Introduction The current definition of chronic obstructive pulmonary disease (COPD) associates persistent airflow limitation and chronic respiratory symptoms. Agricultural work has been associated with an increased risk of developing COPD, but the prevalence and definition of the disease vary greatly between studies. This meta-analysis aimed to assess the association between agricultural work and COPD using the most widely used definitions of the disease.
Methods Inclusion criteria were: (1) design: cross-sectional or longitudinal, (2) groups: at least one group of farmers and a control group of non-farmers, (3) outcome: prevalence or unadjusted OR of COPD, airflow limitation and/or chronic bronchitis, (4) study subjects: groups of exposed subjects comprising ≥30 individuals and with a mean age ≥40 years and (5) language: English and French language, full-length, original publications in peer-reviewed journals.
Results In total, 22 manuscripts were included in the meta-analysis. Eight studies assessed only the prevalence of airflow limitation, nine assessed only the prevalence of chronic bronchitis and four assessed the prevalence of both these parameters. Only one assessed the prevalence of COPD according to its current definition, and this study also provided the prevalence of airflow limitation. Ten studies showed a positive association between farming exposure and airflow limitation or chronic bronchitis, and 12 showed no association (OR (95% CI)=1.77 (1.50 to 2.08), p<0.001). Cattle, swine, poultry and crop farming were associated with either airflow limitation or chronic bronchitis.
Conclusion Although some features of COPD are associated with some agricultural work, well-designed studies with appropriate diagnostic criteria should be conducted to draw strong conclusions about the relationship between COPD and farming.
- chronic obstructive pulmonary disease
- chronic bronchitis
Statistics from Altmetric.com
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.
What is already known about this subject?
Agriculture and farming have been associated with increased prevalence of chronic obstructive pulmonary disease (COPD).
This association varies from one agricultural job to another, and data for some agricultural jobs are conflicting.
What are the new findings?
Only a few epidemiological studies use the appropriate definition of COPD for studying the relationship between farming and COPD.
This meta-analysis shows that cattle farming, swine farming and poultry farming are strongly associated with airflow limitation and chronic bronchitis.
Crop/grain farmers have also a higher prevalence of airflow limitation and of chronic bronchitis compared with unexposed controls.
How might this impact on policy or clinical practice in the foreseeable future?
New epidemiological studies investigating different farming groups and using the appropriate definition of COPD are needed.
Chronic obstructive pulmonary disease (COPD) is a common disease with high global morbidity and mortality.1 COPD affects between 2.5% and 19% of subjects aged 40 years or more depending on the definition of the disease and on the country,2 3 and is now the third leading cause of mortality, accounting for 3 million deaths in 2010.4 The last Global Initiative for Chronic Obstructive Lung Disease (GOLD) statement defines COPD as the association of persistent airflow limitation (postbronchodilator (post-BD) forced expiratory volume in 1 s (FEV1)/forced vital capacity (FVC) ratio<0.70) with chronic respiratory symptoms (chronic dyspnoea, chronic cough, chronic wheeze and/or chronic sputum) in subjects with a history of exposure to risk factors for the disease.5 As such, COPD may be regarded as a syndrome with several underlying mechanisms and aetiologies.6 Although the most frequent aetiological factor for COPD in developed countries is tobacco smoking, some studies strongly support the association between several workplace exposures and COPD, with a global population-attributable occupational risk for COPD estimated at 15%–20%.7 8
Beside mining and some non-mining industrial sectors including welding,9 agriculture and farming are among the occupations that have been associated with increased prevalence of COPD.9 10 Nevertheless, this association varies from one agricultural job to another, and data for some agricultural jobs are conflicting. There are at least three possible reasons for these conflicting results. First, agricultural and farming jobs cover a wide spectrum of activities, and for a given farming job, occupational exposures are dictated by the size and type of the farm, differing between traditional family farms and large and modern farm businesses.11 12 Furthermore, recent data suggest that a given farming job (eg, cattle raising) may lead to different exposures depending on the region of exercise.2 Second, the definition of COPD varies widely between studies: in general, older studies use only clinical criteria of chronic bronchitis to define COPD, and only more recent ones use spirometric criteria of airflow limitation (either pre-BD or post-BD).13 Notably, the definition of airflow limitation may also vary between studies, depending on the cut-off used to define an abnormally low FEV1/FVC ratio.14 Thirdly, results may depend on study designs and on methods of population sampling. For example, several cross-sectional studies found that farming jobs exposing workers to organic dusts were associated with increased prevalence of COPD,15 16 whereas a recent meta-analysis including only longitudinal studies concluded that there was limited evidence for a causal association between occupational exposure to organic dust and decline of lung function.17 In addition, many studies did not include a control group of subjects unexposed to any agricultural environment, which may also lead to some bias.
