Analysis of Clinical Methods Used to Evaluate Dyspnea in Patients with Chronic Obstructive Pulmonary Disease

TAKASHI HAJIRO, KOICHI NISHIMURA, MITSUHIRO TSUKINO, AKIHIKO IKEDA, HIROSHI KOYAMA, and TAKATERU IZUMI
Department of Respiratory Medicine, Graduate School of Medicine, Kyoto University, Kyoto, Japan

When dyspnea must be assessed clinically, there are three methods of assessment: the measurement of dyspnea with activities of daily living using clinical dyspnea ratings such as the modified Medical Research Council (MRC), the Baseline Dyspnea Index (BDI), and the Oxygen Cost Diagram (OCD); the measurement of dyspnea during exercise using the Borg scale; to assess the influence of dyspnea on health-related quality of life (HRQoL) using disease-specific questionnaires such as the St. Georges Respiratory Questionnaire (SGRQ) and the Chronic Respiratory Disease Questionnaire (CRQ). The purpose of the present cross-sectional study was to clarify relationships between dyspnea ratings and HRQoL questionnaires by applying factor analysis. One hundred sixty-one patients with mild to severe COPD completed pulmonary function tests, progressive cycle ergometer testing for exercise capacity, assessment of dyspnea, HRQoL, anxiety, and depression. Factor analysis demonstrated that the MRC, BDI, OCD, and Activity of the SGRQ, and Dyspnea of the CRQ, were grouped into the same factor, and the frequency distribution histograms of these five measures showed virtually the same distribution. The Borg scale at the end of maximum exercise was found to be a different factor. The MRC, BDI, OCD, and Activity in the SGRQ, and Dyspnea in the CRQ demonstrated the same pattern of correlation with physiologic data, and they had significant relationships with FEV1 (correlation coefficients [Rs] 0.31 to 0.48) and maximal oxygen uptake (Rs 0.46 to 0.60). Diseasespecific HRQoL questionnaires, the SGRQ and the CRQ, which contain a specific dimension for evaluating dyspnea, may be substituted for clinical dyspnea ratings in a cross-sectional assessment. Dyspnea rating at the end of exercise may provide further information regarding dyspnea. Hajiro T, Nishimura K, Tsukino M, Ikeda A, Koyama H, Izumi T. Analysis of clinical methods used to evaluate dyspnea in patients with chronic obstructive pulmonary disease.
AM J RESPIR CRIT CARE MED 1998;158:11851189.

The evaluation of dyspnea in patients with COPD should be included among standard measures of physiologic lung function since dyspnea ratings and lung function are separate factors that independently characterize the condition of patients with COPD (1). Dyspnea with activities of daily living (ADL) has been measured using clinical dyspnea ratings such as those of the modified Medical Research Council (MRC) (2), the Baseline Dyspnea Index (BDI) (3), and the Oxygen Cost Diagram (OCD) (4), all of which have been widely used. Dyspnea during exercise testing or with physical activity has been measured with the Borg scale (5), or the Visual Analog Scale (6). Arbitrary choices on the dyspnea scale have been made in clinical trials when clinical evaluation of dyspnea was utilized as one of the outcome measures.

(Received in original form February 19, 1998 and in revised form June 11, 1998) Correspondence and requests for reprints should be addressed to Takashi Hajiro, M.D., Department of Respiratory Medicine, Graduate School of Medicine, Kyoto University, Sakyo, Kyoto, 606-8507, Japan. E-mail: hajiro@chest.kyoto-u.ac.jp Am J Respir Crit Care Med Vol 158. pp 11851189, 1998 Internet address: www.atsjournals.org

On the other hand, an important aim of medical interventions may be to improve not only the quantity of life but also the quality of life of each patient. Because the term healthrelated quality of life (HRQoL) is widely used to describe how the quality of life of patients is affected by health and disease status, a well-designed HRQoL questionnaire should encompass comprehensive dimensions that quantify the overall effect of disease on the patients daily life and sense of well being (7, 8). In patients with COPD, it is essential for a measure of HRQoL to include an assessment of influences attributable to dyspnea. Disease-specific questionnaires for COPD have a specific dimension in which the impacts of dyspnea on patient health status are supposed to be evaluated (8, 9). Presently there are three possible ways to assess dyspnea clinically: the measurement of dyspnea during ADLs using clinical dyspnea ratings, the measurement of dyspnea during exercise testing in laboratories using dyspnea ratings, and the assessment of the influence of dyspnea on HRQoL using a disease-specific questionnaire. We speculate that a disease-specific HRQoL questionnaire can evaluate dyspnea as well as clinical dyspnea ratings can. The purpose of the present crosssectional study was to clarify interrelationships between clini-

