Dyspnea, Ventilatory Pattern, and Changes in Dynamic Hyperinflation Related to the Intensity of Constant Work Rate Exercise in COPD* Luis Puente-Maestu, MD, PhD; Julia Garcı́a de Pedro, MD, PhD; Yolanda Martı́nez-Abad, MD; José Maria Ruı́z de Oña, MD; Daniel Llorente, MD; and José Manuel Cubillo, MD Study objective: We undertook the present study to investigate the perception of dyspnea (with respect to changes in end-inspiratory and end-expiratory lung volumes), during four different levels of high-intensity constant work rate exercise (CWRE) in patients with severe COPD. Design: Crossover descriptive study with consecutively recruited subjects. Setting: Tertiary university hospital. Patients: Twenty-seven subjects with severe COPD (mean [ⴞ SD] age, 65 ⴞ 5 years of age; mean FEV1, 43 ⴞ 8% predicted; and mean inspiratory capacity [IC]; 74 ⴞ 14% predicted). Measurements and results: Subjects randomly performed four high-intensity CWRE tests (conducted at 65%, 75%, 85%, and 95% of their symptom-limited peak work rate). Dyspnea, leg fatigue, and IC were determined every 2 min during exercise with breath-by-breath gas exchange and ventilatory measurements. There was a good correlation between the resting IC percent predicted and the oxygen uptake (V̇O2) peak (r ⴝ 0.64 to 0.69 between the IC percent predicted and V̇O2 peak at the four work rates). There were significant differences (p < 0.01) in mean respiratory rate (33 ⴞ 6, 35 ⴞ 6, 37 ⴞ 6, and 38 ⴞ 6 min), peak dyspnea score (5.9 ⴞ 1.3, 6.3 ⴞ 1.4, 6.8 ⴞ 1.2, and 6.9 ⴞ 1.6), minute ventilation (45.0 ⴞ 8.7, 43.8 ⴞ 7.7, 43.1 ⴞ 8.7, and 42.8 ⴞ 8.0 L/min), leg fatigue (4.8 ⴞ 1.3, 5.1 ⴞ 1.3, 5.7 ⴞ 1.4, and 5.8 ⴞ 1.4), and end-tidal carbon dioxide partial pressure (4.41 ⴞ 0.36, 4.53 ⴞ 0.33, 4.66 ⴞ 0.31, and 4.76 ⴞ 0.24 kPa), respectively, for tests conducted at 65%, 75%, 85%, and 95% of their symptom-limited peak work rate, and in mean end-expiratory lung volume ([EELV] 4.55 ⴞ 0.44, 4.69 ⴞ 0.43, and 4.79 ⴞ 0.43 L), respectively, for tests conducted at 65%, 75%, and 85% of their symptom-limited peak work rate. In multivariable analysis, the factors that were independently correlated with dyspnea (p < 0.05) were EELV, peak inspiratory flow, and leg fatigue/discomfort. Conclusion: In COPD subjects with flow limitation at rest, the perception of dyspnea increased nonlinearly with the magnitude of high-intensity CWRE in association with a faster respiratory pattern and an increase in EELV. At the highest work rates, it appeared that a reduction in tidal volume and ventilation peak may have limited the tolerance to exercise. (CHEST 2005; 128:651– 656) Key words: constant work rate exercise; COPD; dynamic hyperinflation; dyspnea; flow limitation; ventilatory pattern Abbreviations: CWRE ⫽ constant work rate exercise; EELV ⫽ end-expiratory lung volume; EILV ⫽ end-inspiratory lung volume; IC ⫽ inspiratory capacity; PETCO2 ⫽ end-tidal carbon dioxide partial pressure; RR ⫽ respiratory rate; Ti ⫽ inspiratory time; TLC ⫽ total lung capacity; V̇co2 ⫽ carbon dioxide output; V̇e ⫽ minute ventilation; V̇o2 ⫽ oxygen uptake; Vt ⫽ tidal volume OPD, with its high prevalence worldwide, is a C major medical disorder. In Spain, approximately 10% of the population between 50 and 70 years of age1 has COPD. Patients with COPD are often limited in their activities by the sensation of dyspnea. Reductions in functional status, quality of life, and *From the Hospital General Universitario Gregorio Marañón (Drs. Puente-Maestu, Martı́nez-Abad, and Cubillo), Servicio de Neumologı́a, Madrid, Spain; Hospital Virgen de la Torre (Drs. Garcı́a de Pedro and Ruı́z de Oña), Madrid, Spain; and Hospital La Mancha Centro (Dr. Llorente), Aleázar de San Juan Ciudad Real, Spain. This study was supported by a grant from the Fondo de Investigaciones Sanitarias (grant No. 96/2042) and Sociedad Española de Neumologı́a y Cirugı́a Torácica. Manuscript received February 10, 2004; revision accepted January 7, 2005. Reproduction of this article is prohibited without written permission