Occupational Exposures and the Co-Occurrence of Work-Related Skin and Respiratory Symptoms

Transcription

1 Occupational Exposures and the Co-Occurrence of Work-Related Skin and Respiratory Symptoms by Victoria Helen Arrandale A thesis submitted in conformity with the requirements for the degree of Doctor of Philosophy Institute of Medical Science University of Toronto Copyright by Victoria Helen Arrandale (2012)

2 Occupational Exposures and the Co-Occurrence of Work- Related Skin and Respiratory Symptoms Abstract Victoria Helen Arrandale Doctor of Philosophy Institute of Medical Science University of Toronto 2012 Occupational skin and respiratory symptoms, and disease, are common problems. Workers can develop new disease or aggravate existing disease as a result of exposures at work. Many workers are exposed to chemicals that can cause both respiratory and skin responses and there is evidence that some workers experience symptoms in both systems. There is also evidence that skin exposure may lead to sensitization and the development of respiratory disease. There is very little research that has examined both airborne and skin exposures together with lung and skin outcomes. The purpose of this thesis was to further investigate the relationships between occupational exposures, skin symptoms and disease, and respiratory symptoms and disease. Four studies were undertaken to improve our understanding of these complex relationships. Results from a study of clinical patch test data determined that seven of the ten most common occupational contact allergens are also capable of causing occupational asthma and that these common occupational exposures may not be recognized as sensitizers in common reference materials. Exposure-response relationships for skin symptoms were modeled in bakery workers and auto body shop workers using historical data; significant exposure-response relationships were found for auto body workers. In two separate studies of concurrent skin and respiratory symptoms, ii

3 workers did report concurrent skin and respiratory symptoms. In predictive models, subjects reporting a history of eczema were more likely to report concurrent skin and respiratory symptoms. Overall, the results from this thesis provide more evidence that the skin and respiratory systems are associated. This body of work suggests that: (1) several common occupational exposures can cause disease in both the skin and respiratory system; (2) a portion of workers report both skin and respiratory symptoms; and (3) exposure-response relationships do exist for skin symptoms, both work-related and non-work-related. Future studies need to gather detailed information about exposure and response in both systems in order to better determine the role of exposure(s) in the development of skin and respiratory symptoms. Improved understanding of these relationships will allow for more targeted and effective exposure prevention strategies and will ultimately reduce the burden of occupational disease. iii

4 Acknowledgments First and foremost I would like to thank my committee members, Dr. Jeremy Scott, Dr. Susan Tarlo, Dr. Frances Silverman and especially Dr. Linn Holness for their support, guidance and feedback throughout his process. I would like to extend thanks to Dr. Dick Heederik for his warm welcome in Utrecht, The Netherlands, and for allowing me to explore the occupational data that his group has collected. I am also very thankful to Dr. Carrie Redlich and Dr. Allen Kraut for their feedback on several key aspects of this thesis. I must thank the staff at the St Michael s Hospital Occupational Health Clinic for their help in data collection, especially their sense of humour when things didn t go exactly as planned. I m also grateful to the staff at the Gage Occupational and Environmental Health Unit, my academic home for the last five years. I would also like to thank the Massey College community that has served as my home away from home. I am forever grateful for the friendships that I know will endure, and the memories that I will carry with me. I am grateful to the North American Contact Dermatitis Group (NACDG) for providing data access and both the Research Advisory Council of the Ontario Workplace Safety and Insurance Board as well as the Workers Compensation Board of Manitoba for providing operating grant support for the research studies. Both the Canadian Institutes of Health Research and the Centre for Research Expertise in Occupational Disease supported my stipend and this financial support made this thesis possible. I would also like to thank Nigel, Carole and Samantha. We may have rendezvoused in San Francisco, New York, and Paris but nothing beats being together, fireside, at And finally, thank you to Taylor for your patience, love and fierce editing skills. iv

5 Table of Contents Acknowledgments... iv Table of Contents... v List of Abbreviations... x List of Tables... xv List of Figures... xviii List of Appendices... xix Chapter 1 Literature Review Background Asthma Work-Related Asthma (WRA) Diagnosis of WRA Prevalence of WRA Causes of WRA Epidemiological Evidence: Relationship Between Symptoms and Disease Contact Dermatitis Occupational Contact Dermatitis (OCD) Diagnosis of OCD Prevalence of OCD Causes of OCD Epidemiological Evidence: Relationship between Symptoms and Disease Occupational Exposure Exposure-Response Relationships Connecting the Skin and Respiratory Systems Occupational Exposure v

6 1.5.2 Mechanisms of Effect Concurrent Skin and Respiratory Disease in Individuals Epidemiological Evidence Linking Skin and Respiratory Outcomes Possible Cross-System Sensitization Cross-System Interaction in Other Disease Models Additional Considerations Framework Chapter 2 Research Aims and Hypotheses Knowledge Gaps Research Aims Specific Research Aims Hypotheses Chapter 3 Occupational Contact Allergens: Are They Also Associated With Occupational Asthma? Abstract Introduction Methods Diagnosis of Occupational Allergic Contact Dermatitis Determination of Occupationally-Relevant Positive Patch Test Responses Determination of Whether OCAs May Also Cause OA Determination of Skin Sensitizer Notation Status Results Common Occupational Contact Allergens Occupational Contact Allergens as a Cause of Occupational Asthma Sensitizer Notations Discussion vi

7 3.5.1 Limitations Chapter 4 Co-existing Skin and Respiratory Symptoms in Four Occupational Groups Abstract Introduction Methods Results Discussion Chapter 5 Skin Symptoms in Bakery and Auto Body Shop Workers: Associations with Exposure and Respiratory Symptoms Abstract Introduction Methods Exposure Specific IgE and Atopy Symptoms Additional Variables Analyses Results Discussion Chapter 6 Skin and Respiratory Symptoms Among Workers with Suspected Work- Related Disease Abstract Introduction Methods Outcome Variables Predictor Variables vii

8 6.3.3 Statistical Analyses Results Concurrent Symptoms Discussion Limitations Conclusions Chapter 7 General Discussion Revisiting Research Aims and Hypotheses Methodological Considerations Causes of Occupational Skin and Respiratory Disease Surveillance of OCD and OA Knowledge Translation and Communication Modeling Exposure-Response Relationships Workers Do Report Concurrent Skin and Respiratory Symptoms Symptom Progression Predictors of Concurrent Skin and Respiratory Symptoms Barrier Function and Concurrent Skin and Respiratory Symptoms Personal Protective Equipment Smoking Mechanism of Effect Strengths & Limitations Strengths Limitations Contribution to the Literature Chapter 8 Conclusions Chapter 9 Future Directions viii

9 References Appendices Copyright Acknowledgements ix

10 List of Abbreviations AA ACD ACGIH AD ANOVA AOCD ATS ATSQ BADGE BEI BMRC CD CI Conc Derm Df DGEBA DNCB DREAM Allergy/Asthma Allergic Contact Dermatitis American Conference of Governmental Industrial Hygiene Atopic Dermatitis Analysis of Variance Allergic Occupational Contact Dermatitis American Thoracic Society American Thoracic Society Questionnaire Bisphenol A Diglycidyl Ether Biological Exposure Indices British Medical Research Questionnaire Contact Dermatitis Confidence Intervals Concurrent Dermatology Degrees of Freedom Diglycidyl Ether of Bisphenol A Dinitrochlorobenzene Dermal Exposure Assessment Method x

11 ECCS ECRHS EPIDERM ESSCA FEV 1 FROD FVC GCV HDI HDM HEMA HMW HSE ICD IgE IQR IUATLD kda LMW MDI European Community of Coal and Steel European Community Respiratory Health Survey Occupational Skin Surveillance (UK) European Surveillance System on Contact Allergies Forced Expiratory Volume in One Second Finnish Register of Occupational Disease Forced Vital Capacity Generalized Cross Validation Hexamethylene diisocyanate House Dust Mite Hydroxyethyl Methacrylate High Molecular Weight Health and Safety Executive (UK) Irritant Contact Dermatitis Immunoglobin E Inter-quartile Range International Union Against Tuberculosis and Lung Disease Kilodalton Low Molecular Weight Methylene Diphenyl Diisocyanate xi

12 MRC MSDS NACDG NCO NIH NIOSH NLM NOCS NPG NR NS OA OCA OCD OH OHIP OPRA OR OSD PAC Medical Research Council Material Safety Data Sheet North American Contact Dermatitis Group Isocyanate Functional Group (Nitrogen, Carbon, Oxygen) National Institutes of Health National Institute of Occupational Safety and Health (USA) National Library of Medicine National Occupational Classification System NIOSH Pocket Guide to Chemical Hazards Not Reported Not Significant Occupational Asthma Occupational Contact Allergen Occupational Contact Dermatitis Occupational Hygiene Ontario Health Insurance Plan Occupational Physicians Reporting Activity (UK) Odds Ratio Occupational Skin Disease Polycyclic Aromatic Compounds xii

13 PPD PPE PPT PR Pred. PT RADS Resp SABRE SAS sd SEN Sens SIC SLS Spec Sx SWORD TDI TEWL P-phenylenediamine Personal Protective Equipment Positive Patch Test Prevalence Ratio Predicted Patch Test Reactive Airways Dysfunction Syndrome Respiratory Surveillance of Australian Workplace Based Respiratory Events Statistical Analysis Software (program) Standard Deviation Sensitizer Notation Sensitivity Specific Inhalation Challenge Sodium Laurel Sulfate Specificity Symptom(s) Surveillance of Work-Related and Occupational Respiratory Disease (UK) Toluene Diisocyanate Transepidermal Water Loss xiii

14 THOR TLV TMA VITAE WEA WRA WR WSIB The Health and Occupation Research Network (UK) Threshold Limit Value Trimellitic Anhydride Video Imaging Technique for Assessing Occupational Skin Exposure Work-Exacerbated Asthma Work-Related Asthma Work-Related Workplace Safety and Insurance Board (of Ontario) xiv

15 List of Tables Table 1 Agents Potentially Causing Work-Related Asthma (WRA)... 7 Table 2 Common Agents Causing Occupational Contact Dermatitis, Both Irritant and Allergic Table 3 Summary of Literature Investigating the Relationship between Symptom Reporting and Skin Disease Diagnoses Table 4 Correlation Coefficients for the Association Between Skin and Airborne Exposures in Various Occupational Studies Table 5 Exposures Reported in Published Case Studies to Cause Both Occupational Asthma and Occupational Contact Dermatitis Table 6 Summary of Experimental Animal Studies Demonstrating Skin Exposure Resulting in Sensitization and an Asthma-like Response on First Inhalation Challenge Table 7 Basic Descriptive Statistics for the Entire Study Population, Subjects with an Allergic Contact Dermatitis (ACD) Diagnoses and ACD Cases Stratified by Occupational Relatedness Table 8 Ten Most Common Occupational Contact Allergens (OCAs) Table 9 Summary of the Ten Most Frequent Occupational Contact Allergens (OCAs) and the Evidence Linking Each to OA in Asthma in the Workplace and the UK HSE Asthmagen Table 10 Categorization of Whether Each Common Occupational Contact Allergen (OCA) Has the Potential To Cause OA Based on Reference Sources and Systematic Literature Review, Where Necessary Table 11 Summary of Sensitizer Notations for the Ten Most Common OCAs in Common Occupational Hygiene Reference Documents xv

16 Table 12 Skin and Respiratory Symptom Group Distribution (Work-Related and Non- Work-Related) Across Studies and Description of Groups by Age, Sex, Smoking and Pulmonary Function Variables Table 13 Demographics and Symptom Frequencies for Both Auto Body Repair and Bakery Workers Table 14 Results of Generalized Linear Models Describing the Simple Relationship Between Exposure, Skin Symptoms, Atopy and Specific IgE Table 15 Prevalence Ratio (PR) of Symptoms per Inter-Quartile Range (IQR) Increase in Average Exposure Table 16 Association Between Skin Symptoms and Respiratory Symptoms in Both Bakery and Auto Body Repair Workers Table 17 Demographic Description of Study Population, Stratified by Subjects Who Reported Both Skin and Respiratory Symptoms Table 18 Skin and Respiratory Symptom Prevalence, Stratified by Clinical Stream Table 19 Self-Reported Workplace Characteristics, Stratified by Subjects Who Reported Both Skin and Respiratory Symptoms Table 20 Self-Reported Workplace Exposures, Stratified by Subjects Who Reported Both Skin and Respiratory Symptoms Table 21 Multiple Logistic Regression Model Results for Predictors of Reporting Concurrent Skin and Respiratory Symptom Outcomes Table 22 Description of Possible Outcome Groups when Considering Both Skin and Respiratory Symptom Outcomes and their Individual Work-Relatedness Table 23 Results of Generalized Linear Models Describing the Simple Relationship Between Exposure, Respiratory Symptoms, Atopy and Specific IgE xvi

17 Table 24 Comparison Between Participants and Patients Who Refused Study Participation Table 25 Comparison Between Dermatology (Derm) Strem and Asthma/Allergy (AA) Stream Subjects Table 26 Comparison Between Public Insurance Subjects (OHIP) and Workplace Insurance Subjects (ODSP) Table 27 Comparison Between Participants and Non-Participants in the Reliability Testing Study Table 28 Test-Retest Reliability Statistics for Workplace Characteristics Questionnaire. 178 Table 29 Test-Retest Reliability Statistics for Workplace Exposure Questionnaire Items.179 Table 30 Multiple Logistic Regression Model Results for Predictors of Reporting Work-Related Concurrent Skin and Respiratory Symptom Outcomes xvii

18 List of Figures Figure 1 Schematic Describing the Relationships Between Work-Related Asthma (WRA), Work-Exacerbated Asthma (WEA) and Occupational Asthma (OA)... 3 Figure 2 A Proposed Framework for Conceptualizing the Connections Between Skin and Respiratory Symptoms and Disease Figure 3 Auto Body Shop Workers: Associations Between Average Diisocyanate Exposure and Skin Symptoms, Shown in Smoothed Plots, Stratified by Atopy Figure 4 Bakery Workers: Associations Between Average Wheat Exposure and Skin Symptoms, Shown in Smoothed Plots, Stratified by Atopy Figure 5 Flow Chart of Study Progression, Including Sample Sizes at Each Stage Figure 6 Modified Framework for Conceptualizing the Connections Between Skin and Respiratory Symptoms in Occupational Disease Figure 7 Auto Body Shop Workers Associations Between Average Isocyanate Exposure and Respiratory Symptoms, Shown In Smoothed Plots Stratified by Atopy Figure 8 Bakery Workers Associations Between Average Wheat Exposure and Respiratory Symptoms, Shown in Smoothed Plots Stratified by Atopy xviii

19 List of Appendices Appendix 1: Supplemental Figures for Chapter 5 Skin Symptoms in Bakery and Auto Body Shop Workers: Associations with Exposure and Respiratory Symptoms Appendix 2: Interviewer-Administered Questionnaire for Chapter 6 Skin and Respiratory Symptoms Among Workers with Suspected Work-Related Disease Appendix 3: Supplemental Tables for Chapter 6 Skin and Respiratory Symptoms Among Workers with Suspected Work-Related Disease xix

20 1 Chapter 1 Literature Review 1.1 Background Occupational lung disease and occupational skin disease are common problems. Many workers are exposed to chemicals that are thought to cause respiratory and/or skin responses as a result of either airborne or skin exposure. Historically, there has been a significant amount of work focused on individual respiratory or skin outcomes and their association with particular (route specific) exposures, but there has been very little work done examining both airborne and skin exposures together with lung and skin outcomes. This lack of evidence creates a problem for recognition, diagnosis, and prevention of disease. Clinically, disease in the other system may be under-recognized when workers are assessed by either skin or respiratory physicians. From a prevention standpoint, opportunities for exposure control and disease prevention may be missed if research continues to focus only on airborne or skin exposure. There is a need to better understand the relationships between skin and airborne exposures, between skin and respiratory outcomes, and also the complicated relationships between both routes of exposure and outcomes in both systems. 1.2 Asthma Asthma is a heterogeneous inflammatory disorder characterized by variable airflow limitation and/or airway hyper-responsiveness [Holgate. 2008, Lombardo and Balmes. 2000]. The mechanisms that can cause an asthmatic response are varied: antigen induced hypersensitivity, pharmacologic effect, nonspecific inflammation and direct irritation of the airways [Lombardo and Balmes. 2000]. In the general population, approximately 7.7% of adults have asthma [Akinbami et al. 2011]. A review by Toren and Blanc suggests that between 15-20% of all adult asthma (new-onset and exacerbation) is attributable to work-place exposures [Toren and Blanc. 2009].

21 Work-Related Asthma (WRA) Work-related asthma (WRA) is one of the most common occupational lung diseases [Lombardo and Balmes. 2000]. WRA describes all asthma that is caused or made worse by one s work. Under the umbrella of work-related asthma (WRA) there is both occupational asthma (OA) and work-exacerbated asthma (WEA) [Tarlo et al. 2008] Work-Exacerbated Asthma (WEA) Workers who have pre-existing asthma (either current, or quiescent) may experience aggravation of their asthma symptoms, or re-development of asthma, in response to workplace exposures [Banks and Jalloul. 2007, Goe et al. 2004, Pelissier et al. 2006] Occupational Asthma (OA) Occupational asthma (OA) can be defined as asthma due to causes and conditions attributable to a particular occupational environment and not to stimuli encountered outside the workplace [Bernstein et al. 2006b]. Within OA there are two possible mechanisms of response: allergic and irritant. Allergic and irritant OA differ in latency and mechanism of effect Allergic Occupational Asthma Allergic asthma is characterized by a latency period between the onset of exposure and the onset of symptoms, which allows time for the characteristic development of an immunological response [Bernstein et al. 2006b]. Two types of sensitizers are described as causative agents in occupational allergic asthma, high and low molecular weight proteins and chemicals [Toren et al. 2000]. High molecular weight biologic proteins (e.g., animal or plant proteins) stimulate the production of specific IgE antibodies. Some low molecular weight agents (e.g., complex platinum salts and epoxy compounds) act as haptens, binding with a protein to form an antigen, and inducing occupational asthma via a specific IgE-mediated mechanism. Other low molecular weight sensitizers, including diisocyanates and plicatic acid (Western red cedar) act through mechanisms that are not yet fully understood [Tarlo et al. 2006] Irritant Occupational Asthma Historically, irritant induced asthma was thought to occur after a single exposure to irritant agents and was described as reactive airways dysfunction syndrome (RADS) [Bernstein et al.

22 3 2006b, Brooks et al. 1985] but there is increased understanding and recognition of the chronic lung effects of irritant exposures. There is now consensus that irritant asthma may result from repeated exposures, and there may also be a latency period between exposure and symptoms in irritant-induced asthma. Common exposures that can cause RADS include chlorine gas, anhydrous ammonia and fire smoke [Gautrin et al. 1999]. Lime, welding fumes and hydrochloric acid are potential causes of irritant induced asthma [Burge et al. 2011]. Figure 1 Schematic Describing the Relationships Between Work-Related Asthma (WRA), Work-Exacerbated Asthma (WEA) and Occupational Asthma (OA) As Described by Tarlo et al. [Tarlo et al. 2008] Diagnosis of WRA The diagnosis of OA is made based on a combination of work history and clinical evaluation [Tarlo et al. 2008, Tarlo et al. 1998]. The American College of Chest Physicians Consensus Statement outlines the following steps in diagnosis [Tarlo et al. 2008]: 1. Confirm Asthma and Onset: Using medical history, reported symptoms, spirometry results, and medications

23 4 2. Assess Exposures that Cause/Exacerbate Asthma: Using the occupational history, environmental history, reported/confirmed allergies, and atopy. 3. Assess Relationship of Asthma to Work: Based on reported symptoms (onset, severity and timing), physiological findings (peak flows, spirometry, methacholine challenge, specific inhalation challenge), and results of immunologic tests 4. Decide Whether the Asthma is Primary OA or WEA. In order to assess the relationship between asthma and work, clinical investigations may use several diagnostic tools. Repeated peak expiratory flow measures throughout the day (at work and at home), and over several days, will provide a measure of variability in airflow limitation. Nonspecific bronchial hyper-responsiveness can be measured using methacholine challenge, and spirometry can measure standard lung function parameters including the forced expiratory volume in one second (FEV 1 ). Both spirometry and methacholine challenge can be repeated before and after a work shift to measure changes over the course of the workday, providing the subject is still in the workplace and not off due to illness. Where the experimental equipment is available, the patient can also undergo a specific inhalation challenge (SIC), the gold standard in the diagnosis of occupational asthma. In a SIC the patient is exposed to the workplace agent suspected of causing asthma in a controlled environment. A positive response (i.e., decreased FEV 1, decreased peak flow, classical asthmatic reaction ) following the exposure confirms it as causative. The limitations of the SIC include the time and complexity of the testing set-up, but also the fact that only one exposure can be tested at a time; most workplaces have mixed exposures which can make identifying the causal exposure via SIC challenging Prevalence of WRA The burden of work-related asthma is difficult to assess. The criteria for compensation differ by country or region; some regions compensate for work-exacerbated asthma while others do not. In regions where compensation exists, detailed data are generally collected only for cases that are accepted for compensation. There are surveillance schemes that collect information on cases of occupational asthma, but the organization of each scheme differs. Often, participation in

24 5 surveillance or reporting schemes for occupational asthma (and other occupational diseases) is voluntary, leading to under-reporting of disease. In Finland, physicians are required to report any known or suspected occupational diseases to the Finnish Registry of Occupational Disease (FROD). The requirement for mandatory reporting makes the FROD one of the best occupational surveillance systems. From , over 2600 cases of OA were reported to FROD, corresponding to a mean annual incidence rate of 17.4 cases per 100,000 workers [Karjalainen et al. 2000, Karjalainen et al. 2000]. The most recent data available (2002) permits the calculation of an incidence rate for OA of 12.9 cases per 100,000 workers across all industries and occupations [Riihimäki et al. 2004]. In the United Kingdom (UK), chest and occupational physicians voluntarily report cases of occupational asthma to both SWORD (Surveillance of Work-Related and Occupational Respiratory Disease) and OPRA (Occupational Physicians Reporting Activity). In 2001 these schemes were subsumed into a larger program, The Health and Occupational Research Network (THOR). From , 897 reports of occupational asthma were received; the estimated average incidence of occupational asthma was 28 cases per million people (men) and 14 cases per million people (women) [McDonald et al. 2005]. In Australia, the Surveillance of Australian workplace Based Respiratory Events (SABRE) is voluntary reporting scheme that began 1997 and involves both thoracic and occupational physicians [Hannaford-Turner et al. 2010]. From June 2001 to December 2008 the New South Wales region of SABRE received 3856 reports including 89 cases of OA; the majority of reports were made up by pleural plaques (32%), mesothelioma (24%), diffuse pleural plaques (22%), asbestosis (10%) and lung cancer (5%). The prevalence of OA can be converted to a rate of 18 cases per million people, based on approximately 4.8 million NSW Australians of working age [Australian Bureau of Statistics. 2011]. The difference in the population estimates between Finland, the UK and Australia the Finnish estimates are an order of magnitude larger than the UK and Australia - is likely due to the mandatory nature of the Finnish system, compared with the voluntary system in the UK and Australia.

