Abstract
To evaluate respiratory risk in dairy farmers, the present authors conducted a longitudinal study in the Doubs region of France.
From a cohort constituted in 1986 (T1), 157 (62.8%) dairy farmers and 159 (63.6%) controls were re-evaluated in 1998 (T3). The study protocol comprised a medical and occupational questionnaire, spirometric tests at both evaluations, and noninvasive measure of blood oxygen saturation with pulse oximetry (Sp,O2) at T3.
In 1998, the prevalence of chronic bronchitis was higher in dairy farmers. In cross-sectional analyses, all respiratory function parameters and Sp,O2 were significantly lower in dairy farmers. In a multiple linear regression model, farming, age, male sex and smoking were significantly and negatively correlated with Sp,O2. However, the mean annual decline in respiratory function parameters did not differ significantly between groups. After adjustment of covariables, accelerated decline in vital capacity and forced expiratory volume in one second was associated with age, smoking and male sex. Decline in vital capacity was accelerated in dairy farmers working in traditional farms and those currently foddering.
The current study demonstrates that dairy farming is associated with an increased risk of lung disorders and a decrease in blood oxygen saturation and suggests that respiratory function impairment is correlated with cumulated exposure to organic dusts.
Many epidemiological studies listed in recent reviews 1, 2 have shown significant and consistent associations between agricultural occupational exposure and an increase in respiratory symptoms, especially chronic bronchitis (CB). The effects of exposure on respiratory function are less clear. Cross-sectional studies generally show a moderate alteration in respiratory flow in farmers. Very few longitudinal studies with control groups have been conducted to date 3, 4 and, in most cases, the follow-up is short and the interpretation of the results is difficult 5. Nevertheless, accelerated decline of forced expiratory volume in one second (FEV1) and forced vital capacity has been suggested in grain handlers 6, swine confinement workers 7 and dairy farmers 4, 8. The causes and mechanisms of this respiratory involvement, which are probably multifactorial and complex, have not yet been clearly elucidated 1. Exposure to organic dust may have a determinant role 9. Farmers are exposed to many organic particles that induce inflammatory or allergic reactions of the respiratory tract 10. Theoretically, these lung disorders may have repercussions on gas exchange. To the current authors’ knowledge, only one study conducted in dairy farmers 4 has shown blood oxygenation disorders.
In the present authors’ region, two controlled longitudinal studies were conducted in two different geographical zones 4, 8. A cross-sectional analysis of these studies showed a significant excess in respiratory symptoms and, to a lesser degree, in bronchial obstruction among farmers compared with nonfarming controls 11, 12. These two cohorts were re-evaluated at 5 4 and 6 yrs 8, respectively, and the analysis revealed an excess of respiratory symptoms (particularly in CB) and bronchial obstruction among exposed subjects. These studies did not clearly reveal an accelerated decline in respiratory function among dairy farmers. This might be due to an improvement in working conditions leading to a decrease in agricultural exposure over time, as suggested by studies in grain elevator handlers 13.
The present study concerns the earlier of the two cohorts. In 1986 (T1), 250 dairy farmers were compared with 250 nonexposed control subjects, living in the same rural area 11. The study protocol comprised a medical and occupational questionnaire and spirometry. Identical investigations were conducted on both groups (194 farmers and 155 control subjects) 6 yrs later (T2) 8.
All 500 subjects were re-evaluated in 1998 (T3) with the following objectives. 1) To re-evaluate the excess of lung disorders in dairy farmers 12 yrs years after the initial evaluation, and to confirm their consequences on blood oxygenation in another cohort. 2) To compare changes in respiratory function parameters between groups and analyse the role of agricultural exposure in these changes in the exposed group. Long-term (12 yrs) follow-up and noninvasive measurement of oxygen saturation are the main original properties of the present study.
METHODS
Population
The cohort was formed in 1986 and was composed of 250 dairy farmers of both sexes and 250 administrative employees from various agricultural companies, matched on the basis of their sex, age, smoking status and altitude 11. The 500 subjects were selected from Mutualite Sociale Agricole (MSA) medical files by a method described previously 11. The subjects came from six distinct districts (three located on plains at an altitude of 250–400 m and three located on tablelands at an altitude of 400–700 m). The symptoms and lung function of both groups were compared; the characteristics of the initial population and the results of the cross-sectional comparisons were described in the current authors’ previous study 11.
