|Home | About | Journals | Submit | Contact Us | Français|
The mechanism of functional limitation in heart failure with preserved ejection fraction (HFpEF) remains controversial. We examined the contributions of central cardiac and peripheral mechanisms and hypothesized that the pulmonary vascular response to exercise is an important determinant of aerobic capacity among patients with exertional pulmonary venous hypertension (ePVH).
We compared 31 ePVH patients (peak VO2<80% predicted and peak pulmonary arterial wedge pressure≥20 mmHg) with 31 age and gender matched controls (peak VO2>80% predicted) who underwent invasive cardiopulmonary exercise testing for unexplained exertional intolerance. ePVH patients had lower peak cardiac output (73±14 vs 103±18 % predicted; p<0.001) compared to controls, related both to impaired chronotropic response (peak HR 111±25 vs 136±24 bpm; p<0.001) and to reduced peak stroke volume index (47±10 vs 54±15 mL/min/m2; p=0.03). Peak systemic O2 extraction was not different between groups (arterial-mixed venous oxygen content difference: 13.0±2.1 vs 13.4±2.4 mL/dL; p=0.46). ePVH patients had higher resting (150±74 vs 106±50 dyne.s.cm−5; p=0.009), peak (124±74 vs 70±41 dyne.s.cm−5; p<0.001) and isoflow pulmonary vascular resistance (PVR; 124±74 vs 91±33 dyne.s.cm−5 at CO~10.6L/min; p=0.04). PVR decreased with exercise in all control subjects but increased in 36% (n=11) of ePVH patients. Abnormal pulmonary vascular response was not associated with peak VO2.
Reduced cardiac output response, rather than impaired peripheral O2 extraction, constrains oxygen delivery and aerobic capacity in ePVH. Pulmonary vascular dysfunction is common in patients with ePVH at rest and with exercise.
Heart failure (HF) is characterized by the inability to provide adequate cardiac output (CO) to satisfy peripheral metabolic needs at normal intracardiac filling pressures1. In approximately half of HF patients, left ventricular (LV) ejection fraction (EF) is preserved (HFpEF). These patients demonstrate similar symptoms, prognosis, and functional capacity to those with reduced ejection fraction heart failure (HFrEF)2-4. While exertional intolerance in HFrEF is predominantly attributed to the reduced CO augmentation, the primary mechanism of functional limitation in HFpEF remains controversial. Diastolic dysfunction, impaired cardiac contractility, abnormal peripheral vasodilation, chronotropic incompetence and abnormal skeletal muscle may all limit maximal oxygen delivery and consumption5-8. Many studies ascribe exertional intolerance in HFpEF to inadequate CO5, 9-12, with increasing attention towards the role of the right ventricle (RV) and pulmonary circulation in limiting cardiac output13. In HFrEF, exercise performance inversely correlates with the slope of the flow-related increase in pulmonary arterial pressure and pulmonary vascular resistance (PVR)14, 15. The prevalence of pulmonary hypertension (PH) in HFpEF subjects is high and PH is independently associated with poor prognosis in this population16. Nonetheless, the relative impact of abnormalities of the pulmonary circulation on functional capacity in HFpEF is poorly understood. The central limitation paradigm itself, however, has been challenged; others have suggested that abnormal peripheral O2 extraction may be the primary limiting factor to exercise performance17-19. A better understanding of the relative weight of central and peripheral mechanisms to the functional impairment in HFpEF will help to identify relevant pathophysiological subgroups in this heterogeneous disease and to target therapies.
We examined the physiological correlates of impaired aerobic capacity in patients with unexplained dyspnea, preserved LV ejection fraction and elevated left ventricular filling pressure at peak exercise, exploring the specific contribution of central cardiac and peripheral mechanisms. Specifically, we hypothesized that the pulmonary vascular response limits cardiac output and is therefore an important determinant of maximal aerobic capacity among these patients.
This exploratory study included patients with exertional intolerance of indeterminate cause referred to the Dyspnea Clinic at Brigham and Women's Hospital between March 2011 and September 2013 who underwent resting supine right heart catheterization (RHC) followed by upright cycle invasive cardiopulmonary exercise testing (iCPET)20. We excluded those with submaximal exercise testing (respiratory exchange ratio < 1.0), moderate or severe mitral and/or aortic valve disease, LVEF<50%, FEV1/FVC ratio<70% predicted, marked anemia (hemoglobin<11g/dL), or diagnostic testing suggestive of clinically relevant myocardial ischemia.
