Rationale: Outcomes other than spirometry are required to assess nonbronchodilator therapies for chronic obstructive pulmonary disease. Estimates of the minimal clinically important difference for the 6-minute-walk distance (6MWD) have been derived from narrow cohorts using nonblinded intervention.
Objectives: To determine minimum clinically important difference for change in 6MWD over 1 year as a function of mortality and first hospitalization in an observational cohort of patients with COPD.
Methods: Data from the ECLIPSE cohort were used (n = 2,112). Death or first hospitalization were index events; we measured change in 6MWD in the 12-month period before the event and related change in 6MWD to lung function and St. George’s Respiratory Questionnaire (health status).
Measurement and Main Results: Of subjects with change in the 6MWD data, 94 died, and 323 were hospitalized. 6MWD fell by 29.7 m (SD, 82.9 m) more among those who died than among survivors (P < 0.001). A reduction in distance of more than 30 m conferred a hazard ratio of 1.93 (95% confidence interval, 1.29–2.90; P = 0.001) for death. No significant difference was observed for first hospitalization. Weak relationships only were observed with change in lung function or health status.
Conclusions: A reduction in the 6MWD of 30 m or more is associated with increased risk of death but not hospitalization due to exacerbation in patients with chronic obstructive pulmonary disease and represents a clinically significant minimally important difference.
Existing estimates of the minimal clinically important difference (MCID) in 6-minute walking distance in chronic obstructive pulmonary disease (COPD) have been derived from smaller cohorts with a relatively narrow range of airflow obstruction and who have been subjected to interventions intended to be beneficial, which cannot be placebo controlled (e.g., pulmonary rehabilitation or pneumoplasty for inhomogeneous emphysema).
In the present study, the MCID was obtained by analysis of a large, multicenter observational cohort in whom no intervention was applied, using mortality as an index event. This study establishes the MCID of patients with COPD with a wide range of degrees of airflow obstruction. Unlike smaller studies centered on patients selected for unblinded interventions, this study used mortality as the anchor measure to determine the MCID. An annual change of −30 m was found to be the MCID for death. Six-minute walk is an important functional outcome in COPD. This study defines a minimal clinically important difference based on mortality.
For patients with chronic obstructive pulmonary disease (COPD), the 6-minute-walk distance (6MWD) is a potentially useful biomarker of disease severity. Its utility stems from the fact that it is a simple integrative field test in which performance is not only influenced by the severity of lung compromise but also by extrapulmonary manifestations of the condition, such as muscle weakness, pulmonary vascular disease, or depression (1). The 6MWD has been shown to be an important predictor of survival in observational studies (2, 3). In the Evaluation of COPD Longitudinally to Identify Predictive Surrogate Endpoints (ECLIPSE) cohort, we have previously shown that patients with a 6MWD less than 334 and 357 m have an increased risk of death and hospitalization, respectively (4). Conversely, improved 6MWD has been reported after nonbronchodilator interventions of accepted benefit in COPD, most notably pulmonary rehabilitation (5) and responders to lung volume reduction surgery (6) or both (7).
The minimal clinically important difference (MCID) has been defined as the smallest difference in a measurable clinical parameter that indicates a meaningful change in the condition for better or for worse as perceived by the patient, clinician, or investigator (8). In designing studies in which 6MWD is an outcome measure, knowledge of the MCID is essential in interpreting the results, but current data regarding the MCID are incomplete. In particular, all current studies reporting MCID for COPD derived its value by anchoring it to other indices of function in small COPD cohorts or selected patients receiving pulmonary rehabilitation (9, 10) and/or lung volume reduction surgery (7), meaning that the results may not be widely applicable. Second, in all these studies an intervention was used for which it has not been possible to provide a placebo control (7, 9, 10). No study by design sought to relate the change in 6MWD to worsening health outcomes and in particular to death or hospital admission, which are of indisputable importance. The ECLIPSE cohort offers a unique opportunity to explore the MCID for the 6MWD because it is a large, multicenter, prospective observational cohort of patients with a wide range of airflow obstruction who were not selected for likely response to specific interventions (11). In these patients, any fluctuation of their clinical state occurred independently of the investigators and was, temporally, randomly related to the measurement points. Therefore, we sought, using this cohort, to determine an MCID relative to death or hospital admission.