In 1998 (T3), 87 subjects from the initial cohort were lost to follow-up and seven had died. Among the 220 dairy farmers and the 186 control subjects remaining, each subject (not only those investigated 6 yrs earlier) was contacted individually and invited to participate in the third identical investigation. Subjects who agreed were re-evaluated at the same time of year as the two previous analyses. They were examined between 08:30–11:30 h near their home. The medical questionnaire and spirometric tests were the same as those used in T1 and T2; they were completed along with an occupational questionnaire, blood oxygen saturation and cardiac frequency (fC) measures.
The current study respects ERS principles for research involving humans and received approval from the local ethics committee (Besançon, France).
Questionnaires
Medical questionnaires were sent by mail 10 days before the scheduled medical examination and were collected during check-up examinations. They were read and completed, when necessary (in the case of missing data), by the same investigator as in 1986 (T1) and 1992 (T2). The questionnaire was based on the American Thoracic Society (ATS) questionnaire 14 and the long version of the European Community Respiratory Health Survey questionnaire 15. Questions concerning smoking habits, respiratory symptoms and allergy, as well as the definition of CB, dyspnoea and asthma, have been given previously 11. The occupational questionnaire, used in previous studies 16, 17, was completed during the examination. It was performed by the authors in collaboration with engineers and technicians from the local Dept of Agriculture (Besançon, France) and MSA technicians. The main professional exposure indicators concerned the size of the farm, the size of the herd and the method of storing and drying fodder. Exposure to fodder was assessed in terms of handling, the farm's modernisation and the average time spent in the barn. Fodder handling was measured in two ways, first in terms of actual handling and secondly in terms of bale-yrs. There were three handling categories: 1) never having handled fodder, 2) having stopped handling fodder ≥1 yr before completing the questionnaire; and 3) currently handling fodder. Bale-yrs were defined by the number of average density bales of hay (or equivalent when farmers used other methods of storage) fed to cattle per day, multiplied by the number of years of foddering. The modernisation of farms was classified as follows. Traditional: fodder in bulk or packed in average density bales, no drying method for the fodder and no electric ventilation of the barn. Modernised: fodder packed in round bales, no drying method and no electric ventilation. Modern: drying method of fodder and electric ventilation of the barn. These exposure indicators concerned the winter of 1997 and the 12 yrs of follow-up (T1–T3).
Respiratory function tests
Respiratory function tests were performed by the same investigator as 6 and 12 yrs earlier, according to ATS recommendations 18, with a portable pneumotachograph (Autospiro Minato, AS 500; Medical Science Company Ltd, Osaka, Japan). Three measures of slow vital capacity (VC), FEV1, mean forced expiratory flow between 25 and 75% of forced VC (FEF25-75%) and peak expiratory flow were performed, and the best values selected. The spirometer was calibrated daily for atmospheric pressure, hygrometry and temperature. The values were expressed as absolute values and as percentages of European Community for Steel and Coal reference values, calculated in relation to sex, age and height 19.
Oximetry data
Arterial oxygen saturation (Sa,O2) and fC were evaluated for each subject with a finger pulse oximeter (Onyx® model 9500; Nonin Medical Inc., Plymouth, MN, USA). Patients were seated for ≥15 min before three measurements were taken at 30-s intervals on the left index finger of each subject. The highest value of blood oxygen saturation measured with pulse oximetry (Sp,O2) was retained with the corresponding fC. The right index finger was used in left-handed subjects. In case of an injured nail, another finger on the same hand was used. It was also ensured that subjects were not wearing nail polish; if they were it was removed. Measures were taken for both farmers and controls after spending ≥30 min in heated rooms. The pulse oximeter was tested weekly for accuracy by comparing Sp,O2 with the Sa,O2 of arterial blood gases.
