Dyspneic patients with chronic obstructive pulmonary disease (COPD) who are markedly hyperinflated are considered especially likely to display abnormalities in rib cage motion such as paradoxical (inward) inspiratory movement of their lower rib cage (4,6,12–15). Studies in healthy humans have led to the hypothesis that the primary mechanism of abnormal chest wall motion in patients with COPD is probably an abnormal alteration of forces applied to chest wall compartments (9,19) and an increase in airway resistance (18). Chihara et al. (9) have speculated that, when rib cage distortion is present, a greater degree of recruitment of inspiratory rib cage muscles and greater predisposition to dyspnea for a given load and strength do occur. On the other hand, the role of hyperinflation on abnormal chest movement is questionable in healthy subjects (18). Accordingly, it has recently been shown that paradoxical movement of the lower rib cage cannot be fully explained by static lung hyperinflation (6) or dynamic rib cage hyperinflation (30) in patients with COPD. By contrast, Aliverti et al. (4) have shown that lower rib cage paradox yields to an early onset of dynamic hyperinflation as a likely explanation for the increased exertional breathlessness in these patients. Nonetheless, the link between changes in operational lung volumes and exertional breathlessness has not been definitely established in normoxic COPD patients (25,28,32). Intervention such as oxygen supplementation reduces ventilation and the rate of dynamic hyperinflation, but the contribution of reduced lung volumes to dyspnea relief remains uncertain in these patients (25,28,32).
Therefore, the purposes of this study were (i) to determine whether and by which mechanisms hyperoxia would affect exercise dyspnea, chest wall dynamic hyperinflation, and rib cage distortion in normoxic COPD patients; and (ii) to explore if these phenomena are interrelated. Our hypothesis is that they are not. We based the hypothesis on the following observations: (i) significant dyspnea relief and improvement in exercise endurance can occur even in the absence of an effect on dynamic lung hyperinflation (28); (ii) externally imposed expiratory flow limitation is associated with no rib cage distortion during strenuous incremental exercise, with indexes of hyperinflation not being correlated with dyspnea (16); (iii) end-expiratory chest wall volume may either increase or decrease during exercise in patients with COPD, with those who hyperinflate being as breathless as those who do not (3); and (iv) a similar level of dyspnea is associated with different increase in chest wall dynamic hyperinflation at the limit of exercise tolerance (33).
We based the analysis of chest wall kinematics on optoelectronic plethysmography (OEP), which evaluates the motion of chest wall compartments: the upper rib cage, lower rib cage, and abdomen (2,6,21).
To validate the hypothesis, we carried out the present investigation where the effects of supplemental oxygen administration on chest wall dynamic hyperinflation, rib cage distortion, and dyspnea were assessed during high-intensity constant exercise in severely obstructed and hyperinflated normoxic COPD patients.
Sixteen consecutive male patients with severe airway obstruction and hyperinflation participated in the study (Table 1). COPD was diagnosed on the basis of history, physical examination, chest radiograph, and results of pulmonary function studies (5). All patients had a long history of smoking and were clinically stable and on appropriate medication.
The study was conducted during 2 days. After giving written informed consent, patients were familiarized with all the testing procedures and completed a symptom-limited incremental exercise test. In a subsequent visit, they performed two constant-load exercise tests at 75% of their previously determined maximal work rate while breathing either 50% oxygen or room air (21% O2) in randomized order with a 60–90 min washout or recovery period between tests. Subjects were blinded to the oxygen concentration being breathed as was the investigator evaluating subjective responses and performing data analysis. The study was carried out at the Don Gnocchi Foundation, in accordance with the Declaration of Helsinki (2000) WMA and approved by the local ethics committee.
Pulmonary function tests
Routine spirometry and lung volumes, obtained with subjects seated in a comfortable armchair, were measured according to ATS/ERS guidelines (24). Functional residual capacity was measured with a body plethysmograph (Autobox DL, 6200; SensorMedics, Yorba Linda, CA) according to a standardized procedure (35). The normal values for lung function were those of the European Community for Coal and Steel (11). Patients breathed through a mass flow sensor attached to the mouthpiece, and its integral was used to display the flow volume loop and, if any existed, expiratory flow limitation (EFL). EFL was considered present at rest when >50% of the tidal breath met or exceeded the expiratory boundary of the maximal flow–volume loop (17).
