Rehabilitation has been a treatment method that is considered essential for patients with chronic obstructive pulmonary disease (COPD). Traditionally, this has centered on lower limb exercises, such as walking or stationary cycling. However, many patients with COPD report the onset of dyspnea and muscle fatigue when performing activities in which the upper extremities are involved, such as lifting objects, making the bed or gardening.1 At present, upper body exercises have been included in the rehabilitation programs of patients with COPD, but there is still no consensus with respect to how conduct training in these patients and why the shoulder girdle muscles limit exercise in these cases.

In 1986, Celli et al.2 were the first to describe that patients with COPD could develop thoracoabdominal asynchrony during pegboard exercise with the upper limbs. In their study, they observed 42% of patients with asynchrony who also presented major dyspnea, earlier fatigue of the upper extremities and shorter endurance exercise time.

Gigliotti et al.3 observed that patients with COPD who performed exercises on an arm cycle ergometer developed dynamic pulmonary hyperinflation, increased dyspnea and upper limb fatigue.

It seems inconceivable to admit that the restriction experienced by a patient with COPD during exercise performed with the arm could be due to an association between dynamic pulmonary hyperinflation and thoracoabdominal asynchrony. Nevertheless, this association has not been demonstrated in previous studies. It is generally considered that the technique of arm cycle ergometry is the gold standard for arm training; however, it is believed that the diagonal technique recruits the action of more muscles than other types of exercise.

Consequently, we hypothesized that dynamic hyperinflation and thoracoabdominal asynchrony occur in patients with COPD who perform arm exercises. Hence, the aim of this study was to analyze the possible onset of dynamic pulmonary hyperinflation and thoracoabdominal asynchrony in patients with COPD during two types of upper limb exercises: the arm cycle ergometer and the diagonal technique. We also evaluated the most appropriate arm training technique for use in routine clinical practice in relation to dynamic hyperinflation and thoracoabdominal asynchrony.

This study enabled us to obtain the following results: firstly, at least one-third of patients with COPD had significant hyperinflation after an upper limb exercise test. Secondly, thoracoabdominal asynchrony occurred only in patients who presented dynamic pulmonary hyperinflation. Thirdly, the upper limb exercise using the diagonal technique induced less dynamic pulmonary hyperinflation, thoracoabdominal asynchrony and changes in the respiratory pattern than the arm cycle ergometer technique.

Although arm cycle ergometer training is usually considered the gold standard in rehabilitation of the upper limb muscles, the diagonal technique is also used for training the upper half of the body, as it resembles the movements of various activities of daily living, and because it involves recruitment of more muscle groups.14

Ashutosh et al.18 and Tobin et al.19,20 studied the breathing patterns in patients with COPD who performed simple activities of daily living using their arms, such as brushing their hair or tying their shoelaces, and observed that these patients had thoracoabdominal asynchrony. Gigliotti et al.3 were the first to demonstrate that patients with COPD may show dynamic pulmonary hyperinflation during upper limb exercise that leads to an increase in the ventilatory load. It is believed that there is an association between the increase in dyspnea and the increase in thoracoabdominal asynchrony with hyperinflation, but this association has not been demonstrated until now.

The onset of thoracoabdominal asynchrony is expected with an increase in respiratory resistance2,13 (upper or lower airways, lung tissue), decrease in pulmonary distensibility (parenchymal disease) and increase in chest wall distensibility20 (flexible rib cage, neuromuscular disease); this may contribute to producing respiratory muscle fatigue.21–24

Our patients did not present thoracoabdominal asynchrony before the diagonal (MCA/VT=1.07±0.12) and arm cycle ergometer exercise tests (MCA/VT=1.05±0.15). However, during the exercises, approximately 30% of patients with the diagonal technique and 60% of patients with the arm cycle ergometer presented thoracoabdominal asynchrony, with a disproportionate increase in the participation of the abdominal component with respect to the thoracic component. The abdominal motion contributed 75% to the chest wall movement, while the chest motion contributed 50%, so the sum of both compartments reached 125% with respect to the rest period; this clearly indicates thoracoabdominal asynchrony. Thus, we can hypothesize that the asynchrony presented by patients was mainly due to overwork of the diaphragm muscle. Criner and Celli24 demonstrated that, during arm exercise, the rib cage and abdominal wall should be fixed to stabilize the torso, so the rib cage becomes more rigid.

