Secondary Logo

Journal Logo

Adult Circulatory Support

High Right Ventricular Afterload Is Associated with Impaired Exercise Tolerance in Patients with Left Ventricular Assist Devices

Ton, Van-Khue*; Ramani, Gautam; Hsu, Steven; Hopkins, C. Danielle; Kaczorowski, David§; Madathil, Ronson J§; Mak, Susanna; Tedford, Ryan J.

Author Information
doi: 10.1097/MAT.0000000000001169


Survival of patients with end-stage heart failure (HF) has improved thanks to technological progress in left ventricular assist device (LVAD).1 LVAD patients report better functional status measured by New York Heart Association (NYHA) classes.2,3 However, exercise capacity, obtained from 6-minute walk distances or peak O2 consumption (pVO2) from cardiopulmonary exercise test (CPET), remain poor.3–5 Attempts at increasing LVAD speed to augment cardiac output (CO) during exercise have resulted in very modest pVO2 improvement (≈1 ml/kg/min).6,7 While recovered left ventricular function is important in maintaining exercise tolerance,8 much less is known about the role of the right ventricle (RV) during exercise. In patients with chronic HF, ventilatory efficiency, expressed as the slope of minute ventilation divided by CO2 production (VE/VCO2), is a marker of poor prognosis and RV dysfunction.9,10 An analysis of resting hemodynamics has suggested increasing RV sensitivity to afterload following LVAD implant,11 although the impact of exercise has not been well studied. Additionally, the RV response to exercise has not been compared with other conditions of elevated RV afterload, such as pulmonary arterial hypertension (PAH), or healthy controls.

In this study, we aimed to compare exercise hemodynamics among three cohorts: LVAD, PAH, and normal non-athletic controls. We hypothesized that LVAD subjects would have more impaired RV reserve and increased RV afterload during exercise would correlate with markers of poor exercise tolerance (low pVO2) and ventilatory inefficiency (high VE/VCO2 slope).


Between November 2017 and December 2018, at the time of right heart catheterization (RHC) for evaluation of heart transplant, 12 ambulatory adults with HeartWare LVAD (Medtronic, MN) at University of Maryland performed symptom-limited exercise on a supine cycle ergometer, at a work load of 10 watts/2-minute increments. “Pre-LVAD” RHC was retrospectively reviewed and selected for values closest to LVAD implantation. On the day of supine exercise, “rest” hemodynamics were obtained first, then simultaneous right atrial (RAP) and pulmonary artery pressures (PAP) were recorded every two minutes during exercise. All values closest to maximal effort were used in the analysis. Thermodilution CO measurements at peak exercise were done in triplicate, averaged and reported as a single value for each patient. Filling pressures were recorded by the Xper Information Management software (Phillips North America Corp., MA), and retrospectively verified by a cardiologist (V.K.T.) for end-expiration measurements. In cases of wide respiratory variation (5 patients), another cardiologist (G.R.) independently reviewed the waveforms and adjudicated filling pressures. Formulas used to calculate RV afterload and other measures of RV function are shown in Table S1, Supplemental Digital Content 1, Laboratory values within two weeks of RHC were also collected.

On a separate day from RHC, 10 out of 12 patients performed CPET on an upright cycle ergometer to maximal effort (10 watts/minute increments). One patient refused CPET, and another could not perform due to subsequent hospitalization for infection. Breath-by-breath gas exchange was measured by the Vmax Encore metabolic cart (CareFusion, Beckton Dickinson and Co., NJ) per standard protocol.12 There was no change in LVAD setting during CPET or RHC. Age-predicted maximum heart rate (APMHR) was calculated as 220—age in the absence of beta blocker, or as previously described in the presence of beta blocker.13 Heart rate reserve (HRR) was the percentage of heart rate at peak exercise/APMHR. Chronotropic incompetence was defined as HRR < 80% of APMHR.

Rest and exercise hemodynamics from two age- and sex-matched cohorts at University of Toronto (“normal” non-athletic adults without cardiopulmonary limitations) and Johns Hopkins Hospital (ambulatory adults with stable idiopathic PAH, defined according to the World Health Organization diagnostic criteria at the time of study design/enrollment14) were compared to hemodynamics of LVAD subjects. Exercise protocols for these two cohorts were previously published.15,16 Briefly, the “normal” cohort underwent submaximal semi-upright ergometry exercise for 8 to 10 minutes at a target heart rate of 100 beats/minute, without gas exchange measurements.16 The ambulatory PAH cohort performed supine ergometry exercise at 10 watts/2-minute increments until maximal effort, and continuous gas exchange was obtained with the Innocor metabolic cart (Innovision, Denmark).15 For both comparison cohorts, hemodynamic measures were also taken at end-expiration and verified by a cardiologist (S.H. and S.M.). Laboratory values and echocardiographic parameters were collected closest to the time of exercise. All data collection was approved by the Institutional Review Board at each institution, and all data were de-identified.

