Integration of Brain Tissue Saturation Monitoring in Cardiopulmonary Exercise Testing in Patients with Heart Failure

This protocol integrated near-infrared spectroscopy into conventional cardiopulmonary exercise testing to identify the involvement of the cerebral hemodynamic response in exercise intolerance in patients with heart failure.

Cerebral hypo-oxygenation during rest or exercise negatively impacts the exercise capacity of patients with heart failure with reduced ejection fraction (HF). However, in clinical cardiopulmonary exercise testing (CPET), cerebral hemodynamics is not assessed. NIRS is used to measure cerebral tissue oxygen saturation (SctO₂) in the frontal lobe. This method is reliable and valid and has been utilized in several studies. SctO₂ is lower during both rest and peak exercise in patients with HF than in healthy controls (66.3 ± 13.3% and 63.4 ± 13.8% vs. 73.1 ± 2.8% and 72 ± 3.2%). SctO₂ at rest is significantly linearly correlated with peak VO₂ (r = 0.602), oxygen uptake efficiency slope (r = 0.501), and brain natriuretic peptide (r = -0.492), all of which are recognized prognostic and disease severity markers, indicating its potential prognostic value. SctO₂ is determined mainly by end-tidal CO₂ pressure, mean arterial pressure, and hemoglobin in the HF population. This article demonstrates a protocol that integrates SctO₂ using NIRS into incremental CPET on a calibrated bicycle ergometer.

Cardiopulmonary exercise testing (CPET) has been applied in patients with heart failure with reduced ejection fraction (HF) for multiple aims, including the quantification of cardiopulmonary fitness, prognosis, diagnosing causes of exercise limitations, and exercise prescriptions^1,2,3. During testing, hemodynamic variables and data derived from automatic gas exchange are monitored and analyzed. Cerebral tissue oxygen saturation (SctO₂) monitoring has value for grading prognosis and disease severity^4,5.

Near-infrared spectroscopy (NIRS) uses infrared light to penetrate the skull and estimate brain tissue oxygenation continuously and non-invasively⁶. Since oxyhemoglobin and deoxyhemoglobin have different light absorption spectra and are the primary chromophores that absorb light, their concentrations can be measured using light transmission and absorption^6,7. However, background light absorbers also scatter light and may influence the measurement⁸. This study adopted a spatially resolved NIRS to measure SctO₂ from rest to peak exercise⁹. Four wavelengths were emitted to compensate for wavelength-dependent scattering losses and eliminate background interference, thus enhancing accuracy¹⁰.

SctO₂ represents the proportion of oxygen delivery vs. consumption in cerebral tissue. Cerebral desaturation is associated with disrupted cerebral blood flow (CBF), decreased arterial oxygen concentration, and increased cerebral tissue oxygen consumption¹¹. Other than cardiac output insufficiency, advanced HF causes cerebral hypoperfusion during exercise by indirectly inducing cerebral vasoconstriction via diminishing arterial partial pressure of carbon dioxide (PaCO₂) through hyperventilation¹².

The clinical significance of cerebral oxygenation in HF was revealed by Chen et al.⁴. First, SctO₂ was significantly decreased in the HF group compared with healthy controls. SctO₂ is not only diminished at rest but also declined further during exercise. It is not observed in the healthy group. Second, SctO_2rest and SctO_2peak were correlated with VO_2peak, brain natriuretic peptide (BNP), and oxygen uptake efficiency slope (OUES), all of which are established prognostic markers. Therefore, SctO_2rest and SctO_2peak are very likely to be prognostic and reflect disease severity in HF patients. Another study by Koike et al. suggested that the change in cerebral oxyhemoglobin measured at the forehead from rest to peak exercise was significantly lower in non-survivors compared to that in survivors of patients with coronary artery disease⁵. Hence, cerebral oxygenation may be employed to stratify the disease severity and prognosis of patients with HF.

The following protocol was approved by the ethics committee in Chang Gung Memorial Hospital, Linkou, Taiwan. The exercise test was carried out in an air-conditioned laboratory with an atmospheric temperature of 22-25 °C, pressure of 755 to 770 Torr, and relative humidity of 55-65%. Before each test, the gas analyzer was calibrated following the manufacturer's instructions using room air and a gas mixture of known concentration (FO₂: 0.12; FCO₂: 0.05; N₂ as balance). The turbine flow meter of the system was calibrated by the 2-point method with 0.2 L/s and 2 L/s by an automatically-pumping system.

