Suppression of cerebral hemodynamics is associated with reduced functional capacity in patients with heart failure

Abstract

This investigation elucidated the underlying mechanisms of functional impairments in patients with heart failure (HF) by simultaneously comparing cardiac-cerebral-muscle hemodynamic and ventilatory responses to exercise among HF patients with various functional capacities. One hundred one patients with HF [New York Heart Association HF functional class II (HF-II, n = 53) and functional class III (HF-III, n = 48) patients] and 71 normal subjects [older control (O-C, n = 39) and younger control (Y-C, n = 32) adults] performed an incremental exercise test using a bicycle ergometer. A recently developed noninvasive bioreactance device was adopted to measure cardiac hemodynamics, and near-infrared spectroscopy was employed to assess perfusions in the frontal cerebral lobe (Δ[THb]_FC) and vastus lateralis muscle (Δ[THb]_VL). The results demonstrated that the Y-C group had higher levels of cardiac output, Δ[THb]_FC, and Δ[THb]_VL during exercise than the O-C group. Moreover, these cardiac/peripheral hemodynamic responses to exercise in HF-III group were smaller than those in both HF-II and O-C groups. Although the change of cardiac output caused by exercise was normalized, the amounts of blood distributed to frontal cerebral lobe and vastus lateralis muscle in the HF-III group significantly declined during exercise. The HF-III patients had lower oxygen-uptake efficiency slopes (OUES) and greater V̇e-V̇o₂ slopes than the HF-II patients and age-matched controls. However, neither hemodynamic nor ventilatory response to exercise differed significantly between the HF-II and O-C groups. Cardiac output, Δ[THb]_FC, and Δ[THb]_VL during exercise were directly related to the OUES and V̇o_2peak and inversely related to the V̇e-V̇co₂ slope. Moreover, cardiac output or Δ[THb]_FC was an effect modifier, which modulated the correlation status between Δ[THb]_VL and V̇e-V̇co₂ slope. We concluded that the suppression of cerebral/muscle hemodynamics during exercise is associated with ventilatory abnormality, which reduces functional capacity in patients with HF.

reduced exercise capacity negatively affects the ability of patients with chronic heart failure (HF) to perform the activities required for daily life, further decreasing their independence and quality of life (27). Cardiac dysfunction is typically considered to be a primary factor for the impaired functional capacity in the HF population (10). However, a weak correlation between resting ventricular function and exercise tolerance (15) and the failure of inotropic and vasodilatory agents to improve exercise capacity (29) have been reported in clinical investigations. These findings suggest that cardiac dysfunction is not the only contributor to the development of exercise intolerance in patients with HF. Abnormal cerebral metabolism caused by hypoperfusion has been diagnosed using magnetic resonance spectroscopy in patients with advanced HF (21). Whether the impairment of cerebral hemodynamic response to exercise contributes to the decline of functional capacity in patients with HF remains unclear.

Cerebral hypoperfusion occurs during exercise in some cardiac patients in whom the cardiac output fails to increase normally (19, 20), suggesting that the decline in cerebral perfusion is partially caused by the blunted cardiac output increase during exercise. Additionally, ventilatory inefficiency during exercise is frequently diagnosed in patients with HF (6, 24). Exertional hyperventilation that is caused by HF reduces alveolar Pco₂ and the subsequent Pa_CO2 (6, 42), a response that may further induce cerebral vasoconstriction during exercise (3). Furthermore, reduced cerebral perfusion/oxygenation may result in a central limitation to exercise performance (4, 26, 32, 37). Accordingly, we hypothesize that the suppression of cerebral hemodynamics by abnormal perfusion or/and ventilation response(s) to exercise is associated with the decline of functional capacity in patients with HF.

Symptoms and functional exercise capacity are utilized to classify the severity of HF on the basis of the New York Heart Association (NYHA) functional classification (Fc) system and to judge the patient's response to treatment (29). Present treatment guidelines recommend exercise training in patients with HF in NYHA Fc II (HF-II) and Fc III (HF-III) (29). This study was the first to elucidate the underlying mechanisms of functional impairments in patients with HF by simultaneously comparing cardiac-cerebral-muscle hemodynamic and ventilatory responses to exercise among HF patients with various functional capacities. A novel, noninvasive bioreactance device (noninvasive continuous cardiac output monitoring system, NICOM) was applied to assess the cardiac hemodynamics (18, 23, 25), and near-infrared spectroscopy (NIRS) was employed simultaneously to monitor changes of perfusion/oxygenation in cerebral tissue (32) and skeletal muscle (37). Furthermore, whether these exercise-induced central (cardiac)/peripheral (cerebral and muscle) hemodynamic changes were correlated with indices of ventilatory efficiency, including the V̇e-V̇co₂ slope and the oxygen-uptake efficiency slope (OUES) (6), was determined to clarify the relationship between hemodynamic and ventilatory responses to exercise in HF patients.