Statistical Analysis

Variables were expressed as median with 25th–75th interquartile range (IQR). Continuous values were compared by the Wilcoxon signed-rank test for paired values (pre-LVAD versus post-LVAD at rest and on exercise), or the Mann-Whitney U test for unpaired values. The Kruskal-Wallis equality-of-populations rank test was used for multiple group comparisons. Inter-rater variability was estimated by kappa statistics. Pearson’s correlation was used to find relationships between parameters. p < 0.05 was considered statistical significance. Analysis was performed with STATA 15 (StataCorp, TX).


Baseline Characteristics

Left ventricular assist device subjects.

Of 12 LVAD subjects, eight (66.7%) received LVAD implant for bridge-to-transplant indication, while the rest were destination therapy or “bridge-to-decision” subjects. Median time from HF diagnosis to LVAD was 12 months (IQR, 6–84), LVAD to RHC was 305 days (IQR, 233–399), LVAD to CPET was 244 days (IQR, 222–332), and time between RHC and CPET was 45 days (IQR, 26–116) (see Table S2, Supplemental Digital Content 2, Median NYHA class was one (IQR, 1–2). Eight subjects (67%) had non-ischemic cardiomyopathy. Median LVAD speed was 2622 rpms (IQR, 2540–2760). Estimated pump flow at rest was 4.4 L/min (IQR, 4–5) and 5.9 L/min (IQR, 4.9–6.8) during exercise when speed remained fixed. Pump flow estimated by the device correlated well with thermodilution CO. All subjects underwent LVAD insertion via the lateral thoracotomy approach, and no one had concomitant procedures. All LVAD subjects received inhaled pulmonary vasodilators on postoperative day 0 as standard of care, and no one was on phosphodiesterase inhibitors at any point during the study period. No subject required placement of temporary or durable right-sided mechanical support device following LVAD implant. On post-LVAD echocardiograms obtained at time closest to RHC, there was a trend toward less severe RV dysfunction and tricuspid regurgitation (TR) compared with pre-LVAD studies (36% vs. 64% moderate-severe RV dysfunction, 0% vs. 2% moderate-severe TR, not shown).

Pre- and post-LVAD hemodynamics are shown in Table S3, Supplemental Digital Content 3, Biventricular filling pressures, pulmonary pressures, and CO normalized following LVAD insertion. Inter-rater agreement was reasonable for exercise hemodynamics from five patients with large respiratory variations (κ = 0.2–0.4, p < 0.05, Pearson’s r > 0.8, p < 0.05). Postoperatively, among hemodynamic markers of RV function, pulmonary artery pulsatility index (PAPi) improved (2.1 to 4, p = 0.02), while RV stroke work index (RVSWI) and right atrial pressure:pulmonary artery wedge pressure (RAP:PAWP) ratio remained unchanged.

Comparison of left ventricular assist device, pulmonary arterial hypertension, and normal subjects.

Among these three cohorts, there was no difference in median ages, number of male subjects, body mass indices, and mean arterial pressures at rest (Table 1). Compared to normal cohort, resting median heart rates of LVAD and PAH subjects were higher (LVAD: 86, PAH: 74, normal: 63 bpm, p = 0.0042). LVAD subjects had lower hemoglobin (Hb) than normal subjects (12 vs. 15 g/dl, p = 0.0209), while PAH and normal subjects had comparable Hb (13 vs. 15 g/dl, p = 0.2). Laboratory values within two weeks of RHC in LVAD subjects (renal and hepatic function, prealbumin) were within normal limits (see Table S4, Supplemental Digital Content 4,

Table 1. - Comparisons of LVAD, PAH, and Normal Cohorts (Data Are Presented as Median With 25th–75th Interquartile Range)
LVAD PAH Normal p Value
Baseline characteristics
 Age (years) 54 (44–65) 55 (46–63) 51 (47–61) 0.9
 Male sex 92% 92% 92%
 Body mass index (kg/m2) 29 (27–33) 28 (25–31) 27 (24–28) 0.2
 Mean arterial pressure (mmHg) 86 (81–89) 87 (81–96) 88 (81–97) 0.8
 Heart rate (bpm) 86 (72–89) 74 (68–80) 63 (60–70) 0.0042
 Hemoglobin (g/dl) 12 (11–14) 13 (12–15) 15 (14–15) 0.06
CPET parameters of LVAD and PAH subjects
 pVO2 (ml/kg/min) 12.9 (10.4–16.9) 11.8 (8.8–12.5) 0.2
 RER 1.2 (1.2–1.3) 0.9 (0.9–1) 0.0001
 VE/VCO2 slope 34 (31–44) 46 (38–67) 0.01
 Peak heart rate 131 (122–154) 127 (111–137) 0.3
 Percent of predicted HR max 92 (88–99) 74 (69–81) 0.0001
 On beta blockers (n, %) 8 (66.7%) 0 0.0007
Echocardiographic findings of LVAD and PAH subjects closest to time of exercise
LVAD (rest) (n = 11) PAH (n = 12) p Value
LVEF (%, median, IQR) 15 (10–23) 65 <0.0001
RV dysfunction (n, %) 0.1
 Moderate-severe 4 (36.4%) 2 (16.7%)
 Mild 3 (27.3%) 2 (16.7%)
 None 4 (36.4%) 8 (66.7%)
 TR (n, %) 0.4
  Severe 0 0
  Moderate 0 2 (16.7%)
  Mild 5 (45.5%) 5 (41.7%)
  None 6 (54.5%) 5 (41.7%)
CPET, cardiopulmonary exercise test; IQR, interquartile range; LVAD, left ventricular assist device; LVEF, left ventricular ejection fraction; PAH, pulmonary arterial hypertension; RER, respiratory exchange ratio; RV, right ventricle; TR, tricuspid regurgitation.
*p < 0.05: value was compared to that of normal subjects.