25 6 It is also useful to look to population studies that have attempted to estimate the portion of new onset adult asthma that may be work-related as another measure of the burden of work-related asthma. Blanc and Toren have comprehensively reviewed reports since 1966 in two successive publications, one in 1999 and the other in [Toren and Blanc. 2009, Blanc and Toren. 1999]. The most recent summary included seventeen published studies of adult onset asthma and estimated the median population attributable fraction (PAR) of all asthma due to occupation to be 17% [Toren and Blanc. 2009]. This suggests that almost one-fifth of all adult asthma may occur as a result of occupational exposures Causes of WRA More than 300 causes of OA have been reported around the world [Mapp et al. 2005]. Two publications stand out for the thoroughness of their summary: the Appendix compiled by Malo and Chan-Yeung within the text book Asthma in the Workplace, edited by Dr. L. I. Bernstein [Bernstein et al. 2006a] and the publication from van Kampen et al. [van Kampen et al. 2000]. Both provide a thorough list of exposures that have been shown to cause occupational asthma, including the clinical, symptom and diagnostic information reported in each case. Malo and Chan-Yeung published an appended version of their list in 2009, which identifies the most common agents associated with occupational asthma separated into low molecular weight (LMW) (<5 kda) and high molecular weight (HMW) ( 5 kda) antigens. The common LMW agents include anhydrides, metals, diisocyanates, cleaning agents, wood dusts, soldering fluxes, pesticides, pharmaceuticals and reactive dyes. HMW agents include enzymes, cereals, flour, animals and latex [Malo and Chan-Yeung. 2009]. Mapp et al. identify a similar list, but also include epoxy compounds and persulfate as common workplace sensitizers that may cause occupational asthma [Mapp et al. 2005]. Mapp also published a more detailed list that also included the HMW agents of crustacean, arthropods, moulds and plants as well as the LMW agents anhydrides, aliphatic amines, biocides, fungicides, acrylates, metal working fluids, perfumes, and general irritants [Mapp. 2001]. Known causes of occupational asthma are summarized in Table 1 and separated into three categories: LMW, HMW, and Irritant. Using reported surveillance data, previous studies have attempted to identify which of the known causes of OA are the most common. The published studies have presented very similar common causes of OA. Hannaford-Turner et al. reported that the most common agents causing

26 7 occupational asthma in an Australian surveillance program were flour, diisocyanates, and solvents [Hannaford-Turner et al. 2010]. In the UK, diisocyanates, flour, wood dust, metals, solder/colophony, glutaraldehyde, and epoxy were the most common agents reported by physicians as causing OA [McDonald et al. 2005]. Animals, flour, mites, diisocyanates, and welding fumes are the five most common causative agents in the Finnish data from [Karjalainen et al. 2000]. Table 1 Agents Potentially Causing Work-Related Asthma (WRA) Based on Peer- Reviewed Publications [Mapp et al. 2005, Malo and Chan-Yeung. 2009, Mapp. 2001]. Acrylates Aliphatic amines Anhydrides Biocides Cleaning agents Diisocyanates Drugs/Pharmaceuticals Epoxy compounds Fungicides Hardeners Metals Persulfate Pesticides Reactive dyes Solder fluxes Synthetic materials Wood dust or bark LMW HMW Irritant Animal derived allergens Arthropods Biological enzymes Crustaceans, seafood, fish Moulds Plants (Latex) General Irritants Perfumes

27 Epidemiological Evidence: Relationship Between Symptoms and Disease There is a detailed body of literature that has examined the inter-relationships between selfreported symptoms, bronchial hyper-responsiveness, and asthma diagnoses. One important goal of these studies has been to determine if, and how, population studies can measure asthma using questionnaire items rather than expensive and time-consuming diagnostic testing. The result is a better understanding of which symptom questions, or combination of symptom questions, can serve as a reasonable proxy for an asthma diagnosis among study subjects. For the most part, these studies have focused on general asthma, and not occupational asthma. Among adults in an Australian study physician diagnosis was the gold standard. Self-reported asthma with a reported attack in the last 12 months had a higher Youden s index (combination of sensitivity and specificity) than measured bronchial hyper-reactivity [Jenkins et al. 1996]. In New Zealand adults, all of self-reported wheeze, wheeze with dyspnea, and wheeze without cold had better performance (measured as a higher Youden s Index) than non-specific bronchial hyper-responsiveness when compared with self-reported doctor-diagnosed asthma [Sistek et al. 2006]. Pekkanen et al. looked at the same symptoms questions in relation to self-reported everasthma (not necessarily doctor-diagnosed) in the European Community Respiratory Health Study (ECRHS) [Pekkanen et al. 2005]. Though results suggested worse performance (lower Youden s indices) than Sistek et al., the wheeze questions were still more highly associated with asthma than measured non-specific bronchial hyper-responsiveness [Pekkanen et al. 2005]. Vandenplas et al. focused on the ability of symptoms to predict occupational asthma (diagnosed by specific inhalation challenge) and showed that wheezing at work was the strongest single questionnaire item for predicting occupational asthma [Vandenplas et al. 2005]. This association between wheeze at work and occupational asthma diagnosis was even stronger when the population was limited to workers with HMW exposures [Vandenplas et al. 2005]. The research addressing the relationship between questionnaire responses and diagnosis of asthma provides occupational health and population researchers with a methodological

28 9 alternative to diagnostic testing, particularly in large studies. This literature also allows for better understanding of the implications of using a questionnaire tool for identifying subjects with asthma, or occupational asthma. 1.3 Contact Dermatitis Contact dermatitis is defined as an inflammatory skin reaction to direct contact with noxious agents in our environment [Lachapelle. 1995]. Clinical contact dermatitis presents as itching, redness, scaling, erythema, vesiculation, and papulovesicles [Diepgen and Coenraads. 1999]. Causes of contact dermatitis can be grouped into three categories: physical, biological, and chemical [Lushniak. 2004]. For contact dermatitis to be considered as work-related the exposure to the causal agent must occur in the workplace. As in the case of occupational asthma there are two major types of contact dermatitis - irritant and allergic contact dermatitis. Although the mechanisms underlying the development of allergic contact dermatitis and irritant contact dermatitis are different, the physical appearances of the two inflammatory responses are similar in many aspects [Marks et al. 1992]. A number of industrial agents are classified as both irritants and allergens. Individuals who have underlying atopic dermatitis may also develop aggravation of their disease related to workplace exposures [Marks et al. 1992] Occupational Contact Dermatitis (OCD) Allergic Contact Dermatitis (ACD) Allergic contact dermatitis (ACD) is the result of a delayed, cell-mediated (Type IV) immunologic response [Kimber et al. 2002]. Similar to allergic asthma, ACD is characterized by a latency period between the onset of exposure and the onset of symptoms. This latency allows for the induction of sensitization, prior to the elicitation of the allergic response upon subsequent exposure [Kimber et al. 2002] Irritant Contact Dermatitis (ICD) Irritant contact dermatitis (ICD) is the direct toxic effect of a chemical agent on the skin, following either a single application (e.g., an acute response such as a chemical burn) or repeated applications [English. 2004].

29 Work-Exacerbated Dermatitis Workers who have a history of atopic dermatitis (AD) may develop aggravation of their dermatitis from exposure to physical or chemical irritants in the workplace leading to workaggravated atopic dermatitis. Unlike WEA, there is very little published research on work aggravation of atopic dermatitis Diagnosis of OCD The diagnosis of occupational contact dermatitis is based on the occupational exposure history, the temporal relationships between exposure and disease, physical examination, and patch testing [Mathias. 1994]. Patch testing is a specialized technique that involves applying a small amount of a chemical with a known concentration, to the upper back for at least 48 hours [Zug et al. 2009]. The site where the exposure patch has been placed is occluded for 48 hours, at which time the occlusion is removed and the skin is examined. The exposure site then remains un-occluded for at least an additional hours, when the skin is examined a second time. The skin responses at both time points are considered when determining if any reaction has occurred, and whether the reaction is allergic or irritant. In contrast to specific inhalation challenge in the diagnosis of OA, which is complicated and can only be completed for one exposure at a time, the use of patch testing in the diagnosis of contact dermatitis (CD) permits the examination of an individual s reactivity to a large number, often 65 or more, agents at one time [Zug et al. 2009] Prevalence of OCD As in the case of OA, the population burden of OCD is challenging to estimate. Estimates of incidence and prevalence are hampered by under-reporting of disease to practitioners, and a lack of mandatory surveillance schemes in many jurisdictions. Some estimates suggest that occupational skin disease (not just contact dermatitis) may represent up to 30% of all occupational disease worldwide [Diepgen and Coenraads. 1999]. Under the mandatory reporting scheme in Finland (FROD) 643 cases of occupational contact dermatitis (OCD) were reported in 2002 [Riihimäki et al. 2004]. The cases were evenly split between irritant (313) and allergic (330) CD. The 2002 incidence rate for CD in Finland was

30 11 estimated at 27.3 cases per 100,000 workers more than double the rate for OA [Riihimäki et al. 2004]. The Netherlands began a voluntary reporting scheme for occupational skin disease in Under this program, 4516 cases of occupational skin disease were reported from ; 3603 (80%) of cases reported were contact dermatitis (CD) [Pal et al. 2009]. The mean annual incidence rate for CD in the Netherlands was 8.7 cases per 100,000 workers [Pal et al. 2009, United States Department of Labor Bureau of Labor Statistics. 2011]. From the UK surveillance programs (OPRA, THOR) reported an annual average of 2191 cases of CD reported by dermatologists and 1427 cases by occupational physicians [McDonald et al. 2006]. McDonald et al. estimate an annual incidence rate of 74 cases per million for dermatologists, and 510 per million for occupational physicians [McDonald et al. 2006]. A more recent report suggests incidence rates may have decreased; 68 cases per million among dermatologists and 260 cases per million from occupational physicians [Turner et al. 2007]. Based on studies from the 1990s, Diepgen and Coenraads estimate the annual incidence rate in most countries to be around cases per 1000 full-time workers, or approximately cases per 100,000 workers, significantly higher than even the mandatory surveillance scheme in Finland [Diepgen and Coenraads. 1999]. As in the case of WRA we see discrepancies between the reporting schemes with mandatory surveillance programs reporting higher incidence rates than voluntary programs. But compared with WRA, there is much less research focused on estimating the population incidence (or prevalence) of occupational contact dermatitis Causes of OCD The surveillance schemes that provide estimates of population incidence of CD can also provide insight into the most common causal factors. Turner et al. reported the top ten suspected agents causing CD in EPIDERM and OPRA for to be latex, soaps/cleansers, wet work, personal protective equipment (PPE), nickel, preservatives, resin and acrylics, foods, chromes/chromates, and cobalt [Turner et al. 2007]. This list of common causes of CD is relatively unchanged from , except for differences with petroleum products, cutting oils and coolants, solvents and alcohol [McDonald et al. 2006].

31 12 It is also possible to standardize the diagnostic procedure (patch testing) so that data can be compared within and between clinical centers. The North American Contact Dermatitis Group (NACDG) is one consortium that pools patch test results for both occupational and environmental (non-occupational) patients. Among the thirteen contributing members, the ten most common occupational patch test positive allergens in were: nickel, thiuram, carba mix, formaldehyde, quaternium-15, neomycin, cobalt, thimerosol, bacitracin, and balsam of Peru [Rietschel et al. 2002]. A similar group, the European Surveillance System on Contact Allergens (ESSCA) has reported on the common agents in positive patch tests, for all patients, both occupational and nonoccupational across Europe [Uter et al. 2009]. In 2005/2006 the most common contact allergens resulting in positive patch tested included: fragrance mix, nickel sulfate, cobalt chloride, potassium dichromate, colophonium, p-phenylenediamine, and formaldehyde, among others [Uter et al. 2009]. Table 2 summarizes the exposures commonly associated with occupational contact dermatitis, and demonstrates the troubles when comparing different reporting schemes. The UK surveillance schemes for CD (THOR, EPIDERM, OPRA) have grouped causal agents in to chemical groups (likely to ease reporting burden on physicians), while the patch test results from the NACDG are allergen specific. This prevents direct comparison, but still allows for similarities to be identified (i.e., rubber chemicals in THOR, EPIDERM and OPRA and thiuram and carba mix in the NACDG data).

32 13 Table 2 Common Agents Causing Occupational Contact Dermatitis, Both Irritant and Allergic. As Reported by Peer-Reviewed Studies as Cited. NACDG (US) [Rietschel et al. 2002] No Irritants Includes Allergens Carba Mix Cobalt Chloride Epoxy Resin Formaldehyde Glutaraldehyde Glyceryl Thioglycolate Mercaptobenzothiazole Nickel Sulfate Potassium Dichromate Quaternium 15 Thiuram THOR (UK) [Turner et al. 2007] Includes Irritants Chromes/chromates Cobalt Foods Latex materials Nickel PPE Preservative Resins and acrylics Soaps/cleansers Wet work EPIDERM and OPRA (UK) [McDonald et al. 2006] Cutting oils and coolants Foods and flour Nickel Petroleum and products Preservatives Resins and acrylics Rubber chemicals and materials Soaps and cleaners Solvents and alcohols Wet work Epidemiological Evidence: Relationship between Symptoms and Disease The research examining the relationship between symptoms and skin disease is far less thorough than the equivalent body of literature relating to respiratory symptoms and disease. The standardization of questionnaire items is also, arguably, less advanced than in studies of respiratory health. There are a few studies that have specifically addressed the relationships between reported symptoms and doctor-diagnosed skin disease; these studies are summarized in Table 3. Svensson et al. found that asking patients if they had hand eczema had better sensitivity and specificity than asking about specific skin symptoms (fissures, scaling, vesicles, papules, erythema) when compared with the gold standard of dermatologist examination and diagnosis [Svensson et al. 2002]. A study of hand dermatitis in nurses showed high sensitivity and specificity for both a symptom questionnaire-based diagnosis as well as a self-reported diagnosis when compared with the gold standard of physician diagnosis [Smit and Coenraads. 1993].

33 14 In an occupational setting, Meding et al. compared self-reported hand eczema with doctordiagnosed hand eczema among car mechanics, dentists, and office workers and found similar sensitivities (range 53-59%) and specificities (range 96-99%) across the three groups [Meding and Barregard. 2001]. However in another occupational study Carstensen et al. found lower sensitivities (range 22-33%) and specificities (range 76-89%) for both self-reported diagnosis and symptom based diagnosis when compared with doctor diagnosis [Carstensen et al. 2006]. The lack of consistent results in the use of questionnaire-based diagnoses when compared with physician diagnosis may be partly due to the lack of a standardized questionnaire for skin symptoms. Independently, Vermeulen et al. and Carstensen et al. noted that the differences in occupational environments may also require that skin symptom questionnaires be validated in the specific occupational setting prior to large-scale epidemiological use [Carstensen et al. 2006, Vermeulen et al. 2000]. It may also be partly due to the number of terms used to describe skin conditions: eczema, hand dermatitis, atopic dermatitis, skin rash, and others. More information on the study population, as well as more detailed and/or self-explanatory questions may help to clarify the relationship of skin symptom reporting to specific diagnoses. And, as the examination by a dermatologist is more straightforward and less labour intensive than spirometry and specific inhalation challenges, this task should be less daunting than the work already completed in respiratory epidemiology.

34 15 Table 3 Summary of Literature Investigating the Relationship between Symptom Reporting and Skin Disease Diagnoses. Population Ref Gold Standard Outcome Sen Spec Nurses [Smit et al. Doctor-diagnosed Symptom based ] hand dermatitis Self-reported hand dermatitis Metal workers [de Joode et Doctor-diagnosed Symptom based al. 2007] hand dermatitis Picture based questionnaire Clinical Wind turbine production workers Rubber manufacturing Car Mechanics Dentists Office Workers [Svensson et al. 2002] [Carstensen et al. 2006] [Vermeulen et al. 2000] [Meding and Barregard. 2001] Doctor-diagnosed hand eczema Doctor-diagnosed current dermatitis Doctor-diagnosed hand dermatitis Doctor-diagnosed hand eczema Symptom based Patient reported hand eczema Self-reported skin rash Self-reported one or more symptoms, lasting more than 3 weeks Self-reported one or more symptoms Self-reported two or more symptoms, lasting more than 3 weeks Self-reported hand eczema Self-reported hand eczema Self-reported hand eczema Occupational Exposure Over time, the recognition, assessment, and control of occupational exposures has evolved and grown into the science of occupational, or industrial, hygiene. Historically, exposure assessment has focused on airborne exposure rather than skin exposure, and rarely a combination of the two. This may be due in part to the relative complexity of sampling skin exposure when compared with the air samples required for determining potential inhalation exposure.

35 16 It may also be due in part to the reality that exposure limits have focused on airborne exposure; in some cases a biological marker is suggested to monitor systemic burden as a result of uptake from all exposure routes (inhalation, ingestion, skin absorption), but a quantitative skin exposure limit is almost unheard of, and rarely suggested [Bos et al. 1998]. The emphasis on airborne hazards is important because it has resulted in fundamental differences in the approach to prevention of skin and airborne hazards. The American Conference of Governmental Industrial Hygienists (ACGIH) publishes a handbook of exposure limits called threshold limit values (TLVs) [American Conference of Governmental Industrial Hygienists (ACGIH). 2008]. These values have been adopted by many jurisdictions as the regulated occupational exposure limits, including Ontario. In the ACGIH TLVs, quantitative airborne exposure limits are listed in up to three formats per substance: (1) a time weighted average concentration for a conventional 8-hour work day, (2) a short term exposure limit intended as a 15-minute time weighted average, and (3) a ceiling limit which should not be exceeded at any point in work day. In comparison, substances that have relevant skin exposures are given a qualitative skin notation, which denotes a potential significant contribution to the overall exposure by the cutaneous route [American Conference of Governmental Industrial Hygienists (ACGIH). 2008]. The skin notation is not intended to denote chemicals that have only a dermatological effect [Boeniger. 2003]. The qualitative nature of this skin notation and the resulting lack of a measurable exposure limit make it near impossible to declare dermal exposure too high without the use of biological exposure indices (BEIs) which take into account systemic burden integrated over all exposure routes Exposure-Response Relationships The relationship between airborne exposure and respiratory disease has been studied in numerous workplaces for a wide variety of exposures and outcomes ranging from reported symptoms to confirmed clinical diagnoses of respiratory disease [Jaakkola et al. 2009, Heldal et al. 2010, Lillienberg et al. 2010, Smit et al. 2008, Pronk et al. 2007, Jacobs et al. 2008, Cox- Ganser et al. 2009]. Comparatively, studies focusing on skin exposure or on exposure-response studies for skin symptoms and/or disease are rare. This lack of research makes it challenging to identify causal

36 17 links between skin exposure and skin symptoms/disease, and may ultimately hamper efforts to better control skin exposure in the workplace. The few exposure-response studies completed to date suggest that there are observable exposure-response relationships for skin exposure and skin symptoms/disease. In 2009, Sripaiboonkij et al. published two studies that included exposure-response analysis for skin symptoms. The first was a study of glass microfiber production workers, which demonstrated that workers in the factory areas had increased odds (OR 3.89, 95% CI ) of reporting skin symptoms (dryness or flaking of skin, itchy skin, irritation, smarting or redness of skin, sore or tender skin, or urticaria) [Sripaiboonkij et al. 2009b]. When the workers were classified into high and low microfiber exposure groups (airborne exposure), those with high exposures also had greater odds of reporting skin symptoms (OR ) compared with office workers, suggesting a dose-response relationship [Sripaiboonkij et al. 2009b]. The second study by Sripaiboonkij et al. reported on similar relationships, this time among employees in a wood furniture factory [Sripaiboonkij et al. 2009a]. There were no associations between factory work (vs. office work) and skin symptoms or between a high/low categorization of airborne wood dust levels and skin symptoms. But, workers who reported exposures to chemicals on a questionnaire had increased odds of reporting skin symptoms (OR 2.53, 95% CI 1.24 to 5.17); it is unclear whether the questionnaire asked about route of exposure to the chemicals [Sripaiboonkij et al. 2009a]. An earlier study (2007) by Van Wendel do Joode et al. investigated the association between skin symptoms and exposure to semi-synthetic metalworking fluids [de Joode et al. 2007]. Skin exposure was measured in two ways: first by a semi-quantitative dermal exposure assessment tool, DREAM [Van-Wendel-de-Joode et al. 2003], and second using a quantitative tracer method, VITAE [Fenske et al. 1986a, Fenske et al. 1986b]. Skin outcomes were also measured in two ways: first a standard symptoms questionnaire and second, a picture based screening list. In separate models, exposure (yes/no) and workers with high exposure (none/low/high) had increased prevalence ratios (PR) for reporting skin symptoms on their hands, forearms or face (PR range ) [de Joode et al. 2007].

37 18 Chapter 5 of this thesis further contributes to the developing body of knowledge on exposureresponse relationships for skin symptoms in two occupational populations: bakery workers and auto body shop workers. 1.5 Connecting the Skin and Respiratory Systems Occupational Exposure Exposure Recognition Together, Table 1 and Table 2 provide a summary of workplace exposures that may cause occupational asthma and occupational contact dermatitis, respectively. There is some evident overlap between these two lists (i.e., metals causing OA and cobalt/nickel/chromate causing OCD) but a review of the overlap among these causal agents has never been formally completed. Chapter 3 of this thesis will identify common occupational contact allergens in recent Canadian data, and investigate whether these known occupational contact allergens are also capable of causing occupational asthma Exposure Assessment The measurement of airborne exposure is well established. Air samples are collected, usually by drawing a known amount of air through a filter and weighing or analyzing the filter for the contaminant of interest. The pathway from airborne contamination inhaled into the respiratory system is reasonably straightforward and well understood. The contaminants enter the upper respiratory system, deposit at various depths (depending on particle aerodynamic diameter) within the airways, and the smallest particles, gases and vapours deepest into the lung. The pathway from environmental contaminant onto, into, and through the skin is seen to be more complicated. The exposure agent can be deposited directly onto the skin during work tasks, picked up by the skin when touching surfaces (settled airborne exposure), deposited from the airborne compartment directly on to the skin or onto surfaces, or deposited onto a clothing layer covering the skin [Schneider et al. 1999]. From here, the contamination can be transferred between the compartments, ultimately reaching the skin surface where it may or may not penetrate depending on the chemical composition of the exposure and the integrity of the skin barrier. Much of the recent work focusing on dermal exposure and dermal exposure assessment

38 19 has stemmed from the conceptual model published in 1999 by Schneider et al. [Schneider et al. 1999]. In this model, Schneider et al. acknowledge the interaction between the source of contamination, the air component and surface contamination. Additionally, the model acknowledges the complexity that exists due to the exchange of contaminant between the outer clothing layer, the inner clothing layer and the skin itself. All of these layers can receive contaminant from either the airborne component or the surface contaminant layer, and can also exchange contaminant between each other. Given the interconnectedness of the surface and air compartments in the Schneider model, it is likely that when the contribution from the airborne contaminant to the surface contamination layer is high, that skin and airborne exposures in the workplace may be correlated. Previous studies have investigated this question. In 2004, McClean et al. studied highway construction workers exposure to pyrene and polycyclic aromatic compounds (PACs) and found that among paver operators, screedmen, and roller operators skin and airborne exposures to both pyrene (r = 0.58, p = 0.04) and PACs (r = 0.45, p = 0.01) were significantly correlated [McClean et al. 2004]. Burstyn et al. found similar, though not statistically significant, results for correlation between bitumen fume among a small group (n = 7) of Dutch paving workers (r = 0.71, p = 0.08) [Burstyn et al. 2002]. Studies of both skin and airborne exposure in spray painters have measured exposure to xylene, ethyl benzene, 1,6-hexamethylene diisocyanate (HDI) monomers, and isocyanurate. Chang et al. found significant correlation between full shift skin and airborne exposure to both xylene and ethylbenzene [Chang et al. 2007a, Chang et al. 2007b]. Fent et al. collected task based samples for both HDI monomers and isocyanurate and also observed significant correlation between skin and airborne exposure in both cases (HDI r = 0.79, p < ; isocyanurate r = 0.71, p = <0.0001) [Fent et al. 2008]. A similar correlation relationship (r = 0.98) for methylene bisphenyl diisocyanate (MDI) exposure was also observed iron foundry workers [Liljelind et al. 2010]. The correlation between skin and airborne exposure to nickel in foundry workers varied depending on the anatomical site of skin exposure assessment, but ranged from 0.46 to 0.57 (all p-values <0.001) [Hughson et al. 2010]. Though the actual Pearson coefficient was not given,

39 20 there was a significant association between skin and airborne exposure to fentanyl reported in a recent study of pharmaceutical workers [Van Nimmen et al. 2006]. In studies of pesticide exposure, the results have not been consistent. Two studies, Flack et al. studying propiconazole exposure in workers applying the pesticide to peach crops, and Aprea et al. studying imidacloprid exposure in greenhouse workers, both found no correlation between measured skin and airborne exposures [Flack et al. 2008, Aprea et al. 2009]. However, Tsakirakis et al. recently reported skin and airborne exposures that suggest a high level of correlation between skin and airborne exposure to malathion in pesticide applicators [Tsakirakis et al. 2011]. Aprea et al. suggest that the lack of correlation between skin and airborne exposure in the greenhouse workers may indicate that exposed skin was contaminated by a different mechanism, possible accidental contact of the face with contaminated hands, clothes or surfaces and not through deposition of airborne contamination onto exposed skin [Aprea et al. 2009]. It would appear that skin and airborne exposure are correlated in many, but not all exposure scenarios. When the airborne contaminant portion is contributing highly to the skin exposure, the two exposure measurements are likely to be correlated. This may be especially relevant with low vapour pressure (low volatility) substances, but is likely to be dynamic in all scenarios, dependent on exposure, tasks performed, and personal protective equipment utilized. The mechanisms of airborne and skin exposure are complex. Exposures can be correlated and exposure can occur in one system or both (simultaneously or concurrently). The results of studies reporting on the correlation between skin and airborne exposure are summarized in Table 4.

40 21 Table 4 Correlation Coefficients for the Association Between Skin and Airborne Exposures in Various Occupational Studies. NR = Not reported, *=Pearson Correlation Coefficient Was Calculated Based On Published Exposure Data Occupation Exposure Ref. r p Paving Workers Bitumen [Burstyn et al. 2002] Highway Construction - Rakers Polycyclic Aromatic Compounds (PACs) [McClean et al. 2004] Highway Construction Non-Rakers Pyrene Polycyclic Aromatic Compounds (PACs) Pyrene Iron Foundry Workers MDI [Liljelind et al. 2010] 0.69* Nickel Refinery Nickel [Hughson et al. 2010] <0.001 Workers 0.57 Ship Spray Painters Xylene [Chang et al. 2007b] 0.64* Ship Spray Painters Auto Spray Painters Xylene [Chang et al. 2007a] NR <0.001 Ethylbenzene NR <0.001 HDI monomer [Fent et al. 2008] 0.79 < Isocyanurate 0.71 < Farm Workers Propiconazole [Flack et al. 2008] Pesticide Applicators Malathion [Tsakirakis et al. 2011] 0.82* Greenhouse Workers Imidacloprid [Aprea et al. 2009] NR NS Pharmaceutical Fentanyl [Van Nimmen et al. NR Production 2006] Mechanisms of Effect Traditionally the clinical and research communities have investigated disease in silos; respirologists diagnose and treat asthma, while dermatologists diagnose and treat dermatitis. The outcome of this isolated approach is that thorough clinical research into the mechanisms of disease in each system occurs individually, and investigations into occupational exposures and pre-clinical outcomes (i.e., self-reported symptoms) address either respiratory illness or skin illness; rarely, if ever, are both examined together.

