Physiological and psychosocial occupational exposures and intermediate health outcomes in
the general population
Faruque, Md Omar
DOI:10.33612/diss.154938193
IMPORTANT NOTE: You are advised to consult the publisher's version (publisher's PDF) if you wish to cite from it. Please check the document version below.
Document Version
Publisher's PDF, also known as Version of record
Publication date: 2021
Link to publication in University of Groningen/UMCG research database
Citation for published version (APA):
Faruque, M. O. (2021). Physiological and psychosocial occupational exposures and intermediate health outcomes in the general population. University of Groningen. https://doi.org/10.33612/diss.154938193
Copyright
Other than for strictly personal use, it is not permitted to download or to forward/distribute the text or part of it without the consent of the author(s) and/or copyright holder(s), unless the work is under an open content license (like Creative Commons).
Take-down policy
If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately and investigate your claim.
Downloaded from the University of Groningen/UMCG research database (Pure): http://www.rug.nl/research/portal. For technical reasons the number of authors shown on this cover page is limited to 10 maximum.
19
Chapter 2
Occupational exposures and genetic susceptibility to occupational exposures are related to sickness absence in the Lifelines Cohort Study
M.O. Faruque K. de Jong J.M. Vonk H. Kromhout R. Vermeulen U. Bültmann H. M. Boezen
20
ABSTRACT
In this cross-sectional study, we investigated the association between occupational
exposures and sickness absence (SA), the mediating role of respiratory symptoms, and whether genetic susceptibility to SA upon occupational exposures exists. Logistic regression was used to examine associations and structural equation modelling was used for mediation analyses. Genetic susceptibility was investigated by including interactions between occupational exposures and 11 candidate single nucleotide polymorphisms (SNPs). Biological dust, mineral dust, and pesticides exposure were associated with a lower prevalence of any SA (OR(95%CI)=0.72(0.58-0.89), 0.88(0.78-0.99), and 0.70(0.55-0.89), respectively) while gases/fumes exposure was associated with a higher prevalence of long-term SA (1.46(1.11-1.91)). Subjects exposed to solvents and metals had a higher prevalence of any (1.14(1.03-1.26) and 1.68(1.26-2.24)) and long-term SA (1.26(1.08-1.46) and 1.75(1.15-2.67)). Chronic cough and chronic phlegm mediated the association between high gases/fumes exposure and long-term SA. Two of 11 SNPs investigated had a positive interaction with exposure on SA and
one SNP negatively interacted with exposure on SA. Exposure to metals and
gases/fumes showed a clear dose-response relationship with a higher prevalence of long-term SA; contrary, exposure to pesticides and biological/mineral dust showed a protective effect on any SA. Respiratory symptoms mediated the association between occupational exposures and SA. Moreover, gene-by-exposure interactions exist.
21
INTRODUCTION
Sickness absence has a negative effect on production both qualitatively and quantitatively [1]. Studies from different countries and various occupational settings have shown that many factors such as socio-demographic and personal factors, physical and psychosocial working conditions as well as somatic and mental health conditions, are associated with sickness absence [2]. However, whether occupational exposure to biological dust, mineral dust, gases/fumes, pesticides, solvents, and metals is associated with sickness absence in the general working population, has yet to be elucidated.
Several studies have shown that occupational exposure to vapours, dust, gases and fumes (VGDF), organic dust, chlorinated solvents, lead, and occupational chemicals such as detergents, surfactants or pesticides, increases the prevalence of respiratory symptoms (e.g. cough, dyspnea), respiratory diseases, lung function limitation, dizziness, anxiety, abdominal pain, and skin irritation and lesions [3–5]. On inhalation, occupational exposures may impair lung function by triggering immune or inflammatory responses [6-8]. Indeed, a previous study has found that bioaerosol inhalation induced inflammation (increased neutrophils and interleukin-8 level) in the small airways, which in turn reduced lung function among organic waste collectors [9]. VGDF exposure was also strongly associated with a higher prevalence of sickness absence among workers with respiratory symptoms [10]. Therefore, it can be hypothesized that the prevalence of sickness absence is higher among people in “dirty” jobs (e.g. welding, painting, and construction) compared to people with a clean working environment, because airborne occupational exposures may affect the respiratory system. However, not all workers experience respiratory symptoms upon occupational exposure, and also the symptom severity differs between individuals [11-12]. Genetic make-up may play a role in the differential susceptibility to these exposures. Indeed, we have previously shown that specific single nucleotide polymorphisms (SNPs) in biologically plausible genes were associated with the susceptibility to occupational exposures with regard to respiratory health effects, i.e., lung function level [13,14]. For example, subjects carrying the minor allele of SNP rs17490056 had a lower
FEV1 compared to wildtype subjects, yet only in those subjects with high biological
dust exposure and not in subjects with low or no exposure [13]. These SNPs may be plausible candidates to modify the association between occupational exposures and sickness absence.
The main aim of this study was to investigate the association between occupational exposure to biological dust, mineral dust, gases/fumes, pesticides, solvents, and metals, assessed with ALOHA+ job-exposure matrix (JEM) [15], and self-reported
22
sickness absence in active workers in a general population cohort. We further investigated whether the associations were mediated by respiratory symptoms, and we assessed whether workers with a specific genetic make-up are more susceptible to sickness absence upon occupational exposure.
MATERIALS AND METHODS
Study population
In this study, we included adults from the Lifelines Cohort Study and biobank [16]. At the baseline visit, between 2006 and 2013, subjects had a physical examination and completed questionnaires on occupation, health, lifestyle, environment, and psychosocial parameters. A subset (n=13,302) of genetically unrelated Lifelines participants had genome-wide genotyping data. For this subset, we also estimated occupational exposures using a JEM. This study was approved by the Medical Ethical Commission (METC) of the University Medical Center Groningen (Reference number-2007/152). All subjects signed written informed consent. All methods were carried out in accordance with relevant guidelines and regulations for human subjects.
In the current analysis, out of 13 302 subjects, we included 10 087 ‘active workers’, defined as having a paid current job. Of those, 9883 (98%) active workers answered the questions on sickness absence.
Sickness absence
Sickness absence was self-reported (see ‘S1 Appendix: Supplementary Questions). ‘Any sickness absence’ was defined as being absent from work due to illness or problems (except pregnancy) at least one day in the last year (yes/no). ‘Long-term sickness absence’ was defined as being absent from work due to illness or problems (except pregnancy) for two consecutive weeks or more in the last year (yes/no).
Occupational exposures
Occupational exposures were estimated using the job titles as reported in the questionnaire. The self-reported jobs were coded according to the International Standard Classification of Occupations (ISCO-88) [17]. Subsequently, the ALOHA+ JEM (a modified version of the ad hoc JEM for COPD called the ALOHA JEM) [18] was used to classify occupational exposure into no, low, or high exposure categories (0/1/2) for the following occupational exposures: biological dust, mineral dust, gases/fumes, pesticides, solvents, and metals.
23
Respiratory symptoms
The presence of chronic cough, chronic phlegm, and dyspnea was self-reported (see S1 Table for the exact definition).
Candidate SNPs and genotyping
The selection of candidate SNPs was based on two genome-wide interaction studies conducted by our research group [13,14] that identified 11 SNPs in biologically plausible genes that significantly interacted with occupational exposures on lung function, i.e. rs17490056 with biological dust, rs13278529, rs473892, and rs6751439 with mineral dust, rs159497, rs516732, and rs2888674 with gases/fumes [13], and rs4764419, rs10459067, rs482555, and rs2145067 with pesticides [14]. Gene annotation, biological plausibility, and details on how genotyping was performed are described elsewhere [13,14].
Co-variates
Subjects’ age, sex, and body mass index (BMI) were determined during the baseline screening examination. Smoking status, monthly income, and education were taken from the baseline questionnaire. Smoking status was categorized into never, former, and current smoker. Monthly income was categorized into low, medium, high, and don’t know/don’t tell. Finally, education was categorized into low, medium, high, and unclassifiable. (see ‘S2 Table’).
Statistical methods
Chi-Square and Mann-Whitney U test were performed to investigate the univariate association of demographic characteristics, respiratory symptoms, and occupational exposures with sickness absence. To investigate the independent association between occupational exposures and sickness absence, multivariate logistic regression models with adjustment for potential confounders were used. No sickness absence was considered as reference group for both any and long-term sickness absence. Subjects with long-term sickness absence (≥2 weeks) were also included in the analyses on any sickness absence. A two-sided p-value < 0.05 was considered statistically significant.
To assess whether respiratory symptoms mediate the association between occupational exposures and sickness absence, we performed structural equation modeling adjusted for covariates (Fig 1) in MPlus software using the probit link-function [19]. We performed mediation analyses by respiratory symptoms for all models with a significant positive association between exposure (either high or low)
24
and sickness absence. Significant mediation by the respiratory symptom was considered present when the p-value of the indirect effect was < 0.05.
