Reviewer #1 – Didier Saey (Comments to the Author (Required)): Reviewer’s comment Major Comment: 1. As claimed by authors, the aim of this study is to investigate the determinants of functional exercise performance of the CET and 6MWT in a large clinical cohort of COPD patients. This not exactly reflected in the title which focuses only on Performance of cycle ergometry at 75% of peak work rate. Response: We would like to take the opportunity to thank the reviewer for making useful and constructive comments that have improved our manuscript. Following your suggestion we changed the title of the manuscript to “Characteristics and determinants of endurance cycle ergometry and six-minute walk distance in patients with COPD” Reviewer’s comment Major Comment: 2. Moreover the manuscript, result and discussion focus more on gender effect that some others relevant determinant highlighted by the multiple stepwise regression. By example, there is no discussion concerning the importance of GOLD stage (coefficient =-42.09 when CET-Tend was used as the dependent variable and only 15.04 when 6MWD was used as the dependent variable). This is contrast with the coefficient for FEV1 which was only 1.28 when 6MWD was used as the dependent variable and was not retained in the model of CET. It could be relevant to strength a little bit this part of data. Response: We agree with the reviewer that the influence of GOLD stage in the CET and 6MWT performance is important and should be described further. Indeed, disease severity by 1 GOLD stages demonstrated a higher impact to the performance of CET compared to 6MWT, however, the coefficient for FEV1 was significant for the 6MWD but in the CPET was not retained. We observed that the majority (71%) of patients was able to cycle up to 5 minutes (Figure 1). Obviously, the overall physiological stress during the CET seems to be greater for the most of patients compared to the 6MWT and might be affected by an overall multidimensional affliction of COPD more than it might be affected by the inter-individual variation of FEV1 amongst the GOLD stages. Therefore, the 6MWT might be more responsive than the CET in the interindividual variation of FEV1 across the GOLD stages and this could a possible explanation for the presence of FEV1 coefficient in the model of 6MWD. Accordingly, we have added an additional paragraph in the discussion (P13 paragraph 4) which provides an interpretation from a physiological perspective. “CET-Tend seems to be influenced in a high extent from the degree of disease severity classified by GOLD stages more than the inter-individual variation of FEV1 or other specific clinical characteristics compared to 6MWT. It is observed that patients across the GOLD stages are characterized by increased physical deconditioning which leads to a progressive deterioration in lower limb muscle function.[28] The impact of the power/duration relationship on the exercise tolerance indicates that endurance is influenced by the intensity of exercise.[29] Both exercise tests are sub-maximal, however, the 6MWT is a self-paced while the CET at 75% is a constant load exercise test. The termination of exercise in CET is more associated to the inability to maintain the constant work load of exercise and thus CET-Tend is associated to an overall impaired health status, sex of patients and physical deconditioning more than specific lung function characteristics compared to the 6MWT.” Reviewer’s comment Major Comment: 3. In the same line, it is also a source of debate whether (and in which extent) the absolute intensity is a determinant of CETend. It could be relevant to provide some 2 data concerning relationship or (quartile gradation) between absolute exercise intensity (in watt or oxygen consumption, by example) observed in the present study. Response: We agree with the reviewer. Accordingly, we have added a quartile gradation of the peak power output in absolute Watt for total studied population (Figure 3), males (Online Figure 1) and females (Online Figure 2). Also, we referred on the absolute Watt quartile gradation in the paragraph of Examination of performance quartiles (P10). Reviewer’s comment Major Comment: 4. Examination and analyze of performance quartile seems over-considered throughout the manuscript Response: We agree. Therefore, the paragraph of “Examination of performance quartiles” (P.10) has been more comprehensive. Reviewer’s comment Major Comment: 5. Some data presented in examination performance quartile (P10) are somewhat redundant with data presented in figure 3. Consider summarize the main findings in the text. This would facilitate the reading and understanding of this section. Response: We agree. The paragraph of “Examination of performance quartiles” (P.10) is now more comprehensive in order to facilitate the reading and understanding of this section. 