MATERIALS AND METHODS

This study enrolled 101 patients with a diagnosis of HF from the Department of Cardiology, Chang Gung Memorial Hospital. HF was diagnosed if they had 1) a left ventricular ejection fraction (LVEF) >40% and were in NYHA Fc HF-II to -III despite optimal treatment for at least 12 mo according to the American Heart Association/American College of Cardiology guidelines, or 2) an LVEF >40% with episodes of acute pulmonary edema after excluding other noncardiogenic etiologies. According to NYHA Fc, these patients with HF were divided into Fc II (HF-II, n = 53) and Fc III (HF-III, n = 48) groups, with slight [≥4 to <7 metabolic equivalences (MET)] and marked limitations (≥1.5 to <4 MET) of physical activity, respectively (Table 1). Patients were excluded for the presence of atrial fibrillation or flutter, second- or third-degree heart block, history of life-threatening ventricular arrhythmias, recent unstable angina, myocardial infarction, coronary revascularization (<4 wk), uncontrolled diabetes mellitus, severe chronic obstructive pulmonary disease, symptomatic cerebrovascular disease within 12 mo, collagen vascular disease, alcohol or drug abuse within the prior 12 mo, or significant renal or hepatic disease. To avoid the confounding effects of physical activity on hemodynamic regulation, no patient had ever been submitted to cardiovascular rehabilitation. Because patients with HF-II and HF-III were middle-aged to elderly adults whose ages ranged from 50 to 75 years, age- and sex-matched older (O-C, n = 39) and sex-matched young (Y-C, n = 32) control groups were used to elucidate age-related hemodynamic regulation (Table 1). Both O-C and Y-C groups were carefully selected to recruit subjects who had not engaged in regular physical activity (i.e., exercise frequency ≤1 time per week, duration <20 min) for at least one year before the experiment. Moreover, the control subjects were required to be free of chronic pulmonary, cardiovascular, immune, and metabolic diseases. Subjects in the four groups were instructed to fast for at least 8 h and to refrain from exercise for at least 24 h before this study. All subjects arrived at the testing center at 9:00 AM to eliminate any possible diurnal effect. The study was performed according to the Helsinki declaration and was approved by the Institutional Review Board of Chang Gung Memorial Hospital, Taiwan. All subjects gave informed, written consent after the experimental procedures were explained.

Table 1.

Demographic and clinical characteristics

		HF-III	HF-II	O-C	Y-C
Anthropometrics/Clinical Characteristics
Sex	n (M/F)	48 (34/14)	53 (38/15)	39 (27/12)	32 (20/12)
Age	year	67.5 ± 1.5	66.5 ± 1.2	67.2 ± 1.0	24.2 ± 0.4‡
Height	cm	160.1 ± 0.8	162.8 ± 1.2	161.7 ± 0.8	168.4 ± 1.1‡
Weight	kg	63.3 ± 1.8	64.4 ± 1.1	62.4 ± 1.2	67.2 ± 1.5‡
Body mass index	kg/m2	24.7 ± 0.5	24.3 ± 0.4	23.9 ± 0.5	23.7 ± 0.4
SaO2	%	96.6 ± 0.3	96.4 ± 0.3	96.3 ± 0.3	97.6 ± 0.3
HR	beats/min	81 ± 1	80 ± 2	79 ± 1	75 ± 2
Systolic BP	mmHg	141 ± 3	142 ± 4	140 ± 3	125 ± 3‡
Diastolic BP	mmHg	80 ± 2	79 ± 3	80 ± 3	78 ± 2
VE	l/min	8.8 ± 0.3	9.1 ± 0.3	8.4 ± 0.4	7.5 ± 0.4
PETO2	mmHg	116 ± 2	115 ± 2	114 ± 1	115 ± 1
PETCO2	mmHg	32 ± 1	33 ± 1	35 ± 1	35 ± 1
Functional Capacity
MET	unit	3.46 ± 0.08	5.43 ± 0.11†	6.05 ± 0.17*	12.03 ± 0.4.3‡
Echocardiography
LVEF	%	38.1 ± 1.5	38.6 ± 1.1	ND	ND
> 40%	n (%)	28 (58)	30 (57)	ND	ND
≤ 40%	n (%)	20 (42)	23 (43)	ND	ND
Renal Functions
Creatinine	mg/dl	1.13 ± 0.06	1.03 ± 0.5	ND	ND
BUN	mg/dl	7.74 ± 0.38	7.27 ± 0.36	ND	ND
Medications
Digoxin	n (%)	6 (13)	5 (9)	0 (0)	0 (0)
β-blocker	n (%)	44 (92)	46 (87)	3 (8)	0 (0)
ACE/ARB	n (%)	46 (96)	49 (92)	2 (5)	0 (0)
Ca2+ channel blocker	n (%)	6 (13)	8 (15)	0 (0)	0 (0)
Diuretics	n (%)	32 (67)	34 (64)	3 (8)	0 (0)

Applicable values are means ± SE. male; F, female; SaO₂, arterial O₂ saturation; HR, heart rate; BP, blood pressure; V_E, minute ventilation; P_ETO₂ and P_ETCO_2, the end-tidal partial pressures of O₂ and CO₂; MET, metabolic equivalences; VLEF, left ventricle ejection fraction; BUN, blood urine nitrogen; ACE/ARB, angiotensin-converting enzyme inhibitor/angiotensin receptor blocker; ND, not determined. .

^*P < 0.05, the New York Heart Association heart failure functional class II (HF-II) vs. the functional class III (HF-III) group;

^†P < 0.05, the older control (O-C) group vs. the HF-III group;

^‡P < 0.05, the younger control (Y-C) group vs. the O-C group.

Cardiopulmonary exercise test.