To assess whether the association between occupational exposures and sickness absence was dependent on genetic make-up, a multivariate logistic regression, including interactions between SNPs and occupational exposures, was used. SNPs were tested in a co-dominant model. Both SNP by low and SNP by high exposure interactions were assessed, and interaction was considered statistically significant at p-value < 0.05. The interaction models included dummy variables for low and high exposure, for the heterozygous (HZ) and homozygous for the minor allele (HM) genotypes and their interactions i.e.
Sickness absence = low exposure + high exposure + HZ + HM + low exposure*HZ + low exposure*HM + high exposure* HZ + high exposure*HM + covariates
SPSS 22 (IBM Corp. Released 2013. IBM SPSS Statistics for Windows, Version 22.0. Armonk, NY: IBM Corp) was used for the data analysis.
Figure 1. Mediation analyses pathway.
Association between occupational exposures and respiratory symptoms (a). Association between respiratory symptoms and sickness absence (b). Indirect effect is a product of ab. Total effect—association between occupational exposures and sickness absence adjusted for covariates (c). Direct effect—association between occupational exposures and sickness absence additionally adjusted for respiratory symptoms (c').
25
RESULTS
Baseline characteristics
In Fig 2, a flowchart of the subject selection is presented. In the final analyses, 204 workers were excluded because they lacked data on sickness absence. These excluded workers were slightly older, more often female, more often current smokers, and had a lower socioeconomic status compared to workers with data on sickness absence (S3 Table).
More than half of the 9,883 included subjects (53%) reported any sickness absence during the last year (Table 1). Subjects with any sickness absence were younger, more often female, had a higher BMI, were more often current smokers, had higher education but lower income, and had a higher prevalence of respiratory symptoms compared to subjects with no sickness absence. Of 5,235 subjects who reported any sickness absence, 1,230 (≈23%) subjects also reported long-term sickness absence. The subjects with long-term sickness absence were more often female, had a higher BMI, were more often current smokers, had lower education and lower monthly income, and had a higher prevalence of respiratory symptoms compared to subjects with no sickness absence. The distribution of the population characteristics according to the different levels of occupational exposures is given in Tables S4 and S5.
Occupational exposures and sickness absence
Table 1 shows that subjects reporting any sickness absence during the last year were somewhat more often exposed to solvents and metals, while they had a lower prevalence of high exposure to biological dust, mineral dust, and pesticides
26
Table 1: Comparison of demographic factors, respiratory symptoms, and occupational exposures among subjects with no sickness absence, any sickness absence, and long-term sickness absence (≥ 2 weeks). Subjects with long-term sickness absence were also included in the analyses on any sickness absence.
Demographic factors, respiratory symptoms,
and occupational exposures No SA, (N=4,648) Any SA, (5,235) Long-term SA, (1,230) Age (years), Median (min-max) 46 (18-76) 45 (20-77)* 47 (20-71)
Body Mass Index (BMI)
(kg/meter2), Median (min-max) 25 (17-53) 26 (16-52) 26 (17-51)ϴ
Sex Male, N (%) 2,152 (46.3) 2,269 (43.3)* 492 (40.0)ϴ Female, N (%) 2,496 (53.7) 2,966 (56.7) 738 (60.0) Smoking status Never smoker, N (%) 2,064 (44.8) 2,052 (39.5)* 432 (35.3)ϴ Former-smoker, N (%) 1,556 (33.8) 1,837 (35.3) 436 (35.6) Current smoker, N (%) 990 (21.4) 1,311 (25.2) 357 (29.1) Education Low, N (%) 767 (16.5) 782 (15.0)* 271 (22.1)ϴ Medium, N (%) 2,521 (54.3) 2,794 (53.4) 691 (56.3) High, N (%) 1,341 (28.9) 1,647 (31.5) 265 (21.6) Unclassifiable, N (%) 13 (0.3) 6 (0.1) 1 (0.1) Monthly Income Low income, N (%) 390 (8.4) 597 (11.4)* 178 (14.5)ϴ Medium income, N (%) 1,283 (27.7) 1,579 (30.3) 403 (32.9) High income, N (%) 2,197 (47.5) 2,478 (47.5) 479 (39.1)
Don’t know/Don’t tell, N (%) 757 (16.4) 565 (10.8) 164 (13.4)
Chronic cough No, N (%) 4,322 (93.9) 4,699 (90.9)* 1,091(89.6)ϴ Yes, N (%) 283 (6.1) 473 (9.1) 126 (10.4) Chronic phlegm No, N (%) 4,389 (95.0) 4,801 (92.3)* 1,110 (91.1)ϴ Yes, N (%) 230 (5.0) 400 (7.7) 109 (8.9) Dyspnea No, N (%) 3,502 (87.8) 3,525 (82.3)* 733 (76.9)ϴ Yes, N (%) 488 (12.2) 756 (17.7) 220 (23.1) Biological dust No exposure, N (%) 3,188 (68.6) 3,574 (68.3)* 782 (63.6)ϴ Low exposure, N (%) 1,233 (26.5) 1,498 (28.6) 397 (32.3) High exposure, N (%) 227 (4.9) 163 (3.1) 51 (4.1) Mineral dust No exposure, N (%) 3,624 (78.0) 4,198 (80.2)* 922 (75.0)ϴ Low exposure, N (%) 801 (17.2) 790 (15.1) 230 (18.7) High exposure, N (%) 223 (4.8) 247 (4.7) 78 (6.3) Gases/fumes No exposure, N (%) 2,672 (57.5) 3,061 (58.5) 613 (49.8)ϴ Low exposure, N (%) 1,696 (36.5) 1,853 (35.4) 516 (42.0) High exposure, N (%) 280 (6.0) 321 (6.1) 101 (8.2) Pesticides No exposure, N (%) 4,412 (94.9) 5,077 (97.0)* 1,180 (95.9)
27 Low exposure, N (%) 184 (4.0) 124 (2.4) 37 (3.0) High exposure, N (%) 52 (1.1) 34 (0.6) 13 (1.1) Solvents No exposure, N (%) 3,519 (75.7) 3,819 (73.0)* 871 (70.8)ϴ Low exposure, N (%) 972 (20.9) 1,232 (23.5) 318 (25.9) High exposure, N (%) 157 (3.4) 184 (3.5) 41 (3.3) Metals No exposure, N (%) 4,335 (93.3) 4,831 (92.3)* 1,114(90.6)ϴ Low exposure, N (%) 228 (4.9) 263 (5.0) 81 (6.6) High exposure, N (%) 85 (1.8) 141 (2.7) 35 (2.8)
*Statistically significant at two-sided P-value <0.05 between no sickness absence and any sickness absence.
ϴStatistically significant at two-sided P-value <0.05 between no sickness absence and long-term
sickness absence. SA= Sickness absence.
compared to subjects reporting no sickness absence. Subjects reporting long-term sickness absence were (considerably) more often exposed to biological dust, mineral dust, gases/fumes, solvents, and metals compared to subjects reporting no sickness absence. The correlation among different occupational exposures is given in S1 Fig.
After adjustment for covariates, subjects with high exposure to biological dust and low exposure to mineral dust or pesticides had a lower prevalence of any sickness absence compared to subjects without these exposures (Fig 3). No significant associations were found between these exposures and long-term sickness absence. Subjects with high exposure to gases/fumes reported long-term sickness absence more often, but not any sickness absence, compared to subjects not exposed to gases/fumes. Low exposure to solvents was associated with a higher prevalence of both any and long-term sickness absence. High exposure to metals was associated with a higher prevalence of any sickness absence whereas both exposure to low and high metals were associated with a higher prevalence of
long-term sickness absence in a dose-dependent way (see S6 Table).
The significant association between high exposure to gases/fumes and long-term sickness absence was mediated by chronic cough and chronic phlegm (S7 Table). The association between high solvents exposure and any sickness absence was mediated by chronic phlegm, however, the association between high solvents exposure and any sickness absence was not significant. The associations between low and high metals exposure and sickness absence were not mediated by respiratory symptoms.
28
Mediation analyses
Gene-by-exposure interactions on sickness absence
Out of the 11 candidate SNPs, three SNPs had a significant interaction with occupational exposures on sickness absence (see Fig 4 and Tables S10-S13). Two of the SNPs (rs473892 and rs159497) had a positive interaction with exposure to mineral dust and gases/fumes, respectively, on sickness absence. This finding implies that subjects carrying one (for rs159497) or two (for rs473892) minor alleles reported a higher prevalence of long-term or any sickness absence upon the specific occupational exposure compared to subjects carrying two major alleles (Figs 4B and 4C). One SNP (rs2888674) negatively interacted with exposure to gases/fumes on both any and long-term sickness absence (Fig 4A).
DISCUSSION
In this large cross-sectional study, we investigated the association between occupational exposure to biological dust, mineral dust, gases/fumes, pesticides, solvents, or metals and sickness absence. We investigated whether the associations were mediated by respiratory symptoms. In addition, we explored
Figure 3. Associations between occupational exposures and sickness absence.