Likewise, arm exercises can produce stabilization of the rib cage and excessive movement of the abdomen, leading to increased end-expiratory lung volume. Vogiatzis et al.21 and Aliverti22 et al. observed that late hyperinflation during a leg exercise was associated with an increase in movement of the abdominal compartment. Some recent data have indicated that patients who performed an exercise with a leg cycle ergometer had increased end-tidal volume, presumably to cope with the expiratory action of the lower rib cage. One possible mechanism underlying this process is the dynamic hyperinflation that might have occurred in patients with asynchrony.22

The exercise restriction that occurs in patients with COPD is well known and is associated predominantly with ventilation.11,12 We observed an increase in the pulmonary ventilation, respiratory rate, tidal volume, mean inspiratory flow and dyspnea, and a decrease in the expiratory time in the maximal and sub-maximal exercise performed using the diagonal and arm cycle ergometer techniques. We also observed that the technique that enhanced the pulmonary ventilation and decreased the expiration time to a greater extent was the arm cycle ergometer technique, compared to the diagonal technique (Table 3). No difference was observed in the heart rate between the upper limb exercises, indicating that patients performed the exercises with the same workload (Table 3). We can probably start from the assumption that hyperventilation and shortening of the expiratory time lead to dynamic pulmonary hyperinflation that places a burden on the system, increasing the elastic work of breathing, which may predispose to fatigue. Patients with hyperinflation showed an increase in the mean inspiratory flow rate, which reflects the greater respiratory drive that these patients had to generate.

Furthermore, we observed that thoracoabdominal asynchrony only occurred in patients with hyperinflation. Regardless of the exercise, patients who presented pulmonary hyperinflation and thoracoabdominal asynchrony showed higher pulmonary ventilation, respiratory rate, central respiratory drive and dyspnea values in a manner associated with a shorter expiratory time, compared to patients without hyperinflation (Table 4). However, the percentage of patients who had these alterations was lower during the diagonal technique than during the arm cycle ergometer technique. We believe that the lower prevalence of onset of dynamic hyperinflation with the diagonal exercise is due to the increased expiratory time of the patients, and that it allows the inspiratory tidal volume to be completely expelled, thereby maintaining a more stable end-tidal lung volume. Thus, the diagonal technique may be more appropriate for exercising the upper limbs, as it not only induces less dynamic hyperinflation and thoracoabdominal asynchrony, but its introduction in a daily rehabilitation program is cheaper. However, to be precise, this observation should be evaluated using a complete rehabilitation program with the upper limbs, since long-term continuous physical training may reduce the dynamic pulmonary hyperinflation that these patients have during arm exercise testing.

Our study has three limitations: Firstly, we did not measure the total static lung volumes in our patients, and it could be argued that these measurements would help to improve evaluation of the resting respiratory status of our patients. Instead, we focused mainly on evaluating the dynamic inspiratory capacity and asynchrony at the end of the exercises and not on the degree of static pulmonary hyperinflation. Moreover, we designed the study to evaluate dynamic hyperinflation at the end of the exercise test and, unfortunately, we did not measure the inspiratory capacity when the exercises were carried out. This evaluation would have provided information relative to the exact moment of onset of dynamic hyperinflation; however, this was not the study objective and our aim was to specifically observe pulmonary hyperinflation at the end of the exercise tests. Secondly, we recruited only patients with COPD GOLD III and IV; in fact, however, we wanted to recruit these types of patients, as in our pulmonary rehabilitation clinical practice we observed that these patients presented higher dyspnea scores while they were performing arm exercises. Hence, from this clinical observation, we hypothesized that there would be differences in the primary outcome variables (inspiratory capacity and thoracoabdominal asynchrony) in these selected patients. Thirdly, we did not determine the metabolic load of either type of arm exercise; however, the workload of the diagonal exercise and the arm cycle ergometer were matched by the heart rate reached in the maximal exercise, which is a representative variable of the amount of work produced in each exercise.

As far as we know, the novel observation in this study is that, for the first time, it has been demonstrated that dynamic hyperinflation and thoracoabdominal asynchrony occurred simultaneously in patients with COPD who performed an upper limb exercise using an arm cycle ergometer or diagonal technique.

Our conclusion is that dynamic hyperinflation occurs due to superficial respiration which is detrimental to the respiratory pattern, as patients with COPD develop thoracoabdominal asynchrony and higher levels of dyspnea during the application of the diagonal and arm cycle ergometer techniques. The diagonal exercise technique may be the best technique for use in clinical practice compared to the arm cycle ergometer, as it induces less hyperinflation and thoracoabdominal asynchrony in patients with COPD.

None of the authors had any conflict of interests to declare in relation to this paper. We did not have any financial or personal relationships with other people or organizations that could inappropriately influence our work, such as employment, consultancies, stock ownership, honoraria, paid expert testimony, patent applications/registration and grants or other funding. No tobacco companies have financed any part of this manuscript.