3 Reviewer’s comment Minor Comments: There are some mistakes in the text, by example: 1. P5: I'm not sure that reference 12 is appropriate 2. P10: It is claimed that age, gender, BMI, FEV1%, FVC%, FVC/FEV1, DLCO, GOLD stage and leg fatigue Borg score at rest were retained in the model when 6MWD was the dependent variable". However, table 4 and discussion show that Gender and FVC% were not retained in the model but FRC % yes. Please check carefully if there are no others throughout the manuscript 3. P6 L11, please specify that was during the progressive maximal cardiopulmonary exercise test. 4. P11 The statement that space of improvement may be greater in CET compared to 6MWT after PR is speculative 5. P11 L13 second paragraph: In fact, from the regression model, it seems that gender is not a determinant of six-minute walk distance 6. P11 L15 (second paragraph): men were not only older but also more severe compare to women. This is an important point and it should be emphasis in the discussion. Response: Thanks the reviewer for the suggestions. Accordingly, we have provided the following changes: 1. The reference 12 (which corresponds to reference 20 after revision) has been replaced with the following reference. P6: Ref. 20 “Spruit MA, Vanderhoven-Augustin I, Janssen PP, Wouters EF: Integration of pulmonary rehabilitation in COPD. Lancet 2008, 371:12-13” 2. We have updated the data description of table 4 and checked for others inconsistencies throughout the manuscript. P11 second paragraph: “Age, BMI, FFMI, FEV1%, FVC/FEV1, GOLD stage, FRC, DLCO, and leg fatigue Borg score at rest were retained in the model when 6MWD was the dependent variable (Table 4).” 4 3. We have specified the cardiopulmonary exercise test as incremental and maximal. Methods P7 first paragraph: “.....reached a physiological maximum during the incremental maximal cardiopulmonary exercise test...” 4. We have restated with the following sentence; P15 second paragraph “Moreover, in daily life the majority of patients have been adapted to walking but not to cycling exercise and that may results in a greater potential of improvement in CET compared to 6MWT after PR.” 5. We have corrected the sentence and stated that; P12 L1: “Gender appears to be a strong determinant of CET performance in contrast to 6MWD.” 6. We edit the second paragraph as follows: P12 first paragraph: “Gender appears to be a strong determinant of CET-Tend in contrast to 6MWD. We found that men on average achieved a higher peak work rate and cycled longer during the CET than women; in spite of the fact the men were older with more impaired lung function (Table 1 and 2). Previous studies have demonstrated that quadriceps muscle strength[22] and peak VO2[23, 24] have small reductions in COPD patients over a 1 to 5 year period. Moreover, quadriceps muscle strength is positively related to DLCO in patients with COPD[25] and the studied male population had a greater diffusing capacity which could be accountable for the better cycle performance.” 5 Reviewer #2 – Scotty Butcher (Comments to the Author (Required)): Reviewer’s comment Major Comment: 1. Background: much of the introduction is focused on the difference in endurance with an 81% max test vs 75%, however this does not reflect the methods/hypotheses of the study. Gender is not mentioned, but is a large part of the analysis. In addition, there should be a more thorough comparison of cycle vs walking tests to set up the hypothesis. The introduction should be re-written to include background related to the purposes: ie. the gender differences and performance times of endurance testing. Response: We appreciate your comments and we would like to thank you very much for the opportunity you provide us to improve our manuscript. We agree that the introduction should include gender differences and a more thorough comparison of cycle vs walking test. Therefore, we added the following paragraph into the introductory; Background P5 third paragraph: “Constant work rate endurance protocols in CET are based on the maintenance of exercise effort until the exhaustion.[14] Although CET is considered as sub-maximal exercise, it is aimed to assess the maximal aerobic capacity of patients and it is suggested as a highly repeatable test across run-in visits.[15] However, the utility of CETs at 75% peak work rate comes into doubt as more patients would cycle to the obligatory stopping point of 20 minutes.[10] Similarly, the 6MWT even though selfpaced, it can be considered as a maximal sustainable exercise in patients with COPD[16]. Both testing modalities for assessing exercise tolerance may provoke different physiological responses and cause different levels of dyspnea and fatigue during the exertion to patients with COPD.[17, 18] Furthermore, the sex of patients has been related to dyspnea in the 6MWT [19] and may also determine the exertion during the CET. From a physiological point of view, CET and 6MWT may not interchangeable exercise tests and the determinants of exercise performance between the tests can differ.” 