from the American College of Chest Physicians (www.chestjournal. org/misc/reprints.shtml). Correspondence to: Luis Puente-Maestu, MD, PhD, Hospital General Universitario Gregorio Marañón, Servicio de Neumologı́a, c/o Doctor Ezquerdo 46, 28007 Madrid, Spain; e-mail: lpuente@separ.es www.chestjournal.org Downloaded From: http://journal.publications.chestnet.org/ on 02/02/2016 CHEST / 128 / 2 / AUGUST, 2005 651 disability are frequently consequences of this symptom.2 Dyspnea is also a core feature of panic attacks in patients with COPD.3 Patient “fear of dyspnea” may lead to avoidance of otherwise achievable physical activity and cause additional deconditioning.4 Thus, knowledge of the factors contributing to dyspnea during different intensities and types of exercise will be important to the development of therapeutic interventions and counseling for these patients. It is well-established that dyspnea follows a linear relationship with work rate and oxygen uptake (V̇o2) during incremental exercise in both healthy subjects and patients with respiratory diseases.5 There is little information, however, about the perception of dyspnea when the patients undergo highintensity constant work rate exercise (CWRE). Typically, high-intensity CWRE poses a more abrupt stimulus than incremental tests and entails a faster response of the integrated central and peripheral systems involved in exercise. Therefore, it may be a better model for some daily activities that suddenly require a sizable effort by the legs. Furthermore, CWRE is frequently used in leg training in COPD patients.6 In high-intensity CWRE, V̇o2, minute ventilation (V̇e), and metabolic acidosis inexorably increase to the limit of tolerance, reaching approximately the same end values.7,8 Some investigators8 have described differences in the respiratory frequency with CWRE of different intensities. Such differences may have an impact on dynamic hyperinflation9 and dyspnea.10 Therefore, we undertook the present study to investigate the perception of dyspnea with respect to changes in end-inspiratory lung volume (EILV) and end-expiratory lung volume (EELV) during four high-intensity CWRE tests in severe COPD patients. Materials and Methods Study Design, Study Group, and Settings The present investigation was designed as a crossover study in which each patient randomly performed four high-intensity CWRE tests. The subjects of the study were recruited from a population of patients with severe COPD according to American Thoracic Society guidelines11 with no significant acute reversibility (ie, ⬍ 12% increase in FEV1 15 min after the administration of 400 ␮g of salbutamol) who stopped smoking at least 6 months before enrolling in the study. From this population, those who developed severe desaturation during the preliminary incremental exercise test (⬍ 85%) were excluded. All of the participants signed an informed consent form, and the protocol was approved by the Committee for Ethics in Human Research of our institution. Our center is a tertiary university hospital. Measurements Spirometry, consisting of resting flow-volume loops and singlebreath carbon monoxide diffusion tests, was performed in the 652 Downloaded From: http://journal.publications.chestnet.org/ on 02/02/2016 seated position according to international guidelines12,13 with a gas exchange system (Compactransfer; Jaeger; Hochberg, Germany). The spirometric values used in the study are those obtained 15 min after the inhalation of 400 ␮g of salbutamol. Total lung capacity was measured by plethysmography in a constant-volume chamber (Jaeger). Exercise tests were performed on an electromagnetically braked cycle-ergometer (ER900; Jaeger). Ventilation and pulmonary gas exchange were measured breath-by-breath by a gas exchange system (Quark-b2; Cosmed; Rome, Italy). Patients performed an initial incremental exercise test following a ramp protocol14 at 10 W/min to a symptom-limited maximum. Prior to the CWRE tests in which measurements were made, a practice CWRE