Hemodynamic Responses to Exercise

Cardiopulmonary exercise test.

LVAD and PAH subjects shared similarly reduced exercise capacity, with median pVO2 of 12.9 and 11.8 ml/kg/min, respectively (Table 1). Compared with PAH, LVAD subjects achieved higher respiratory exchange ratio (1.2 vs. 0.9, p = 0.0001) and had a lower VE/VCO2 slope (34 vs. 46, p = 0.01). LVAD subjects did not exhibit chronotropic incompetence (median peak heart rate at 92% of APMHR), despite beta blocker use in eight subjects (67%). None of the PAH subjects was on beta blockers. As expected, left ventricular ejection fraction (LVEF) was significantly lower in LVAD versus PAH subjects, although the severity of RV dysfunction and TR by echocardiographic measurements were similar (Table 1).

Filling pressures and cardiac output.

Rest and exercise hemodynamics are shown in Table 2. All values in tables and text herein are presented as median with 25th–75th IQR. Duration of supine exercise in LVAD subjects was significantly shorter than PAH subjects (3.4 minutes [2.5–6.6] vs. nine minutes [4–10], p = 0.0175). LVAD subjects achieved lower work load than PAH subjects (17 watts [12.5–33] vs. 50 watts [25–55], p = 0.04). Normal subjects exercised at submaximal effort for 8–10 minutes per protocol.16

Table 2. - Rest and Exercise Hemodynamics of the Three Cohorts
LVAD (IQR) (n) PAH (IQR) (n) Normal (IQR) (n) p Value
Filling pressures and CO
 PAWP (mmHg) Rest 6 (4 to 16) (11) 10 (7 to 13) (12) 11 (9 to 12) (12) 0.3
Exercise 20 (12 to 25) (11) 14 (10 to 15) (12) 15 (14 to 17) (12) 0.4
 Mean PAP (mmHg) Rest 16 (14 to 18) (12) 32 (27 to 54) (12) 17 (15 to 19) (12) 0.0001
Exercise 34 (26 to 37) (12) 63 (44 to 83) (12) 24 (22 to 29) (12) 0.0001
 RAP (mmHg) Rest 3 (2 to 8) (11) 8 (4 to 8.5) (12) 6 (4.5 to 7) (12) 0.3
Exercise 12 (8 to 17) (11) 12 (5 to 18) (12) 6 (4.5 to 8) (12) 0.03
 DPG (mm Hg) Rest 4 (0 to 6) (11) 13 (9 to 29) (12) 0 (−1 to 2) (12) 0.0001
Exercise 9 (0 to 10) (11) 25 (5 to 56) (12) 0.5 (−3 to 2) (12) 0.005
 CO (L/min) Rest 4.9 (4.5 to 5.4) (11) 4.4 (3.8 to 5.4) (12) 5.2 (4.7 to 5.8) (12) 0.2
Exercise 6.2 (5.3 to 8) (11) 9.3 (7.6 to 11.5) (12) 10.7 (8.7 to 12.1) (12) 0.0029
Right ventricular afterload
 PVR (Wood units) Rest 2.2 (1.2 to 2.5) (11) 6.1 (4.1 to 9.5) (12) 1.1 (0.8 to 1.3) (12) 0.0001
Exercise 2.6 (1.4 to 3.2) (11) 5.2 (2.7 to 9.3) (12) 1 (0.7 to 1.2) (12) 0.0001
 PAC (ml/mmHg) Rest 3.3 (2.8 to 4.8) (10) 1.6 (1.2 to 2.4) (12) 6.1 (5.1 to 7.7) (12) 0.0001
Exercise 1.9 (1.2 to 3) (10) 1.8 (1.1 to 2.5) (12) 4.6 (3.8 to 6.8) (12) 0.0003
 Ea (mmHg/ml) Rest 0.5 (0.3 to 0.5) (11) 1 (0.7 to 1.5) (12) 0.3 (0.2 to 0.3) (12) 0.0001
Exercise 0.9 (0.7 to 1.6) (11) 1.1 (0.8 to 2) (12) 0.4 (0.3 to 0.4) (12) 0.0001
Right ventricular response to exercise
 RVSWI (mmHg/L/m2) Rest 0.3 (0.3 to 0.4) (10) 1 (0.8 to 1.5) (12) 0.5 (0.4 to 0.5) (12) 0.0001
Exercise 0.6 (0.3 to 0.7) (10) 1.8 (1.6 to 2.4) (12) 1.1 (0.8 to 1.2) (12) 0.0001
 PAPi Rest 4 (2.7 to 11.4) (12) 6.8 (3.5 to 7.6) (12) 2.6 (2 to 3.9) (12) 0.03
Exercise 1.7 (1.3 to 4.4) (12) 5.2 (3.3 to 7.7) (12) 3.1 (2.5 to 5.5) (12) 0.05
 RAP:PAWP Rest 0.5 (0.3 to 0.8) (11) 0.7 (0.5 to 0.8) (12) 0.6 (0.5 to 0.6) (12) 0.09
Exercise 0.6 (0.4 to 1.1) (11) 1 (0.3 to 1.4) (12) 0.4 (0.4 to 0.5) (12) 0.1
CO, cardiac output; CPET, cardiopulmonary exercise test; DPG, diastolic pulmonary gradient; IQR, interquartile range; LVAD, left ventricular assist device; LVEF, left ventricular ejection fraction; PAC, pulmonary artery compliance; PAH, pulmonary arterial hypertension; PAP, pulmonary artery pressure; PAPi, pulmonary artery pulsatility index; PAWP, pulmonary artery wedge pressure; PVR, pulmonary vascular resistance; RAP, right atrial pressure; RER, respiratory exchange ratio; RV, right ventricle; RVSWI, right ventricular stroke work index; TR, tricuspid regurgitation.
*p < 0.05: paired comparison between subjects’ rest and exercise hemodynamics.