From a workplace health perspective in the agricultural setting, it is very important to identify those jobs that are associated with increased prevalence of COPD. Identification will help to organise both primary and secondary prevention for workers employed in jobs identified as ‘at risk’ of COPD. In view of current knowledge regarding the relationship between COPD and agriculture, we believe that there is a need for a systematic review and meta-analysis aiming to study the association between agricultural work and COPD, taking into account the various definitions of COPD used in the available literature.
The literature search of published population-based studies was conducted using the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA), as currently recommended.18 NCBI PubMed, Embase and the Cochrane Library were searched for English and French-language peer-reviewed publications from January 1980 to July 2018. The research equation was constructed using the PubMed Search Builder and was then translated into Embase and Cochrane languages. The terms used to identify COPD were: ‘chronic obstructive disease’, ‘COPD’, ‘chronic airway disease’, ‘airway obstruction’, ‘chronic respiratory disease’ and ‘lung function’. The terms used to identify agricultural employment and associated exposures were: ‘agriculture’, ‘farm’, ‘occupational exposure’, ‘rural health’ and ‘dust’.
Once the list of articles was available, duplicates were eliminated and the relevant studies were identified by reading first the title, then the abstract and finally the full articles.
Criteria for study selection
The criteria for study selection were the following: (1) design: either cross-sectional or longitudinal. When more than one article analysed the same population (for longitudinal cohorts), only the last published article was retained in the final analysis; (2) groups: at least one group of farmers (a priori considered as the exposed group) and a control group. The control groups had to comprise subjects who worked and/or had worked only in non-agricultural jobs and who had had no occupational exposure of any kind. The exposed group had to be composed of subjects who reported at least one farming activity; (3) outcome: prevalence or unadjusted OR of COPD, airflow limitation and/or chronic bronchitis. Prevalence for both exposed and control groups and/or unadjusted ORs had to be clearly presented in the Results sections of the articles. As most articles that report data on airflow limitation in farmers use the GOLD criterion (FEV1/FVC<0.70), when an article presented prevalence of airflow limitation with both the GOLD and the American Thoracic Society/European Respiratory Society (ATS/ERS) (FEV1/FVC<lower limit of normal (LLN)) criteria, we used only the prevalence defined with the GOLD criterion; (4) study subjects: groups of exposed subjects composed of at least 30 individuals and a mean age of at least 40 years (in order to avoid the risk of confusion between COPD and asthma); (5) language: English and French language, full-length, original publications in peer-reviewed journals.
The quality of the included studies was assessed using the Newcastle-Ottawa Quality Assessment Scale for cohort and cross-sectional studies.19
For all analyses, the item ‘airflow limitation’ was retained according to spirometric criteria (either GOLD or ATS/ERS); the item ‘chronic bronchitis’ was defined by the presence of chronic cough and sputum production for at least 3 months a year for two consecutive years; COPD was a priori defined by the association of persistent airflow limitation with chronic respiratory symptoms (chronic dyspnoea, chronic cough, chronic wheeze and/or chronic sputum). Prevalences of airflow limitation, chronic bronchitis or COPD were collected for farmers and controls, and ORs for each eligible study were computed.
We found only one article in which COPD was defined as the association of persistent airflow limitation with chronic respiratory symptoms.20 In all other selected articles, COPD was defined either according to a spirometric criterion (airflow limitation, either pre-BD or post-BD) or according to a clinical criterion (chronic bronchitis). In some articles, both prevalence of airflow limitation and prevalence of chronic bronchitis were given, but not the prevalence of the combination of persistent airflow limitation with chronic bronchitis.