Data analysis
First, a cross-sectional analysis of the 1998 data was performed to compare dairy farmers with control subjects. A multiple logistic regression was used to compare the odds ratios for respiratory symptoms in order to correct for imbalances in age, sex and smoking. Relationships between lung function, Sp,O2 and fC, and farming were assessed using multiple linear regression models. Adjustments were made for smoking (current smoker, ex-smoker or nonsmoker) in the respiratory function model, and for age (as a continuous variable), sex (female = 0, male = 1), smoking (pack-yrs), FEV1/VC ratio (as a continuous variable) and altitude (plain = 0, tableland = 1) in the Sp,O2 models. Farming was added to a model analysing Sp,O2. Secondly, a longitudinal analysis of respiratory function was performed. The effect of farming on the annual change in respiratory parameters between T1 and T3 (1998 value–1986 value/number of years between T1 and T3 examinations) was tested by using a multiple linear regression model adjusted for the 1998 value of farming, age, sex, smoking, altitude and re-evaluation in 1992 (not re-evaluated = 0, re-evaluated = 1). Thirdly, a cross-sectional analysis was performed on T3 data to compare the annual change in respiratory parameters and blood oxygen saturation with occupational exposure indicators in 1998 and their evolution between T1 and T3. Univariate analyses were used, including unpaired t-test and ANOVA for discrete variables and simple linear regression for continuous variables. All tests were two-sided and p-values ≤0.05 were considered statistically significant.
RESULTS
Population characteristics
In 1986 (T1), the initial cohort included 500 subjects. By 1998 (T3), 87 subjects had been lost to follow-up and seven had died. Among the 220 dairy farmers and the 186 control subjects remaining, 316 (77.8%) agreed to participate in the present study and were re-evaluated. Among the 90 subjects who were not re-evaluated, 66 (37 dairy farmers and 29 controls) were contacted by telephone and answered a questionnaire. Reasons for refusal included lack of time, lack of interest in the study and inability to attend the medical evaluation for occupational reasons. Only seven cases refused for medical reasons (five dairy farmers and two controls). Individual characteristics, respiratory symptoms and lung function at T1 were compared between re-evaluated subjects and subjects not re-evaluated. Except for a decrease in females re-evaluated in the control population in 1998 (p<0.05), no significant difference was found for age, respiratory symptoms and function. Among the population re-evaluated in 1998, 246 subjects (133 farmers and 113 controls) had been re-evaluated in 1992. The individual characteristics of the study population at T3 are presented in table 1⇓. Statistically significant differences between both groups were observed for smoking (p<0.05) and altitude (p<0.001). The occupational characteristics in dairy farmers at T3 are detailed in table 2⇓.
Description of the study population in 1998
Occupational characteristics in dairy farmers in 1998
Cross-sectional analyses
The prevalence of asthma, asthma-related symptoms and dyspnoea were the same in both groups. The prevalence of atopy was lower in farmers after adjustment for age, sex and smoking (36.9 versus 50.3% (odds ratio 0.57, 95% confidence interval 0.36–0.91)). Conversely, the prevalence of CB was significantly higher in dairy farmers after adjustment for the same covariables (17.2 versus 5.7% (2.22, 1.09–4.53)). The results of respiratory function and Sp,O2 at T3 are shown in table 3⇓. All respiratory function parameters were significantly lower in dairy farmers. The Sp,O2 was also lower (p<0.001) among dairy farmers, with a mean difference of 0.7%. In a linear multiple regression model, farming, age and smoking were inversely correlated with Sp,O2 (p<0.025; table 4⇓). A positive correlation was found between FEV1/VC ratio and Sp,O2 (p<0.001). Blood oxygen saturation was significantly lower among males (p<0.001).
