Association between occupational exposure and the clinical characteristics of COPD

The present study is based on the analysis of cross-sectional data from a previously described cohort of COPD patients recruited between January 2005 and August 2008 in 17 university respiratory medicine departments located throughout France [12]. Respiratory physicians prospectively recruited smokers or ex-smokers (>10 pack-years cumulative tobacco consumption) in stable condition (no history of exacerbation requiring medical treatment for the previous 4 weeks) with a GOLD (Global Initiative on Obstructive Lung Disease)-based diagnosis of COPD (post-bronchodilator FEV₁/FVC ratio < 70%) [13]. Patients with a main diagnosis of bronchiectasis, asthma or any significant respiratory disease other than COPD were excluded. The study was approved by the Ethics Committee of Versailles (France) and all patients provided informed written consent.

We used a standardized case report form that covered demographic data, cumulative tobacco smoking, and clinical and lung function characteristics in stable condition. Symptoms of chronic bronchitis and dyspnea were evaluated using questions derived from the European Community Respiratory Health Survey (ECRHS). Current (during the previous year) history of sputum production and wheezing was assessed [14]. Pulmonary function tests were performed according to international standards [15]. The number of self-reported acute exacerbations of COPD was assessed and exacerbations were classified as mild, moderate (use of antibiotics or oral corticosteroids) or severe (hospitalization or emergency department visit). Health related quality of life was evaluated using the Saint George’s Respiratory Questionnaire (SGRQ) [14].

Prior occupational exposures were assessed by the following ECRHS-derived question: “Have you ever worked in a job which exposed you to vapors, dust, gas, or fumes?” [5]. A history of transient work-related respiratory disability was assessed by the question: “Have you ever had to stop transiently your job because it affected your breathing? “Permanent work-related respiratory handicap was defined by a positive answer to the question “Have you ever had to stop definitely your job because it affected your breathing?”. Patients who reported VDGF exposure were questioned on the job position that they occupied during the longest period of time. Answers were analyzed using a previously published specific job exposure matrix (JEM) that provided a semiquantitative estimate (none, low or high) of exposure to three broad categories of agents: organic dust, mineral dust, or gas/fumes [16, 17].

Patients were categorized (i) as current smokers or ex-smokers (smoking cessation at least 12 months ago) and (ii) as smokers or ex-smokers of ≥ or < 20 pack-years [16].

Patients were considered as atopic when they reported a history of hay fever.

Since some variables were not continuously distributed, data are reported as median [Q1-Q3] for quantitative variables. Percentages were used to describe categorical variables. Univariate comparisons between patients exposed or not to VDGF were performed using Chi-square test for qualitative variables and the two-sided non-parametric Wilcoxon test for quantitative variables.

A p value <0.05 was considered for statistical significance. Analysis was performed with SAS software (version 9.2).

The research protocol was approved by the ethics committee of Versailles (France) and all patients provided informed written consent.

Among the 615 COPD patients included in the cohort, required data were available in 591 (96%). Exposure to VDGF was reported by 209 subjects (35%) and was more common among males than females (90% vs 10%, Table 1). The median age was 65 years, patients reporting VDGF exposure being about 4 years older than VDGF- subjects. Altogether, 27% of patients were still current smokers at inclusion in the cohort. Cumulative smoking was 41 (26–57) pack-years. Current and cumulative smoking were also less in COPD patients with VDGF exposure, despite similar severity of airflow obstruction (see following table). Cumulative smoking and FEV1 (% predicted) were not correlated (r = 0.08). With regard to atopy, hay fever was almost twice more frequent in the VDGF + population. A trend towards a similar association was found for life-time asthma. FEV1 and GOLD stage did not differ between exposed and non-exposed subjects. They were not influenced by the level of cumulative smoking.

Among patients reporting VDGF exposure, job-exposure matrix (JEM) analysis found that 61% were exposed to mineral dusts, 34% to biological pollutants and 73% to gas, vapors or fumes. Altogether, 3% of the whole patient population were exposed to one category of pollutants, 20% to two and 8% to three categories. Current smokers were less represented among subjects exposed to mineral dusts (17% vs 30% in the remaining population) and to vapors, gas and fumes (20% vs 29%).