test was performed at 65% of the peak work rate that was reached in the incremental test. The test allowed us to confirm that in every subject the V̇o2 changed ⬎ 100 mL from minute 4 to the end, with an average of 141 mL/min. Subsequently, patients performed the four CWRE tests at a pedaling rate of 60 revolutions per minute in a randomized sequence at the four work intensities (ie, 65%, 75%, 85%, and 95% of the peak incremental work rate). The CWRE tests were performed on different days and are designated as 65%, 75%, 85%, and 95%. During the testing, the patient wore a face mask with a dead space of approximately 20 mL. The CWRE tests consisted of 3 min of rest and 3 min of unloaded pedaling, and then the preset work rate was abruptly instituted. The tests were terminated when, after standardized encouragement (ie, “keep up the good work” or “keep on a bit more”), the patients were unable to continue because of symptoms. Breath-by-breath values of ⬎ 3 SD were filtered, and the remaining measurements were collected into five-breath averages. Oxygen saturation was continuously measured by means of a pulse-oximeter (N-180; Nellcor N180; Pleasanton, CA). Before each exercise test, the subjects inhaled two puffs (200 ␮g) of salbutamol, and the inspiratory capacity (IC) maneuver was demonstrated to the subjects. They were specifically instructed after the verbal signal, and at the end of the next normal breath out, they were to continue the next breath until the lungs were full and then try to give an extra effort to fill them up even more. They were asked to do this fairly quickly so as not to interrupt breathing for very long. The maneuver ended with a normal, unforced exhalation. During exercise, the subjects were prompted to perform the inspiratory IC and received verbal encouragement to inspire maximally.15,16 Linear regression of the EELV as function of time for five breaths (10 to 15 s) preceding the IC prompt was used to calculate the drift slope of EELV.16 The volume was integrated from the flow signal. Predicted values for IC were calculated from the predicted values for total lung capacity (TLC) minus the predicted values for functional residual capacity. Because there were increases with marked changes in tidal volume (Vt) and respiratory rate (RR) during the first 3 min of high-intensity CWRE ventilation, we started the IC maneuvers after the third minute and every 2 min thereafter. IC could not be measured reliably in the 95% work-rate constant exercise tests, because the test duration was only 3 min in many patients. EELV was derived as TLC IC, and EILV as EELV ⫹ Vt. Dyspnea and leg fatigue were measured every 2 min using a numerical scale of 0 to 10 with descriptors (modified from the Borg scale)17 just before the IC maneuver. Statistical Analysis Interval variables are summarized as the mean ⫾ SD within parentheses, unless otherwise specified. Comparisons between mean responses for the different tests were made by repeated analysis of variance measurements. Incremental test results were considered as one additional level. Post hoc pairwise contrasts for the “work-rate- level” factor were made by the Tukey method.18 Clinical Investigations A p value of ⬍ 0.05 for detecting type I error was considered to be significant. Correlations were sought with the linear correlation coefficient and multiple linear correlation coefficients. Variables were selected in a stepwise forward fashion for the multiple model. These analyses were performed using a statistical software package (SPSS 9.0, Hispanoportugesa SPSS; Madrid, Spain). Results Twenty-seven patients participated in this study, and their physical characteristics are presented in Table 1. As a group, they had a mean body weight slightly below normal, with moderate hypoxemia and no CO2 retention at rest. The patients had moderate hyperinflation and moderately reduced diffusing capacity of the lung for carbon monoxide. They also had moderately impaired aerobic capacity. Nineteen of the 27 patients (70%) had