Lower-limb endurance training program influences thoracoabdominal motion of patients with COPD ?

Introduction: Thoracoabdominal-TA asynchrony is an important sign of Chronic Obstructive Pulmonary Disease (COPD). Studies investigating the influence of endurance training on TA asynchrony have not been found. Objective: To analyze lower-limb endurance training effects on TA asynchrony in patients with COPD. Materials and methods: Two patients with severe COPD were evaluated in a single-subject design AB (A-baseline for six weeks, B-training on cycle ergometer with intensity of 70% of baseline peak load, for 12 weeks) with repeated measures of variables: phase inspiratory relation (PhRIB), phase expiratory relation (PhREB) and phase angle (PhAng). These variables were assessed by respiratory inductive plethysmography during incremental exercise tests on a cycle ergometer (same load and peak load of each test). Statistical methods included visual analysis, two-standard deviation band test and split middle line test, considering significant p < 0.05. It was considered the results for variables with agreement of at least two


Introduction
An important sign of COPD, related to disease severity, is the presence of asynchronous thoracoabdominal motion (TAM) during the respiratory cycle (1-4).Asynchronous TAM can be described in two categories: asynchrony, which reflects the delay between thoracic and abdominal compartments in expansion or retraction and paradoxical movement which consists of opposite movement between rib cage (RC) and abdomen (AB), also reported as complete asynchrony (5,6).
Several studies have assessed TAM in patients with COPD (1-3, 5, 7-10).According to the literature, these patients present more thoracoabdominal asynchrony compared to healthy subjects, both at rest (5,8) and during exercise (9).Exercise may reinforce the asynchrony in patients with COPD (7,9), and the presence of asynchronous TAM in these patients was associated with higher severity of the disease, increased risk of respiratory failure and worse prognosis (1-3).However, in most of the studies objective measurements were not used (1-3, 9).
Another disturbing sign of COPD is the intolerance to exercise (11).One of the primary symptoms that limits the exercise in patients with COPD is dyspnea (11,12), which can be related to the dynamic hyperinflation (4,(12)(13)(14)(15).
Dyspnea and exercise capacity limitation form a "vicious circle" with social and psychosocial consequences for these patients.Pulmonary rehabilitation can interrupt this circle since that the lower-limb endurance training can provide important positive effects in patients with COPD (16)(17)(18)(19)(20)(21).However, studies of two bands (Teflon ® -coated inductance) used to measure the changes in cross-sectional area of the RC and AB.Signals were calibrated using the Qualitative Diagnostic Calibration (26) and the following variables were analyzed: 1-Phase Angle (PhAng): it reflects the delay between RC and AB excursions and has been studied in patients with COPD (5,7,8).It ranges from 0° (perfect synchrony) to 180° (paradoxical movement) (26).This variable has the advantage of incorporating data from the whole respiratory cycle, but in the presence of irregular or "8 figure" Konno-Mead loops, PhAng values may be wrong (27).
2-Inspiratory Phase Relation (PhRIB) and Expiratory Phase Relation (PhREB): express the percentage of agreement between RC and AB movements directions during inspiratory or expiratory time, respectively.Range from 0% (perfect synchrony) to 100% (paradoxical movement).These parameters have the advantage of quantifying thoracoabdominal asynchrony at each point of the respiratory cycle and of not depending on calculations derived from Konno-Mead loops (26).
3-Cross-Correlation Function (CCF): it determines the delay in seconds (s) between the signals of RC and AB at each respiratory cycle.CCF equal 0s reflects perfect thoracoabdominal synchrony (28).Its calculation does not depend on Konno-Mead loops (28,29).

Intervention
Patients underwent a 12-week exercise training, three times per week, on a cycle ergometer at 70% (16) -target load -of the mean peak loads achieved in baseline tests.The goal was to reach 30 minutes of cycling at this intensity.Pedal frequency was maintained at 60 rpm.SpO 2 , BP, HR and perceived exertion were evaluated every five minutes.Recovery intervals where the workload was decreased to the basal load were allowed, according to participant tolerance or occurrence of any exercise interruption criterion (23).

Data reduction
The peak work load was defined as the last load at which the participant was able to complete at least 20s of pedaling.Plethysmography data processing was made by a blinded investigator for the addressing the influence of lower-limb endurance training on thoracoabdominal asynchrony of patients with COPD have not been found.

Materials and methods
It was conducted a single-subject experimental design AB aiming at documenting lower-limb endurance training effects on TA asynchrony in patients with COPD.The study consisted of weekly assessments for six weeks (Phase A) followed by lower-limb endurance training for 12 weeks with evaluations every 15 days (Phase B).The University Ethics Research Committee approved the protocol and the subjects gave informed consent.

Subjects
Two patients with severe COPD (4), negative results on bronchodilator tests (22), clinical stability, aged 73 (participant 1) and 64 (participant 2) years old, normal body mass index (BMI), former smokers, without cardiac and metabolic diseases and not performing physical exercise on a regular basis.