ECLIPSE was a 3-year, multicenter, longitudinal, prospective study to identify novel endpoints in COPD (12). Participants were 40 to 75 years of age and had a smoking history of ≥10 pack-years, a postbronchodilator FEV1 of <80% of the predicted value, and baseline postbronchodilator FEV1/FVC of ≤0.7 (13). From the spirometry we determined the FEV1 in absolute values and as percent predicted (14). At baseline, at 3 and 6 months, and every 6 months thereafter, in addition to the lung function from which FEV1 was determined, we measured health status (using the COPD-specific St. George’s Respiratory Questionnaire [SGRQ-C]) and performe other tests as described previously (1, 4, 11, 12). This study was approved by the ethical review boards of the participating centers, and all participants provided written informed consent.
The 6-minute-walk test (6MWT) was performed indoors along a flat, straight, 30-m walking course supervised by a trained researcher according to the ATS guidelines (15); a practice 6MWT was not completed. Patients were encouraged every minute of the 6MWT using two phrases: “You are doing well” or “Keep up the good work.” Patients were allowed to stop and rest during the test but were instructed to resume walking as soon as they felt able to do so. Baseline and longitudinal 6MWD data of the ECLIPSE study have been reported elsewhere (1, 4).
In ECLIPSE, exacerbations were defined as events that led a care provider to prescribe antibiotics and/or corticosteroids or that led to hospitalization (16). However, for the purpose of this study, the analysis was limited to exacerbations that required hospitalization because these are of unequivocal relevance to patients. The number of exacerbations during the 3 years after the baseline visit was recorded prospectively at clinic visits and by monthly telephone calls, using a structured interview scheme. Mortality was determined up to Day 1,060 of the study. All-cause mortality was used as the outcome; no attempts were made to determine cause of death.
Anchor-based methodology was used to investigate the MCID of the change in 6MWD with clinically relevant functional anchors (FEV1 and SGRQ-C) to replicate derivations of the MCID reported in the literature.
For the purpose of this analysis, we considered a priori that the quantity of interest was a change in the 6-minute walking distance (Δ6MWD) in the 1 year preceding a biological outcome of relevance, specifically death or first exacerbation requiring hospitalization. By definition, these events had to occur after the second 6MWD had been recorded at 12 months. Comparator data were obtained from participants who did not have either index event, and the last available Δ6MWD was used for this purpose. Precise definitions for the choice of each Δ6MWD are given in Table 1. The last two 6-minute walks from which the interval was defined were within the last 2 years of the index event for 79% of those who died, for 89% of those admitted to hospital, and for 88% for the composite endpoint of admission or death.
|1. If the subject did not die or have a hospitalization, the last nonmissing 1-yr Δ6MWD is used.|
|2. If the subject died but did not have a hospitalization, the last nonmissing 1-yr Δ6MWD is used.|
|3. If the subject did not die but had a hospitalization, the last nonmissing 1-yr Δ6MWD that is prior to the first hospitalization is used.|
|4. If the subject died and had a hospitalization, the last nonmissing 1-yr Δ6MWD that is prior to the first hospitalization is used.|
Because we expected the number of deaths to be small in Years 2 and 3, we not only evaluated the relationship between Δ6MWD and death and hospitalization but also with a composite index of the two outcomes combined. We then applied Cox proportional analysis to determine the hazard ratios associated with these outcomes.