Subjects performed a graded exercise on a bicycle ergometer (150P; Ergoselect, Lindenstrasse, Germany) to evaluate their aerobic fitness and hemodynamic function (41). The exercise test was carried out in an air-conditioned laboratory with an atmospheric temperature of 22–25°C, a barometric pressure of 755 to 770 Torr, and a relative humidity of 55–65%. The exercise protocol in the Y-C subjects comprised 2 min of unloaded pedaling followed by a continuous increase of work rate of 20 W every minute until exhaustion (progressive exercise to peak oxygen consumption, V̇o_2peak). The work rate of the O-C, HF-II, and HF-III subjects was increased by 10 W every minute. Minute ventilation (V̇e), the end-tidal partial pressures of O₂ (Pet_O2) and CO₂ (Pet_CO2), oxygen consumption (V̇o₂), and carbonic dioxide production (V̇co₂) were measured breath by breath using a computer-based system (MasterScreen CPX; Cardinal-Health, Hoechberg, Germany). Heart rate was determined from the R-R interval on a 12-lead electrocardiogram; blood pressure was measured by an automatic blood pressure system (Tango; SunTech Medical, Eynsham, UK), and arterial O₂ saturation was monitored by finger pulse-oximetry (model 9500; Nonin Onyx, Plymouth, MN). The V̇o_2peak was defined by the following criteria: 1) the level of V̇o₂ increased less than 2 ml/kg per minute over least 2 min, 2) heart rate exceeded 85% of its predicted maximal value, 3) the respiratory exchange ratio exceeded 1.15, or 4) some other symptom/sign limitations, according to the guidelines of American College of Sports Medicine for exercise testing (5). The value of ventilatory threshold was determined by two experienced, independent reviewers using the V-slope method and verified by ventilatory criteria as follows: 1) the V̇e/V̇o₂ ratio increased without a corresponding increase in the V̇e/V̇co₂ ratio, 2) Pet_O2 increased without a decrease in the Pet_CO2, or 3) departure from linearity for V̇e (5).

Ventilation and V̇co₂ responses, acquired from the initiation of exercise to the peak values, were used to calculate the V̇e-V̇co₂ slope using least-squares linear regression (y = mx + b, m = slope). The OUES was derived from the slope of a natural logarithm plot of V̇e vs. V̇o₂. As such, the OUES is an estimation of the efficiency of ventilation with respect to V̇o₂, with greater slopes indicating higher ventilatory efficiency (6, 7).

Cardiac hemodynamic measurements.

The NICOM (Cheetah Medical, Wilmington, DE) was employed to evaluate cardiac hemodynamic response to exercise, which analyzes the phase shift (ΔΦ) created by alternating electrical current across the subjects' chest, as described in our previous study (41). Four dual surface electrodes were placed on the back to establish electrical contact with the body and to avoid the interference of upper body motion to the electrical cables during exercise test (25). Stroke volume was estimated using the equation, Stroke volume = C·VET·dΦ/dtmax, where C is a constant of proportionality, and VET denotes the ventricular ejection time, as determined from the NICOM and electrocardiogram signals. The cardiac output, systemic vascular conductance, and arteriovenous O₂ difference were then calculated using the following equations: cardiac output = stroke volume·heart rate; systemic vascular conductance = cardiac output/mean arterial pressure; arteriovenous O₂ difference = V̇o₂/cardiac output.

Skeletal muscular and cerebral hemodynamic measurements.

Two pairs of NIR probes were attached to each subject to monitor the absorption of light across the left frontal cortex region and vastus lateralis muscle (Oxymon, Artinis, The Netherland) during exercise test, as described in our previous study (41). The Beer-Lambert law was used to calculate micromolar changes in tissue oxygenation (Δ[O₂Hb] and Δ[HHb]) using received optical densities from two NIR wavelengths of 780 and 850 nm. The differential path length factors of muscle and cerebral tissues were set in 4.95 and 5.93, respectively (13, 40). Total Hb concentration (Δ[THb]) was calculated as the sum of Δ[O₂Hb] and Δ[HHb] and used as an index of change in regional blood volume (31, 36). Data were recorded at 10 Hz and filtered with a Savitzky-Golay smoothing algorithm before analysis.

Statistical analysis.

Data were expressed as means ± SE. The statistical software package StatVeiw was employed to analyze the data. To elucidate the underlying mechanisms of functional impairments in patients with HF, the ventilatory and cardiac/cerebral/muscle hemodynamic parameters of the O-C, HF-II, and HF-III groups during exercise test were compared by repeated-measures ANOVA [3 groups × 5 time sample points (rest status and 25%, 50%, 75%, and 100% of V̇o_2peak)] followed by Bonferroni's post hoc test. Additionally, this study also adopted 2 (O-C and O-Y groups) × 5 (the time points of resting and various exercise intensities) repeated-measures ANOVA followed by Bonferroni's post hoc test to evaluate the effect of age on ventilatory and hemodynamic responses to exercise in normal individuals. Multiple-regression analysis was used to elucidate the relationship between ventilatory and hemodynamic variables during exercise test. The criterion for significance was P < 0.05.

RESULTS

The anthropometric characteristics of HF patients (HF-II and -III) did not differ significantly from those of the age-matched controls (O-C). The Y-C subjects had similar body mass indexes to those of the O-C subjects despite differences in height and weight. Although the HF-III group had a lower value of METs than the HF-III group, the two HF groups exhibited no significant difference in resting ventricular ejection fraction (Table 1).

Aerobic fitness and ventilatory response to exercise.

The Y-C group had a higher aerobic capacity than the O-C group, revealed by higher levels of work rate, V̇e, V̇o₂, and V̇co₂ at peak exercise (Table 2). The HF-III patients exhibited larger V̇e, V̇e/V̇co₂ ratios and lower Pet_CO2 (Table 1) at the same work rate (30 W), compared with the HF-II patients and age-matched controls. Moreover, the HF-III group also had smaller levels of work rate, V̇e, V̇o₂, and V̇co₂ at ventilatory threshold and peak exercise than both HF-II and O-C groups (Table 2).