At rest, subjects in all three cohorts had normal and similar PAWP (LVAD: 6, PAH: 10, normal: 11 mmHg, p = 0.3) (Table 2). Pulmonary arterial hypertension subjects had the highest mean PAP (LVAD: 16, PAH: 32, normal: 17 mmHg, p = 0.0001). Right atrial pressures of all subjects were in the normal range (LVAD: 3, PAH: 8, normal: 6 mmHg, p = 0.3). Diastolic pulmonary gradient (DPG) was the highest in PAH subjects, while LVAD and normal subjects had normal DPG (LVAD: 4, PAH: 13, normal: 0 mmHg, p = 0.0001).

During exercise, compared with rest hemodynamics, all cohorts had significant increases in PAWP (LVAD: 20, PAH: 14, normal: 15 mmHg, p = 0.4) and mean PAP (LVAD: 34, PAH: 63, normal: 24 mmHg, p = 0.0001). However, only LVAD and PAH subjects had significant rises in RAP, while RAP was unchanged in normal subjects (LVAD: 12, PAH: 12, normal: 6 mmHg, p = 0.0326). DPG of LVAD and PAH cohorts numerically increased during exercise but the changes were statistically insignificant, while normal subjects had no change (LVAD: 9, PAH: 25, normal: 0.5, p = 0.0050). Cardiac output at rest was normal and similar in all subjects (LVAD: 4.9, PAH: 4.4, normal: 5.2 L/min, p = 0.2). On exercise, CO increased in all cohorts, but to a greater extent in PAH and normal subjects (LVAD: 6.2, PAH: 9.3, normal: 10.7 L/min, p = 0.0029).

To further understand the magnitudes of hemodynamic changes, we compared the absolute differences between exercise and rest values among the three cohorts (Δ = exercise − rest values) (Figure 1A). Although all groups had PAWP increases during exercise, the changes were similar (ΔPAWP in LVAD: 8, PAH: 5, normal: 5 mmHg, p = 0.2). Increases in mean PAP were larger in LVAD and PAH groups compared with normal (ΔmPAP in LVAD: 17, PAH: 22, normal: 8 mmHg, p = 0.0177). Uncoupling of PAD and PAWP on exercise was numerically larger in LVAD and PAH cohorts compared to normal but p > 0.05 (ΔDPG in LVAD: 5, PAH: 13, normal: 0 mmHg, p = 0.3, not shown). Notably, the rise in RAP was largest in LVAD, followed by PAH while there was minimal change in normal subjects (ΔRAP in LVAD: 7, PAH: 4; normal: 1 mmHg, p = 0.0179). Left ventricular assist device subjects could not augment CO as much as PAH and normal subjects (ΔCO in LVAD: 1.5, PAH: 4.3, normal: 5.7 L/min, p = 0.0014).

Figure 1.
Figure 1.:
Hemodynamic differences (Δvalues) between exercise and rest. A: Intracardiac filing pressures and cardiac output (CO) (inset); (B) Right ventricular afterload measurements; (C): Right ventricular response to exercise. Line at 0 (y-axis) denotes “no change.”

Right ventricular afterload.