In terms of symptoms, although chronic bronchitis was not more frequent among VDGF + COPD patients (overall prevalence: 73%), they reported current sputum production more often. Similarly, current wheeze and life-time wheeze were more frequent in patients with VDGF exposure, which was associated with poorer health status (Table 2).

COPD patients with VDGF exposure reported significantly more often an impact of their respiratory situation on their ability to work: 28.5% of VDGF-exposed subjects had to stop working definitely due to breathing difficulties vs 20.3% in non-exposed patients. In addition, 34.7% (vs 25.4%) had to stop working transiently because they experienced respiratory problems. Altogether, only 36.8% of VDGF-exposed subjects (vs 54.3% of non-exposed subjects) reported no impact of their breathing situation on work.

Main differences between VDGF + and VDGF- COPD patients are summarized in Figure 1.

Since the late 1960s, it has been known that, although cigarette smoking is a risk factor for COPD, only 20 to 30% of cigarette smokers develop airways disease, suggesting that genetic susceptibility plays an important role as a determinant of disease occurrence.

In our cohort, occupational exposures were associated with atopy, despite the fact that all COPD patients were smokers or ex-smokers. This observation suggests that atopy and at least some occupational exposures interact to favour the development of the disease. Some decades ago, the Dutch hypothesis suggested that asthma and COPD are different manifestations of a single disease entity, called chronic nonspecific lung disease. It was suggested that endogenous factors (e.g., sex and age), environmental factors (e.g., allergens, occupation, and smoking) and genetic factors (including those predisposing to atopy and airway hyperresponsiveness) all play a role in the pathogenesis of the disease [18, 19].

Indeed, there is evidence that atopy, independently of its association with bronchial hyperesponsiveness, may be associated or have a role in the pathogenesis of COPD [18, 20]. In a general population study, COPD was associated not only with smoking but also with occupational exposures and hay fever [21]. Other studies found an inverse association between atopy, as defined by IgE level, and the FEV1/FVC ratio, independently of smoking status [22]. Some genes, such as those coding IL-13 and IL-17F, might be involved in a global model of shared genetic factors for atopy, asthma and COPD [11, 23]. Lastly, it could be hypothesized that, in predisposed subjects, tobacco smoking may facilitate the development of atopy through its effect on IgE levels [24]. In patients subsequently exposed to occupational sensitizing agents, this may lead to the development of asthma-like features.

Persistent wheeze, which was associated with VDGF exposure in our series of COPD patients, can be a feature of airway hyperresponsiveness (AHR) [25]. AHR is a cardinal feature of asthma [26] and may contribute to the development of COPD [27]. Indeed, several studies found an increase in the risk of COPD in patients reporting a personal history of asthma or AHR [8, 28].

A longitudinal study of males with early COPD suggested that occupational exposure to fumes could be associated with an increased rate of decline of lung function [29]. However, in our population, lung function of exposed patients did not differ from that of patients with no reported occupational exposure. This discrepancy could relate to the lower cumulative smoking in exposed patients, or to an improvement in occupational conditions with aging, related to seniority in the job [30].

In subjects with COPD, exposure to VDGF appears to promote transient or even permanent work loss due to respiratory disability. Only a few studies reported work loss associated with COPD. In the ECRHS, which included adults aged 20 to 44 years, job change due to breathing difficulties at work was reported by 4% of the whole studied population. This figure increased to 11% in subjects reporting either asthma or chronic bronchitis [31]. In the confronting COPD international survey (mean age 63.3 years), more than one third of persons with COPD (35.7%) reported that their condition kept them from working [32]. In a community based cohort study using structured telephone interviews of 234 COPD patients, 25% reported respiratory disability at work and 16% reported both VDGF exposures and respiratory-related work disability [6]. In a patient series of 185 male patients, 34 had become unemployed (18%) due to COPD [9]. Our patient series is in agreement with these studies, with 28.5 % of patients reporting cessation of work due to breathing, which is likely associated to both social and economical consequences.

