The influence of increased venous return on right ventricular dyssynchrony during acute and sustained hypoxaemia

What is the central question of this study? Right ventricular dyssynchrony is a marker of function that is elevated in healthy individuals exposed to acute hypoxia, but does it remain elevated during sustained exposure to high altitude hypoxia, and can it be normalised by augmenting venous return? What is the main finding and its importance? For the first time it is demonstrated that (i) increasing venous return in acute hypoxia restores the synchrony of right ventricular contraction and (ii) dyssynchrony is evident after acclimatisation to high altitude, and remains sensitive to changes in venous return. Therefore, the interpretation of right ventricular dyssynchrony requires consideration the prevailing haemodynamic state.

In a study of healthy volunteers, Pezzuto et al. (2018) have shown that RVD occurs during acute hypoxia (F iO 2 = 0.12), a stimulus known to increase pulmonary artery pressure (Pezzuto et al., 2018). Interestingly however, RVD did not occur during exercise, despite observing a greater increase in pulmonary pressure during exercise than during acute hypoxia (Pezzuto et al., 2018). The absence of RVD during exercise was attributed to systemic factors associated with global hypoxaemia. However, a change in RV geometry and increase in contractility that would result from the augmentation of venous return during exercise (Naeije & Badagliacca, 2017) may better explain the homogeneous contraction, although this is yet to be investigated. Moreover, little is known about whether RVD occurs in healthy individuals when the hypoxic exposure is sustained, such as during sojourn to high altitude. In this setting, as well as the increase in pulmonary pressure, venous return will be simultaneously decreased due to plasma volume constriction , and may exacerbate RVD if venous return does indeed influence the timing of regional contraction in the right ventricle. We therefore sought to examine the independent and combined effects of increased pulmonary artery pressure and venous return on RVD in healthy volunteers by performing two separate studies, (i) in acute and (ii) in sustained hypoxia (5-10 days), both with and without augmented RV venous return. We hypothesised that increased pulmonary artery pressure in hypoxia would increase RVD in both the acute and sustained setting. Furthermore, RVD will be reduced by increasing venous return (iii) in acute and (iv) in sustained hypoxic exposure. This work will provide novel insight into the mechanisms contributing to RVD, thus enhancing its utility as a marker of right heart function.

New Findings
• What is the central question of this study?
Right ventricular dyssynchrony is a marker of function that is elevated in healthy individuals exposed to acute hypoxia, but does it remain elevated during sustained exposure to high altitude hypoxia, and can it be normalised by augmenting venous return?
• What is the main finding and its importance? hypoxia with increased venous return (LBPP). All conditions consisted of a sphygmomanometer blood pressure (BP) recording followed by an echocardiographic evaluation while simultaneously recording heart rate (HR) and peripheral oxygen saturation (S pO 2 ). Measurements during LBPP were started after 2 min wash-in at the desired positive pressure and a 30-min hypoxic wash-in period was applied before hypoxic measurements were taken.

Lower body positive pressure
Participants assumed the supine position with their lower torso and legs in a custom-made box sealed by a flexible neoprene cover around the waist, enclosed at the level of the iliac crest. The pressure in the box was increased to 10 mmHg, based on previous research demonstrating a 5.9 mmHg increase in central venous pressure at this level of LBPP (Norsk et al., 1986). Participants were asked to keep their legs relaxed to prevent any discrepancies in venous return due to leg movement.

Hypoxic breathing
Hypoxia was administered by a hypoxic generator (PowerBREATHE Altitude Systems, High Performance Pro System, POWERbreathe International Ltd, Southam, UK), producing a fraction of inspired oxygen (F iO 2 ) of 12%. This degree of hypoxia corresponds to an altitude of 4500 m and has been shown to be well tolerated with minimal changes in arterial P CO 2 (Pezzuto et al., 2018).

Study 2: sustained hypobaric hypoxia
To examine RVD in sustained hypoxia, a retrospective analysis was performed on data collected in the autumn of 2015. Some of the data reported herein have been reported previously ; however, a distinct research question is addressed with a novel primary outcome variable (RVD) that has not been previously reported.

Study population
Ten male participants, aged 27 ± 6 years, were recruited from the expedition team who were normotensive, non-smokers with no previous history of cardiovascular or respiratory diseases and were taking no prescription medications. At the time of testing, none of the participants were experiencing any symptoms of acute mountain sickness, as assessed by the Lake Louise questionnaire, when at high altitude (3800 m).

Experimental design
The data were collected during two experimental visits: one at sea

Plasma volume expansion
Saline infusions were performed using an infusion pump following cannulation of a peripheral vein. The volume infused was tailored to each individual in order to restore absolute blood volume to sea level values assessed via carbon monoxide rebreathing, as previously used by the research team (Williams et al., 2016). The average infusion volume required to restore plasma volume was 398 ± 248 ml, as previously described in detail . Infusion duration varied between individuals (20-40 min) and haemoglobin concentration was checked following infusion to ensure sufficient volume had entered the vascular space to normalise plasma volume to sea level.