Right ventricular afterload values at rest and during exercise are shown in Table 2. At rest, LVAD subjects had higher pulmonary vascular resistance (PVR) than that of normal but lower than PAH subjects’ (LVAD: 2.2, PAH: 6.1, normal: 1.1 Wood units, p = 0.0001). The same hierarchy held true for pulmonary artery compliance (PAC) (LVAD: 3.3, PAH: 1.6, normal: 6.1 ml/mmHg, p = 0.0001), and pulmonary arterial elastance (Ea) (LVAD: 0.5, PAH: 1.0, normal: 0.3 mmHg/ml, p = 0.0001). During exercise, PVR numerically increased in LVAD, significantly dropped in PAH while remained unchanged in normal subjects (LVAD: 2.6, PAH: 5.2, normal: 1 Wood units, p = 0.0001). Absolute PVR change was the highest in LVAD subjects (ΔPVR in LVAD: 1.3, PAH: −0.9, normal: −0.2 Wood units, p = 0.0074) (Figure 1B). Conversely, PAC lowered more in LVAD than PAH and normal subjects (ΔPAC in LVAD: −1.5, PAH: 0, normal: −0.6 ml/mmHg, p = 0.0023). Ea behaved similarly (ΔEa in LVAD: 0.4, PAH: 0.1, normal: 0.1 mmHg/ml, p = 0.0024).

Right ventricular reserve on exercise.

To examine RV reserve during exercise, we measured RVSWI, PAPi, and RAP:PAWP ratio. All subjects augmented RVSWI from rest to exercise (Table 2), although LVAD subjects had the lowest RVSWI at rest (LVAD: 0.3, PAH: 1, normal: 0.5 mmHg/L/m2, p = 0.0001), and their exercise-induced RVSWI augmentation was also the smallest (ΔRVSWI in LVAD: 0.2, PAH: 1, normal 0.6 mmHg/L/m2, p = 0.003) (Figure 1C). Pulmonary artery pulsatility index (PAPi) in LVAD and PAH subjects was higher than that of normal subjects at rest (LVAD: 4, PAH: 6.8, normal: 2.6, p = 0.0264). On exercise, only normal subjects significantly augmented PAPi (ΔPAPi in LVAD: −1.3, PAH: −0.6, normal: 0.7, p = 0.024) (Figure 1C, inset). Right atrial pressure:pulmonary artery wedge pressure ratio (RAP:PAWP) was similar at rest in all cohorts (LVAD: 0.5, PAH: 0.7, normal: 0.6, p = 0.09). During exercise, this ratio significantly decreased in normal subjects but were unchanged in LVAD and PAH (LVAD: 0.6, PAH: 1, normal: 0.4, p = 0.1). ΔRAP:PAWP in LVAD was higher than in normal subjects (0.2 vs. −0.2, p = 0.040) (Figure 1C).

Correlation Between Right Ventricular Afterload and Exercise Capacity in Left Ventricular Assist Device Patients

Of the filling pressures during exercise in LVAD patients, only RAP correlated significantly with pVO2 (r = −0.7, p = 0.0287) and VE/VCO2 slope (r = 0.8, p = 0.0174), while PAP and PAWP did not correlate with either. Exercise CO trended to correlate with pVO2 (r = 0.6, p = 0.08) and VE/VCO2 slope (r = −0.6, p = 0.07). There was also a trend of positive correlation between PAC and pVO2 (r = 0.6, p = 0.07), and between PVR and VE/VCO2 slope (r = 0.5, p = 0.09). Notably, lower pVO2 correlated significantly with higher exercise Ea (r = −0.8, p = 0.0101), while the opposite was true between VE/VCO2 slope and Ea (r = 0.7, p = 0.0132) (Figure 2). Exercise PAPi trended to correlate with pVO2 (r = 0.6, p = 0.08), and exercise RVSWI correlated with VE/VCO2 slope (r = −0.8, p = 0.01). There was no correlation between exercise RAP:PAWP and pVO2 or VE/VCO2 slope.

Figure 2.
Figure 2.:
Correlations between exercise capacity and right ventricular afterload. Correlation between pVO2 (A) and VE/VCO2 slope (B) with pulmonary arterial elastance (Ea) at peak exercise in left ventricular assist device (LVAD) patients. Shaded area: 95% CI.


In this study, we compared exercise hemodynamics between three age- and sex-matched cohorts: LVAD, PAH, and normal subjects. We found that while PAH subjects had the highest RV afterload at rest, during exercise, only LVAD subjects had the largest relative RV afterload increase, the lowest CO augmentation, and the poorest RV reserve. High RV afterload in LVAD subjects correlated with poor exercise capacity.

Baseline RV impairment in LVAD subjects was evidenced by preoperative hemodynamics (RAP:PCWP > 0.5,17 Table S3, Supplemental Digital Content 3, However, postoperatively, RV function appeared to improve (less severe RV dysfunction and TR compared to pre-LVAD echocardiograms [not shown], PAPi increasing from 2.1 to 4 [Table S3, Supplemental Digital Content 3,]). It has been shown that risk scores predicting postoperative RV dysfunction performed modestly when applied to external cohorts,18,19 indicating that our ability to predict RV failure post-LVAD using pre-LVAD parameters remains limited. Therefore, we elected to analyze exercise tolerance based on postoperative RV function at the time of exercise, as opposed to RV function before LVAD insertion.