Operational chest wall volume measurements
OEP provides an accurate three-dimensional computation of the volume of the chest wall (Vcw) based on coordinates from surface markers attached to the chest wall surface. In brief, six TV cameras (three placed 4 m behind and three placed 4 m in front of the subject) tracked the three-dimensional movements of 89 small surface markers attached to the skin of the trunk with double-sided adhesive tape and lit by infrared light-emitting diodes coaxial with lenses of the cameras. The OEP data were recorded at a sampling frequency of 60 Hz. Details of this technique have been thoroughly described previously (6,21,29). For methodological details, see Supplemental Digital Content 1, https://links.lww.com/MSS/A134.
Cardiopulmonary exercise testing
Symptom-limited cardiopulmonary exercise testing was conducted on an electronically braked cycle ergometer (Ergometrics 800; SensorMedics) using the Vmax 29c Cardiopulmonary Exercise Testing System (SensorMedics), according to procedures previously described (22). During the experimental visit, each subject performed two symptom-limited cycle exercise tests at the same constant power equal to 75% of their previously established peak work rate at a pedaling rate of 50–60 revolutions per minute. Subjects wore a nose-clip and breathed through a low-dead-space mouthpiece. Flow was measured with a mass flow sensor (Vmax 229; SensorMedics) near the mouthpiece, and lung volume changes were obtained by integrating the flow signal. A gas mixture (inspiratory oxygen fraction of 0.50 balanced with nitrogen) was inspired by the patients from a Douglas bag through a two-way nonrebreathing valve (model 27900; Hans-Rudolph, Kansas City, MO: 115 mL of dead space). The flow into the Douglas bag was constant, and patients breathed the gas mixture at the rate they demanded. We carefully reduced the impedance of the tubing by increasing its width and minimizing its length. To ascertain the linearity of the analyzer, we used a 0.50 oxygen calibration cylinder. During the test, the flow rate at the mouth and gas exchange were recorded breath-by-breath (Vmax 229; SensorMedics). Expired gas was analyzed for oxygen uptake (V˙O2), and carbon dioxide production (V˙CO2). Cardiac frequency was continuously measured using a 12-lead ECG; oxygen saturation was measured using a pulse oxymeter (NPB 290; Nellcore Puritan Bennett, Pleasanton, CA). Equipment was calibrated immediately before each test. V˙CO2 and V˙O2 were expressed as standard temperature, pressure, and dry.
The flow signal was synchronized to that of the motion analysis used for OEP and sent to a personal computer for subsequent analysis (further methodological details are in the Supplemental Digital Content 1 https://links.lww.com/MSS/A134).
The rest signals at quiet breathing were recorded during a 3-min period after a 10-min period of adaptation to equipment. The volume tracings were normalized with respect to time in each patient to allow ensemble averaging over three reproducible (i.e., with stable PETCO2) (10) consecutive breaths randomly chosen within the period of interest (rest, warm-up, each minute exercise, iso-time, end-exercise) and to derive an average respiratory cycle over each of the data acquisition periods. Inspiratory and expiratory phases of the breathing cycles were derived from the Vcw signal.
Because most patients were unable to relax their respiratory muscles enough to yield accurate and meaningful relaxation volume–pressure curves of the thorax, the presence of rib cage distortion was established by 1) comparing the time courses of Vrcp versus Vrca and 2) the phase shift between Vrca and Vrcp when these two volumes were plotted against each other. This was measured as the ratio of distance delimited by the intercepts of Vrcp versus Vrca dynamic loop on line parallel to the x axis at 50% of RCp tidal volume (m), divided by RCa tidal volume (s), as &thetas; = sin−1 (ms−1), a previously adopted approach (1,4,18,30) (Fig. 1). In this system, a phase angle of zero represents a completely synchronous movement of the compartments and 180° total asynchrony. Rib cage to abdomen displacement was assessed by the ratio of changes in Vrc to change in Vab. (Supplemental Digital Content 1, https://links.lww.com/MSS/A134).
Assessment of dyspnea
Patients were familiarized with a 10-point Borg category scale (7). Patients were asked to describe their perception of dyspnea before exercise testing and at the end of every minute during tests.
Breathing patterns and volume changes were assessed using two-way ANOVA. The Bonferroni test was used for multiple comparisons.
