity (40% MVC) SBP and DBP were lower than those reported previously (Balmain et al. 2016; Weiner et al. 2012). Sustained IHG may have had a greater effect on afterload, explaining the impairment evident following 3 min of IHG in those studies compared to the preserved twist presented here. Thus, this suggests short duration knee extension may not have elicited sufficient afterload to influence twist. At 40% MVC enhanced cardiac output and MAP is solely driven by tachycardia via vagal withdrawal, whilst sympathetic activation only assists HR after 10 s at 75% MVC (Maciel et al. 1987). Given the short duration (15 s) of effort in this study, and the biphasic blood pressure response to isometric exercise, sympathetic activation could have further increased MAP towards the latter end of contraction. However, the relative contribution of this sympathetic drive is likely to be minimal relative to vagal withdrawal. Previous work has demonstrated that total peripheral resistance remains unchanged with short duration (20 s) IHG at progressively higher intensities despite increased muscle sympathetic nerve activity. Therefore, in short duration isometric exercise, increased MAP is predominantly a consequence of augmented HR, and thus cardiac output (Lalande et al. 2014). It is plausible that longer contractions, such as the 3 min IHG utilised previously (Balmain et al. 2016; Weiner et al. 2012), may accentuate afterload and explain the reduced LV twist in those studies. The implications of such reductions remain incompletely understood. However, chronic hypertensive and aortic stenosis patients showed increased baseline twist (Santoro et al. 2014b), which may be demonstrative of a compensatory mechanism to maintain normal cardiac function. Thus, it is clear that greater understanding surrounding how afterload inducing exercise effects acute twist mechanics may have important implications for chronic health and disease. A second explanation concerns a limited vascular response during the isometric exercise. Twist decreased after partial and full arm arterial occlusion (Balmain et al. 2016; van Mil et al. 2016), indicating that twist may be more sensitive to peripheral resistance and/or vascular-mediated rather than flow-mediated (cardiac output) blood pressure changes (Balmain et al. 2016), even within a normal afterload range and independent of HR and ischemia (van Mil et al. 2016). Indeed, LV twist and systemic vascular resistance are negatively associated (van Mil et al. 2016). In conjunction with the aforementioned maintenance of peripheral resistance during short duration IHG (Lalande et al. 2014) and although we observed rapidly increased MAP, an insufficient vascular response may have prevented changes in twist. It may be possible that either a lack of afterload or vascular effect or a combination of both were responsible for the preserved twist in this study. Nevertheless, this study is unique in that it is the first to investigate LV twist mechanics during isometric exercise utilising a large muscle mass. It would be beneficial for future work to replicate exercise training and explore the influence of multiple, consecutive repetitions and/or longer duration lower body isometric efforts on twist mechanics. The exercise pressor response would likely be larger and potentially evoke differing cardiac responses than observed in our study.

PTV progressively increased with incremental exercise, whereas PUV remained unchanged from rest to 40% MVC which agrees with previous IHG work (Balmain et al. 2016; Weiner et al. 2012). Our study extends these observations showing significantly increased PUV to 75% MVC. The magnitude and rate of untwisting are increased with pharmacological inotropic interventions (Rademakers et al. 1992). However, increased PUV at 75% MVC is likely due to predominantly greater withdrawal of vagal tone, as this is the primary mechanism for tachycardia at the initiation of isometric contractions (Maciel et al. 1987). Full circulatory occlusion reduced HR via vagal tone restoration, resulting in decreased PUV which indicates a HR response may confound the ‘true’ effect of arterial pressure on PUV (Balmain et al. 2016). Thus, it is plausible that limited increases in arterial pressure and sympathetic activation, the increased twisting mechanics as shown in this study be mediated solely by accentuated chronotropic effect. Similar to the aforementioned reduced systolic and diastolic durations, altered twisting velocities as opposed to absolute twist, may contribute a pivotal role in supporting LV function during such physiological conditions.