At rest, LVAD subjects had lower PAWP than values obtained from prior studies. For example, in a cohort of 44 patients with HeartMate II LVAD (Abbott, IL) undergoing hemodynamic ramp test, Jung et al.20 showed a mean baseline PAWP of 10.8 ± 6.7 mmHg. Muthiah et al.21 exercised 19 patients with HeartWare LVAD on a supine ergometer and showed a resting PAWP of 13.9 ± 6.9 mmHg, which increased to 24.9 ± 6.3 mmHg on exercise.21 In our study, we observed that LVAD, PAH, and normal subjects all had similar exercise-induced increases in PAWP. Although not statistically significant, numerically ΔPAWP in LVAD subjects was higher (8 mmHg, IQR, 4–15) than PAH (5 mmHg, IQR, 0–8) or normal subjects (5 mmHg, IQR, 2–6). In our highly selected cohort, inadequate LV unloading was unlikely to account for our observations of higher RV afterload and impaired RV reserve during exercise.

Compared with PAH and normal cohorts, LVAD subjects faced a unique problem of increased PVR during exercise. We found a proportional decline in PAC and increase in total RV afterload (Ea) suggesting the change in PVR was real. On the contrary, in PAH and normal subjects, despite a PAC decrease on exercise, PVR either dropped or remained constant, leading to less significant increases in Ea. The mechanism for exercise-induced PVR increase in LVAD subjects remains unknown, but has previously been observed in patients with HF and reduced EF,22 and most likely stemmed from low exercise CO augmentation.

Left ventricular assist device subjects had the highest rise in RAP during exercise relative to PAH and normal subjects. They also had the lowest CO and RVSWI augmentation. Exercise PAPi in LVAD subjects trended to worsening, and RV adaptation to load was also abnormal, with RAP:PAWP ratio unchanged on exercise versus decreased in normal subjects. It is worth noting that LVAD subjects at baseline were well-compensated, with low biventricular filling pressures, normal CO, low NYHA class, and normal laboratory values (except for mildly decreased Hb, which could have contributed to lower pVO2). Additionally, LVAD subjects did not have worsening RV hemodynamics following pump implant (Table S3, Supplemental Digital Content 3, Despite this compensation at rest, our data suggested they have more impaired RV reserve compared to PAH and normal subjects. We observed a trend toward worsening pulmonary decoupling (i.e. larger DPG increase) on exercise in LVAD and PAH patients compared to normal. Elevated DPG after LVAD implant has been implicated in RV failure and poor outcomes.23,24 It remains to be seen if pulmonary decoupling would be associated with impaired RV response to exercise in LVAD subjects. Finally, our findings of large exercise-induced RAP rise and poor CO augmentation were consistent with findings in Muthiah et al.,21 although our LVAD subjects’ median RAP appeared to be lower (3 vs. 10 mmHg), suggesting better compensated RV at rest.

For patients with chronic HF, RV dysfunction significantly contributes to poor exercise tolerance.25 The right heart’s role during exercise in LVAD patients is less clear. In two previous manuscripts, one reported a significant association between RVEF and pVO2,26 but another did not.27 However, EF is a poor measure of RV contractile function, given the load dependence of the RV.28 A previous study by Houston et al.11 reported that while RV afterload decreased following LVAD implantation, RV adaptation to load (RAP:PAWP ratio) worsened and remained abnormal over time. In this study, we showed that high RV afterload (Ea) strongly correlated with exercise intolerance (low pVO2) and ventilatory inefficiency (high VE/VCO2 slope). The correlation between high Ea and high VE/VCO2 slope was plausible, as the latter would increase if pulmonary hypoperfusion occurred due to RV dysfunction. It remains unclear whether LVAD subjects’ poor RV performance stemmed from preexisting RV dysfunction, and/or adaptation to continuous-flow physiology following pump implant. While LVAD patients may have low pVO2 due to anemia, chronotropic incompetence, or inefficient peripheral O2 extraction,4 we suggest that impaired RV reserve may be added as a predictor of poor exercise performance. “Sub-phenotyping” exercise intolerance in LVAD subjects and determining the relative contributions of poor RV reserve, inadequate LV unloading, chronotropic incompetence, and/or impaired peripheral O2 extraction could prove useful. This may allow for better study of targeted therapeutics, such as the role of pulmonary vasodilators in those with disproportionately high RV afterload during exercise, speed modulation in those with inadequate LV unloading, beta blocker reduction or rate-responsive pacing in the setting of chronotropic incompetence.


