Respiratory symptoms in adults are related to impaired quality of life, regardless of asthma and COPD: results from the European community respiratory health survey

The ECRHS I [20] was undertaken in 1991 through 1993, and comprised representative population samples of adults aged 20-44 years in 48 centres in 22 countries. Altogether 137 619 subjects completed a postal questionnaire on respiratory health, giving a median response rate of 78% (stage I). Randomly selected representative subsamples of subjects from 36 centres participated in interviews and standardised clinical testing (ECRHS I stage II). The ECRHS II [21] was a follow-up study of respondents to ECRHS I stage II, conducted from 1998 through 2002. Of the invited, a total of 11 168 responded (response rate 76%). The respondents completed interviews and standardised clinical testing with blood tests, spirometry, and methacholine challenge. A total of 6009 subjects from 24 centres in 11 countries completed data regarding respiratory symptoms, HRQoL, and lung function in the ECHRS II. The regional ethics committees approved the study, and all participants signed informed consent. The full protocol is available at http://​www.​echrs.​org.

Information on respiratory symptoms was obtained in ECRHS II by questions 1-11 listed in appendix 1. Asthma was defined as ever having had asthma, question 12 in appendix 1 (ECRHS II).

Smoking was categorized as never-, ex- and current, where the current smokers were the ones that reported having smoked during the last month in the ECRHS II. The longest held job during the follow up period was used as a marker of socioeconomic status, and classified into 5 groups as 1) managers and professionals; 2) technicians; 3) other non-manual workers; 4) manual workers; and 5) non-classified workers. Data on comorbidities were obtained by the questions listed in appendix 2. Presence of comorbidity was included as a dichotomized variable.

The Short Form-36 (SF-36) questionnaire was used to assess HRQoL in the ECRHS II. The SF-36 is a standardized, widely used and validated general HRQoL questionnaire, which is translated into several languages [22]. It includes 36 questions from which the physical component summary (PCS) and the mental component summary (MCS) were computed as described in the SF-36 User's Manual [23]. The scores were transformed so that the mean score was 50 and the standard deviation was 10 [24]. A higher score indicates better quality of life. In clinical trials, a 5 point change in the MCS and PCS are regarded as clinically relevant, whereas statistical significance can be reached with changes in 1-2 points in the scores.

Pre-bronchodilator lung function (forced expiratory volume in one second (FEV₁) and forced vital capacity (FVC)) and BHR were measured by standard methods [21]. BHR was defined as a 20% reduction in FEV₁ (PD₂₀) after methacholine challenge, with a maximum cumulative methacholine dose of 2 mg [20]. COPD was defined and classified according to the Global initiative for Obstructive Lung Disease (GOLD) [25], i.e FEV₁/FVC < 0.70. Because the number of subjects with FEV₁ < 50% predicted were small (COPD GOLD stage III and IV), these were grouped together. Specific IgE levels were measured with the Pharmacia CAP System [26], and atopy was defined as assay results > 0.35 kU/L for any of the following allergens: house dust mite (Dermatophagoides Pteronyssinus), timothy grass, cat and mould (Cladosporium Herbarum).

The statistical software Stata version 10.1 was used for the statistical analyses [27]. The bivariate relationships between the outcome variables (PCS and MCS) and the respiratory symptoms were described by means and standard deviations and tested with Kruskal Wallis and Wilcoxon rank-sum tests. Nonparametric test were used due to skewed distributions of the SF-36 scores. The relationship between the SF-36 scores and each of the symptom variables were investigated in robust multiple linear regression models, with adjustments for age, gender, smoking, occupation, body mass index (BMI), comorbidity, and study centre. Stratified analyses were performed with respect to diagnosis (neither asthma nor COPD; and asthma and/or COPD). Respiratory symptoms often occur together, causing problems of co-linearity if included together, in a mutually adjusted model. Therefore, we used exploratory factor analysis, where symptoms with high level of covariance were summarized in a factor. This yields an estimate of the association between having one of the symptoms in the factor, regardless of which, and HRQoL. The factors substituted the observed symptom variables in multiple linear regression analysis with respect to the SF-36 scores, adjusted for atopy, BHR, and FEV₁ in addition to age, gender, smoking, occupation, BMI, comorbidity, and study centre.