1. Preparation: Placement of sensors and recorders

1. Clean the forehead twice with an alcohol pad to remove sweat and dirt from the skin.
2. Place NIRS sensors on the forehead bilaterally. Use a large sensor in which the distance between emitter and detector is 5 cm. The estimated measuring depth is 2.5 cm. Ensure that the sensors are securely attached.
3. Attach patches of electrocardiography to the anterior chest, bilateral acromioclavicular joints, and lower back.
4. Have the patient sit on the bicycle ergometer.
5. Place the armband of the sphygmomanometer.
6. Instruct the patient to wear the mask for the gas analysis. Make sure that no gas leaks through the brim of the mask.
7. Place the sensors of the pulse oximeter on the patient's ear lobe and index finger.

2. CPET and SctO₂ monitoring

1. Tell the patient to rest for at least 2 min to obtain a stable baseline value, including SctO₂ and respiratory exchange ratio.
2. Have the patient complete the warm-up stage at a work rate of 10 W for 1 min on the cycle ergometer.
3. Increase the rate by 10 W/min and ask the patient to pedal at around 60 rpm until failing to keep up with a cadence >50 rpm despite strong encouragement (symptom-limited exercise testing).
4. Average the SctO₂ value every second automatically from data scanned at the frequency of 100 Hz.
5. Measure the blood pressure every 2 min automatically by the sphygmomanometer.
6. Analyze the gas component breath by breath, including VO₂ and end-tidal carbon dioxide pressure (P_ETCO₂).
7. Have the patient complete the recovery stage at a work rate of 0 W for 2-6 min.

Thirty-four HF patients and 17 healthy controls were enrolled at Linkou Chang Gung Memorial Hospital, Taiwan. Each subject underwent cardiopulmonary exercise testing that incorporated SctO₂ monitoring by NIRS. Briefly, SctO₂ (rest; peak) values were significantly lower in the HF group (66.3 ± 13.3%; 63.4 ± 13.8%,) than in the control (73.1 ± 2.8%; 72 ± 3.2%) group (Figure 1). In the HF group, SctO₂ at rest (SctO_2rest) and peak SctO₂ (SctO_2peak) were linearly correlated with brain natriuretic peptide (BNP), VO_2peak, and OUES (r of -0.561 to 0.677, p < 0.001) (Figure 2). Notably, SctO_2rest was determined by the partial pressure of P_ETCO₂ at rest (P_ETCO_2rest), hemoglobin, and mean arterial pressure at rest (MAP_rest) (adjusted R = 0.681, p < 0.05 on stepwise linear regression) (Table 1). The main findings are illustrated in Figure 3.

Figure 1: Box plots of SctO_2rest and SctO_2peak in HF and control groups. Both SctO₂ values at rest and peak exercise were significantly lower in the HF group than in the control group. Adapted from Chen et al.⁴. * p < 0.05, HF vs. control, repeated measure ANOVA. # p < 0.05, SctO_2rest HF vs. control, paired t-test. † p < 0.05, SctO_2peak HF vs. control, paired t-test. SctO_2rest: cerebral tissue oxygen saturation at rest; SctO_2peak: cerebral tissue oxygen saturation at peak exercise. Please click here to view a larger version of this figure.

Figure 2: Scatter plots of SctO_2rest and SctO_2peak versus VO_2peak, BNP, and OUES. SctO_2rest (left panel) and SctO_2peak (right panel) values were linearly correlated with VO_2peak, BNP, and OUES. Adapted from Chen et al.⁴ SctO₂: cerebral tissue oxygen saturation; VO₂: oxygen consumption; BNP: brain natriuretic peptide; OUES: oxygen uptake efficiency slope. Please click here to view a larger version of this figure.