1186

AMERICAN JOURNAL OF RESPIRATORY AND CRITICAL CARE MEDICINE

VOL 158

1998

cal dyspnea ratings, dyspnea ratings at maximal exercise, and HRQoL questionnaires by applying factor analysis (10), and to clarify the suitability of evaluating patient dyspneic status using the various methods.

preted and named. We incorporated the following variables in this factor analysis: MRC, BDI, OCD, the Borg scale, three dimensions of the SGRQ, the four components of the CRQ, and Anxiety and Depression of the HAD.

RESULTS METHODS
Patients with stable COPD defined by the American Thoracic Society (11) were recruited at the outpatient clinic. All patients had more than 6 mo of outpatient management before entry to avoid substantial changes in subjective parameters brought about by new medical interventions (12). Entry criteria for the study were: (1) a FEV1/FVC ratio of less than 0.7 and FEV1 of less than 80% of the predicted value for all measurements made during the previous 6 mo; (2) smoking history of more than 20 pack-yr; (3) no history suggestive of asthma; (4) no exacerbation of airflow limitation in the preceding 6 wk; and (5) no changes in treatment regimen during the preceding 4 wk. All eligible patients underwent the following examinations on the same day. Inhalation of bronchodilators was withheld for at least 12 h before pulmonary function tests were performed. FEV1 and FVC were assessed before and 60 min after the inhalation of 400 g of salbutamol and 80 g of ipratropium bromide using a metered-dose inhaler with a spacer device (InspirEase; Schering-Plough K.K, Osaka, Japan). All spirometric flowvolume curves were recorded according to the method described in the American Thoracic Society 1994 update (13). The predicted values for FEV1 and vital capacity (VC) were calculated according to the proposal of the Japan Society of Chest Diseases (14). The residual volume (RV) was measured by the closed-circuit helium method and the diffusion capacity for carbon monoxide (DLCO) was measured using the single-breath technique (CHESTAC65V; Chest, Tokyo, Japan). The progressive exercise test to a symptom-limited maximum was performed using an electrically braked cycle ergometer (Corival WLP-400; Lode, Groningen, The Netherlands) at 60 min after the inhalation of the bronchodilators as described by Ikeda and colleagues (15). The workload was increased automatically, and patients maintained a pedaling frequency above 40 cycles/min throughout the test. Exercise data were recorded using an automated exercise testing system (Desktop Diagnostics/CPX; Medical Graphic Corp., St. Paul, MN). At the end of each exercise test, maximum oxy gen uptake ( VO2max) was calculated and breathlessness was scored with the Borg scale (0 to 10) (5). To assess dyspnea, the Japanese version of the modified MRC (2), the BDI (3, 16), and the OCD (4) were used. The MRC is a five-point scale based on degrees of various physical activities that precipitate dyspnea. The BDI recognizes five grades for each of the following categories: functional impairment, magnitude of task, and magnitude of effort. The OCD is a visual analog scale that corresponds to oxygen requirements at different activity levels, which was represented as a value ranging from zero to 100, with a score of 100 indicating no impairment. The HRQoL was assessed by the Japanese version of the St. Georges Respiratory Questionnaire (SGRQ) (8) and the Chronic Respiratory Questionnaire (CRQ) (9), which have been precisely validated (17). The SGRQ has three dimensions: Symptoms, concerned with respiratory symptoms; Activity, concerned with activities that evoke or are limited by dyspnea; and Impacts, a measure of the overall disturbance. The CRQ has four components: Dyspnea, Emotional Function, Mastery, and Fatigue. In the Dyspnea dimension, each patient defined the five items in terms of ADL limited by disease. The Japanese version of the Hospital Anxiety and Depression Scale (HAD) (18) was used for evaluating patients anxiety and depression status. All results are presented as mean SD. The relationship between two sets of data was analyzed by Spearmans rank correlation test. A p value of less than 0.01 was considered to be statistically significant. Factor analyses was used to determine the dimensions underlying the pattern of interrelationships and to reduce a large set of variables to smaller sets of factors (10). Factor analysis may be regarded as a data reduction technique. In factor analysis, a matrix of correlations between variables is created, and then data are transformed into linear combinations of variables that share common variance between measures. The correlations between the original variables and the linear combinations or factors are called factor loadings. By examining which variables loaded highly on each factor, the factors were inter-