Values are means ± SD unless otherwise reported. The level of significance was set at P < 0.05. All statistical procedures were carried out using the Statgraphics Plus 5.0 statistical package (Manugistics, Rockville, MD).
Results on air
An early increase in Vcw,ee (chest wall dynamic hyperinflation) was found in 10 patients during exercise (Fig. 2). At iso-time (Table 2), these patients, called hyperinflators, exhibited a 0.47 ± 0.35 L increase in Vcw,ee (P < 0.000), of which 0.38 L was accommodated within the rib cage (Vrc,ee, P = 0.000) and 80 mL was within the abdomen (Vab,ee, P = not significant [NS]). The increase in Vcw,ei (1.63 ± 0.48 L, P < 0.000) was distributed in the three compartments. The 0.413 L increase in VTcw from rest to end-exercise (P = 0.000) was shared between 0.284 L of VTab and 0.129 L of VTrc. Individual compartmental data (Supplemental Digital Content 2, individual time courses; https://links.lww.com/MSS/A135) show the increases in Vrca,ee (eight patients), Vrcp,ee (5) and Vab,ee (3).
In six patients called non-hyperinflators, chest wall volumes did not significantly change (Fig. 2 and Table 2). Baseline function did not differ between hyperinflators and non-hyperinflators (Table 1); flow limitation ≥85% VT was found in each of the 16 subjects with no difference between the two groups. Hyperinflators exhibited higher values of Vcw,ee (P < 0.002), Vrca,ee (P < 0.01), Vab,ee (P < 005), and phase angle degree (P < 0.000) at iso-time on air and lower values of VE (P < 0.02), VTrcp (P < 0.05), and VTcw/Ti (P < 0.02) than non-hyperinflators but higher phase angle degree (P < 0.000) on oxygen (Table 2).
Effects of oxygen
Oxygen did not significantly modify chest wall operational volumes at rest but changed chest wall kinematics at iso-time during exercise in the 10 hyperinflators (Fig. 2 and Table 2). A lower increase in Vcw,ee (about 200 mL, P = 0.000) was shared between Vrc,ee and Vab,ee, and a lower increase in Vcw,ei (270 mL, P < 0.002) was mostly accommodated within the abdomen (Vab,ei). In turn, oxygen shifted the operational volumes to lower values without affecting the increase in VTcw, VTrc, and VTab on air; in contrast, VE, RF, and VT/Ti decreased, whereas Te increased. At iso-time, oxygen also increased the endurance time (from 3 ± 1.5 to 8.6 ± 5.9 min) and decreased the Borg score and leg effort. By contrast, in the six non-hyperinflators, oxygen did not significantly change the breathing pattern and chest wall kinematics, reduced the Borg score (from 9 ± 0 to 5.6 ± 1.03, P < 0.05) and leg effort (from 8 ± 0 to 5.7 ± 1.0, P < 0.01), but not endurance time (from 3.25 ± 0.5 to 4.6 ± 1.3 min).
Rib cage distortion
The time course of Vcw swings in a representative patient is shown in Figure 3: the two vertical lines indicate the beginning and end of inspiratory volume of the chest wall. Early inspiration at rest and during exercise, while the upper rib cage and abdomen inflate, the lower rib cage exhibits a paradoxical inward displacement. Data from our laboratory (using at least 2SD above the mean value for the normal subjects) gave a threshold for the upper limit of normal of 18° phase angle. The phase angle degrees at rest were as follows: 69.6° ± 35.5° in seven patients with rib cage distortion and 17.4° ± 2.8° in three patients with no distortion among hyperinflators; 49.1° ± 2.5° in three patients who distorted and 14.9° ± 1.2° in three patients who did not (χ2 = NS) among non-hyperinflators. Irrespective of the attained level of hyperinflation, rib cage distortion was less evident at the end-exercise in five hyperinflators (Supplemental Digital Content 2, individual time courses; https://links.lww.com/MSS/A135). Collectively, these data indicate that the occurrence of rib cage distortion is independent of changes in Vcw,ee.