6 Reviewer’s comment Major Comment: 2. Discussion: in reference to the first comment, the first major paragraph of the discussion doesn't really relate to the results of the study and should be moved to a supplementary paragraph later in the discussion (with less emphasis). Please focus the main first paragraph on interpreting the results of your data in reference to your hypothesis. Response: We fully appreciate your suggestion. We placed a new main first paragraph which focuses on interpreting the results of our data in reference to our hypothesis while we moved as 3rd paragraph in the discussion and shorten the previous first paragraphs. The new main first paragraph of the discussion (P11) is the following: “In current study, we aimed to investigate the characteristics of functional exercise performance and the determinants which can influence the exercise performance in CET and the 6MWT in a large cohort of patients with moderate to severe COPD (Table 1). The main finding of this study is that CET-Tend can be determined by different physiological and clinical characteristics in comparison with the 6MWT. Of particular clinical importance, the sex of patients influenced CET-Tend but not in the 6MWD (Table 2-4). Moreover, 6MWD seems to be more determined by lung function compared to CET-Tend (Tables 3-4). Also, a number of independent physiological and clinical characteristics demonstrated a different pattern in performance quartile gradation between CET and 6MWT (Figure 3).” We moved and shorten the previous first paragraph (P14 current 5th paragraph). After modification, the paragraph is the following: “In contrast to the findings of Neder and colleagues [10], we found that the vast majority of patients with COPD did not reach the 20 minute time point during a CET at 75% of peak work rate. This discrepancy between two studies is likely the result of differences in methodology and study populations. We retrospectively examined a 7 clinical cohort of patients who were referred for pulmonary rehabilitation, whereas Neder and colleagues [10] studied a proscribed, smaller cohort with the intent of defining the power duration asymptote. In the present study, CET-Tend varied from 2 to 5min for the most of the patients (Figure 1A). This is similar to the findings of Vivodtzev et al. [12] where most of the patients sustained a CET duration of 3 to 5 minutes. Although the proportion of patients with CET-Tend of 20 minutes was very small (2.5%), we evaluated the possibility of an underestimation in peak work rate that may result in unduly prolonged CET. Nevertheless, peak heart rate, ventilation and dyspnea scores following a cardiopulmonary exercise test indicated a maximum effort for those patients (Online table 3).” Reviewer’s comment Major Comment: 3. Discussion: in the last paragraph ("in summary..."), the first sentence is inappropriate. The mere fact patients can perform the test does not indicate that it is appropriate. Appropriateness would imply there was some sort of intervention or analysis that it reflects some other measure. This was not demonstrated in this study, so should be reworded. Response: The reviewer is right. We have changed the first sentence as follow: P18 last paragraph “In summary, CET at 75% of peak work rate appears to be a useful clinical exercise tool in COPD.” 8 Reviewer’s comment Major Comment: 4. Discussion and Conclusions: Again, the authors are referring to previous data on the appropriateness of the CET test from previous work; however, their direct conclusions are not appropriate for the results of the work presented here. Could they not focus specifically on the regression analysis and the gender differences and provide an interpretation from a physiological perspective? Response: We fully appreciate the concern of the reviewer regarding the physiological perspective of exercise tolerance. Therefore, we added the following paragraph in the discussion (PP 13-14). “CET-Tend seems to be influenced in a high extent from the degree of disease severity classified by GOLD stages more than the inter-individual variation of FEV1 or other specific clinical characteristics compared to 6MWT. It is observed that patients across the GOLD stages are characterized by increased physical deconditioning which leads to a progressive deterioration in lower limb muscle function.[28] The impact of the power/duration relationship on the exercise tolerance indicates that endurance is influenced by the intensity of exercise.[29] Both exercise tests are sub-maximal, however, the 6MWT is a self-paced while the CET at 75% is a constant load exercise test. The termination of exercise in CET is more associated to the inability to maintain the constant work load of exercise and thus CET-Tend is associated to an overall impaired health status, sex of patients and physical deconditioning more than specific lung function characteristics compared to the 6MWT.” Additionally, we added the sentence below in the conclusions (P19): “CET and 6MWT are not interchangeable exercise tests.” 