Table 2.

Comparisons of cardiopulmonary fitness during exercise test in various groups

		HF-III			HF-II			O-C			Y-C
		SWR	Vth	Peak	SWR	Vth	Peak	SWR	Vth	Peak	SWR	Vth	Peak
Work-rate	Watt	30 ± 0	28 ± 2	55 ± 3	30 ± 0	51 ± 2†	106 ± 4†	30 ± 0	54 ± 2*	112 ± 5*	30 ± 0	105 ± 3‡	192 ± 4‡
Pulmonary
VE	l/min	26.8 ± 0.3	23.1 ± 0.8	33.2 ± 1.9	20.7 ± 0.3†	27.2 ± 1.1†	55.2 ± 2.1†	18.4 ± 0.5*	29.2 ± 1.0*	58.1 ± 2.1*	18.3 ± 0.4‡	48.2 ± 2.1‡	118.2 ± 3.1‡
VO2	ml/min/kg	9.7 ± 0.2	8.7 ± 0.2	12.1 ± 0.3	9.7 ± 0.1	12.0 ± 0.2†	20.2 ± 0.4†	10.2 ± 1.0*	12.8 ± 1.0*	22.3 ± 0.3*	10.3 ± 0.4‡	25.5 ± 1.0‡	42.1 ± 1.5‡
VCO2	ml/min/kg	9.8 ± 0.4	8.6 ± 0.2	13.3 ± 0.4	8.9 ± 0.1†	12.2 ± 0.2†	24.1 ± 0.5†	9.1 ± 0.2*	12.6 ± 0.8*	26.1 ± 1.0*	8.4 ± 0.3‡	25.2 ± 1.2‡	51.2 ± 1.6‡
VE/VO2	ratio	43.6 ± 1.2	41.9 ± 1.1	43.3 ± 1.5	36.1 ± 1.6†	35.3 ± 1.2†	42.3 ± 2.1	29.0 ± 1.8*	36.6 ± 2.1*	41.8 ± 1.1	26.4 ± 1.4‡	28.5 ± 1.0‡	41.5 ± 1.4
VE/VCO2	ratio	43.2 ± 1.5	42.4 ± 1.2	40.1 ± 1.4	33.1 ± 1.8†	34.6 ± 1.7†	35.8 ± 1.5†	29.5 ± 1.4*	37.1 ± 1.9*	35.7 ± 1.9*	32.4 ± 1.2‡	28.1 ± 1.1‡	34.5 ± 1.5‡
PETO2	mmHg	114 ± 1	113 ± 1	117 ± 1	112 ± 1	110 ± 1	118 ± 2	109 ± 1	113 ± 1	118 ± 2	112 ± 1	114 ± 1	117 ± 1
PETCO2	mmHg	32 ± 1	33 ± 1	32 ± 1	35 ± 1†	37 ± 1†	36 ± 1†	40 ± 1*	41 ± 1*	37 ± 1*	41 ± 1	43 ± 1	36 ± 1
Cardiovascular
HR	beats/min	110 ± 2	102 ± 2	115 ± 2	102 ± 2†	108 ± 2†	146 ± 2†	99 ± 2*	109 ± 2*	153 ± 2*	97 ± 2‡	140 ± 2‡	195 ± 2‡
SV	ml	58 ± 2	61 ± 2	58 ± 3	65 ± 4†	68 ± 2†	78 ± 2†	75 ± 3*	66 ± 2	72 ± 2*	94 ± 2‡	99 ± 3‡	81 ± 2‡
CO	l/min	6.4 ± 0.3	6.0 ± 0.2	7.6 ± 0.3	6.5 ± 0.4	7.2 ± 0.3†	11.0 ± 0.2†	7.3 ± 0.3*	7.4 ± 0.3*	12.1 ± 0.3*	9.2 ± 0.3‡	13.5 ± 0.4‡	14.9 ± 0.4‡
SVC	ml/min/mmHg	62.7 ± 1.8	60.2 ± 1.8	77.4 ± 4.3	66.3 ± 2.1†	70.7 ± 2.1†	89.2 ± 4.2†	73.2 ± 2.7*	71.2 ± 2.3*	93.1 ± 2.1*	95.8 ± 2.5‡	121 ± 2.3‡	130.8 ± 4.9‡
MAP	mmHg	102 ± 3	100 ± 3	104 ± 3	98 ± 2	101 ± 2	118 ± 2†	100 ± 2	103 ± 2	119 ± 3*	96 ± 3	108 ± 3	117 ± 3
SaO2	%	96.5 ± 0.2	96.2 ± 0.2	96.0 ± 0.3	96.7 ± 0.2	96.6 ± 0.5	96.6 ± 0.4	96.6 ± 0.2	96.5 ± 0.2	96.3 ± 0.3	97.5 ± 0.3	97.4 ± 0.4	97.2 ± 0.2
Da-vO2	ml/dl	9.6 ± 0.4	8.9 ± 0.3	11.0 ± 0.4	9.6 ± 0.5	10.5 ± 0.6†	17.1 ± 0.4†	8.9 ± 0.6	11.2 ± 0.5*	18.6 ± 0.3*	7.5 ± 0.3‡	13.6 ± 0.8‡	21.9 ± 1.2‡

Values are means ± SE. SWR, same work-rate; V_th, ventilatory threshold; Peak, peak exercise; VO₂, O₂ consumption; VCO₂, CO₂ production; SV, stroke volume; CO, cardiac output; SVC, systemic vascular conductance; MAP, mean arterial pressure; Da-VO₂, arteriovenous O₂ difference.

