INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

In patients with chronic obstructive pulmonary disease (COPD), the relationship between their physiological impairments and the health-related quality of life (HRQoL) measurements is generally weak (1). Quality of life questionnaires can provide information on the patients' status which physiological evaluations cannot (4), and therefore such information should be obtained for the clinical assessments of patients with COPD. Various HRQoL questionnaires have been developed mainly for clinical research, and such instruments have to be comprehensive in order to assess various domains relevant to the functional status, emotional tendencies, social interactions, and symptoms (4). Disease-specific questionnaires focus on those domains most relevant to certain diseases in order to describe disease-specific impairments in the patients' quality of life (5). The St. George's Respiratory Questionnaire (SGRQ) (3) and the Chronic Respiratory Disease Questionnaire (CRQ) (6) are two representative measures, and have been used in many clinical trials. However, these instruments are generally time-consuming and complicated both in calculating the scores and in applying them in a situation where there is limited time. The mean completion times for these questionnaires have been reported to be 10 min for the SGRQ and 20 min for the CRQ (7), not including the time required for scoring.

In light of the growing importance of measuring the HRQoL, a short and simple method may be more important for routine use (8). Quirk and Jones developed a new and simple questionnaire, the Airway Questionnaire 20/30 (AQ20/ AQ30), which was designed to facilitate the application of HRQoL measurements in patients with asthma and COPD (9). The Airways Questionnaire 30 (AQ30) has 30 items with yes/no responses, and the Airways Questionnaire 20 (AQ20) contains 20 of the items present in the AQ30. They reported that the AQ20/AQ30 should take only a few minutes to complete and score. Barley and colleagues recently demonstrated that the AQ20/AQ30 produced a valid evaluation of the HRQoL in patients with asthma, and that the AQ30 had no advantage over the AQ20 (10).

The purpose of the present study was to investigate whether the AQ20 is capable enough of describing the HRQoL of patients with COPD, and to measure longitudinal changes over time. We evaluated the discriminative properties and responsiveness of the AQ20 by comparing it against two representative disease-specific HRQoL questionnaires, the SGRQ and CRQ. We also evaluated the contributions of pulmonary function, exercise capacity, the dyspnea rating, and the psychological status to the HRQoL of COPD patients to highlight the characteristics of the AQ20 (11).

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Cross-sectional Study

Patients. A total of 165 patients with stable COPD, who had at least 6 mo of outpatient management at our institute, were recruited from January 1996 to August 1997. The diagnosis of COPD was based on the definition provided by the American Thoracic Society (12). The entry criteria for this study were as follows: (1) a maximal forced expiratory volume in one second (FEV₁)/forced vital capacity (FVC) ratio of less than 0.7 and an FEV₁ of less than 80% of the predicted value; (2) a smoking history of more than 20 pack-years; (3) no history suggestive of asthma; (4) no exacerbations of airflow limitation over the preceding 6 wk; and (5) no changes in treatment regimen over the preceding 4 wk. All eligible patients finished the following examinations on the same day: pulmonary function tests, a progressive cycle ergometer test, assessments of their HRQoL, and assessments of their dyspnea, anxiety, and depression. Verbal informed consent was obtained from all patients. Their pulmonary function tests, exercise performance, and demographic details are shown in Table 1. One hundred sixty patients were treated with the inhalation of 400 µg salbutamol plus 80 µg ipratropium bromide more than four times a day, and five patients received no medication. Seven patients were managed with long-term domiciliary oxygen therapy.

Pulmonary function tests and symptom-limited progressive cycle ergometer exercise test. The inhalation of bronchodilators was withheld for at least 12 h before the pulmonary function tests were performed. The FEV₁ and FVC were assessed before and at 15 and 60 min after the inhalation of 400 µg salbutamol plus 80 µg ipratropium bromide using a metered-dose inhaler with a spacer device (InspirEase). All spirometric flow-volume curves were recorded according to the method described in the American Thoracic Society 1994 update (13). The predicted values for the FEV₁ and vital capacity (VC) were calculated according to the Japan Society of Chest Diseases' proposal (14). The residual volume (RV) was measured by the closed-circuit helium methods, and the diffusion capacity for carbon monoxide (DL_CO) was measured using the single-breath technique (CHESTAC-65V; 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 bronchodilators, as described by Ikeda and colleagues (15). The workload was increased automatically, and the patients maintained a pedaling frequency above 40 cycles/min throughout the test. The exercise data were recorded using an automated exercise testing system (Desktop Diagnostics/CPX; Medical Graphic Corporation, St. Paul, MN). At the end of each exercise test, the maximal oxygen uptake (O₂max) was calculated.