Assessment
Patients underwent a maximal incremental symptom-limited exercise test on cycle ergometer (Ergo Cycle167, Pirmasens, Germany) consisting of: 1) three minutes resting, 2) one minute cycling at the basal bicycle work load (15 watts), 3) 10-watt increments per minute until exhaustion or occurrence of any exercise interruption criterions (23), 4) three minutes cycling at the basal work load.Throughout testing, except during the recovery phase, pedal cadence was kept at 60 rpm.Electrocardiographic signs were monitored continuously.Transcutaneous oxygen saturation (SpO 2 ), heart rate (HR) and perceived exertion were evaluated every one minute, while blood pressure (BP) was measured every two minutes.
The respiratory inductive plethysmography (Respitrace ® 204, NIMS, Miami, FL, USA) was used to assess TAM during the incremental exercise tests.The accuracy of the plethysmography has been evaluated in patients with COPD at rest and during exercise (24,25).This noninvasive system consists

Results
Table 1 shows the spirometric and anthropometric data for the two participants.
Figure 1 (panel I) presents a typical example of waveforms at rest (A) and during exercise (B) seen through the software (RespEvents, NIMS, USA).At rest, the TAM can be considered almost synchronic, with delay occurring between thoracic and abdominal compartments only in some respiratory cycles (indicated by dotted lines).During exercise, in addition to the delay between compartments, it can be observed asynchrony within the cycles (indicated by arrows).It was also observed that tidal volume waveform during exercise was non-sinusoidal and irregularities appear to be arising out of the abdominal compartment.Similar waveforms were observed in both participants in all tests.The bottom of Table 2 presents TAM data at rest and during exercise.Comparisons showed significant increases in PhRIB, PhREB and CCF during exercise, in both participants, without significant differences between the two levels of exercise.
Figure 2 presents charts of TAM variables throughout the study phases at same load and peak load which were used as the basis for Visual Analysis.Phase Angle was not analyzed due to the high number of respiratory cycles presenting Konno-Mead loops with "8 figure", particularly during exercise.
Phase Angle was not analyzed due to the high number of respiratory cycles presenting Konno-Mead loops with "8 figure", particularly during exercise.study phase.In each evaluation, the following periods were considered: 1) at rest -30s during steady state; 2) same load -the last 30s of the highest load tolerated for one minute in all baseline tests (19); 3) peak load -the last minute before peak load of each assessment.
Respiratory cycles with loops in "8 figure" were excluded from PhAng analysis (26).In order to decrease data variability, PhRIB and PhREB were submitted to angle transformation for the data analysis.

Data analysis
Comparisons of TAM among rest and the two levels of exercise (same load and peak load) were performed using ANOVA for repeated measures.Comparisons of TAM between baseline and intervention phases at two levels of exercise (same load and peak load) were carried out using three methods: Two Standard Deviation Band and Celeration Line tests (30,31), and Visual Analysis, performed by three independent and blinded investigators.For assessing the concordance between assessors (30), Kappa statistics was used (32).It was considered as response to intervention results with concordance between Visual Analysis and at least one statistical test.
For statistical tests, the level of significance was set at α = 0.05.For ANOVA, Bonferroni correction was used, modifying the level of significance to 0.017 (33).Kappa statistics and ANOVA were performed using the Statistical Package for the Social Sciences (SPSS 13.0, Chicago, IL, USA).Data are present as mean ± standard deviation when appropriate.variables and for only one variable there was substantial agreement (Kappa from 0.61 to 0.80) (31).

Discussion
The main results of this study were: 1) Comparisons of TAM between baseline and intervention phase showed significant decrease in thoracoabdominal asynchrony at same load and peak load; 2) Comparisons between rest and exercise showed significant increase in asynchronous TAM at same load and peak load, without significant differences between these two levels of exercise.
To the best of our knowledge, this study was the first one to assess the impact of a lower-limb endurance training program on variables that reflect thoracoabdominal asynchrony during exercise in COPD patients.Comparisons of TAM variables between baseline and training phase showed significant decrease in thoracoabdominal asynchrony at two levels of exercise.This reduction can be considered as a positive impact of endurance training, since the presence of asynchronous TAM in patients with COPD has been associated with higher severity of the disease, increased risk of respiratory failure and worse prognosis (1-3).
Table 3 presents the results regarding Visual Analysis, Celeration Line and Two Standard Deviation Band, and the response to intervention of each variable at same load and peak load.Participant 1 showed significant decrease in PhRIB at same load and in PhREB at peak load, without significant differences in other variables.Participant 2 showed a significant decrease in PhRIB and PhREB at same load and in PhRIB at peak load.No significant changes were observed in other variables.With respect to Visual Analysis, it was observed almost perfect agreement (Kappa from 0.81 to 1.00) between two of the three examiners for most variables and for only one variable there was substantial agreement (Kappa from 0.61 to 0.80) (31).
Table 3 presents the results regarding Visual Analysis, Celeration Line and Two Standard Deviation Band, and the response to intervention of each variable at same load and peak load.Participant 1 showed significant decrease in PhRIB at same load and in PhREB at peak load, without significant differences in other variables.Participant 2 showed a significant decrease in PhRIB and PhREB at same load and in PhRIB at peak load.No significant changes were observed in other variables.With respect to Visual Analysis, it was observed almost perfect agreement (Kappa from 0.81 to 1.00) between two of the three examiners for most  subjects of each study, which can be verified by the standard deviation values.This variability may be due to the analysis of respiratory cycles based on altered Konno-Mead loops since those authors did not report the exclusion of "8 figure" Konno-Mead loops during PhAng analysis.Thus, a rigorous assessment of Konno-Mead loops with exclusion of "8 figures" is an important step to be followed in PhAng calculation.
In the present study, the large number of "8 figure" Konno-Mead loops was determinant for the exclusion of this variable from data analysis.
Both participants studied increased their peak exercise capacity at the end of 12 weeks of training.It may be evidenced by the increase in the work load reached in the last incremental test when compared to mean peak load of baseline tests: 17 and 10 watts in participants 1 and 2, respectively.These values are in agreement with results observed in previous studies (17,19,21) and may be related to the improvement of the physical performance provided by the lower-limb endurance training program.