^*P < 0.05, the HF-II vs. the HF-III group;

^†P < 0.05, the O-C group vs. the HF-III group;

^‡P < 0.05, the Y-C group versus the O-C group.

The value of V̇o_2peak was positively correlated with the OUES (Fig. 1A) and was negatively correlated with the V̇e-V̇co₂ slope (Fig. 1B) in the HF-III, HF-II, and O-C subjects. The normal young subjects exhibited higher OUES levels (Fig. 1C) and lower V̇e-V̇co₂ slopes (Fig. 1D) than the normal older subjects. Furthermore, the HF-III group reflected a lower OUES level and a steeper V̇e-V̇co₂ slope than the HF-II and O-C groups (Fig. 1, C and D). However, neither the aerobic capacity nor ventilatory response to exercise differed significantly between the HF-II and O-C groups (Table 2 and Fig. 1, A and B).

Fig. 1.

The relationships between V̇o_2peak and the oxygen-uptake efficiency slopes (OUES) (A) and the V̇e-V̇co₂ slope (B), as well as, comparisons of the OUES (C) and V̇e-V̇co₂ slope (D) among various groups. Values are means ± SE. *P < 0.05, the New York Heart Association heart failure functional class II (HF-II) vs. the functional class III (HF-III) group; +P < 0.05, the older control (O-C) group vs. the HF-III group; #P < 0.05, the younger control (Y-C) group vs. the O-C group.

Cardiovascular hemodynamics.

At ventilatory threshold and peak exercise, the Y-C group had higher levels of heart rate, stroke volume, cardiac output, mean arterial pressure, systemic vascular conductance, and arteriovenous O₂ difference than the O-C group (Table 2). Conversely, these cardiovascular responses to the two exercise statuses in HF-III group were smaller than those in both HF-II and O-C groups (Table 2). Moreover, the HF-III patients also exhibited higher heart rate, lower stroke volume, and smaller systemic vascular conductance during exercise of 30 W, compared with the HF-II patients and age-matched controls. The analysis of the differences in cardiovascular responses of the various groups to relative exercise intensities revealed that the Y-C group had higher levels of cardiac output (Fig. 2A), systemic vascular conductance (Fig. 2B), and arteriovenous O₂ difference (Fig. 2C) at 25% to 100% of V̇o_2peak than the O-C groups. During exercise from 50–100% of V̇o_2peak, the HF-III group exhibited lower levels of cardiac output, systemic vascular conductance, and arteriovenous O₂ difference than the O-C and HF-II groups (Figs. 2, A–C). However, the cardiovascular response to exercise did not significantly differ between the O-C and HF-II groups (Table 2 and Fig. 2, A–C).

Fig. 2.

Comparisons of cardiac output (CO) (A), systemic vascular conductance (SVC) (B), and arteriovenous O₂ difference (D_a-vO₂) (C) during exercise among various groups. Values are means ± SE. *P < 0.05, the HF-II vs. the HF-III group; +P < 0.05, the O-C group vs. the HF-III group; #P < 0.05, the Y-C group vs. the O-C group.

Cerebral hemodynamics.

During exercise test, perfusion (Δ[THb]_FC) (Fig. 3A) and oxygenation (Δ[O₂Hb]_FC) (Fig. 3B) in the frontal cerebral lobe significantly increased from 50 and 100% of V̇o_2peak in the Y-C group (P < 0.001) and the two O-C/HF-II groups (P < 0.01), whereas the frontal cerebral hemodynamic variables in the HF-III group modestly decreased at peak exercise (P < 0.05). The extent of cardiac output change was significantly positively related to the level of Δ[THb]_FC at peak exercise (Fig. 4A). When the cardiac output response to exercise was normalized, the level of blood distributed to the frontal cerebral lobe (Δ[THb]_FC/Δcardiac output) in the HF-III group significantly declined during incremental exercise despite the fact that Δ[THb]_FC/Δcardiac output increased during exercise for other groups (Fig. 4C). Furthermore, the O-C group was less effective than the Y-C group in terms of the perfusion of the frontal cerebral lobe and the distribution of blood from the heart to the frontal cerebral lobe during exercise (Fig. 4C). However, the hemodynamic responses of the frontal cerebral lobe to exercise did not significantly differ between the HF-II group and the O-C group (Fig. 4C).

Fig. 3.

Comparisons of perfusion and oxygenation in the frontal cerebral lobe [Δ[THb]_FC (A) and Δ[O₂Hb]_FC (B)] and vastus lateralis muscle [Δ[THb]_VL (C) and Δ[O₂Hb]_VL (D)] during exercise among various groups. Values are mean ± SE. *P < 0.05, the HF-II vs. the HF-III group; +P < 0.05, the O-C group vs. the HF-III group; #P < 0.05, the Y-C group vs. the O-C group.

Fig. 4.

The relationships between the changes of CO and the perfusions in the frontal cerebral lobe (Δ[THb]_FC) (A) and vastus lateralis muscle (Δ[THb]_VL) (B) at peak exercise, as well as, comparisons of Δ[THb]_FC/CO (C) and Δ[THb]_VL/CO (D) during exercise among various groups. Values are means ± SE. *P < 0.05, the HF-II vs. the HF-III group; +P < 0.05, the O-C group vs. the HF-III group; #P < 0.05, the Y-C group vs. the O-C group.

Skeletal muscle hemodynamics.

