Cardiopulmonary Exercise Testing in Patients With Left Bundle Branch Block and Preserved Ejection Fraction

Background: Left bundle branch block (LBBB) has prognostic significance in patients with congestive heart failure. However, its influence is not well established in patients with preserved systolic ventricular function. Objective: To evaluate the implications of LBBB presence in the cardiovascular performance of patients with preserved left ventricular systolic function (LVEF). Methods: 26 LBBB patients (61.3 ± 8.2 years of age) and 23 healthy individuals (58 ± 6.8 years of age) with LVEF > 0.5 underwent cardiopulmonary exercise testing (CPET). Results: CPET analysis revealed: peak oxygen consumption (VO 2 ) predicted in the LBBB group was 87.2 ± 15.0% versus 105.0 ± 15.6% (p < 0.0001); peak oxygen pulse predicted in LBBB group was 98.6 ± 18.6% vs 109.9 ± 13.5% (p = 0.02); VO 2 predicted anaerobic threshold in LBBB group was 67.9 ± 13.6% vs 70.2 ± 12.8% (p = 0.55); ΔVO 2 /Δload in the LBBB group was 15.5 ± 5.51 versus 20.7 ± 7.3 ml.min -1 .watts -1 (p = 0.006); ventilation / carbon dioxide production (VE/VCO 2 slope) in LBBB group was 29.8 ± 2.9 versus 26.2 ± 2.9 (p = 0.0001) and VO 2 recovery time in the LBBB group was 85.2 ± 11.8 vs. 71.5 ± 11.0 seconds (p = 0.0001). LBBB was an independent marker for VE/VCO 2 slope increase. Conclusion: LBBB presence in individuals with preserved LVEF did not affect cardiovascular performance, but there was an increase of the VE/VCO 2 slope in comparison to the control group. (Int J Cardiovasc Sci. 2017;30(1):11-19)


Introduction
The isolated presence of LBBB, regardless of cardiopathies, seems to be a slow progression marker of degenerative cardiac diseases, ischemic or not, affecting the myocardial conduction system and contractile performance. 1 LBBB can induce cardiomyopathy since individuals with this condition, when submitted to cardiac resynchronization therapy, may present left ventricular reverse remodelling (LV). 2,3 Thus, it should not be considered a mere electrocardiographic finding, but a "cardiac clinical entity", as suggested by Kumar et al. 4 Despite reports of LV structural alterations caused by LBBB, there is, in literature, a lack of predicting factors in the reduction of cardiovascular performance in these patients. CPET is a non-invasive method used for the diagnosis and prognostic stratification of congestive heart failure (CHF) patients, through the analysis of gases expired during stress tests. The main variables assessed with this methodology are: peak oxygen consumption (VO 2 peak), VO 2 at the anaerobic threshold, oxygen pulse (PO 2 ), ΔVO 2 /Δload, relation between ventilation and carbon dioxide production (VE/VCO 2 Cardiopulmonary testing and left bundle branch block Int J Cardiovasc Sci. 2017;30(1): [11][12][13][14][15][16][17][18][19] Original Article slope) and VO 2 recovery kinetics. On the other hand, it has been documented that LVEF assessed through echocardiogram does not show a satisfactory relation to VO 2 , 5,6 and thus does not constitute a good predictor of functional capacity.
We can then hypothesize that CPET could show, in asymptomatic LBBB patients, cardiovascular behavior alterations that precede LVEF compromise. The present research investigated if isolated LBBB in patients with normal LVEF interferes in the cardiovascular performance evaluated through CPET.

Design and study population
This is an observational cross-section analytical study, in which we evaluated 26 LBBB patients (LBBB group) and 23 individuals without LBBB (control group). LBBB was defined according to the following electrocardiographic criteria: enlarged QRS (with duration ≥ 120 ms); wide S waves in V1 and V2 with the absence of R waves, or its occurrence in an embryonic form; and R waves predominant in leads D1, AVL, V5, and V6. 7 All volunteers presented LVEF, obtained through transthoracic echocardiogram, superior to 0.50 and underwent CPET. Myocardial ischemia was excluded through stress echocardiogram. All exams were performed in the sector of graphic methods of a hospital specialized in cardiology, which has a level 3 accreditation (IQG -Health Accreditation Services).

Exclusion criteria
Exclusion criteria included patients with: LVEF < 0.50 calculated by the Simpson method, previous coronary artery disease (CAD), moderate to severe valvulopathy, cardiac arrhythmia that hindered PO 2 analysis, pulmonary disease, and anemia (hemoglobin < 10 mg/dl).