Sickness absence presented as (A) Any sickness absence—subjects with long-term sickness absence were also included in the analyses on any sickness absence. (B) long-term sickness absence (≥ 2 weeks). The multivariate logistic regression model was adjusted for age, sex, BMI, education, smoking status, and monthly income.
29
whether subjects with a specific genotype were more susceptible to the effects of occupational exposure on sickness absence.
The results showed that subjects with high exposure to biological dust or low exposure to mineral dust and especially to pesticides significantly less often reported any sickness absence. Subjects with high exposure to gases/fumes reported long-term sickness absence significantly more often, and subjects with low exposure to solvents and high exposure to metals reported a significantly higher prevalence of both any and long-term sickness absence. For exposure to metals, long-term sickness absence increased with the intensity of exposure. The results further showed that chronic cough and chronic phlegm significantly mediate the association between high exposure to gases/fumes and long-term sickness absence.
Figure 4. Associations between SNPs and sickness absence in subjects with no, low, and high gases and fumes and mineral dust exposure.
In both any and long-term sickness absence, no exposure was considered as reference group. SNPs presented are (A) rs2888674, (B) rs473892, and (C) rs159497. Interactions analysis adjusted for age, sex, BMI, education, smoking status, and monthly income.
30
The main route of exposure to dust and fumes is through inhalation, and this specifically affects the respiratory system. Dust and fumes exposure is a strong predictor of respiratory symptoms [3]. Also our previous study found that high exposure to dust and gases/fumes was associated with airway obstruction [4]. In addition, another study showed that any exposure to dust and fumes is strongly associated with sickness absence in subjects with respiratory symptoms [10]. Our results showed indeed positive associations between occupational exposure to gases/fumes and sickness absence, especially with long-term sickness absence and this association (partly) runs via respiratory symptoms. However, for both biological dust (high exposure) and mineral dust (low exposure), we found a lower prevalence of any sickness absence in the exposed subjects. This lower prevalence of sickness absence is also seen in subjects with low pesticide exposure. A recent meta-analysis showed negative effects of exposure to biological and mineral dust on lung function level [20] and a recent review showed that pesticides exposure is associated with a higher incidence of chronic diseases [21]. One explanation for our unexpected findings could be that a significant proportion of subjects that were exposed to pesticides, high biological dust, and low mineral dust were self-employed workers (among those with pesticide exposure ~ 50%, high biological dust exposure ~ 40% and low mineral dust exposure ~ 10%). Farmers had a very low prevalence of any sickness absence, i.e., 22% (sickness absence prevalence in the total study sample was 53%), which may be the result of being employed. Previous studies showed that self-employed workers tend to have a lower prevalence of sickness absence compared to employed workers [22,23]. Possible reasons for this may be lack of compensation, high time demands, or difficulties in finding a replacer [23,24].
An alternative explanation for our unexpected findings could be the ‘healthy worker effect’ [25,26]. This implies that people with respiratory diseases or sensitive to low exposure to biological dust, mineral dust, or pesticides did not take up a job with these types of exposure, or switched to a job with less occupational exposure. As a result, only the workers who did not experience negative health effects from these exposures stayed in their job.
Occupational exposure to solvents or metals was associated with a higher prevalence of sickness absence in an exposure intensity depending way. Previous studies showed that occupational exposure to solvents and metals was associated with a broad spectrum of diseases, such as pulmonary diseases, brain diseases, and kidney diseases [27,28]. Given these broad ranges of health consequences of exposure to solvents and metals, it is not surprising that we found a higher prevalence of sickness absence in exposed subjects.
31
In this study, we found that chronic cough and chronic phlegm mediate the association between high gases/fumes exposure and long-term sickness absence. This means that high gases/fumes exposure is a risk factor for chronic cough and chronic phlegm, which in turn lead to sickness absence, especially long-term sickness absence.
However, the mediation effects of these symptoms only partly explain the association between exposure and sickness absence. This indicates that other factors could mediate the association between occupational exposures and sickness absence. Indeed, studies found that chronic diseases and mental disorders are strong predictors of sickness absence [2,29]. Future studies should consider these factors as potential mediators in the association between occupational exposures and sickness absence.
In the current study, we investigated whether our previously identified SNPs modify the association between occupational exposures and sickness absence. Earlier, we observed effect modification by SNPs on the association between occupational exposures and lung function [13,14] suggesting that these genetic variants make subjects more susceptible to the health effects of occupational exposures. Therefore, we expected effect modification by these identified biologically plausible SNPs on the association between occupational exposure and sickness absence. Indeed, we found several SNPs that interacted with mineral dust and gases/fumes exposure on sickness absence.
Subjects homozygous for the minor allele of rs2888674 had a lower prevalence of sickness absence upon gases/fumes exposure compared to subjects homozygous for the major allele. The minor allele of rs2888674 may thus be protective against the effects of gases/fumes exposure. In our previous study, we observed a
protective effect of the rs2888674 minor allele on FEV1 level upon gases/fumes
exposure [13]. The minor allele of rs2888674 is associated with a higher TMEM176A expression compared to the major allele [13]. A higher expression of TMEM176A attenuates co-stimulatory molecules expression and thereby, weakens inflammatory response [30]. Thus, we hypothesize that the protective effect against occupational exposure of the minor allele of rs2888674 (i.e., less sickness absence and less affected lung function level) may be the result of this lower inflammatory response to environmental triggers.
Subjects who were exposed to mineral dust and homozygous for the minor allele of rs473892 reported a higher prevalence of any sickness absence compared to exposed subjects who were homozygous for the major allele. In our previous study, rs473892 showed the same protective effect against exposure as the TMEM176A SNP described in the previous paragraph [13]. This implies that the result of the
32
current study on sickness absence is contradicting our previous results on lung function. SNP rs473892 is located near the oligodendrocyte transcription factor 3 (OLIG3) gene, and the biological function of OLIG3 is largely unknown, more research is required to explain this finding.
Subjects heterozygous for rs159497 had a higher prevalence of long-term sickness absence upon gases/fumes exposure compared to exposed subjects homozygous for the major allele. Rs159497 is located near the phosphodiesterase-4D (PDE4D) gene, and the minor allele is associated with a higher PDE4D-expression [13]. The PDE4D-enzyme has a degrading and inactivating role on cyclic adenosine monophosphate (cAMP) [31]. cAMP attenuates immune and inflammatory responses and leads to airway smooth muscle relaxation and bronchodilation [32]. Therefore, a higher PDE4D-level may increase inflammation, and subsequently make the subject vulnerable to the harmful effects of environmental substances. Our observation that minor allele carriers exposed to gases/fumes had a higher prevalence of sickness absence is in line with this.
We did not find any significant gene-by-biological dust or gene-by-pesticides interactions on sickness absence.
Strengths and limitations
To our knowledge, this is the first study that investigated the association between several (airborne) occupational exposures (i.e., biological dust, mineral dust, gases/fumes, pesticides, solvents, and metals) and sickness absence in the general working population. We used information from almost 10,000 extensively characterized Lifelines subjects. In addition, we investigated whether subjects with a specific genetic make-up are more susceptible to sickness absence upon occupational exposures. In developed countries such as The Netherlands, strict occupational safety and health guidelines have been developed to protect workers. Despite this, we still found a strong association between airborne occupational exposure and sickness absence. Hence, it could be questioned whether, in practice, workers fully comply with the provided preventive measures.
The JEM is an easy-to-use tool for assessing occupational exposure with several advantages. The JEM converts coded occupational titles into estimated exposures, which is advantageous in many instances when it is difficult to measure exposure at the individual level [33]. In the self-reported approach, workers often struggle to estimate exposure level when an agent is not seen or smelled [34], and difficulty in recalling the correct exposure duration influences the validity and reliability of the report [35]. A JEM estimates occupational exposure independent of workers' perception of exposure, and thus eliminates the chance of differential
33
misclassification or recall bias [36]. However, a JEM may result in non-differential misclassification bias [36] and thereby dilutes the effect estimates towards null or insignificant values [37]. A disadvantage is that a JEM does not assess exposure at the individual chemical or biological agent level. Furthermore, this study is cross-sectional in design, so it does not infer any causality, nor does it take lifetime cumulative exposure into account. Finally, we adjusted for well-known covariates (also covariates that are available in the LifeLines Cohort Study) to overcome confounding effects. We did not adjust for other potential confounders such as stress, physical workload, or type of employment contract. So we cannot rule out the effect of these unmeasured confounders in our analysis. We also could not adjust for co-exposure since several occupational exposures, e.g. biological dust, mineral dust, and gases/fumes are highly correlated which will lead to multicollinearity in the model and thus biased estimates.
CONCLUSIONS
In conclusion, high exposure to gases/fumes, low exposure to solvents, and metals exposure are associated with a higher prevalence of sickness absence and especially with long-term sickness absence. Chronic cough and chronic phlegm mediate the association between high gases/fumes exposure and long-term sickness absence. Although many preventive measures are applied to control occupational exposure levels, still an association with sickness absence exists. Studying gene-by-occupational exposure interactions may help to understand underlying cellular and molecular pathways. Future research should focus on the causal association between the identified genes and health effects. A thorough understanding of the gene-by-exposure effect on health will enable us to identify susceptible subjects and set health-based and personalized recommended exposure limits for all exposed workers.