Vastus lateralis muscle perfusion (Δ[THb]_VL) (Fig. 3C) and oxygenation (Δ[O₂Hb]_VL) (Fig. 3D) in the Y-C group increased in the period of exercise from 50 to 100% of V̇o_2peak (P < 0.001). Conversely, both HF-II and -III groups exhibited decreases in Δ[THb]_VL (Fig. 3C) and Δ[O₂Hb]_VL (Fig. 3D) from 25 to 100% of V̇o_2peak (P < 0.05). At peak exercise, the extent of cardiac output change was significantly positively correlated with the level of Δ[THb]_VL (Fig. 4B). As adjusting the change of cardiac output caused by exercise, the amount of blood distributed to the vastus lateralis muscle (Δ[THb]_VL/Δcardiac output) in the HF-III group significantly fell, whereas exercise did not alter Δ[THb]_VL/Δcardiac output in the Y-C, O-C, and HF-II groups (Fig. 4D).

Relationship between ventilatory and hemodynamic responses to exercise.

Pearson correlation coefficients of the main ventilatory and hemodynamic variables of the O-C, HF-II, and HF-III subjects were presented in Table 3. In the perfusion variables, the levels of cardiac output, Δ[THb]_FC, and Δ[THb]_VL at peak exercise were positively correlated with the values of V̇o_2peak (Fig. 5, A–C) and OUES (Fig. 5, D–F) and were negatively correlated with the V̇e-V̇co₂ slope (Fig. 5, G–I). The significant correlation also existed between cerebral/muscle oxygenation and ventilatory responses to exercise; i.e., arteriovenous O₂ difference as well as Δ[O₂Hb]_FC and Δ[O₂Hb]_VL at peak exercise were directly related to the levels of V̇o_2peak (Fig. 6, A–C) and OUES (Fig. 6, D–F), and inversely related to the V̇e-V̇co₂ slope (Fig. 6, G–I).

Table 3.

Correlation coefficients of the ventilatory and hemodynamics variables of the O-C, HF-II, and HF-III subjects

	VO2peak	OUES	VE-VCO2 slope	CO	Δ[THb]FC	Δ[THb]VL
VO2peak	—	0.840*	−0.595*	0.729*	0.725*	0.618*
OUES	0.840*	—	−0.545*	0.691*	0.648*	0.540*
VE-VCO2 Slope	−0.596*	−0.545*	—	−0.485*	−0.570*	−0.374*
CO	0.729*	0.691*	−0.485*	—	0.520*	0.476*
Δ[THb]FC	0.725*	0.648*	−0.570*	0.520*	—	0.532*
Δ[THb]VL	0.618*	0.540*	−0.374*	0.476*	0532*	—

VO_2peak, peak O₂ consumption; OUES, oxygen-uptake efficiency slopes; V_E-VCO₂ Slope, ventilation-CO₂ production slope; Δ[THb]_FC, frontal cerebral lobe perfusion; Δ[THb]_VL, vastus lateralis muscle perfusion.

^*Significantly correlated at α<0.5 level.

Fig. 5.

The relationships between peak CO and frontal cerebral (Δ[THb]_FC)/vastus lateralis (Δ[THb]_VL) perfusions and indices of ventilatory efficiency, including the V̇o_2peak (A–C), V̇e-V̇co₂slope (D–F), and OUES (G–I) in the O-C, HF-II, and HF-III subjects.

Fig. 6.

The relationships between peak arteriovenous O₂ difference (D_a-vO₂) and frontal cerebral (Δ[O₂Hb]_FC)/vastus lateralis (Δ[O₂Hb]_VL) oxygenations and indices of ventilatory efficiency, including the V̇o_2peak (A–C), V̇e-V̇co₂ slope (D–F), and OUES (G–I) in the O-C, HF-II, and HF-III subjects.

The final multiple-regression model for the O-C, HF-II, and HF-III subjects is presented in Table 4. The assumptions for constructing the model are that the reduced ventilatory efficiency is associated with the suppression of hemodynamic response to exercise in these subjects. The results revealed that variables cardiac output, Δ[THb]_FC, and Δ[THb]_VL were significantly correlated with V̇o_2peak, OUES, or V̇e-V̇co₂ slope in the univariate analysis. Moreover, after control for other factors, variables cardiac output, Δ[THb]_FC, and Δ[THb]_VL were also significantly associated with V̇o_2peak and OUES. This model explained 71.8% and 59.8% of the variations of V̇o_2peak and OUES, respectively. However, only variables cardiac output and Δ[THb]_FC were significantly associated with the V̇e-V̇co₂ slope. Variable cardiac output or Δ[THb]_FC was an effect modifier (or confounding factor), which modulated the significant correlation status between Δ[THb]_VL and V̇e-V̇co₂ slope. The adjusted R² of the regression model was 0.360.

Table 4.

The multiple-regression model of ventilatory vs. and hemodynamics variables in the O-C, HF-II, and HF-III subjects

	Coefficients	t-Value	P Value	Adjusted R2
VO2peak vs. Hemodynamics Variables
Intercept	9.195	9.784	<0.0001	0.718
CO	0.735	7.698	<0.0001
Δ[THb]FC	0.555	6.868	<0.0001
Δ[THb]VL	0.251	3.191	0.0030
OUES vs. Hemodynamics Variables
Intercept	319.2	8.960	<0.0001	0.598
CO	24.28	6.729	<0.0001
Δ[THb]FC	15.06	4.928	<0.0001
Δ[THb]VL	5.746	2.240	0.0268
VE-VCO2 Slope vs. Hemodynamics Variables
Intercept	39.06	26.12	<0.0001	0.360
CO	−0.456	−3.012	0.0031
Δ[THb]FC	−0.628	−4.898	<0.0001
Δ[THb]VL	−0.038	−0.352	0.7251

The reliabilities of ventilatory and hemodynamic responses to exercise.