Experimental measures: study 1 and study 2
All measurements were recorded continuously, with the exception of echocardiography assessment and manual BP assessment in study 1.
Oxygen saturation (S pO 2 ) was monitored via fingertip pulse oximetry were assessed continuously by breath-by-breath online gas analysis measured in all participants in study 2, but only 14 participants in study 1 due to a technical error (n = 1) (Oxycon Mobile, Carefusion, San Diego, CA, USA).

Conventional measurements
Echocardiographic evaluation was performed with a 4.5-MHz phased array transducer using a commercially available ultrasound system (GE, Vivid E9, GE Healthcare, Milwaukee, WI, USA). Digital greyscale twodimensional (2D) and tissue Doppler cine loops from five consecutive beats were obtained at functional residual capacity, with a brief relaxed end-expiratory breath hold. Cardiac parameters were obtained according to the American Society of Echocardiography guidelines (Lang et al., 2015;Rudski et al., 2010) and were taken by two Right atrial pressure was estimated by diameter and collapsibility of the inferior vena cava, according to Rudski et al. (2010). Systolic pulmonary artery pressure (sPAP) was estimated by adding the peak systolic right ventricle-right atrium pressure gradient to right atrial pressure; mean pulmonary artery pressure (mPAP) was estimated from sPAP using Chemla's formula (0.61 × sPAP + 2 mmHg) (Chemla et al., 2004;Rudski et al., 2010). Pulmonary vascular resistance was calculated according to Abbas's formula [(peak velocity of the tricuspid regurgitant jet/RV outflow tract velocity-time integral) × 10 + 0.16] (Abbas et al., 2003). RV stroke volume (SV) was calculated from the RV outflow tract diameter (RVOTd) and RV outflow tract velocity-time integral (RVOT VTI) with the formula SV = 0.785 x RVOTd 2 x RVOT VTI (Zoghbi et al., 2017). RV end-diastolic area (RVEDA) and RV endsystolic area (RVESA) were determined from a modified apical fourchamber view in line with the American Society of Echocardiography guidelines for the assessment of the right heart (Rudski et al., 2010).

Speckle tracking
Peak strain and time to peak strain from each of the six segments were taking the standard deviation of the times to peak-systolic strain for the four mid-basal RV segments, corrected to the R-R interval between two QRS complexes, according to Bazett's formula, and called RVSD4 (Badagliacca et al., 2015a(Badagliacca et al., , b, 2017Kalogeropoulos et al., 2008;Pezzuto et al., 2018). The two apical segments were excluded because of their excessive variability, not only in pulmonary hypertension patients but also in healthy volunteers, as is conventional (Badagliacca 2015a(Badagliacca , b, 2017Pezzuto et al., 2018;Lamia et al., 2017). The time to peak contraction response was determined by combining the basal and mid segments of the RV free wall and the interventricular septal wall.
Our intraobserver coefficients for speckle-tracking echocardiography have previously been reported for the left ventricle (Stembridge et al., 2015) and range from 3.4% to 6.6% for RV areas and 4.7 to 6.0 for RV strain and strain rate. For visual representation, mean segmental strain curves were generated for each condition using cubic spline interpolation, as described previously (Stembridge et al., 2014).

Statistical analysis
All data of study 1 and 2 were analysed using SPSS version 22 (IBM Corp., Armonk, NY, USA) and expressed as means ± SD. Statistical significance was defined as P ≤ 0.05 and distribution normality was confirmed using repeated Shapiro-Wilk W tests.

Study 1
The effects of hypoxia and LBPP on RVSD4 were assessed by a twoway repeated measures ANOVA. Student's paired, two-tailed t-test or a Wilcoxon signed rank test, when not normally distributed, was used to assess the difference between the measures in the normoxia baseline, normoxia LBPP, hypoxia baseline and hypoxia LBPP. Cohen's D was used to assess the differences between time to peak contraction.

Study 2
Differences in RVSD4 between sea level, high altitude and high altitude combined with plasma volume expansion were analysed using a oneway repeated measures ANOVA. A post hoc Tukey test was used to determine the difference in RVSD4 between sea level, high altitude and high altitude combined with plasma volume expansion. Student's paired, two-tailed t-test or the Wilcoxon signed rank test when not normally distributed was used to assess the difference between the measures in sea level, high altitude and high altitude combined with plasma volume expansion. Cohen's D was used to assess the differences between time to peak contraction.

