Relationship between disease severity and quality of life in patients with chronic obstructive pulmonary disease

1Departamento de Fisioterapia, Centro Universitário Católico Salesiano Auxilium (Unisalesiano), Araçatuba, SP, Brasil 2Faculdades Adamantinenses Integradas, Adamantina, SP, Brasil 3Departamento de Fisioterapia, Unisalesiano de Lins, Lins, SP, Brasil 4Departamento de Medicina Interna, Faculdade de Medicina de Botucatu, 5Departamento de Bioestatística, Instistuto de Biociências, Universidade Estadual Paulista, Botucatu, SP, Brasil


Introduction
The assessment of health-related quality of life is an important tool for determining the impact of disease severity in patients with chronic obstructive pulmonary disease (COPD) and for monitoring their response to treatment (1).A recent study showed that health status was significantly predictive of mortality in COPD patients, independent of airflow limitation and age (2).
The St. George's Respiratory Questionnaire (SGRQ) and the Airways Questionnaire 20 (AQ20) are examples of specific instruments used for the assessment of quality of life in patients with lung diseases (3,4).The average time to complete and score the SGRQ is estimated to be 10-13 min, compared with 3-4 min for the AQ20 (3,5).Since it is short and can be quickly completed and scored, the AQ20 could be a useful instrument in situations when the time allotted for quality of life assessments is limited.www.bjournal.com.brVarious studies have evaluated the relationship between SGRQ scores and physiological outcomes, as well as between SGRQ scores and systemic markers of COPD.It has been shown that forced expiratory volume in 1 s (FEV 1 ), nutritional status and the 6-min walk test (6MWT) distance present weak correlations with SGRQ scores in COPD patients (6)(7)(8)(9)(10).However, various investigators have reported a strong relationship between dyspnea and SGRQ scores, performing multiple regression analyses in which dyspnea was identified as the greatest contributing factor to variations in questionnaire scores in patients with COPD (6,(8)(9)(10)(11).
The discriminatory power and responsiveness of the AQ20 in comparison with the SGRQ was first established by Hajiro et al. (12) in a study involving 165 patients with stable COPD.However, few studies have evaluated physiological outcomes and systemic markers of COPD in relation to quality of life as assessed using the AQ20 in COPD patients (3,5,12).Therefore, the aim of the present study was to determine the relationship of FEV 1 , body mass index (BMI), fat-free mass index (FFMI), 6MWT results, dyspnea sensation and peripheral oxygen saturation (SpO 2 ) with the AQ20 scores in a cohort of COPD patients.

Patients and Methods
Between August 2004 and August 2005, 99 COPD patients, consecutively recruited from an outpatient clinic, were evaluated.Patients were included if they met the criteria for COPD according to the guidelines established in the global initiative for chronic obstructive lung disease (13): a post-bronchodilator FEV 1 /forced vital capacity (FVC) ratio <0.70, and an increase in FEV 1 after inhalation of a ß 2 agonist <15% or 200 mL.Additional inclusion criteria were as follows: being clinically stable, having no history of infections or exacerbation of respiratory symptoms or changed medication within the 2 months preceding the study onset, and presenting no clinical signs of edema.The following exclusion criteria were applied: inability to perform the lung function tests and 6MWT, inability to understand or complete quality of life questionnaires, myocardial infarction within the preceding 4 months, unstable angina, and congestive heart failure (New York Heart Association class II or IV).All patients were made aware of the proposed study procedures and freely gave written informed consent.All procedures were approved by the Research Ethics Committee of the Universidade Estadual Paulista (UNESP), School of Medicine at Botucatu, in the State of São Paulo, Brazil.

Pulmonary function and peripheral oxygen saturation
Pulmonary function and reversibility tests were per-formed using a spirometer (Ferraris KOKO, USA) according to the criteria established by the American Thoracic Society (14).Values of FEV 1 are reported in liters, in percent of FVC and as percent of reference values (15).SpO 2 was evaluated using an oximeter (Onyx 9500; Nonin Medical, USA).

Nutritional assessment
Body weight and height were measured.BMI (weight/ height 2 ) was also calculated.Body composition was evaluated using a bioelectrical impedance analyzer (BIA 101A; RJL systems, USA).Resistance was measured on the right side of the body with the patient in the supine position in accordance with the European Society for Clinical Nutrition and Metabolism (ESPEN) guidelines (16).Fat-free mass (FFM) was estimated (in kg) using a group-specific regression equation developed by Kyle et al. (17), and the FFMI was also calculated (expressed as FFM/height 2 ).The muscle depletion criteria used was FFMI ≤16 kg/m 2 (for men) or ≤15 kg/m 2 (for women) (18).