an IC that was ⬍ 80% predicted. Dyspnea followed a different time course during the CWRE tests than that observed during the incremental exercise tests, corresponding with the rate of increase in metabolic load and ventilation (Fig 1). Analysis of variance demonstrated that the peak scores of dyspnea were statistically significantly different among the CWRE tests (Table 2). Post hoc analysis showed significant differences (p ⬍ 0.05) between the scores on the 65% test and those on the other three levels and the scores on the 75% test and the other three levels, but the dyspnea scores at the end of the 85% and 95% tests were not significantly different. Dyspnea measured at the end of the incremental exercise test was significantly different from dyspnea at the end of the 65%, 85%, and 95% tests (p ⫽ 0.04, 0.011, and ⬍ 0.001, respectively) but Table 1—Description of the Sample (n ⴝ 27)* Variables Mean SD Age, yr BMI, kg/m2 FEV1, L % predicted FEV1/VC, % Pao2, kPa Paco2, kPa TLC, L % predicted RV/TLC, % Dlco, % predicted IC, % predicted V̇o2peak, % predicted WRpeak, W Dyspnea Leg fatigue 65.2 23.1 1.2 42.5 47.0 8.4 5.1 6.68 108.7 59.3 61.5 74.4 66.5 96 6.7 5.3 4.5 3.2 0.2 8.3 5.6 0.5 0.5 0.65 10.1 4.3 8.5 16.5 13.7 24 1.3 1.0 *BMI ⫽ body mass index; VC ⫽ vital capacity; V̇o2peak ⫽ peak V̇o2; WRpeak ⫽ peak work rate; RV ⫽ residual volume. www.chestjournal.org Downloaded From: http://journal.publications.chestnet.org/ on 02/02/2016 Figure 1. Time course of dyspnea during the four CWRE tests and the incremental exercise test. Maximum dyspnea scores were greater as the intensity of the CWRE tests increased. Time course of dyspnea was clearly different between the CWRE and the incremental exercise tests. not from the 75% test. Dyspnea at the end of the CWRE tests correlated with most of the variables characterizing the respiratory pattern and the ventilatory mechanical load (ie, RR, Vt, EILV, and EELV), as shown in Table 3. However, in a multiple first-order linear regression model, EELV as a percentage of measured TLC (p ⬍ 0.001), peak inspiratory flow (p ⬍ 0.001), and leg fatigue/discomfort (p ⬍ 0.001) remained significant. V̇e peak was slightly lower at the two highest work rates, but the differences were only statistically significant (p ⬍ 0.05) between the 65% CWRE test and both the 85% and 95% CWRE tests. Neither the V̇e peak of the 75% constant test nor the V̇e peak of the incremental test were different from those in the other four tests. The mean (⫾ SE) CO2 output (V̇co2) peak was decreased in proportion to ventilation (r ⫽ 0.93 ⫾ 0.37). A trend for increasing endtidal carbon dioxide partial pressure (PETCO2) with the intensity of the test was seen (Table 2), and the differences reached statistical significance (p ⬍ 0.001) between the 65% and 95% CWRE tests. The pattern of ventilation was found to be significantly different among the CWRE tests (p ⬍ 0.001), with a lower Vt and more rapid RR at the more intense work rates. Both RR and Vt reached statistical significance (p ⬍ 0.05) between the results of the 65% CWRE test and both the 85% and 95% CWRE tests, and between the 75% and 95% CWRE tests. Differences between the incremental and the 95% CWRE test were also significant. The reduction in Vt was almost entirely caused by an increase in EELV, with little change in EILV, as can be seen in Table 2. The mean peak EILV values were 88.1 ⫾ 3.5%, 88.6 ⫾ 4.4%, and 88.8 ⫾ 4.6%, whereas CHEST / 128 / 2 / AUGUST, 2005 653 Table 2—Changes in Exercise Response Related to the Intensity and Type of Exercise* Incremental Test 65% Test 75% Test 85% Test 95% Test Variables Units Mean (SD) Mean (SD) Mean (SD) Mean (SD) Mean (SD) p Value Dyspnea Leg fatigue Duration WR V̇epeak V̇o2peak V̇co2peak PETCO2 RR Vt Vt/Ti Ti/Ttot ratio EILV EILV/TLCm ratio EELV EELV/TLCm ratio Borg Borg min W L/min L/min L/min kPa min L L/s 6.2 5.3 9.4 94 43.7 1.23 1.28 4.53 34.7 1.30 2.00 0.38 (1.3) (1.0) (2.4) (24) (7.7) (0.22) (0.26) (0.64) (6.6) (0.32) (0.51) (0.08) (1.3) (1.3) (4.5) (17.5) (8.7) (0.26) (0.26) (0.36) (5.6) (0.36) (0.60) (0.09) (0.56) (3.5) 