Haemodynamic response
Hypoxia elicited a significant increase in HR and mPAP compared to normoxia, with no change in RVEDA (Figure 1). LBPP in hypoxia further increased mPAP compared to hypoxia baseline, simultaneous to a significant increase in RVEDA (Figure 1). Left ventricular (LV) eccentricity index was increased in hypoxia but unchanged by LBPP (Table 1).

RVSD4 response
RVSD4 was significantly increased in hypoxia compared to normoxia, but was significantly lower in hypoxia + LBPP compared to hypoxia alone (Figure 2). Time to peak contraction of the RV free wall and interventricular septal wall was significantly longer (i.e. prolonged contraction) during hypoxia and hypoxia + LBPP compared to normoxia. Moreover, the time to peak contraction was significantly prolonged in the interventricular septal wall during hypoxia + LBPP compared to hypoxia alone, with no change in the RV free wall ( Figure 2). This suggests the increase in RV filling was altering septal and not restoring RV free wall function, despite the difference in time to peak contraction between normoxia and hypoxia being larger in the RV free wall (Figure 2). These changes happened independent of changes in RV strain rate, which remained unaltered in response to either LBPP or hypoxia.

Haemodynamic response
HR and mPAP both significantly increased at high altitude compared to sea level (see below, Figure 4). RVEDA significantly decreased at high altitude compared to sea level, but was restored at high altitude following plasma volume expansion (Figure 3) to be comparable to sea level values. In contrast to our preload intervention in study 1 (LBPP), mPAP was not increased following plasma volume expansion (Figure 3). LV eccentricity index increased at high altitude compared to sea level, and remained elevated following plasma volume expansion. Further haemodynamic responses are depicted in Table 2.

RVSD4 response
RVSD4 was significantly increased at high altitude compared to sea level, but was lowered once blood volume was restored at high altitude via infusion (Figure 4). Time to peak contraction of the RV free wall was significantly longer at high altitude compared to sea level, and significantly shorter at high altitude with volume expansion compared to high altitude alone. The time to peak contraction of the interventricular septal wall was significantly longer at high altitude with volume expansion compared to sea level (Figure 4), but in contrast to acute hypoxia, the reduction in RVD with volume expansion was likely mediated by a decrease in time to peak in the RV free wall. Similar to acute hypoxia, there were no significant changes in RV strain rate in response to either LBPP or hypoxia.

DISCUSSION
This study is the first to examine the effect of changes in venous return on RVD in acute and sustained hypoxia. In relation to our four hypotheses, our primary novel findings are that both (i) acute and (ii) sustained hypoxia increased RVSD4, but the increase was attenuated

4.1
The influence of increased pulmonary artery pressure in hypoxia on RVD Pezzuto et al. (2018) were the first to show that an increase in RV afterload secondary to acute hypoxia increases RVD in healthy volunteers (Pezzuto et al., 2018). Herein, we confirm and extend these findings by demonstrating that both acute and sustained hypoxia result in a robust increase in RVD in healthy volunteers. The results of this study are also in line with earlier work demonstrating that RVD is elevated in conditions with a chronic increase in right ventricular afterload in pulmonary hypertension patients (Badagliacca et al., 2015a(Badagliacca et al., , b, 2017Kalogeropoulos et al., 2008).
There are three main mechanisms that could cause a delay of cardiac contraction: (i) an increased QRS duration (Badagliacca et al., 2015a;Morita et al., 2019) RV (Badagliacca et al., 2015b;Morita et al., 2019). Considering (i), given that there is no relationship between the QRS duration and the delay in contraction of the RV free wall or the QRS duration with the onset of the RV contraction in pulmonary patients or in healthy volunteers (Marcus et al., 2008;Pezzuto et al., 2018), one could assume that a delay in QRS duration is not involved in RVD.
In relation to geometric changes (ii), we found no evidence that the changes in RVD were linked to septal flattening, as the eccentricity index increased in hypoxia but remained elevated during LBPP and plasma volume expansion despite RVD decreasing. Therefore, whilst changes in RV geometry occur in hypoxia, such alterations do not appear to result in dyssynchronous contraction per se. Therefore, the most likely explanation for the increase in RVD during an increase in hypoxia is an increase and non-uniform distribution of wall stress, induced by an increase in RV pressure and workload (Badagliacca et al., 2015a, b;Morita et al., 2019). The importance of wall stress on timing of the contraction of myocytes has been described by van Heuningen et al. (1982), who showed that the contraction of rat myocytes was slower when wall stress was increased (van Heuningen et al., 1982).
Thus, when RV afterload is increased, wall stress is elevated in specific segments and causes a slower region-specific contraction, resulting in a dyssynchronised contraction. This is supported by the observation in the current study and previous work in pulmonary hypertension patients that demonstrates time to peak contraction of the RV free wall to be delayed more than the interventricular-septal wall (Lopez-Candales et al., 2005a;Marcus et al., 2008).