A total of 161 consecutive patients (160 men) with mild to severe COPD, and having a wide range of FEV1 predicted values, were studied (Table 1). One hundred fifty-seven patients were treated with inhalation of both 400 g of salbutamol and 80 g of ipratropium bromide more than four times a day. Seven patients were managed with long-term domiciliary oxygen therapy. Four patients received no medication. A summary of the varimax rotation of the factor analysis is provided in Table 2. Three factors accounted for 70.0% of the total variance of the data. The MRC, the BDI, the OCD, Activity in the SGRQ, and Dyspnea in the CRQ loaded predominantly on Factor 1, which appeared to be the dyspnea factor. Factor 2 included scores of anxiety and depression on the HAD, and appeared to be related to psychological status. Symptoms in the SGRQ and Fatigue, Emotion, and Mastery in the CRQ loaded predominantly on Factor 3. Using the Borg scale, however, the dyspnea rating after maximal exercise was found not to be grouped with any of the factors above. On the basis of these results, we focused on Factor 1, the dyspnea factor, in which the MRC, the BDI, the OCD, Activity in the SGRQ, and Dyspnea in the CRQ were grouped. The frequency distribution histograms of the MRC, the BDI, the OCD, Activity in the SGRQ, and Dyspnea in the CRQ are represented in Figure 1. These histograms demonstrated virtually normal distributions.

TABLE 1 PATIENT CHARACTERISTICS IN 161 PATIENTS WITH COPD

Mean Sex, M/F Age, yr FEV1, L FEV1, % pred Postbronchodilator FEV1, % pred RV/TLC, % DLCO, % pred Clinical dyspnea ratings MRC score (04) BDI score (012) OCD (0100), mm Exercise test V O2max, mL/min Borg score at the end of exercise HAD Anxiety (021) Depression (021) HRQoL questionnaires SGRQ Symptoms (0100) Activity (0100) Impact (0100) CRQ Dyspnea (535) Fatigue (428) Emotional function (749) Mastery (428) SD Range 4889 0.172.65 6079.8 15.787.6 26.171.2 30.5131.2 04 212 8100 2021,625 310 017 016

160/1 69 7 1.06 0.49 40.0 17.5 47.9 17.4 46.9 10.3 65.1 19.5 1.0 7.9 62 829 6.0 3.9 5.0 0.8 2.6 19 270 1.5 3.3 3.7

52.5 47.5 25.7 26.2 19.8 39.5 22.0

19.9 21.9 16.9 5.7 5.0 7.0 4.3

7.8100 0100 075.6 1135 428 1949 1128

Definition of abbreviations: RV residual volume; MRC Medical Research Council; BDI Baseline Dyspnea Index; OCD Oxygen Cost Diagram; VO2max maximum oxygen uptake; HAD Hospital Anxiety and Depression Scale; HRQoL health-related quality of life; SGRQ St. Georges Respiratory Questionnaire; CRQ Chronic Respiratory Disease Questionnaire. * Numbers in parentheses represent the theoretical score range.

Hajiro, Nishimura, Tsukino, et al.: Dyspnea and Health-related Quality of Life

TABLE 2 ROTATED FACTOR LOADINGS FROM FACTOR ANALYSIS IN 161 PATIENTS WITH COPD
Factor 1 Factor 1: dyspnea MRC BDI OCD Activity, SGRQ Dyspnea, CRQ Factor 2: psychological status Anxiety, HAD Depression, HAD Factor 3: other HRQoL Symptoms, SGRQ Fatigue, CRQ Emotional function, CRQ Mastery, CRQ Other factors Impact, SGRQ Borg score at the end of the exercise
For definition of abbreviations, see Table 1.