6.2 5.1 6.9 73.0 43.8 1.22 1.26 4.53 34.9 1.29 2.04 0.37 5.99 88.6 (1.4) (1.3) (1.9) (19.2) (7.7) (0.23) (0.23) (0.33) (6.4) (0.31) (0.40) (0.07) (0.54) (4.4) 6.8 5.7 4.8 81 43.1 1.21 1.24 4.66 36.6 1.21 2.01 0.35 6.01 88.8 (1.2) (1.4) (1.3) (20.8) (8.7) (0.23) (0.24) (0.31) (5.8) (0.32) (0.51) (0.05) (0.61) (4.6) 6.9 5.8 3.6 90 42.8 1.19 1.22 4.76 38.4 1.15 1.98 0.34 (1.6) (1.4) (1.16) (22.8) (8.0) (0.27) (0.23) (0.24) (5.8) (0.28) (0.44) (0.09) L % 5.9 4.8 12.3 64.3 45.0 1.26 1.30 4.41 33 1.41 2.22 0.39 5.95 88.1 ⬍ 0.001 0.018 ⬍ 0.001 ⬍ 0.001 ⬍ 0.001 0.069 ⬍ 0.001 ⬍ 0.006 ⬍ 0.001 ⬍ 0.001 ⬍ 0.001 ⬍ 0.001 0.526 0.530 L % 4.55 67.2 (0.44) (2.9) 4.69 69.4 (0.43) (3.7) 4.79 71.0 (0.43) (3.8) ⬍ 0.001 ⬍ 0.001 *WR ⫽ work rate; V̇epeak ⫽ peak V̇e; V̇co2peak ⫽ peak V̇co2; Ttot ⫽ total breathing cycle time; TLCm ⫽ measured TLC; WR ⫽ work rate. See Table 1 for abbreviations not used in the text. the peak EELV values were 67.2 ⫾ 2.9%, 69.4 ⫾ 3.7%, and 71.0 ⫾ 3.8% of the measured TLC, respectively, for the 65%, 75%, and 85% constant tests. The differences in peak EELV (but not in peak EILV) were statistically significant (p ⬍ 0.05) for each of the three work-rate levels. There were good correlations between the Vt peak and the V̇o2 peak (r ⫽ 0.74 ⫾ 0.13, 0.80 ⫾ 0.12, 0.72 ⫾ 0.14, 0.71 ⫾ 0.14, and 0.79 ⫾ 0.12, respectively, for the 65%, 75%, 85%, and 95% CWRE tests, and the ramp tests). This was also observed for the mean IC percent predicted at rest and with peak V̇o2 percent predicted (0.65 ⫾ 0.15, 0.69 ⫾ 0.14, 0.66 ⫾ 0.14, 0.63 ⫾ 0.15, and 0.68 ⫾ 0.14, respectively, for the 65%, 75%, 85%, 95% CWRE tests, and the ramp tests). The perception of leg fatigue/discomfort (Table 2), as with dyspnea, was greater as the work-rate inten- Table 3—Univariate Correlations of the Ventilatory Variables With Dyspnea at the Constant Work Rate Tests* Variables Units r Value SE No. p Value V̇epeak RR Vt Ti/Ttot Vt/Ti EILV/TLCm EELV/TLCm Leg fatigue L/min min L 0.327 0.591 ⫺ 0.437 ⫺ 0.439 0.419 0.442 0.715 0.375 (0.092) (0.078) (0.087) (0.087) (0.088) (0.102) (0.079) (0.090) 108 108 108 108 108 81 81 108 ⬍ 0.001 ⬍ 0.001 ⬍ 0.001 ⬍ 0.001 ⬍ 0.001 ⬍ 0.001 ⬍ 0.001 ⬍ 0.001 L/s % % Borg *See Table 2 for abbreviations not used in the text. 654 Downloaded From: http://journal.publications.chestnet.org/ on 02/02/2016 sity increased (p ⬍ 0.001). Paired comparisons showed significant statistical differences (p ⬍ 0.05) between the mean scores of each test and those of the other four tests, except for the results of the 85% CWRE test compared with those of the 95% CWRE test (p ⫽ 0.84), and for the results of the incremental test compared with those of the 75% CWRE test (p ⫽ 0.98). Peak leg fatigue measurements were higher than dyspnea measurements at the end of the CWRE tests in 10 patients (37.0%), 10 patients (37.0%), 11 patients (40.7%), and 13 patients (48.1%), respectively, for the 65%, 75%, 85%, and 95% CWRE tests. Discussion There were several relevant findings in this study of CWRE in patients with severe COPD. First, patients stopped the test at peak V̇e levels that decreased in a progressive, statistically significant fashion. In contrast, the perception of dyspnea at the end of exercise progressively increased with the intensity of the CWRE. The relationship between dyspnea and V̇e was, therefore, different for each CWRE intensity. Furthermore, if we consider the set of four work-rate peak dyspnea points obtained from the CWRE tests, dyspnea increased nonlinearly with the work rate, in striking contrast with the linear manner described during the incremental ramp exercise protocols in the literature5 and observed as well in our subjects (Fig 1). Second, compared with the Clinical Investigations progressive exercise test, we found that the breathing pattern was more shallow and faster at the most intense work rates. Finally, we report that the perception of dyspnea was mainly correlated with EELV; however, the Vt/inspiratory time (Ti) and