Conclusion
Using a single-subject experimental design, the effects of a lower-limb endurance training program on variables reflecting thoracoabdominal motion were evaluated in two patients with severe COPD.The results suggest that despite the exercise augmented the asynchrony compared to rest, the endurance training program influenced positively the thoracoabdominal motion leading to significant decrease in asynchrony during exercise in patients assessed.However, it was observed that TAM variables responses were not the same for the participants evaluated.This difference could be the result of distinct dynamic hyperinflation patterns, as reported in the literature (21,(34)(35)(36)(37), since respiratory muscles dysfunction, related to lung hyperinflation, has been described as the main factor responsible for thoracoabdominal asynchrony in patients with COPD (5,8,(36)(37)(38).However, it cannot be confirmed as dynamic hyperinflation was not assessed in this study.Different thoracoabdominal asynchrony patterns (1, 9) and progression of these patterns during exercise in some patients with COPD (9) may also have contributed to this difference between the participants.

Funding acknowledgements
Comparisons between rest and the two levels of exercise showed significant increase in asynchronous TAM at same load and peak load, without significant differences between these exercise levels.This result is similar to that observed by Alves et al. (7), who compared the TAM of 22 COPD patients at rest and during three levels of progressive exercise (30-50%, 70-80% and 100% of the peak work load) performed on a cycle ergometer.These authors observed significant increase in asynchronous TAM during all exercise levels, without difference between them.
During exercise, non-sinusoidal waves were observed, showing irregularities from abdominal compartment.This finding was observed in both participants.Similar waves have been observed in COPD patients during spontaneous breathing (1,10) and exercise (9,38).Ashutosh et al. (1) assessed TAM during spontaneous breathing at rest in 16 stable patients and in 14 patients with exacerbation.Asynchronous TAM was observed in 13 patients, 1 stable and 12 exacerbated.Three types of thoracoabdominal asynchrony were observed, all involving abdominal compartment irregularities.Delgado et al. (9) observed, during progressive exercise on a treadmill, similar movements of RC among patients but distinct AB movement patterns.It was also found that asynchronous TAM was related to underlying pulmonary abnormality and exercise limitation.
Regarding TAM asynchrony at rest, studies presenting PhRIB, PhREB and CCF values in patients with COPD have not been found.PhAng is the TAM variable most often reported in these patients (5,7,8).It was observed a large variability in PhAng mean values among these studies and important variability among Figure 1 (panel II) shows examples of Konno-Mead loops with (a, b and c) and without (d) "8 figure".

Figure 1 -
Figure 1 -Example of typical waveforms at rest (A) and during exercise (B) on panel I, and Konno-Mead loops with "8 figure" (a, b and c) and without "8 figure" (d) on panel II Source: Research data.Notes: Vt = Tidal Volume; RC = Rib Cage; AB = Abdomen.The dotted line allows to verify the compartments delay.The arrows indicate the irregularities waves.

Figure 2 -
Figure 2 -Graphics of TAM variables during baseline (A) and intervention (B) phases at rest, same load and peak loak on the two participants studied Source: Research data.Notes: PhRIB = inspiratory phase relation in percentage; PhREB = expiratory phase relation in percentage; CCF = cross-correlation function in seconds; A = baseline phase with periodic assessments weekly during six weeks; B = intervention phase with periodic assessments each 15 days, during 12 weeks.

Table 1 -
Spirometric and anthropometric data for the two participants

Table 1 -
Spirometric and anthropometric data for the two participants

Table 2 -
All tests mean of asynchronous TAM data at rest and during two exercise levels (same load and peak load) for both participants Notes: Data are presented as mean ± standard deviation.PhRIB = inspiratory phase relation; PhREB = expiratory phase relation; CCF = cross-correlation function; * = statistical significant difference between rest and the two levels of exercise, considering significant p < 0.017, according to Bonferroni's correction.

Table 3 -
Response to intervention according to Visual Analysis, Celeration Line and Two Standard Deviation Band during same load and peak load