Demographic data for subjects with baseline 6MWD, for subjects with the defined Δ6MWD, and for subjects with the index events of interest are shown in Table 2. The subset of subjects eligible for the analyses (n = 1,847) was similar to subjects who had a baseline 6MWD (n = 2,112). At baseline, nonsurvivors walked an average 51.8 m less than survivors (P < 0.001). They were also slightly older, had more smoking history, had a higher incidence of self-reported cardiovascular disease, were more dyspneic, had more severe airflow obstruction, and had more impaired quality of life. Similar differences were seen between hospitalized patients and those that were not (Table 2), with the exception of smoking history and self-reported cardiovascular disease.
|Subject Death (Years 2 and 3)||Exacerbation Hospitalization (Years 2 and 3)|
|Characteristic||COPD Subjects with 6MWT at Baseline||COPD Subjects with 1-Year Pair||No||Yes||P Value||No||Yes||P Value|
|Age, yr||63.4 (7.1)||63.3 (7.0)||63.1 (7.0)||66.1 (6.9)||<0.001||63.1 (7.1)||64.0 (6.4)||0.029|
|Male, %||1,381 (65%)||1,204 (65%)||1135 (65%)||69 (73%)||0.086||834 (65%)||201 (62%)||0.317|
|Current smoker, %||764 (36%)||652 (35%)||622 (35%)||30 (32%)||0.481||462 (36%)||110 (34%)||0.489|
|Pack-years||48.7 (27.1)||48.7 (27.3)||48.2 (27.0)||56.9 (31.8)||0.003||48.9 (28.6)||48.0 (24.3)||0.565|
|FEV1% predicted (postbronchodilator)||48.4 (15.7)||49.1 (15.7)||49.4 (15.7)||43.7 (14.4)||<0.001||52.0 (15.2)||44.9 (14.7)||<0.001|
|Cardiovascular disease||1,174 (56%)||1,022 (55%)||955 (54%)||67 (71%)||0.001||716 (56%)||174 (54%)||0.495|
|BMI, kg/m2||26.5 (5.6)||26.6 (5.6)||26.6 (5.5)||26.6 (6.5)||0.976||26.9 (5.7)||26.3 (5.1)||0.139|
|FFMI, kg/m2||17.2 (2.8)||17.2 (2.8)||17.2 (2.8)||17.6 (3.5)||0.188||17.3 (2.9)||17.0 (2.6)||0.078|
|6MWD, m||369.1 (121.7)||378.5 (117.0)||381.1 (116.7)||329.3 (112.9)||<0.001||390.7 (117.1)||356.5 (118.8)||<0.001|
|mMRC score||1.7 (1.1)||1.6 (1.0)||1.6 (1.0)||1.9 (1.0)||0.002||1.5 (1.0)||1.9 (1.1)||<0.001|
|mMRC score ≥2||1,098 (53%)||914 (51%)||855 (50%)||59 (65%)||0.006||548 (44%)||190 (61%)||<0.001|
|CES-D score||11.4 (9.3)||11.1 (9.2)||11.1 (9.3)||11.7 (8.4)||0.530||10.6 (9.1)||11.3 (9.1)||0.252|
|CES-D score ≥16||544 (26%)||449 (25%)||423 (25%)||26 (28%)||0.462||287 (23%)||85 (27%)||0.163|
|FACIT-F score||35.2 (10.6)||35.7 (10.4)||35.9 (10.4)||32.6 (10.9)||0.003||36.7 (10.1)||35.3 (10.3)||0.024|
|FACIT-F Score ≤30||653 (32%)||528 (29%)||491 (29%)||37 (39%)||0.027||323 (26%)||97 (30%)||0.100|
|SGRQ-C total score||50.0 (20.2)||48.8 (20.2)||48.6 (20.2)||53.5 (19.7)||0.023||45.2 (20.0)||53.6 (18.8)||<0.001|
|SGRQ activity domain||63.8 (25.0)||62.5 (25.1)||62.2 (25.1)||69.1 (24.7)||0.009||57.9 (25.3)||69.2 (22.5)||<0.001|
|SGRQ impacts domain||38.4 (21.9)||37.2 (21.7)||37.0 (21.7)||42.0 (21.5)||0.031||33.9 (21.2)||41.3 (20.9)||<0.001|
|SGRQ symptoms domain||60.4 (21.6)||59.4 (21.7)||59.3 (21.8)||61.4 (19.3)||0.367||56.7 (21.7)||62.9 (21.4)||<0.001|
There was a modest but statistically significant association between baseline 6MWD and the functional anchoring variables (FEV1 and SGRQ) (Table 3). The strength of the correlation between change in 6MWD and the functional anchors was weak and less than the previously quoted figure of ≥0.3 for a useful anchor (7, 10), and therefore MCID determination by anchor-based methodology was not taken forward for these functional anchors.