the leg fatigue/discomfort also showed smaller but statistically significant (p ⬍ 0.05) independent correlations. The reasons for the observed nonlinear increase of dyspnea with respect to V̇e at the different work rates cannot be fully elucidated from our study. We found an excellent correlation between dynamic hyperinflation (ie, increase in EELV) and dyspnea, and it is well known that dynamic lung hyperinflation is an important contributor to the intensity and quality of dyspnea.19,20 The increase in EELV may be the corollary of tachypnea in the severe COPD patient9; thus, differences in dyspnea could be considered secondary events that are associated with more rapid breathing. Several mechanisms could have played some role in the observed differences in respiratory pattern at the different CWRE intensities,21 such as sensorial inputs from the legs, lungs, or respiratory muscles secondary to more rapid accumulation of metabolic by-products, more pronounced stretching, faster changes in shape,22 differences in catecholamine blood levels,23 or chemical stimuli, such as hypercapnia or lactic acidemia.24,25 For example, patients with pulmonary disease have been shown to generate more tachypnea in response to CO2 compared with healthy subjects.26,27 We did not directly measure Paco2, pH, or lactate levels, because the manipulations associated with such measurements could have disturbed the breathing pattern and influenced the maximum exercise tolerance, but PETCO2 was significantly higher in the more intense tests. The actual differences in Paco2 may have been even higher, because the gap between Paco2 and PETCO2 becomes more pronounced as the RR increases.28 Another finding in this study was that leg fatigue was independently correlated with dyspnea. To our knowledge, there has not been previous reports of a possible cross-effect between these two sensations; however, other distressing sensations, such as pain, have been shown to augment dyspnea.29 An alternative way to explain the association between dyspnea and EELV is that both were not related causally, each one being an epiphenomenon of an underlying common cause, such as inspiratory muscle fatigue or the failure to rapidly increase V̇e to match the increasing metabolic load (ie, V̇co2) in the presence of airflow limitation. We also cannot ignore the possibility that inspiratory muscles fatigued or approached fatigue at the highest work rates. The tension-time index might have reached or approached the fatigue point as EELV increased www.chestjournal.org Downloaded From: http://journal.publications.chestnet.org/ on 02/02/2016 (EELV increases are associated with augmented elastic load and decreased ability to generate peak force because of the shortening of inspiratory muscle fibers) and the Ti/total time was reduced. Diaphragmatic fatigue has not been consistently found in COPD after progressively (ie, not of sudden onset) increasing exercise,30 but it could be that the sensation of dyspnea acts as a protective mechanism, reaching intolerable levels before the muscles actually fatigue. An alternative explanation to inspiratory muscle fatigue is that the ventilatory system of these subjects could not respond as needed to the metabolic demands. Several types of evidence point to inspiratory and expiratory flow limitation in these subjects. We found a good correlation between resting IC percent predicted and V̇o2 peak, a characteristic finding of COPD patients with flow limitation;31 furthermore, PETCO2 increased with exercise. The increase in Paco2 (for which PETCO2 is a surrogate) is deemed as another indicator of flow limitation32 resulting from a decreased efficiency to excrete CO2 by the lungs when Vt is smaller, and, consequently, physiologic dead space/Vt ratio is larger for a given V̇e and V̇co2. Finally, as can be seen in Table 2, the maximal Vt/Ti ratio was about the same at all of the work rates, likely indicating that the “ceiling” inspiratory flow had been attained. An unexpected finding of the present