Figure 3: Illustration of the possible prognostic value of cerebral desaturation and its physiological basis in patients with HF. SctO₂, especially at peak exercise, was correlated with VO_2peak, BNP, and OUES. SctO_2rest was determined by P_ETCO_2rest, hemoglobin, and MAP_rest, while the main determinant of SctO_2peak was VCO_2peak. Please click here to view a larger version of this figure.

Table 1: Stepwise regression of SctO_2rest in the HF group. Adapted from Chen et al.⁴. SctO₂: cerebral tissue oxygen saturation. HF: heart failure.

Cerebral oxygenation monitored noninvasively and continuously by NIRS has been applied in various scenarios, including cardiovascular surgery¹³ and brain functional analyses such as those that estimate neural activity¹⁴. This protocol integrated NIRS into conventional CPET to identify the involvement of the cerebral hemodynamic response in exercise intolerance in patients with HF. It increases the value of exercise testing in determining prognosis and disease severity.

Cardiac dysfunction was considered the main cause of exercise intolerance in patients with HF¹⁵. Nonetheless, clinical studies demonstrated that inotropic or vasodilatory agents failed to increase exercise capacity¹⁶, and the association between resting cardiac function and peak oxygen consumption is weak¹⁷. Accordingly, cardiac dysfunction is not the sole cause of exercise intolerance in patients with HF.

The decreased cerebral perfusion and oxygenation during exercise have been demonstrated in patients whose cardiac output fail to increase normally^18,19, suggesting that the cerebral hypoperfusion is partially caused by the blunted cardiac output increase during exercise. Reduced frontal cortex oxygenation impaired the force-generating capacity of the peripheral working muscle, thus limit the exercise performance²⁰. Moreover, the suppressed cerebral hemodynamics during exercise is associated with ventilatory abnormality, which reduces the functional capacity of patients with HF²¹. Furthermore, SctO₂ is correlated with peak VO₂ and OUES as well as BNP, all of which are well-recognized markers for HF severity and prognosis⁴.

Exertional cerebral hypoperfusion²¹ is caused by cardiac output insufficiency as well as exertional hyperventilation, which reduces alveolar PCO₂ and the subsequent PaCO₂, a response that may further induce cerebral vasoconstriction during exercise^22,23,24. A previous study showed that PaCO₂ is positively linearly correlated with a CBF of 15-60 mmHg²⁵. In fact, it is the main physiologic determinant of SctO₂⁴. SctO₂ is also affected by hemoglobin and MAP⁴, influencing arterial oxygen concentration and cerebral perfusion, respectively^26,27. One may argue that anemia leads to increased mean optical path length and might influence the validity of SctO₂ measurement by NIRS. A previous study has already demonstrated that SctO₂ measured by phase-resolved spectroscopy did not alter significantly in response to the change of hemoglobin concentration in a cardiac surgery²⁸.

Despite the high reproducibility and validity of NIRS measurements in the resting state, the validity of this device in the HF population during exercise has not been established. Nonetheless, the different combinations of end-tidal O₂ and CO₂ were simulated in a previous validation study, which is in part similar to exercise status²⁹. The increase or decrease of skin blood flow at peak exercise in patients with HF might overestimate or underestimate the true value of cerebral oxygenation at the forehead^30,31. No matter what, the fact that low SctO₂ measured by NIRS on the forehead is a potentially negative prognostic factor based on the present result, has been established, except that the measured SctO₂ value not only represents cerebral oxygenation at frontal lobe, but also skin blood flow at forehead to a certain degree. Besides, extracranial melanin may absorb light and thus attenuate the signal, although SctO₂ was calculated from the concentration of oxy- and deoxyhemoglobin and is affected less by skin melanin⁸. Time-resolved spectroscopy - NIRS may solve the above problem to some extent. However, standard NIRS is rather easier for clinical application. Finally, a longitudinal study is required to confirm the prognostic value of SctO₂ in patients with HF.

The authors have nothing to disclose.

The patient who participated in exercise testing is deeply appreciated. This research was supported by National Science Council, Taiwan (NMRPG3G6231/2/3), Chang Gung Memorial Hospital (Grant No. CMRPG3G0601/2), and Healthy Aging Research Center, Chang Gung University and the Taiwan Ministry of Education's Higher Education Deep Plowing Program (Grant Numbers EMRPD1H0351 and EMRPD1H0551).