We performed a first analysis assessing the OR of either COPD or airflow limitation or chronic bronchitis in farmers compared with non-exposed subjects. For this first analysis, in studies providing prevalence of airflow limitation and of chronic bronchitis, only the prevalence of airflow limitation was used, and for the study who provided both prevalence of COPD and airflow limitation, we used the prevalence of COPD. We performed a second and a third analysis assessing the ORs of airflow limitation and of chronic bronchitis in farmers compared with non-exposed control subjects, respectively. These three analyses were displayed using forest plots. Pooled ORs were computed using random effect models. For each analysis, sensitivity analyses were performed for the following types of activities: cattle farmers, poultry farmers, swine farmers, livestock farmers (this term was retained for studies where no details on the animals raised were provided or for studies where farmers raised several types of animals), crop/grain farmers (including farmers exposed to crop protection products) and other farmers (for studies where farming activities were not specified or where subjects had multiple farming activities). As only one study provided prevalence of COPD according to the GOLD definition, this result was not used for the sensitivity analyses.
We assessed heterogeneity between studies using Higgins I2 statistics and publication bias by funnel plots.21 22
All analyses were performed using Review Manager V.5.3 software.
Submission of the research equation to the three databases (Medline, Embase and the Cochrane Library) identified 11 839 manuscripts. After exclusion of 6515 duplicates, the 5324 remaining publications were screened based on title and abstract, leading to exclusion of 4053 based on their title and 1145 based on their abstract. Based on the full-text review of the remaining 126 articles and on the inclusion criteria applied, 22 manuscripts were included in the final meta-analysis: eight studies assessed only the prevalence of airflow limitation, nine assessed only the prevalence of chronic bronchitis and four assessed the prevalence of both these parameters. Only one assessed the prevalence of COPD according to its current definition, and this study also provided the prevalence of airflow limitation (figure 1).
The main characteristics of the 22 included manuscripts are presented in table 1 in chronological order and the quality of each is assessed in online supplementary table 1. Publication bias of each analysis is presented in online supplementary figure 1 and sensitivity analysis by quality score is presented in online supplementary figure 2. The mean±SD year of publication of the 22 included studies was 2004±11, ranging from 1986 to 2017. Studies in which COPD was defined as chronic bronchitis were published much earlier than those defining COPD as airflow limitation (1998±10 vs 2009±9, respectively; p=0.01). The majority of studies were performed in Europe (n=14; 64% of the total), while others were performed in the USA (n=3), Canada (n=2), Nigeria (n=1), India (n=1) and New Zealand (n=1).
Sample sizes of the included studies ranged from 77 to 34 616 subjects, resulting in a total sample size of 63 572 subjects. Seven studies had a group of cattle farmers,2 20 23–27 4 had a group of swine farmers,2 28–30 2 had a group of poultry farmers,2 31 2 had a group of livestock farmers,2 32 6 had a group of crop/grain farmers or farmers exposed to crop protection products2 20 30 33–35 and 10 had a group of mixed farming activities or gave no detail about farming activities.2 25 27 36–42 All studies had a control group of subjects unexposed to any farming activity, the size of these control groups ranging from 39 to 31 935.
Prevalence of COPD defined either by airflow limitation or by chronic bronchitis
After pooling of the 22 studies, the prevalence of COPD defined either by airflow limitation or by chronic bronchitis was 7.5% (948/12 702) among farmers and 2.3% (1149/50 870) among controls (OR (95% CI)=1.77 (1.50 to 2.08), p<0.001). Ten studies showed a positive association between farming exposure and presence of COPD/airflow limitation/chronic bronchitis, and 12 showed no association (see online supplementary figure 3).
COPD defined by airflow limitation/chronic bronchitis was significantly associated with the following: cattle farmers, swine farmers, poultry farmers, crop farmers and mixed groups of farmers, with ORs (95% CI) ranging from 1.59 (1.30 to 1.95) to 2.64 (1.95 to 3.58) (figure 2). In contrast, livestock farming were not associated with either persistent airflow limitation or chronic bronchitis (figure 2).