41 22 As outlined in sections 1.2 and 1.3, the diagnostic categories within asthma and contact dermatitis are similar; in both cases allergic and irritant forms of the diseases are recognized. In contact dermatitis the irritant form is more prevalent, whereas in the case of asthma the allergic form is considered more prevalent [English. 2004, Maestrelli et al. 2009]. The mechanisms of immune response for occupational contact dermatitis, and for most cases of occupational asthma, are known and are accepted to be different. In the case of allergic asthma the mechanism is usually a Type I (IgE) hypersensitivity response, while in allergic contact dermatitis it is a Type IV (cell-mediated) delayed hypersensitivity reaction Concurrent Skin and Respiratory Disease in Individuals Case reports in the literature indicate that workers do present with concurrent contact dermatitis and asthma that are both work-related in a variety of occupations including manufacturing, construction, and animal work [Moulin et al. 2009, Valks et al. 2003, De Raeve et al. 1998, Kanerva et al. 1995, Estlander et al. 1993, Paggiaro et al. 1979]. Table 5 summarizes case reports of concurrent asthma and contact dermatitis, stratified by cases where results of specific inhalation challenge (SIC) testing and patch testing were reported, and those where one, or both, of SIC or PT results were not reported.

42 23 Table 5 Exposures Reported in Published Case Studies to Cause Both Occupational Asthma and Occupational Contact Dermatitis. Limited to Case Reports Where OA was Diagnosed Using Specific Inhalation Challenge (SIC) and Occupational Contact Dermatitis (OCD) was Diagnosed Using Patch Testing (PT). Exposure Occupation Ref. Diagnosed OA (SIC) and OCD (Patch Test): 2-hydroxyethyl methacrylate Beautician [Moulin et al. 2009] (HEMA) Diglycidyl Ether of Bisphenol Resin applier [Moulin et al. 2009] A (DGEBA) Diphenylmethane-4,4 - Manufacturing (Automotive [Valks et al. 2003] diisocyanate (MDI) Industry) Potassium Dichromate Cement Floorer [De Raeve et al. 1998] Aziridine Hardener Painter and varnisher [Kanerva et al. 1995] Onion Homemaker [Valdivieso et al. 1994] Nickel Manual grinding of metal [Estlander et al. 1993] castings Spiramycin Poultry breeder [Paggiaro et al. 1979] Diagnosed OA and OCD (No SIC and/or No Patch Test): Nematode (Anisakis simplex) Fish processing [Barbuzza et al. 2009] Limolene Labourer [Guarneri et al. 2008] Peptide Coupling Reagents Laboratory workers [Vandenplas et al. 2008] Ortho-phthalaldehyde Nurse [Fujita et al. 2006] Sapele Wood Carpenter [Alvarez-Cuesta et al. 2004] Ammonium Persulfate Hairdresser [Krautheim et al. 2004] Compositae Florist [Uter et al. 2001] Leek Agricultural worker [Cadot et al. 2001] Aziridine Cross Linker Spray painter [Leffler and Milton. 1999] Sodium Metabisulfite Photographic technician [Jacobs and Rycroft. 1995] Green Bean Homemaker [Igea et al. 1994] Diglycidyl Ether of Bisphenol A (DGEBA) Insulation manufacturing [Kanerva et al. 1991]

43 Epidemiological Evidence Linking Skin and Respiratory Outcomes Despite the case reports of co-occurring skin and respiratory disease (Table 5) attributed to occupational exposures, there is still very limited research in working populations regarding the co-existence of these outcomes. As mentioned previously, there are many studies of respiratory symptoms in working populations [Jaakkola et al. 2009, Heldal et al. 2010, Lillienberg et al. 2010, Smit et al. 2008, Pronk et al. 2007, Jacobs et al. 2008, Cox-Ganser et al. 2009] but a comparatively small body of literature on dermal symptoms [de Joode et al. 2007, Sripaiboonkij et al. 2009a, Sripaiboonkij et al. 2009b]. There are several population level studies that have collected and reported information on both skin and lung symptoms individually [Sripaiboonkij et al. 2009b, Sripaiboonkij et al. 2009a, Lindgren et al. 2002, Fantuzzi et al. 2010, Nettis et al. 2002, Kujala and Reijula. 1995, Holter et al. 2002, Friis et al. 1999, Holness and Nethercott. 1989, Holness et al. 1989, Nethercott and Holness. 1988, Huusom et al. 2011]. However, there are surprisingly few published studies that have reported on co-existing skin and respiratory symptoms. A clinical study by Moulin et al. reported on 234 patients with diagnosed WR contact dermatitis (CD) in an occupational health clinic. When the CD patients were asked about both their respiratory symptoms only 10 (4%) reported work-related respiratory symptoms [Moulin et al. 2009]. Lynde et al. focused on skin and respiratory symptoms in a group of professional cleaners [Lynde et al. 2009]. Of the 549 cleaners, 460 (84%) were male. Among male cleaners, 43 (9.3%) had a current rash and 86 (18.6%) reported having a rash in the last 12 months. In total, 33 (7.2%) male cleaners had a current rash and also reported three or more respiratory symptoms; 17 (3.7%) had skin rash plus three or more work-related respiratory symptoms. In logistic regression models, those cleaners with a current rash and those reporting a rash in the last twelve months were at significantly increased odds of reporting work-related respiratory symptoms compared with cleaners who did not have a rash [Lynde et al. 2009]. These two studies were the only studies that commented on the prevalence in concurrent skin and respiratory symptoms in populations. Though Lynde et al. reported on a working population and Moulin et al. on a clinical population, the prevalence of concurrent symptoms in both groups was similar. This thesis will further contribute to the literature on concurrent skin and respiratory symptoms. Data from historical occupational studies in a variety of workplaces will be pooled

44 25 and analyzed to determine if there is a group of workers who report concurrent symptoms (Chapter 4). Additionally, this thesis will consider how workers (or patients) with concurrent skin and respiratory symptoms may be different from those with only one symptom, or those who are asymptomatic. This thesis will be the first study to investigate the differences in personal, environmental and occupational factors between subjects with concurrent skin and respiratory symptoms and those with symptoms only in one system in a clinical study population (Chapter 6) Possible Cross-System Sensitization Animal Studies As previously stated, the mechanisms of immunologic (allergic) response differ between the skin and the respiratory systems. Allergic contact dermatitis occurs through a Type IV T-cell mediated response whereas allergic asthma most often through a Type I IgE mediated response. Despite these differences in mechanisms of action, there is an increasing interest in examining whether each route of exposure is capable of inducing sensitization and immune response(s) in the other system. It has been demonstrated in animal models of occupational asthma that skin sensitization may be relevant to respiratory outcomes in the case of exposure to some chemical agents (Table 6). These experiments generally follow structured sensitization (induction) and elicitation (challenge) exposure protocols. The skin of naive experimental animals is exposed to the agent in order to develop sensitization, and upon subsequent (first) inhalation challenge to the same agent, an asthma-like response is measured in the study animals. This skin induction and inhalation challenge experimental system has been demonstrated for various exposures relevant to occupational asthma including: TDI [Tarkowski et al. 2007, Vanoirbeek et al. 2004], MDI [Pauluhn. 2008, Rattray et al. 1994], TMA [Vanoirbeek et al. 2006, Arts et al. 2004, Zhang et al. 2004, Arts et al. 1998], and latex [Woolhiser et al. 1999, Woolhiser et al. 2000].

45 26 Table 6 summarizes the published studies where animal models were used to determine whether topical skin exposure can cause sensitization that results in an asthma-like response on first respiratory challenge (intratracheal, intranasal or inhalation). Each study reported in Table 6 measured a ventilatory response to determine whether the experimental animals were having a physiological asthma-like ventilatory response; Table 6 does not summarize experiments that included only cellular outcomes. The experimental animal studies reported in Table 6 involve the use of genetically identical animals in controlled exposure scenarios. Despite the high degree of control, the different studies still follow slightly different protocols. There are between-study differences in the location of the dermal exposure (e.g., flank vs. ear), the preparation of the exposure site (e.g., hair removal methods), the number of exposures in the sensitization period, and the delay between sensitization and challenge. This lack of standardization in the animal studies has been noted by other researchers who have attempted to summarize a large body of evidence [Arts et al. 2006]. The lack of standardization will be amplified when moving from animal studies into humans, as there will be an exponential increase in variability. Human populations are genetically diverse, with large variability in both exposure (within and between individuals) and physiological response (symptoms and diagnosed disease).

46 27 Table 6 Summary of Experimental Animal Studies Demonstrating Skin Exposure Resulting in Sensitization and an Asthma-like Response on First Inhalation Challenge. Reference Exposure Species Sensitization Challenge day 21: inhalation challenge + [Kuper et al. 2011] Oxazolone Norway rats day 0, 7: topical application to flank OR day 0, 4, 11, 14: topical application alternating flank and ear [Tarkowski et al. 2007] Toluene Balb/c mice day 1, 7: dermal application diisocyanate (TDI) on ear [Vanoirbeek et al. TDI Balb/c mice day 1, 2, 3, 7: dermal 2004] application on ear [Pauluhn. 2008] Methylene diphenyl diisocyanate (MDI) Norway rats day 0, 7: topical application to flank [Rattray et al. 1994] MDI Guinea Pigs day 1: topical, intradermal or inhalation application [Vanoirbeek et al. 2006] Trimellitic anhydride (TMA) Balb/c mice day 1, 7: dermal application on ear [Arts et al. 2004] TMA Norway rats day 0, 7: topical application to flank [Zhang et al. 2004] TMA Norway rats day 0, 7, 14, 21: topical application on dorsum [Arts et al. 1998] TMA Norway rats day 0, 7: dermal application on flank [Franko et al. 2011] Furfuryl Alcohol Balb/c mice Day 1-4: dermal application on ear, increasing dose day 10: intranasal challenge + day 10: intranasal challenge + day 20, 35, 50, 65: inhalation challenge day 21: inhalation challenge + day 10: intranasal challenge + day 21: inhalation challenge + day 28: challenge with clean air; day 35, 42: intranasal challenge day 21 or 22: respirator challenge + Day 5, 9, 13, 17: pharyngeal aspiration Asthma response + + +

47 28 Asthma Reference Exposure Species Sensitization Challenge response [Woolhiser et al. 1999, Latex Balb/c mice 5 days a week, for 7 weeks: day 28: inhalation challenge + Woolhiser et al. 2000] topical application to dorsum or lumbar region + [Klink and Meade. 2003] [Lastbom et al. 1998, Lastbom et al. 2000] 3-amino-5- mercapto-1,2,4- triazole (AMT) Balb/c mice 35 consecutive days: topical application on dorsum day 28, 35: intratracheal challenge 3-carene Guinea Pigs four dermal exposures day 21-28: perfused lung challenge [Arts et al. 2004] TMA Wistar rats day 0, 7: topical application to flank [Arts et al. 1998] TMA Wistar rats day 0, 7: dermal application on flank [van Triel et al. 2010] Dinitrochlorobenzene (DNCB) Wistar rats day 0: topical application on flank; day 7: topical application on ear [Kuper et al. 2008] DNCB Norway rats day 0: topical application on flank; day 7: topical application on ear [Vanoirbeek et al. 2006] DNCB Balb/c mice day 1, 7: dermal application on ear [Arts et al. 1998] DNCB Norway rats day 0, 7: dermal application on flank [Arts et al. 1998] DNCB Wistar rats day 0, 7: dermal application on flank [Lee et al. 1984] Formaldehyde Guinea Pigs day 0, 2: topical application to dorsum day 21: inhalation challenge - day 21 or 22: respirator challenge - day 21, 24, 27, 30, 34, 37, 41, 43: inhalation challenge day 21: inhalation challenge - day 10: intranasal challenge - day 21 or 22: respirator challenge - day 21 or 22: respirator challenge - day 7, 22, 29: inhalation challenge + - -

48 Human Studies & Epidemiological Evidence Perhaps due to the challenges of large exposure and response variability in humans, but also due to ethical considerations, there is very little experimental evidence of skin sensitization leading to respiratory disease in humans. De Zotti reported on a latex challenge test in a health care worker where the wearing of a latex glove on one hand was followed by an immediate decrease in peak flow and FEV 1 as well as welts and flares on the head and neck [De Zotti et al. 1992]. In a population study Kujala et al. reported that among 534 Finnish health care workers, wearing gloves for two hours per day was associated with having at least one respiratory disorder (p<0.001) [Kujala and Reijula. 1995]. In the case of health care workers, there was clearly a skin exposure to the gloves, but there may have also been an airborne exposure to the glove powder, containing latex. However, no quantitative exposure measurements were collected as part of this study. Two human studies have tested experimentally whether inhalation challenges can elicit a skin response. Tupker et al. exposed twenty subjects with confirmed atopic dermatitis (AD) to airborne house dust mite (HDM) allergen to determine whether this exposure could exacerbate existing dermatitis or induce new dermatitis lesions [Tupker et al. 1996]. In nine subjects (45%) a skin reaction was observed; in eight patients the skin symptoms were preceded by a reaction in the respiratory system (decreased FEV 1 ). Isaksson et al. attempted to reactivate positive patch test responses to budesonide with an inhalation challenge (budesonide) six weeks after the initial positive patch test [Isaksson and Bruze. 2002]. Fifteen subjects with a positive patch test to budesonide were enrolled. Seven subjects were given an inhalation challenge with budesonide, eight received a control inhalation challenge. Of the seven who received the experimental challenge, four exhibited a reactivation of the positive patch test response. In working populations there is less work examining the role of skin exposure in the development of occupational asthma. In 1993, Nemery and Laenerts commented on the use of methylene diphenyl isocyanate (MDI) use in coal mines and the potential for skin exposure to cause respiratory complaints [Nemery and Lenaerts. 1993]. More recently, Petsonk et al. completed a longitudinal study of workers in a wood products plant that had exposure to MDI. In these workers, those who reported skin staining or clothing staining (a possible marker of skin

49 30 exposure) were significantly more likely to report new onset asthma symptoms at the follow-up visits (after reporting no asthma symptoms at the initial visit) [Petsonk et al. 2000] Cross-System Interaction in Other Disease Models The concept of cross-system interaction, particularly between the skin and the lungs is not limited to occupational asthma. There are other examples of cross-system interaction in human disease. The experiment by Tupker et al. was previously mentioned. In this case, subjects allergic to dust mites were exposed to house dust mite allergen by inhalation, and skin responses were observed in 45% of the subjects [Tupker et al. 1996]. In the case of chronic beryllium disease (a respiratory disease) there is evidence that the skin is an important route of exposure for sensitization and disease progression [Tinkle et al. 2003, Day et al. 2007, Day et al. 2006]. Additionally, respiratory symptoms have been observed in cases of food allergy in both children and adults [Chiang et al. 2010, Bjorksten. 1996]. Cullinan et al. reported increased respiratory symptoms in sensitized subjects after a blind food challenge test (bread containing enzymes), compared with a control test (bread without enzymes) [Cullinan et al. 1997] Additional Considerations The Skin as a Barrier When thinking about the role of skin exposure in the development of sensitization and potentially the development of subsequent respiratory disease, the role of the skin barrier must be considered. It is possible that damage to the skin either through physical, chemical, or biological processes may hamper the skin s ability to block penetration of allergens. Nielsen chemically damaged human skin samples with known concentrations of sodium laurel sulfate (SLS), a known skin irritant. The penetration of five pesticides was tested on a series of skin samples with varying levels of SLS induced damage. Results suggested the penetration rate, total penetration, and lag-time were all greater in the damaged skin samples compared with control (undamaged) skin samples [Nielsen. 2005]. The ability of the SLS treatment to disrupt the barrier function was confirmed by measuring the penetration of tritiated water across SLS treated skin samples. Skin treated with higher concentrations of SLS had greater overall penetration and a greater rate of penetration compared with undamaged skin [Nielsen. 2005].

50 31 An in vivo study of penetration in human skin demonstrated that the normal skin barrier can be disrupted using physical (tape stripping) and chemical (sodium laurel sulfate) treatment, and also showed that the penetration of salicylic acid is greater in skin with the most severely disrupted barrier [Benfeldt et al. 1999]. Jakasa et al. studied volunteer subjects with atopic dermatitis (AD) and healthy skin (controls) to test whether the skin penetration of sodium laurel sulfate differed between the two groups [Jakasa et al. 2006]. Results suggested that the penetration was greater among subjects with active AD compared with healthy controls, but there were no differences between subjects with inactive AD and the healthy controls [Jakasa et al. 2006]. In an epidemiological study of dermal and airborne exposure to potentially carcinogenic substances in rubber workers, Vermeulen et al. collected information on the skin condition of the exposed workers [Vermeulen et al. 2003]. A dermatologist assessed the skin of each subjects and recorded the presence of active hand dermatitis, minor dermatitis, and skin injuries. The results from Vermeulen et al. suggested that workers who were observed to have minor skin aberrations (mild dermatitis) had higher levels of urine mutagenicity (measured using a Salmonella typhimurium assay) than subjects with normal skin, implying that the subjects with skin disease had great uptake of occupational exposure [Vermeulen et al. 2003]. This result suggests a role for barrier function in modifying the uptake of dermal exposure. Similarly, Hino et al. studied toluene and xylene exposure in auto body painters. They found no correlation between airborne exposure levels and urinary biomarkers for either toluene or xylene [Hino et al. 2008]. However, they did find a strong and significant correlation between the urinary biomarkers of toluene and xylene exposure and poor skin condition (toluene: r = 0.61, p < ; xylene: r = 0.34, p = 0.004) as measured by a dermatologist on a skin severity index [Hino et al. 2008]. A clinical study by Bremmer et al. examined 491 patients with atopic dermatitis (AD) and ichthyosis vulgaris (IV), those with higher severity of IV were more likely to report asthma, even after adjustment for AD severity, age, sex, and season of symptom reporting [Bremmer et al. 2008]. Bremmer suggests that the presence of severe IV could be used as a marker for patients who are more likely to develop allergic respiratory disease. These findings lend more support to

51 32 the hypothesis that skin barrier dysfunction influences the degree of allergic sensitization, though this mechanism remains unclear [Bremmer et al. 2008]. Together, these studies suggest that a compromised skin barrier is a modifying factor in the uptake of skin exposure into the human body. These results also suggest that the barrier function is not an innate, static characteristic within an individual. The application of physical stress (tape stripping) or chemical stress (SLS) can induce a change in barrier function. It is possible that some workplace exposures will act to disrupt the skin barrier, increasing the permeability of workers skin and, in turn, the uptake of workplace exposures through the skin. 1.6 Framework A simple framework for conceptualizing the possible connection between skin and respiratory outcomes in terms of both exposure and response is proposed in Figure 2. Figure 2 uses the common exposure to health effects pathway and incorporates the exposures, intermediate health effects, and specific outcomes of interest for this thesis. In four separate studies, this thesis will further investigate three of the relationships outlined in Figure 2. The results from this thesis will contribute to knowledge of each of these relationships individually and more broadly to the understanding of the framework as a whole. The three specific relationships that this thesis seeks to investigate are: Relationship between airborne exposure and skin exposure, Relationship between skin symptoms and respiratory symptoms, Relationship between exposure (airborne) and skin symptoms.

52 33 Figure 2 A Proposed Framework for Conceptualizing the Connections Between Skin and Respiratory Symptoms and Disease. With Contributions from V.H. Arrandale, J. Cherrie and D. Heederik.

53 34 Chapter 2 Research Aims and Hypotheses 2.1 Knowledge Gaps Occupational contact dermatitis and occupational asthma are common problems in the workplace. Workers can develop new disease, or aggravate existing disease, as a result of exposures at work. Both diseases have an allergic and an irritant form, and both are associated with symptoms reported by the workers. Though we have a good understanding of exposure response relationships in relation to respiratory symptoms and occupational asthma, we do not understand exposure-response relationships for skin symptoms and skin disease to the same extent. We also do not have a good understanding of which exposures may be risks to both the skin and respiratory symptoms, nor have we evaluated our current systems of identifying potential skin and/or respiratory allergens in the workplace. Additionally, there is evidence that some workers experience symptoms in both systems, and these symptoms may be a result of workplace exposures. There is mounting evidence that skin exposure may lead to sensitization that is relevant to the development of respiratory disease. Neither of these two areas, concurrent symptoms and cross-system sensitization, have been explored thoroughly in occupational studies. 2.2 Research Aims The overarching purpose of this thesis is to further investigate the relationships between occupational exposures, skin symptoms and disease, and respiratory symptoms and disease. There is particular focus on the interaction between the two systems in terms of co-existing exposures and concurrent symptoms Specific Research Aims The specific research aims of this thesis are as follows, arranged by thesis chapter: Chapter 3: Occupational Contact Allergens: Are They Also Associated With Occupational Asthma?

54 35 Aim 1. To identify the most common occupational contact allergens in the available Canadian patch test data. Aim 2. To identify which of the most common occupational contact allergens in Canadian data are also associated with causing occupational asthma. Chapter 4: Co-existing Skin and Respiratory Symptoms in Four Occupational Groups Aim 3. To determine whether workers report symptoms in both their skin and respiratory systems. Chapter 5: Skin Symptoms in Bakery and Auto Body Shop Workers: Associations With Exposure and Respiratory Symptoms Aim 4. To determine whether exposure-response relationships can be observed for skin symptoms in two separate occupational groups: bakery workers and auto body shop workers. Chapter 6: Predictors of Concurrent Skin and Respiratory Symptoms among Workers with Suspected Work-Related Skin or Respiratory Disease Aim 5. To determine how many patients with suspected work-related skin or respiratory disease report symptoms in both systems, and whether or not these symptoms are work-related. Aim 6. To identify predictors of reporting both skin and respiratory symptoms among a clinical population with suspected work-related skin or respiratory disease. 2.3 Hypotheses A. There is significant overlap in agents/exposures that cause occupational asthma and occupational contact dermatitis. Previous research has studied the causes of occupational contact dermatitis and occupational asthma separately. Results from these studies show that there are similarities in the type of exposure that cause these two outcomes, however differences in the way these exposures are identified hamper the comparison. This thesis will identify recent common occupational contact

55 36 allergens in Canadian patch test data and systematically review the literature to determine whether they are also capable of causing occupational asthma. B. Exposure-response relationships exist for occupational exposures and skin symptoms in a similar manner to those observed for occupational exposures and respiratory symptoms. In general, the exposure-response relationships for respiratory symptoms are studied more often, and are better understood, than exposure-response relationships for skin symptoms. Exposureresponse relationships for respiratory symptoms have been previously described in two working populations, bakery workers and auto body shop workers. The data describing skin symptoms in these two groups has not been examined. This thesis will examine whether exposure-response relationships for skin symptoms exists in these two working populations. C. Some workers are experiencing and will report both skin and respiratory symptoms. Very few studies have examined the prevalence of concurrent skin and respiratory symptoms and disease. We know that skin and airborne exposures are likely to occur in the workplace and that these exposures can be correlated. Despite this, it is not know whether concurrent symptoms are rare or common, or whether they occur in some occupations or all occupations. This thesis will pool data from historical occupational studies and analyze the data to determine what portion of workers (form a variety of workplaces) report concurrent skin and respiratory symptoms. This thesis will also measure the prevalence of concurrent skin and respiratory symptoms in a clinical population of patients with suspected work-related disease. D. Patients with concurrent skin and respiratory symptoms will differ from patients with symptoms in only one system based on exposure, occupation and/or personal protective equipment use. It is suspected that a clinical population with suspected work-related disease (either skin or respiratory) will have higher symptom prevalence than healthy working populations. The increased prevalence in the clinical population will permit for analyses to identify predictors of concurrent symptoms. This analysis will allow for the identification of occupations and workplaces that are at higher risk of concurrent skin and respiratory symptoms. This thesis will

56 37 explore whether patients with concurrent symptoms differ from patients with symptom(s) only in one system based on occupation, exposure and personal protective equipment use.

57 38 Chapter 3 Occupational Contact Allergens: Are They Also Associated With Occupational Asthma? Victoria H. Arrandale 1 ; Gary M. Liss 1,2 ; Susan M. Tarlo 1,3,6 ; Melanie Pratt 4 ; Denis Sasseville 5 ; Irena Kudla 6 ; D. Linn Holness 1, 6 1 University of Toronto, Toronto, Canada 2 Ontario Ministry of Labour, Toronto, Canada 3 Toronto Western Hospital, Toronto, Canada 4 University of Ottawa, Ottawa, Canada 5 McGill University, Montreal, Canada 6 St. Michaels Hospital, Toronto, ON This manuscript was published in the American Journal of Industrial Medicine (Am J Ind Med Apr; 55 (4):353-60). Copyright permission has been granted (see Copyright Acknowledgements).

58 Abstract Background: Workplace exposures that can potentially cause both allergic occupational contact dermatitis (AOCD) and occupational asthma (OA) are not clearly identified. Methods: Occupational contact allergens (OCAs) were identified using North American Contact Dermatitis Group (NACDG) data. Reference documents and systematic reviews were used to determine whether each OCA had been reported to potentially cause OA. The presence or absence of a sensitizer notation in occupational hygiene reference documents was also examined. Results: The 10 most common OCAs were: epoxy resin systems*, thiuram, carba mix, nickel sulfate*, cobalt chloride*, potassium dichromate*, glyceryl thioglycolate, p-phenylenediamine*, formaldehyde* and glutaraldehyde*. Seven (indicated by *) were determined to be possible causes of OA. Information on sensitizing potential from OH reference materials contained conflicting information. Conclusions: Several common OCAs can also potentially cause OA. Inhalation and dermal exposures to these agents should be controlled and both OA and AOCD should be considered as possible health outcomes. Increased consistency in sensitizer notations is needed.

