Effects of Left Ventricular Assist Device on Cardiopulmonary Exercise Performance

Over the last years, left ventricular assist devices (LVADs) have become a possible destination therapy for patients with advanced heart failure (HF).1 According to heart transplant guidelines, exercise performance is a major parameter in the decision-making process. Cardiopulmonary exercise test (CPET) is the gold standard for exercise evaluation in HF. At present, however, there is a lack of evidence on the contribution of LVAD in improving exercise capacity as measured by CPET, since in many cases a complete CPET may not be performed in the pre-LVAD evaluation and/or a maximal effort is considered too demanding – if not dangerous – for severe HF patients. Exercise capacity before and after LVAD has been investigated in few reports analysing a very limited number of cases, and, moreover, data are contradictory; indeed, some authors reported that exercise capacity remains severely reduced, while others found a substantial increase in peak oxygen uptake (VO2) and in overall exercise performance.2, 3 With the aim of analysing the effect of LVAD on exercise performance, we evaluated the CPETs of 15 consecutive patients who had CPET during pre-LVAD evaluations and underwent a successful continuous-flow LVAD implantation as a destination therapy (Jarvik 2000, Jarvik Heart Inc., New York, NY, USA). All CPETs were performed using Quark PFT-CPET (Cosmed, Rome, Italy). Our study complies with the Declaration of Helsinki, our ethics committee has approved the research protocol (R517/16 –CCM546), and informed consent was obtained from all subjects. The mean age of our cohort was 68 ± 4.3 years and the majority of patients were male (14/15). All patients had chronic HF with a mean left ventricular ejection fraction of 22.47 ± 5.07% without significant right ventricular dysfunction. Mean Interagency Registry for Mechanically Assisted Circulatory Support (INTERMACS) class was 3.47 ± 0.52. Ten patients had a post-ischaemic disease (67%), four had an idiopathic dilated cardiomyopathy (26%), and one patient had myocarditis-related cardiomyopathy. Pharmacological treatment included amiodarone (87% of patients), beta-blockers (93%), angiotensin-converting enzyme inhibitors (60%), ivabradine (13%), and diuretic therapy (furosemide, in 100% of patients, and spironolactone, in 80%), with no major changes at post-LVAD evaluation, except for a significant decrease in diuretic dosage (mean furosemide daily dose decreased from 133 to 40 mg, P = 0.01). Baseline evaluation included standard spirometry testing (Table 1). We compared exercise at baseline (mean 50 ± 31 days before implantation) and at least 4 months after implantation (mean 7.6 ± 4.7 months), a time lag needed to allow a stable post-surgical clinical condition while avoiding a prolonged training effect. Indeed, all patients undertook a customized rehabilitation programme, which varied according to patient conditions. No significant difference was found in haemoglobin values nor in body mass index before and after surgery. CPETs were performed and analysed using a standard methodology: specifically, CPETs were performed on a cycle-ergometer, measuring respiratory gases breath by breath and applying a ramp protocol set to achieve peak exercise in 10 minutes. All patients had a custom pump speed of 3, both at rest and during exercise. All patients started with virtually 0 W, which increased gradually as per ramp protocol. Six patients were evaluated with the same ramp and two patients with a higher one. Seven patients were evaluated with a lower one (mean ramp pre 6.29 ± 2.05, post 4.71 ± 1.79, P = 0.01); this subgroup reached 41.71 ± 11.95 W and 46.29 ± 16.31 W pre and post-LVAD, respectively (P = 0.49). Globally, no significant difference was found in mean ramp pre- and post-LVAD implantation (5.93 ± 1.90 and 5.40 ± 1.88, respectively, P = 0.15). Exercise duration (7.1 vs. 9.5 min, P < 0.05), workload (42 vs. 51 W, P < 0.05), and peak VO2 (10.5 vs. 11.8 mL/kg/min, P < 0.05) increased, while ventilatory efficiency slope by means of CO2 production/ventilation relationship (VE/VCO2) (44.3 vs. 36.9, P < 0.05) significantly decreased (Table 1). Moreover, at pre-LVAD CPET, exercise-induced periodic breathing was observed in eight patients, while it was observed only in two cases after LVAD implantation (P < 0.05). Anaerobic threshold (AT) was identified in 7/15 and in 14/15 patients at the pre- and post-LVAD CPETs, respectively (P < 0.05). In the patients who had AT identified both at the pre- and post-LVAD CPETs (n = 7), VO2 at AT increased significantly (Δ 2.24 ± 1.35 mL/kg/min, P < 0.01), while peak increase was Δ 1.96 ± 1.77 mL/kg/min (P < 0.0001). In the entire population, peak VO2 increase was Δ 1.37 ± 2.29 mL/kg/min (P < 0.05). The higher VO2 at AT and the increased number of subjects with an identifiable AT reflect an improved haemodynamic pattern, likely due to increased cardiac output (CO), since AT is strictly CO-dependent.4 AT changes reflect daily activity improvement. It is important to highlight the role and importance of AT in the evaluation of everyday activity. Indeed, AT represents the moment when oxygen delivery to muscles becomes insufficient for aerobic ATP production, and VO2 at AT is strictly related to everyday activity. The increased number of patients who had an identifiable AT and the increase in VO2 at AT are in line with the reported improvement in exercise performance, and, most importantly, in quality of life with LVAD.5 We know that changes in peak values alone, although significant, are not sufficient to define the real improvement in the quality of life of patients. Indeed, peak VO2 is one of the few objective parameters we can evaluate. It is also important to underline the poor baseline conditions of our cohort (mean left ventricular ejection fraction 22.47 ± 5.07%, mean New York Heart Association class 3.53 ± 0.51, mean INTERMACS class 3.47 ± 0.52), in which even a small increase reflects an improvement in symptoms (Table 1). Finally, we noticed a reduction in VE/VCO2 slope. This finding is of great importance as it reflects an improvement in ventilatory efficiency during exercise. Moreover, we studied only patients who underwent LVAD implantation as a destination therapy with Jarvik 2000. Given the unique peculiarities of each LVAD device, our results cannot be translated in a straightforward manner to other devices which have their own different characteristics regarding pre- and afterload sensitivity and response. In brief, although several causes may limit exercise performance improvement in LVAD recipient patients, such as chronotropic incompetence, right ventricular dysfunction and peripheral factors,6 we showed that exercise performance as measured by peak VO2 improves. The postponed AT suggests a haemodynamic improvement, while the reduction in VE/VCO2 slope suggests an improvement in ventilation efficiency during exercise. Conflict of interest: none declared.