One of the advantages of using a general HRQoL questionnaire like the SF-36 is that it allows for comparison between different populations. In a Norwegian study where the HRQoL measured by SF-36 of patients with rheumatoid arthritis, epilepsy, angina pectoris, asthma and COPD were compared, and the patients with COPD and rheumatoid arthritis had the lowest physical scores [28]. The mean PCS of patients with COPD was 45, in other words in the same magnitude as the score decrements as the respiratory symptoms investigated in the present study. In a German study evaluating the HRQoL of patients in general practices, the adjusted mean PCS decrements of patients with COPD and asthma were 4.2 points, also comparable to the impairment of PCS scores we found in our study [29]. Our study elaborates on the association between respiratory disease and HRQoL by highlighting the importance of respiratory symptoms, and not only diagnosed disease in this association.

Chronic respiratory symptoms can be caused by other conditions than asthma and COPD. Obesity, heart diseases, hypertension with congestive heart failure, and cancer with metastasis to the lungs can have similar symptomatology. In the current study, both BMI and comorbidities were included as adjustment variables in the analyses. Of the 6009 subjects, 45.8% had no comorbidities 33.2% had one or more, and in 21.0% data on comorbidities were missing. To evaluate if the observed associations between respiratory symptoms and HRQoL could be due to residual confounding by comorbidity, the analyses presented in additional file 1 were also conducted in subjects without comorbidities. In this subgroup, there were still significant associations between respiratory symptoms and HRQoL, both in subjects with and without asthma and/or COPD (data not shown).

In the current analyses, we chose to join self-reported asthma and spirometrically defined COPD in one group. Firstly, aging is an established risk factor for COPD [17, 30, 31]. The risk of developing COPD is substantially higher in subjects over 45 years of age [17] compared to younger subjects, and the majority of our study population was under 45 years. Secondly, among our study participants, 336 subjects (5.6%) had airway obstruction measured with pre-bronchodilatator spirometry, and the majority of these had FEV₁ > 80% of predicted. Most likely, a large number of these subjects have reversible airway obstruction [32], indicating asthma and not COPD. Thirdly, there where 102 subjects that reported ever having had asthma in addition to having FEV₁/FVC < 0.70. Thus, in a population study as this classification error between asthma and COPD is difficult to avoid. Since the main focus of the study was the subjects without obstructive lung disease, we chose to group asthma and COPD together.

A priori, respiratory symptoms without a known cause will probably be more worrying than symptoms with a recognized pathological origin. A study of asthma and COPD patients in the Netherlands showed that knowledge about both asthma and COPD was related to better adaptation to these diseases [33]. Knowledge of disease is likely to vary in the group of asthma and/or COPD in our study, since asthma was self-reported and COPD was based on spirometric measurements of airflow limitation. To account for an eventual positive effect of knowledge of disease we conducted analyses stratified by asthma and COPD. However, the HRQoL scores did not differ significantly between the subjects with self-reported asthma and respiratory symptoms and the subjects with spirometrically confirmed COPD and respiratory symptoms.

Some symptoms had a greater impact on HRQoL than other symptoms. Subjects with breathlessness without asthma and/or COPD had lower both PCS and MCS than subjects with asthma and/or COPD. It could be that breathlessness indicated by subjects without asthma or COPD represents a different sensation than in subjects with asthma and/or COPD; dyspnoea is a term describing many sensations [34], and the underlying mechanisms of dyspnoea are many [35].

Subjects with cough/phlegm symptoms and asthma/COPD had lower PCS than subjects with these symptoms and neither asthma nor COPD. This difference might be explained by the frequency and intensity of these symptoms, which most likely are higher in subjects with asthma and/or COPD. The symptom questionnaire did not assess symptom frequency and intensity, so this remains unsolved. COPD is a heterogeneous disease, and the classical way of describing the different phenotypes have been through the Snider venn diagram. Snider described the variety of obstructive lung disease is through the phenotype with asthma, the one with chronic bronchitis, the one with emphysema, and varying degrees of overlap between the three phenotypes. The hallmark symptoms of the chronic bronchitis phenotype often are cough and phlegm. The decreased PCS accompanying the cough and phlegm in our analyses of subjects with asthma and/or COPD may indicate that subjects with the chronic bronchitis phenotype have lower HRQoL than the asthma and emphysema phenotypes.

When atopy, BHR and FEV₁ was included in the regression model with the symptom factors, the regression coefficients did not change overtly. Garcia-Marcos et al evaluated HRQoL in asthmatic children and found that atopy was not a risk factor for decreased HRQoL [36]. We have found no previous general population studies investigating BHR in the context of HRQoL. Better lung function was related to higher HRQoL as found in another study [31], but the difference in FEV₁ associated with one point change in the PCS score was more than 1000 ml. In other words, a change in PCS score barely statistically significant corresponded to five times the FEV1 change regarded as clinically relevant spirometrically.