34
References
1. Valirad, F. et al. Interaction of Physical Exposures and Occupational Factors on Sickness Absence in Automotive Industry Workers. Glob. J. Health Sci. 7, 276, (2015).
2. Duijts, S.F., Kant, I., Swaen, G.M., van den Brandt, P.A. & Zeegers, M.P. A meta-analysis of
observational studies identifies predictors of sickness absence. J. Clin. Epidemiol. 60,1105-1115, (2007).
3. Eagan, T.M., Gulsvik, A., Eide, G.E. & Bakke, P.S. Occupational airborne exposure and the
incidence of respiratory symptoms and asthma. Am. J. Respir. Crit. Care Med. 166, 933-938, (2002).
4. De Jong, K. et al. Pesticides and other occupational exposures are associated with airway
obstruction: the LifeLines cohort study. Occup. Environ. Med. 71, 88-96, (2014).
5. Anderson, S.E. & Meade, B.J. Potential health effects associated with dermal exposure to
occupational chemicals. Environ. Health Insights 8, 51-62, (2014).
6. Sigsgaard, T., Bonefeld-Jørgensen, E.C., Hoffmann, H.J., Bønløkke, J. & Krüger, T. Microbial
cell wall agents as an occupational hazard. Toxicol. Appl. Pharmacol. 207, 310-319, (2005).
7. Ulvestad, B. et al. Gas and dust exposure in underground construction is associated with signs
of airway inflammation. Eur. Respir. J. 17, 416–421, (2001).
8. Prince, P., Lemière, C., Dufour, M.H., Chaboillez, S. & Boulet, L.P. Airway Inflammatory
Responses Following Exposure to Occupational Agents. Chest 141, 1522–1527, (2012).
9. Heldal, K.K., Halstensen, A.S., Thorn, J., Eduard, W. & Halstensen, T.S. Airway inflammation
in waste handlers exposed to bioaerosols assessed by induced sputum. Eur. Respir. J. 21, 641–645, (2003).
10. Kim, J.L. et al. Predictors of respiratory sickness absence: An international population-based
study. Am. J. Ind. Med. 56, 541–549, (2013).
11. Hur, G.Y. et al. Association of β2-adrenergic receptor polymorphism with work-related
symptoms in workers exposed to wheat flour. Yonsei. Med. J. 52, 488–494, (2011).
12. Cho, H.J. et al. Effect of Toll-like receptor 4 gene polymorphisms on work-related respiratory
symptoms and sensitization to wheat flour in bakery workers. Ann. Allergy Asthma Immunol. 107, 57–64, (2011).
13. De Jong, K. et al. Genome-wide interaction study of gene-by-occupational exposure and
effects on FEV1 levels. J. Allergy Clin. Immunol. 136, 1664-1672, (2015).
14. De Jong, K. et al. NOS1: A susceptibility gene for reduced level of FEV1 in the setting of
pesticide exposure. Am. J. Respir. Crit. Care. Med. 190, 1188-1190, (2014).
15. Matheson, M.C. et al. Biological dust exposure in the workplace is a risk factor for chronic
obstructive pulmonary disease. Thorax 60, 645-651, (2005).
16. Scholtens, S. et al. Cohort Profile: LifeLines, a three-generation cohort study and biobank. Int.
J. Epidemiol. 44, 1172-1180, (2015).
35 International Labour Organization. ISCO-88 1990.
18. Sunyer, J. et al. Pulmonary ventilatory defects and occupational exposures in a
population-based study in Spain. Am. J. Respir. Crit. Care. Med. 157, 512-517, (1998).
19. Muthén, L.K. & Muthén, B.O. Mplus version 7.4 software. Los Angeles, CA: Muthén & Muthén 1998.
20. Alif, S.M. et al. Occupational exposure and risk of chronic obstructive pulmonary disease: a
systematic review and meta-analysis. Expert. Rev. Respir. Med. 10, 861-872, (2016).
21. Mostafalou, S. & Abdollahi, M. Pesticides: an update of human exposure and toxicity. Arch.
Toxicol. 91, 549–599, (2017).
22. Rosta, J., Tellnes, G. & Aasland, O.G.. Differences in sickness absence between
self-employed and self-employed doctors: a cross-sectional study on national sample of Norwegian doctors in 2010. BMC Health Serv. Res. 14, 199, (2014).
23. Lechmann, D.S. & Schnabel, C. Absence from Work of the Self‐Employed: A Comparison with
Paid Employees. Kyklos 67, 368-390, (2014).
24. Nordenmark, M., Hagqvist, E. & Vinberg, S. Sickness Presenteeism among the Self-employed and Employed in Northwestern Europe—The Importance of Time Demands. Saf. Health Work 10, 224-228, (2019).
25. Chénard, L. et al. Lung function and farm size predict healthy worker effect in swine farmers.
Chest 131, 245-254 (2007).
26. Elholm, G. et al. Become a farmer and avoid new allergic sensitization: adult farming
exposures protect against new-onset atopic sensitization. J. Allergy Clin. Immunol. 132, 1239, (2013).
27. Niaz, K., Bahadar, H., Maqbool, F. & Abdollahi, M. A review of environmental and occupational
exposure to xylene and its health concerns. EXCLI J. 14, 1167, (2015).
28. Jaishankar, M., Tseten, T., Anbalagan, N., Mathew, B.B. & Beeregowda, K.N. Toxicity,
mechanism and health effects of some heavy metals. Interdiscip. Toxicol. 7, 60-72, (2014).
29. Casimirri, E. et al. Chronic diseases are strongly associated with sickness absences in a
sample of Italian public employees. Int. J. Occup. Med. Environ. Health 27, 343-354, (2014).
30. Condamine, T. et al. Tmem176B and Tmem176A are associated with the immature state of
dendritic cells. J. Leukoc. Biol. 88, 507-515, (2010).
31. Yoon, H.K. et al. Polymorphisms in PDE4D are associated with a risk of COPD in
non-emphysematous Koreans. COPD-J. Chronic Obstr. Pulm. Dis. 11, 652-658, (2014).
32. Billington, C.K., Ojo, O.O., Penn, R.B. & Ito, S. cAMP regulation of airway smooth muscle
function. Pulm. Pharmacol. Ther. 26, 112-120, (2013).
33. Fischer, H.J. et al. Developing a job-exposure matrix with exposure uncertainty from expert elicitation and data modeling. J. Expo. Sci. Environ. Epidemiol. 27, 7, (2017).
34. Sadhra, S.S. et al. Development of an occupational airborne chemical exposure matrix. Occup.
36
35. Teschke, K. et al. Occupational exposure assessment in case–control studies: opportunities for
improvement. Occup. Environ. Med. 59, 575-594, (2002).
36. Kauppinen, T.P., Mutanen, P.O. & Seitsamo, J.T. Magnitude of misclassification bias when
using a job-exposure matrix. Scand. J. Work Environ. Health 18,105-112, (1992).
37
Chapter 2
38
S1 Appendix: Supplementary Questions
1. In the last year. how many days/weeks did you not work due to illness or problems (excluding pregnancy)? Answer: number of days or weeks. 2. In the last year. did you not go to work for one or more periods of at least
two consecutive weeks because of illness or problems (excluding pregnancy)? Answer categories: Yes/No.
39
S1 Table: Definition of the respiratory symptoms used in the main analyses.
S2 Table: Co-variates adjusted in the regression analyses. Respiratory
symptoms Definition
Chronic cough Usual coughing first thing in the morning, or during the day, or at night in winter for at least three months in a year. Chronic phlegm Usual bringing up any phlegm first thing in the morning, or during the day, or at night in winter for at least three months in a year. Dyspnea Having shortness of breath when hurrying on level ground, or walking up a slight hill, or stairs at a normal pace.
Covariates Operational definition
Age Subjects’ age in years
Body Mass Index
(BMI) Subjects’ BMI calculated as weight in kilogram/ (height in meter)2 Sex Gender of the study subjects
Male
Female
Smoking status Smoking status of the subjects
Never smoker Never smoked or smoked for < 1 year
Former-smoker Smoked for ≥ 1 year and stopped smoking for ≥ 1 month
Current smoker Current smoker or stopped smoking < 1 month
Education Highest level of completed education
Low No training. primary education. or lower or pre-vocational education
Medium General secondary education. secondary vocational or professional guiding. or pre-university education
High higher professional or university degree
Unclassifiable Subjects with other than above-mentioned education.
Monthly income Subjects monthly income in euros
Low income Monthly income ≤ €1500
Medium income Monthly income between €1500 and €2500
High income Monthly income ≥ €2500
40
S3 Table: Comparison of demographic characteristics between workers with and without data on sickness absence.