9 Reviewer’s comment Minor Comments: 1. Title: I think the title could be more reflective of the purpose of the study i.e. adding "performance and determinants of endurance tests" rather than specifying CET at 75% which doesn't also include the 6MWT. My suggestion would be something like: "Performance and determinants of endurance exercise tests in patients with COPD". 2. Abstract: the purpose focuses only on 'determinants' but the results are very performance based. I would add "performance and" prior to determinants. 3. Abstract: there should be a comma between "others" and "respiratory" in the second sentence of the methods. Response: Thank you for these suggestions. Accordingly, we have applied changes as follow: 1. We have changed the title of the manuscript to “Characteristics and determinants of endurance cycle ergometry and six-minute walk distance in patients with COPD” 2. We have changed the sentence of the aim of the study in the abstract as follows: “We sought to investigate the characteristics of functional exercise performance and determinants of the CET and 6MWT in a large clinical cohort of COPD patients.” 3. We have put a comma between "others" and "respiratory" in the second sentence of the methods: “Patients underwent, amongst others, respiratory function evaluation; ...” 10 Reviewer #3 – Pierantonio Laveneziana (Comments to the Author (Required)): Reviewer’s comment Major Comment: 1. My major concern is the lack of information on operating lung volumes (inspiratory capacity reduction and indices of ventilatory constraints) and ventilatory profile and response to constant work rate which could have helped authors explain the variability of CET-Tend and perhaps differences between men and women. If the authors have the data, please add this information which is crucial! Response: We would like to thank the reviewer for his time reviewing the manuscript, and for the valuable comments – always greatly appreciated. We regret that unfortunately we have no data for the operating lung volumes of patients during the constant work rate test (CET). However, we do have the ventilation (VE) patterns of patients during the maximal cycle ergometry. The increase in ventilation is an effort to maintain O2 and CO2 homeostasis and may partly provide an insight of the ventilatory profile in those patients. Accordingly, we added the Online Figure 3 which is referred to the following additional text: P15 in the beginning “Moreover, a quartile gradation of ventilation pattern derived by the maximum cardiopulmonary test indicates the effort to maintain O2 and CO2 homeostasis and may partly provide an insight of the ventilatory profile in patients of current study (Online Figure 3).” Reviewer’s comment Major Comment: 2. Background and Discussion: "In fact, the first report of minimal clinically important difference for the CET has only been published in 2008 [7]." I encourage the authors to include: 1) the Puente-Maestu paper on the clinical relevance of constant power exercise duration in COPD (Eur Respir J. 2009; 34: 340-5) and 11 discuss their results in light of that paper which reported that baseline/pre-rehab CETTend at 75% of the peak work rate was 397±184s in 105 COPD patients; and 2) the O'Donnell paper on reliability of ventilatory parameters during cycle ergometry in multicentre trials in COPD (Eur Respir J. 2009; 34: 866-74) and discuss their results in light of that paper which reported that CET-Tend at 75% of the peak work rate was highly repeatable across run-in visits: 7.9+/-4.8 and 8.4+/-5.1 min (R = 0.84) in 463 COPD patients. Response: We added the reference of Puente-Maestu in background and discussion. Background P5 L3-4: “CET endurance time (CET-Tend) can be used in the evaluation of exercise tolerance.[4]” Discussion: P15 second paragraph L13 “....the minimal clinically important difference has been reported to be between 100 and 200 seconds[9] or ~33% increased from baseline[4].” We added the reference of O’Donnell in background and discussion. Background: P5 third paragraph: “Although CET is considered as sub-maximal exercise, it is aimed to assess the maximal aerobic capacity of patients and it is suggested as a highly repeatable test across run-in visits.[15]” Discussion: Second paragraph PP15-16: “Nevertheless, O’Donnell et al.