This is a pilot study of a small LVAD cohort. Results need to be confirmed in a larger group before they can be generalized. Our findings may have been impacted by selection bias, given the exceptionally normal resting hemodynamics in LVAD subjects. Another important limitation was the low number of female subjects by chance. Exercise hemodynamics and CPET were obtained on two separate days. In between, although patients’ clinical courses remained stable, we could not account for all physiologic changes such as volume status that might have affected exercise tolerance. Large respiratory variations could affect RHC values, although we attempted to control for this error with interobserver variability analysis. Previous studies have suggested exercise hemodynamic measurements be averaged across several respiratory cycles.29 We elected to obtain end-expiratory values for LVAD subjects to be consistent with values obtained at the other centers.

Regarding RV imaging, we could not reliably visualize the RV with echocardiogram for the entire cohort due to poor image quality during exercise. In fact, prior studies in exercising LVAD patients had to rely on nuclear imaging or cardiac computed tomography to measure RV function.26,27

The three LVAD, PAH, and normal cohorts, though matched for sex, age, and BMI, underwent slightly different exercise protocols at three separate institutions, therefore unknown confounders could not be accounted for. Notably, LVAD and PAH subjects exercised on a supine ergometer, while normal subjects did so in the semi-upright position. Compared with the supine position, PAP, PAWP, and CO can be lower when the body is upright, although the differences diminish at higher levels of exercise.29 Given this finding, the similar exercise-induced PAWP increases among the three cohorts were even more surprising since the normal cohort should have had the least PAWP increase based on positioning. In previous studies, LVAD subjects were thought to resemble patients with HF with preserved EF in their response to exercise (i.e. large exercise-induced PAWP rise).21,30 One reason could have been selection bias, as two out of 12 LVAD subjects reached PAWP of 27 and 30 mmHg at peak exercise. Alternatively, all LVAD subjects had pump implant via the lateral thoracotomy approach as part of routine practice at University of Maryland. This technique was thought to be associated with improved inflow cannula position,31 which could lead to better LV unloading and less PAWP rise during exercise. This hypothesis requires further investigation. While we could not directly compare the mechanisms of RV response to afterload in LVAD versus PAH subjects due to disparate disease pathophysiology, we used the PAH cohort to illustrate the markedly poor RV reserve in our LVAD cohort.


In this study, we showed that well-compensated LVAD subjects, compared to those with PAH or normal non-athletes, had exercise-induced increases in RV afterload and evidence of poor RV reserve. These changes occurred in the context of poor CO augmentation, increased PVR, decreased PAC and increased total RV afterload (Ea). We demonstrated a strong correlation between high RV afterload during exercise and low pVO2 as well as elevated VE/VCO2 slope.