Respiratory function, arterial oxygen saturation measured by pulse oximetry(Sp,O2) and cardiac frequency (fC) in dairy farmers and controls in 1998
Multiple regression model for arterial oxygen saturation measured by pulse oximetry(Sp,O2)
Longitudinal analyses
The longitudinal analysis of pulmonary function tests was performed on 289 subjects (145 dairy farmers and 144 controls) whose explorations were concordant with the ATS criteria 18 at both T1 and T3. Mean annual changes in respiratory function parameters between 1986 and 1998 are presented in table 5⇓. The mean annual decline in VC and FEV1 was 23.7 and 15.09 mL·yr−1, respectively, for the whole cohort. The decline in lung function was similar in both groups (table 6⇓). The factors associated with an accelerated decline in VC and in FEV1 were age (p<0.001), smoking (p<0.001 and p<0.01 for FEV1 and VC, respectively) and male sex (p<0.001). The subjects re-evaluated in 1992 (133 dairy farmers re-evaluated at T1, T2 and T3) had a lower decline in VC (p<0.001) than those only re-evaluated in 1998.
Mean annual changes in respiratory function parameters between 1986 and 1998
Regression models for annual changes in lung function
Relationship between occupational exposure and respiratory parameters: lung function decline and Sp,O2
Results are detailed in table 7⇓. Dairy farmers who were officially retired showed a significant decline in VC (p<0.001) and FEV1 (p<0.025) and a decrease in Sp,O2 (p<0.001). Decline in lung function was significantly accelerated in dairy farmers working on traditional farms (for VC p<0.05), and in those already having handled fodder (p<0.001). Foddering had a deleterious effect on Sp,O2. The mean duration of exposure was significantly associated with the decline in FEV1 (p<0.05) and with a decrease in Sp,O2 (p<0.025). In a multiple regression model, after adjustment for these same covariables, the decline in VC was accelerated (p<0.001) in dairy farmers who had stopped handling fodder and who were currently foddering versus those who had never handled fodder.
Univariate analysis of annual decline in respiratory function between 1986 and 1998 and occupational exposure
DISCUSSION
The results of the present study are consistent with those of the 1986 and 1992 cross-sectional analyses 8, 11, and with those of other studies conducted in the same region 4, 12. Dairy farmers present a persistent excess of CB with bronchial obstruction. These lung disorders have a significant impact on blood oxygenation. Conversely, an accelerated decline of respiratory function parameters was not observed in this cohort, possibly due to a decrease in exposure with over time.
In the current study, the prevalence of CB was higher in dairy farmers, but the prevalence of asthma was not. In addition, the prevalence of atopy was lower in exposed subjects. These results concur with studies recently reviewed 20. In cross-sectional analyses, all respiratory function parameters were significantly lower among farmers, particularly FEV1 and FEV1/VC ratio, with the same difference between the two groups as 12 yrs earlier at T1 11.
One of the unique properties of the present study was the measurement of Sp,O2 as, to date, there have been no published recommendations for the use of pulse oximeter in epidemiological studies. A recent study evaluating the potential relationship between air pollution and Sa,O2 in elderly adults suggested that the accuracy and reproducibility of Sp,O2 measures allowed the use of a pulse oximeter for epidemiological studies 21. In the current study, a recent pulse oximeter was used, in accordance with the 1995 European regulations. The best of three measures was retained. It was ensured that subjects were not wearing nail polish 22; if they were it was removed. The current results demonstrate that Sp,O2 was lower in dairy farmers (p<0.001). Fodder handling and a longer time spent in the barn were negatively associated with Sp,O2. In a recent study, the first to measure Sp,O2 in farmers, the current authors observed the same difference of Sp,O2 between farmers and nonexposed subjects groups (0.7%), but with a lower level of significance 4. In the current study, the fact that Sp,O2 was, as expected, negatively correlated to smoking and age, and positively correlated to all respiratory function parameters argues for both the accuracy and the relevance of the tool. After adjustment for FEV1/VC ratio, farming remained associated with a decreased Sp,O2; this suggests that mechanisms other than bronchial obstruction are involved. Desaturation can be due to alveolitis (possibly infra-clinical) as in hypersensitivity pneumonitis. Indeed, this affection is frequent in this region of France 23. Dairy farmers are usually exposed to a great quantity of organic particles, including endotoxins, which may induce inflammatory reactions of alveoli and small airways, and therefore alter the alveolo-capillary diffusion 8, 24–26. This may partially explain the decrease in Sp,O2. However, it cannot be totally excluded that more callous hands in farmers play a role in the observed results. As the present study was conducted in winter and spring, pulse oximetry readings may also been affected by physiological processes, such as vasoconstriction in reaction to the cold. However, in the present study, subjects first completed the questionnaires with the physician, then the Sp,O2 measures were performed after at least 30 min spent in a heated room. Furthermore, both farmers and controls were investigated in the same heated rooms. Subjects accustomed to working outdoors also presented with cold acclimatisation in the hands. Among these outdoor workers, cold vasodilatation occurred earlier than in nonacclimatised subjects working indoors 27, 28. Hence, in a cold room, higher Sp,O2 measures should have been observed in farmers than in controls, given earlier vasodilatation due to cold acclimatisation in farmers.