[15] demonstrated that CET at 75% peak work rate was highly repeatable across run-in visits: 7.9+/-4.8 and 8.4+/-5.1min (R = 0.84) in 463 COPD patients indicating the appropriateness of CET in the exercise assessment of patients with COPD.” Reviewer’s comment Major Comment: 3. Methods: What reference values were used for peak work rate? Jones 1988, Jones 1989, Wasserman, Neder, Hansen...or others? Please specify this point. In addition, was the predicted work rate corrected for the (ideal) body weight? This seems 12 important, given the differences in body weight between males and females (see Table 1). Response: The reference values of Wasserman were used for the peak work rate. The predicted work rates of Wasserman are corrected for the body weight. Reviewer’s comment Minor Comments: 1. TITLE: "gender differences"....it should be "sex differences". 2. Keywords: "Pulmonary rehabilitation" does not reflect the topic of the manuscript; please replace it with "cycle endurance test" which seems more appropriate in this context. 3. Abstract and Discussion: A dataset of "2053" COPD patients was analyzed retrospectively. Please specify that the final dataset for the statistical analysis was based on 2002 COPD patients. 4. Abstract, Methods, Results and Discussion: please replace "exercise capacity" with "exercise tolerance". 5. Abstract and Discussion: "The percent of predicted peak work rate achieved differed significantly between men (48±21 % predicted) and women (67±31 % predicted). In contrast, CET-Tend was longer in men (286±174s vs 250±153s, p<0.001)". Is this related to the fact that the work rate during CET was greater in women? Or are there any other factors, not considered in this paper such as operating lung volumes (inspiratory capacity reduction and indices of ventilatory constraints) and ventilatory abnormalities (profile and response), that could explain the abovementioned finding? 6. Abstract and Discussion: exercise tolerance was quantified by using peak work rate: why did not the authors use peak oxygen consumption (VO2), which is known to correlate (to some extent) with the distance covered during a 6MWT? 13 7. Methods: "inclusion criteria: "no evidence" that they had reached a physiological maximum during the cardiopulmonary exercise test on a stationary cycle ergometer (i.e., ventilation #85% of maximal voluntary ventilation; heart rate #80% of the agepredicted heart rate; symptom score for dyspnea #7/10 points; and/or symptoms score for fatigue #7/10 points)" it should be "evidence" otherwise the CPET is deemed to be "submaximal". Please comment on it and correct accordingly. 8. Methods: "inclusion criteria: "CET done at 75% of predicted"....are the authors sure of this? It should be "CET done at 75% of peak work rate". 9. Discussion: "Nevertheless, peak heart rate, ventilation and dyspnea scores following a cardiopulmonary exercise test indicated a maximum effort for those patients (Online table 3)". I can't seem to find the Online Table 3. Response: Thanks the reviewer for the suggestions. Accordingly, we have provided the following changes: 1. We edited the title of the manuscript as follows: “Characteristics and determinants of endurance cycle ergometry and six-minute walk distance in patients with COPD” 2. We deleted the keyword "Pulmonary rehabilitation" (P4). The "cycle endurance test" abbreviated as "CET" is included in keywords. 3. We added in the Abstract: Results “A total of 2002 patients had CET time of duration (CET-Tend) less than 20min while only 51 (2.5%) of the patients achieved 20min of CET-Tend”. Additionally, we added the number of patients in the beginning of results P8 “On average, patients (n=2002; CETend <20min) had....” 4. We replaced "exercise capacity" with "exercise tolerance" throughout the manuscript. 5. Women had significantly lower FFMI compared to men (15.5 vs 18.1 kg/m2) and lower DLCO (50.9 vs 55.3%) which has been found to be positively related to quadriceps strength in patients with COPD (Ref 25). Moreover, the absolute peak 14 work rate were greater in men than women (we added also the absolute watt in quartile gradation for males and females; Online Figure 1 and 2, respectively) which suggest that even though men had a lower peak work rate %predicted, the absolute values of achieved watt indicating stronger quadriceps muscles can explain the longer CET-Tend in men. 6. The high rate of oxygen consumption cannot be often attainable during maximal exercise in patients with COPD or during the CET for practical reasons. Patients usually perform a symptom-limited CET and terminate the exercise effort before reaching the anaerobic threshold during the maximal cycle test or practically the measurement of VO2 during the CET is difficult for both assessors and the patients. In our dataset, unfortunately the VO2 during the CET was not available. We assumed that peak work rate can be also a valid tool for quantifying exercise tolerance in a big cohort of COPD patients. 7. Reviewer is right. We corrected this typing mistake. We changed the “no evidence” to “evidence” (P7 L7). 8. We corrected it. We edited it to "CET done at 75% of peak work rate" (P7 L.11) 9. The online table 3 was missing. We added it to OLS. 15