To evaluate the reliabilities of ventilatory and hemodynamic responses to exercise, 9 elderly subjects with/without HF (3 HF-III, 3 HF-II, and 3 O-C subjects; age = 65.2 ± 2.3 yr; sex = 5 males and 4 females) were tested twice at 1-day intervals. The results concerning the cardiac output, Δ[THb]_FC, and Δ[THb]_VL at peak exercise were highly reproducible day to day, and the single-measure intraclass corrections (ICCs) were 0.901, 0.872, and 0.813, respectively. Moreover, the ICCs of the V̇o_2peak, OUES, and the V̇e-V̇co₂ slope were 0.946, 0.926, and 0.903. Additionally, the results of ventilatory and hemodynamic responses to exercise in young subjects (n = 6; age = 23.2 ± 2.1 yr; sex = 3 males and 3 females) had also high reproducibility, and ICC were from 0.827 to 0.925 for the test-retest reliability.

DISCUSSION

Recently, preclinical and clinical data have established the feasibility of using blood flow-related phase shifts of transthoracic electric signals to monitor cardiac output continuously and noninvasively (18, 25, 35). Additionally, the accuracy, precision, and responsiveness of this bioreactance-based device have been verified by thermodilution method, which is the gold standard for cardiac output determination in the clinical setting (18, 23, 35). This study was the first to integrate novel bioreactance-based measurement, NIRS, and automatic gas analysis to identify the involvement of ventilatory and cardiac-cerebral-muscle hemodynamic responses to exercise in functional impairment in patients with HF. The present investigation adopted this integrated system to demonstrate a decline in aerobic capacity with aging, which was caused mainly by the decreases in O₂ delivery to, and in its utilization by, the tissue, as revealed by the lower exercise cardiac output, systemic vascular conductance, and arteriovenous O₂ difference levels in the O-C group than in the Y-C group. Previous investigations showed that age declined cardiovascular tolerance to oxidative stress (1) and reduced myocardial contractile response to β-adrenoceptor stimulation (14), which are associated with development of cardiovascular disorders with aging. Although no significant differences between these O₂ delivery-to-utilization variables of the O-C and HF-II groups existed in this study, the HF-III group exhibited lower levels of cardiac output, systemic vascular conductance, and arteriovenous O₂ difference during exercise than did either the O-C or the HF-II groups. Therefore, the severely impaired cardiovascular response to exercise in patients with HF may be associated with the accelerated decline in aerobic fitness with aging. A clinical investigation has reported that the severity of functional impairment is an effective predictor of mortality in elderly patients with HF; i.e., mortality increases with frailty in these patients (9).

In addition to inadequate cardiac output response to exercise, the HF-III patients also exhibited lower Δ[THb]_FC and Δ[THb]_VL during exercise than the HF-II patients and age-matched controls did. As the change of cardiac output caused by exercise adjusted further, the extent of blood distributed to frontal cerebral lobe (Δ[THb]_FC/Δcardiac output) or vastus lateralis muscle (Δ[THb]_VL/Δcardiac output) in the HF-III group substantially reduced during exercise test. The results may reflect a phenomenon of exercise-induced cerebral/muscle vasoconstriction. Moreover, the suppressed cerebral/muscle hemodynamic responses to exercise in the HF-III patients were accompanied by the reductions in cerebral/muscle oxygenation (revealing decreases in Δ[O₂Hb]_FC and Δ[O₂Hb]_VL during exercise). These findings further verify that the HF patients with severely impaired physical fitness frequently suffer from cerebral/muscle hemodynamic malfunctions (10, 19, 20).

Hyperpena upon exertion is a hallmark of HF, and the V̇e-V̇co₂ slope (6) or the V̇e/V̇co₂ ratio (24) during exercise increases with the severity of HF. This investigation also observed that the HF-III group had higher V̇e-V̇co₂ slope and V̇e/V̇co₂ ratio during exercise, compared with those of the HF-II and O-C groups. Moreover, peak cardiac output was inversely related to V̇e-V̇co₂ slope in these tested subjects. A decline in cardiac output influences both left- and right-sided circulation in HF, reducing pulmonary perfusion and CO₂ exchange, thus increasing the ratio of V̇e to V̇co₂ (24, 43). Adachi et al. (2) identified a relationship between an elevated V̇e-V̇co₂ slope and depressed nitric oxide production in patients with HF, further supporting the mechanistic importance of ventilation-perfusion mismatching, in this instance, by blunted vasodilation in pulmonary vasculature during exercise (2). Additionally, a heightened sensitivity to central/peripheral chemoreceptors and muscle egoreceptors appear to cause the abnormal ventilatory response to exercise in patients with HF (28, 30, 31). These results suggest that ventilation-perfusion mismatching is responsible for intolerance of exercise in HF.

Abnormal hyperventilation reduces Pa_CO2 in patients with advanced HF (6). Cerebral blood flow is known to be positively correlated with Pa_CO2; a fall in Pa_CO2 level results in cerebral hypoperfusion (17). In the present study, the HF-III group had a lower Pet_CO2 level and a higher V̇e/V̇co₂ ratio at ventilatory threshold and therefore a lower value of Δ[THb]_FC at peak exercise than those of the HF-II and O-C groups. On the basis of the steep V̇e-V̇co₂ slope, high V̇e/V̇co₂ ratio, and low Pet_CO2 level during exercise, the HF-III patients may rapidly enter a state of lactic acidosis and further develop pulmonary congestion in early stage of exercise, which causes hyperventilation and subsequently reduces the Pa_CO2 level, eventually leading to cerebral hypoperfusion.