Cardiopulmonary exercise test protocol
Analysis of minute ventilation (VE), VO 2 , and production of carbon dioxide (VCO 2 ) was performed every 10 seconds through the gas analyser Cortex Metalyser 3B (Micromed) connected to the computer, equipped with the software Elite. We used: Micromed digital electrocardiograph to register and analyse the electrocardiogram during effort and an Inbrasport treadmill, model Super ATL. Negative chronotropic medications were suspended three days before CPET.
Participants were encouraged to exercise to exhaustion, according to the ramp protocol, which is characterized by a duration between 8 and 12 minutes, with constant gradual increase in speed and inclination. The test was considered maximum when RER > 1.05 was reached. The test was interrupted according to criteria established by the III Brazilian Society of Cardiology Guidelines on Exercise Tests. 8 The following CPET variables were evaluated: a. Peak VO 2 is the highest VO 2 value, reached in the last 30 seconds of effort and expressed in ml.kg -1 .min -1 . The predicted peak VO 2 was calculated with base on age, gender, weight, and level of physical activity using the Wasserman equation. 9, 10 We evaluated the percentage reached by the patient of the predicted VO 2 .
b. VO 2 at anaerobic threshold was determined by the "V slope method". If that was not possible, it was performed through the analysis of the plot graph of ventilator equivalents. We also assessed the percentage of VO 2 predicted at anaerobic threshold. 8,9,11,12 c. Peak VO 2 was the highest value reached in the last 30 seconds and was expressed in ml/beats. We analysed the percentage of the predicted peak PO 2 , obtained by the division of the predicted VO 2 max by the maximum heart rate predicted by age. 8,9,11,12 d. The relation ΔVO 2 /Δload was calculated by the difference between VO 2 max and at rest, divided by the maximum load and expressed in ml.min -1 .watts -1 . For practical purposes, we considered VO 2 at rest as 3.5 ml.kg -1 .min -1 . 8,9,11,12 Since the exam was performed on a treadmill, we adapted the load to Watts. Power calculation for the treadmill was: W (kgm/ minute) = mass (patient weight -Kg) x speed (meters/minute) x sine of the alpha angle (which is the angle between the treadmill and the ground).
The relation between kilogram-metres and watts is: 1 watt corresponds to 6.1 kilogram-metres/minute, and the transformation of the load from Kgm/min to Watts is done.
e. VE/VCO 2 slope corresponds to the inclination of the representative regression line between VE and VCO 2 . 8,9,11,12 f. T 1/2 VO 2 , the necessary time for a 50% drop of peak VO 2 in the recovery period, was quantified in seconds. 8,9,11,12

Statistical analysis
Statistical analyses were processed through the software SPSS Statistic 19.0 (IBM Corporation, 2010). Quantitative variables were described as mean ± standard deviation, fulfilled the assumption of normality, and the comparison between the groups was done through unpaired Student's t test. Categorical variables were summarized as percentages and compared between the groups through chi-square test (X 2 ) or Fisher's exact test, when appropriate. Significance level was set at 5% for α error and tests were two-tailed. Relative risks with confidence intervals (CI) of 95% were estimated. For the multivariate analysis of covariance (MANCOVA), we used the Trace of Pillai, power (≥ 0.8) and partial ETA 2 (effect dimension) as statistical tests. Partial ETA 2 was applied to infer clinical significance, considering the greatly elevated effect dimension if values are > 0.5, and small effect dimension if values are ≤ 0.05.

Ethical aspects
The project was submitted to the Research Ethics Committee and approved under protocol number 0770.0.000.107-11. All patients who participated in the research signed a free consent form.

Results
The sample was composed by 49 consecutive individuals (18 men and 31 women; mean age of 59.8 ± 7.7 years) who underwent CPET, distributed into: LBBB group (26 patients) and control group (23 patients).
LBBB group present significantly higher values of LV systolic and diastolic diameter in comparison to the control group, with a mean difference of 0.34 ± 0.11 cm and 0.39 ± 0.13 cm, respectively. Left ventricular mass index (LVMI) was also significantly higher in the LBBB group, with a mean difference of 13.19 ± 6.18 g/m 2 (p = 0.039). LVEF was lower in the LBBB group, with a significant mean difference of 0.12 ± 0.02% (p < 0.0001) ( Table 2). No significant difference was found between the groups regarding the occurrence of diastolic dysfunction (Table 3). Table 4 shows the hemodynamic and ventilatory variables of the CPET between LBBB and the control group. Through MANCOVA, we verified a significant association between LBBB and the six-variable set of the cardiopulmonary test, with an elevated effect dimension ( 0,578), and power above recommended (β BRE = 0,997). Adjustment was done for sedentary life style and the following covariables: age, gender, BMI, arterial hypertension, LVMI, and LVEF (Table 5).  There was a second multivariate analysis to identify if LBBB had an effect on one or more of the six outcome variables of CPET (Table 6). It was observed that LBBB interfered only in the analysis of VE/VCO 2 slope, with elevated effect dimension ( 0,504) and elevated power (β > 0,80). Figure 1 shows the study sensitivity analysis, considering the sample dimension (n), α = 0.05, and power of 0.80 (1-β). With the sample size (we considered a n = 48, value multiple of 4, corresponding to the number of groups to be compared: LBBB and SAH) and the assumed power and statistical significance, it was verified that the study was able to detect effect dimensions above 0.18 through MANOVA.