Overall, heterogeneity between the 22 studies, assessed by I², was high (44%; p=0.01), but was significant only for studies dealing with livestock farmers (I²=77%, p=0.04).
Prevalence of COPD defined by airflow limitation
In the 13 studies that defined COPD according to spirometric criteria, the following different definitions of airflow limitation were used: post-BD FEV1/FVC<0.70 in three studies20 26 41; post-BD FEV1/FVC<LLN in two studies34 42; pre-BD FEV1/FVC<0.70 in two studies31 33; pre-BD FEV1/FVC<0.70 and FEV1<80% of the predicted value in one study40; post-BD FEV1/FVC<90% of the predicted value in one study24; pre-BD FEV1<85% of the predicted value in one study.28 Moreover, two studies gave no details of the spirometric definition of COPD,35 38 and one study used two different definitions of COPD (post-BD FEV1/FVC<0.70 and post-BD FEV1/FVC<LLN).2
The prevalence of airflow limitation was 8.5% (579/6799) among farmers and 4.0% (602/14 981) among controls (p<0.001). The pooled OR of these 13 studies was statistically significant (OR (95% CI)=1.82 (1.53 to 2.16), p<0.001) and heterogeneity between studies was relatively low (I²=18%, p=0.26) (see online supplementary figure 4).
All categories of farming activities had increased risk of airflow limitation in comparison with non-exposed controls: OR (95% CI) ranged from 1.70 (1.34 to 2.15) in mixed groups of farmers to 2.22 (1.39 to 3.56) in swine farmers (figure 3).
Prevalence of COPD defined by chronic bronchitis
A total of 13 included studies defined COPD as chronic bronchitis, with the following definitions for chronic bronchitis: cough and/or expectoration/phlegm/sputum for 3 months of the year or more for at least two consecutive years in 11 studies23–25 27–30 33 34 37 39; self-reported doctor-diagnosed chronic bronchitis32; one study gave no details of the definition of chronic bronchitis.36
The prevalence of chronic bronchitis was 7.3% (497/6815) among farmers and 1.7% (612/37 027) among controls (OR (95% CI)=1.94 (1.46 to 2.56), p<0.001). Heterogeneity between these 13 studies was I²=65% (p<0.001) (see online supplementary figure 5).
Among the different farming activities, exposure was significantly associated with chronic bronchitis in cattle farmers, swine farmers and crop/grain farmers (OR (95% CI)=2.38 (1.75 to 3.25), 2.62 (1.86 to 3.70), 2.41 (1.29 to 4.47), respectively) but not in mixed groups of farmers (figure 4).
The great variability that exists from one epidemiological study to another regarding the definition of COPD in farmers may lead to some misinterpretation of the relationship between farming and risk of the disease. We present for the first time a systematic review and meta-analysis that takes into account these different definitions of COPD used in the literature. We were unable to find a single study that used the current appropriate definition of COPD for investigating this question. We found that most studies published during the 80 s and the 90 s used only clinical criteria to define COPD, while most recent studies used only spirometric criteria. Overall, we found a significant association between the presence of at least one criterion of COPD (either clinical criteria or spirometric criteria) and some farming activities (cattle farmers, swine farmers and poultry farmers). This finding is based on results that were homogeneous from one study to another. In contrast, we found discordant results depending on the definition of COPD for crop/grain farmers and for mixed groups of farmers. Finally, we found that livestock farmers (farmers who raised multiple types of animals or farmers who raised unspecified animal types) had the same prevalence of chronic bronchitis and of airflow limitation as unexposed controls.
