Chronic Obstructive Pulmonary Disease Stage and Health-Related Quality of Life

Methods

Study Sample

Between April 1993 and July 1994, we recruited all consecutive male patients with clinical symptoms of chronic obstructive pulmonary disease who were attending outpatient respiratory clinics of participating centers. Two university public referral hospitals and one primary health care center for the population of Barcelona, Spain (an urban area); a public referral hospital for the population of Osona County, Spain (a semirural area in Barcelona Province); and a public referral hospital for the inhabitants of Castellon (an urban area) participated in the study.

Inclusion criteria were 1) chronic airflow impairment [defined as FEV₁ < 80% of the predicted value, a ratio of FEV₁ to FVC ≤ 70%, and clinical stability of respiratory disease for at least 1 month before study entry with neither acute clinical decline nor a hospital admission] and 2) an increase in FEV₁ less than both 200 mL and 15% after bronchodilator therapy. The study protocol was approved by the institutional review boards of the participating centers.

Seventeen of 352 patients recruited were ineligible: Nine had airflow obstruction reversibility, 5 had an FEV₁ greater than 80%, 2 had a ratio of FEV₁ to FVC greater than 70%, and 1 was mentally incapacitated. Of the 335 patients who met the inclusion criteria, 14 (4.2%) refused to participate. Thus, 321 patients participated in the study.

Patient Evaluation

We measured FEV₁ and FVC by using standard techniques [3] in the 2 months before or after the patient interview. For 90% of patients, questionnaires were administered and spirometry was performed no more than 23 days apart. Results of blood gas analysis done for diagnostic or therapeutic purposes up to 6 months before study enrollment were obtained from patient medical records; these values were available for 98% of patients with an FEV₁ of 49% of the predicted value or less and 29% of patients with an FEV₁ greater than 49% of the predicted value.

Dyspnea was assessed by using an adapted version of the American Thoracic Society dyspnea questionnaire [4, 5] and a 10-point visual analogue scale [6]. The presence of comorbid conditions was determined by asking patients if they had any of 11 chronic conditions. Social class was assigned according to occupation by using an adapted version of the British Registrar General's Social Classes [7]: class I (professional), class II (intermediate occupations [such as nurse, manager, or schoolteacher]), class III (skilled nonmanual occupations), and classes IV and V (manual occupations).

Most patients completed the Spanish versions of the St. George's Respiratory Questionnaire [8], the Nottingham Health Profile [9], and the 5-item Mental Health Inventory of the Medical Outcome Study 36-item short form health survey [10] on their own. Trained interviewers administered questionnaires to those patients (27%) who had vision problems or were functionally illiterate. Questionnaires were randomly ordered: Half of the study sample responded to the Nottingham Health Profile first, and the other half responded to the St. George's Respiratory Questionnaire first.

The St. George's Respiratory Questionnaire is a standardized questionnaire that is designed to be completed without assistance. It measures health status and perceived well-being in persons with obstructive airway diseases. The Spanish version of the St. George's Respiratory Questionnaire has been shown to be conceptually equivalent to the original instrument and similarly valid and reliable [8]. It contains 50 items (76 levels) divided into three sections: “Symptoms” deals with the frequency and severity of respiratory manifestations, “activity” relates to activities that cause or are limited by breathlessness, and “impacts” covers aspects of social function and psychosocial disturbances that result from respiratory diseases. Scores on the St. George's Respiratory Questionnaire range from 0 (no disturbance of health-related quality of life) to 100 [11]. Mean scores obtained from a sample of persons (n = 74) between 17 and 80 years of age (mean age, 46 years) who had no history of respiratory disease (mean FEV₁, 95%) served as reference values (Jones PW. Scoring Manual of the St. George's Respiratory Questionnaire).

The Nottingham Health Profile is a multidimensional health status questionnaire that has been found to be appropriate for Spanish patients with chronic obstructive pulmonary disease [12]. It contains 38 items divided into six aspects of health (energy, pain, emotional reactions, sleep, social isolation, and physical mobility). A total score on the Nottingham Health Profile is calculated as the proportion of affirmative answers and ranges from 0 (no perceived distress) to 100 (maximum perceived distress). Scores from a representative sample of 610 men older than 40 years of age from the general population of Barcelona served as reference values [13].

Mental Health Inventory scores range from 0 (worst psychological well-being) to 100 (best psychological well-being) [14].