Quality of life and baseline dyspnea
Translated versions of the SGRQ and AQ20, both validated for use in Brazil (3,19), were used to evaluate quality of life.Translated versions of the baseline dyspnea index (BDI) (20,21) and of the Modified Medical Research Council (MMRC) dyspnea scale were used to evaluate the degree of dyspnea (22,23).

Six-minute walk test
The 6MWT was performed according to the American Thoracic Society guidelines (24).Patients were instructed to walk, attempting to cover as much ground as possible within 6 min.A research assistant timed the walk, and standardized verbal encouragement was given.Patient SpO 2 was monitored throughout the test.Patients who were hypoxic at baseline and patients whose SpO 2 decreased to <85% during the test were given oxygen by a physical therapist who wheeled an oxygen tank in a handcart alongside the patient.Before and after the test, data were obtained for SpO 2 , heart rate, respiratory rate, dyspnea sensation (Borg scale dyspnea score) and blood pressure.The distance covered is reported in meters.

BODE index and comorbidity index
For the calculation of the BMI/airflow obstruction/dyspnea/exercise capacity (BODE) index, we used the model of Celli et al. (25).For each threshold value of FEV 1 , 6MWT distance, and MMRC dyspnea grade, the patients received points ranging from 0 (lowest value) to 3 (maximal value).For BMI, the values were 1 (BMI ≤21 kg/m 2 ) or 0 www.bjournal.com.br(BMI >21 kg/m 2 ).The points of each variable were added so the BODE index ranged from 0 to 10 points for each patient.The Charlson index was used to determine the degree of comorbidity (26).

Statistical analysis
Results are reported as mean ± SD unless otherwise indicated.Correlations between variables were assessed using Spearman's rank order correlation.Multiple linear regression analysis was performed using SGRQ and AQ20 scores as dependent variables.Variables significantly associated with quality of life score or included in the BODE index (FEV 1 , BMI, 6MWT distance, MMRC grade) were used as explanatory variables.Two models of multiple regression analysis were used.In model 1, all variables tested in the correlation study, except the BODE index, were included.In model 2, the BODE index replaced BMI, FEV 1 , MMRC grade and 6MWT.Due to the high colinearity between BDI and MMRC scores (r > 0.60, P = 0.024), they were alternatively included in the regression model.Results of the multiple regression analysis are reported in terms of the estimated regression coefficient with corresponding confidence intervals.Values of P < 0.05 were considered to be statistically significant.Statistical analysis was performed using the Sigma Stat statistical software (version 2.03, SPSS, Inc., USA).

Results
Patient characteristics are shown in Table 1.The degree of COPD was classified as very severe in 25 patients (25.3%), as severe in 17 (17.2%),as moderate in 37 (37.3%) and as mild in 20 (20.2%).Mean age was 64.2 ± 9.2 years, and mean FEV 1 was 60.4 ± 25.2% of predicted.As shown in Table 1, mean BMI values were within the normal range; however, individual analyses showed that 15% of the patients presented BMI <21 kg/m 2 , and that 44% of them presented FFMI values lower than the minimum (≤16 kg/m 2 for men or ≤15 kg/m 2 for women).Patients with very severe COPD presented significantly lower SpO 2 , reduced exercise capacity (assessed using the 6MWT), and greater dyspnea sensation compared with mild COPD patients.
Table 3. Table 3. Table 3. Table 3 sity, assessed using the BDI or MMRC scale (model 1) as well as using the BODE index (model 2), remained predictive of AQ20 and SGRQ scores.Dyspnea scores explained 38% of the AQ20 and 36% of the SGRQ variance.