It is worth noting that the average phase angle shift did not change throughout the exercise both on air and on oxygen (Fig. 4). Plots of individual data points at iso-time during exercise (Supplemental Digital Content 3, phase shift at iso-time exercise; https://links.lww.com/MSS/A136) show that average phase angle lay close to the identity line, with only one outlier patient exhibiting reduction in phase angle degree on oxygen. Baseline function did not differ between patients with and without rib cage distortion (Supplemental Digital Content 4, baseline lung function in patients with and without rib cage distortion; https://links.lww.com/MSS/A137). The former experienced the same level of dyspnea as the latter (Borg score = 8.2 ± 1.7 vs 9 ± 0.6, respectively) and stopped exercise because of either dyspnea (six patients), leg effort (two patients), or both (two patients); patients without rib cage distortion stopped because of dyspnea (four patients) or both symptoms (two patients). Oxygen did not change the reasons for stopping.
On air ΔBorg/ΔV˙E did not differ between hyperinflators and non-hyperinflators (0.43 ± 0.26 and 0.4 ± 0.2 AU·L−1·min−1, respectively, P = NS) or between patients with and without rib cage distortion (0.47 ± 0.29 and 0.56 ± 0.38 AU·L−1·min−1, respectively, P = NS). In patients as a whole, oxygen breathing did not change the relationship ΔBorg/ΔV˙E (0.42 ± 0.23 and 0.49 ± 0.3 AU·L−1·min−1, on air and oxygen, respectively, P = NS) so that a lower V˙E was associated with a lower Borg rate on oxygen. Change in Borg at iso-time during exercise with oxygen was not correlated with change in Vcw,ee (r < 0.4, P = NS) or with a change in V˙E. Endurance time (ΔBorg/Δtime) was not significantly correlated with change in Vcw,ee per unit increase in V˙E (ΔVcw,ee/ΔV˙E). As mentioned above, the change in Borg score on exercise was associated with unchanged phase angle degree on air or oxygen.
This study shows that lower dyspnea during exercise with oxygen was associated with a decrease in ventilation regardless of whether patients distorted the rib cage, dynamically hyperinflated, or deflated the chest wall. Changes in chest wall with oxygen were accommodated within the lower rib cage and abdomen.
Chest wall kinematics
Dynamic hyperinflation of the chest wall (increase in Vcw,ee) was accommodated within the rib cage (Vrc,ee) and, to a lesser extent, the abdomen (Vab,ee). By contrast, a decrease in Vab,ee balanced by an increase in Vrc,ee resulted in unchanged Vcw,ee in the non-hyperinflators. The two kinematic patterns are similar to those reported by Aliverti et al. (3) during incremental exercise with a decrease in Vcw,ee being noticed in the less obstructed, hyperinflated patients with marked abdominal pressure generation. However, there are many arguable reasons why some COPD patients deflated the chest wall and others dynamically hyperinflated. The following explanations might be considered: (i) dynamic hyperinflation attenuates the expected increase in expiratory flow limitation as ventilation increases. It is worth noting that the locus of pressure associated with expiratory flow limitation (Pmax) is a function of lung inflation; and above functional residual capacity, expiratory pressure never exceeded Pmax during exercise (27). (ii) Evidence from studies in healthy subjects suggests that, regardless of the level of imposed expiratory flow limitation on exercise, some subjects hyperinflate and some deflate the chest wall (16). (iii) The possibility of associating rib cage hyperinflation with abdominal volume deflation in non-hyperinflators (Fig. 2 and also Fig. 2 in Vogiatzis et al. ) supports the suggestion that abdominal muscle recruitment is unlikely to have a significant impact on expiratory flow rate in these patients (23,34). Accepted evidence remains, however, to be proven in a larger COPD population.
Differences in the relative amount of emphysema might also be relevant because patients with a more emphysematous clinical profile (low DLCO) have faster rates of dynamic hyperinflation and greater constraints of tidal volume expansion during exercise and are expected to have a greater propensity to expiratory flow limitation because of reduced lung recoil and airway tethering (28). We postulate, however, that the observed differences in Vcw,ee during exercise were likely due to different mechanisms as DLCO was similar in hyperinflators and non-hyperinflators (Table 1).