Demographic factors Workers with no data on sickness absence. 204 (2%)
Workers with data on sickness absence.
9883 (98%) p-value Age (years) Median (min-max) 48 (22-73) 46 (18-77) <0.001 Body Mass Index (BMI)
(kg/meter2) Median (min-max) 26 (17-45) 26 (16-53) 0.277
Sex Female. N (%) 138 ( 67.6) 5462 ( 55.3) <0.001 Smoking status Never smoker, N (%) 72 (35.6) 4116 (42.0) 0.030 Former-smoker, N (%) 67 (33.2) 3393 (34.6) Current smoker, N (%) 63 (31.2) 2301 (23.5) Education Low, N (%) 66 (33.0) 1549 (15.7) <0.001 Medium, N (%) 105 (52.5) 5315 (53.8) High, N (%) 29 (14.5) 2988 (30.3) Unclassifiable, N (%) 0 (0.0) 19 (0.2) Monthly Income Low income, N (%) 38 (19.3) 987 (10.0) <0.001 Medium income, N (%) 63 (32.0) 2862 (29.1) High income, N (%) 45 (22.8) 4675 (47.5)
41 S4 T ab le : D is tr ib ut io n of th e po pu la tio n ch ar ac te ris tic s ac co rd in g to th e di ffe re nt le ve ls o f o cc up at io na l e xp os ur es (b io lo gi ca l d us t, m in er al d us t, an d ga se s/ fu m es ). Po pu la tio n ch ar ac te ris tic s B io lo gi ca l d us t M in er al d us t G as es /fu m es N o Lo w H ig h N o Lo w H ig h N o Lo w H ig h Ag e, m ed ia n (m in -m ax ) 46 (1 8-77 ) 46 (2 0-75 ) 45 (2 2-71 ) 46 (1 8-77 ) 46 (2 1-75 ) 45 (2 0-73 ) 46 (1 8-77 ) 46 (2 0-75 ) 45 (2 3-70 ) B M I, m ed ia n (m in -m ax ) 26 (1 6-53 ) 26 (1 7-52 ) 26 (1 8-49 ) 25 (1 7-53 ) 26 (1 6-52 ) 27 (1 7-48 ) 25 (1 7-53 ) 26 (1 7-52 ) 26 (1 6-47 ) Se x M al e, N (% ) 34 47 (5 0. 0) 71 5 (2 5. 6) 32 5 (8 0. 6) 31 14 (3 9. 1) 93 3 (5 7. 1) 44 0 (9 1. 1) 24 91 (4 2. 7) 14 41 (3 9. 6) 55 5 (9 0. 7) Fe m al e, N (% ) 34 43 (5 0. 0) 20 79 (7 4. 4) 78 (1 9. 4) 48 56 (6 0. 9) 70 1 (4 2. 9) 43 (8 .9 ) 33 47 (5 7. 3) 21 96 (6 0. 4) 57 (9 .3 ) Sm ok in g st at us N ev er s m ok er , N (% ) 29 12 (4 2. 5) 10 85 (3 9. 2) 19 1 (4 8. 0) 33 96 (4 2. 9) 61 1 (3 7. 6) 18 1 (3 7. 8) 25 56 (4 4. 1) 14 21 (3 9. 5) 21 1 (3 4. 6) Fo rm er -s m ok er , N (% ) 23 43 (3 4. 2) 10 08 (3 6. 4) 10 9 (2 7. 4) 27 80 (3 5. 2) 53 3 (3 2. 8) 14 7 (3 0. 7) 20 14 (3 4. 7) 12 45 (3 4. 6) 20 1 (3 3. 0) C ur re nt s m ok er , N (% ) 15 90 (2 3. 2) 67 6 (2 4. 4) 98 (2 4. 6) 17 32 (2 1. 9) 48 1 (2 9. 6) 15 1 (3 1. 5) 12 31 (2 1. 2) 93 6 (2 6. 0) 19 7 (3 2. 3) Ed uc at io n Lo w , N (% ) 89 5 (1 3. 0) 59 1 (2 1. 2) 12 9 (3 2. 2) 76 7 (9 .6 ) 63 3 (3 8. 9) 21 5 (4 4. 7) 46 9 (8 .0 ) 87 5 (2 4. 1) 27 1 (4 4. 4) M ed iu m , N (% ) 35 28 (5 1. 3) 16 66 (5 9. 8) 22 6 (5 6. 4) 43 38 (5 4. 5) 84 9 (5 2. 1) 23 3 (4 8. 4) 28 99 (4 9. 7) 22 12 (6 1. 0) 30 9 (5 0. 6) H ig h, N (% ) 24 44 (3 5. 5) 52 7 (1 8. 9) 46 (1 1. 5) 28 39 (3 5. 7) 14 5 (8 .9 ) 33 (6 .9 ) 24 49 (4 2. 0) 53 7 (1 4. 8) 31 (5 .1 ) U nc la ss ifi ab le , N (% ) 16 (. 2) 3 (.1 ) 17 (. 2) 2 (.1 ) 16 (. 3) 3 (.1 ) M on th ly In co m e Lo w in co m e, N (% ) 53 5 (7 .8 ) 42 2 (1 5. 2) 68 (1 6. 9) 70 7 (8 .9 ) 23 2 (1 4. 3) 86 (1 7. 9) 40 9 (7 .0 ) 54 5 (1 5. 1) 71 (1 1. 7) M ed iu m in co m e, N (% ) 19 00 (2 7. 7) 90 9 (3 2. 7) 11 6 (2 8. 8) 21 22 (2 6. 7) 62 8 (3 8. 6) 17 5 (3 6. 4) 14 57 (2 5. 1) 12 09 (3 3. 4) 25 9 (4 2. 5) H ig h in co m e, N (% ) 36 00 (5 2. 5) 10 31 (3 7. 1) 89 (2 2. 1) 41 73 (5 2. 6) 43 8 (2 6. 9) 10 9 (2 2. 7) 33 13 (5 7. 0) 12 41 (3 4. 3) 16 6 (2 7. 3) D on ’t kn ow /D on ’t te ll, N (% ) 82 5 (1 2. 0) 41 8 (1 5. 0) 13 0 (3 2. 3) 93 4 (1 1. 8) 32 8 (2 0. 2) 11 1 (2 3. 1) 63 6 (1 0. 9) 62 4 (1 7. 2) 11 3 (1 8. 6)
42 S5 T ab le : D is tr ib ut io n of th e po pu la tio n ch ar ac te ris tic s ac co rd in g to th e di ffe re nt le ve ls o f o cc up at io na l e xp os ur es (p es tic id es , s ol ve nt s, a nd m et al s) . Po pu la tio n ch ar ac te ris tic s Pe st ic id es So lv en ts M et al s N o Lo w H ig h N o Lo w H ig h N o Lo w H ig h Ag e, m ed ia n (m in -m ax ) 46 (1 8-77 ) 46 (2 2-73 ) 45 (2 3-70 ) 46 (1 8-77 ) 46 (2 0-73 ) 44 (2 0-73 ) 46 (1 8-77 ) 44 (2 2-66 ) 45 (2 0-70 ) B M I, m ed ia n (m in -m ax ) 26 (1 6-53 ) 26 (1 9-44 ) 26 (2 0-35 ) 26 (1 7-53 ) 26 (1 6-52 ) 27 (1 7-51 ) 25 (1 7-53 ) 26 (1 8-48 ) 26 (1 6-37 ) Se x M al e, N (% ) 41 72 (4 3. 1) 23 7 (7 4. 8) 78 (8 6. 7) 34 27 (4 5. 7) 83 7 (3 7. 2) 22 3 (6 4. 5) 38 07 (4 0. 7) 45 7 (9 0. 5) 22 3 (9 7. 8) Fe m al e, N (% ) 55 08 (5 6. 9) 80 (2 5. 2) 12 (1 3. 3) 40 65 (5 4. 3) 14 12 (6 2. 8) 12 3 (3 5. 5) 55 47 (5 9. 3) 48 (9 .5 ) 5 (2 .2 ) Sm ok in g st at us N ev er s m ok er , N (% ) 39 86 (4 1. 5) 16 0 (5 1. 0) 42 (4 6. 7) 31 28 (4 2. 0) 94 1 (4 1. 8) 11 9 (3 4. 5) 39 11 (4 2. 1) 19 4 (3 8. 6) 83 (3 6. 6) Fo rm er -s m ok er , N (% ) 33 52 (3 4. 9) 83 (2 6. 4) 25 (2 7. 8) 25 48 (3 4. 3) 78 3 (3 4. 8) 12 9 (3 7. 4) 32 33 (3 4. 8) 14 9 (2 9. 7) 78 (3 4. 4) C ur re nt s m ok er , N (% ) 22 70 (2 3. 6) 71 (2 2. 6) 23 (2 5. 6) 17 63 (2 3. 7) 50 4 (2 2. 4) 97 (2 8. 1) 21 39 (2 3. 0) 15 9 (3 1. 7) 66 (2 9. 1) Ed uc at io n Lo w , N (% ) 14 98 (1 5. 5) 92 (2 9. 1) 25 (2 8. 1) 11 98 (1 6. 0) 30 3 (1 3. 5) 11 4 (3 3. 0) 13 49 (1 4. 4) 15 9 (3 1. 5) 10 7( 47 .1 ) M ed iu m , N (% ) 51 88 (5 3. 7) 17 5 (5 5. 4) 57 (6 4. 0) 38 15 (5 1. 0) 13 83 (6 1. 6) 22 2 (6 4. 3) 50 43 (5 4. 0) 26 3 (5 2. 1) 11 4 (5 0. 2) H ig h, N (% ) 29 61 (3 0. 6) 49 (1 5. 5) 7 (7 .9 ) 24 51 (3 2. 8) 55 7 (2 4. 8) 9 (2 .6 ) 29 28 (3 1. 4) 83 (1 6. 4) 6 (2 .6 ) U nc la ss ifi ab le , N (% ) 19 (. 2) 18 (. 2) 1 (.0 ) 19 (. 2) M on th ly In co m e Lo w in co m e, N (% ) 95 1 (9 .9 ) 58 (1 8. 3) 16 (1 8. 0) 70 6 (9 .5 ) 27 6 (1 2. 3) 43 (1 2. 5) 94 5 (1 0. 1) 54 (1 0. 7) 26 (1 1. 5) M ed iu m in co m e, N (% ) 28 20 (2 9. 3) 80 (2 5. 2) 25 (2 8. 1) 21 34 (2 8. 6) 66 2 (2 9. 5) 12 9 (3 7. 4) 26 45 (2 8. 4) 17 8 (3 5. 3) 10 2 (4 4. 9) H ig h in co m e, N (% ) 46 32 (4 8. 1) 70 (2 2. 1) 18 (2 0. 2) 36 31 (4 8. 7) 99 6 (4 4. 4) 93 (2 7. 0) 44 77 (4 8. 1) 18 6 (3 6. 9) 57 (2 5. 1) D on ’t kn ow /D on ’t te ll, N (% ) 12 34 (1 2. 8) 10 9 (3 4. 4) 30 (3 3. 7) 98 5 (1 3. 2) 30 8 (1 3. 7) 80 (2 3. 2) 12 45 (1 3. 4) 86 (1 7. 1) 42 (1 8. 5)