59 Introduction Traditionally, research related to occupational contact dermatitis (OCD) and occupational asthma (OA) has been done in separate organ system silos, with the work focusing on either lung disease or skin disease, but rarely the two together. In the workplace, workers are exposed to chemicals that may cause both irritant and allergic effects, in the skin and lung, by both dermal and inhalation routes of exposure. Agents causing responses in both systems are not well documented, and are not always recognized in the occupational or clinical settings. In addition, there is an emerging body of research that examines whether each route (dermal and inhalation) of exposure is capable of inducing sensitization and response in the other system. It has been shown in animal models that dermal exposure can cause sensitization that upon first inhalation exposure (in a naïve animal) results in an asthma-like response [Vanoirbeek et al. 2004, Zhang et al. 2004, Klink and Meade. 2003, Herrick et al. 2002, Lehto et al. 2005]. There are also case reports in the literature of workers having both AOCD and OA in response to the same chemical [Moulin et al. 2009, De Raeve et al. 1998, Guarneri et al. 2008]. The evidence from animal models and human case reports shows that there is a need to better understand the inter-relationships between inhalation and dermal routes of exposure, and between respiratory and skin responses, to workplace chemicals in order to develop appropriate prevention strategies. Patch testing is an important clinical tool in the diagnosis of allergic contact dermatitis (ACD). The North American Contact Dermatitis Group (NACDG) is a group of dermatologists across the United States and Canada who have agreed to patch test patients to standard sets of allergens according to a standardized protocol and pool their data for surveillance and research purposes [Zug et al. 2009]. Members are dermatologists trained in patch testing and experienced in the diagnosis of occupational skin disease; two of the thirteen members of the NACDG are located in Canada. The group uses standardized methods for patch testing and interpretation of results and regularly reviews the results of the pooled patch test results from their clinics The NACDG database offers an opportunity to examine contact allergens in a large clinical population. There is no equivalent database containing data on airway sensitizers.

60 41 The aim of this study was to first identify the most common OCAs in the Canadian portion of the NACDG data, and to determine whether these agents were also airway sensitizers using the peerreviewed occupational asthma literature. The secondary aim was to identify whether or not the common OCAs were noted as skin or respiratory sensitizers in common occupational hygiene (OH) reference documents. 3.3 Methods De-identified patient data were obtained for the two Canadian NACDG sites for the period January 1, 2001 to December 31, This data represented three 2-year cycles of NACDG data collection. All patients were patch tested with a standard screening series of 65 allergens ( standard tray ); a standardized technique was followed in all cases [Pratt et al. 2004]. Informed consent was obtained from all patients as required by each study center s human research committee. This study was approved by the St. Michael s Research Ethics Board. During each 2-year study period the standard patch test tray included sixty-five (65) allergens; changes to this standard tray are made between cycles. Generally, the allergens removed have low rates of positive response, and the allergens added are emerging as more common allergens. Fifty-eight (58) allergens were common to all three NACDG cycles included in this analysis. None of the fourteen allergens excluded were known, common occupational allergens. For each patient, the physician assigned up to three diagnostic codes, which indicate the diagnosis. The work-relatedness of the diagnosis was also recorded by the physician as a separate variable (yes, no, unsure or not tested). For all patients, their response to each of the individual 58 allergens was recorded as allergic, unknown, negative or not tested. The workrelatedness for each individual allergen (separate from the diagnosis) was recorded as yes, no, unsure or not tested. Additional information on the subjects age, sex and race was included in the data Diagnosis of Occupational Allergic Contact Dermatitis Subjects were considered to have allergic contact dermatitis (ACD) if any of the three diagnostic codes were allergic dermatitis. Subjects were considered to have occupational-related ACD (AOCD) if there was a diagnostic code of allergic dermatitis and if the overall diagnosis was

61 42 coded as occupationally-related ( yes ). Subjects where the overall occupationally-related variable was coded as unsure were not included as cases of AOCD for this analysis Determination of Occupationally-Relevant Positive Patch Test Responses Occupational contact allergens (OCAs) were identified using the response code ( allergic ) and the work-relatedness code ( yes ) assigned to the individual allergens tested. The number of individual positive patch test (PPT) responses that were coded both as occupational (allergen specific) and allergic (allergen specific) were counted for each allergen. As the OCAs were determined based on individual allergen responses, it was possible for a single subject to contribute multiple times to the OCA frequency table. Differences between occupational and non-occupational ACD cases were tested using ANOVA (continuous variables) and chi-square (categorical variables) tests. All analyses were conducted using SAS v9.2 [SAS Institute Inc. 2008] Determination of Whether OCAs May Also Cause OA Whether each of the ten most common OCAs was also a potential cause of work-related asthma was determined using two reference documents that are considered authoritative sources of information among the occupational asthma research community. Where no information was available in the reference sources, the peer-reviewed literature was searched systematically. The first source for determining whether an OCA has been associated with OA was Asthma in the Workplace, edited by Bernstein et al. (2006). If the allergen was listed in Appendix Agents Causing Occupational Asthma with Key References (p.825), the allergen was considered to have been associated with OA. Second, the United Kingdom Health and Safety Executive s 2001 publication Asthmagen? was consulted [UK Health and Safety Executive,. 2001]. Asthmagen? reported on all substances that were suspected of being respiratory sensitizers under the European Union s (EU) 1996 criteria. If the allergen was listed as meeting the criteria it was considered to be associated with OA.

62 43 In cases where a search of the peer-reviewed literature was required, a systematic review strategy was followed. All alternate names for the allergen listed in the manufacturer s documentation (Chemotechnique Diagnostics, Vellinge, Sweden) as well as alternate names listed in the HazMap on-line database [National Institutes of Health. 2009] were used. Searches were completed for each allergen (all alternate names) combined with occupational asthma. Searches were limited to human studies published on or before December 31, The systematic search strategy was completed in three separate databases: Ovid MEDLINE, TOXLINE and EMBASE. Results from the three database searches were combined for each OCA and all articles were reviewed. Based on data gathered from the two reference materials and the systematic review of the peerreviewed literature, each of the ten most common OCAs was classified into one of four categories describing the level of evidence linking the agent with OA: Established, the OCA is listed in either Asthma in the Workplace or Asthmagen as a potential cause of OA; Possible, the OCA is not listed in either Asthma in the Workplace or Asthmagen, but there is some evidence in the peer-reviewed literature (e.g. case reports), OR the evidence in Asthma in the Workplace and Asthmagen is conflicting; Negative, studies have investigated the OCA in OA aetiology and found negative results; No evidence located, no evidence of the role of the OCA in the aetiology of OA was located Determination of Skin Sensitizer Notation Status Three sources of information available to practicing occupational hygienists were reviewed to assess whether the 10 most common OCAs identified were listed as possible sensitizers. The sources reviewed were the American Conference of Governmental Industrial Hygienists (ACGIH) Threshold Limit Value (TLV) handbook [American Conference of Governmental Industrial Hygienists (ACGIH),. 2008], the National Library of Medicine s HazMap Occupational Exposure to Hazardous Chemical database [National Institutes of Health. 2009] and the NIOSH Pocket Guide to Chemical Hazards [National Institute for Occupational Safety and Health. 2007].

63 44 The notations indicating sensitizing potential differed between the OH reference documents. In some cases sensitizing potential was not explicitly noted. The notations that were accepted to indicate the possibility of sensitization resulting from exposure were as follows: NLM HazMap Database: adverse effects Skin and Asthma as well as potential disease outcomes of Asthma and Contact Dermatitis. ACGIH TLV Handbook: specific sensitizer notation, SEN, though this does not differentiate between the systems in which sensitization may occur. NIOSH Pocket Guide: possible symptoms as Respiratory sensitizer, Asthma or Skin Sensitizer. 3.4 Results Data received from the NACDG contained 3676 patch test records for the period January 1, 2001 December 31, In total 1808 (49%) subjects had a diagnosis of allergic contact dermatitis and 619 (17%) subjects had a diagnosis of work-related dermatitis. Overall, 397 subjects (11% of all subjects) met the definition of AOCD. The ACD and overall population were not observed to have any striking differences in the demographic variables (Table 7). There were 354 (10%) subjects who had both an irritant and an allergic diagnosis (data not shown). Among the occupational cases, 20% had both an allergic and an irritant diagnosis (n=120) (data not shown). Among the ACD cases, occupational cases were more likely to be male (p<0.0001) and to be younger (p<0.0001) than the non-occupational cases (Table 7) Common Occupational Contact Allergens The ten most frequent OCAs (PPT reactions coded as occupational and allergic) were: epoxy resin, thiuram, carba mix, nickel sulfate, cobalt chloride, potassium dichromate, glyceryl thioglycolate, p-phenylenediamine (PPD), glutaraldehyde, and formaldehyde (Table 8). In total forty-one allergens were coded as both work-related and allergic in at least one subject.

64 45 Table 7 Basic Descriptive Statistics for the Entire Study Population, Subjects with an Allergic Contact Dermatitis (ACD) Diagnoses and ACD Cases Stratified by Occupational Occupational ACD Reported as n (%) of ACD Cases. P-values Shown for Comparison Between AOCD and Non-Occupational ACD Groups. All Patients ACD AOCD Non-Occup. ACD n (%) n (%) n (%) n (%) p-value Overall 3676 (100%) 1808 (49%) 397 (22%) 1411 (78%) - Physician A 2167 (59%) 1286 (59%) 244 (19%) 1042 (81%) B 1509 (41%) 522 (35%) 153 (29%) 369 (71%) < Relatedness. ACD Frequencies Reported as n (%) of All Patients. AOCD and Non- Work- Related Sex Race Yes 619 (17%) 397 (64%) - - No 2854 (78%) 1321 (46%) - - Unsure 203 (5.5%) 90 (44%) - - Male 1155 (31%) 554 (48%) 200 (36%) 354 (64%) Female 2521 (69%) 1254 (50%) 197 (16%) 1057 (84%) White 3265 (89%) 1601 (49%) 339 (21%) 1262 (79%) Black 88 (2.4%) 45 (51%) 15 (33%) 30 (67%) Asiatic 205 (5.6%) 105 (51%) 27 (26%) 78 (74%) Hispanic 35 (1.0%) 11 (31%) 4 (36%) 7 (64%) - < Other 83 (2.3%) 46 (55%) 12 (26%) 34 (74%) Age (mean, sd) 45.5 (15.9) 45.4 (15.9) 39.3 (11.2) 47.1 (16.6) < NS Table 8 Ten Most Common Occupational Contact Allergens (OCAs). (Frequency That Individual Responses to the Allergen Were Coded Both as Work-Related and Allergic.) # Allergen n 1 Epoxy Resin 49 2 Thiuram 39 3 Carba Mix 36 4 Nickel Sulfate 22 5 Cobalt Chloride 21 6 Potassium Dichromate 20 6 Glyceryl Thioglycolate 20 8 P-phenylenediamine (PPD) 19 8 Formaldehyde Glutaraldehyde 16

65 Occupational Contact Allergens as a Cause of Occupational Asthma Of the ten most frequent OCAs in our patient population, seven were listed in Asthma in the Workplace (all except thiuram, carba mix and glyceryl thioglycolate) and five were listed in the HSE Asthmagen document (some components of epoxy resin systems, nickel sulfate, cobalt chloride, potassium dichromate, and glutaraldehyde) as exposures capable of causing OA (Table 9). Three of the ten most common OCAs in our patient population were not listed in either document: thiuram, carba mix and glyceryl thioglycolate; the systematic review strategy was completed for each of these agents. The systematic search strategy retrieved the following number of citations: carba mix - 9 articles; thiuram mix -19 articles; and glyceryl thioglycolate - 1 article. After review (VA, ST, GL, LH), none of the retrieved articles contained objective diagnostic test results that supported an association between exposure to the OCA and occupational asthma. Table 9 Summary of the Ten Most Frequent Occupational Contact Allergens (OCAs) and the Evidence Linking Each to OA in Asthma in the Workplace and the UK HSE Asthmagen. 1 Components of Epoxy Resin Systems Asthma in Workplace Yes* HSE Asthmagen Yes* 2 Thiuram Not Listed Not Listed 3 Carba Mix Not Listed Not Listed 4 Nickel Sulfate Yes Yes 5 Cobalt Chloride Yes Yes 6 Potassium Dichromate Yes Yes 7 Glyceryl Thioglycolate Not Listed Not Listed 8 P-phenylenediamine (PPD) Yes Watch List 9 Formaldehyde Yes No 10 Glutaraldehyde Yes Yes * Epoxy resin systems contain epoxy monomers and polymers (e.g., bisphenol A, bisphenol F) as well as hardeners (e.g., acid anhydrides and amines). The hardeners are confirmed causes of OA, some epoxy monomers, such as bisphenol A have been associated with OA in a few cases [Moulin et al. 2009, Hannu et al. 2009, Kanerva et al. 2000].

66 47 Table 10 Categorization of Whether Each Common Occupational Contact Allergen (OCA) Has the Potential To Cause OA Based on Reference Sources and Systematic Literature Review, Where Necessary. Asthma Causing Agent? Establ. Possible Negative No Current Evidence 1 Components of Epoxy Resin Systems X* 2 Thiuram X 3 Carba Mix X 4 Nickel Sulfate X 5 Cobalt Chloride X 6 Potassium Dichromate X 7 Glyceryl Thioglycolate X 8 P-phenylenediamine (PPD) X 9 Formaldehyde X 10 Glutaraldehyde X * Epoxy resin systems contain epoxy monomers and polymers (e.g., bisphenol A, bisphenol F) as well as hardeners (e.g., acid anhydrides and amines). The hardeners are confirmed causes of OA, some epoxy monomers, such as bisphenol A have been associated with OA in a few cases [Moulin et al. 2009, Hannu et al. 2009, Kanerva et al. 2000]. Epoxy resin, nickel sulfate, cobalt chloride, potassium dichromate, and glutaraldehyde were classified as having established associations with OA. P-phenylenediamine (PPD) and formaldehyde were classified as possibly associated with OA due to discrepancies between the two reference documents. Thiuram, carba mix and glyceryl thioglycolate were classified as having no current evidence of causing OA. These categorizations are summarized in Table Sensitizer Notations Lastly, the presence or absence of a sensitizer notation, or a reasonable equivalent as described in the Methods, was determined for each of the ten most common OCAs in the three OH reference documents (Table 11). The NLM HazMap data base was the only source to assign all of the OCAs a skin or contact dermatitis notation (indicating sensitizing or irritant effects); seven were assigned an asthma notation (Table 11). All three sources agreed that epoxy resin (or

67 48 components of epoxy resin systems) and glutaraldehyde were potential sensitizers (either skin or respiratory) but for all other OCAs there was disagreement between the sources. Table 11 Summary of Sensitizer Notations for the Ten Most Common OCAs in Common Occupational Hygiene Reference Documents. Components of Epoxy Resin Systems ACGIH TLVs NLM HazMap Health Effects NIOSH Pocket Guide Sensitizer Notation Asthma Skin OR Contact Dermatitis Resp Sensitizer OR Asthma Skin Sensitizer Hardener Yes Yes Yes Yes No Monomer No Yes Yes No No Thiuram Not listed No Yes No No Carba Mix Not listed No Yes Not listed Nickel Sulfate No Yes Yes Yes Yes Cobalt Chloride No Yes Yes Not listed Potassium Dichromate No Yes Yes No Yes Glyceryl Thioglycolate Not listed No Yes Not listed p-phenylenediamine (PPD) No Yes Yes Not listed Formaldehyde Yes Yes Yes No No Glutaraldehyde Yes Yes Yes Yes Yes Not listed = Agent not listed in the reference material. 3.5 Discussion This study generated a list of the ten most common OCAs in the Canadian portion of the NACDG population. In 2002 Rietschel et al published an analysis of the complete NACDG data from and reported that the most common occupational contact allergens in this period were, in descending frequency: thiuram, epoxy resin, carba mix, nickel sulfate, formaldehyde, potassium dichromate, quaternium 15, cobalt chloride, glutaraldehyde, glyceryl thioglycolate and mercaptobenzothiazole [Rietschel et al. 2002]. With the exception of quaternium 15 and mercaptobenzothiazole, the results presented here are strikingly similar to

68 49 those of Rietschel et al. This suggests that the results from the Canadian data are at least somewhat representative of the NACDG data as a whole. A second comparison can be made with the UK s surveillance system data: EPIDERM and the Occupational Physicians Reporting Activity (OPRA). These surveillance schemes group allergens into more holistic groups but results from both indicate that the most common OCAs in the Canadian NACDG population map well to the published EPIDERM and OPRA data [McDonald et al. 2006]. For example, rubber chemicals and materials, which would include thiuram and carba mix, were the materials most frequently reported as causing cases of contact dermatitis in both OPRA and EPIDERM from 1996 to 2001 [McDonald et al. 2006]. Resins and acrylics, nickel and preservatives were also in the ten most common agents reported as causes of contact dermatitis during this period. OPRA and EPIDERM do not differentiate between irritant and allergic contact dermatitis; this both explains why wet work appears as the second leading cause of dermatitis, and suggests why more of the agents identified in the current study were not more highly ranked in the OPRA/ EPIDERM results. These comparisons suggest that, although the small sample size may have caused the results to be influenced by local industry and case clusters, the most common agents observed in the Canadian NACDG data are common occupational contact allergens in other parts of the world. Determining whether an OCA is related to occupational asthma presented a few challenges, specifically in the case of epoxy resin. The epoxy resin on the NACDG standard tray contains epichlorhydrin and bisphenol A that react to form the epoxy resin monomer diglycidyl ether of bisphenol A (DGEBA or BADGE). However, a worker using epoxy resin systems in the workplace will likely have exposure to both the epoxy resin (DGEBA or BADGE) as well as at least one hardener to cure the epoxy. Several common hardeners, including amine and acid anhydride formulas are known to cause occupational asthma; these agents were listed in both reference texts [Bernstein et al. 2006a, UK Health and Safety Executive,. 2001] as known causes of OA. Epoxy resin monomer, specifically DGEBA, has also been associated with asthma in the peer-review literature [Moulin et al. 2009, Hannu et al. 2009, Kanerva et al. 2000] but was not listed as a cause of OA in either text. It is clear that components of epoxy resin systems are capable of causing AOCD and OA, and that a worker using epoxy resins systems may be at risk for both OCD and OA from separate chemical exposures during the use of epoxy resin systems.

69 50 For this reason we included both the epoxy resin monomer and the hardeners in our review of the reference documents and literature. When the common OCAs were cross-referenced with the OA literature, seven of the ten OCAs were determined to be established or possible causes of OA (epoxy resin, nickel sulfate, cobalt chloride, potassium dichromate, PPD, formaldehyde and glutaraldehyde). A wide variety of workers may be exposed to these common OCAs in various occupational settings [National Institutes of Health. 2009]. Epoxy resins are commonly used in coatings for metal, wood, concrete, or plastic, but are also found in adhesives, castings and electrical components. Nickel sulfate is used as a dye and in the process of electroplating; it can also be found in both coatings and ceramics and is commonly used in the production of other nickel compounds. Cobalt chloride is found in a wide variety of applications including as a laboratory reagent, in the process of electroplating, and as a glass or porcelain pigment. Potassium dichromate can be used in the processes of dyeing, bleaching and leather tanning and can also be found in some types of cement. It can also be present in the health care setting as an antiseptic or astringent, as well as in painting, printing, coating or staining. P-phenylenediamine (PPD) is present in dark hair dyes used by hairdressers, but can also be used in photographic development. Formaldehyde is used in the production of formaldehyde resins, plywood, particle board and urea-formaldehyde foam. Glutaraldehyde is used for sterilization purposes in health care settings. Embalmers and laboratory workers may have exposure to both formaldehyde and glutaraldehyde. Three OCAs were found to have no evidence of an association with OA: thiuram, carba mix and glyceryl thioglycolate. Glyceryl thioglycolate is found in permanent wave solution used by hairdressers. Thiuram and carba mix are accelerants used in rubber manufacturing, and can be present in the final manufactured rubber product (i.e. rubber elastic components, rubber gloves, rubber tool hand grips). Components of the carba mix allergen can also be found in carbamate pesticides. The results suggesting an association between epoxy resin and OA are largely based on the link between amine and anhydride hardeners (used to cure the epoxy resin) with OA (rather than from the epoxy monomers themselves, though BADGE (or DGEBA) has been shown in three case studies to be related to OA [Moulin et al. 2009, Hannu et al. 2009, Kanerva et al. 2000].

70 51 When the ten most common OCAs identified in the Canadian NACDG data were crossreferenced in the occupational hygiene resources only the NLM NIH HazMap database identified all ten as causing skin effects. The ACGIH TLV Hand Book reported only epoxy resin hardeners, formaldehyde and glutaraldehyde as potential sensitizing agents, but does not specify the system of effect. The NIOSH Pocket Book identified epoxy resin hardeners, nickel sulfate and glutaraldehyde as potential respiratory sensitizers as well as nickel sulfate, potassium dichromate and glutaraldehyde as possible dermal sensitizers. In the ACGIH TLV Hand Book the sensitizer (SEN) notation does not distinguish between the respiratory, dermal or conjunctival organ systems. In both the NLM NIH HazMap database and the NIOSH Pocket Book respiratory and dermal effects are separate notations. However, the HazMap database does not differentiate between sensitizing and irritant, thus a limitation is that the notations could refer to either allergic or irritants exposure effects. Though each OH reference document has a slightly different purpose and scope, the contradictory information provided by these sources is problematic. Neither a practicing occupational hygienist nor occupational physician should be required to reference multiple sources for comprehensive information on an occupational exposure. More work is needed to ensure consistency in the application of sensitizer notations if workers are to be sufficiently protected; this process has been proposed in other jurisdictions [Schnuch et al. 2002]. This study analyzed six years of patient data from the two Canadian contributors to the NACDG. The results could reflect local clusters of exposure as one of the centers has a large telecommunications industry with known epoxy exposure. Due to the small number of AOCD cases, the data is perhaps more sensitive to these regional clusters. However, the common OCAs that were identified in this study have previously been reported as some of the most frequent OCAs [Rietschel et al. 2002] Limitations There are some limitations with our study to be noted. It must be emphasized that the basis for the list of common OCAs was a database of patch test data focused exclusively on contact sensitizers. The OCAs identified as possible causes of OA should not be interpreted as common causes of OA, but rather as common contact sensitizers (or common causes of ACD) that are

71 52 also capable of causing OA. No inference on each OCA s likelihood of causing OA, or frequency of cases, can be made from this study. There is no source of information on occupational respiratory sensitizers that is comparable to the North American Contact Dermatitis Group database, making it challenging to identify the most common asthmagens at the population level. Although the agents discussed in this study are common contact sensitizers they are not necessarily the most common respiratory sensitizers, they are simply capable of causing OA in some workers. The practical importance of the current findings are 2-fold: first, there are many OCAs that can also cause OA. Awareness of this by occupational hygienists and clinicians alike needs to be improved. Second, the designation of workplace chemicals as potential skin and/or lung sensitizers is lacking in consistency across databases and agencies. Results from this study highlight the importance of considering both dermal and inhalation routes of exposure. Analysis of the data from the NACDG Canadian data showed that of the ten most common OCAs, seven have been associated with OA in the literature. Unrecognized or uncontrolled exposure to these agents through either dermal or inhalation routes of exposure can potentially lead to work-related allergic disease in both the skin and the respiratory system.

72 53 Chapter 4 Co-existing Skin and Respiratory Symptoms in Four Occupational Groups Victoria H. Arrandale 1 and D. Linn Holness 1,2 1 University of Toronto, Toronto, ON, Canada 2 St. Michael s Hospital, Toronto, ON, Canada This short report manuscript was peer-reviewed and ultimately rejected. This manuscript is currently under revision in preparation for re-submission.

73 Abstract Background: There is evidence in animals of an interaction between the skin and respiratory system in the development of allergic skin and respiratory disease. The evidence supporting a similar association in humans is limited. The goal of this study was to determine if workers report both skin and respiratory symptoms related to their work. Methods: Data from four studies were pooled. Information on self-reported skin and respiratory symptoms was obtained using an interviewer-administered questionnaire; pulmonary function was measured in all subjects. Results: A total of 113 (46%) workers reported at least one respiratory symptom; 42 (17%) reported a skin rash. Overall, 26 (11%) workers reported both skin and respiratory symptoms (range 6-17% across groups). Only 2 workers reported both work-related skin and work-related respiratory symptoms. Conclusions: As the potential interaction between the skin and the respiratory system is further investigated thorough information about exposure and response in both systems must be collected.

74 Introduction There is some evidence in animals of an interaction between the skin and respiratory system with respect to the development of allergic skin and respiratory disease. Results from animal studies have shown that the skin is a viable route of sensitization for some exposures [Rattray et al. 1994, Zhang et al. 2004]. The evidence that there is a similar association in humans is still quite limited, however there are case reports of co-existing work-related allergic contact dermatitis and allergic asthma [Lockman. 2002]. Occupational disease researchers are beginning to consider the skin and the lungs as important routes of exposure [Redlich and Herrick. 2008]. In the case of experimental animal models, if dermal exposure is the route of sensitization, an inhalation exposure is still required to trigger an asthma-like response. However, once sensitized, the inhalation exposure required to trigger an asthma-like response may be much lower than the exposure level required to sensitize the airways [Arts et al. 2006]. We hypothesize that if dermal and airborne exposures do occur in the workplace there will be a portion of the working population that experiences symptoms in both systems. These symptoms could be irritant or allergic in nature. The goal of this study was to use previously collected data from four occupational groups to determine if, and how many, workers experience both skin and respiratory symptoms that they relate to their work. 4.3 Methods Previously collected data from four cross-sectional studies were pooled for this analysis. The four studies included workers from a soda ash production facility [Holness et al. 1989] (exposure: ammonia), a softwood planing mill (exposure: softwood dust), embalming [Holness and Nethercott. 1989] (exposure: formaldehyde and glutaraldehyde) and cabinet making [Sass- Kortsak et al. 1986] (exposure: hardwood dust). All studies were initially designed to study work-related skin problems and/or work-related lung problems. The study of respiratory effects in softwood planing mill workers has not been previously published. Briefly, this was a crosssectional study of fifty softwood sawmill workers. The workers respiratory and cutaneous statuses were evaluated and dust exposure was measured over a work-week, allowing exposureeffect relationships to be assessed. The research ethics board at St Michael s Hospital in Toronto, Canada, approved all original studies.

