Feasibility of Strain Assessment of Right Ventricular Contractile Reserve in Patients Undergoing Lung Resection

Lung resection causes substantial exertional dyspnea postoperatively. While it is intuitive that the functional limitation experienced by a patient will be dictated by pulmonary function, postoperative lung function only accounts for a small proportion of this exertional dyspnea1. Beyond lung function, cardiac function is also impaired after lung resection. Specifically, resting right ventricular (RV) function is impaired after lung resection, with no change in left ventricular function, and may be partly responsible for the functional limitation observed1. It is important therefore to develop a means of assessing preoperative RV function in order to identify which patients may be most at risk postoperatively2,3. Right ventricular contractile reserve (RVCR) describes the physiological increase in RV function during exercise3. Speckle tracking RV free-wall longitudinal strain (RVFWLS) analysis is a novel echocardiographic technique to assess RV function that is well associated with gold standard cardiac magnetic resonance ejection-fraction where other RV echocardiography parameters are not4. We therefore performed a feasibility study of RVFWLS analysis in patients undergoing exercise stress echocardiography pre and post lung resection. While previous work has shown that both left ventricular global longitudinal strain (LVGLS) and LVGLS-rate linearly increase with exercise, the relationship between RVFWLS (and RVFWLS-rate) and exercise is unexplored5,6. We therefore sought to investigate the use of RVFWLS (and RVFWLS-rate) and its relationship with level of exercise intensity as a surrogate measure of RVCR. We hypothesized that preoperatively RVCR would be demonstrated by a linear relationship between RVFWLS/RVFWLS-rate and exercise intensity, and that this would be impaired postoperatively. We performed a prospective single center observational feasibility study, undertaken with ethical approval (RECref 17/EE/0134), including consenting patients >16 years of age, undergoing planned elective video-assisted thoracoscopic lobectomy. Patients who were pregnant or undergoing pneumonectomy, sub-lobar, or isolated right middle lobectomy were excluded. We performed exercise stress echocardiography pre- and two-months postoperatively on a semi-supine cycle ergometer at three-minute incremental workloads (0, 15, 30, 45 and 60 Watts). At each workload, we used the percentage of age-predicted maximal heart rate (ppHRmax) as a measure of the exercise intensity7,8. Focused transthoracic echocardiography was performed by British Society of Echocardiography accredited echocardiographers. Two-dimensional speckle tracking analysis of RVFWLS and RVFWLS-rate was performed offline using TomTec 2D-Cardiac Performance Analysis in accordance with recent consensus guidelines9. For easier interpretation we reported RVFWLS as a positive percentage, with higher percentages indicating better RV function. We assessed the peak tricuspid regurgitant (TR) jet velocity as a surrogate of RV afterload. Patient tolerability, technical feasibility of RVFWLS assessment, and overall feasibility (a combination of tolerability and technical feasibility) were reported. To evaluate RVCR, we assessed for linear association between RVFWLS and ppHRmax. We performed analysis of covariance (ANCOVA) to account for within subject correlation. Seven patients were enrolled between November 2017 to August 2018. Mean (SD) age was 67 (13). Five (71%) patients were female. Two patients were American Society of Anesthesiologists (ASA) status two, the other five patients were ASA status three. Six patients had a smoking history, and three had a diagnosis of chronic obstructive pulmonary disease. No patients were taking beta-blocker medication. All patients underwent lobectomy, with a mean (SD) one-lung ventilation time of 156 (28) minutes. Patients were followed up at two-months postoperatively to assess for RVCR using semi-supine cycle ergometer with incremental workload. Exercise was tolerated in all patients preoperatively up to 45W, but two patients were unable to exercise at 60W due to dyspnea and joint pains (Figure 1A). Tolerability was poorer postoperatively at 60W, with two patients experiencing intolerable dyspnea. Technical feasibility was higher preoperatively than postoperatively, with 86% technical feasibility up to 45W preoperatively. This was possibly due to the changes in