1187

Factor 2 0.10 0.05 0.11 0.06 0.12 0.91 0.86 0.23 0.17 0.35 0.21 0.06 0.00

Factor 3 0.23 0.07 0.24 0.17 0.34 0.00 0.00 0.66 0.66 0.70 0.73 0.44 0.00

1.00 0.85 0.92 0.81 0.55 0.00 0.15 0.28 0.15 0.00 0.06 0.47 0.00

0.48), andthe OCD showed the strongest correlation. Measurement of VO2max were correlated significantly with the dys0.46 to 0.60). There were no significant pnea factor (Rs correlations between RV/TLC and the BDI or Dyspnea in the CRQ. In contrast, the Borg scale at the end of exercise had no significant relationship with FEV1, RV/TLC, or VO2max.

DISCUSSION
This cross-sectional study showed that factor analysis grouped clinical dyspnea ratings (the MRC, the BDI, and the OCD) and two dimensions of the disease-specific HRQoL questionnaires under the same factor, and these five measures performed almost the same in assessing dyspnea in patients with COPD. The Borg scale was found to evaluate a different aspect of dyspnea using the methods mentioned above. HRQoL questionnaires that contain a specific dimension for evaluating dyspnea may be substituted for clinical dyspnea ratings in a cross-sectional assessment. The factors grouped by factor analysis facilitated the selection of individual variables that may most closely represent the conceptual meaning of the composite variable (10). In the present study, it is reasonable to suppose that the MRC, the BDI, the OCD, Activity of the SGRQ, and Dyspnea of the CRQ, which constituted Factor 1, have the same conceptual meaning in common. These five methods may evaluate the same aspect of dyspnea even though different approaches are taken in each method to quantify dyspnea. The BDI and Dyspnea of the CRQ present questions regarding the level of dyspnea that is provoked by patients ADL, and unidimensional measures (the MRC, the OCD) detect the threshold of activities that bring physical limitations caused by breathlessness (16). In

Correlations between dyspnea ratings, HRQoL, and physiological data are shown in Table 3. Selected physiological vari ables were FEV1, RV/TLC, and VO2max representing airflow limitation, hyperinflation, and exercise capacity, respectively. The MRC, the BDI, the OCD, Activity in the SGRQ, and Dyspnea in the CRQ displayed an almost identical pattern in correlations with physiologic data. All of these had significant relationships with FEV1 (correlation coefficient [Rs] 0.31 to

Figure 1. Frequency distribution histograms represent the scores of dyspnea assessments. In the Medical Research Council (MRC) and Activity of the St. Georges Respiratory Questionnaire (SGRQ), higher scores indicate more severe dyspnea. In the Baseline Dyspnea Index (BDI), the Oxygen Cost Diagram (OCD), and the Chronic Respiratory Disease Questionnare (CRQ), higher scores indicate less impairment. Numbers in parentheses represent the theoretical score range.

1188
TABLE 3

AMERICAN JOURNAL OF RESPIRATORY AND CRITICAL CARE MEDICINE

VOL 158

1998

SPEARMANS RANK CORRELATIONS BETWEEN DYSPNEA RATINGS, HRQoL, AND PHYSIOLOGICAL VARIABLES IN 161 PATIENTS WITH COPD
MRC FEV1 RV/TLC V O2max 0.39 0.23 0.60 BDI 0.31 0.14* 0.46 OCD 0.48 0.31 0.56 Activity, SGRQ 0.36 0.25 0.55 Dyspnea, CRQ 0.35 0.19* 0.46
0.01) except the

For definition of abbreviations, see Table 1. All values listed represent statistically significant relationships (p two values indicated with asterisks.