43
S6 Table: Associations between occupational exposures and sickness absence. Occupational
exposures
Any SAa Long-term SAb
OR (95% CI) P-value OR (95% CI) P-value Biological dust No exposure Low exposure 1.07 (0.97-1.18) 0.165 1.11 (0.95-1.28) 0.186 High exposure 0.72 (0.58-0.89) 0.003 0.90 (0.65-1.25) 0.533 Mineral dust No exposure Low exposure 0.88 (0.78-0.99) 0.036 0.98 (0.82-1.17) 0.814 High exposure 1.04 (0.85-1.28) 0.702 1.31 (0.98-1.76) 0.071 Gasfumes No exposure Low exposure 0.96 (0.89-1.05) 0.393 1.13 (0.98-1.30) 0.096 High exposure 1.07(0.89-1.28) 0.483 1.46 (1.11-1.91) 0.006 Pesticides No exposure Low exposure 0.70 (0.55-0.89) 0.004 0.82 (0.57-1.19) 0.305 High exposure 0.72 (0.46-1.13) 0.157 1.00 (0.53-1.89) 0.991 Solvents No exposure Low exposure 1.14 (1.03-1.26) 0.009 1.26 (1.08-1.46) 0.003 High exposure 1.13 (0.90-1.41) 0.296 1.02 (0.71-1.46) 0.924 Metals No exposure Low exposure 1.11 (0.92-1.35) 0.271 1.44 (1.09-1.91) 0.011 High exposure 1.68 (1.26-2.24) 0.000 1.75 (1.15-2.67) 0.009
No exposure as reference group.
a SA=Sickness absence; Association between occupational exposures and any sickness absence
adjusted for age. sex. BMI. education. smoking status. and monthly income (n=9756).
b Association between occupational exposures and long-term sickness absence adjusted for age. sex.
44 S7 T ab le : M ed ia tio n an al ys is o f r es pi ra to ry s ym pt om s in th e as so ci at io n be tw ee n oc cu pa tio na l e xp os ur es a nd s ic kn es s ab se nc e. O cc up at io na l e xp os ur es An y si ck ne ss a bs en ce Lo ng -te rm s ic kn es s ab se nc e To ta l e ffe ct D ire ct e ffe ct In di re ct e ffe ct To ta l e ffe ct D ire ct e ffe ct To ta l i nd ire ct e ffe ct O R (9 5% ) p va lu e O R (9 5% ) p va lu e O R (9 5% ) p va lu e O R (9 5% ) p va lu e O R (9 5% ) p va lu e O R (9 5% ) p va lu e C hr on ic c ou gh G as es /fu m es Lo w e xp os ur e 1. 15 (0 .9 5-1. 35 ) 0. 12 9 1. 15 (0 .9 5-1. 35 ) 0. 13 9 1. 00 (1 .0 0-1. 01 ) 0. 23 5 H ig h ex po su re 1. 56 (1 .0 5-2. 07 ) 0. 00 0 1. 54 (1 .0 4-2. 04 ) 0. 00 0 1. 01 (1 .0 0-1. 03 ) 0. 00 0 So lv en ts Lo w e xp os ur e 1. 14 (1 .0 3-1. 26 ) 0. 00 8 1. 14 (1 .0 3-1. 25 ) 0. 00 9 1. 00 (1 .0 0-1. 01 ) 0. 38 9 1. 28 (1 .0 6-1. 51 ) 0. 02 7 1. 28 (1 .0 5-1. 51 ) 0. 02 9 1. 00 (1 .0 0-1. 01 ) 0. 40 7 H ig h ex po su re 1. 12 (. 87 -1 .3 8) 0. 31 1 1. 12 (. 86 -1 .3 7) 0. 32 8 1. 00 (. 99 -1 .0 2) 0. 50 4 1. 00 (0 .5 8-1. 43 ) 0. 98 6 1. 00 (0 .5 7-1. 42 ) 0. 99 6 1. 00 (0 .9 9-1. 02 ) 0. 51 4 M et al s Low e xp os ur e 1. 11 (0 .9 0-1. 33 ) 0. 28 2 1. 11 (0 .8 9-1. 32 ) 0. 28 9 1. 0 (0 .9 9-1. 01 ) 0. 75 8 1. 52 (1 .0 1-2. 03 ) 0. 04 8 1. 52 (1 .0 1-2. 03 ) 0. 04 9 1. 00 (0 .9 9-1. 01 ) 0. 75 4 H ig h ex po su re 1. 72 (1 .2 1-2. 22 ) 0. 00 0 1. 73 ( 1. 22 -2 .2 4) 0. 00 0 0. 99 (0 .9 81 .0 1) 0. 38 4 1. 92 (0 .9 7-2. 88 ) 0. 05 2 1. 94 (0 .9 8-2. 90 ) 0. 04 9 0. 99 (0 .9 8-1. 01 ) 0. 40 0 C hr on ic p hl eg m G as es /fu m es Lo w e xp os ur e 1. 15 (0 .9 5-1. 35 ) 0. 12 9 1. 15 (0 .9 5-1. 35 ) 0. 12 9 1. 00 (1 .0 0-1. 01 ) 0. 35 1 H ig h ex po su re 1. 56 (1 .0 5-2. 07 ) 0. 00 0 1. 54 (1 .0 4-2. 05 ) 0. 03 1 1. 03 (1 .0 1-1. 05 ) 0. 02 1 So lv en ts Lo w e xp os ur e 1. 14 (1 .0 3-1. 26 ) 0. 00 8 1. 14 (1 .0 3-1. 26 ) 0. 00 9 1. 00 (0 .9 9-1. 01 ) 0. 80 3 1. 28 (1 .0 6-1. 51 ) 0. 02 7 1. 29 (1 .0 6-1. 52 ) 0. 02 7 1. 00 (0 .9 9-1. 01 ) 0. 80 6 H ig h ex po su re 1. 12 (. 87 -1 .3 8) 0. 31 1 1. 11 (0 .8 5-1. 36 ) 0. 37 0 1. 01 (1 .0 0-1. 03 ) 0. 04 8 1. 00 (0 .5 8-1. 43 ) 0. 98 6 1. 00 (0 .5 7-1. 42 ) 0. 99 3 1. 02 (1 .0 0-1. 04 ) 0. 09 6 M et al s Low e xp os ur e 1. 11 (0 .9 0-1. 33 ) 0. 28 2 1. 11 ( 0. 89 -1 .3 2) 0. 30 0 1. 01 (1 .0 0-1. 01 ) 0. 23 9 1. 52 (1 .0 1-2. 03 ) 0. 04 8 1. 52 (1 .0 1-2. 03 ) 0. 04 9 1. 01 (0 .9 9-1. 02 ) 0. 26 7 H ig h ex po su re 1. 72 (1 .2 1-2. 22 ) 0. 00 0 1. 69 (1 .2 0-2. 19 ) 0. 00 0 1. 01 (1 .0 0-1. 03 ) 0. 07 6 1. 92 (0 .9 7-2. 88 ) 0. 05 2 1. 92 (0 .9 6-2. 87 ) 0. 05 3 1. 02 (1 .0 0-1. 04 ) 0. 11 4 D ys pn ea G as es /fu m es Lo w e xp os ur e 1. 15 (0 .9 5-1. 35 ) 0. 12 9 1. 16 (0 .9 6-1. 36 ) 0. 11 7 1. 00 (0 .9 9-1. 01 ) 0. 68 6 H ig h ex po su re 1. 56 (1 .0 5-2. 07 ) 0. 00 0 1. 54 (1 .0 4-2. 05 ) 0. 03 2 1. 03 (0 .9 9-1. 03 ) 0. 45 4 So lv en ts Lo w e xp os ur e 1. 14 (1 .0 3-1. 26 ) 0. 00 8 1. 15 (1 .0 3-1. 26 ) 0. 00 7 1. 00 (0 .9 1 .0 0) 0. 44 0 1. 28 (1 .0 6-1. 51 ) 0. 02 7 1. 30 (1 .0 7-1. 53 ) 0. 02 2 0. 99 (0 .9 8-1. 01 ) 0. 43 7 H ig h ex po su re 1. 12 (. 87 -1 .3 8) 0. 31 1 1. 12 (0 .8 6-1. 38 ) 0. 31 8 1. 00 (0 .9 8-1. 02 ) 0. 92 0 1. 00 (0 .5 8-1. 43 ) 0. 98 6 1. 01 (0 .5 8-1. 44 ) 0. 94 9 1. 00 (0 .9 7-1. 03 ) 0. 92 2 M et al s Low e xp os ur e 1. 11 (0 .9 0-1. 33 ) 0. 28 2 1. 11 (0 .8 9-1. 32 ) 0. 29 4 1. 00 (0 .9 91 .0 2) 0. 58 8 1. 52 (1 .0 1-2. 03 ) 0. 04 8 1. 52 (1 .0 1-1. 94 ) 0. 04 9 1. 01 (0 .9 8-1. 03 ) 0. 55 0 H ig h ex po su re 1. 72 (1 .2 2 .2 2) 0. 00 0 1. 69 (1 .2 0-2. 19 ) 0. 00 0 1. 01 (0 .9 9-1. 04 ) 0. 15 5 1. 92 (0 .9 7-2. 88 ) 0. 05 2 1. 90 (0 .9 2 .7 0) 0. 05 5 1. 02 (0 .9 9-1. 05 ) 0. 18 2