The severity of chronic obstructive pulmonary disease was staged according to the American Thoracic Society guidelines [15] as follows: stage I, FEV₁ greater than 49% of the predicted value; stage II, FEV (₁) 35% to 49% of the predicted value; and stage III, FEV₁ less than 35% of the predicted value. Predicted FEV₁ values were taken from a sample of Mediterranean persons [16]. Categories of PaO₂ included no hypoxemia (PaO₂ >87 mm Hg), mild hypoxemia (PaO₂, 75 to 87 mm Hg), and moderate to severe hypoxemia (PaO₂ <75 mm Hg).

Statistical Analysis

The Kruskal-Wallis test (with correction for ties when necessary) was used to compare health-related quality-of-life scores with clinical and functional categories of chronic obstructive pulmonary disease. The Spearman correlation coefficient (r) was calculated to assess the association between health-related quality-of-life scores and clinical or functional variables. Differences in health-related quality-of-life scores and other continuous variables according to the presence of comorbid conditions were tested by using the t-test. The Statistical Package for the Social Sciences [17] was used for calculations.

Multivariate linear regression was used to identify variables that were associated with total scores on the Nottingham Health Profile and the St. George's Respiratory Questionnaire. Residual values from parametric regression were distributed normally. We used SAS software [18] to assess the adjusted least-squares means.

Results

Demographic and clinical characteristics of the study sample are shown in Table 1. The mean age of the patients was 64.9 ± 9.6 years; more than two thirds of the patients were retired. One hundred thirty-one patients (41%) had stage I disease (mean percentage of predicted FEV₁ ±SD, 62.9% ± 8.4%), 76 patients (24%) had stage II disease (mean percentage of predicted FEV₁, 41.8% ± 4.2%), and 114 patients (35%) had stage III disease (mean percentage of predicted FEV₁, 25.3% ± 6.0%). Eighty-four percent of patients reported at least one coexisting chronic condition; osteoarthritis was the most prevalent (37.7% of patients).

Both specific and generic health-related quality-of-life instruments showed decreased health-related quality of life with increased stage of chronic obstructive pulmonary disease (Table 2). This pattern was shown most clearly and consistently by the St. George's Respiratory Questionnaire scores (Figure 1). In all sections of the St. George's Respiratory Questionnaire, scores were moderately to strongly associated with FEV₁ categories (r = 0.27 to 0.51). Of note, values for patients with stage I disease showed substantial and statistically significant impairment compared with reference values in all sections of the St. George's Respiratory Questionnaire (total score, 34 compared with 6; P < 0.001) and the Nottingham Health Profile (total score, 11 compared with 21; P < 0.001). Dyspnea also decreased in a statistically significant manner as FEV₁ worsened (Table 2). The association between level of hypoxemia and staging categories for chronic obstructive pulmonary disease was statistically significant only for the activity section of the St. George's Respiratory Questionnaire.

View larger version:

• In this window
• In a new window

Figure 1. White bars represent reference values; striped bars indicate stage I disease (FEV > 49% of predicted value); bars striped in the other direction indicate stage II disease (FEV 35%-49% of predicted value); and hatched bars represent stage III disease (FEV < 35% of predicted value). Possible scores ranged from 0 (best possible health) to 100 (worst possible health). NHP = Nottingham Health Profile; SGRQ = St. George's Respiratory Questionnaire. Mean health-related quality-of-life (HRQL) scores (95% CI) in 321 male patients according to the three consecutive stages of chronic obstructive pulmonary disease and reference values.111

Patients with comorbid conditions had dyspnea and gasometric values that were similar to those of patients without comorbid conditions, but they were older and reported greater mental health dysfunction (Table 3). Health-related quality-of-life scores increased (that is, health-related quality of life worsened) with increased stage of disease severity in patients with and without comorbid conditions, but the association between chronic obstructive pulmonary disease stage and health-related quality of life was stronger for patients without comorbid conditions (for the St. George's Respiratory Questionnaire, r = 0.68 for patients without comorbid conditions and 0.42 for patients with comorbid conditions; for the Nottingham Health Profile, r = 0.40 for patients without comorbid conditions and 0.25 for patients with comorbid conditions). This finding indicates that comorbid conditions only partly influence the deterioration of health-related quality of life across the stages of chronic obstructive pulmonary disease.

To assess the interaction between chronic obstructive pulmonary disease staging and comorbid conditions after adjustment for other relevant variables, we calculated the adjusted means of health-related quality of life. As in the bivariate analysis, the influence of comorbid conditions on the total score on the St. George's Respiratory Questionnaire in stages II and III was statistically significant. The effect of comorbid conditions on St. George's Respiratory Questionnaire scores depended on the severity of disease; this relation was not observed for Nottingham Health Profile scores (Table 4 , Table 5).