Activity of the SGRQ (8), patients were asked to indicate with yes-no answers whether or not activities such as sitting, walking, or playing sports usually make them feel breathless. Additionally, score distributions showed that the MRC, the BDI, the OCD, Activity of the SGRQ, and Dyspnea of the CRQ have almost the same level of discriminatory power in accordance with a wide range of disease severity despite differences in the number of items, grading scales, and scoring of each method. Dyspnea of the CRQ (9) was developed as an evaluative and not as a discriminative measure because patients quantify their dyspnea during five self-selected activities and the score in this dimension is individualized. However, our study showed that Dyspnea in the CRQ was also discriminative. With regard to the relationships of these methods with physiologic parameters, expected information was obtained such as mild correlations with airflow limitation and moderate-to-strong correlations with exercise capacity (19, 20). These correlations further demonstrated minor differences, and this suggests that these five measures may be similar in the validity. These observations can lead us to suppose that arbitrary choices of measures within the dyspnea factor, that is, the MRC, the BDI, the OCD, Activity of the SGRQ, and Dyspnea of the CRQ may not make any significant difference when clinical cross-sectional assessment for dyspnea is mandatory. The statistical method of factor analysis has been used previously to clarify the relationships between various parameters in patients with COPD, and our results partly support the findings of previous studies. Mahler and Harver (1) demonstrated that clinical dyspnea ratings consisted of an independent factor, using principal-component factor analysis, although they did not evaluate patients with an HRQoL questionnaire. Wegner and colleagues (20) also used factor analysis in patients with moderate-to-severe COPD and demonstrated that dyspnea ratings (the MRC, the OCD, and the BDI) and quality of life evaluated with the CRQ represented one independent entity. It is not clear from their analysis how Dyspnea in the CRQ and the other three dimensions correlated with dyspnea ratings and other parameters because they adopted the total score of the CRQ in their analysis instead of the scores of the four separate dimensions. In fact, Fatigue, Emotional Function, and Mastery of the CRQ were grouped as a different factor from the dyspnea factor in the present study. Incorporating each dimension of the HRQoL questionnaires separately into factor analysis may be more appropriate in providing insight into interrelationships between clinical dyspnea ratings and HRQoL. It is of interest to note that the Borg scale at the end of symptom-limited exercise was found to be a completely separate entity in this factor analysis. This suggests that the Borg scale may evaluate a different aspect of dyspnea from that evaluated by other dyspnea ratings. The reasons for this may