75 56 In each study, information on self-reported skin and respiratory symptoms was obtained using an interviewer-administered questionnaire. Data on basic demographics (e.g., age, sex, smoking) as well as self-reported symptoms were collected. Pulmonary function was measured and is reported as percent of predicted forced expiratory volume in one second (FEV 1 ) based on the equations of Crapo et al. [Crapo et al. 1981]. Respiratory symptom questions were based on the American Thoracic Society Questionnaire (ATSQ). The question employed for skin rash was as follows, Do you have a skin rash?. A subject was considered to have at least one respiratory symptom if they said yes to any of the cough, wheeze, shortness of breath or phlegm questions. In order to determine the respiratory symptom s relation to the subjects work, subjects were asked Is your [cough/wheeze/shortness of breath/phlegm] better, worse, unchanged when you re at work? For skin rash, subjects were asked Is your [skin] rash better, worse, unchanged when you re [at work/on holiday/on layoff]? Differences between occupational groups were investigated using chi-square tests and ANOVA with post-hoc Scheffe. The prevalence of self-reported respiratory and skin symptoms was determined using basic descriptive statistics. Subjects were divided into four possible symptom groups: no symptoms, only skin symptoms, only respiratory symptoms and both skin and respiratory symptoms. Between group differences in age, sex, pulmonary function and smoking were examined. All analyses were completed using SAS v Results A total of 247 workers were included in this analysis (Table 12). The individual studies had sample sizes ranging from 50 to 86 (average 62). The entire population was male, except for 9 female embalmers. The mean age of subjects was 36.4 years (sd 12.7). Cabinet makers were significantly older than all other groups (p<0.0001). There was no difference in the proportion of ever smokers between the studies though among smokers, cabinet makers had more pack-years of smoking (p=0.0023), likely due to their older age. Pulmonary function, reported as percentpredicted FEV 1, was approximately normal for all groups; there were no differences between studies in measures FEV 1 percent predicted. A total of 113 (46%) workers reported at least one respiratory symptom; 42 (17%) reported a skin rash. Overall, dyspnea was the least common self-reported symptom (n=33, 13%) and

76 57 phlegm was the most common (n=65, 26%). Cabinet makers were more likely than other workers to report a cough (p=0.0444); no other differences were observed between study groups. Of the 113 workers who reported at least one respiratory symptom, 34 (30%) reported at least one respiratory symptom that was work-related. Wheeze was the most common work-related symptom reported (n=15, 35% of self-reported wheeze). Overall, embalmers were significantly less likely to report work-related respiratory symptoms compared with the other study groups (p=0.0002). Overall, 26 (11%) workers reported both skin and respiratory symptoms (range 6-17% across occupational groups). Embalmers were the most likely to report both skin and respiratory symptoms. Soda ash production workers had the highest prevalence of reporting skin symptoms only and the lowest of respiratory symptoms only, however, these differences did not reach statistical significance. Subjects reporting both skin and respiratory symptoms were more likely to be female (p<0.0001). No differences between the symptom groups were observed in age, smoking or FEV 1 percent predicted (Table 12). There were four workers who reported work-related skin symptoms but no work-related respiratory symptoms while 32 workers reported work-related respiratory symptoms and no work-related skin symptoms. Only 2 workers reported both work-related skin and work-related respiratory symptoms. These two workers were both male and worked in the Soda Ash Production facility; neither reported allergies or asthma. One of the two workers was a smoker with 21 pack years of smoking history; the other had never smoked. One worker was 36 years old with normal lung function results (FEV 1 92% pred., FVC 95% pred.), the other was 56 years old with below predicted lung function (FEV 1 71% pred., FVC 64% pred.).

77 58 Table 12 Skin and Respiratory Symptom Group Distribution (Work-Related and Non- Work-Related) Across Studies and Description of Groups by Age, Sex, Smoking and Pulmonary Function Variables. Symptom Frequencies Given as Row n (%). Demographic Variables Given as Column n (%). None Skin only Respiratory only Both skin & respiratory p-value Symptoms Overall 116 (47%) 16 (6%) 87 (35%) 26 (11%) Soda Ash Prod 32 (55%) 7 (12%) 14 (24%) 5 (9%) NS Cabinet 25 (50%) 2 (4%) 20 (40%) 3 (6%) Softwood 23 (43%) 3 (6%) 23 (43%) 4 (8%) Embalming 36 (43%) 4 (5%) 30 (36%) 14 (17%) Female 2 (2%) 0 (0%) 0 (0%) 7 (27%) < Age (yrs), mean (sd) 36.2 (12.3) 35.2 (14.0) 37.0 (13.0) 36.3 (13.1) NS Ever Smokers 75 (64%) 11 (68%) 64 (74%) 22 (84%) NS Pack-yrs, mean (sd) 16.6 (18.2) 16.9 (19.7) 20.2 (18.8) 14.5 (18.3) NS FEV 1 % pred., mean (sd) 95.5 (12.8) 91.5 (16.6) 93.6 (11.7) 90.8 (12.6) NS Work-Related Symptoms Overall 207 (84%) 4 (1.6%) 32 (13%) 2 (0.8%) - Soda Ash Prod 47 (81%) 1 (1.7%) 8 (14%) 2 (3.4%) Cabinet 39 (78%) 1 (2.0%) 10 (20%) 0 (0%) Softwood 41 (77%) 1 (1.9%) 11 (21%) 0 (0%) Embalming 80 (95%) 1 (1.2%) 3 (3.6%) 0 (0%) Female 8 (3.8%) 1 (25%) 0 (0%) 0 (0%) NS b Age, mean yr (sd) 35.3 (12.2) a 42.8 (13.5) 42.1 (14.3) a 46.0 (14.1) Ever Smokers 145 (70%) 4 (100%) 22 (69%) 1 (5%) NS Pack-yrs, mean (sd) 17.1 (18.5) 22.8 (14.8) 20.8 (19.8) 21.6 (-) NS FEV 1, % pred., mean (sd) 94.2 (12.4) 80.9 (3.9) 95.8 (13.4) 79.4 (22.1) NS b a significantly different groups in post hoc comparisons; b 0.05 < p < 0.10

78 Discussion Among workers in four different occupational groups the prevalence of co-existing skin and respiratory symptoms was 11%. These subjects were more likely to work in embalming and be female, though these variables were highly correlated. No differences in smoking, age or FEV 1 percent predicted were observed between symptom groups. The prevalence of skin symptoms (14%) and respiratory symptoms (range 18-35%) is similar to other studies in occupational populations by the same research group [Nethercott and Holness. 1988, Holness et al. 1984]. Very few studies have reported the prevalence of co-existing skin and respiratory symptoms. A recent study by Lynde et al. among professional cleaners reported a much higher prevalence of co-existing skin and respiratory symptoms. In this study, among workers with a current rash almost two-thirds reported two or more work-related respiratory symptoms [Lynde et al. 2009]. The reported prevalence of co-existing symptoms in the dermal and respiratory systems does not indicate whether the symptoms are related to the same exposure, nor does it provide an indication of whether the symptoms are due to allergic or irritant mechanisms. The reported prevalence simply indicates that in working populations that have demonstrated airborne and/or dermal exposure to agents capable of causing allergic and/or irritant responses there exists a group of workers who report symptoms in both systems, and a small portion of these workers identify these symptoms as being associated with their jobs. The workers with co-existing skin and respiratory symptoms make up a small portion (11%) of the workers in this sample. This group of workers is important because they demonstrate that in workplaces where one route of exposure may be recognized as the most hazardous, some workers still report symptoms in other systems. However, of these workers only a small number reported that their symptoms were related to their work. The hypotheses of how the skin and the lung may interact in terms of exposure and outcomes are complex. If a worker is sensitized due to dermal exposure, a respiratory response may be triggered by relatively low airborne exposure. This has implications for exposure control because airborne exposure limits may prevent airways sensitization but may not be protective for elicitation in the airways of a sensitized worker.

79 60 This study shows that there is a portion of workers who experience both skin and respiratory symptoms that they attribute to their work. As we continue to further investigate the potential interaction between the skin and the respiratory system in terms of exposure and disease we need to ensure that we collect thorough information about exposure and response in both systems, using quantitative measurements in addition to self-reported symptoms where possible. Although there is a need to understand how the skin and respiratory exposure routes contribute to occupational respiratory disease, it is clear that both dermal and airborne exposures must be minimized in the occupational setting. The techniques for achieving reduction and elimination of occupational exposures are well understood and should be applied in the workplace.

80 61 Chapter 5 Skin Symptoms in Bakery and Auto Body Shop Workers: Associations with Exposure and Respiratory Symptoms Victoria H. Arrandale 1, Tim Meijster 2, Anjoeka Pronk 2, Gert Doekes 3, Carrie A. Redlich 4, D. Linn Holness 1, Dick Heederik 3 1 Centre for Research Expertise in Occupational Disease, University of Toronto, Canada 2 TNO Quality and Safety, Zeist, The Netherlands 3 Environmental Epidemiology Department, Institute for Risk Assessment Sciences, Utrecht University, The Netherlands 4 Yale University School of Medicine, New Haven, CT USA This manuscript was e-published ahead of print by the International Archives of Occupational and Environmental Health on March 13, Copyright permission has been granted (see Copyright Acknowledgements).

81 Abstract Background: Despite the importance of skin exposure, studies of skin symptoms in relation to exposure and respiratory symptoms are rare. The goals of this study were to describe exposureresponse relationships for skin symptoms, and to investigate associations between skin and respiratory symptoms in bakery and auto body shop workers. Methods: Data from previous studies of bakery and auto body shop workers were analyzed. Average exposure estimates for wheat allergen and diisocyanates were used. Generalized linear models were constructed to describe the relationships between exposure and skin symptoms, as well as between skin and respiratory symptoms. Results: Data from 723 bakery and 473 auto body shop workers were analyzed. In total, 5.3% of bakery and 6.1% of auto body shop workers were female; subjects mean age was 39 and 38 years, respectively. Exposure-response relationships were observed in auto body shop workers for itchy or dry skin (PR 1.55, 95% CI ) and work-related itchy skin (PR 1.97, 95% CI ). A possible exposure-response relationship for work-related itchy skin in bakery workers did not reach statistical significance. In both groups reporting skin symptoms was strongly and significantly associated with reporting respiratory symptoms, both work-related and non-workrelated. Conclusions: Exposure-response relationships were observed for skin symptoms in auto body shop workers. The lack of significant exposure-response associations in bakery workers should be interpreted cautiously. Workers who reported skin symptoms were up to four times more likely to report respiratory symptoms. Improved awareness of both skin and respiratory outcomes in exposed workers is needed.

82 Introduction The connection between skin and respiratory systems in occupational disease is a growing area of research interest [Redlich and Herrick. 2008]. Specifically, there is interest in determining whether the skin can be an important route of sensitization for occupational allergens and subsequent development of occupational respiratory symptoms, including asthma. Research in this area is challenging, in part due to the organ system silos that have historically existed in medicine and epidemiological research. Recent evidence from animal models suggests that after sensitization through skin exposure to some high (e.g., latex) and low (e.g., trimellitic anhydride, toluene diisocyanate (TDI)) molecular weight agents, an asthma-like response can be elicited upon inhalation exposure [Vanoirbeek et al. 2004, Zhang et al. 2009]. Evidence of possible cross-system sensitization and elicitation in humans is scarce. Among methylene diphenyl diisocyanate (MDI) exposed workers, Petsonk et al. observed that subjects reporting skin staining (a proxy for skin exposure) were more likely to report asthma-like symptoms [Petsonk et al. 2000]. Despite the possibility that skin exposures can contribute to the burden of respiratory disease, studies focusing on skin exposure, and specifically on exposure-response studies for skin symptoms and/or sensitization, are rare. This lack of evidence limits the ability to infer causality between skin exposure and response, and may ultimately hamper efforts to better control both skin exposure as well as skin and respiratory symptoms in the workplace. Studies on skin symptoms in relation to exposure do exist [de Joode et al. 2007, Sripaiboonkij et al. 2009b, Sripaiboonkij et al. 2009a], but even less information is available on the associations between exposure, skin and respiratory symptoms as well as the relationship between skin and respiratory effects. Many occupational studies report the prevalence of both skin and respiratory symptoms but rarely explore the relationship between the two, or the prevalence of these symptoms coexisting. Lynde et al. reported that among male cleaners, those with skin symptoms were more likely to report respiratory symptoms [Lynde et al. 2009]. The mechanisms of airborne and skin exposure are complex. Airborne and skin exposures can be related if they share sources, but these associations are so far poorly studied [Schneider et al. 1999]. Associations between skin and airborne exposures have been reported for bitumen and

83 64 pyrene in road pavers, 1,6-hexamethylene diisocyanate (HDI) in spray painters, methylene bisphenyl isocyanate (MDI) in foundry works, solvents in spray painters and nickel exposure in primary industries [McClean et al. 2004, Burstyn et al. 2002, Chang et al. 2007a, Fent et al. 2008, Liljelind et al. 2010, Hughson and Cherrie. 2005]. In two other studies, both involving pesticide exposure, there was no association found between skin and airborne exposure. The authors attribute this lack of association to the fact that the primary source of skin exposure was likely contact with contaminated foliage rather than the settling of airborne pesticide [Flack et al. 2008, Aprea et al. 2009]. Bakery and auto body shop workers have both skin and respiratory exposures to known occupational allergens, making them good candidates for further study of exposure-response relationships for skin symptoms, as well as the relationship between skin and respiratory symptoms. Bakery and auto body shop workers are at increased risk of occupational asthma (OA) as well as occupational skin disease (OSD) due to their workplace exposures: flour dust and diisocyanates, respectively [McDonald et al. 2005, McDonald et al. 2006]. Flour dust is a common cause of occupational asthma in bakers. Flour dust, which includes wheat and α- amylase allergens among others, contains high molecular weight (HMW) antigens which act through an IgE mediated (Type I) immunological pathway to cause OA and contact urticaria, and can also cause contact dermatitis through a Type IV (cell-mediated) mechanism [Nethercott and Holness. 1989]. Isocyanates are a heterogeneous group of compounds, including monomers and oligomers, categorized as low molecular weight (LMW) antigens. The mechanism of action leading to isocyanate-induced OA is not yet fully understood and though IgE (Type I) mediated processes appear to play a role in some cases, other unrevealed mechanisms play a role in respiratory sensitization [Maestrelli et al. 2009, Wisnewski. 2007]. Similar to flour dust, diisocyanates can also cause contact dermatitis (Type IV) [Donovan et al. 2009, Frick et al. 2003]. The goals of this study are to describe the exposure-response relationships for skin symptoms in both bakery workers and auto body shop workers, and to investigate the association between skin and respiratory symptoms in these two groups.

84 Methods Reports on respiratory outcomes in both the bakery and auto body shop workers studies have been published previously [Pronk et al. 2007, Jacobs et al. 2008]. Workers were asked to complete a questionnaire on respiratory and skin symptoms, an exposure questionnaire and also to provide a blood sample for analysis. For this analysis subjects were required to have complete data for both respiratory and skin symptoms, as well as atopy and workplace allergen specific IgE. In total, 723 bakery workers and 472 auto body shop workers were included in this analysis, which is a slightly different study population than previous publications [Pronk et al. 2007, Jacobs et al. 2008] Exposure In both groups (bakery and auto body shop workers), exposure was estimated based on existing data sets of personal airborne exposure measurements [Pronk et al. 2006a, Meijster et al. 2007]. Cumulative monthly hexamethylene diisocyanate (HDI) exposure was estimated using taskbased measurements of airborne diisocyanates combined with self-reported monthly frequencies of task completion as was described previously [Pronk et al. 2007]. This exposure metric was then divided by the self reported average number of hours worked per month to determine the long-term average diisocyanate exposure of these workers (µg-nco*m-3), which facilitated comparison with the bakery workers. Average wheat exposure for bakery workers was estimated using subjects work characteristics (exposure determinants) reported on the questionnaire combined with an exposure model constructed by Meijster et al., to predict average wheat exposures (µg-dust*m-3) for each subject [Meijster et al. 2007]. A relatively small number of task based skin exposure measurements were available for diisocyanate exposure in auto body shops, but no comparable exposure measurements were available in bakery workers. As a result this study investigates the exposure-response relationships for skin symptoms, using airborne exposure as a proxy for skin exposure in both working populations. In auto body shop workers, airborne exposure was not significantly associated with having a detectable skin diisocyanate exposure (OR 1.34, ), but the analysis was limited by small number of samples and a direct correlation was not calculated [Pronk et al. 2006b].

85 Specific IgE and Atopy Specific IgE was measured using commercially available kits as previously described [Pronk et al. 2007, Jacobs et al. 2008]. In bakery workers specific IgE was measured for wheat protein (Bakery, Pharmacia, Unicap System, Pharmacia Diagnostics, Uppsala, Sweden); in auto body shop workers specific IgE to HDI oligomers (N100-HSA) was measured (Diisocyanates: Phadia, Uppsala Sweden). All samples were also tested for specific IgE to common aeroallergens (house dust mite, cat, dog, grass or birch pollen) [Doekes et al. 1996]. Analytical results were dichotomized and IgE (work-related or common allergens) was considered elevated if above 0.35kU/L. Subjects were classified atopic if they had elevated IgE in response to at least one of the common aeroallergens Symptoms Respiratory symptoms and skin symptoms were reported on a self-completed questionnaire derived from the International Union Against Tuberculosis and Lung Disease (IUATLD) and the Medical Research Council - European Community of Coal and Steel (MRC-ECCS) for the bakery workers, and from the British Medical Research Council (BMRC) respiratory questionnaire for auto body shop workers [Burney et al. 1989, van der Lende and Orie. 1972, Medical Research Council on the Aetiology of Chronic Bronchitis. 1960]. Information on cough, phlegm, wheeze, chest tightness, shortness of breath and self reported asthma were included. A variable describing asthma-like symptoms (wheezing, chest tightness, current/previous asthma) was constructed using the individual symptom responses. Skin itch and dry skin were reported on the questionnaire; a dichotomous variable describing the presence of either itchy or dry skin was constructed. Work-related symptoms were explicit items on the questionnaire. Subjects were asked directly whether they have itchy skin at work and whether they experience asthma-like symptoms at work. No work-related symptom variables were constructed post-hoc Additional Variables Age, sex, smoking (current and historical) as well as years working were self-reported on the questionnaire.

86 Analyses Iterative non-parametric regression models (smoothing splines) with generalized additive models (PROC GAM) were first used to explore the shape of the exposure-response relationships for skin outcomes at the population level. These models were used to explore unadjusted non-linear relationships between estimated exposure and symptoms outcomes. Generalized cross validation (GCV) was used to select the smoothing parameter degrees of freedom (df); the df selected were limited to four to avoid large fluctuations that are likely not biologically relevant [Hastie. 1990]. Generalized linear models (SAS PROC GENMOD) with a log function were used to estimate unadjusted and adjusted prevalence ratios (PR) for the associations between exposure, atopy, specific sensitization and symptoms. Adjusted models included atopy, work-related specific IgE sensitization, age and sex; respiratory symptom models were additionally adjusted for smoking status. Sensitivity analyses were completed to explore whether atopy and specific sensitization were modifying the exposure-response relationships. Exposure-response relationships were investigated in models where atopic and specific sensitized subjects were excluded. All PR estimates for exposure effects are reported as the PR associated with an inter-quartile range (IQR) increase in exposure. Additionally, relationships between skin and respiratory symptoms were explored using generalized linear models (PROC GENMOD) as described above with the same covariates and including sensitivity analyses to explore the effect of atopy and work-related specific sensitization. All analyses were completed in SAS v.9 software (SAS Institute Inc., Cary, NC, USA). 5.4 Results Both the auto body shop and bakery workers were predominantly male with an average age of approximately 38 and 39 years respectively (Table 13). The distribution of smoking status was similar between the two groups, though there were more never-smokers among the bakery workers. The prevalence of atopy among bakery and auto body shop workers was similar (34% vs. 36%, respectively) but the prevalence of specific sensitization to workplace allergens was higher

87 68 among bakery workers (Table 13). Eleven percent of bakery workers had wheat-specific IgE; only 2% of auto body shop workers had HDI-specific IgE. Differences between the bakery and auto body shop workers were observed in symptom frequencies (Table 13). We observed slightly more respiratory symptoms in auto body shop workers and more skin symptoms in bakery workers. Estimated average exposure among auto body repair shop workers ranged from 0 353µg-NCO*m-3 (IQR 21.4), and among bakery workers from µg-wheat*m-3 (IQR 32.9) based on the previously collected exposure measures. Smoothing splines (Figure 3 and Figure 4) show the shape of the exposure-response distribution for skin symptoms at a population level, stratified by atopy. Among bakers the exposureresponse relationship for skin symptoms appears to be linear in both the atopic and non-atopic groups. However, in auto body shop workers a bell-shaped distribution is supported (df=3.7; p<0.05) in non-atopic subjects. Similar analyses for respiratory symptoms have been previously reported for both the bakery and auto body shop workers [Pronk et al. 2007, Jacobs et al. 2008]. Graphs for respiratory symptom models directly comparable to the skin symptom models presented here are provided for comparison in Appendix 1 (Figure 7 and Figure 8).

88 69 Table 13 Demographics and Symptom Frequencies for Both Auto Body Repair and Bakery Workers. All Frequencies Given as Column n (%). Auto Body Repair Bakery Workers Workers Demographics Overall, n Female, n (%) 29 (6.1%) 38 (5.3%) Age, years, mean (sd) 38.0 (11) 39.0 (11) Current smoker, n (%) 173 (37%) 238 (33%) Former smoker, n (%) 130 (28%) 157 (22%) Never smoker, n (%) 170 (36%) 328 (45%) Years working, mean (sd) 17.6 (11%) 14.4 (11%) Symptoms, n (%) Cough 65 (14%) 83 (12%) Wheeze, ever 111 (24%) 111 (15%) Asthma, ever 72 (15%) 71 (9.8%) Asthma Symptoms 134 (28%) 174 (24%) Work-related asthma symptoms 20 (4.2%) 15 (2.1%) Dry Skin in the last 12 months 113 (24%) 188 (26%) Itchy Skin in the last 12 months 50 (11%) 208 (29%) Either itchy or dry skin in the last 12 months 134 (28%) 265 (37%) Work-related itchy skin 40 (8.5%) 122 (17%) Atopy and Specific IgE, n (%) Atopy 169 (36%) 245 (34%) HDI specific IgE 10 (2.1%) - Wheat specific IgE - 82 (11%)

89 70 Figure 3 Auto Body Shop Workers: Associations Between Average Diisocyanate Exposure and Skin Symptoms, Shown in Smoothed Plots, Stratified by Atopy. Data rug indicates the distribution of observations by exposure level. (a) Itchy or dry skin in atopic subjects (linear: NS; spline: NS), (b) Work-related itchy skin in atopic subjects (linear: NS; spline: NS), (c) Itchy or dry skin in non-atopic subjects (linear: NS; spline: df=1.05, p<0.05), (d) Work-related itchy skin in non-atopic subjects (linear: NS; spline: df=3.71, p<0.05).

90 71 Figure 4 Bakery Workers: Associations Between Average Wheat Exposure and Skin Symptoms, Shown in Smoothed Plots, Stratified by Atopy. Data rug indicates the distribution of observations by exposure level. (a) Itchy or dry skin in atopic subjects (linear: NS; spline: NS), (b) Work-related itchy skin in atopic subjects (linear: NS; spline: NS), (c) Itchy or dry skin in non-atopic subjects (linear: NS; spline: NS), (d) Work-related itchy skin in non-atopic subjects (linear: NS; spline: NS).

91 72 In auto body shop workers (Table 14), statistically significant exposure-response relationships were observed for itchy or dry skin (PR 1.56, 95% CI ) and work-related itchy skin (PR 1.97, 95% CI ); a similar trend was observed in the bakery workers for work-related skin symptoms but this did not reach significance (Table 14). In auto body shop workers (Table 14), exposure was significantly related to specific HDI sensitization (PR 10.0, 95% CI ), with wide confidence limits likely due to the small number of sensitized subjects. HDI specific sensitization was associated with itchy or dry skin (PR 1.86, 95% CI ) but not work-related itchy skin. Atopy predicted itchy or dry skin in auto body shop workers (PR 1.26, 95% CI ) but not work-related itchy skin. Among bakery workers (Table 14), wheat exposure was not related to having wheat specific sensitization, but wheat specific sensitization was associated with work-related itchy skin (PR 2.17, 95% CI ). Atopy was associated with both itchy or dry skin (PR 1.45, 95% CI ) and work-related itchy skin (PR 1.67, 95% CI ). In both groups, exposure was negatively associated with atopy, though this relationship only reached significance in the auto body shop workers (Table 14). When atopy and specific sensitization were added to exposure-response models for skin symptoms, the effect on prevalence ratios due to exposure remained relatively unchanged in both groups (Table 15). Removing the atopic and sensitized (work-related specific IgE) subjects also did not change the exposure relative risk estimates (results not shown). The association between reporting skin symptoms and reporting respiratory symptoms was investigated separately (Table 16). In both auto body shop and bakery workers, reporting itchy/dry skin and work-related itchy skin was significantly associated with reporting wheeze and asthma-like symptoms. Both work-related and non-work-related skin symptoms were significantly associated with work-related chest tightness in auto body shop workers. In bakery workers, work-related itchy skin was not significantly associated with work-related chest tightness.