Definitions of COPD
Our results highlight the fact that one of the challenges faced in interpreting the currently available literature investigating the relationships between farming and COPD relates to the definitions of the disease. As mentioned above, the most recent GOLD statement defines COPD as the association of persistent airflow limitation (assessed by a post-BD FEV1/FVC ratio<0.70) with at least one of the following respiratory symptoms: dyspnoea, chronic cough and/or chronic sputum production.5 To the best of our knowledge, only one of the currently published studies carried out in farmers used this definition.20 Many studies used an exclusive clinical criterion of chronic bronchitis to define COPD. Unfortunately, it is well established that the main COPD symptoms have poor specificity for the diagnosis of persistent airflow limitation.43 In agreement with this statement, we have recently shown that the items ‘chronic cough’ and ‘chronic sputum’ were not discriminant enough to be retained in a questionnaire aiming to screen farmers for persistent airflow limitation.44 Equally, persistent airflow limitation may develop without any history of the main COPD symptoms.5 The reasons for this are multiple. Some subjects with persistent airflow limitation may be truly asymptomatic if they adopt a sedentary lifestyle, while others may attribute progressively increasing symptoms to growing older.45 In other words, it is likely that a purely clinical definition of COPD, as used in about half the studies included here, is neither specific nor sensitive for the diagnosis of COPD according to the current definition of the disease.
Although persistent (ie, post-BD) airflow limitation is a prerequisite for the diagnosis of COPD, only a minority of studies defining COPD with a spirometric criterion included a post-BD test. This is another major limitation of the available literature. A BD test is necessary to distinguish COPD from asthma, particularly among younger adults.46 47 In addition, some subjects with a pre-BD FEV1/FVC ratio above 0.70 and/or below LLN may have a post-BD value under these thresholds due to an increase in FVC without any further change in FEV1 with BDs.48
Another factor that may influence the epidemiology of COPD in farmers is related to the definition of airflow limitation used in the literature in terms of cut-off for the FEV1/FVC ratio. It is obvious that the fixed cut-off of 0.70 that is recommended by the GOLD committee may be regarded as inappropriate as it underestimates prevalence of airflow limitation in younger people (age 45 years or less) and overestimates prevalence in older people.14 49 In the current review and meta-analysis, we included only subjects aged 40 years or older. It is therefore likely that studies that used the fixed cut-off of 0.70 overestimated the true prevalence of airflow limitation, as the true prevalence is better reflected by the age-dependent LLN cut-off.49
For all the above reasons, we need to keep in mind that the epidemiology of COPD in farmers is currently impossible to determine, and only the prevalence/risk of either clinical or spirometric patterns that may be associated with COPD can be determined with the current meta-analysis.
Respiratory outcomes and farming activities
The increased prevalence of chronic bronchitis and of airflow limitation found in our meta-analysis in farmers raising cattle, swine or poultry is consistent with pathophysiological considerations. Despite a lower prevalence of asthma and allergic rhinitis reported in these populations,50 51 farmers working inside confined buildings, especially poultry and swine workers, are exposed to indoor air contaminants (carbon dioxide, ammonia, dust and endotoxin) which are determinants of COPD traits, with a clear dose-response relationship with the number of hours working inside the buildings.52 53 For example, in a longitudinal study investigating 63 poultry workers, Guillam et al concluded that exposure to only 0.1 mg/m3 of respirable dust present in poultry houses induces a fourfold increase in the risk of chronic bronchitis.54 Moreover, a Danish study comparing decline of lung function between swine workers and dairy farmers showed that working in swine-confinement buildings causes a mean annual decline of 53 mL of FEV1, a value that was greater than in the group of dairy farmers (36 mL; p=0.02 between groups).55