Based on prior studies reporting peak upright exercise PAWP of healthy subjects21-23, we defined our study sample as: pulmonary artery wedge pressure (PAWP)≥20mmHg at peak exercise and peak VO2<80% predicted24. Despite having symptoms compatible with HF, preserved LVEF and abnormal PAWP increase at peak exercise in the absence of reduced LVEF and significant valvular disease, we acknowledge a limited ability to apply our findings to community-based HFpEF patients because of the referral bias and the use of the exercise hemodynamics criterion; therefore we will herein refer to them as exertional pulmonary venous hypertension (ePVH). Because our aim was to study the physiological correlates of impaired aerobic capacity, our age and gender matched control group comprised patients referred for iCPET that were found to have normal aerobic capacity (peak VO2≥80% predicted) irrespective of central hemodynamics. LVEF was estimated by transthoracic echocardiography by board certified staff echocardiographers at Brigham and Women's Hospital by visual estimation, Teichholtz, or biplane Simpson's methods. Charts were reviewed for demographic, anthropometric and clinical baseline characteristics. Data from RHC and hemodynamic, respiratory and metabolic data at peak exercise testing from iCPET were also collected. The Partners Human Research Committee approved this retrospective chart review and waived the requirement for informed consent.
A flow-directed, balloon-tipped, 4-port pacing pulmonary artery catheter (Edwards Lifesciences, Irvine, CA) was placed into the pulmonary artery, with fluoroscopic guidance as necessary. An arterial line was inserted into the radial artery using a 20-gauge IV or 5-French sheath. End expiratory systemic arterial, right atrial (RAP), right ventricular (RVP), pulmonary artery pressures (PAP) and PAWP were measured using a hemodynamic monitoring system (Xper Cardio Physiomonitoring System, Philips, Andover, Massachusetts) calibrated before each study. The pressure transducer was leveled using as references the mid axillary line (supine) and 5cm below the axillary fold (upright)25.
All exercise tests were performed in the Brigham and Women's Hospital cardiopulmonary exercise laboratory using an upright cycle ergometer with the subject breathing room-air. Two minutes of rest were followed by 2 min of unloaded cycling at 55-65 RPM. Work rate was continuously increased using at 5, 10, 15, or 20 W/min, chosen on the basis of exertional tolerance history, to a symptom-limited maximum. Minute ventilation (VE), pulmonary gas exchange, heart rate (HR), radial arterial blood pressure (BP), RAP, RVP and PAP were measured continuously while PAWP and a 12-lead EKG were obtained at rest and each minute of exercise. Blood samples were simultaneously drawn from the radial artery and pulmonary artery during the last minute of the rest period and during the last 15 s of each minute during exercise, during a 2 minute unloaded cycling recovery period immediately following peak exercise. Systemic arterial and PA samples were analyzed for PO2, PCO2, pH and O2 saturation (SaO2), hemoglobin concentration (Hb), and O2 content (CaO2 and CvO2, respectively) by co-oximetry. Breath-by-breath pulmonary gas exchange was measured using a commercially available metabolic cart (MGC Diagnostics, St. Paul MN).
Resting ventilatory and gas exchange data were obtained from the averaged final 30-s interval of the 2-minute rest period. Exercise ventilatory and gas exchange data were averaged over contiguous 30-s intervals. Peak VO2 was defined as the highest 30-s averaged VO2 during the last minute of the symptom-limited exercise test. CO was calculated using the Fick principle (CO = VO2/[CaO2 - CvO2]), and stroke volume (SV) as CO/HR. Arterial-mixed venous oxygen content difference (CavO2diff) was calculated as the difference between CaO2 and CvO2. Predicted maximal CO was calculated from predicted peak VO2 and an assumed maximal arterial-mixed venous O2 content difference equivalent to hemoglobin concentration ([Hb]) for healthy subjects26. PVR, transpulmonary gradient (TPG), diastolic pressure gradient, systemic vascular resistance (SVR), percentage of predicted maximal HR, and heart rate reserve (HRR) were calculated using standard formulas (see supplement 1). To assess exercise-induced hemodynamic changes, we used the upright resting data (using measured VO2 to calculate Fick CO), instead of supine RHC data (which use thermodilution or estimated VO2 to calculate Fick CO) to avoid the error of using different methods to measure CO and the error of the hemodynamic changes due to different body position.