A novel finding of this article is the effect of supplemental oxygen on chest wall kinematics during high-intensity constant exercise in normoxic COPD patients (Fig. 2 and Table 2). Oxygen reduced the increase in Vcw,ee and Vcw,ei, in hyperinflators; a similar increase in VTcw at iso-time during exercise was obtained at operational volumes lower than on air. This deflationary activity of oxygen was mainly accommodated within the lower rib cage (Vrca,ee) and, to a lesser extent, the abdomen (Vab,ee). Also, the reduction of Vcw,ei was shared between the lower rib cage (Vrca,ei) and abdomen (Vab,ei). In contrast, oxygen did not affect chest wall operational volumes in the six non-hyperinflators. Because supplemental oxygen is unlikely to affect airway resistance in COPD patients (8,25), the mechanism for reducing hyperinflation was likely due to decreased ventilatory drive (VE, VT/Ti), lower RF, and prolonged Te, which, in turn, facilitates expiration and increases chest wall dynamic deflation (8,28,31,32).
We have recently shown that paradoxical movements of lower rib cage cannot be fully explained by static lung hyperinflation at rest (6) or even partially by dynamic rib cage hyperinflation (30) during arm exercise in COPD patients. Extending those data, we have here shown that chest wall dynamic hyperinflation is not strictly coupled with rib cage distortion as shown by the following: (i) a phase angle >18° indicated rib cage distortion in patients who hyperinflated and in those who did not; (ii) an early Vcw,ee increase at warm-up was associated with decreased phase angle on air and at the maximum Vcw,ee the increase in phase angle did not change from rest to end-exercise (Fig. 4); (iii) despite decrease in Vcw,ee, the phase angle lay unmodified at iso-time during exercise with oxygen (Table 2; and Supplemental Digital Content 3, phase shift at iso-time exercise; https://links.lww.com/MSS/A136); and (iv) a lower rib cage paradox was associated with an evident increase in Vcw,ee in only three patients (As, An, Te in Supplemental Digital Content 2, individual time courses; https://links.lww.com/MSS/A135). Collectively, our results do not mimic those by Aliverti et al. (4) who have shown that lower rib cage paradox yields to an early onset of chest wall hyperinflation and that an increase in hyperinflation is associated with a decrease in phase angle shift at end-exercise. Differences in COPD populations and exercise protocols (endurance here vs incremental there) have certainly contributed to the different results of the two studies.
The coordinated action of operating muscle forces is constant during exercise, explaining the very low rib cage distortion in healthy subjects (2,10,19). Conversely, a non–uniformly distributed pressure on the rib cage or different values of upper and lower rib cage compliance (9) might account for constant rib cage uncoupling throughout exercise. The present data do not allow us to define this point.
In keeping with previous results (25,28,32), the administration of supplemental oxygen increased exercise endurance, reduced chest wall dynamic hyperinflation, and decreased both dyspnea and leg effort in hyperinflators. Besides its effects on dynamic hyperinflation, the multifactorial effects of hyperoxia involve many integrated mechanisms, e.g., decreased impedance of the respiratory system and improved respiratory muscle function, cardiovascular function, and central nervous system function (25,31). This might be the reason why neither end-expiratory volume of the chest wall here nor inspiratory capacity in other studies (25) are correlated with reduced breathlessness, despite being reduced.
Our findings that the Borg score on air did not differ between patients with and without rib cage distortion (lower rib cage paradox) and that changes in Borg with oxygen were associated with no change in phase angle shift do not provide evidence that rib cage distortion plays a major role in the perceived intensity of breathlessness. But that does not mean that it could not contribute because there is no evidence that phase shift accurately reflects the different pressures acting on lower and upper rib cage (9,19). Recent studies (2,29) have shown that, to minimize distortion, all that is necessary is for the pressure developed by the rib cage muscles to be directly proportional to the pressure developed by the abdominal muscles and 180° out of phase. Despite a large increase in rib cage muscle (Prcm) and abdominal pressures (Pabm), minimal increases in rib cage distortion have been found at 180° out of phase of Prcm and Pabm during leg or arm exercise in healthy humans (29). We now know that the gradual relaxation of abdominal muscles is crucial in minimizing rib cage distortion (2).