92 73 Table 14 Results of Generalized Linear Models Describing the Simple Relationship Between Exposure, Skin Symptoms, Atopy and Specific IgE. Each Reported Prevalence Ratio (PR) Was Estimated From a Separate Model. Models Adjusted for Age and Sex. (WR=Work-related) Independent Variable Dependant Variable PR (95% CI) Auto Body Repair Workers (n=473): Average Diisocyanate Exposure (µg- Itchy or Dry Skin 1.56 ( ) NCO*m-3) WR Itchy Skin 1.97 ( ) Atopy 0.83 ( ) HDI-Specific IgE 10.0 (1.4-73) Atopy Itchy or Dry Skin 1.26 ( ) WR Itchy Skin 0.80 ( ) HDI-Specific IgE Itchy or Dry Skin 1.86 ( ) WR Itchy Skin 1.03 ( ) Bakery Workers (n=723): Average Wheat Exposure (µg*m-3) Itchy or Dry Skin 0.96 ( ) WR Itchy Skin 1.16 ( ) Atopy 0.91 ( ) Wheat-Specific IgE 1.12 ( ) Atopy Itchy or Dry Skin 1.45 ( ) WR Itchy Skin 1.67 ( ) Wheat -Specific IgE Itchy or Dry Skin 1.22 ( ) WR Itchy Skin 2.17 ( )

93 74 Table 15 Prevalence Ratio (PR) of Symptoms per Inter-Quartile Range (IQR) Increase in Average Exposure. Multivariate Models Adjusted for Atopy and Specific Sensitization in Addition to Age, Sex and Smoking As Described. Outcome Covariates PR (95% CI) Auto Body Repair Workers (n=473) Either itchy or dry skin in last 12 months A, S, Atp, IgE 1.55 ( ) Work-related itchy skin A, Atp, IgE 1.97 ( ) Bakery Workers (n=723) Either itchy or dry skin in last 12 months A, S, Atp, IgE 0.96 ( ) Work-related itchy skin A, S, Atp, IgE 1.14 ( ) A=Age, S=Sex, Atp=Atopy, IgE=Work-related Specific IgE Table 16 Association Between Skin Symptoms and Respiratory Symptoms in Both Bakery and Auto Body Repair Workers. Reported as Prevalence Ratio of Respiratory Symptoms, Adjusted for Age, Sex, Smoking and Atopy with 95% Confidence Intervals. Auto Body Repair Workers Bakery Workers Predictor Outcome PR (95% CI) PR (95% CI) Wheeze, ever 2.01 ( ) 1.94 ( ) Itchy or dry skin Asthma-like symptoms 1.83 ( ) 1.78 ( ) in last 12 months Work-related itchy skin WR asthma symptoms 4.06 (1.6-10) 3.90 (1.2-12) Wheeze, ever 2.50 ( ) 1.60 ( ) Asthma-like symptoms 2.12 ( ) 1.54 ( ) WR asthma symptoms 3.61 ( ) 2.15 ( )

94 Discussion Significant exposure-response relationships were observed between estimated exposure to diisocyanates (µg-nco*m-3) and skin symptoms in auto body shop workers. Such associations have not been previously reported. Though similar trends were observed between wheat exposure and work-related skin symptoms in bakery workers, the associations did not reach statistical significance. Both auto body repair and bakery workers who reported skin symptoms were consistently and significantly more likely to report work-related and non-work-related respiratory symptoms. These findings are comparable with results of Lynde et al. who showed that male cleaners with a skin rash were more likely to report respiratory symptoms, particularly work-related respiratory symptoms [Lynde et al. 2009]. The prevalence of skin symptoms reported in auto body shop workers and bakery workers is similar to previous studies of skin outcomes in these populations. Randolph et al. reported that 32% of HDI exposed spray painters reported hand dermatitis, while Daftarian found 35% of TDI exposed workers to have skin symptoms [Randolph et al. 1997, Daftarian et al. 2002]. Cullinan et al. found that 11% of bakery and flour mill workers had skin symptoms [Cullinan et al. 2001]. Steiner et al. reported that 19% of all bakers and 31% of high-risk (higher likelihood of exposure) bakers reported at least one skin symptom in the last 12 months [Steiner et al. 2011]. Previous research supports that self-reported skin symptoms are predictive of skin disease. However, some results suggest that self-reported skin symptoms may over estimate [Smit et al. 1992, Lynde et al. 2009] or underestimate [Holness et al. 1995] the prevalence of disease when compared with a physician examination. The use of picture based questionnaires and selfreported doctor-diagnosed dermatitis may provide a prevalence estimate closer to that of physician diagnoses, but may also underestimate prevalence [Smit et al. 1992]. Skin symptoms may be due to irritant or different immunologic (Type I or Type IV) mechanisms. Though it is possible to differentiate between these outcomes in the clinical setting, it is not possible to differentiate using symptoms reported on the questionnaire alone. The strong relationship between wheat specific IgE and work-related itchy skin supports a role for the IgE mediated (Type I) allergy in the development of work-related skin symptoms in bakery workers.

95 76 Parallel results for respiratory symptoms (Supplemental Material) also demonstrate strong relationships between wheat specific IgE and both asthma-like symptoms and work-related chest tightness. It is not possible to model the potential role of Type IV allergy or irritant mechanisms in symptom development in this study. The bell shaped (non-linear) distribution observed for non-atopic auto body shop workers in the smoothing splines (Figure 3) may be the result of a healthy worker effect, with fewer symptomatic subjects at the higher exposure levels. The negative association between exposure and atopy in both the auto body shop and bakery workers also suggests a healthy worker effect (Table 14). The prevalence of work-related allergen specific sensitization was five times higher in bakery workers (11%) compared to auto body shop workers (2%). The low prevalence of HDI specific- IgE sensitization is well documented in other studies and is commonly interpreted as indicating mechanisms other than IgE sensitization are responsible for the development of symptoms in exposed workers [Maestrelli et al. 2009, Wisnewski. 2007]. Atopy and work-related sensitization were strongly associated in both auto body shop workers (PR 13.8, 95% CI ) and bakery workers (PR 2.62, 95% CI ). The correlation between these two variables necessitated caution when offering both variables to the same model. Models where adjustment for atopy and specific sensitization was desired were first constructed separately and estimates were compared with those from models including both variables. In the end, estimates from the separate models were comparable and both variables were offered into all of the combined models. In general, auto body shop workers tended to report more respiratory symptoms, while bakery workers tended to report more skin symptoms. This could be due, in part, to differences in exposure prevention activities. Unfortunately self-reported use of personal protective equipment was only available for auto body shop workers, preventing a comparison of this effect. Observations by the researchers in the field suggest that differences did exist between the two populations, specifically that bakery workers did not use hand or respiratory protection while auto body shop workers tended to use both. A significant exposure-response relationship was

96 77 observed in the auto body shop workers, the group observed to use PPE, suggesting that in these workers PPE use did not reduce exposure to a level that was trivial with respect to health effects. Estimates of airborne exposure were used in the exposure-response models as a crude proxy for skin exposure, so results should be interpreted as airborne exposure-skin symptom associations. It is plausible that the airborne exposure estimates provide a good surrogate of skin exposure. Results from previous studies have shown a relatively strong association between skin and airborne exposures in auto body shop workers [Fent et al. 2008, Liljelind et al. 2010]. No reports comparing skin and airborne exposures in bakery workers were located. It is possible that airborne exposure may be a better surrogate for skin exposure in the auto body shops, resulting in less exposure misclassification among auto body shop workers compared to bakery workers. It may also be that average diisocyanate exposure (µg-nco*m-3), or another exposure which was correlated with diisocyanates, was the causal exposure for skin symptoms in auto body shop workers, but that an exposure other than average wheat exposure (µg-wheat*m-3) was responsible for skin symptoms among bakery workers (i.e., wet work, oils etc.). Despite the observed associations between atopy, specific sensitization and skin symptoms the exposure-response relationships remained unchanged in sensitivity analyses. When atopic and specifically sensitized subjects were excluded from the models the exposure-response relationships for skin symptoms in auto body shop workers persisted and the effect estimates were not attenuated. This provides support for the existence of an exposure-response relationship between NCO exposure and skin symptoms (work-related and non-work-related) in auto body shop workers. In the second analysis, reported skin symptoms were predictive of reporting respiratory symptoms in both occupational groups regardless of the symptom combination, an association that has rarely been investigated [Lynde et al. 2009]. Results were unchanged after adjustment for age, sex, smoking and atopy. The persistence of the association after adjustment for these variables suggests that there are other factors that lead to the co-existing skin and respiratory symptoms (i.e., exposure). These results highlight the importance of considering both skin and respiratory outcomes in exposed workers as well as the importance of properly assessing both skin and airborne exposure in the workplace.

97 78 In conclusion, reporting skin symptoms was strongly and consistently associated with reporting respiratory symptoms in both bakery and auto body shop workers. Additionally, exposureresponse relationships for skin symptoms were observed in auto body shop workers, similar relationships for work-related skin symptoms in bakery workers did not reach statistical significance. There are several reasons why an association may have been missed in bakery workers, including poor correlation between airborne and skin exposure for the particulate exposure and the lack of information on other, potentially causal, exposures in the workplace. The lack of observed association in bakery workers should be interpreted cautiously; exposureresponse relationships for skin symptoms require more investigation in all occupations. These relationships must be better understood before more complex relationships are investigated, however the overall goal remains the reduction of both airborne and skin exposure.

98 79 Chapter 6 Skin and Respiratory Symptoms Among Workers with Suspected Work-Related Disease Victoria H. Arrandale 1, Irena Kudla 2, Allen G. Kraut 3, Jeremy A. Scott 1, Susan M. Tarlo 1,2,4, Carrie A. Redlich 5, D. Linn Holness 1,2 1 University of Toronto, Toronto, Ontario, CANADA 2 St. Michael s Hospital, Toronto, Ontario, CANADA 3 Faculty of Medicine, University of Manitoba, Winnipeg, Manitoba, CANADA 4 Toronto Western Hospital, Toronto, Ontario, CANADA 5 Yale University School of Medicine, New Haven, Connecticut, USA This manuscript is currently under review by Occupational Medicine (Oxford).

99 Abstract Background: Many workers are exposed to chemicals that can cause both respiratory and skin responses. Although there has been much work on respiratory and skin outcomes individually, there are few published studies examining lung and skin outcomes together. The objective of this study was to identify predictors of reporting concurrent skin and respiratory symptoms in a clinical population. Methods: Patients with possible work-related skin or respiratory disease were recruited. An interviewer-administered questionnaire collected data on skin and respiratory symptoms, health history, smoking habits, workplace characteristics and occupational exposures. Predictors of concurrent skin and respiratory symptoms were identified using multiple logistic models. Results: In total, 204 subjects participated; 46% of the subjects were female and the mean age was 45.4 years (sd 10.5). Most subjects (n=167, 82%) had possible work-related skin disease, compared to only 37 (18%) subjects with possible work-related respiratory disease. Subjects with a history of eczema (OR 3.68, 95% CI ), those from larger (>499 employees) workplaces (OR 2.82, 95% ) and those reporting respirator use at work (OR 2.44, 95% CI ) had significantly greater odds of reporting both work-related skin and respiratory symptoms. Current smoking was also associated with reporting concurrent skin and respiratory symptoms (OR 2.57, 95% CI ). Conclusions: Workers do report symptoms in both systems; this may be under-recognized both in the workplace and in the clinic. The association between history of doctor-diagnosed eczema and concurrent skin and respiratory symptoms suggests role for impaired barrier function and this finding needs further investigation.

100 Introduction Many workers are exposed to chemicals that can cause lung and skin responses as a result of both inhalation and skin exposure. Although there has been significant work focused on individual lung or skin outcomes and their association with exposures, there are few published studies examining both airborne and skin exposures with lung and skin outcomes together. This is important clinically, as disease in the one system may be under-recognized when workers are assessed by physicians who are specialist in the other system. It is also important from a prevention standpoint, because opportunities for exposure control and prevention may be missed if research focuses on either airborne or skin exposure. There is also increasing interest in the role that skin exposure may play in sensitization and the development of respiratory symptoms and/or disease [Day et al. 2006, Redlich. 2010]. In humans, the role of skin exposure in the development of occupational asthma has been examined almost exclusively in diisocyanate exposure scenarios [Petsonk et al. 2000, Bello et al. 2008] though studies in animal models suggest this pathway may exist for other exposures [Vanoirbeek et al. 2006, Arts et al. 2004]. Animal studies demonstrate that in order for the skin to act as a relevant route of sensitization in the development of occupational asthma, both a skin and an inhalation exposure are required [Vanoirbeek et al. 2004, Kuper et al. 2011]. It is also known that occupational airborne and skin exposures may be correlated [Fent et al. 2008, Liljelind et al. 2010]. The skin exposure model proposed by Schneider et al. suggests that there is a compartmental connection between airborne and skin exposure based on contribution of one to the other (and vice versa) [Schneider et al. 1999]. This connection means it is plausible that when the contribution of airborne exposure to skin exposure via deposition is high, the two exposures (airborne and skin) will be correlated. In addition to correlations between skin and airborne exposure levels, there is evidence that several common contact sensitizers are also associated with occupational asthma [Arrandale et al. 2012]. Individually, exposure-response relationships have been reported between occupational exposure and skin symptoms [de Joode et al. 2007, Sripaiboonkij et al. 2009a], though studies of exposure-response for respiratory symptoms are far more common. There has been little research on causal exposures or risk factors for reporting concurrent skin and respiratory symptoms,

101 82 despite a number of case studies of workers with allergic contact dermatitis and occupational asthma in response to the same occupational exposure [Moulin et al. 2009, De Raeve et al. 1998]. Given that some workers are likely to have both skin and inhalation exposures, and that these exposures may be correlated, it is plausible that some workers may experience both skin and respiratory symptoms. Aside from one clinical study and one occupational study, this issue has received little attention in previous research [Moulin et al. 2009, Lynde et al. 2009]. The primary objective of this study was to estimate the prevalence of concurrent skin and respiratory symptoms in a clinical population, and to determine how workers with concurrent skin and respiratory symptoms differ from those with symptoms in only one system (skin or respiratory). 6.3 Methods Subjects were recruited consecutively from a hospital-based outpatient occupational health clinic between July 2009 and June Patients with possible work-related respiratory disease were seen through the Allergy/Asthma (AA) clinical stream and those with possible work-related skin disease were seen through the Dermatology (Derm) clinical stream; both were eligible to participate. In both streams, patients were either referred through the public health care insurance system (Ontario Health Insurance Plan - OHIP), or through the workers compensation insurance system (Workplace Safety Insurance Board - WSIB). Patients provided written informed consent. An interviewer-administered questionnaire was completed during the clinic visit by one of two trained interviewers. If there was insufficient time to complete the questionnaire, subjects were given the option of completing the questionnaire via telephone. The questionnaire contained questions on respiratory symptoms (modified ATS questionnaire [Ferris. 1978]), skin symptoms, health history, smoking history, workplace characteristics and workplace exposures. Following the clinic visit(s), the final physician diagnosis was abstracted from the subjects medical chart.

102 Outcome Variables The main outcome of interest was concurrent, skin and respiratory symptoms. This was conceptualized as subjects who reported any respiratory symptom (any of cough, phlegm, wheeze, shortness of breath, or chest tightness) as well as current skin rash. The portion of subjects reporting concurrent skin and respiratory symptoms that were both work-related was also determined. In descriptive analysis, both the prevalence of current hand or arm rash as well as the prevalence of asthma symptoms (defined as wheeze accompanied by shortness of breath but without a cold/flu occurring in the last 12 months [Pekkanen et al. 2005]) were also investigated Predictor Variables Variables considered as possible predictors of concurrent symptoms included: history of asthma, history of eczema, workplace size, use of gloves at work, use of a respirator at work, occupation, insurance scheme (workers compensation vs. public system), clinical stream (asthma/allergy vs. dermatology), presence of a union in the workplace, workplace education on personal protective equipment, workplace education on occupational disease(s) and a series of possible workplace exposures (cement, dander, dust, fumes, isocyanates, paints, pesticides, wet work and others). Age and sex were reported on the questionnaire. Smoking included any reported tobacco smoking. Atopy was defined based on subjects positive responses to having allergies to dust, dust mites or other animals, or having had doctor-diagnosed hay fever [Lakwijk et al. 1998]. Previous asthma and eczema were recorded as a positive response to both have you ever had asthma/eczema? and, was it confirmed by a doctor? Workplace size was categorized into four groups based on the number of employees (<20, 20-99, and >500) [Industry Canada. 2011]. Self-reported job title and industry information was coded using the National Occupational Classification System (NOCS) [Statistics Canada. 2011]. Occupations were further condensed into 6 groups (trades, sales/service, office, manufacturing, health, other) due to small sample sizes in the major NOCS groupings. Workplace education variables included education on skin and/or respiratory personal protective equipment and occupational disease.

103 Reliability Testing We also measured the test-retest reliability of the workplace characteristics and workplace exposure sections of a detailed occupational hygiene questionnaire that was already used in the clinic setting. Subjects who consented to participate in further research were selected randomly and contacted by telephone. Subjects were required to have no significant changes at their job since first completing the questionnaire in order to participate. Results of the test-retest reliability portion of this study are shown in Appendix Statistical Analyses Differences between participants and patients who refused to participate were tested using chi square for categorical variables and Student s t-test continuous variables. Cohen s Kappa, percent observed agreement, percent positive agreement and percent negative agreements were calculated to assess the reliability of the workplace questionnaire. Possible predictors of concurrent symptoms were first investigated using simple logistic regression. The outcome in these models was concurrent skin and respiratory symptoms; the comparison group was subjects who reported either skin or respiratory symptoms but not both. Predictors where the parameter estimate in simple logistic regression results had a p < 0.20 were offered into multiple logistic models. Multiple logistic regression models were adjusted for age, sex, smoking, atopy and interviewer. A nested model comparing subjects with respiratory symptoms only to subjects with skin symptoms only was also constructed in the same manner. All analyses were completed in SAS v.9 software (SAS Institute Inc., Cary, NC, USA). The study was approved by the St Michael s Hospital Research Ethics Board (Toronto, ON, Canada). 6.4 Results In total, 218 subjects were successfully recruited with a response rate of 81%. Figure 5 shows a flow chart of the study progression with the sample sizes at each stage. Thirty-three patients were not approached at the discretion of the staff/physician or due to higher than usual activity in the clinic. Fourteen subjects were excluded from analyses as they reported neither skin nor respiratory symptoms. The sample size for the reported analysis is 204 subjects.

104 85 Figure 5 Flow Chart of Study Progression, Including Sample Sizes at Each Stage. Patients who refused study participation did not differ from the participants in terms of age or sex (Appendix 3) but those who declined participation were more likely to be from the AA clinical stream (32% vs. 13%, p = ) (Appendix 3). Patients who declined participation also tended to be referred more often through the provincial worker s compensation system rather than the main public health care system (21% vs. 13%, p = 0.07) though this difference did not reach statistical significance. Of the total 204 subjects, most (n = 167, 82%) had possible work-related skin disease and were seen through the dermatology (Derm) stream; only 37 (18%) subjects were seen in the asthma/allergy (AA) stream with possible work-related respiratory disease. This distribution of participants between the Derm and AA streams is reflective of the overall patient breakdown in the clinic. The participation rates were 86% and 68% for the Derm and AA streams, respectively. Twenty-two (59%) AA subjects and 117 (70%) Derm subjects were diagnosed with work-related disorders after their assessment. An additional 4 (11%) AA subjects and 31 (19%) Derm subjects were diagnosed with a possible work-related disorder. Just under half (46%) of the subjects were female and the mean age was 45.4 years (sd 10.5) (Table 17). Fifty-one percent of subjects had a smoking history, either former or current. Almost one-quarter of the study population selfreported doctor-diagnosed asthma (Table 17), though asthma was more common in the AA stream (40% vs. 11%, p < ).

105 86 Table 17 Demographic Description of Study Population, Stratified by Subjects Who Reported Both Skin and Respiratory Symptoms. Overall Values Reported as Column n (%), Symptom Values Reported as the Row n (%). Comparison between Both Skin and Respiratory Symptoms and Skin Symptoms Only or Respiratory Symptoms Only. NS Not Significant (p > 0.05). Overall Skin Symptoms Only or Respiratory Symptoms Only Both Skin and Respiratory Symptoms n Female 94 (46%) 59 (63%) 35 (37%) Male 110 (54%) 62 (56%) 48 (44%) p-value Age in years, mean (sd) 45.4 (10.5) 45.7 (10.0) 45.1 (11.2) NS Age <35 years 38 (19%) 20 (53%) 18 (47%) Age years 88 (43%) 55 (62%) 33 (38%) Age 50 years 78 (38%) 46 (59%) 32 (41%) Never Smoker 98 (48%) 64 (65%) 34 (35%) Former Smoker 50 (24%) 31 (62%) 19 (38%) Current Smoker 56 (27%) 26 (46%) 30 (54%) Interviewer (56%) 69 (60%) 46 (40%) Interviewer 2 89 (44%) 52 (58%) 37 (42%) Telephone Questionnaire 5 (2%) 2 (40%) 3 (60%) In Clinic Questionnaire 198 (98%) 118 (60%) 80 (40%) AA Clinical Stream 37 (18%) 31 (84%) 6 (16%) Derm Clinical Stream 167 (82%) 90 (54%) 77 (46%) Workplace Insurance 91 (45%) 59 (65%) 32 (35%) Public Insurance 113 (55%) 62 (55%) 51 (45%) NS NS NS NS NS Atopy 68 (33%) 42 (62%) 26 (38%) NS History of Asthma 50 (24%) 29 (58%) 21 (42%) NS History of Eczema 54 (26%) 20 (37%) 34 (63%) NS

106 87 Among the study subjects, symptom prevalence was high, as expected in a clinical population (Table 18). General concurrent skin and respiratory symptoms were reported by approximately one-third of the subjects (Table 18) but work-related concurrent symptoms were only reported by 18% of subjects. Among the subjects being assessed for possible work-related skin disease (Derm stream), 30% reported at least one work-related respiratory symptoms and 6% reported work-related asthma symptoms. Work-related skin symptoms were less common in the AA stream, but were still reported (11% work-related rash, 8% work-related hand/arm rash) (Table 18). Table 18 Skin and Respiratory Symptom Prevalence, Stratified by Clinical Stream. All Frequencies Reported as Column n (%). Clinical Stream Overall Dermatology Asthma/Allergy n General Symptoms: Current Skin Rash 146 (72%) 140 (84%) 6 (16%) Hand/Arm Rash 134 (66%) 130 (78%) 4 (11%) Any Respiratory Symptom 141 (69%) 104 (62%) 37 (100%) Asthma-like Symptoms 48 (24%) 24 (14%) 24 (65%) Work-Related Symptoms: Current Skin Rash 129 (63%) 125 (75%) 4 (11%) Hand/Arm Rash 120 (59%) 117 (70%) 3 (8%) Any Respiratory Symptom 83 (41%) 50 (30%) 33 (89%) Asthma-like Symptoms 33 (16%) 11 (6%) 22 (59%) Concurrent Skin and Respiratory Symptoms 83 (41%) 77 (46%) 6 (16%) Concurrent Work-Related Skin and Respiratory Symptoms 40 (20%) 36 (22%) 4 (11%)

107 88 Reported glove use at work was higher (90%) than the reported use of a respirator (39%) (Table 19). More than half of the subjects (53%) reported having had workplace education about personal protective equipment, but only 14% reported having education that addressed the topic of work-related disease. Dust was the most commonly reported occupational exposure (68%); exposure to wet work (65%), fumes (55%) and paint (41%) were also common exposures (Table 20). None of the selfreported exposures were associated with reporting concurrent skin and respiratory symptoms. Table 19 Self-Reported Workplace Characteristics, Stratified by Subjects Who Reported Both Skin and Respiratory Symptoms. Overall Values Reported as Column n (%), Symptom Values Reported as the Row n (%). Comparison between Both Skin and Respiratory Symptoms and Skin Symptoms Only or Respiratory Symptoms Only. NS Not Significant (p > 0.05). Skin Symptoms Overall Only or Respiratory Symptoms Only Both Skin and Respiratory Symptoms p-value Unionized 98 (48%) 62 (63%) 36 (37%) NS < 20 Employees 52 (25%) 31 (60%) 21 (40%) NS Employees 57 (28%) 38 (67%) 19 (33%) Employees 52 (25%) 34 (65%) 18 (35%) >499 Employees 43 (21%) 18 (42%) 25 (58%) Gloves Use 183 (90%) 105 (57%) 78 (43%) NS Respirator Use 80 (39%) 39 (49%) 41 (51%) Trades 49 (24%) 28 (57%) 21 (43%) NS Sales and Service 42 (20%) 25 (60%) 17 (40%) Manufacturing 38 (19%) 25 (66%) 13 (34%) Health Related 30 (15%) 12 (40%) 18 (60%) Office Workers 27 (13%) 19 (70%) 8 (30%) Other Occupations 18 (9%) 12 (67%) 6 (33%) PPE Education 108 (53%) 59 (55%) 49 (45%) NS WR Disease Education 28 (14%) 14 (50%) 14 (50%) NS

