Exercise-induced cardiac fatigue after a 45-minute bout of high-intensity running exercise is not altered under hypoxia

32 Background. Acute exercise promotes transient exercise-induced cardiac fatigue (EICF), which affects 33 the right ventricle (RV) and to a lesser extent the left ventricle (LV). Hypoxic exposure induces an 34 additional increase in RV afterload. Therefore, exercise in hypoxia may differently affect both 35


INTRODUCTION 56
It is well established that exercise is associated with potent cardioprotective effects 1-3 , however, 57 acute exercise can lead to a paradoxical short-term increase in cardiac events. 4-6 One potential 58 explanation is that exercise performed under demanding conditions (i.e. exercise at high-intensity 59 and/or during prolonged duration) may lead to an acute reduction in cardiac function. [7][8][9][10][11][12][13] This 60 transient decline in cardiac function after strenuous exercise is typically referred to as exercise-61 induced cardiac fatigue (EICF). EICF may affect both left (LV) and right ventricles (RV), with possibly a 62 larger impact on the RV due to the disproportionately higher wall stress experienced by the RV 63 relative to the LV during exercise. 11, 14, 15 64 Previous studies have demonstrated that hypoxia increases the demands on the cardiovascular 65 system. 16 Specifically, acute exposure to hypoxia induces a decrease in systemic vascular resistance 66 at rest, which may contribute to a decrease in LV afterload. 17,18 In contrast, hypoxia leads to a resting 67 increase in pulmonary artery resistance, and subsequently to an increase in pulmonary vascular 68 resistance (PVR) and pulmonary artery pressure (PAP). 19 Exercise in normoxic conditions results in 69 additional load challenges and an increased PAP secondary to the mismatch of elevated stroke 70 volume to inadequate pulmonary vascular distension. 20 This is exacerbated when exercising in 71 hypoxic conditions, leading to an even greater PAP and RV wall stress and potentially further 72 increasing the risk of RV EICF. 19-23 73 To non-invasively examine right heart haemodynamics, studies have examined conventional and 74 Doppler based echocardiographic indices at rest and during exercise. [24][25][26] Recently, the strain-area 75 loop has been introduced assessing simultaneous structure and strain across the cardiac cycle. 5 76 Previously, we found that RV loop characteristics relate to PVR in patients with pulmonary arterial 77 hypertension (PAH) whilst also demonstrating value in the assessment of EICF. 27, 28 Therefore, these 78 non-invasive characteristics may provide additional insight in understanding exercise-induced 79 changes in hypoxia. 80

J o u r n a l P r e -p r o o f
In view of this, the aim of this study, was to investigate the acute effects of a bout of high-intensity 81 exercise under hypoxia versus normoxia in healthy individuals on right-and left-sided cardiac 82 function and mechanics (i.e. longitudinal strain and strain-area loops). Based on the presumed higher 83 workload of the RV during hypoxic versus normoxic exercise, we hypothesize that exercise under 84 hypoxia exaggerates RV to a greater extent than LV compared to exercise under normoxia. To 85 investigate EICF, we examined pre-and post-exercise echocardiography at rest, but also during a 86 standardized low-to-moderate-intensity recumbent exercise challenge ('stress'). As the post-exercise 87 recovery period is associated with persistent sympathoexcitation and peripheral vasodilation 17, 18 , 88 evaluation of EICF could be confounded when evaluated solely at rest. Therefore, evaluation during 89 stress echocardiography may better reflect cardiac function during exercise and offsets the key 90 limitation of (para)sympathetic imbalance associated with echocardiographic assessment in 91 recovery. 14 92 93