Either dyspnea or leg effort or both may be the principal complaints for stopping exercise in patients with COPD (20,25,26,28). Regardless of whether patients distorted the rib cage, dynamically hyperinflated, or deflated the chest wall, the primary symptom limiting exercise is dyspnea. These data are in keeping with those of Iandelli et al. (16) who found that externally imposed expiratory flow limitation does not necessarily lead to dynamic hyperinflation, does not impair exercise performance in subjects who do not hyperinflate the chest wall, and does not contribute to dyspnea in subjects who hyperinflate until the highest expiratory flow limitation exercise level is reached. Collectively, all these data are not in line with a previous report (4) showing that an early onset of dynamic hyperinflation of the chest wall is the most likely explanation of predominance of dyspnea in patients with rib cage distortion, and that when paradox is absent, the sense of leg effort becomes a more important symptom-limiting exercise. In turn, data in healthy subjects may or may not reflect the response in COPD patients. Also, the effort-dependent nature of different exercise tests, the underlying multifactorial mechanisms, and the subjective nature of dyspnea scaling might account for different results in patients.
Critique of method
We have recently applied OEP for the first time to quantify the degree of rib cage distortion at rest in COPD patients (6,30). The limitation of OEP in assessing the relative changes in Vrcp to Vrca has been thoroughly discussed in those articles. Briefly, a limiting factor of OEP might be the changes in the cephalic margin of the zone of apposition, i.e., in the area over which the rib cage is effectively exposed to abdominal pressure (9). If exercise induces sufficient dynamic hyperinflation, the area of apposition of the diaphragm to rib cage might disappear, converting a two-compartment rib cage to a single compartment. Nonetheless, monitoring the upper border of the area of apposition (i.e., the border between RCp and RCa) during exercise by ultrasound has shown that the area is well maintained even in the presence of Starling resistor–induced dynamic hyperinflation (16). In the circumstances of the present investigation, inasmuch as the area of apposition was diminished in patients, the abdominal rib cage region was considerably smaller than normal.
In conclusion, dyspnea, rib cage distortion, and changes in chest wall dimension do not seem to be closely interrelated phenomena during constant-load cycle exercise in COPD patients.
No funding was received. Each author declares no conflict of interest.
The results of the present study do not constitute endorsement by the American College of Sports Medicine.
1. Agostoni E, Mognoni P. Deformation of the chest wall during breathing efforts. J Appl Physiol. 1966; 21: 1827–32.
2. Aliverti A, Cala SJ, Duranti R, et al.. Human respiratory muscle action and control during exercise. J Appl Physiol. 1997; 83: 1256–69.
3. Aliverti A, Stevenson N, Dellacà RL, Lo Mauro A, Pedotti A, Calverley PM. Regional chest wall volumes during exercise in chronic obstructive pulmonary disease. Thorax. 2004; 59: 210–6.
4. Aliverti A, Quaranta M, Chakrabarti B, Albuquerque AL, Calverley PM. Paradoxical movement of the lower ribcage at rest and during exercise in COPD patients. Eur Respir J. 2009; 33: 49–60.
5. American Thoracic Society. Definitions, epidemiology, pathophysiology, diagnosis and staging. Am J Respir Crit Care Med. 1995; 152: S78–121.
6. Binazzi B, Bianchi R, Romagnoli I, et al.. Chest wall kinematics and Hoover’s sign. Respir Physiol Neurobiol. 2008; 160: 325–33.
7. Borg GAV. Psychophysical basis of perceived exertion. Med Sci Sports Exerc. 1982; 14: 377–81.
8. Casaburi R, Porszasz J. Reduction of hyperinflation by pharmacologic and other interventions. Proc Am Thorac Soc. 2006; 3 (2): 185–9. Review.
9. Chihara K, Kenyon CM, Macklem PT. Human rib cage distortability. J Appl Physiol. 1996; 81: 437–47.
10. Duranti R, Sanna A, Romagnoli I, et al.. Walking modality affects respiratory muscle action and contribution to respiratory effort. Pflugers Arch. 2004; 448 (2): 222–30.
11. European Community for Coal and Steel. Standardization of lung function tests. Eur Respir J. 1993; (suppl 16): 1–100.
12. Garcia-Pachon E. Paradoxical movement of the lateral rib margin (Hoover sign) for detecting obstructive airway disease. Chest. 2002; 122: 651–5.
13. Gilmartin JJ, Gibson GJ. Abnormalities of chest wall motion in patients with chronic airflow obstruction. Thorax. 1984; 39: 264–71.