|Measurement||n||Correlation* with 6MWD Distance at Baseline||n||Correlation* with Change in 6MWD|
|SGRQ-C total score||2,033||−0.469||1,403||−0.212|
|SGRQ-C activity domain score||2,056||−0.514||1,381||−0.258|
Table 4 shows that the Δ6MWD in the year before death was more negative and statistically significant (P < 0.001) in nonsurvivors compared with survivors. In contrast, the Δ6MWD was not significant for Δ6MWD before first exacerbation leading to hospitalization or for the composite index of both endpoints.
|Death and/or hospitalization||1,228||−4.2||82.3||374||−3.6||85.6||−0.7||83.1||0.892|
Based on the mortality endpoint, it proved rational to go forward with a decrease in 6MWD of 30 m. A reduction in 6MWD of more than 30 m yielded a hazard ratio (HR) of 1.93 (95% confidence interval [CI], 1.29–2.90; P = 0.001) for the endpoint of death, a HR of 1.18 (95% CI, 0.93–1.49; P = 0.179) for hospitalization, and a HR of 1.24 (95% CI, 0.99–1.54; P = 0.056) for the composite endpoint of death or hospitalization (Figure 1). Because for an individual (or their clinician) absolute change is of more interest than differences between two groups, we repeated this analysis for a threshold of −39.6 m, which was the last recorded mean change in 6MWD in patients who died (Table 4). The corresponding HRs for this change were 2.08 (95% CI, 1.38–3.12; P < 0.001) for the endpoint of death, 1.14 (95% CI, 0.89–1.46; P = 0.303) for hospitalization, and 1.22 (95% CI, 0.97–1.53; P = 0.092) for the composite endpoint of death or hospitalization.
Data from the ECLIPSE large cohort of patients with COPD show that a decrease in 6MWD over 1 year of 30 m or more is significantly related to the risk of death over the subsequent 12 months. The value of −30 m can be considered a useful MCID because it is associated with an indisputably important outcome and is consistent with the estimates of MCID obtained in the literature using other methods, thus reinforcing the validity of its use.
Prior reports proposing MCID values for the 6MWD have all been obtained from scenarios where therapies were given with the aim of improving outcome, typically lung volume reduction surgery or pulmonary rehabilitation (7, 9, 10, 17). In the present study, the MCID was determined in the context of clinical deterioration, specifically death or hospitalization due to exacerbation, which are important clinical outcomes. The value of −30 m, which corresponds to increased mortality, parallels similar estimates associated with patient-perceived improvement, which have been in the range 26 to 54 m.
The ECLIPSE cohort differs from all prior studies by representing a COPD population that is generally included in clinical trials (moderate to severe airflow obstruction, recruited from specialist centers) but with very few exclusions. In contrast, prior studies, even where a control group was used, recruited patients in a selective fashion in an attempt to achieve the primary aims of the study; for example, the patients participating in the NETT study, from which an MCID of 26 m was obtained (7), were selected by the presence of emphysema and lack of important comorbidities to be included in the study. In ECLIPSE there were no specified treatment interventions, and clinicians caring for the patients were permitted to change treatment in line with local guidelines and resources. This means that changes in the patient’s condition due to treatment interventions were randomly associated with measurement time points, making the emergence of an MCID more robust.