Assessment of the HRQoL, dyspnea, anxiety, and depression. The HRQoL was assessed by the Japanese versions of the AQ20, the SGRQ (3), and the CRQ (7). These questionnaires were translated into Japanese in accordance with standardized methodology. The Japanese versions of the SGRQ and CRQ have been previously evaluated (11). The translation of the AQ20 followed an established forward-backward translation procedure, with independent translations and countertranslations. The AQ20 uses the first 20 items of the AQ30 with three responses (yes, no, and not applicable) (10). "Yes" responses are scored as 1, and "No" and "Not applicable" are scored as 0. The AQ20 scores range from 0 to 20, with a score of zero indicating no impairment. The total CRQ score, calculated as the sum of the scores on the four dimensions of the CRQ, was calculated for comparison with the other questionnaires, although Guyatt and associates have not introduced the idea of a total score in their analysis of the CRQ (6). To assess dyspnea, the Oxygen Cost Diagram (OCD) (16) was used. The Japanese version of the Hospital Anxiety and Depression Scale (HADS) was used for evaluating the patients' anxiety and depression status (17). The three HRQoL questionnaires, the OCD, and the HADS were self-administrated in the same order in booklet form.

Longitudinal Study

Patients. The criteria for patient selection included a new diagnosis of COPD based on the definition of the American Thoracic Society (12), a smoking history of more than 20 pack-years, and a ratio of the maximal FEV₁ to the FVC of less than 0.7. Patients who visited the emergency room for an acute exacerbation of COPD were excluded form the study. The analytical criteria for the study were as follows: a best FEV₁/FVC for 3 mo of less than 0.7; regular visits; and possibility of treatment and examination. This study was performed as part of our standard outpatient care, and verbal informed consent was obtained from all patients. A total of 82 male and 4 female, consecutive patients with COPD who had been newly diagnosed and completed the following examinations were recruited from our outpatient clinic from April 1996 to April 1998.

Outcome measures. The eligible patients were assessed with the AQ20, the SGRQ, and the CRQ at baseline and again 3 mo after the initiation of treatment. The HRQoL was evaluated when the treatment regimen had not been changed for at least 4 wk prior to the assessment. The three questionnaires were self-administered in booklet form. The patients received the following stepwise treatments in accordance with their symptoms: step 1: 2 or 4 puffs of anticholinergic agents (20 µg per puff ipratropium bromide) plus ₂-agonists (100 µg per puff salbutamol) four times per day; step 2: step 1 plus high doses of inhaled corticosteroids (beclomethasone dipropionate 400 µg four times per day); step 3: step 2 plus daily oral corticosteroids (prednisolone 5 to 15 mg/d). The FEV₁ and FVC were then measured 15 min after the inhalation of the currently prescribed bronchodilators. On each visit, the patients were advised to stop smoking, to increase their physical exercise, and also received education regarding the proper use of the inhalation devices. The patients were also taught how to detect the early signs and symptoms of an acute exacerbation, and how to contact the health care system when an exacerbation occurred.

Statistical Analysis

All of the results are presented as mean ± SD. Changes in the physiological parameters were analyzed with a paired t test. The Wilcoxon sign-rank test was used to analyze the changes in the HRQoL scores. The relationship between two sets of data was analyzed by Spearman's rank correlation test, and a p value of less than 0.05 was considered to be statistically significant.

Calculating Cronbach's coefficient  enabled us to assess the internal consistency. In the cross-sectional study, forward and backward stepwise multiple regression analyses were performed to identify the variables most influential on the HRQoL, as previously described (11). The results of the regression analysis highlighted the characteristics of the AQ20 by comparing it against the SGRQ and CRQ. Variables that were considered to be important for COPD were used as independent variables for the regression analyses. These independent variables were the FEV₁ for airflow limitation, the O₂max for exercise capacity, the OCD for the dyspnea rating, and the Anxiety on the HADS for psychological status. Dependent variables for this statistical model were HRQoL scores such as the AQ20 score, the total score on the SGRQ, and the total score on the CRQ.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Internal Consistency and Score Distributions

The scores on the AQ20, SGRQ, and CRQ are presented in Table 1. The internal consistency of the AQ20 was high (Cronbach's coefficient  = 0.81), and the total scores for the SGRQ and CRQ were also high ( = 0.93, 0.94, respectively). Figure 1 shows the frequency distributions of the scores from the AQ20, SGRQ, and CRQ. Although the SGRQ and CRQ scores were normally distributed, the AQ20 score was skewed toward the milder end of the quality of life scale. The best possible score ("ceiling effect") on the AQ20 was noted in 21 patients (13%), whose average FEV₁ was 1.44 L (53% of the predicted value). The number of patients who had the ceiling effect on the SGRQ and CRQ was 0 and 1, respectively.