14. Gilmartin JJ, Gibson GJ. Mechanisms of paradoxical rib cage motion in patients with chronic pulmonary disease. Am Rev Respir Dis. 1986; 134: 683–7.
15. Hoover CF. The diagnostic significance of inspiratory movements of the costal margins. Am J Med Sci. 1920; 159: 633–46.
16. Iandelli I, Aliverti A, Kayser B, et al.. Determinants of exercise performance in normal men with externally imposed expiratory flow limitation. J Appl Physiol. 2002; 92: 1943–52.
17. Johnson BD, Weisman IM, Zeballos RJ, Beck KC. Emerging concepts in evaluation of ventilatory limitation during exercise: the excessive tidal flow-volume loop. Chest. 1999; 116: 488–503.
18. Jubran A, Tobin MJ. The effect of hyperinflation on rib cage-abdominal motion. Am Rev Respir Dis. 1992; 146: 1378–82.
19. Kenyon CM, Cala SJ, Yan S, et al.. Rib cage mechanics during quiet breathing and exercise in humans. J Appl Physiol. 1997; 83: 1242–55.
20. Killian KJ, Leblanc P, Martin D, Summers E, Jones NL, Campbell EJM. Exercise capacity and ventilator, circulatory, and symptom limitation in patients with chronic airflow limitation. Am Rev Respir Dis. 1992; 146: 935–40.
21. Lanini B, Bianchi R, Romagnoli I, et al.. Chest wall kinematics in patients with hemiplegia. Am J Respir Crit Care Med. 2003; 168: 109–13.
22. Laveneziana P, Lotti P, Coli C, et al.. Mechanisms of dyspnea and its language in patients with asthma. Eur Respir J. 2006; 27: 742–7.
23. Macklem PT. Exercise in COPD: damned if you do and damned if you don’t. Thorax. 2005; 60: 887–8.
24. Miller MR, Hankinson J, Brusasco V, et al.. Standardisation of spirometry. Eur Respir J. 2005; 26: 319–38.
25. O’Donnell DE, Bain DJ, Webb K. Factors contributing to relief of exertional breathlessness during hyperoxia in chronic airflow limitation. Am J Respir Crit Care Med. 1997; 155: 530–5.
26. O’Donnell DE, Travers J, Webb KA, et al.. Reliability of ventilatory parameters during cycle ergometry in multicentre trials in COPD. Eur Respir J. 2009; 34: 866–74.
27. Olaffson S, Hyatt RE. Ventilatory mechanisms and expiratory flow limitation during exercise in normal subjects. J Clin Invest. 1969; 48: 564–73.
28. Peters MM, Webb KA, O’Donnell DE. Combined physiological effects of bronchodilators and hyperoxia on exertional dyspnoea in normoxic COPD. Thorax. 2006; 61: 559–67.
29. Romagnoli I, Gorini M, Gigliotti F, et al.. Chest wall kinematics, respiratory muscle action and dyspnoea during arm vs. leg exercise in humans. Acta Physiol. 2006; 188: 63–73.
30. Romagnoli I, Gigliotti F, Lanini B, et al.. Chest wall kinematics and breathlessness during unsupported arm exercise in COPD patients. Respir Physiol Neurobiol. 2011; 178: 242–9.
31. Scano G, van Meerhaeghe A, Willeput R, Vachaudez JP, Sergysels R. Effect of oxygen on breathing during exercise in patients with chronic obstructive lung disease. Eur J Respir Dis. 1982; 63: 23–30.
32. Somfay A, Porszasz J, Lee SM, Casaburi R. Dose–response effect of oxygen on hyperinflation and exercise endurance in nonhypoxaemic COPD patients. Eur Respir J. 2001; 18: 77–84.
33. Vogiatzis I, Georgiadou O, Golemati S, et al.. Patterns of dynamic hyperinflation during exercise and recovery in patients with severe chronic obstructive pulmonary disease. Thorax. 2005; 60 (9): 723–9.
34. Vogiatzis I, Stratakos G, Athanassopoulos D, et al.. Chest wall volume regulation during exercise in COPD patients with Gold stages II to IV. Eur Respir J. 2008; 32: 42–52.
35. Wanger J, Clausen JL, Coates A, et al.. Standardisation of the measurement of lung volumes. Eur Respir J. 2005; 26 (3): 511–522.