Assessment of occupational exposure was relatively crude in this study, since self-reported exposure based on a single item is subject to recall bias or subjective influences, in contrast to job-exposure matrix (JEM), which is considered as the “gold standard”. However, previous studies performed in the general population suggest that, when compared to complex assessments such as job-exposure matrices, self-reported exposures are accurate to identify associations between occupation and disease [33]. For instance, in two cohort studies of adults with asthma, self-reported VDGF exposure was fairly sensitive (71%) when compared to JEM-defined exposure [34] and performed well against a checklist of 16 specific exposures [35]. In addition, several studies found increased respiratory symptoms in association with self-reported VDGF exposure [36, 37]. For instance, one study of subjects with established COPD showed that prior exposure to VDGF was associated with increased symptoms over a 1 year follow-up [6]. Finally and importantly, the job-exposure matrix analysis in our subjects reporting VDGF exposure confirmed the reality of exposure. In the present study, statistical analyses using job-exposure matrix were performed to explore the relationship between the type of exposure and the presence of atopy, hay fever, asthma and wheezing; however, due to the relatively low number of patients in each individual category, it was not possible to draw any firm conclusion (data not shown).

Since patients recruited in the real-life observational Initiatives BPCO cohort are all followed in tertiary care centers, they cannot be considered as representative of the general COPD population. In addition, although centers were asked to include all consecutive COPD patients visiting their clinic, it is likely that recruitment was not exhaustive and varied with local resources affected to the study, competitive studies (subjects could not be included in the cohort if they participated to another study) and the effective duration of each center’s participation to the cohort constitution. However, although we cannot formally test the issue of selection bias, it must be noted that (i) the study protocol did not mention any specific guidance on risk factors and (ii) the proportion of patients with occupational exposures is similar to what has been reported in other cohorts. Therefore, a selection bias appears quite unlikely to influence the results of the present analyses.

Pre-bronchodilator FEV1 was not available for all subjects, which prevented us from including reversibility data in the studied variables. Patients exposed to VDGF could have more reversible airway obstruction, corresponding to the asthma-like symptoms (increased frequency of current wheeze) that were reported. However, we have no way to test this hypothesis and, in general, the relationship between symptoms and lung function variables including reversibility is poor. Along the same line, it would have been interesting to compare lung volumes, diffusing capacity, and non-specific bronchial hyperresponsiveness between exposed and unexposed patients. However, these measurements were unavailable in most subjects, as explained by the real-life nature of the cohort. Similar concerns can be expressed for skin prick tests.

A particular aspect of this patient series is that, to limit the risk of including predominantly asthmatic subjects with fixed airflow obstruction, it happened that all centers included only smokers or ex-smokers. This was both a strength since it “secured” the diagnosis of COPD, and a limitation in that it precluded analyzes of the role of occupational exposures in never-smokers. However, it did not prevent from identifying specific clinical features in exposed patients. Similar studies in never-smoking subjects with COPD would be of interest to determine whether our observations are the results of VDGF only, or of VDGF-cigarette smoke interactions.

Another area of interest is the respective weight of risk factors as determinants of the severity of airflow obstruction. Although presented analyses were not initially designed to address this research question, we explored this issue using two multivariate models: one was a logistic regression analysis with GOLD stage as dependent variable, the other was a multilinear regression with FEV₁ (% predicted) as the dependent variable. Smoking was accounted for using cumulative consumption in pack-years and smoking status at inclusion (present vs past smoking). Variables reflecting atopy were patient-reported hay fever and familial history of atopy. With the first model, only less than 2% (R² = 0,018, p value of the model: 0.46) of variations in the GOLD classification could be explained by risk factors. In the second model, R² was even lower (0.0089, with a p value for the model of 0.28). Thus, it can be concluded that, in this population of smokers and ex-smokers, risk factors (age, smoking, occupational exposure, atopy) are very poorly correlated with the severity of airflow obstruction, which makes it impossible to draw any conclusion as to their respective weight.