Discussion

Our results indicate that the effect of chronic obstructive pulmonary disease on health-related quality of life increases as disease severity increases. An unexpected finding was that values for health-related quality of life in patients with stage I disease was substantially lower than reference values. Comorbid conditions may account for some of this impairment, but patients with stage I disease who did not have coexisting chronic conditions had St. George's Respiratory Questionnaire scores that were threefold worse than the reference values. These findings support the FEV₁-based staging system advocated by the American Thoracic Society but counter the widely held opinion that patients with an FEV₁ of 50% or less than the predicted value have minimal impairment in health-related quality of life [1].

Previous studies of the relation between FEV₁ and health-related quality of life had important limitations. Prigatano and colleagues [19], McSweeny and coworkers [20], and Alonso and associates [12] reported that health-related quality of life was not significantly associated with the percentage of predicted FEV₁; however, those investigators used only generic measures of health-related quality of life that may be less sensitive than the St. George's Respiratory Questionnaire to pulmonary symptoms. Okubadejo and coworkers [21] showed a linear relation between the total score on the St. George's Respiratory Questionnaire and PaO₂ but found no association between FEV₁ and the total score on the St. George's Respiratory Questionnaire. However, that study included only 41 patients, and stratification into FEV₁ categories was not possible. Ketelaars and colleagues [22] studied 126 patients with chronic obstructive pulmonary disease by using the St. George's Respiratory Questionnaire but did not include patients with stage I disease [22]. Furthermore, these studies [19-22] reported only mean correlations between the percentage of predicted FEV₁ and health-related quality of life; thus, it was impossible to determine from their results how health-related quality of life is distributed between the stages of severity of chronic obstructive pulmonary disease according to the American Thoracic Society guidelines.

The clinical importance of the degree of health-related quality of life impairment in patients with stage I disease is difficult to assess, although it seems striking from a quantitative point of view. Jones and coworkers [11] have suggested that a four-point difference in the St. George's Respiratory Questionnaire total score may indicate a clinically relevant difference between populations. Using the Asthma Quality of Life Questionnaire, Juniper and associates [23] suggested that a difference in score of about 20% of the theoretical range represents a large relevant change in persons with asthma and that even a 10% difference in score is clinically relevant. In our study, patients with stage I disease who did not have comorbid conditions had a St. George's Respiratory Questionnaire total score of 19 compared with 6 for the reference group of healthy persons (Jones PW. Scoring Manual of the St. George's Respiratory Index), that is, a difference of about 13% of the theoretical range of St. George's Respiratory Questionnaire total score. Although what constitutes a minimal important change in health-related quality of life remains to be definitively established, the lower-than-expected health-related quality of life that we found in patients with stage I disease seems clinically relevant.

Although cut-off points for the percentage of predicted FEV₁ are necessary for a disease staging system, they are arbitrary to some extent. We used the classification criteria proposed by the American Thoracic Society (FEV₁ cut-off points of 50% and 35%) [15], which differ from the FEV₁ cut-off points in the staging categories adopted by the European Respiratory Society (70% and 50%) [2]. Data analysis done by using the European Respiratory Society staging system showed no differences in health-related quality of life scores between patients in the first two European Respiratory Society stages. This suggests that the American Thoracic Society staging system describes the deterioration of health-related quality of life more accurately than does the European Respiratory Society system and that the American Thoracic Society staging system of chronic obstructive pulmonary disease is suitable for research and clinical management.

Nevertheless, as expected, health-related quality of life varied greatly within each stage of severity, even after we stratified for comorbid conditions. As a result, health-related quality of life and FEV (₁) were only moderately correlated. Furthermore, optimization of health-related quality of life is a relevant goal for care of patients with chronic obstructive pulmonary disease. On the basis of current knowledge, we suggest that both FEV₁ and health-related quality of life should be used to evaluate these patients whenever feasible. Further studies should examine whether health-related quality of life is a better predictor of death than is FEV₁ in patients with chronic obstructive pulmonary disease and the extent to which the current criteria for staging can be improved.