Continuous variables are expressed as mean ± standard deviation or median [25th, 75th percentiles] as appropriate for distribution. Categorical variables are expressed as number of subjects and proportion [n (%)]. Comparisons between groups were performed using 2-sided parametric or non-parametric tests (unpaired or paired t or Wilcoxon rank sums) for normally and non-normally distributed data respectively. Fisher's exact test was applied to compare proportions. One-way ANOVA with the Bonferroni correction was used to perform multiple group comparisons. Univariate linear regression analysis was performed to study associations between aerobic capacity and clinical and physiological variables. Correlations between hemodynamic and metabolic variables were determined using Pearson or Spearman correlation, as appropriate. A two sided p-value <0.05 was considered significant. Statistical analysis was performed using Stata software Version 12.1 (Stata Corp LP, College Station, TX, USA).
Demographic data, comorbidities and medications among controls (n=31) and ePVH patients (n=31) are presented in Table 1. Controls were well matched for age and gender. Patients with ePVH had higher BMI and a higher prevalence of coronary artery disease. They were also more likely to be prescribed β-blockers, angiotensin-converting enzyme inhibitors and diuretics. There were no significant differences in hemoglobin concentration, estimated glomerular filtration rate or N-terminal pro-brain natriuretic peptide levels.
Resting supine hemodynamic data are presented in Table 1. ePVH patients had higher PAWP than controls (15±5 vs 12±5 mmHg; p=0.02), with a PAWP>15 mmHg in 13 (42%) ePVH patients versus 5 (16%) controls (p=0.03). ePVH patients also had higher mPAP, RAP and PVR compared to control group. 11 (36%) ePVH patients presented a hemodynamic profile compatible with group 2 pulmonary hypertension defined by a mPAP≥25 mmHg and PAWP>15mmHg, compared to 3 (10%) of control group. Among ePVH patients, 4 (13%) had a PVR>240 dyne.s.cm−5, 16 (52%) had a TPG>12 mmHg and 2 (7%) had a diastolic pressure gradient (diastolic PAP – PAWP) > 7 mmHg. Resting PVR did not correlate with resting PAWP (r=0.25; p=0.18). Resting CI and HR were similar between groups (see Table 1). None of the supine resting hemodynamic variables was associated with % predicted peak VO2. The hemodynamic and metabolic data of resting upright are displayed in Table 2. Differences in right and left heart filling pressures persisted with the transition to upright posture, except that PVR increased modestly in the control group.
Exercise hemodynamic and metabolic data are displayed in Table 2. By study design, ePVH patients had lower peak VO2 than controls. Peak PAWP in ePVH correlated modestly with % predicted peak VO2 (r=−0.33; p=0.07). Among ePVH patients, β-blocker treatment was the only baseline feature associated with % predicted peak VO2 (p=0.02).
Compared to controls, peak CO was decreased in ePVH patients (Figure 1), and this difference persisted after adjusting individually for age, gender and body size. For each liter of O2 consumed, patients from both groups had the same increase in CO, as there was no significant difference in ΔCO/ΔVO2 slope (p=0.38). At peak exercise, we observed similar absolute peak CavO2diff values between groups (13.4±2.4 vs 13.0±2.1 mL/dL; p=0.46).
ePVH patients demonstrated impaired chronotropic response, with lower absolute peak HR, % of predicted peak HR and HRR (Table 2). HRR was significantly, though modestly, correlated with peak VO2 (r=0.37; p=0.04). Likewise, ePVH patients had a lower peak SVi (Table 2) and decreased SVi augmentation (ΔSVi: 14±12 vs 21±12 mL/m2; p=0.05) compared to controls.
Excluding those on β-blockade, ePVH patients and controls (12 and 23, respectively) had similar HR response to exercise (peak HR: 84±10 vs 88±14 % predicted; p=0.29). CO remained significantly lower among ePVH patients in this subset of patients (78±9 vs 103±18 % predicted; p<0.001), primarily related to lower peak SVi (46±10 vs 56±16 mL/m2; p=0.06). Peak CavO2diff was similar between ePVH and controls in this subset of patients (13.0±1.7 vs 13.6±2.7 mL/dL; p=0.48), consistent with no peripheral limitation.