Cross-sectional results in 1986 (T1) showed a significant bronchial obstruction in dairy farmers which might predict an accelerated decline in lung function in exposed subjects. However, no such accelerated decline in farmers was observed, consistent with previous results reported in the European Community Respiratory Health Survey (ECRHS) 29. In the ECRHS, subjects exposed to dust, gases and fumes during the follow-up did not have accelerated decline in FEV1, although they did have an excess of chronic phlegm similar to the present study subjects. Various hypotheses have been stated to explain these discrepancies between longitudinal and cross-sectional results 30–32. In the current study, ill subjects were not lost in follow-up. Respiratory function parameters at T1 in re-evaluated subjects and subjects not re-evaluated, in dairy farmers and controls, were similar. Therefore, a selective follow-up can be excluded, but a selection effect in farmers at inclusion cannot be excluded. In the current study, the mean annual decline was higher between T1 and T2 than between T1 and T3 8. The deceleration in decline might be explained by a learning effect 30, 31 or by systematic errors in measurements of lung function at T2 or T3. In the present study, mean annual changes were positively correlated with smoking and age. This shows the relevance of the current measures and, therefore, current results. Glindmeyer et al. 31 also suggested that past noxious exposures may influence sectional analyses, whereas longitudinal analyses are only sensitive to influences that continue to affect annual decline during the study period. The decrease in respiratory risk exposure over time might explain this absence of accelerated decline. In the current study, occupational characteristics in farmers showed that, between 1986 and 1998, 40% of the farmers claim to have reduced their occupational exposure, 41.4% developed prevention and, in 1998, ∼40% worked on a modern farm. Furthermore, subjects with bronchial obstruction at T1 (1986) might have been more likely to decrease their exposure and to develop prevention. The respiratory function in farmers had probably decreased significantly before 1986, when occupational exposure was greater. Furthermore, during the first 6 yrs 8, the decline of respiratory function parameters in dairy farmers accelerated slightly, whereas it slowed down over the following 6 yrs. The present authors speculate that this type of course over the follow-up period may result from a decrease in the offending exposure. Decline in respiratory function parameters seemed to decrease in dairy farmers foddering in modern farms, especially those with a barn drying system (table 7⇑). Previous studies have suggested that the barn drying fodder system is an important criterion of modernisation, known to lower exposure to aero-contaminants by reducing the proliferation of micro-organisms 16, 33. Another recent study has shown working on a modern farm to be associated with a decrease in risk of lung impairment 34. In addition, accelerated decline in respiratory function was correlated with fodder handling and average time spent in the barn (p<0.001 and p<0.05 for the decline in FEV1 and decline in FEF25-75%, respectively). All these relationships between exposure and lung function impairment fit with the current authors’ speculation that the observed improvement of lung function in farmers is related to a significant improvement in working conditions over time.
Conclusion
The present results demonstrate an excess of chronic bronchitis with obstructive airway disorders and a decrease in blood oxygen saturation in farmers. No significant accelerated decline in respiratory function parameters was found in dairy farmers, possibly due to a decrease in the risk of occupational exposure over time. However, in farmers, an accelerated decline in respiratory function was correlated with fodder handling and the average time spent in the barn.
- Received December 21, 2005.
- Accepted March 20, 2007.
- © ERS Journals Ltd
References