study was the slight but significant decline in V̇e at the higher CWRE test associated with a reduction of Vt. Vt was very well correlated with V̇o2 suggesting that it may play an important role in the decision to terminate exercise. Breathing models predict that subjects with flow obstruction of the magnitude observed in our subjects may have a decrease in Vt and consequently of V̇e because of dynamic hyperinflation when RRs are high.9 Diaz et al.32 also postulated that in COPD patients with inspiratory flow limitation during exercise, tachypnea would make it impossible to additionally reduce the inspiratory reserve volume in the available Ti, that is, the ceiling lung volume would be smaller than the TLC. In support of this concept, we found (Table 2) that EILV reached similar submaximal values compared with the TLC in spite of the increasing EELV. Other factors might have played some role as well, for example, early peripheral muscle fatigue33 or more rapid attainment of intolerable leg discomfort at higher work rates. But, as mentioned above, several findings clearly pointed to flow limitation and the associated mechanical impairment as major contributing factors. In conclusion, the perception of dyspnea experienced by COPD subjects in our study worsened nonlinearly with the magnitude of sudden onset, high-intensity CWRE. This appeared to be casually related to changes in the respiratory pattern and CHEST / 128 / 2 / AUGUST, 2005 655 increased dynamic hyperinflation. The reasons for faster breathing at the highest work might have been of sensory or cortical origin or just an adjustment to the failed attempt to increase ventilation to match an abrupt and overwhelming metabolic load (ie, V̇co2) when expiratory and inspiratory flow limitation were present. In any case, this lead to an increase in EELV and a reduction in Vt and V̇e. Such behavior hampered the ventilatory response at the highest work rates, apparently limiting exercise tolerance. Our results contribute to a better understanding of the physiology of high-intensity CWRE exercise in COPD patients and may help determine the best target work rate intensity for pulmonary rehabilitation, as well as provide scientific support for counseling these patients on how to better tolerate activities of daily living that require sudden onset, sustained, high-intensity CWRE. 14 15 16 17 18 19 20 ACKNOWLEDGMENT: We thank Dr. William W. Stringer for his help with the manuscript. This study was funded by Fondo de Investigaciones Sanitarias Grant 96/2042 and by the Spanish Pulmonary and Thoracic Surgery Society. References 1 Sobradillo V, Miravitlles M, Jimenez CA, et al. Epidemiological study of chronic obstructive pulmonary disease in Spain (IBERPOC): prevalence of chronic respiratory symptoms and airflow limitation. Arc Bronconeumol 1999; 35:159 –166 2 Jones NL, Killian KJ. Limitation of exercise in chronic airway obstruction. In: Cherniack NS, ed. Chronic obstructive pulmonary disease. Philadelphia, PA: WB Saunders, 1991; 196 –207 3 Anderson DJ, Noyes R, Jr, Crowe RR. A comparison of panic disorder and generalized anxiety disorder Am J Psychiatry 1984; 141:572–575 4 Agle DP, Baum GL. Psychological aspects of chronic obstructive pulmonary disease. Med Clin North Am 1977; 61:749 –758 5 Killian KJ. The objective measurment of breathlessness. Chest 1985; 88:84S–90S 6 Ries AL, Carlin BW, Carrieri-Kohlman V, et al. Pulmonary rehabilitation: joint ACCP/AACVPR evidence-based guidelines. Chest 1997; 112:1363–1396 7 Vandewalle H, Vautier JF, Kachouri M, et al. Work-exhaustion time relationships and the critical power concept: a critical review. J Sports Med Phys Fitness 1997; 37:89 –102 8 Puente-Maestu L, Santacruz A, Vargas T, et al. Effects of training on the tolerance to high-intensity exercise in patients with severe COPD. Respiration 2003; 70:367–370 9 Younes M. Determinants of thoracic excursion during exercise. In: Exercise pulmonary physiology and physiopathology. Whipp BJ, Wasserman K, eds. New York, NY: Marcel