Study population 95
Twenty-one participants (22.2±0.6 years, fourteen men, 24.0±0.6 kg/m 2 , VO 2 max/kg/min 52.4±1.8 96 mL/min/kg) completed the study. Baseline characteristics are shown in Table 1. Participants were 97 eligible to take part in this study if they were able to run on a treadmill and if they trained <2 hours a 98 week at moderate-to-high-intensity for the last six months. Exclusion criteria were a body mass index 99 (BMI) <18 or >30 kg/m 2 , active smoker, any possibility of pregnancy, personal history of 100 cardiovascular disease, positive family history of cardiovascular death (<55y), exercise-limiting 101 respiratory disease, physical (i.e. musculoskeletal) complaints making completion of a bout of high-102 intensity running exercise impossible, abnormal resting 12-lead electrocardiogram (ECG) and 103 abnormalities observed on resting transthoracic echocardiography. The procedures were in 104 accordance with institutional guidelines and conformed to the declaration of Helsinki. The study was 105 J o u r n a l P r e -p r o o f approved by the Ethics Research Committee of Liverpool John Moores University (18/SPS/065). 106 Participants gave full written and verbal informed consent before participation. 107 108

Study design 109
In this randomized crossover trial, participants attended the laboratory on three separate occasions 110 (Figure 1). During the first visit, a medical screening was performed to determine eligibility of the 111 potential participants. After signing informed consent, baseline measurements were performed. 112 Visits two and three included performance of a bout of 45-minute high-intensity running exercise 113 under normobaric hypoxia or normoxia, which were performed in a randomized order. Participants 114 were blinded for the order of test days and abstained from exercise for a minimum of 48 hours, and 115 from alcohol and caffeine consumption 24 hours before the test days. 116 Baseline measurements. Participants were examined for height (SECA stadiometer, SECA GmbH, 117 Germany), weight (SECA scale, SECA GmbH, Germany), oxygen saturation (SpO 2 , pulse oximetry; Ana 118 Pulse 100, Ana Wiz Ltd., UK), 12-lead ECG (Cardiovit MS-2010, Schiller, Switzerland) and maximal 119 oxygen consumption (VO 2 max). Resting heart rate (HR, Polar, Kempele, Finland) and resting blood 120 pressure (BP, Dinamap V100, GE Medical, Norway) were determined at the end of ten minutes of 121 quiet rest in a supine position. A standardized maximal cardiopulmonary exercise test (CPET, Oxycon 122 pro, CareFusion, VS) for VO 2 max assessment was conducted on a motorized treadmill (HP Cosmos, 123 Nussdorf, Germany) after a 10-min warm-up and familiarization. VO 2 max was defined as the highest 124 value of a 30-s average 29 , and attainment was verified according to previous recommended criteria. 30