45
S8 Table. Association between occupational exposures and respiratory symptoms. The logistic regression model was adjusted for age, sex, BMI, smoking, education, and monthly income.
S9 Table: Association between respiratory symptoms and sickness absence. The logistic regression model was adjusted for age, sex, BMI, smoking, education, and monthly income.
Occupational exposures
Chronic cough Chronic phlegm Dyspnea
OR (95% CI) P
value OR (95% CI) value P OR (95% CI) value P
Gases/fumes
No Reference Reference Reference
Low 1.09 (.96-1.24) .170 1.04 (.90-1.21) .567 1.01 (.88-1.17) .842
High 1.18 (.94-1.50) .159 1.29 (.99-1.67) .055 1.21 (.91-1.62) .191
Solvents
No Reference Reference Reference
Low .99 (.86-1.13) .835 .95 (.81-1.12) .539 .95 (.81-1.10) .489
High 1.07 (.80-1.43) .667 1.37 (1.00-1.88) .049 1.05 (.74-1.50) .771
Metals
No Reference Reference Reference
Low 1.24 (.98-1.57) .080 1.14 (.87-1.50) .350 1.11 (.81-1.52) .505
High .88 (.60-1.27) .488 1.18 (.80-1.75) .406 1.46 (.95-2.24) .086
Respiratory
symptoms OR (95% CI) Any sickness absence P-value Long-term sickness absence OR (95% CI) P-value Chronic Cough No Reference Reference Yes 1.52 (1.35-1.71) 0.000 1.46 (1.22-1.73) .000 Chronic Phlegm No Reference Reference Yes 1.59 (1.38-1.83) 0.000 1.74 (1.43-2.13) .000 Dyspnea No Reference Reference Yes 1.46 (1.28-1.67) 0.000 1.70 (1.40-2.05) .000
46 S1 0 Ta bl e: S N Ps -b y-B io lo gi ca l d us t/M in er al d us t/G as es a nd F um es in te ra ct io ns o n an y si ck ne ss a bs en ce . Va ria bl es in eq ua tio n B io lo gi ca l d us t M in er al d us t G as es a nd fu m es rs 17 49 00 56 C C : n = 24 49 (re fe re nc e) TC : n =4 78 9 TT : n =2 46 2 rs 13 27 85 29 TT : n = 71 05 (re fe re nc e) TG : n =2 39 0 G G : n =2 05 rs 47 38 92 C C : n = 29 07 (re fe re nc e) C T: n =4 80 4 TT : n =1 98 9 rs 67 51 43 9 G G : n = 73 51 (re fe re nc e) G A: n =2 17 7 AA : n =1 72 rs 28 88 67 4 G G : n = 28 03 (re fe re nc e) G A: n =4 82 2 AA : n =2 07 5 rs 15 94 97 TT : n = 27 63 (re fe re nc e) TC : n =4 88 2 C C : n =2 05 5 rs 51 67 32 TT : n = 27 80 (re fe re nc e) TC : n =4 81 8 C C : n =2 10 2 O R (C I 9 5% ) O R (C I 9 5% ) O R (C I 9 5% ) O R (C I 9 5% ) O R (C I 9 5% ) O R (C I 9 5% ) O R (C I 9 5% ) Lo w e xp os ur e 1. 00 (0 .8 3-1. 20 ) p = 0. 98 0 0. 86 (0 .7 5-0. 99 ) p = 0. 0 33 0. 75 (0 .6 1-0. 93 ) p = 0. 00 8 0. 87 (0 .7 6-1. 00 ) p = 0. 04 7 1. 17 (1 .0 0-1. 38 ) p = 0. 05 6 0. 88 (0 .7 4-1. 03 ) p = 0. 11 6 0. 93 (0 .7 9-1. 10 ) p = 0. 42 3 H ig h ex po su re 0. 54 (0 .3 6-0. 83 ) p = 0. 00 5 0. 99 (0 .7 8-1. 26 ) p = 0. 93 4 0. 94 (0 .6 6-1. 33 ) p = 0. 72 1 1. 04 (0 .8 3-1. 32 ) p = 0. 70 7 1. 40 (0 .9 9-1. 97 ) p = 0. 05 4 1. 00 (0 .7 2-1. 38 ) p = 0. 99 4 0. 93 (0 .6 7-1. 30 ) p = 0. 67 7 SN P-H Z 0. 97 (0 .8 6-1. 09 ) p = 0. 57 4 0. 94 (0 .8 4-1. 04 ) p = 0. 23 6 0. 96 (0 .8 6-1. 06 ) p = 0. 40 2 1. 03 (0 .9 3-1. 16 ) p = 0. 53 7 1. 14 (1 .0 1-1. 30 ) p = 0. 03 3 0. 95 (0 .8 4-1. 07 ) p = 0. 40 2 0. 93 (0 .8 2-1. 05 ) p = 0. 25 1 SN P-H M 0. 93 (0 .8 1-1. 07 ) p = 0. 31 3 1. 13 (0 .8 1-1. 58 ) p = 0. 46 9 0. 92 (0 .8 9-1. 05 ) p = 0. 21 4 0. 91 (0 .6 4-1. 30 ) p = 0. 61 8 1. 29 (1 .1 1-1. 50 ) p = 0. 00 1 0. 84 (0 .7 2-0. 97 ) p = 0. 02 1 0. 92 (0 .7 9-1. 07 ) p = 0. 27 8 Lo w e xp os ur e by S N P-H Z 1. 13 (0 .9 6-1. 42 ) p = 0. 27 5 1. 16 (0 .8 9-1. 50 ) p = 0. 27 1 1. 16 (0 .8 9-1. 49 ) p = 0. 26 8 1. 11 (0 .8 5-1. 45 ) p = 0. 42 9 0. 82 (0 .6 7-1. 01 ) p = 0. 05 7 1. 11 (0 .9 1-1. 36 ) p = 0. 29 5 1. 02 (0 .8 3-1. 25 ) p = 0. 83 8 Lo w e xp os ur e by S N P-H M 1. 05 (0 .8 1-1. 35 ) p = 0. 70 9 0. 83 (0 .4 1-1. 65 ) p = 0. 59 3 1. 58 (1 .1 5-2. 18 ) p = 0. 00 5 0. 77 (0 .3 4-1. 76 ) p = 0. 54 3 0. 64 (0 .5 0-0. 81 ) p = 0. 00 0 1. 21 (0 .9 5-1. 55 ) p = 0. 12 7 1. 11 (0 .8 7-1. 42 ) p = 0. 40 8 H ig h ex po su re by S N P-H Z 1. 47 (0 .8 8-2. 48 ) p = 0. 14 3 1. 40 (0 .9 0-2. 17 ) p = 0. 13 1 1. 12 (0 .7 2-1. 75 ) p = 0. 60 5 1. 00 (0 .6 3-1. 60 ) p = 0. 98 3 0. 67 (0 .4 5-1. 02 ) p = 0. 06 0 1. 08 (0 .7 2-1. 61 ) p = 0. 71 0 1. 23 (0 .8 2-1. 85 ) p = 0. 32 3 H ig h ex po su re by S N P-H M 1. 68 (0 .9 1-3. 07 ) p = 0. 09 4 0. 26 (0 .0 7-1. 04 ) p = 0. 05 6 1. 29 (0 .7 5-2. 21 ) p = 0. 35 3 1. 19 (0 .3 3-4. 30 ) p = 0. 79 1 0. 67 (0 .4 0-1. 12 ) p = 0. 13 0 1. 03 (0 .6 3-1. 71 ) p = 0. 89 5 1. 10 (0 .6 8-1. 78 ) p = 0. 68 9 Inte ra ct io ns a na ly si s ad ju st ed fo r a ge . s ex . B M I. ed uc at io n. s m ok in g st at us . a nd m on th ly in co m e. Bi ol og ic al d us t ( no e xp os ur e: n = 6 65 0 (re fe re nc e) . l ow e xp os ur e: n = 2 67 0. h ig h ex po su re : n = 3 80 ) M in er al d us t ( no e xp os ur e: n = 7 67 9 (re fe re nc e) . l ow e xp os ur e: n = 1 56 1. h ig h ex po su re : n = 4 60 ) G as es a nd fu m es (n o ex po su re : n = 5 64 1 (re fe re nc e) . l ow e xp os ur e: n = 3 46 9. h ig h ex po su re : n = 5 90 ) SN P-Si ng le n uc le ot id e po ly m or ph is m ; H Z-H et er oz yg ou s; H M -H om oz yg ou s fo r m in or a lle le .