the anatomical location of the heart that occur after lung resection, resulting in poorer acoustic windows for echocardiography. Technical feasibility trended downwards at higher workloads both preoperatively and postoperatively due to difficulty visualising the RV free wall and apex at higher workloads. There was high overall feasibility preoperatively (≥86%) up to 45W, falling to 57% at 60W. Overall feasibility was poorer postoperatively, falling to 29% at 60W. Given the poor overall feasibility at 60W, we excluded this data-point from RVCR assessment. As we acquired peak TR velocity for only one patient at all workloads, we excluded it from further analysis. There was no association between RVFWLS and ppHRmax pre- or postoperatively (r=0.02 p=0.95 and r=0.16 p=0.58 respectively, Figure 1B). RVFWLS-rate was associated with ppHRmax preoperatively (r=0.58 p=0.03, Figure 1C), but not postoperatively (r=0.05 p=0.87). Exercise stress echocardiography was well tolerated preoperatively, but less so postoperatively. Overall feasibility of RVFWLS postoperatively was challenging, meaning it may be better suited as a preoperative assessment tool. Importantly, before surgery is where we envisage its clinical utility. We did not identify an association between RVFWLS and exercise intensity pre- or postoperatively, suggesting RVFWLS was unable to detect the presence of RVCR. Although this is surprising, previous research has shown that RVFWLS only modestly improves with exercise, with a study of 121 healthy volunteers demonstrating an absolute change in RVFWLS of just 1.1% when exercising compared to rest8. An alternative explanation is that preoperatively our lung resection cohort may not reflect a healthy population and may already have an impaired RVFWLS response to exercise. RVFWLS-rate was associated with exercise intensity preoperatively (r=0.58 p=0.03), this association was lost postoperatively, suggesting that RVFWLS-rate may have identified the presence of RVCR preoperatively, but this was impaired postoperatively. Previous studies investigating RV global longitudinal strain (RVGLS)-rate during dobutamine stress echocardiography demonstrated that RVGLS-rate significantly improved with increased inotropy, with no change in RVGLS10,11. Schlangen et al performed dobutamine stress echocardiography with contemporaneous invasive measurement of pulmonary hemodynamics to assess RV systolic elastance (Es- a load independent measure of contractility)11. RVGLS-rate was significantly associated with Es, with no association between RVGLS and Es, suggesting RVGLS-rate may be a superior correlate of contractility in this setting. Similarly, our results suggest that RVFWLS-rate may be more suited than RVFWLS to assess RVCR perioperatively. There has been only one previous (observational) study of RVCR in patients undergoing lung resection using invasive right heart catheterisation which identified that impaired preoperative RVCR was associated with increased risk of postoperative cardiorespiratory outcomes and longer length of stay12. Given that exercise stress echocardiography was well tolerated and highly feasible preoperatively, an appealing avenue of further research would be to assess if impaired preoperative RVCR (using RVFWLS-rate) is associated with postoperative patient outcomes. Limitations of our study include that we could not report surrogates of pulmonary vascular resistance, and were therefore unable to assess RV-pulmonary artery coupling. Since this was a feasibility study with a small sample size, our outcomes are purely explorative. In summary, our findings suggest that in patients undergoing lung resection, RVFWLS-rate is feasible and examining its relationship with exercise intensity may be a useful surrogate measure of RVCR. In the present study the relationship between RVFWLS-rate and exercise intensity identified the presence of RVCR preoperatively that is lost postoperatively. Further studies are needed to assess the clinical significance of impaired RVCR diagnosed by RVFWLS-rate in these patients and explore the potential for its use in preoperative assessment to identify a cohort of high-risk patients in whom perioperative RV protective interventions may be warranted. This work was supported by the National Institute of Academic Anaesthesia 2016 Ernest Leech Research Fund. The funding body had no input in the design of the study and collection, analysis, and interpretation of data. The funding body had no role in writing the manuscript.