1. Kormos RL, Cowger J, Pagani FD, et al. The Society of Thoracic Surgeons Intermacs database annual report: Evolving indications, outcomes, and scientific partnerships. J Heart Lung Transplant. 2019; 38:114–126
2. Pagani FD, Milano CA, Tatooles AJ, et al. HeartWare HVAD for the treatment of patients with advanced heart failure ineligible for cardiac transplantation: Results of the ENDURANCE Destination Therapy Trial. J Heart Lung Transpl. 2015; 34:S9
3. Mehra MR, Goldstein DJ, Uriel N, et al.; MOMENTUM 3 Investigators. Two-year outcomes with a magnetically levitated cardiac pump in heart failure. N Engl J Med. 2018; 378:1386–1395
4. Jung MH, Gustafsson F. Exercise in heart failure patients supported with a left ventricular assist device. J Heart Lung Transplant. 2015; 34:489–496
5. Cowger JA, Naka Y, Aaronson KD, et al.; MOMENTUM 3 Investigators. Quality of life and functional capacity outcomes in the MOMENTUM 3 trial at 6 months: A call for new metrics for left ventricular assist device patients. J Heart Lung Transplant. 2018; 37:15–24
6. Jung MH, Houston B, Russell SD, Gustafsson F. Pump speed modulations and sub-maximal exercise tolerance in left ventricular assist device recipients: A double-blind, randomized trial. J Heart Lung Transplant. 2017; 36:36–41
7. Apostolo A, Paolillo S, Contini M, et al. Comprehensive effects of left ventricular assist device speed changes on alveolar gas exchange, sleep ventilatory pattern, and exercise performance. J Heart Lung Transplant. 2018; 37:1361–1371
8. Noor MR, Bowles C, Banner NR. Relationship between pump speed and exercise capacity during heartMate II left ventricular assist device support: Influence of residual left ventricular function. Eur J Heart Fail. 2012; 14:613–620
9. Methvin AB, Owens AT, Emmi AG, et al. Ventilatory inefficiency reflects right ventricular dysfunction in systolic heart failure. Chest. 2011; 139:617–625
10. Arena R, Myers J, Abella J, et al. Development of a ventilatory classification system in patients with heart failure. Circulation. 2007; 115:2410–2417
11. Houston BA, Kalathiya RJ, Hsu S, et al. Right ventricular afterload sensitivity dramatically increases after left ventricular assist device implantation: A multi-center hemodynamic analysis. J Heart Lung Transplant. 2016; 35:868–876
12. Wasserman K, Hansen JE, Sue DY, et al. Principles of Exercise Testing and Interpretation. 2012. 5th ed. Philadelphia, PA: Lippincott Williams & Wilkins
13. Brawner CA, Ehrman JK, Schairer JR, Cao JJ, Keteyian SJ. Predicting maximum heart rate among patients with coronary heart disease receiving beta-adrenergic blockade therapy. Am Heart J. 2004; 148:910–914
14. Galiè N, Humbert M, Vachiery JL, et al. 2015 ESC/ERS Guidelines for the diagnosis and treatment of pulmonary hypertension: The joint task force for the diagnosis and treatment of pulmonary hypertension of the European Society of Cardiology (ESC) and the European Respiratory Society (ERS): Endorsed by: Association for European Paediatric and Congenital Cardiology (AEPC), International Society for Heart and Lung Transplantation (ISHLT). Eur Respir J. 2015; 46:903–975
15. Hsu S, Houston BA, Tampakakis E, et al. Right ventricular functional reserve in pulmonary arterial hypertension. Circulation. 2016; 133:2413–2422
16. Wright SP, Granton JT, Esfandiari S, Goodman JM, Mak S. The relationship of pulmonary vascular resistance and compliance to pulmonary artery wedge pressure during submaximal exercise in healthy older adults. J Physiol. 2016; 594:3307–3315
17. Soliman OII, Akin S, Muslem R, et al.; EUROMACS Investigators. Derivation and validation of a novel right-sided heart failure model after implantation of continuous flow left ventricular assist devices: The EUROMACS (European Registry for Patients with Mechanical Circulatory Support) right-sided heart +failure risk score. Circulation. 2018; 137:891–906
18. Kalogeropoulos AP, Kelkar A, Weinberger JF, et al. Validation of clinical scores for right ventricular failure prediction after implantation of continuous-flow left ventricular assist devices. J Heart Lung Transplant. 2015; 34:1595–1603
19. Houston BA, Shah KB, Mehra MR, Tedford RJ. A new “twist” on right heart failure with left ventricular assist systems. J Heart Lung Transplant. 2017; 36:701–707
20. Jung MH, Gustafsson F, Houston B, Russell SD. Ramp Study hemodynamics, functional capacity, and outcome in heart failure patients with continuous-flow left ventricular assist devices. ASAIO J. 2016; 62:442–446
21. Muthiah K, Robson D, Prichard R, et al. Effect of exercise and pump speed modulation on invasive hemodynamics in patients with centrifugal continuous-flow left ventricular assist devices. J Heart Lung Transplant. 2015; 34:522–529
22. Butler J, Chomsky DB, Wilson JR. Pulmonary hypertension and exercise intolerance in patients with heart failure. J Am Coll Cardiol. 1999; 34:1802–1806
23. Imamura T, Chung B, Nguyen A, et al. Decoupling between diastolic pulmonary artery pressure and pulmonary capillary wedge pressure as a prognostic factor after continuous flow ventricular assist device implantation. Circ Heart Fail. 2017; 10:e003882
24. Alnsasra H, Asleh R, Schettle SD, et al. Diastolic pulmonary gradient as a predictor of right ventricular failure after left ventricular assist device implantation. J Am Heart Assoc. 2019; 8:e012073
25. Murninkas D, Alba AC, Delgado D, et al. Right ventricular function and prognosis in stable heart failure patients. J Card Fail. 2014; 20:343–349
26. Lairez O, Delmas C, Fournier P, et al. Feasibility and accuracy of gated blood pool SPECT equilibrium radionuclide ventriculography for the assessment of left and right ventricular volumes and function in patients with left ventricular assist devices. J Nucl Cardiol. 2018; 25:625–634
27. Mirza KK, Jung MH, Sigvardsen PE, et al. Computed tomography-estimated right ventricular function and exercise capacity in patients with continuous-flow left ventricular assist devices. ASAIO J. 2020; 66:8–16
28. Morimont P, Lambermont B, Ghuysen A, et al. Effective arterial elastance as an index of pulmonary vascular load. Am J Physiol Heart Circ Physiol. 2008; 294:H2736–H2742
29. Kovacs G, Herve P, Barbera JA, et al. An official European Respiratory Society statement: Pulmonary haemodynamics during exercise. Eur Respir J. 2017; 50:1700578
30. Burrell A, Hayward C, Mariani J, Leet A, Kaye DM. Clinical utility of invasive exercise hemodynamic evaluation in LVAD patients. J Heart Lung Transplant. 2015; 34:1635–1637
31. McGee E Jr, Danter M, Strueber M, et al. Evaluation of a lateral thoracotomy implant approach for a centrifugal-flow left ventricular assist device: The LATERAL clinical trial. J Heart Lung Transplant. 2019; 38:344–351

left ventricular assist device; right ventricular afterload; exercise hemodynamics; cardiopulmonary exercise stress test; pulmonary arterial hypertension

Supplemental Digital Content

Copyright © 2020 by the ASAIO