4.2
The influence of increased venous return on RVD in acute and sustained hypoxia For the first time, we demonstrate that the hypoxia-induced increase in RVD can be attenuated in healthy volunteers when venous return is augmented, in both acute and sustained hypoxaemia. Thus, RVD is related to both RV preload and afterload, rather than solely RV afterload, and may explain why Pezzuto et al. (2018) did not report an increase in RVD during exercise, despite higher pulmonary artery pressures than during hypoxia.
Although RVD was decreased when right ventricular venous return was increased in both acute and sustained hypoxia, the way in which it was achieved differed. Increasing venous return via LBPP in acute hypoxia increased the time to peak contraction in the interventricular septal wall so that it was comparable to the RV free wall, thereby reducing dyssynchrony. This suggests that increasing RV venous return when PAP is elevated lengthens the time to peak contraction in the septum. In contrast to acute hypoxia, restoring venous return via saline in sustained hypoxia did in fact decrease the time to peak contraction in the RV free wall, lowering RVD through a direct effect on RV function. The contrasting findings may, in part, be due to the nature of our two scenarios. During acute hypoxia prior to LBPP, no change in ventricular volumes was observed, nor would it be expected (Boulet, Stembridge, Tymko, Tremblay, & Foster, 2016). With sustained hypoxic exposure, but it is well established that ventricular volumes decrease (Stembridge & Levine, 2019), especially in the left heart. The saline infusion used in the current study was designed only to normalise ventricular volumes to sea level values, whereas LBPP increased ventricular filling beyond baseline levels. Nevertheless, in both scenarios, increasing RV filling decreased RVD, demonstrating the interaction between RVD and filling.

Clinical perspective
In conditions of chronically elevated pulmonary artery pressure, RVD is regarded as a marker of both the functional capacity and prognostic outcome (Badagliacca et al., 2015a(Badagliacca et al., , b, 2017Lopez-Candales et al., 2005a, b;Marcus et al., 2008;Murata et al., 2017;Smith et al., 2014). Based on the current study, one could argue that RVD may be related to the prevailing haemodynamics and not just the pulmonary pressure. Moreover, further information is required on how changes in RV structure with disease progression influence RVD. For example, RV coupling (contractility to arterial afterload) can be maintained in patients with pulmonary hypertension via an increase in workload (Vonk Noordegraaf et al., 2017). In the early stages of the disease, RV uncoupling is prevented by an increase in RV contractility, facilitated by remodelling of RV wall thickness and altering muscular properties of the myocardium (Vonk Noordegraaf et al., 2017). When the RV is no longer able to adapt to the increase in workload by cardiac remodelling, the RV will dilate to increase preload to maintain coupling (Sano et al., 2007;Vonk Noordegraaf et al., 2017). Only in advanced stages of pulmonary hypertension will uncoupling occur, when the adaptability of the RV is insufficient to cope with an increase in workload (Trip et al., 2015).

Limitations
The number of participants in this study is relatively small, especially in our study conducted at high altitude. Expeditions of this nature are financially and logistically difficult to organise, and small sample sizes are often unavoidable (Ainslie, 2014). Notwithstanding, all participants showed a similar response to the conditions of this study, and we report individual data points for transparency. Similarly, our method for assessing right atrial and right ventricular pressure was indirect, but invasive measures were not possible, especially in remote locations used in the sustained hypoxia trial. Despite both males and females being actively recruited for our studies, only three females volunteered for our sea level study and no female members of the expedition team were able to participate due to competing demands from other experiments. Therefore, regrettably, we report data on a sample that consists of predominantly males. Despite the small number of females, their RVD response was similar compared to males, indicating that sex has no influence on RVD response to alterations in preload and afterload. The stimuli used to increase preload differed between the acute and sustained setting. However, both stimuli sufficiently increased preload in our study, as well as in previous work (Greenway & Lautt, 1986;Grimminger et al., 2017;Norsk et al., 1986;Pezzuto et al., 2018).
The method employed to detect RV dyssynchrony (RVSD4) does not discriminate between systolic and post-systolic contraction. As the occurrence of post-systolic shortening is an important determinant in clinical prognosis in pulmonary hypertension patients, the timing of contraction should be considered when determining RV function (Badagliacca et al., 2015a, b). Lastly, speckle tracking is an analysis method that is subjective to human variability, as the tracking area has to be set manually. To ensure reproducibility of the tracking area, all tracking was performed by the same researcher (M.E.).

CONCLUSION
Collectively, our findings demonstrate the dependence of RVD on ventricular filling as well as afterload, and highlight the need of haemodynamic state and stage of structural remodelling to be considered in the interpretation of RV function in health and disease.