125
Test days. Figure 1 outlines the details of a single test day. One of the test days was performed at 126 normoxia (sea level, equivalent to fraction of inspired oxygen [FiO 2 ] 20.9%) and the other at 127 normobaric hypoxia (FiO 2 14.5%; equivalent to a simulated altitude of 3,000m), separated by at least 128 48 hours of rest. Participants were subjected to 30 minutes of acclimation in a seated position 129 followed by 45-minute of high-intensity (85% of maximum achieved HR during CPET) endurance 130 running exercise on a motorized treadmill (HP Cosmos, Nussdorf, Germany) and 60 minutes of 131 recovery in seated position. HR was measured continuously throughout (Polar, Kempele, Finland), 132 and rate of perceived exertion (RPE) was monitored during the 45-minutes high-intensity running 133 exercise. 31 134 In total four echocardiographic assessments were performed per test day. After acclimation and 135 prior to the 45-minute exercise, echocardiography was performed under resting conditions ('rest') 136 and during recumbent cycling to elevate heart rate to directly assess cardiac function during exercise 137 ('stress', target HR 110-120 bpm). The 'stress' echocardiogram was repeated directly after the 45-138 minute exercise, to prevent sympathetic withdrawal (i.e. a drop in BP and HR). 32 Finally, images were 139 obtained at the end of the 60 minutes of recovery in a resting state. During every echocardiography 140 assessment, BP measurements were performed. Measurements were performed at the same time 141 on both days to control for diurnal variation. Fluid intake was controlled by providing the same 142 amount of water to participants during both testing days. 143 Environmental chamber and safety. All exercise tests were conducted in an environmental chamber 144 (TISS, Alton, UK; Sportingedge, Basingstoke, UK). Normobaric hypoxia was achieved by a nitrogen 145 dilution technique. Ambient temperature, carbon dioxide (CO 2 ) and oxygen (O 2 ) levels were 146 controlled in all sessions (20°C; FiO 2 14.5%; CO 2 0.03%), whilst a Servomex gas analysis system 147 (Servomex MiniMP 5200, Servomex Group Ltd., UK) was used inside the chamber to provide the 148 researcher continuous information regarding O 2 and CO 2 levels. Acute mountain sickness symptoms 149 (AMS, measured by Lake Louise Score 33 (LLS)) were monitored during testing and training sessions 150 every 20 minutes. The subject was removed from the environmental chamber if oxygen saturation 151 levels dropped below 80% or severe AMS was suspected (LLS≥6). Both the right and left heart had normal geometry and all structural measurements were within 222 normal ranges (  indicative for EICF, which was mainly expressed during a low-to-moderate-intensity exercise 280 challenge ('stress') compared to resting conditions. Earlier studies primarily investigated EICF after 281 prolonged exercise (>180minutes) 4, 27 , however, recent research has revealed a dose-response 282 relationship between EICF and the duration and intensity of exercise. 14, 39 Our study adds the novel 283 knowledge that EICF also occurs after relatively short periods of high-intensity exercise in both the 284 RV and LV. Interestingly, in contrast to other short-term high-intensity EICF studies 10, 14, 39 , we showed 285 also marked reductions in LV function which may be due to the different type of exercise (running vs. 286 cycling). An explanation for our ability to detect EICF after a relatively short duration of exercise may 287 relate to the post-exercise assessment of cardiac function during 'stress', i.e. low-to-moderate-288 intensity exercise Indeed, some of the indices for systolic function were primarily/only reduced when 289 echocardiography was performed during the low-to-moderate-intensity exercise challenge. For 290 example, a reduction in RVLS post-exercise was only apparent during the low-to-moderate-intensity 291 exercise challenge (Figure 4A). We believe the echocardiography assessment under low-to-292 moderate-intensity exercise is more likely to detect EICF. The recovery phase post-exercise is 293 associated with a change in autonomic tone and vasodilation, which may result in post-exercise 294 tachycardia and hypotension, respectively. These (para)sympathetic imbalance factors likely 295 influence cardiac function measurements such as strain, and therefore potentially mask the presence 296 of EICF. Evaluation of cardiac function during the high-intensity exercise, therefore, is preferred. 297 However, one should consider the practical aspects (e.g. echocardiography is impossible during 298 running) and that reliable speckle tracking is extremely challenging with higher heart rates (i.e. 70% 299 of maximum HR). 40 Low-to-moderate intensity cycling exercise at a semi-recumbent bike is both 300 feasible and reliable, and allows to examine cardiac function during exercise. Utilising this approach, 301 our data indicates that, with short durations of high-intensity exercise, EICF occurs when assessment 302 of cardiac function is performed during an exercise challenge. 303 304 J o u r n a l P r e -p r o o f