Our approach, although novel for COPD, is consistent with the MCID thresholds that have been derived for other lung diseases, such as idiopathic pulmonary fibrosis, where the outcome used to anchor the MCID was death or a composite death and hospitalization outcome (18). Our results indicate that, using a similar approach in patients with COPD and considering death, a change of −30 m is a reasonable threshold that falls between the 25 and 54 m that have been reported and is close to Puhan and coworkers’ initial figure in the context of improvement of 35 m (10). Our data significantly strengthen the case for an MCID in the range 25 to 35 m because they are derived from data obtained in a quite different context from other reports.
Despite the similarity between the current MCID and previously proposed values, we speculate that more work needs to be done to better define the effect of context and, by inference, placebo interventions on 6MWD. In particular, a patients’ perception of change could be heightened by the presence of an intervention, coupled with measurements close to that intervention, so a MCID might be easier to achieve. If this argument is accepted, it may be appealing to assess the MCID in the context of a placebo–controlled, double-blind intervention because this falls between the current data and prior work. The literature supports a strong placebo effect on 6MWD in the case of heart failure (for discussion see Reference 19), underlining the importance of the current dataset.
In ECLIPSE, a practice 6MWT was not undertaken. It is accepted that a second 6MWT done the day after the first is associated with a significant learning effect, probably approximately 6% (20). The effect of a practice walk on Δ6MWD measured over a 12-month period is unknown. However, although our data must be interpreted in this context, the absence of a practice walk was present equally in patients who suffered an index event and those who did not, so although we doubt this detracts from our findings, this may require further evaluation. The use of a single threshold value to express an MCID when the variance is large, as is the case with the 6MWD, offers problems of accuracy. However, the values, until now reported using different methods in different populations, fall within a relatively narrow range. No improvement in accuracy or variance was obtained when the changes were analyzed as percent of predicted values (to normalize for age, gender, height, and weight) and as percent of baseline (results not shown). This is consistent with the finding by Cote and coworkers, who followed over 1,400 patients with COPD and found the predictive value of the 6MWD expressed in meters to have equal or better performance than corrections derived from the use of three different normalizing equations (3).
We observed only weak associations between changes in FEV1 and SGRQ and changes in 6MWD. We believe this is likely due to the observational nature of the study whereby no interventions were applied, thus decreasing the likelihood that the outcomes would be linked. The associations were very similar to those reported by DuBois and colleagues (18) in the determination of MCID for patients with idiopathic pulmonary fibrosis. In contrast to the findings in that study, the association between the change in 6MWD and death was statistically significant in our study, whereas in that study, only the combined end-point (hospitalization/death) reached statistical significance.
The observation that we could not define a Δ6MWD in relation to exacerbation requiring hospitalization may reflect differing impacts of treatment and admission criteria in different health care systems given the international nature of ECLIPSE. Put simply, Δ6MWD does not seem to change the likelihood of having a first hospitalization, which in most cases has an infectious trigger (21), whereas the chance of dying (i.e., not responding to therapy) may be more directly related to frailty, measured as muscle weakness (22), or aspects of lung disease beyond the FEV1 (23). However, although the nonsurvivors had a higher prevalence of self-reported cardiovascular disease and prior smoking than survivors, this difference was not observed between those who were hospitalized and those who were not, even though more patients suffered hospitalization than death. Thus, an additional hypothesis would be that these patients had cardiovascular disease that rendered hospitalization nonsurvivable; this hypothesis would be consistent with the observation that troponin is a marker of impaired survival during acute exacerbation of COPD (24). The cause of death was not adjudicated in ECLIPSE, so it is not possible to explore this further.
The question arises whether the value evaluated for the MCID value should be the mean change in those who died (39.6 m in the current study) or the mean between-group change for those who died and survivors (29.7 m in the current study). Even though the difference between these values is relatively small, we suspect that the latter is most useful for clinical trials in secondary care where two groups are typically compared, and the former might be more useful for an individual or a clinician. From the data presented herein, however, it is apparent that very similar HRs are observed if 39.6 m is used; this higher value tends to marginally weaken rather than strengthen the relationship with exacerbation.