SVR was similar at rest between ePVH and controls, and decreased in both groups. With exercise, a less pronounced fall in SVR was noted in ePVH (−862±617 vs −1263±471 dyne.s.cm−5; p=0.008), resulting in a higher peak exercise SVR (782±222 vs 589±208 dyne.s.cm−5; p=0.001). In ePVH group, peak SVR modestly correlated with % predicted peak VO2 (r=−0.39; p=0.04).
In addition to the higher resting PVR, ePVH patients had higher peak PVR (124±74 vs 70±41 dyne.s.cm−5; p<0.001). Thirteen (43%) ePVH patients had a peak PVR above 120 dyne.s.cm−5 compared to 3 (10%) controls; 11 (36%) ePVH patients had an increase in PVR during exercise compared to none of control subjects (Figure 2). Using as reference the peak CO of ePVH group, we compared an isoflow (10.6±1.1 vs 10.6±3.5 L/min; p=0.98) PVR between control and ePVH groups. ePVH patients had an increased isoflow PVR (124±74 vs 91±33 dyne.s.cm−5; p=0.04) when compared to controls. The control group had a greater decrease in PVR during exercise (ΔPVR: −70±51 vs −24±66 dyne.s.cm−5; p=0.004). The VE/VCO2 slope modestly correlated with peak PVR (r=0.39; p=0.03) in ePVH group.
Among ePVH patients, peak PVR was not associated with % predicted peak VO2 (r=−0.20; p=0.29). There were no significant differences in peak VO2 of ePVH patients divided by peak PVR tertiles (68±8 vs 69±10 vs 63±10 % predicted; p=0.34). Likewise, the presence of resting group 2 PH (p=0.74), a TPG>12 mmHg (p=0.12) and an abnormal increase in PVR (p=0.28) during exercise were not associated with the degree of aerobic impairment (% predicted peak VO2). Also notable, peak PVR was not correlated with peak PAWP in ePVH group (r=0.10; p=0.43).
These data demonstrate that reduced aerobic capacity in ePVH patients is constrained by a central cardiac limit rather than a peripheral one, and that impaired CO reserve is determined by reduced SV augmentation and inadequate chronotropic response during exercise, with the latter related to β-blocker treatment. Systemic O2 extraction in ePVH was the same as for controls and did not contribute to decreased peak VO2. Also, pulmonary vascular dysfunction is present in a sizable subset of ePVH patients, as evidenced by the increased resting, peak and isoflow PVR, and also by the abnormal increase in PVR during exercise observed in a significant subset of patients.
Low peak VO2 may be the result of inadequate CO augmentation or abnormal CavO2diff, which is a function of pulmonary blood oxygenation, hemoglobin O2 transport capacity (which together comprise arterial O2 content) and systemic O2 extraction by working muscles from systemic capillaries. In the current study, peak CO (whether considered as absolute value, BSA normalized, or adjusted to age, gender and body size) was reduced in ePVH patients compared to age and gender-matched controls. In contrast, peak CavO2diff was normal for both groups. Therefore, consistent with several previous reports5, 6, 9, 12 but contrary to others17, 18, these data do not support an important role for abnormal peripheral O2 extraction in aerobic limitation in HFpEF. This may be explained by both methodological differences and how the population of interest was sampled in various studies. Direct Fick is the reference standard for measurement of cardiac output. Non-invasive methods to estimate CO reported in some studies are imprecise and introduce variability27 when used to indirectly estimate peripheral O2 extraction. In addition, HFpEF is a heterogeneous disease that may present subgroups of patients whom central and peripheral mechanisms have different weights in constraining exercise capacity28. Our sample may represent a mild phenotype of HFpEF whom peak systemic oxygen extraction is still intact; it is possible that patients with longstanding HFpEF develop secondary changes in O2 extraction and that this contributes importantly to the pathophysiology of chronic disease.
Chronotropic incompetence is an important correlate of peak VO2 in HFpEF5, 6, 29-31. In the current study, the severity of chronotropic impairment was similar to recent HFpEF trials30, 32 and it was associated with lower exercise capacity. The ELANDD trial observed a lowered exercise capacity due to reduced HR response related to nebivolol treatment in HFpEF patients33. The current comprehensive physiological assessment during exercise draws attention to the potentially adverse impact of β-blockers on exercise capacity in patients with HFpEF. The effect of withdrawal negative chronotropic drugs requires further study.