47 S1 1 Ta bl e: S N Ps -b y-B io lo gi ca l d us t/M in er al d us t/G as es a nd F um es in te ra ct io ns o n lo ng -te rm s ic kn es s ab se nc e. Va ria bl es in eq ua tio n B io lo gi ca l d us t M in er al d us t G as es a nd fu m es rs 17 49 00 56 C C : n =1 46 1 (re fe re nc e) TC : n =2 81 9 TT : n =1 47 7 rs 13 27 85 29 TT : n =4 20 6 (re fe re nc e) TG : n =1 43 2 G G : n =1 19 rs 47 38 92 C C : n =1 69 1 (re fe re nc e) C T: n =2 87 4 TT : n =1 19 2 rs 67 51 43 9 G G : n =4 34 3 (re fe re nc e) G A: n =1 29 7 AA : n =1 17 rs 28 88 67 4 G G : n =1 70 4 (re fe re nc e) G A: n =2 85 0 AA : n =1 20 3 rs 15 94 97 TT : n =1 61 5 (re fe re nc e) TC : n =2 89 4 C C : n =1 24 8 rs 51 67 32 TT : n =1 62 9 (re fe re nc e) TC : n =2 88 9 C C : n =1 23 9 O R (C I 9 5% ) O R (C I 9 5% ) O R (C I 9 5% ) O R (C I 9 5% ) O R (C I 9 5% ) O R (C I 9 5% ) O R (C I 9 5% ) Lo w e xp os ur e 1. 20 (0 .9 1-1. 58 ) p = 0. 20 3 0. 96 (0 .7 8-1. 19 ) p = 0. 71 2 0. 90 (0 .6 5-1. 24 ) p = 0. 51 8 1. 00 (0 .8 1-1. 23 ) p = 0. 99 6 1. 29 (1 .0 1-1. 66 ) p = 0. 04 4 0. 85 (0 .6 5-1. 11 ) p = 0. 22 8 1. 08 (0 .8 4-1. 40 ) p = 0. 54 8 H ig h ex po su re 0. 95 (0 .5 3-1. 72 ) p = 0. 87 0 1. 27 (0 .9 0-1. 79 ) p = 0. 17 5 1. 14 (0 .6 7-1. 93 ) p = 0. 63 6 1. 22 (0 .8 6-1. 72 ) p = 0. 26 6 1. 79 (1 .0 9-2. 94 ) p = 0. 02 1 1. 28 (0 .8 0-2. 06 ) p = 0. 30 1 1. 23 (0 .7 6-2. 00 ) p = 0. 39 9 SN P-H Z 1. 04 (0 .8 5-1. 26 ) p = 0. 73 1 0. 95 (0 .8 0-1. 13 ) p = 0. 53 6 1. 07 (0 .9 0-1. 27 ) p = 0. 46 9 1. 14 (0 .9 6-1. 36 ) p = 0. 13 6 1. 03 (0 .8 4-1. 27 ) p = 0. 75 3 0. 88 (0 .7 1-1. 08 ) p = 0. 21 5 0. 94 (0 .7 6-1. 16 ) p = 0. 57 2 SN P-H M 0. 92 (0 .7 3-1. 15 ) p = 0. 46 3 1. 19 (0 .7 0-2. 02 ) p = 0. 51 4 1. 07 (0 .8 7-1. 33 ) p = 0. 50 5 1. 42 (0 .8 6-2. 36 ) p = 0. 17 3 1. 12 (0 .8 7-1. 44 ) p = 0. 39 6 0. 81 (0 .6 3-1. 04 ) p = 0. 09 8 0. 90 (0 .6 9-1. 16 ) p = 0. 40 4 Lo w e xp os ur e by S N P-H Z 0. 87 (0 .6 2-1. 22 ) p = 0. 41 6 1. 17 (0 .8 0-1. 73 ) p = 0. 42 0 1. 06 (0 .7 1-1. 57 ) p = 0. 78 8 0. 95 (0 .6 3-1. 41 ) p = 0. 78 4 0. 90 (0 .6 6-1. 23 ) p = 0. 50 3 1. 51 (1 .1 0-2. 08 ) p = 0. 01 2 1. 04 (0 .7 6-1. 43 ) p = 0. 80 5 Lo w e xp os ur e by S N P-H M 0. 94 (0 .6 4-1. 40 ) p = 0. 77 3 0. 23 (0 .0 5-1. 09 ) p = 0. 06 4 1. 32 (0 .8 1-2. 14 ) p = 0. 26 6 0. 48 (0 .1 3-1. 84 ) p = 0. 28 6 0. 67 (0 .4 5-0. 98 ) p = 0. 04 1 1. 39 (0 .9 3-2. 06 ) p = 0. 10 5 1. 11 (0 .7 5-1. 63 ) p = 0. 61 0 H ig h ex po su re by S N P-H Z 0. 73 (0 .3 4-1. 60 ) p = 0. 43 5 1. 30 (0 .6 9-2. 43 ) p = 0. 41 3 1. 17 (0 .6 1-2. 24 ) p = 0. 63 8 1. 21 (0 .6 4-2. 28 ) p = 0. 55 9 0. 70 (0 .3 9-1. 28 ) p = 0. 25 0 1. 24 (0 .7 0-2. 21 ) p = 0. 46 8 1. 31 (0 .7 2-2. 37 ) p = 0. 37 0 H ig h ex po su re by S N P-H M 1. 42 (0 .6 1-3. 32 ) p = 0. 41 7 0. 28 (0 .0 3-2. 39 ) p = 0. 24 3 1. 37 (0 .6 3-2. 99 ) p = 0. 42 5 1. 81 (0 .4 1-8. 00 ) p = 0. 43 6 0. 84 (0 .4 1-1. 73 ) p = 0. 64 5 1. 01 (0 .4 8-2. 15 ) p = 0. 97 3 1. 13 (0 .5 6-2. 29 ) p = 0. 73 3 In te ra ct io ns a na ly si s ad ju st ed fo r a ge . s ex . B M I. ed uc at io n. s m ok in g st at us . a nd m on th ly in co m e. Bi ol og ic al d us t ( no e xp os ur e: n = 3 89 7 (re fe re nc e) . l ow e xp os ur e: n = 1 59 2. h ig h ex po su re : n = 2 68 ) M in er al d us t ( no e xp os ur e: n = 4 45 5 (re fe re nc e) . l ow e xp os ur e: n = 1 00 8. h ig h ex po su re : n = 2 94 ) G as es a nd fu m es (n o ex po su re : n = 3 22 6 (re fe re nc e) . l ow e xp os ur e: n = 2 15 7. h ig h ex po su re : n = 3 74 )