View larger version (17K): [in this window] [in a new window]   Figure 1.   Frequency distribution histograms of the total scores on the health-related quality of life questionnaires. In the AQ20 and the SGRQ, higher scores indicate poor health. A higher score indicates better health on the CRQ. The numbers in parentheses represent the theoretical score range.

Comparison of the AQ20 with Other Questionnaires

Table 2 shows that the AQ20 correlated well with its two counterparts. The AQ20 score had significant correlations with the three components and the total score on the SGRQ (Spearman's rank correlation coefficient [r_s] = 0.51 ~ 0.80). There were also significant correlations between the AQ20 and all four domains and the total score on the CRQ (r_s = 0.40 ~ 0.72).

Relationships to Clinical and Physiological Parameters

The values of r_s between the AQ20 and the other variables are shown in Table 3. The AQ20 also showed significant correlations with the DL_CO, O₂max, OCD, and Anxiety and Depression on the HADS. The total score on the SGRQ was significantly correlated with all of the variables shown in Table 3. There was no significant correlation between the total score on the CRQ and the FEV₁.

Stepwise multiple regression analyses were used to identify those variables which may best predict the AQ20 score (Table 4). Variables such as the FEV₁, O₂max, OCD, and Anxiety on the HADS were adopted as independent variables in the regression analyses. The OCD and Anxiety accounted for 43% of the variance in the AQ20 score and in the total score on the CRQ. The OCD, Anxiety on the HADS, and O₂max accounted for 61% of the variance in the total score on the SGRQ.

Responsiveness of the AQ20

Of all 86 eligible patients, 62 received the step 1 treatment (inhaled bronchodilators) at the time of the second evaluation. Seventeen patients received the step 2 treatment (with high-dose inhaled corticosteroids) and two patients received the step 3 treatment (with oral prednisone). None of the patients were given any specific treatment at the initial assessment. The FEV₁ showed a significant improvement over the 3-mo period (45 ± 19% of predicted to 54 ± 19% of predicted, t test p < 0.001). The FVC showed a significant improvement as well (2.39 ± 0.79 L to 2.84 ± 0.79 L, t test p < 0.001). As for the HRQoL questionnaires, the AQ20, SGRQ, and CRQ all showed significant improvements in their total scores (8 ± 5 to 5 ± 4, 43 ± 19 to 37 ± 16, and 98 ± 21 to 111 ± 16, respectively, Wilcoxon sign-rank test; p < 0.01). The scores on each component or dimension of the SGRQ and CRQ were also significantly improved.

The frequency distributions of the change in the scores for the three questionnaires are shown in Figure 2. In general, the AQ20, SGRQ, and CRQ were normally distributed. The change in the AQ20 score showed a moderate correlations with the change in the total score on the SGRQ and CRQ (r_s = 0.56, 0.52, respectively) (Figure 3). The correlation between the change in total score on the SGRQ and the change on the CRQ was also found to be statistically significant (r_s = 0.63). On the other hand, the changes in the scores on the AQ20 and SGRQ had no significant correlations with the changes in the FEV₁ or FVC (r_s = 0.01 ~ 0.20). As for the CRQ, only the change in the Dyspnea dimension was correlated significantly with the change in the FVC (r_s = 0.25, p < 0.05).

View larger version (15K): [in this window] [in a new window]   Figure 2.   Frequency distribution histograms of the change in the scores () on the HRQoL questionnaires. In the AQ20 and the SGRQ, negative changes in the scores represent improvements in the HRQoL. A positive change in the CRQ scores indicates an improvement in the HRQoL. All changes are expressed as a percentage of the scoring range to facilitate comparisons between the questionnaires.