Impact of exercise under hypoxia 305
Under hypoxic conditions, less oxygen is bound to haemoglobin, and will, therefore, increase the 306 demand on the cardiovascular system. In our population, this was reflected by a higher resting HR 307 under hypoxia versus normoxia and the less distance covered under hypoxia versus normoxia during 308 the exercise despite it being matched for relative intensity. More importantly, hypoxia has been 309 shown to induce vasoconstriction of the pulmonary vasculature, leading to higher relative PVR 310 resulting in a higher PAP, and consequently a higher RV wall stress. Elevated PAP has been previously 311 demonstrated at conditions at 3000m altitude. 23 Although we were unable to directly measure PAP, 312 we demonstrated shorter PAT and a larger RA size which indirectly supports the presence of an 313 increase in PAP and, therefore potentially wall stress. Also, the strain-area loop showed less 314 uncoupling in late diastole and a trend for a less steep systolic slope under hypoxia. In line with a 315 previous study in PAH patients, these changes are associated with a higher PVR at rest. 28 Although 316 we adopted a non-invasive approach and one should consider alternative explanations (i.e. related to 317 the assessment), these findings support the presence of an elevated wall stress in our study under 318 hypoxia. That aside, our hypothesis was rejected as the 45-minute high-intensity running exercise 319 under hypoxia did not exaggerate RV EICF compared to exercise under normoxia. This suggests that 320 changing cardiac workload does not necessarily change the magnitude of RV EICF and may not be the 321 principle mechanism for RV EICF. One potential explanation for the lack of an impact of hypoxia on 322 EICF may be that the exaggerated loading conditions under hypoxia were not sufficient enough at 323 3000m of simulated altitude, and/or the exposure time to the raised afterload of the RV was not long 324 enough to contribute to the EICF magnitude. There are also indications that hypoxia itself may induce 325 cardiac dysfunction due to sustained low oxygen availability, however, this seems mainly during 326 prolonged exposure. 41 327 Our hypothesis originated from the accepted phenomenon of disproportionately higher relative wall 328 stress in the RV compared to the LV during exercise, but also based on observations suggesting a 329 larger magnitude of EICF in the RV compared to the LV. 11, 14, 15 For example, Stewart et al. examined 330 the influence of high-intensity exercise on RV free wall and segmental LV strain EICF following 90 331 minutes cycling 10 , and found that the reduction in strain was more profound in the RV than in the LV. 332 In their study they demonstrated a relative reduction in RV strain of -17.5% compared to -9.8% in our 333 study, which supports a dose-response relationship. Our study is the first to our knowledge to 334 directly compare normoxic and hypoxic conditions on EICF, and demonstrated similar changes in 335 both RV and the LV. Although mechanical changes in the RV and LV are independent of each other 27 , 336 and likely differ during exercise, our work suggests that (after)load dependency may be a less 337 contributing factor to EICF as previously suggested. Alternatively, intrinsic myocardial factors such as 338 β-adrenergic receptor desensitization 7, 42 and oxidative stress 43 may play a more substantial role. Our 339 study, however, is unable to provide further insight into these other possible mechanisms. 340 It is also of interest that following the 45-minute high-intensity exercise, this study showed a lack of 341 any RV dilation (no rightward shift strain-area loop, Figure 4)  is not augmented and thus it may be less likely that a role for elevated RV wall stress is relevant. 360 Although knowledge about the clinical long-term consequences of these temporary post-exercise 361 reductions in cardiac function is lacking, it has been hypothesized that this may be associated with 362 myocardial damage and worse clinical outcome. The absence of an effect in EICF between exercising 363 at sea level (normoxia) and 3000m altitude (hypoxia) is interesting, but long-term studies that link 364 these findings to prolonged follow-up is needed to better understand these findings. The novel 365 strain-area loop, introduced to assess haemodynamics non-invasively, provided substantial added 366 value in this study where it was sensitive enough to detect changes due to hypoxia. This novel 367 technique seems promising in providing physiological and pathophysiological insight and might be of This study implemented a standardized exercise challenge to prevent a pre-and post-exercise 372 (para)sympathetic imbalance during echocardiographic evaluation. Instead of the methodology of 373 Stewart et al. 14 (aiming at 100 bpm), we set our target HR at 110-120 bpm during the exercise 374 challenge, to better mimic cardiac function during exercise. This higher HR may impede speckle 375 tracking quality. With current frame rates used, we experienced that tracking was still good to 376 excellent for LV global longitudinal strain and RV free wall strain. A further limitation is that we did 377 not obtain direct measures of RV wall stress as this would require invasive procedures. Alternatively, 378 we used only non-invasive echocardiographic, indirect measures to estimate any potential difference 379

557
J o u r n a l P r e -p r o o f