Although their working conditions and exposure differ from those of farmers working in confined spaces, dairy and cattle farmers also have an increased risk of impaired lung function and respiratory symptoms compared with unexposed controls.56–59 The seven studies of cattle farming included in this meta-analysis reported homogeneous results (I²=0%, p=0.61). The four studies with the highest weight showed a significant association between cattle farming and prevalence of either chronic bronchitis or airflow limitation (total weight=94.1%) while the two other studies only showed a trend towards significant association, with a global OR (95% CI) as high as 2.27 (1.81 to 2.86). These results remained unchanged when the prevalence of chronic bronchitis and of airflow limitation were analysed separately. It is noteworthy that this increased risk of COPD traits among dairy farmers may have decreased during the last few years. Our group recently investigated the impact of farm modernity on respiratory outcomes such as airflow limitation and chronic bronchitis in dairy farmers.12 This study demonstrated that working in a modern farm (with a separation between the house and the cowshed and using a loose housing system) decreased the risk of airflow limitation twofold after adjustment for age and for smoking habits.12
Although there are homogeneous data suggesting that raising one specified type of animal increases the risk of COPD traits, we found here that raising more than one type of animal or raising unspecified types of animals was not associated with increased risk of chronic bronchitis and/or airflow limitation. This finding is in contradiction with the conclusions of studies that were not included in the current meta-analysis. In a cross-sectional study of Norwegian farmers, Eduard et al assessed respiratory symptoms and pre-BD lung function in 994 exclusive plant producers and 3741 livestock farmers.60 The authors reported an increased prevalence of chronic bronchitis, airflow limitation and reduced FEV1 among livestock farmers. In another study performed in the Netherlands, living in an area with high density of livestock was shown to be a major risk factor for exacerbations in patients with COPD.61 The lack of significant association found in our meta-analysis between COPD traits and raising more than one animal type or raising unspecified types of animals could be explained by the very small number of studies included in this group (n=2) and the fact that these two studies used different respiratory outcomes. In addition, between-studies heterogeneity was as high as 77% (p=0.04), indicating that these results must be interpreted with caution.
In the current literature, the group of grain/crop farmers is the most controversial regarding the prevalence of airflow limitation and/or chronic bronchitis.62 63 Among the studies included in the current meta-analysis for this group, one study showed a strong association, one study reported a weak and non-significant association and four showed no association between crop/grain farming and prevalence of airflow limitation and/or chronic bronchitis in comparison with non-exposed subjects. The heterogeneity of results between these farmers could be explained by the wide range of activities (grain elevator workers, fruit growers, etc), working methods (use of greenhouse, use of closed and well-ventilated tractors, etc) and the type of climate. When analyses were stratified on respiratory outcome, crop/grain farmers showed a weak but significant association with prevalence of airflow limitation (OR (95% CI)=1.72 (1.01 to 2.94), p=0.049) and heterogeneity between studies was high but not significant (I²=54%, p=0.07). Similarly, crop/grain farming was strongly associated with prevalence of chronic bronchitis (OR (95% CI)=2.41 (1.29 to 4.47), p=0.006). This finding is of importance as some studies used crop/grain farmers as a control group, considering that the risk of COPD in this group was similar to that of unexposed subjects.60 Our observation suggests that crop/grain farmers should be investigated more closely in order to understand the pathophysiology of chronic bronchitis and of airflow limitation found in this group.
In the current meta-analysis, we reviewed the risk of airflow limitation and/or chronic bronchitis among mixed, unspecified groups of farmers. A large number of studies included groups of farmers working in more than one activity, whereas other studies did not describe their type of activities. Among mixed, unspecified groups of farmers, we found a significantly increased prevalence of either airflow limitation or chronic bronchitis (OR (95% CI)=1.59 (1.30 to 1.95)), or airflow limitation alone (OR (95% CI)=1.70 (1.34 to 2.15)), but not of chronic bronchitis alone (OR (95% CI)=1.43 (0.95 to 2.16)). However, it is difficult to interpret this result, given that the definition of exposed groups varies greatly from one study to another.
Quality of the studies included
The sensitivity analysis of all the studies included, ranked according to their quality score, revealed that the OR estimated with the five studies with the best quality scores (OR (95% CI)=1.72 (1.39 to 2.14)) was very similar to the global OR of our meta-analysis (OR (95% CI)=1.77 (1.50 to 2.08)) (see online supplementary figure S2). This result is reassuring since it means that the global OR is mainly based on high-quality studies, rather than on studies whose quality could skew the results obtained.