In summary, we conclude that −30 m represents a clinically meaningful MCID for the Δ6MWD in the context of worsening clinical status in patients with a wide range of severity of COPD where no intervention is applied. Further studies to evaluate the MCID for COPD patients in placebo-controlled drug studies of effective medicinal agents may help further refine knowledge of the MCID for different interventions in this condition.
Editorial support limited to collation of conflict of interest statements and assistance with submission of the manuscript draft approved by all authors was provided by Claudia Russell at Gardiner-Caldwell Communications, Macclesfield, UK.
|1.||Spruit MA, Watkins ML, Edwards LD, Vestbo J, Calverley PM, Pinto-Plata V, Celli BR, Tal-Singer R, Wouters EF. Determinants of poor 6-min walking distance in patients with COPD: the ECLIPSE cohort. Respir Med 2010;104:849–857.|
|2.||Pinto-Plata VM, Cote C, Cabral H, Taylor J, Celli BR. The 6-min walk distance: change over time and value as a predictor of survival in severe COPD. Eur Respir J 2004;23:28–33.|
|3.||Cote CG, Pinto-Plata VM, Marin JM, Nekach H, Dordelly LJ, Celli BR. The modified BODE index: validation with mortality in COPD. Eur Respir J 2008;32:1269–1274.|
|4.||Spruit MA, Polkey MI, Celli B, Edwards LD, Watkins ML, Pinto-Plata V, Vestbo J, Calverley PM, Tal-Singer R, Agusti A, et al.. Predicting outcomes from 6-minute walk distance in chronic obstructive pulmonary disease. J Am Med Dir Assoc 2012;13:291–297.|
|5.||Troosters T, Gosselink R, Decramer M. Short- and long-term effects of outpatient rehabilitation in patients with chronic obstructive pulmonary disease: a randomized trial. Am J Med 2000;109:207–212.|
|6.||Fishman A, Martinez F, Naunheim K, Piantadosi S, Wise R, Ries A, Weinmann G, Wood DE. A randomized trial comparing lung-volume-reduction surgery with medical therapy for severe emphysema. N Engl J Med 2003;348:2059–2073.|
|7.||Puhan MA, Chandra D, Mosenifar Z, Ries A, Make B, Hansel NN, Wise RA, Sciurba F. The minimal important difference of exercise tests in severe COPD. Eur Respir J 2011;37:784–790.|
|8.||Kiley JP, Sri Ram J, Croxton TL, Weinmann GG. Challenges associated with estimating minimal clinically important differences in COPD: the NHLBI perspective. COPD 2005;2:43–46.|
|9.||Holland AE, Hill CJ, Rasekaba T, Lee A, Naughton MT, McDonald CF. Updating the minimal important difference for six-minute walk distance in patients with chronic obstructive pulmonary disease. Arch Phys Med Rehabil 2010;91:221–225.|
|10.||Puhan MA, Mador MJ, Held U, Goldstein R, Guyatt GH, Schunemann HJ. Interpretation of treatment changes in 6-minute walk distance in patients with COPD. Eur Respir J 2008;32:637–643.|
|11.||Agusti A, Calverley PM, Celli B, Coxson HO, Edwards LD, Lomas DA, MacNee W, Miller BE, Rennard S, Silverman EK, et al.. Characterisation of COPD heterogeneity in the ECLIPSE cohort. Respir Res 2010;11:122.|
|12.||Vestbo J, Anderson W, Coxson HO, Crim C, Dawber F, Edwards L, Hagan G, Knobil K, Lomas DA, MacNee W, et al.. Evaluation of COPD longitudinally to identify predictive surrogate end-points (ECLIPSE). Eur Respir J 2008;31:869–873.|
|13.||Rabe KF, Hurd S, Anzueto A, Barnes PJ, Buist SA, Calverley P, Fukuchi Y, Jenkins C, Rodriguez-Roisin R, van Weel C, et al.. Global strategy for the diagnosis, management, and prevention of chronic obstructive pulmonary disease: GOLD executive summary. Am J Respir Crit Care Med 2007;176:532–555.|