Discussion
In the studied population, VE/VCO 2 significantly increased in the LBBB group in relation to the control group. The ventilatory equivalent of carbon dioxide represents how much it is necessary to ventilate to eliminate a certain amount of produced carbon dioxide. This relation is analysed by linear regression, which reflects the slope of the line. VE/VCO 2 slope is already established in literature as a marker of poor prognosis in CHF patients, when it reaches a value > 34. 13 There is a relation between VE/VCO 2 slope increase and the presence of LBBB in dilated cardiomyopathy patients, regardless of CAD. 14 However, the literature is still scarce in the analysis of this variable for those without ventricular dysfunction.      Chemoreflexes are the main control and regulation mechanisms of ventilatory responses to changes in oxygen and carbon dioxide concentrations. Peripheral chemoreceptors located in the carotid bodies respond primarily to hypoxia, and central chemoreceptors, located in the ventral surface of the spinal cord, respond primarily to hypercapnia. It has been proven that there is a potential selection of central chemosensitivity in CHF patients functional class I and II. 17 Ponikowski et al. 18 studied CHF patients with preserved tolerance to exercise and observed that the normal increase of ventilatory response to exercise occurred due to the high sensitivity of the cardiorespiratory system control reflex. They showed the increase of central and peripheral chemosensitivity, compromise of the sympathovagal system with sympathetic predominance, hyperactivity of peripheral muscles ergoreceptors, and reduction of circulation baroreflex control.
The present work showed a VE/VCO 2 slope increase in the LBBB group with preserved LVEF. It is more likely that the mechanism responsible for this alteration is the initial imbalance of the autonomic system, because this population does not present hemodynamic compromise.
Echocardiographic results suggest that LBBB presence may trigger LV remodelling. These data were also presented in studies, such as the one by Melek et al., 19 who evaluated LBBB patients without cardiomyopathy though echocardiography and observed that the end systolic diameter was larger and LVEF was smaller in the LBBB group (54 ± 7%) in relation to the control group (61 ± 6%, p < 0.001). Valenti et al. 20 studied LV and LVEF through cardiac MRI, and demonstrated that LBBB patients presented higher LV and LVMI volumes and lower LVEF in comparison to control individuals.
It is established in the literature that there is no good correlation between peak VO 2 and ventricular function. 5,6 VO 2 during exercise provides an objective parameter of functional capacity, and indirectly reflects cardiovascular function. Impairment of exercise capacity is an independent predictor of adverse prognosis. In our work, we evaluated the main CPET variables related to cardiovascular performance. LBBB patients reached a lower percentage of the predicted peak VO 2 , but in the multivariate analysis, it was shown that the predicting factor for the referred reduction was sedentary life style, and not LBBB. Duncan et al. 14 studied functional capacity predictors in dilated cardiomyopathy patients, and LBBB was also not an independent predictor of the percentage of the peak VO 2 predicted in the multivariate analysis. PO 2 is representative of LV systolic volume and of the arteriovenous difference of O 2 during effort, and it reflects O 2 supply to the myocardial and the cardiac functional reserve under physiological stress. Generally, it increases gradually and linearly with the load increase until it reaches its highest value. In ventricular dysfunction patients, PO 2 may be reduced, despite the increase in load, or reach an early plateau even before reaching effort exhaustion. 11,12 In LBBB patients, LV shape is distorted during pre-ejection, as a result of the flattening of the septal curvature and simultaneous stretching of the lateral wall activated late. 21 The present work showed that the LBBB group reached peak PO 2 > 85% of the predicted value, and presented ascending behavior of the PO 2 curve during effort. Even though there are structural alterations in the LV, LVEF was preserved, without cardiovascular performance deficit.
It was demonstrated that isolated LBBB did not interfere in the ΔVO 2 /Δload -VO 2 was adequate for the applied workload. There was also no LBBB interference in T 1/2 VO 2 . T 1/2 of VO 2 is a prognostic marker, and its value increases according to the severity of CHF. It was demonstrated that the association of T 1/2 VO 2 with peak VO 2 improves the prognosis evaluation and identifies higher risk groups. 22 The groups had T 1/2 VO 2 lower than 90 seconds, a time considered to be within normality.
Our results show that CPET is a non-invasive, physiological, low cost method that could contribute in the follow-up of LBBB patients. We have demonstrated that, even though the LBBB group with preserved LVEF presented LV structural alterations within thresholds considered normal, cardiovascular performance was not impaired in the analysis of CPET metabolic variables, but there was an increase in the VE/VCO 2 slope in relation to the control group. It is not yet possible to identify which individual with isolated LBBB will develop ventricular dysfunction. Further studies should evaluate the impact of VE/VCO 2 slope increase by following up this sample of the population.
During this research, there were difficulties in the selection of LBBB patients, due to this pathology's concomitance with SAH, ventricular dysfunction and/or myocardial ischemia. SAH was not considered an exclusion criterion due to its high prevalence in this population, making it necessary to adjust its influence through statistical analysis. Another limitation was the exclusion of myocardial ischemia through echocardiography during physical stress, for which the ideal exam would be coronary angiotomography or coronary cineangiography,