Additional value to the literature
A review paper recently published by Fontana et al shares some conclusions with ours, especially regarding the negative effects of farming on respiratory health.15 Nevertheless, we believe that our study adds new findings on respiratory health in farmers, for the following reasons. First, our meta-analysis includes a greater number of articles than the review published by Fontana et al (22 instead of 14) and the literature search has been completed more recently (July 2018 instead of December 2016). Second, the major added value of our current study comes from the subanalyses of the relationship between COPD and farming considering different definitions of COPD. As stated in the ’Introduction' section, the definition of COPD varies widely between studies and in general, older studies use only clinical criteria of chronic bronchitis to define COPD, and only more recent ones use spirometric criteria of airflow limitation. In addition, we performed a meta-analysis instead of a systematic review. Combining these two approaches allowed us to demonstrate for the first time a significant (although weak) relationship between crop farming on the one hand, and airflow limitation and chronic bronchitis on the other hand.
Limitations of the study
We acknowledge that our meta-analysis carries several limitations. First, the between-studies heterogeneity could have biased interpretation of some of our results. The global effects of all included studies are significant for the analyses of the criteria ‘either airflow limitation or chronic bronchitis’ and ‘chronic bronchitis’, but both analyses had significant between-studies heterogeneity. This could be due to the large number of studies included and the wide range of exposure and respiratory outcome definitions. Nevertheless, for the criterion ‘airflow limitation’, the pooled OR did not have significant between-studies heterogeneity. This could be explained by the fact that studies using a spirometric criterion to define COPD are more recent, better designed and/or with a more robust outcome than studies using a clinical criterion. Second, some inclusion criteria, in particular, age limit over 40 years and the number of 30 individuals per group, can be regarded as arbitrary. Nevertheless, the age limit of 40 was chosen in order to avoid confusion between COPD and asthma. This confusion may exist when COPD is defined by a pre-BD FEV1/FVC ratio below either 0.70 or age-dependent LLN. In addition, COPD is very infrequent before the age of 40 years. The number of individuals was chosen in order to avoid the inclusion of non-representative studies. The limit of 30 was taken by analogy statistical practice, where a sample of 30 individuals is considered to be representative of the whole group. Third, we excluded many studies that did not fulfil the inclusion criteria of our meta-analysis. Many studies were not included because of the lack of a control group of non-exposed subjects.60 64–70 Others were not included because of the respiratory outcome: a large number of studies assessing respiratory status presented only the mean FEV1/FVC values instead of the proportion of subjects with a FEV1/FVC ratio below 0.70 or below LLN, and some studies reported ‘respiratory symptoms’ with no further detail, so that it was not possible to estimate the proportion of subjects with chronic bronchitis.50 71 72 Fourth, we used only crude OR for respiratory outcomes, without taking into account confounding factors such as gender, age or smoking status. We did so because several studies that we included did not provide ORs adjusted for these confounding factors. We acknowledge that this choice could introduce a confounding bias, particularly when the population of agricultural workers was different to the controls for these factors. Nevertheless, for better interpretation of our results, in this paper we present the distribution of all confounding factors in all the groups of agricultural workers and controls included.
This meta-analysis demonstrates that cattle farming, swine farming and poultry farming are strongly associated with airflow limitation and chronic bronchitis. Our analysis also demonstrates for the first time that crop/grain farmers have a higher prevalence of airflow limitation and of chronic bronchitis compared with unexposed controls. However, the results concerning farmers breeding undefined livestock depend on the chosen respiratory outcome and, perhaps, on the real exposure of the different groups of farmers studied. The current meta-analysis highlights the fact that at this stage, it is not possible to draw firm conclusions regarding the relationships between farming and COPD, as there are very few studies in which COPD has been defined according to appropriate criteria. It seems therefore crucial to conduct new epidemiological studies investigating different farming groups (by job titles and with details regarding type of farming, working activities as well as qualitative and quantitative measurement of airborne exposure) and using the appropriate definition of COPD (ie, persistent airway obstruction according to age-dependent LLN threshold along with symptoms).
Contributors AG and BD has contributed to the literature search, the identification of relevant articles, full-text screening, drafting the article and final approval of the version to be published. TS and J-CD have contributed to revising the article critically for important intellectual content and final approval of the version to be published. AG and BD are responsible for the overall content as guarantor(s).
Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.
Competing interests None declared.
Patient consent Not required.
Provenance and peer review Not commissioned; externally peer reviewed.