Comorbid conditions in patients with chronic obstructive pulmonary disease were very common; in our study, 84% of patients reported coexisting chronic conditions. This proportion is consistent with the results of the Medical Outcomes Study [24], in which 66.5% of patients with chronic lung problems had comorbid conditions. It is interesting to note that the differences in health-related quality of life between patients with and without comorbid conditions decreased as chronic obstructive pulmonary disease stage increased. It should also be noted that most patients with stage I disease reported having comorbid conditions and that the total score on the St. George's Respiratory Questionnaire for this group was 34 points, a score that indicates substantially worse quality of life than that in healthy persons. This observation is consistent with the results of the Medical Outcomes Study [24], which included 731 patients with chronic lung problems recruited at the time of office visits to physicians in three cities in the United States. The authors concluded that the impact of these conditions on health was substantial and involved all aspects of function and well-being.

Despite its high prevalence, comorbidity has often been neglected in clinical studies of patients with chronic obstructive pulmonary disease. Many clinical trials and descriptive studies of health-related quality of life in patients with chronic obstructive pulmonary disease have excluded important coexisting illnesses [25]. Comorbid conditions are rarely specified and probably include severe diseases (such as cancer) rather than more common conditions (such as osteoarthritis). In addition, the extent to which the exclusion of patients with comorbid conditions limited the external validity of these studies has not been addressed. Our results support the view that assessment of comorbid conditions should be included in the clinical management of patients with chronic obstructive pulmonary disease and that future studies should include this type of data as an informative variable or as a potential confounder, depending on the purpose of the study. Obviously, under specific research conditions, exclusion may be a legitimate design option.

Our study had some limitations. First, its cross-sectional design limits our ability to describe how progression of disease stage relates to changes in health-related quality of life. Second, because we recruited patients from outpatient respiratory clinics, the study sample may not represent patients with undiagnosed chronic obstructive pulmonary disease or patients who do not have access to health care. The presence of comorbid conditions may have increased the probability that a patient in stage I would receive a diagnosis and be included in our study sample, as has previously been suggested for patients with cancer [26]. However, the fact that we found substantial impairment of health-related quality of life in patients with stage I disease who did not have comorbid conditions strongly suggests that a bias in recruitment, if present, does not account for our observation. Third, the reference values for the St. George's Respiratory Questionnaire and the Nottingham Health Profile were obtained from different populations. Reference values in the Nottingham Health Profile were obtained from a study of 1220 persons from the general population of Barcelona, Spain [13]; thus, persons with chronic conditions were included. In contrast, the sample used to obtain normal ranges in the St. George's Respiratory Questionnaire was small (n = 74) and composed of healthy persons in England (Jones PW. Scoring Manual of the St. George's Respiratory Index). Finally, because our study included only men, the generalizability of our results to women with chronic obstructive pulmonary disease is uncertain.

Several expert groups in the United States, Europe, and Canada recently released guidelines for the assessment and management of chronic obstructive pulmonary disease. These guidelines recommend various diagnostic and therapeutic options for specific levels of disease. The American Thoracic Society staging system may facilitate categorization of the heterogeneous population of patients with chronic obstructive pulmonary disease and merits widespread attention. Our results show that classifying patients with chronic obstructive pulmonary disease into three stages of severity on the basis of the percentage of predicted FEV₁, as proposed by the American Thoracic Society standards of care, also divides these patients, on average, into three groups with progressively more severe impairment of health-related quality of life. Furthermore, even patients with stage I disease have health-related quality of life that is lower than that seen in the general population. Our study helps validate the American Thoracic Society staging system for chronic obstructive pulmonary disease and supports the idea that this staging system is suitable for use in research and clinical settings. The assessment and management of comorbid conditions in these patients merit more attention in future guidelines.

Dr. Morera: Department of Pneumology, Hospital de Badalona “Germans Trias i Pujol,” Carretera del Canyet s/n, E-08915 Badalona, Spain.

Dr. Marrades: Department of Pneumology, Hospital Clinic i Provincial, Carrer Villarroel 170, E-08036 Barcelona, Spain.

Dr. Khalaf: Department of Pneumology, Hospital la Magdalena, Partida Bovalar s/n, E-12004 Castellon, Spain.

Dr. Aguar: Department of Pneumology, Hospital General de Castellon, Avenida Benicasim s/n, E-12004 Castellon, Spain.

Dr. Plaza: Department of Pneumology, Hospital de la Santa Creu i Sant Pau, Carrer Sant Antoni M. Claret 167, E-08025 Barcelona, Spain.

Dr. Anto: Respiratory and Environmental Health Research Unit, Institut Municipal d'Investigacio Medica, Carrer del Doctor Aiguader 80, E-08003 Barcelona, Spain.