The redistribution of blood flow caused by the blunted increase in cardiac output during exercise may also reduce blood flow to cerebral tissues in HF-III patients. Recent clinical investigations have found that patients with left ventricular dysfunction suffer from cerebral hypoperfusion (19, 20). Results obtained from this study showed that the cardiac output level was closely related to cerebral perfusion and oxygenation at peak exercise. We posit that the advanced HF causes cerebral hypoperfusion during exercise, possibly by 1) directly blunting the increase in cardiac output associated with exercise and 2) indirectly inducing cerebral vasoconstriction through lowering Pa_CO2 by hyperventilation. A reduction of frontal cortex oxygenation reportedly reduces the force-generating capacity of a muscle, critically limiting exercise performance (4). Accordingly, the deterioration of functional capacity in patients with HF III is, at least partially, attributable to the depression of central motor drive by exertional cerebral hypoperfusion.

This investigation demonstrated that the levels of Δ[THb]_VL and Δ[O₂Hb]_VL in the contracting muscles of the normal and HF subjects affected their ability to perform incremental exercise. During exercise, the extent of cardiac output change was significantly positively related to the level of Δ[THb]_VL. Conversely, the V̇e/V̇co₂ slope also had a modest negatively association with the change of the muscle perfusion during exercise. We further found that cardiac output or cerebral perfusion was an effect-confounding factor, which influenced the significant correlation status between muscle perfusion and V̇e-V̇co₂ slope. These results suggest that the reduction in the amount of blood distributed to contracting muscle results in abnormal ventilatory response to exercise, which is associated with cardiac/cerebral hemodynamic disturbance. Dempsey et al. (12) have indicated that the reduction of O₂ transport caused by desaturation and respiratory muscle fatigue may accelerate the development of peripheral muscle fatigue, further resulting in limitation of exercise (12).

During exercise, the decrease in Δ[THb]_VL and Δ[O₂Hb]_VL below their rest levels that were identified in the HF-III patients may evoke sympathetic and neurohumoral overdrives via exaggerated metaboreflex activity (11, 36). Additionally, hyperventilation and lower Pet_CO2 in patients with advanced HF contributes to partial vasoconstriction via sensitized carotid chemoreceptor (36). This reflex overactivity results in vasoconstriction and a consequent increase in total peripheral resistance during exercise, which may contribute to the intolerance of these patients with HF-III to exercise (11, 36).

Limitations of the study and technical considerations.

NIRS measurements have been closely correlated with intracellular oxygen tension in muscle and venous oxygen saturation in both muscle and cerebral tissue (16, 38). Moreover, this device has been applied to measure regional blood volume and oxygen delivery for utilization by skeletal muscles in patients with chronic HF (34). Despite high reproducibility in NIRS measurements, the validity of this device in the HF population has been not established. Additionally, the relative contribution of the flow of blood in skin to NIRS tissue signals has been questioned because NIRS probes are generally placed on the surface of the skin. Skin blood flow has been estimated to account for 15∼23% of the attenuation of NIR light (8), and therefore underlying tissues are the primary determinants of the resulting NIRS measurements. This study used skin-fold measurement to calculate thickness of skin and adipose, and no significant differences in the thicknesses of the tested surface areas were observed among the various groups (data not shown). Accordingly, the skin blood flow detected by NIRS measurements may not alter the conclusions in this investigation.

It is well known that chronic HF augments sympathetic-mediated vasoconstriction and impairs endothelial and mitochondrial functions, which may cause the abnormalities of perfusion and oxygenation in peripheral tissues (10). This study only observed the global phenomenon of cerebral/muscle perfusion and oxygenation during exercise, and the underlying cellular/molecular mechanisms of the impaired hemodynamic response to exercise in HF need to be further explored.

No direct measurement of cardiac output was made with which to compare the NICOM device in patients with HF despite the high reproducibly of the bioreactance-based measurement in a multicenter investigation (23). However, Maurer et al. (23) had evaluated a cross validation between NICOM and an inert gas rebreathing technique by the Innocor device. Their results clearly showed a good correlation between the measures of the two devices (23). Additionally, this study only evaluated the value of Pet_CO2 during exercise but did not directly measure cerebral Pa_CO2 response to exercise. Therefore, the direct evaluation of the cardiac output and Pa_CO2 responses to exercise using gold-standard measurements requires further work.

Conclusions.

Although abnormal cerebral metabolism at rest has been diagnosed in patients with advanced HF (21), this investigation is the first to demonstrate clearly that attenuated cerebrovascular response to exercise contributes to the decline of functional capacity in the HF patients. Reduced cerebral perfusion/oxygenation caused by exercise in patients with HF-III may result from blunted cardiac output and heightened ventilatory responses, which may substantially limit exercise performance. Furthermore, the suppressions in muscle perfusion/oxygenation partially contribute to exercise intolerance in patients with HF, as governed by reactions that depend on the severity of cardiac hemodynamic disturbance during exercise. These findings suggest that the O₂ delivery-to-utilization mismatching that is caused by abnormal central and peripheral perfusion/oxygenation responses to exercise is associated with reduced functional capacity in patients with HF.

GRANTS

This work was supported by the National Science Council of Taiwan (grant number NSC 96-2314-B-182-001) and Chang Gung University Research Program (grant number CMRPG 280241).

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.