Discussion
The main finding of this study was that the degree of dyspnea, as assessed using the BDI or MMRC scale, was the best predictor of AQ20 scores.The multidimensional (BODE) index, which incorporates BMI, FEV 1 , 6MWT distance and MMRC grade, although presenting a lower predictive value than those obtained for each dyspnea measure in isolation, was also shown to be an independent predictor of AQ20 score.To our knowledge, this is the first time that a relationship between the BODE index and AQ20 scores has been established.In addition, our data show that, for the SGRQ, the predictor factors and the tion variables correlated significantly with SGRQ or AQ20 scores.
In the multiple regression analyses of the potential predictors of health-related quality of life, dyspnea intenwww.bjournal.com.brcorrelations identified between clinical variables and quality of life scores were very similar to the findings for AQ20.
Only two previous studies have evaluated the contribution of dyspnea sensation to AQ20 scores (3,12).One evaluated the discriminatory power and responsiveness of the AQ20 compared with the SGRQ and Chronic Respiratory Questionnaire (CRQ), showing that dyspnea and anxiety accounted for 43% of the variance in AQ20 and CRQ scores (12).In a study performed to evaluate the linguistic and cultural translation of the AQ20 into Brazilian Portuguese, a statistical correlation between the BDI and AQ20 score was shown for the first time (3).The authors found that the strength of that correlation ranged from -0.36 to -0.64.
The results of the present study showed that MMRC and BDI scores are responsible for approximately 38% of the AQ20 and 36% of the SGRQ score variance.This is consistent with the results of studies showing that dyspnea perception can explain between 25 to 54% of the quality of life score variance in COPD patients (10,27).In a recent study assessing dyspnea via the Oxygen Cost Diagram (OCD) and quality of life via the SGRQ in 83 patients with COPD (10), a significant correlation between OCD-assessed dyspnea and SGRQ score was detected (r = 0.50, P < 0.001).Using multiple regression analysis, the authors showed that dyspnea was the greatest contributing factor to the total SGRQ score, explaining 25% of the SGRQ total score.Using factor analysis, González et al. (27) found that dyspnea scale scores correlated with the scores for all dimensions of the CRQ.The authors used multivariate regression analysis and demonstrated that BDI was the best predictor of CRQ score, explaining 54% of the quality of life variance in 40 emphysema patients.In addition, other studies have shown a strong relationship between SGRQ-assessed quality of life and dyspnea perception (9,28).
In the present study, although the FEV 1 and 6MWT distance presented a significant correlation with all domains of the SGRQ and AQ20 scores, in the multiple regression analysis, both variables were ruled out as predictive variables.It has been shown that FEV 1 correlates weakly or not at all with SGRQ and AQ20 scores, and that, in multiple regression analyses, FEV 1 is not identified as a predictor of health-related quality of life in COPD patients (3,6,29,30).Some studies have shown a significant correlation between 6MWT distance and quality of life scores (2,7,27,29).The results of a recent study confirm the relationship between 6MWT distance and SGRQ score (10).However, consistent with our results, the authors of that study found that 6MWT distance was ruled out as a predictive variable in the multiple regression model.
In the present study, however, neither BMI nor FFMI correlated significantly with SGRQ or AQ20 scores.This is in agreement with a study showing no relationship between AQ20 scores and BMI (3).However, significant correlations between body composition attributes and SGRQ have been found in patients with COPD (9,10).Assessing 50 patients with COPD, Shoup et al. (9) showed that underweight patients presented higher scores (indicating lower quality of life) in the activity and impact domains of the SGRQ, as well as higher total SGRQ scores, than did patients presenting normal body weight.The authors showed that low FFM was also associated with greater impairment of quality of life.However, when dyspnea was added to the model, neither weight nor FFM remained significantly related to quality of life.A recent study assessing 49 normal-weight and 34 underweight COPD patients showed that scores in all SGRQ domains, as well as total SGRQ scores, were significantly higher in the underweight group; in addition, BMI correlated significantly with all SGRQ scores except the SGRQ symptom score (10).However, stepwise multiple regression analysis used to identify variables that predict SGRQ total score showed that dyspnea assessed by OCD was the greatest contributing factor to the total SGRQ score (10).
The BODE index, which includes physiological variables other than dyspnea and has been shown to be superior to FEV 1 alone in predicting risk of death among the COPD patients, is also a better predictor of hospitalization in these patients (25,31).In the multivariate analyses conducted in the present study, the BODE index presented a lower coefficient than did MMRC grade or BDI.Nevertheless, it did correlate significantly with the quality of life questionnaire scores and was shown to be an independent predictor of health-related quality of life.
A limitation of the present study was that other factors known to be associated with impaired quality of life in COPD patients such as anxiety and depression were not assessed (32).
The relationships between AQ20 score and markers of disease severity are similar to those previously described for the SGRQ.Therefore, the AQ20, a simple and rapidly applied instrument to evaluate the general impact of disease, can be very useful when the time allotted for measurement of quality of life is limited.

Table 1 .Table 1 .Table 1 .Table 1 .Table 1 .
Characteristics of the patients with chronic obstructive pulmonary disease (COPD) as a function of disease stage.

Table 3 .
. Multiple regression analysis with the Airways Questionnaire 20 score as the dependent variable.