View larger version (15K): [in this window] [in a new window]   Figure 3.   Scatterplots of the change in the AQ20 score ( AQ20) and the change in the total scores on the St. George's Respiratory Questionnaire ( SGRQ) and the Chronic Respiratory Disease Questionnaire ( CRQ). The numbers in the plots represent rs for changes between two questionnaires.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The present study demonstrated that a novel and simple questionnaire, the AQ20, was valid and responsive compared with representative disease-specific questionnaires such as the SGRQ and CRQ. The AQ20 may have similar discriminative properties and responsiveness as the other two questionnaires, and the convenience and simplicity of the AQ20 may facilitate its use in situations with limited time for HRQoL assessments.

The cross-sectional analysis showed that the AQ20 evaluated the functional and psychological aspects of patients with COPD as well as the SGRQ and CRQ. We anticipated certain results on the AQ20 such as the weak correlation with the severity of the airflow limitation and moderate to strong correlations with the dyspnea rating and psychological status, and these were obtained as expected. The pattern of correlation between the AQ20 score and the functional and psychological parameters was found to be very similar to what was noted for the SGRQ and CRQ; these patterns were consistent with our previous report on the SGRQ, CRQ, and the Breathing Problem Questionnaire (11). Multiple regression analysis also suggested that certain characteristics of the AQ20 were similar to those of SGRQ and CRQ; for example, dyspnea and anxiety impacted the HRQoL. These findings indicate that the AQ20, which has fewer items than its counterparts, may maintain the important aspects of instrumental validity on the cross-sectional evaluation.

The responsiveness of the AQ20 was evaluated by comparing it against the SGRQ and CRQ, both of which have already been shown to be responsive (3, 6). The change in the scores on the AQ20 showed a normal distribution, as did the change in the scores on the SGRQ and CRQ, and also was moderately correlated with its counterparts, as shown in Figure 3. Given that the SGRQ and CRQ are standard, disease-specific instruments for evaluating changes in the HRQoL, these findings demonstrate that the AQ20 was responsive. Weak correlations between the changes in the three questionnaires and changes in pulmonary function were noted in the present study; this is consistent with previous reports (3, 6). In addition, Jones and colleagues demonstrated that changes in the SGRQ over time had a weak correlation with most physiological variables (3). Newly diagnosed patients with COPD were enrolled in the longitudinal study because we expected a large change in the quality of life secondary to the medical interventions. We designed the study in this manner to avoid small changes in the quality of life that would be not amenable to statistical analysis. For example, nonsignificant or small changes in the SGRQ and the Sickness Impact profile were recognized in the 1-yr observation (18).

One of the disadvantages of the AQ20 may be a high ceiling effect compared with the other two questionnaires. Because those patients who had a ceiling effect in the present study had milder disease severity, the AQ20 may have less discriminative power for those patients with a mild quality of life impairment than the SGRQ or CRQ, and may not be useful in differentiating the HRQoL among patients with mild COPD. The moderately skewed score of the AQ20 can be explained by its simplicity; there are fewer question items (20 items) and yes/no responses only. For example, the SGRQ contain 50 items and 76 weighted responses, which gives it high discriminative power. The problem of how many questions are sufficient to measure the quality of life has yet to be resolved (8). Nevertheless, the ease and simplicity of the AQ20 may outweigh its disadvantages in the clinical setting, where the major purposes of measuring the HRQoL would be to identify those patients with moderate to severe decrements in their HRQoL, and to identify patients with a worsening of their quality of life in order to target them for intervention.

Some limitations of the present study should be mentioned. First, we tested the Japanese version of the AQ20 in a Japanese population, and concerns as to whether the original English version of the AQ20 behaves in the same manner in patients with COPD are reasonable. However, the original AQ20 has been tested and validated in an English population with asthma (10). We believe that a careful translation and back-translation procedure can produce relevant versions of HRQoL questionnaires, which have similar discriminative properties and responsiveness as original versions (19). Second, because our study included predominantly men, generalizations of the results to women with COPD may be unwarranted. Furthermore, only those patients who were able to perform a progressive cycle ergometer test were enrolled in this study, and therefore some of the more severe patients, with correspondingly worse HRQoL scores, may have been excluded. In the present longitudinal study, we evaluated the ability of the questionnaires to respond to improvements rather than to deteriorations, because comprehensive medical interventions were given to the patients. Therefore, a bias of assessing improvements should be mentioned (19).

In conclusion, the AQ20 is similar to more complex questionnaires such as the SGRQ and CRQ in terms of discriminative properties and responsiveness. This short and simple questionnaire may be useful in studies in COPD patients with limited time for HRQoL assessments. The AQ20 will yield new perspectives on measuring the HRQoL, and on using the results of these assessments for the benefit of the patients.