108 89 Table 20 Self-Reported Workplace Exposures, Stratified by Subjects Who Reported Both Skin and Respiratory Symptoms. Overall Values Reported as Column n (%), Symptom Values Reported as the Row n (%). Comparison between Both Skin and Respiratory Symptoms and Skin Symptoms Only or Respiratory Symptoms Only. NS Not Significant (p > 0.05). Overall Skin Symptoms Only or Respiratory Symptoms Only Both Skin and Respiratory Symptoms p-value Cement 41 (20%) 21 (51%) 20 (49%) NS Animal Dander 25 (12%) 14 (56%) 11 (44%) NS Dust 139 (68%) 85 (61%) 54 (39%) NS Fume 113 (55%) 69 (61%) 44 (39%) NS Isocyanate 32 (16%) 17 (53%) 15 (47%) NS Paint 84 (41%) 52 (62%) 32 (38%) NS Pesticide 22 (11%) 13 (59%) 9 (41%) NS Wet Work 133 (65%) 78 (59%) 55 (41%) NS Concurrent Symptoms Models were constructed to compare subjects with concurrent skin and respiratory symptoms to those with symptoms in only one system. Larger than acceptable variance inflation factors for occupation variable meant that occupation did not remain into the multiple logistic regression models. The results of the multiple logistic regression models are shown in Table 21. Results showed that subjects with a history of eczema (OR 3.68, 95% CI ) had significantly greater odds of reporting both work-related skin and respiratory symptoms. Additionally, subjects from larger (>499 employees) workplaces (OR 2.82, 95% ) and those who reported wearing a respirator while at work (OR 2.44, 95% CI ) were more likely to report concurrent skin and respiratory symptoms. Current smoking was found to be associated with reporting concurrent skin and respiratory symptoms (OR 2.57, 95% CI ). In the nested model comparing subjects with respiratory symptoms only to subjects with skin

109 90 symptoms only smoking was not significantly associated with reporting respiratory symptoms (Table 21). In this model a history of doctor-diagnosed eczema was associated with reporting skin symptoms only (OR 0.28, 95% CI ). Table 21 Multiple Logistic Regression Model Results for Predictors of Reporting Concurrent Skin and Respiratory Symptom Outcomes. Models Adjusted for Age, Sex and Interviewer. Predictor Both Skin and Respiratory Symptoms Model Description Respiratory Symptoms Only vs. Skin Symptoms Only Never Smoker Former Smoker 1.22 ( ) 1.36 ( ) Current Smoker 2.57 ( ) 1.25 ( ) Atopy 0.86 ( ) 2.49 ( ) Doctor-diagnosed Eczema 3.68 ( ) 0.28 ( ) < 20 Employees 0.96 ( ) Employees Employees 1.00 ( ) - >499 Employees 2.82 ( ) - AA Clinical Stream (vs. Derm) 0.25 ( ) - Respirator Worn at Work 2.44 ( ) - Isocyanate Exposure at Work ( ) # Outcomes Model n Discussion Concurrent symptoms were more common among the Derm stream subjects (46%) compared with the AA stream subjects (16%), but both groups had a higher prevalence of concurrent skin and respiratory symptoms than previous studies [Moulin et al. 2009, Lynde et al. 2009]. In another clinical study, Moulin et al. reported on 234 patients with diagnosed contact dermatitis, of whom 10 (4%) had both work-related contact dermatitis and respiratory symptoms [Moulin et al. 2009]. This is one-fourth the prevalence observed here in the Derm patients; 88% of whom

110 91 were diagnosed with contact dermatitis. Lynde et al. studied working professional cleaners but required more than one respiratory symptom to be reported [Lynde et al. 2009]. Lynde et al. found that 7.2% reported both a current skin rash and two or more respiratory symptoms; 5.9% reported current skin rash as well as two or more work-related respiratory symptoms [Lynde et al. 2009]. Work-related symptoms were common, as would be expected in a clinical population being assessed for possible work-related disease. A large proportion of workers also reported workrelated symptoms in the system other than the one they were being assessed for; 30% of Derm stream subjects reported work-related respiratory symptoms and 11% of AA subjects reported work-related skin rash. In models exploring the difference between subjects with concurrent symptoms (both skin and respiratory symptoms) and those with only one symptom (either skin or respiratory symptoms) none of the specific exposures investigated were significant risk factors for concurrent symptoms. A history of doctor-diagnosed eczema was a risk factor for reporting concurrent skin and respiratory symptoms, but a history of asthma was not. Childhood eczema has been reported as a risk-factor for developing adult onset asthma [Martin et al. 2011], but the relationship between eczema and concurrent symptoms is less clear. Studies by Vermeulen et al. in rubber workers and Hino et al. in car spray painters found that subjects with abnormal skin (mild dermatitis or hand eczema) had elevated biomarkers of exposure, suggesting that they had greater uptake of exposure through their skin [Vermeulen et al. 2003, Hino et al. 2008]. In a clinical population with atopic dermatitis, Bremmer et al. reported that patients who also had ichthyosis vulgaris, a skin disease that disrupts the barrier function of the skin, were significantly more likely to report asthma symptoms [Bremmer et al. 2008]. The reported association between eczema and concurrent skin and respiratory symptoms suggests that impaired barrier function may play a role in modifying the uptake of exposures through the skin, and potentially also in the development of both skin and respiratory symptoms [Nielsen. 2005]. However, the data available in this study does not address the extent (if any) of barrier function impairment in the current study subjects, or the location of the eczema and its relevance to occupational exposures.

111 92 Atopy was considered a possible confounder in the relationship between eczema and symptoms and was included in all models. The observed association between a history of eczema and concurrent skin and respiratory symptoms persists in the model after adjusting for atopy, suggesting that the association is not a result of confounding. The association between large workplaces and reporting concurrent skin and respiratory symptoms is more challenging. There may be social factors (i.e., a lack of a personal relationship with their employer) that make workers from larger workplaces more likely to report symptom as associated with their work, but these underlying factors cannot be determined from this study. Neither having a union in the workplace nor receiving occupational health and safety education (PPE or work-related disease) was associated with reporting concurrent skin and respiratory symptoms, though both were more likely in larger workplaces (results not shown). No associations between workplace exposures and reporting concurrent skin and respiratory symptoms were observed. This was a small study population with diverse occupations and exposures that lacked power to explore the relationship between single exposures and the concurrent symptom outcome. However, subjects reporting respirator use at work were observed to have significantly greater odds of reporting concurrent skin and respiratory symptoms. In this case, the use of a respirator at work may serve as proxy for general exposure at work. The observed association may suggest an association between higher exposure, or perhaps higher risk exposure, and concurrent symptoms. It is also possible that, regardless of the other workplace exposures, the respirator is actually an important exposure. In cases of contact dermatitis and patch testing, gloves, and the chemical components of gloves, are often implicated as causal occupational exposures, this is less common in the case of respirators but it is still plausible [Warshaw et al. 2008]. Current smokers had increased odds of concurrent skin and respiratory symptoms. Interestingly in the nested model, which compared subjects with respiratory symptoms only to subjects with skin symptoms only, smoking was not associated with reporting respiratory symptoms. This is surprising because the association between smoking and respiratory symptoms is well established [Higgins. 1959]. In this nested model subjects with atopy were more likely to report respiratory symptoms while subjects with a history of eczema were more likely to report skin symptoms.

112 Limitations The main limitation of this study is that it was cross-sectional and able only to describe association rather than causation. Secondly, this study was completed in a selective clinical population of patients with suspected work-related disease and may not be generalizable to all workers. The aim of this study was to determine if workers who have concurrent skin and respiratory symptoms are different from those with symptoms only in one system that was more easily addressed in a population with higher symptom prevalence. The mechanism(s) underlying the reported symptoms cannot be ascertained from this study. The reported symptoms could be due to either allergic or irritant mechanisms; in the case of concurrent symptoms it is possible that one symptom is allergic in nature, while the other is irritant. Additionally the relative infancy of dermatological epidemiology is a limitation of this research area. Respiratory epidemiology is well-developed, with established questionnaire tools and an accepted understanding of which reported symptoms relate to formal diagnoses [Pekkanen et al. 2005, Ferris. 1978]. The epidemiology of skin disease is significantly less developed; much work remains to be done Conclusions In conclusion, while no specific exposures were found to predict the reporting of concurrent skin and respiratory symptoms, workers who reported respirator use at work and those who worked in larger workplaces had increased odds of reporting concurrent skin and respiratory symptoms, as did current smokers and subjects with a history of eczema. The association between a history of doctor-diagnosed eczema and concurrent skin and respiratory symptoms, including the potential role of impaired barrier function in the development of concurrent symptoms is interesting and needs further investigation. Future occupational research should aim to include measures of both skin and respiratory exposures and outcomes, particularly in studies of occupational exposures that are known to cause effects in both systems.

113 94 Chapter 7 General Discussion The broad aim of this thesis, as stated in Section 2.2, was to further investigate the relationships between occupational exposures, skin symptoms and disease, and respiratory symptoms and disease. This overarching purpose was achieved through four studies, which investigated the connection between the skin and respiratory systems in terms of exposure, response (symptoms or disease), and exposure-response relationships. 7.1 Revisiting Research Aims and Hypotheses Data available from surveillance reporting schemes (both mandatory and voluntary) provides information on the causes of occupational contact dermatitis and occupational asthma in reported cases [Karjalainen et al. 2000, Riihimäki et al. 2004, McDonald et al. 2005, Hannaford-Turner et al. 2010, Pal et al. 2009, McDonald et al. 2006, Turner et al. 2007]. We hypothesized that in a systematic review of common occupational contact allergens there would be many that could also cause work-related asthma. Results from the North American Contact Dermatitis Group (NACDG) patch test data and the subsequent systematic literature review demonstrated that this hypothesis was supported (Chapter 3). Seven of the ten most common contact allergens from the NACDG data ( ) were also found to be associated with OA. Though the hypothesis was shown to be correct, there is still an important gap in knowledge, as this study reported only on the ability for contact allergens to potentially also cause asthma. It did not address the potential for the most common causes of occupational asthma to also cause contact dermatitis. Identifying the common causes of occupational asthma in the same general population is more challenging as there is no structure in place for collecting information on causes of occupational asthma in Canada. Additionally, the diagnosis of occupational asthma is not as clear-cut as the diagnosis of occupational contact allergy. The need for better surveillance and record keeping around causes of occupational asthma will be addressed in this discussion. While investigating the common occupational contact allergens in the NACDG data, conflicting notations regarding the sensitization potential of the common contact allergens was discovered in three common sources of occupational hygiene information. This discordance will also be described in further detail.

114 95 The proposed framework for conceptualizing possible connections between the skin and respiratory systems (Figure 1) led to the hypothesis that, when asked, some workers would report both skin and respiratory symptoms. And, that a portion of these workers (with concurrent skin and respiratory symptoms) would report both symptoms as being associated with their job (work-related). Results from a small data pooling study suggested that this hypothesis was also supported (Chapter 4). Historical data from four occupational studies (n=247) that focused on skin or respiratory symptoms (or both) were pooled and the prevalence of concurrent skin and respiratory symptoms was calculated. The mean prevalence of concurrent symptoms was 11% (range by occupation: 6-17%) though only two subjects (<1%) reported both skin and respiratory symptoms that were both work-related. There is little work on underlying exposure-response relationships that lead to skin symptoms despite that fact that exposure-response is considered to be one of the necessary conditions for demonstrating causation [Hill. 1965]. In order to better understand the exposure-response relationships for skin symptoms, the fourth research aim was to determine whether exposureresponse relationships could be observed in two different occupational groups: bakery workers and auto body shop workers (Chapter 5). Both groups have occupational exposures that are known to cause asthma and contact dermatitis. In both cases there has been research into the prevalence of skin symptoms, but little research into exposure-response relationships. The hypothesis was that exposure-response relationships for skin symptoms would be observed in both of these occupational groups. Results from the analyses suggested that the hypothesis was not supported; exposure-response relationships were observed in auto body shop workers, but not in bakery workers. There are many explanations for why an exposure-response relationship among the bakery workers may not have been observed, if in fact one exists. Wheat exposure was the exposure estimated in the study of bakery workers, but wheat may not have been the causal exposure. The study also used airborne exposure as a proxy for skin exposure, which may have led to exposure misclassification if there was low correlation between skin and airborne exposure to wheat. These limitations will be discussed further. The final section (Chapter 6) of this thesis investigated the prevalence and predictors of concurrent skin and respiratory symptoms in a clinical population with suspected occupational disease. The first hypothesis investigated in this study was the same as in the data pooling study (Chapter 4), namely that subjects would report concurrent skin and respiratory symptoms and

115 96 that a portion would associate these symptoms with their job (work-related). We expected that the prevalence of concurrent symptoms would be higher in the clinical population than in the pooled occupational studies (Chapter 4). With this group we were able to explore possible predictors of concurrent symptoms, including health history, workplace characteristics, prevention practices, and occupational exposures. We hypothesized that there would be differences between subjects who reported only one symptom (either skin or respiratory), and those who reported both skin and respiratory symptoms in these variables. Results from the study of workers with suspected disease (n=204) confirmed that a relatively large portion, 41%, reported both skin and respiratory symptoms, and that 20% reported both skin and respiratory symptoms that were work-related. Workers with a history of eczema as well as workers from larger workplaces and those who reported wearing a respirator at work had increased odds of reporting concurrent skin and respiratory symptoms. No specific occupational exposures were associated with reporting concurrent skin and respiratory symptoms. The importance of the observed relationship between eczema and concurrent symptoms will be discussed in greater detail. There were a series of research aims that guided the work contained in this thesis. These aims were simple in nature, reasonable, and manageable. Individually, they each contribute a small piece of knowledge to an incomplete picture of possible connections and interactions between the skin and respiratory systems in occupational disease. Specifically, these results confirm that some workplace exposures can cause both skin and respiratory disease, that workers do experience both skin and respiratory symptoms at the same time, and that there are observable exposure-response relationships for skin symptoms in some working populations. Together the findings from this thesis support the broad hypothesis that there are connections between the skin and respiratory systems in terms of occupational exposure and disease, and that this area of research both needs and deserves further attention. 7.2 Methodological Considerations Beyond the specific research findings from this thesis, there were two main themes that emerged from the work. The first theme was the existence of silos in respiratory research and dermatology/skin research. Second was the challenge in bridging this gap in studies where the outcome was actually two different, yet concurrent outcomes.

116 97 At a basic level there were differences in the history and body of literature on occupational skin and respiratory disease. There has been a large amount of effort in the past to standardize the questionnaire tools related to respiratory symptoms and disease [Jenkins et al. 1996, Sistek et al. 2006, Pekkanen et al. 2005, Vandenplas et al. 2005, Medical Research Council on the Aetiology of Chronic Bronchitis. 1960, Ferris. 1978]. Researchers have also spent a considerable amount of time investigating the relationships between specific symptoms and respiratory diagnoses (e.g., wheeze, asthma and bronchial hyper-responsiveness) which permits the use of questionnaire items in place of diagnostic testing in research studies [Jenkins et al. 1996, Sistek et al. 2006, Pekkanen et al. 2005, Vandenplas et al. 2005, Medical Research Council on the Aetiology of Chronic Bronchitis. 1960, Ferris. 1978]. There is currently no equivalent body of literature relating skin symptoms to skin diagnoses. Several studies have confirmed that skin symptoms are positively related to physician diagnosis of skin disease. However, these studies used different questionnaire items to predict different clinical outcomes [Svensson et al. 2002, Meding and Barregard. 2001, Carstensen et al. 2006, Smit et al. 1994]. There is a need for more methodological work to better determine if specific skin symptoms questions are predictive of dermatological diagnoses, and if it is possible to use questionnaire items to differentiate between different dermatological diagnoses (i.e., irritant/allergic contact dermatitis). As the studies of concurrent skin and respiratory symptoms and disease were undertaken, it became clear that conceptualizing the concurrent outcomes (skin and respiratory), and addressing the work-relatedness of two outcomes at once, was more complicated than originally considered. The outcomes of interest were two separate reported symptoms, but either, both or neither could be work-related. This resulted in nine possible outcomes groups that are described with respect to each symptom and its work-relatedness in Table 22. Separating subjects who had concurrent symptoms (skin and respiratory) was a straightforward task, but when the layer of work-relatedness was added this forced a decision about whether one work-related symptom (of two reported) was sufficient, or whether both reported symptoms needed to be work-related. In either case, when selecting a definition of work-related concurrent symptoms, there was always a group of subjects (group 6 from Table 22) who were placed in the reference group despite having concurrent skin and respiratory symptoms. This meant that any

117 98 comparison between work-related concurrent symptoms and those without was no longer a comparison between concurrent symptoms and symptoms only in one system as was originally described in the research objectives. It was decided that both symptoms were required to be work-related for membership in the work-related concurrent symptom group (Appendix 3, Table 30). This decision was made in part because the difference between the group with both symptoms (n=83) and the group with both symptoms where only one symptom was work-related (n=77) were small (Table 18) and because the interpretation was simply more straightforward when the concurrent symptoms were both required to be work-related. In the end the models that were included in Chapter 6 focused only on general concurrent symptoms. Table 22 Description of Possible Outcome Groups when Considering Both Skin and Respiratory Symptom Outcomes and their Individual Work-Relatedness. Under Symptoms, X = Symptom is Present in the Symptom Group. Under Model A and Model B, Yes= Symptom Groups are Considered to Have the Outcome of Interest and No = Symptom Groups are Considered Not to Have the Outcome of Interest. Conc = Concurrent, WR = Work-Related, Resp = Respiratory. Symptom Group n Skin Symptoms WR Skin Resp WR Resp Model A: Conc Sx Model B: Both WR Conc Sx (1) No Symptoms (Sx) 14 No No (2) Skin Sx, No Respiratory Sx 7 X No No (3) Work-related Skin Sx, No Respiratory Sx 19 X X No No (4) No Skin Sx, Respiratory Sx 56 X No No (5) No Skin Sx, Work-Related Respiratory Sx 39 X X No No (6) Skin Sx, Respiratory Sx 6 X X Yes No (7) Work-Related Skin Sx, Respiratory Sx 33 X X X Yes No (8) Skin Sx, Work-Related 4 X X X Yes No Respiratory Sx (9) Work-Related Skin Sx, Work- Related Respiratory Sx 40 X X X X Yes Yes

118 Causes of Occupational Skin and Respiratory Disease The results from Chapter 3 demonstrate that seven of the top ten contact allergens were also associated with occupational asthma in the literature or reference materials investigated ( Asthma in the Workplace, UK HSE Asthmagen?, MEDLINE, TOXNET, EMBASE). As emphasized in Chapter 3, it is important to reiterate that we started with a list of contact allergens and investigated their association with OA. The result is a list of contact sensitizers that may also be respiratory sensitizers. Due to the structure of this process there are some common respiratory sensitizers that are missing from the list (e.g., isocyanates, animals, flour). The seven common occupational contact allergens that were found to be capable of causing occupational asthma were: epoxy resin, nickel sulfate, cobalt chloride, potassium dichromate, glutaraldehyde p-phenylenediamine (PPD), and formaldehyde. Thiuram, carba mix, and glyceryl thioglycolate were classified as having no current evidence of causing OA (Table 10). The ten most common occupational contact allergens were similar to previously published studies that reported on the common contact allergens in other regions, including the UK [McDonald et al. 2006], Finland [Riihimäki et al. 2004] and the entire NACDG database [Rietschel et al. 2002]. Though the UK and Finnish surveillance systems use chemical groupings rather than the specific chemical names that are tested in the patch testing process, the similarities between all of the findings is evident. The Canadian NACDG contained carba mix and thiuram, while rubber chemicals were listed in the UK data [McDonald et al. 2006]. Aliphatic aldehydes were in the top ten chemical groups causing allergic contact dermatitis in Finland [Riihimäki et al. 2004]; similarly, formaldehyde and glutaraldehyde were present in the Canadian NACDG data. The consistency of the common allergens among the Canadian centres in the NACDG with the entire NACDG, as well as with other countries suggests that the results from the Canadian centres are not overly sensitive to local industry or case clusters and are good indicators of trends in contact allergy. Additionally, some of the common contact allergens identified in Chapter 3 were the same casual agents implicated in cases studies of workers with both occupational asthma and occupational contact dermatitis (Table 5). There have been reports in the literature of workers with exposure to diglycidyl ether of bisphenol A (DGEBA) a component in epoxy resin systems [Moulin et al. 2009, Kanerva et al. 1991], potassium dichromate [De Raeve et al. 1998], and nickel [Estlander

119 100 et al. 1993]. The combination of peer-reviewed literature supporting the causal link between these contact allergens and occupational asthma, combined with the case studies of occupational contact dermatitis and occupational asthma from the same exposures, provide a good direction for future research of concurrent skin and respiratory outcomes. It would be ideal to complete the reverse process, and investigate common respiratory sensitizers and their association with contact dermatitis. However, there is currently no equivalent to the NACDG patch test data for respiratory sensitizers Surveillance of OCD and OA The history of clinical standardization and data pooling amongst dermatologists specializing in contact dermatitis and patch testing has resulted in a large amount of standardized diagnostic data from the USA, Canada, and Europe that allow for surveillance studies of common contact allergens (both work-related and non-work-related) within these regions [Zug et al. 2009, Uter et al. 2009]. The studies are valuable, and their existence is due in part to the large number of allergens that can be tested at one time in one subject. The North America Contact Dermatitis Group (NACDG) tests a standard tray of 65 allergens, in addition to any occupation-specific allergens the physician deems necessary, including custom allergens in some cases [Zug et al. 2009]. It would be beneficial to have a similar pooling of information on the diagnostic tests used in occupational asthma, but the relative infrequency of specific inhalation challenge testing and the fact that only one allergen is tested at a time in a SIC means that a database would likely contain a fraction of the information. The best data potentially available in Ontario would be workers compensation claims data from the Workplace Safety and Insurance Board (WSIB), but this would limit the exposures to those cases of OA that were accepted for compensation, rather than all work-related asthma at a clinic or population level, which is likely a significantly larger group. A data pooling scheme for SIC testing may be more feasible in jurisdictions where SIC is more common (i.e., the province of Quebec, Canada) but very few workers compensation claims for OA undergo SIC testing in other jurisdictions.

120 101 The development of a system that permitted pooling among the few physicians who complete SIC test would be a huge asset to occupational asthma research in Canada. A voluntary system in the province of Ontario was recently tested for feasibility but the future of this endeavor is unknown [To et al. 2011]. The lessons learned from this experience should be applied to any future attempts to improve the system, or launch a similar program in another jurisdiction. Further to this, the two Canadian centres whose data were utilized in Chapter 3 are part of an American-based organization, the NACDG. An idea championed by Dr. Holness is the development of a Canadian patch test pooling group, similar to the NACDG and the European Surveillance System on Contact Allergies (ESSCA). There are many clinics across the country that offer patch testing, together their pooled data would be a huge asset for skin research, both occupational and non-occupational, in Canada Knowledge Translation and Communication Results from Chapter 3 demonstrated that among identified occupational contact allergens there was significant discordance in how these known sensitizers are described and annotated in common occupational hygiene references materials. Practicing occupational hygienists, occupational physicians, and nurses, as well as other health practitioners often look to the NIOSH pocket guide, the ACGIH TLV book and/or the HazMap database for guidance on the handling of specific chemicals [American Conference of Governmental Industrial Hygienists (ACGIH). 2008, National Institutes of Health. 2009, National Institute for Occupational Safety and Health. 2007]. When the ten most common occupational contact allergens were referenced in each of the three documents, only the HazMap database identified them all as possibly having skin effects. The major issue in interpretation of these results is that the notations provided by the three sources of information are fundamentally different. The NLM HazMap database uses notations for the adverse effects Skin and Asthma as well as potential disease outcomes of Asthma and Contact Dermatitis [National Institutes of Health. 2009]. The ACGIH TLV Handbook includes a specific sensitizer notation, SEN, but this notation does not include information on the route of sensitization through which sensitization may occur [American Conference of Governmental Industrial Hygienists (ACGIH). 2008]. Thirdly, the NIOSH Pocket Guide

121 102 includes possible symptoms as Dermatitis or Asthma as well as notations that are specific to sensitization, Respiratory sensitizer or Skin Sensitizer [National Institute for Occupational Safety and Health. 2007]. To better understand the differences in the notations, the purpose or guiding statement for each document was investigated. The US National Library of Medicine states that: HazMap is an occupational health database designed for health and safety professionals and for consumers seeking information about the health effects of exposure to chemicals and biologicals at work. HazMap links jobs and hazardous tasks with occupational diseases and their symptoms [National Institutes of Health. 2009]. The ACGIH writes that the TLVs are: guidelines designed for use by industrial hygienists in making decisions regarding safe levels of exposure to various chemical substances and physical agents found in the workplace. In using these guidelines, industrial hygienists are cautioned that the TLVs and BEIs are only one of multiple factors to be considered in evaluating specific workplace situations and conditions [American Conference of Governmental Industrial Hygienists (ACGIH). 2008]. Thirdly, the CDC states that the: NIOSH Pocket Guide to Chemical Hazards (NPG) is intended as a source of general industrial hygiene information on several hundred chemicals/classes for workers, employers, and occupational health professionals. The NPG does not contain an analysis of all pertinent data, rather it presents key information and data in abbreviated or tabular form for chemicals or substance groupings that are found in the work environment [National Institute for Occupational Safety and Health. 2007]. Both the ACGIH and the NIOSH pocket guide state that they do not contain all of the relevant data, and should be used in concert with other sources of information when evaluating workplace exposure scenarios. The NLM HazMap database makes no such statement but also employs the

122 103 broadest notations in the reference materials. The broad notation used by the HazMap database does not specify sensitization, rather just specifies skin (skin or contact dermatitis) or respiratory effects (asthma). It is unknown how these reference documents are used in practice; whether occupational hygienists consult only one, or a selection, of the sources of information on hazards. What is clear is that a practitioner who only consulted one of these documents would likely receive an incomplete picture of the exposure hazard. Even though the HazMap database would be more likely to identify a potential hazard, further research and investigation would be required to better identify the specific type of hazard. In Table 28 and Table 29 the reliability statistics for workplace characteristics questionnaire items are reported. It is relevant to this discussion of sensitizer notations to note that in the small reliability study, the least reliable (lowest Kappa) questions were the two asking about allergens or irritants in the workplace. The majority of questionnaire items asking about specific occupational exposures (Table 28 and Table 29) had better reliability than the general questions about allergens and irritants. This suggests that workers may have struggled to categorize their exposures as either allergic or irritants. This may be a result of not understanding the meaning of allergen or irritant when asked during the questionnaire procedure, but also may be due to a lack of knowledge about which exposures are allergens and which are irritants. Given the differing information in the three occupational hygiene documents, it is possible that there is confusion even among the health and safety professionals, and that this confusion and misinformation has trickled down to workers and managers. Workers on the job, occupational hygienists in the field, and health practitioners in the clinic all need to be aware of the potential health hazards associated with a workplace exposure. These groups cannot be required to learn and retain all of this information, but instead need a source of information that provides an up-todate description of workplace exposures, and directs the reader to other sources when necessary. In this digital age, it seems that there should be a way to merge information from multiple sources into one interface, perhaps even one that can be accessed on mobile devices that would allow for on-demand access to health and safety information.