Dekker, 1991; 1– 65 10 O’Donell DE, Webb KA. Exertional breathlessness in patients with chronic airflow limitation: the role of lung hyperinflation. Am J Respir Crit Care Med 1993; 148:1351–1357 11 American Thoracic Society. Standards for the diagnosis and care of patients with chronic obstructive pulmonary disease (COPD) and asthma. Am Rev Respir Dis 1987; 134:239 –243 12 American Thoracic Society. Standardization of spirometry: 1994 update. Am J Respir Crit Care Med 1995; 152:1107–1136 13 American Thoracic Society. Single breath carbon monoxide diffusing capacity (transfer factor): recommendations for a 656 Downloaded From: http://journal.publications.chestnet.org/ on 02/02/2016 21 22 23 24 25 26 27 28 29 30 31 32 33 standard technique; 1995 update. Am J Rrespir Crit Care Med 1995; 152:2185–2198 Whipp BJ, Davis JA, Torres F, et al. A test to determine parameters of aerobic function during exercise. J Appl Physiol 1981; 50:217–221 Yan S, Kaminski D, Sliwinski P. Reliability of inspiratory capacity for estimating end-expiratory lung volume changes during exercise in patients with chronic obstructive pulmonary disease. Am J Respir Crit Care Med 1997; 156:55–59 Dolmage TE, Goldstein RS. Repeatability of inspiratory capacity during incremental exercise in patients with severe COPD. Chest 2002; 121:708 –714 Borg GAV. Psychophysical bases of perceived exertion. Med Sci Sports Exerc 1982; 14:377–381 Kleinbaum DG, Kupper LL, Muller KE. Applied regression analysis and other multivariable methods. Boston, MA: PWSKent, 1987:365–368 O’Donell DE, Bertley JC, Chau KL et al. Qualitative aspects of exertional breathlessness in chronic airflow limitation: pathophysiological mechanisms. Am Rev Respir Crit Care Med 1997; 155:109 –115 Marı́n JM, Carrizo JS, Gascón M, et al. Inspiratory capacity, dynamic hyperinflation, breathlessness, and exercise performance during 6-minute-walk test in chronic obstructive pulmonary disease. Am J Respir Crit Care Med 2001; 163:1395–1399 Schartzstein RM, Manning J, Weiss JW, et al. Dyspnea: a sensory experience. Lung 1990; 168:185–199 Adams L, Guz A. Dyspnea on exertion. In: Exercise pulmonary physiology and physiopathology. Whipp BJ, Wasserman K, eds. New York, NY: Editorial Marcel Dekker, 1991; 449 – 494 Clark AL, Galloway S, MacFarlane N, et al. Catecholamines contribute to exertional dyspnoea and to the ventilatory response to exercise in normal humans. Eur Heart J 1997; 18:1829 –1833 Cowley DS, Arana GW. The diagnostic utility of lactate sensitivity in panic disorder. Arch Gen Psychiatry 1990; 47:277–284 Gorman JM, Asknazi J, Liebowitz MR et al. Response to hyperventilation in a group of patients with panic disorder. Am J Psychiatry 1994; 151:547–553 Gorman JM, Papp LA, Coplan JD, et al. Anxiogenic effects of CO2 and hyperventilation in patients with panic disorders. Am J Psychiatry 1994; 151:547–553 Perna G, Battaglia M, Garberi A, et al. Carbon dioxide/oxygen challenge test in panic disorder. Psychiatry Res 1994; 52:159 – 171 Whipp BJ. The bioenergetics and gas exchange basis of exercise testing. Clin Chest Med 1994; 15:173–193 Nishino T, Shimoyama N, Ide T, et al. Experimental pain augments experimental dyspnea, but not vice versa in human volunteers. Anesthesiology 1999; 91:1633–1638 Mador MJ, Kufel TJ, Pineda LA, et al. Diaphragmatic fatigue and high-intensity exercise in patients with chronic obstructive pulmonary disease. Am J Respir Crit Care Med 2000; 161:118 –123 Diaz O, Villafranca C, Ghezzo H, et al. Role of inspiratory capacity on exercise tolerance in COPD patients with and without tidal expiratory flow limitation at rest. Eur Respir J 2000; 16:269 –275 Diaz O, Villafranca C, Ghezzo H, et al. Breathing pattern and gas exchange at peak exercise in COPD patients with and without tidal flow limitation at rest. Eur Respir J 2001; 17:1120 –1127 Mador JM, Kufel TJ, Pineda LA. Quadriceps and diaphragmatic function after exhaustive cycle exercise in the healthy elderly. Am J Respir Crit Care Med 2000; 162:1760 –1766 Clinical Investigations