be as follows: the intensity of an exercise task as measured by the ergometer does not reflect patients ADL since they tend to avoid strenuous exercise in their daily life once they recognize it causes dyspnea; lower-extremity symptoms can limit exercise performance and affect the sensation of dyspnea as Mahler and Harver (21) demonstrated; the Borg scale is described as a direct measure since the patients are asked to score sensations while they are actually experiencing dyspnea, whereas patients reflecting on their daily life answer clinical dyspnea ratings indirectly (22). The present study has some limitations. First, the cross-sectional analysis does not address the responsiveness of each measure over time. Second, because our study included predominantly men, generalization of the results to women with COPD may be uncertain. In conclusion, clinical dyspnea ratings (the MRC, the BDI, the OCD) and Activity of the SGRQ and Dyspnea of the CRQ were grouped under the same factor, and these five methods performed virtually identically in evaluating dyspnea in patients with COPD. On the other hand, dyspnea at the end of maximal exercise may provide a different type of information regarding dyspnea. Disease-specific HRQoL questionnaires may be substituted for clinical dyspnea ratings in the cross-sectional assessment. Taken in the light of the growing prevalence of HRQoL and its questionnaires as one measure of outcome, these observations may afford another perspective on the measurement of dyspnea in clinical trials.
References
1. Mahler, D. A., and A. Harver. 1992. A factor analysis of dyspnea ratings, respiratory muscle strength, and lung function in patients with chronic obstructive pulmonary disease. Am. Rev. Respir. Dis. 145:467470. 2. Brooks, S. M. 1982. Surveillance for respiratory hazards. ATS News 8: 1216. 3. Mahler, D. A., D. H. Weinberg, C. K. Wells, and A. R. Feinstein. 1984. The measurement of dyspnea: contents, interobserver agreement, and physiologic correlates of two new clinical indexes. Chest 85:751758. 4. McGravin, C. R., M. Artvinli, and H. Naoe. 1978. Dyspnoea, disability, and distance walked: comparison of exercise performance in respiratory disease. B.M.J. 2:241243. 5. Borg, G. A. V. 1982. A category scale with ratio properties for intermodal and interindividual comparisons. In H. G. Geissler and P. Petzold, editors. Psychophysical Judgment and the Process of Perception. Veb Deutscher Verlag der Wissen Schaften, Berlin. 2534. 6. Stervus, S. S., and E. H. Galunter. 1957. Ratio scales and category scales for a dozen perceptual continua. J. Exp. Psychol. 54:377411. 7. Ware, J. E., and C. D. Sherbourne. 1992. The MOS 36-Item Short-Form Health Survey (SF-36). Med. Care 30:474483. 8. Jones, P. W., F. H. Quirk, C. M. Baveystock, and P. Littlejohns. 1992. A self-complete measure of health status for chronic airflow limitation: The St. Georges Respiratory Questionnaire. Am. Rev. Respir. Dis. 145: 13211327. 9. Guyatt, G. H., L. B. Berman, M. Townsend, S. O. Pugsley, and L. W. Chambers. 1987. A measure of quality of life for clinical trials in chronic lung disease. Thorax 42:773778. 10. Ries, A. L., R. M. Kaplan, and E. Blumberg. 1991. Use of factor analysis to consolidate multiple outcome measures in chronic obstructive pulmonary disease. J. Clin. Epidermiol. 6:497503. 11. American Thoracic Society. 1987. Standards for the diagnosis and care of patients with chronic obstructive pulmonary disease (COPD) and asthma. Am. Rev. Respir. Dis. 136:225244. 12. Harlan, L. C., B. F. Polk, S. Cooper, T. P. Blaszkowski, S. J. Ignatius, M. Stromer, and H. Muller. 1986. Effects of labeling and treatment of hypertension on perceived health. Am. J. Prev. Med. 2:256265. 13. American Thoracic Society. 1994. Standardization of spirometry. 1994 update. Am. Rev. Respir. Dis. 152:11071136. 14. Japan Society of Chest Diseases. 1993. The predicted values of pulmonary function testing in Japanese. Jpn. J. Thoracic Dis. 31:Appendix [in Japanese]. 15. Ikeda, A., K. Nishimura, H. Koyama, M. Tsukino, M. Mishima, and T. Izumi. 1996. Dose response study of oxitropium bromide aerosol on

Hajiro, Nishimura, Tsukino, et al.: Dyspnea and Health-related Quality of Life

maximum exercise performance in stable patients with chronic obstructive pulmonary disease. Thorax 51:4853. Mahler, D. A., G. H. Guyatt, and P. W. Jones. 1997. Clinical measurement of dyspnea. In Dyspnea. D. A. Mahler, editor. Marcel Dekker, New York. 149198. Hajiro, T., K. Nishimura, M. Tsukino, A. Ikeda, H. Koyama, and T. Izumi. 1998. Comparison of discriminative properties among diseasespecific questionnaires for measuring health-related quality of life in patients with chronic obstructive pulmonary disease. Am. J. Respir. Crit. Care Med. 157:785790. Zigmund, A. S., R. P. Snaith, and J. A. Wedzicha. 1983. The Hospital Anxiety and Depression Scale. Acta Psychiatr. Scand. 67:361370. Mahler, D. A., R. A. Rosiello, A. Harver, T. Lentine, J. F. McGovern,

1189

16.

17.

18. 19.

and J. A. Daubenspeck. 1987. Comparison of clinical dyspnea ratings and psychophysical measurements of respiratory sensation in obstructive airway disease. Rev. Dis. 135:12291233. 20. Wegner, R. E., R. A. Jorres, D. K. Kirsten, and H. Magnussen. 1994. Factor analysis of exercise capacity, dyspnoea ratings and lung function in patients with severe COPD. Eur. Respir. J. 7:725729. 21. Mahler, D. A., and A. Harver. 1988. Prediction of peak oxygen consumption in obstructive airway disease. Med. Sci. Sports Exerc. 20:574 578. 22. Jones, P. W. 1997. Dyspnea and quality of life in chronic obstructive pulmonary disease. In D. A. Mahler, editor. Dyspnea. Marcel Dekker, New York. 199220.