Importantly, there was no difference in chronotropic response between controls and the subset of HFpEF patients without β-blockade; these HFpEF patients, though, still had reduced peak VO2 due to lower peak CO and peak SVi. This predominant role of SV in exercise capacity was previously described in patients with LV diastolic dysfunction10, 34. Load dependent and independent mechanisms had been put forward to explain constrained SV augmentation in HFpEF such as higher LV elastance, impaired contractile reserve and afterload mismatch due to abnormal systemic vasodilation6, 7. Abnormalities in systemic vascular function influence SV by increasing LV afterload35. Consistent with previous studies5, 9, 12, we observed reduced decline in SVR and higher peak SVR in HFpEF patients. This has been ascribed to systemic endothelial and autonomic dysfunction6, 36.
In HFpEF, the RV faces elevated pulmonary artery impedance driven by increased pulmonary venous pressure. A subset of patients develop pulmonary vascular disease (arteriolar remodeling) with a further increase in impedance to RV ejection37. During exercise, RV afterload increases more steeply as the remodeled pulmonary vessels have limited capacity to accommodate increased CO38. Several findings support the presence of pulmonary vascular disease in ePVH patients. Resting PVR was higher in ePVH compared to controls despite similar CO, and more than a third had PVR>240 dyne.s.cm−5. The increased peak and isoflow PVR and the subgroup of ePVH patients with a PVR augmentation during exercise, rather than the normal decrease, further support the presence of an abnormal pulmonary circulation. We did not find an association between the extent of pulmonary vascular dysfunction and the degree of exercise capacity impairment. It is possible this is because we are detecting earlier disease in a younger patient population and that a relationship between depressed peak VO2 and blunted PVR fall during exercise would have been more apparent in an older population with established disease. The extent of pulmonary vascular disease of the studied patients might not be severe enough to impact RV-pulmonary artery coupling; the RV may be able to adapt to the increased PVR and maintain an adequate SV augmentation during exercise.
By directly measuring left heart filling pressures during upright exercise, our approach overcomes the limited sensitivity of the echocardiographic and circulating surrogate markers for detecting clinically relevant left heart diastolic dysfunction in patients with exertional intolerance39, 40.
Another methodological strength of our study lies on the simultaneous measurement of VO2 and CavO2diff. The former is particularly important because several non-cardiovascular factors (motivation, subjective dyspnea, fitness level, obesity) may determine different load intensity and confound the comparison of hemodynamics between controls and heart failure patients. The determination of the CavO2diff avoids the measurement errors of its indirect estimation from VO2 and non-invasive CO assessment.
This sample may represent an earlier and more mild phenotype of HFpEF, given that resting standard data yielded inconclusive diagnostic information leading to referral for iCPET. Therefore, the generalization to community-based HFpEF patients should be cautious. The cross-sectional design of our study precludes insight into time-varying pathophysiology in different stages of HFpEF. In addition, we can only speculate about the causal nature of the observed association between β-blockade and chronotropic incompetence. The use of multiple statistical hypothesis testing increases the chance of type 1 error. We did not measure inotropic reserve with exercise, and cannot comment of the role of this to limited peak SVi. Nevertheless, this does not affect the conclusion that in this group of patients there is a central limit to exercise and a markedly abnormal pulmonary vascular response to exercise while peripheral oxygen extraction is preserved.
Reduced CO due to chronotropic incompetence and decreased SV augmentation, rather than impaired peripheral O2 extraction, constrain aerobic capacity in ePVH patients. The association between chronotropic incompetence and β-blockers stresses the need of further research regarding the management of HR in HFpEF. There is a significant burden of pulmonary vascular disease in ePVH patients apparent both at rest and during exercise. Together, the current study findings uphold the multifactorial nature of exercise intolerance in HFpEF and emphasize the value of detailed exercise evaluation to discriminate between these distinct pathophysiological mechanisms in patients with unexplained exertional intolerance.
The authors would like to acknowledge David Yang for his technical expertise and Manyoo Agarwal for his helpful assistance.
FUNDING SOURCES: Mário Santos receives funds from HMSP-ICJ/0013/2012 grant from the Portuguese Foundation for Science and Technology. Amil M. Shah receives funds from grant 1K08HL116792-01A1 from the National Institutes of Health. Alexander R. Opotowsky receives support from the Dunlevie Family Fund.
DISCLOSURES: Amil M. Shah receives research support from Novartis. Alexander R. Opotowsky has received research support from Actelion and Merck.