|14.||Quanjer PH, Tammeling GJ, Cotes JE, Fabbri LM, Matthys H, Pedersen OF, Peslin R, Roca J, Sterk PJ, Ulmer WT, et al.. Symbols, abbreviations and units: working party standardization of lung function tests, European community for steel and coal. Eur Respir J Suppl 1993;16:85–100.|
|15.||American Thoracic Society. ATS statement: guidelines for the six-minute walk test. Am J Respir Crit Care Med 2002;166:111–117.|
|16.||Hurst JR, Vestbo J, Anzueto A, Locantore N, Mullerova H, Tal-Singer R, Miller B, Lomas DA, Agusti A, Macnee W, et al.. Susceptibility to exacerbation in chronic obstructive pulmonary disease. N Engl J Med 2010;363:1128–1138.|
|17.||Redelmeier DA, Bayoumi AM, Goldstein RS, Guyatt GH. Interpreting small differences in functional status: the six minute walk test in chronic lung disease patients. Am J Respir Crit Care Med 1997;155:1278–1282.|
|18.||du Bois RM, Weycker D, Albera C, Bradford WZ, Costabel U, Kartashov A, Lancaster L, Noble PW, Sahn SA, Szwarcberg J, et al.. Six-minute-walk test in idiopathic pulmonary fibrosis: test validation and minimal clinically important difference. Am J Respir Crit Care Med 2011;183:1231–1237.|
|19.||Olsson LG, Swedberg K, Clark AL, Witte KK, Cleland JG. Six minute corridor walk test as an outcome measure for the assessment of treatment in randomized, blinded intervention trials of chronic heart failure: a systematic review. Eur Heart J 2005;26:778–793.|
|20.||Hernandes NA, Wouters EF, Meijer K, Annegarn J, Pitta F, Spruit MA. Reproducibility of 6-minute walking test in patients with COPD. Eur Respir J 2011;38:261–267.|
|21.||Wedzicha JA, Seemungal TA. COPD exacerbations: defining their cause and prevention. Lancet 2007;370:786–796.|
|22.||Swallow EB, Reyes D, Hopkinson NS, Man WD, Porcher R, Cetti EJ, Moore AJ, Moxham J, Polkey MI. Quadriceps strength predicts mortality in patients with moderate to severe chronic obstructive pulmonary disease. Thorax 2007;62:115–120.|
|23.||Moore AJ, Soler RS, Cetti EJ, Amanda Sathyapala S, Hopkinson NS, Roughton M, Moxham J, Polkey MI. Sniff nasal inspiratory pressure versus IC/TLC ratio as predictors of mortality in COPD. Respir Med 2010;104:1319–1325.|
|24.||Hoiseth AD, Neukamm A, Karlsson BD, Omland T, Brekke PH, Soyseth V. Elevated high-sensitivity cardiac troponin t is associated with increased mortality after acute exacerbation of chronic obstructive pulmonary disease. Thorax 2011;66:775–781.|
*These authors contributed equally to this work.
This work was supported by the NIHR Respiratory Disease Biomedical Research Unit at the Royal Brompton and Harefield NHS Foundation Trust and Imperial College London (M.I.P.) and by GlaxoSmithKline.
Author Contributions: M.I.P., M.A.S., and V.P.-P. contributed to acquisition of data and analysis and interpretation of data. M.L.W., B.E.M., and J.Y. contributed to analysis and interpretation of data. L.D.E., R.T.-S., and C.C. contributed to conception and design of the study and analysis and interpretation of data. J.V., P.S.B., P.M.A.C., A.A., H.O.C., D.A.L., W.M., S.R., E.K.S., E.F.M.W., and and B.C. contributed to conception and design of the study, acquisition of primary data, and analysis and interpretation of data. All authors contributed to drafting the article or revising it critically for important intellectual content. All authors approved the final version to be published. M.I.P. attests that the authors had access to all the study data, takes responsibility for the accuracy of the analysis, and had authority over manuscript preparation and the decision to submit the manuscript for publication.
Originally Published in Press as DOI: 10.1164/rccm.201209-1596OC on December 21, 2012