123 Modeling Exposure-Response Relationships Historically, occupational studies of health outcomes have tended to measure either skin or airborne exposures. Studies in which skin and airborne exposures were measured in parallel (Table 4) focused only on exposure assessment and not on potential health outcomes. Many studies have collected information about both skin and respiratory symptoms, but have not reported on the prevalence of these symptoms co-existing [Sripaiboonkij et al. 2009b, Sripaiboonkij et al. 2009a, Lindgren et al. 2002, Fantuzzi et al. 2010, Nettis et al. 2002, Kujala and Reijula. 1995, Holter et al. 2002, Friis et al. 1999, Holness and Nethercott. 1989, Holness et al. 1989, Nethercott and Holness. 1988, Huusom et al. 2011]. Compared with collecting exposure measurements of two routes, including questions on both outcomes is relatively simple and may explain why there is such a wealth of data on symptoms, and a lack of exposure data. As a first step in determining whether there may be a connection between skin and respiratory systems, previously collected data was analyzed to determine whether there was a group of workers who reported concurrent skin and respiratory symptoms. These analyses (Chapter 4) benefitted from a series of previous studies that had asked about both skin and respiratory symptoms but had never analyzed the two together. When the exposure-response relationships for skin symptoms were explored in bakery workers (exposure: wheat allergen) and auto body shop workers (exposure: diisocyanates), again the symptom data had been collected for both skin and respiratory outcomes but there was incomplete skin exposure information in both groups. For this reason the airborne exposures were used as a proxy of skin exposure in both groups (Chapter 5) The use of airborne exposure as a proxy for skin exposure was justified based on the evidence in the literature that skin and airborne exposures are correlated in many occupational exposure scenarios (Table 4). Specifically there was evidence of correlation between skin and airborne diisocyanate exposure in an auto body shop population [Fent et al. 2008]. However, there was no evidence located to support or refute a correlation between skin and airborne exposures in bakery workers. In fact, there were no located studies of measured skin exposure in bakery workers. The lack of skin exposure measures in bakery workers may be due to the particulate nature of flour dust, a major exposure in bakeries. The measurement of skin exposure to particulates

124 105 introduces unique challenges. A dermal sampling strategy for particulate exposures has been developed but does not appear in any published studies of bakery workers to date [Lundgren et al. 2006, Ashley et al. 2007]. The results in Chapter 5 showed that there were exposure-response relationships for skin symptoms in auto body shop workers; the same relationships were not observed in bakery workers. There may not be an exposure-response relationship in bakery workers. But, if one does exist, the lack of an observed relationship in this study may be due to any of several factors. The first possibility is that the exposure investigated (wheat allergen) may not have been the causal exposure for skin symptoms. Other possible exposures in the bakery environment include flour dust, enzymes, preservatives, oils, and wet work. The relationships of total dust and α-amylase with skin symptoms were investigated separately and neither showed a significant exposureresponse relationship (results not shown). A second possibility is that the measured airborne exposures to wheat allergen may have been a poor proxy for skin exposure in this exposure scenario. The study by Hughson et al. was the only study in Table 4 to examine skin and airborne exposure to a particulate exposure (nickel). In this case they found strong correlation (r = ) between inhalable nickel concentrations and the amount of nickel on skin (using skin wipes as a removal method). Further understanding of exposure-response relationships for skin symptoms in a wider variety of occupational settings under different exposure scenarios is needed. 7.5 Workers Do Report Concurrent Skin and Respiratory Symptoms Results reported in this thesis (Chapter 4, Chapter 5 and Chapter 6) demonstrate that a portion of workers in a variety of occupations do report both skin and respiratory symptoms. The small data pooling study in Chapter 4 was a pilot study designed to determine whether or not workers from a variety of occupations (ammonia processing, cabinet making, softwood planing, and embalming) reported concurrent skin and respiratory symptoms. Among this group, 26 workers (11%) did report concurrent skin and respiratory symptoms, though few (n=2, < 1%) associated both their skin and their lung symptoms as being caused by work. This study confirmed our hypothesis that concurrent symptoms would exist among workers, but suggested that the

125 106 prevalence of concurrent work-related symptoms was much lower than either work-related skin symptoms or work-related respiratory symptoms alone. In Chapter 6, concurrent skin and respiratory symptoms were examined in a clinical population where the prevalence of both symptoms and work-related symptoms was expected to be higher than in the data used in Chapter 4. This hypothesis was supported; 38% of subjects reported concurrent skin and respiratory symptoms and 18% reported that they had both skin and respiratory symptoms that were work-related. In Chapter 5 the primary goal was to identify exposure-response relationships for skin symptoms among bakery and auto body shop workers. Additionally, the relationships between skin symptoms and respiratory symptoms were also investigated in these two populations. Results suggested that workers who reported skin symptoms were at significantly greater risk of reporting respiratory symptoms. This result was true for both bakery and auto body shop workers, for both work-related and non-work-related symptoms. The results from Chapter 4, Chapter 5, and Chapter 6 are similar to the findings of two previous studies on concurrent skin and respiratory symptoms. Moulin et al. briefly reported on subjects with diagnosed occupational contact dermatitis, 4% of these patients also reported work-related respiratory symptoms [Moulin et al. 2009]. In an occupational study of professional cleaners, Lynde et al. reported than 7.2% of subjects reported a skin rash as well as three or more respiratory symptoms; 3.7% of subjects reported a rash plus two or more work-related respiratory symptoms [Lynde et al. 2009]. The clinical study from Moulin et al. has a low prevalence of work-related respiratory symptoms among a group of occupational contact dermatitis patients. Few details about this population are available, but the fact that they all have occupational contact dermatitis suggests that they are all exposed (via contact) at work. Lynde et al. report similar prevalence of work-related respiratory symptoms (3.7%) among professional cleaners with a current rash. There is no clinical diagnosis of the rash, so this group may contain subjects with contact dermatitis, atopic eczema, or other skin conditions that may or may not be work-related. Comparing between these two studies is difficult as they were conducted in different environments (clinical vs. occupational) with different groups of subjects (patients with varied exposures vs. workers with similar exposures) and different symptom definitions.

126 107 The prevalence of work-related concurrent skin and lung symptoms among the four occupational groups in Chapter 4 are lower than those reported by both Moulin et al. and Lynde et al. [Moulin et al. 2009, Lynde et al. 2009]. Reasons for this are not entirely clear. The data that was used in Chapter 4 were historical (1980s) data collected in the occupational setting. The clinical population in Chapter 6 can be reasonably compared to the study by Moulin et al. Among the Derm (possible work-related skin disease) subjects in Chapter 6, 88% were diagnosed with contact dermatitis and 69% with work-related contact dermatitis. Among all Derm subjects, 28% reported work-related respiratory symptoms and 6% reported work-related asthma-like symptoms. These rates of work-related respiratory symptoms are much higher than those reported my Moulin et al. There may be differences in the clinical population that we cannot determine from the information reported, including occupational distribution, exposures at work, or prevention strategies employed at work, among others. With so few studies, we do not yet have a good estimate of the prevalence of concurrent symptoms among workers or among patients. But it is clear that a portion of subjects in all studies did report concurrent symptoms. There is a need to at least consider the possibility of concurrent symptom outcomes and the mechanisms that might be leading to these outcomes in workers. It would be beneficial if future research considering either skin or respiratory outcomes could collect information on outcomes in both systems, as well as the work-relatedness of both symptoms, and if possible collect information (or measurements) to estimate exposure in both systems. 7.6 Symptom Progression When considering the concurrent skin and respiratory symptom outcomes in Chapter 4, Chapter 5 and Chapter 6 it would have been interesting to be able to group subjects based on which symptom was experienced first. This information was not collected as part of these studies. The clinical stream was initially considered as a proxy for the first, or at least more serious, symptom. But upon further consideration this was deemed inappropriate. The clinical stream under which a patient was seen at the clinic is highly variable and depends in large part on the referring physician. In addition, it is possible that some subjects would have underlying disease (i.e., asthma or dermatitis) that is long-standing and under management by another physician. If

127 108 the primary symptom could not be determined from the clinical stream, and subjects were not asked about the timing of symptoms in the questionnaire, it was not possible to determine the temporality of symptom onset. Though the order that symptoms are experienced in workers may not matter, it cannot be determined without investigation. In future studies it would be beneficial for the timing or order of symptoms onset to be investigated and included in greater depth in the questionnaire items. 7.7 Predictors of Concurrent Skin and Respiratory Symptoms Chapter 6, the final thesis data chapter, investigated the specific predictors of reporting concurrent skin and respiratory symptoms among the clinical subjects. Contrary to our hypotheses, results from the clinic-based study suggested that none of the specific workplace exposures or occupational groups were strong predictors of concurrent skin and respiratory symptoms. In this population, workers with a history of eczema, current smokers and workers from larger workplaces were at higher risk of reporting concurrent symptoms. Despite the fact that these results did not support our hypotheses, they are, nonetheless, interesting Barrier Function and Concurrent Skin and Respiratory Symptoms The finding that workers with a history of eczema had increased odds of reporting concurrent skin and respiratory symptoms suggests a role for impaired barrier function in the development of the concurrent symptoms. The hypothesis is that workers with an impaired barrier function may be more likely to report concurrent symptoms because they have increased uptake of skin exposure through a compromised skin barrier. Previous studies have shown that workers with impaired skin barrier function have increased uptake of exposure. In a study of carcinogen exposed rubber workers, Vermeulen et al. reported that those workers with hand dermatitis had higher levels of mutagenicity in their urine [Vermeulen et al. 2003]. Hino et al. found a better correlation between skin condition and urine biomarkers of exposure than between airborne exposure and urine levels of the biomarkers of exposure [Hino et al. 2008]. Evidence from Bremmer et al. suggested that the presence of ichthyosis vulgaris in patients with atopic dermatitis (AD) is predictive of subsequent development of respiratory symptoms [Bremmer et al. 2008]. Results presented in Chapter 6

128 109 showed that subjects with a history of eczema (doctor-diagnosed) were more likely to report concurrent skin and respiratory symptoms (both work-related and non-work-related). The subjects in Chapter 6 reported whether they ever had eczema, and also reported current skin rash. Jakasa et al. showed that the uptake of SLS (a strong irritant) was greater among subjects with active AD, compared with subjects with normal skin and those with inactive AD [Jakasa et al. 2006]. In this case, the AD had to be active in order for the increased permeation to be observed. Jakasa et al. also measured the penetration of a strong irritant, one that has itself been known to disrupt barrier function even in healthy skin [Nielsen. 2005]. As the subjects in Chapter 6 reported current skin symptoms, it is plausible that they did have current barrier dysfunction, but no objective measure of skin condition was collected as part of this study. There was no association with having a history of asthma and reporting concurrent skin and respiratory symptoms. In univariate analyses there was no significant relationship between a history of asthma and a history of eczema, suggesting there was no confounding between these variables. Additionally, the multiple logistic regression models were adjusted for atopy, and the effect of eczema remained significant after adjustment for atopy, suggesting that this relationship was also not confounded by atopy. Beyond having (or having had) eczema, there are other factors that could disrupt the skin barrier in a manner meaningful for the uptake of occupational exposures. Impaired skin barrier function can result from disease (eczema or atopic dermatitis [Jakasa et al. 2006]), genetics (filaggrin mutations [Nemoto-Hasebe et al. 2008]), exposure effects (irritant exposures [Nielsen. 2005]), or a combination of these factors. The measurement of skin condition and barrier function deserves attention in future studies of skin exposure and both skin and respiratory outcomes. This could take the form of dermatological examination, the use of pictorial questionnaires for identification of skin disease or the direct measurement of skin condition using transepidermal water loss (TEWL) and other physiological measures of skin integrity (i.e., hydration and ph) Personal Protective Equipment In the first two studies of symptom outcomes (Chapter 4 and Chapter 5) there was not consistent information on the use of personal protective equipment (PPE) to be able to consider its use in

129 110 pooled analyses or exposure-response models. As a result, any effect of wearing PPE (i.e., gloves or respirator) was missed in these studies. It is possible that PPE use results in lower exposure and in turn a lower probability of reporting symptoms. However is it also possible that symptoms could result as an effect of wearing PPE. Thirdly, from the progression perspective it is possible that workers did not wear PPE, were consequently exposed and went on to developed symptoms, and then, after the onset of symptoms, began wearing PPE. In Chapter 6 subjects were explicitly asked about their use of PPE, both yes/no use and frequency of use, however no information was collected on the temporal relationship between symptoms and PPE use. The result is that, in the same way that the timing and order of symptom onset cannot be determined, neither can the order of PPE use in relation to symptom onset. Models describing the predictors of reporting both skin and respiratory symptoms indicated that workers who wore respirators at work were more likely to report general skin and respiratory symptoms (Table 21). Several explanations are possible. It is possible that workers who are more highly exposed, or are exposed to higher risk compounds are more likely to report symptoms. In this case, respirator use would be acting as a proxy for exposures. However, it is also possible that the subjects who had symptoms were more likely to wear a respirator at work. As in the case of the symptom timing, we are unable to determine which came first (the symptoms or the PPE use) with the data collected in this cross-sectional study. It would be ideal if future studies of occupational exposure-response relationships could include questions directly about the role of PPE in symptoms development, and the temporality of symptom onset with respect to PPE use (or changes in PPE) in the workplace. This will allow for better understanding of the role of PPE in the pathway from exposure to response Smoking In Chapter 6, current smoking was a significant predictor of reporting concurrent skin and respiratory symptoms (not work-related) in multiple logistic regression models (Table 21). The differences between subjects reporting respiratory symptoms only and subjects reporting skin symptoms only were explored in models nested within the concurrent symptom models (Table

130 111 21). Results from the nested models indicated that smoking was not a significant predictor of reporting respiratory symptoms only (Table 21). This result from the nested model was surprising because the association between smoking and respiratory symptoms is established [Higgins. 1959], and in some cases, most commonly with asbestos exposures and lung cancer, occupational exposures and smoking are known to act in a synergistic manner [Frost et al. 2011]. The lack of an association suggests that the risk of respiratory symptoms among smokers is not statistically different than the risk of skin symptoms. In previous studies of skin symptoms, some researchers have chosen to adjust for smoking [Sripaiboonkij et al. 2009b, Sripaiboonkij et al. 2009a] while others have not [de Joode et al. 2007]. The former studies also examined exposure-response relationships that may explain why they included smoking as a covariate, but there is no explicit discussion of leaving the variable in models that described exposure-response relationships for skin symptoms. Models exploring the exposure-response relationships for skin symptoms in bakery and auto body shop workers in Chapter 5 did not include smoking as a covariate. Smoking was included in the models presented in Chapter 6 due to the known association between smoking and respiratory symptoms, which were part of the outcome variable. The result that smoking was associated with concurrent symptoms and not with respiratory symptoms (in the nested model) was surprising. This result suggests that the relationships between smoking, skin symptoms, and respiratory symptoms may be more complicated than anticipated, and that smokers may be at higher risk of skin symptoms than previously thought. This association deserves further attention in future studies. Researchers investigating skin symptoms should be sure to include smoking, at least in preliminary analyses, to ensure that there is not an unrecognized association Mechanism of Effect As noted in each of Chapter 4, Chapter 5, and Chapter 6 the mechanisms of self-reported symptoms (skin or respiratory) cannot be determined without immunological testing. Without this information it is not possible to differentiate between symptoms resulting from irritant and allergic mechanisms. Despite this limitation in studies of self-reported symptoms, it is important to recognize that when considering the prevention of concurrent skin and respiratory symptom outcomes the

131 112 specific mechanism(s) underlying the symptom may not be as relevant. In the case of workers with concurrent symptoms there may be unrecognized or uncontrolled exposures in either/both systems. These simultaneous exposures may be to the same exposure agent, or to different (and multiple) exposure agents that may act independently to cause concurrent symptoms in the separate systems. In the workplace, the primary concern should be exposure control regardless of the agent or route of exposures. Additionally, many occupational allergens can also act as irritants. And, workers with exposures to sensitizing agents in the workplace will also likely have exposure to irritants (i.e., hand washing and wet work). As seen in previous studies [Vermeulen et al. 2003, Hino et al. 2008], co-existing exposures to irritants and allergens can interact to potentially increase the uptake of exposure. Work from Nielsen et al. mentioned previously shows that irritant exposures (SLS) can disrupt skin barrier function and that the penetration of pesticides is increased in the irritant disrupted skin [Nielsen. 2005]. Benfeldt et al. showed similar effects for the uptake of salicylic acid through skin damaged by tape stripping [Benfeldt et al. 1999]. Together with the results from Vermeulen et al. and Hino et al. which demonstrate increased uptake of occupational exposures in workers with compromised skin barrier [Vermeulen et al. 2003, Hino et al. 2008], there is good reason to believe that damaged skin barrier leads to an increase in subsequent occupational exposures through the skin. The potential for exposure to skin damaging irritants to increase the uptake of occupational sensitizers has not been investigated. There are many interesting studies of allergen and irritant exposure, skin barrier integrity, and sensitization that could be undertaken as a starting point for exploring these relationships. 7.8 Strengths & Limitations Strengths The biggest strength of this thesis is that it investigated the connections between skin and respiratory systems from several different perspectives: the similarities in causal exposures, the association between skin and respiratory symptoms, and exposure-response relationships for skin symptoms. In all of these cases, the results support that there is a connection between the skin

132 113 and the respiratory system in terms of exposure and outcome. These results provide strong evidence that there is a need for further study of these connections. The research aims of this thesis outlined a plan that involved the study of multiple research questions, each approaching a slightly different aspect of the possible connection between the skin and respiratory systems. The results provide some insight into a number of the relationships described in the proposed framework (Figure 2). Overall, the series of papers that make up this thesis are part of small body of research that has considered the skin and respiratory systems together in human studies. Two of the studies included in this thesis were able to take advantage of existing data to explore skin and respiratory symptoms in working populations. The availability of quality previously unanalyzed data relevant to the research questions provided an opportunity to confirm that there was a portion of workers with skin and respiratory symptoms (Chapter 4). The work presented in Chapter 3, the study of common occupational contact allergens and their possible associations with occupational asthma, shed light on workplace exposures that may be able to cause both skin and respiratory disease in workers. Prior to this study there were no systematic comparisons of exposures causing occupational asthma and exposures causing occupational contact dermatitis located in the literature. Secondarily this study provided a review of occupational hygiene notations for each of the common occupational contact allergens. The review on the notations highlighted important differences in the purpose of the three occupational hygiene reference documents, and also clearly demonstrated discrepancies in the information that is acquired from each of the different sources. The dissemination of these results will ideally lead to more informed interpretation of these notations and perhaps even more thorough consideration of skin effects in future versions of these documents. In Chapter 5 the study modeled exposure-response for skin symptom in two working populations, bakery workers and auto body shop workers. Specifically, this study contributes to a relatively small body of literature on exposure-response relationships for skin symptoms in working populations. The results from this study add to previous findings that exposure-response relationships exist for skin symptoms in wood workers, glass microfiber workers and metal workers [de Joode et al. 2007, Sripaiboonkij et al. 2009b, Sripaiboonkij et al. 2009a].

133 114 In addition, this study demonstrated that both bakery workers and auto body shop workers who reported skin symptoms were more likely to report respiratory symptoms than those who did not report skin symptoms. This finding confirmed previous work by Lynde et al. in professional cleaners [Lynde et al. 2009]. Together these studies suggest that concurrent symptoms may be a significant issue in some working populations. Chapter 6 described the first study of co-existing skin and respiratory symptoms among a clinical population of workers with suspected work-related disease. In this study there was a very high response rate (81%) for eligible subjects. Among workers being assessed for work-related skin or respiratory disease, a surprising portion also reported work-related symptoms in the other system. This is an important finding as the majority of clinical investigations are siloed and tend to focus on one problem at a time. This study broke down the silos and demonstrated a need to inquire about exposure and symptoms in other systems, particularly in cases where the exposures are known to cause disease in more than one system Limitations All of the studies reported in this thesis were cross-sectional in nature. Due to the limitations of the cross-sectional design, no temporal relationships were investigated and causality cannot be inferred from any reported associations. In hindsight, it may have been possible to address some of the temporality issues that were raised in the discussion (i.e., order of symptom onset, use of PPE before or after symptom onset) with carefully designed questionnaires. Unfortunately this was not possible in studies using historical data, and was unfortunately not incorporated into tools used for collecting new data. Another aspect of this thesis that might be perceived as a limitation is the strong focus on reported symptoms rather than clinical diagnoses. The lack of reliance on diagnostic testing is due in large part to the fact that subjects were being clinically investigated for either skin or respiratory disease, or had previously participated in a research that was focused on either skin or respiratory symptoms, and therefore only received the diagnostic testing associated with one system. In addition, the studies focused on concurrent symptoms were not particularly concerned with the mechanism of action. Instead the goal was to determine the prevalence and predictors of concurrent symptoms regardless of underlying mechanism. Thus it is possible that the concurrent symptoms we observed arose from simultaneous exposure through both the skin and the

134 115 respiratory routes and that these potential simultaneous exposures may have included multiple different exposure agents. It would be interesting to undertake complete diagnostic testing for both biological systems of interest (i.e., spirometry, methacholine, and patch testing) in all subjects in a future study, but the cost is high and recruiting subjects for a significant amount of time and additional testing might decrease the response rate. A large limitation of studying concurrent symptoms is the lack of accepted methods for investigating these two concurrent outcomes. In addition, the merging of data on respiratory and skin data proved more challenging than anticipated. The limitations of individual studies were described in each chapter. Briefly, the study in Chapter 6 included subjects from a wide variety of occupations and exposures that resulted in low power for investigating predictors of interest and also issues of data sparseness. It was known a priori that the group of patients seen in the clinic was varied, but the true extent of variety in exposures and occupation had never been described. In future it may be preferable to limit the recruitment of subjects by occupation or industry to reduce this variability. The much discussed limitation within Chapter 5 is the use of airborne exposure as a proxy for skin exposure. This choice may have resulted in exposure misclassification for skin exposure, and may have contributed to the lack of an observed exposure-response for skin symptoms in bakery workers. In studies utilizing historical data, the limitations of the data are inescapable. Prospective studies of skin and respiratory outcomes should consider measuring, or at least estimating, both skin and airborne exposure wherever possible. Chapter 4 was again limited by the historical data utilized. No detailed analysis on exposures in these groups was undertaken as it was not feasible to combine the exposures measured in the individual occupational groups into a single exposure metric. Additionally, the health outcomes and diagnostic testing completed were different between the four groups, which again limited the ability to pool these results. Currently, further research into the group of embalmers is underway. The workers in this group are exposed to both glutaraldehyde and formaldehyde, exposure was measured for both the skin and inhalation routes, and subjects were clinically investigated for both skin and respiratory outcomes.

135 116 The main limitation of the study identifying common contact allergens and their ability to also cause OA (Chapter 3) was that this process was not completed for the opposite scenario: common causes of OA that can also cause CD. This additional work was not undertaken due to the inability to access sufficient data for the causes of occupational asthma cases. This limitation could potentially be resolved through the development of a surveillance system for OA, or by identifying the common causes of OA from workers compensation claims on a province-byprovince effort, though both of these may be limited by under-reporting of occupational asthma. 7.9 Contribution to the Literature Overall, the results from this thesis provide evidence from several perspectives that the skin and respiratory systems are more closely related in the occupational setting than previously realized. The exact nature of this connection is not fully understood, and will likely differ depending on the exposure hazard and exposure scenario, but clearly needs to be investigated in future research. This body of work provides evidence that several common occupational exposures can cause disease in both the skin and respiratory system as a result of skin and airborne exposures, respectively. In general, there is a need to better communicate the risks of both skin and respiratory exposure to workers, occupational health and safety practitioners, and clinicians so that occupational disease does not continue to be under-recognized and under-reported. The results from this thesis demonstrate that among working populations and clinical populations, a portion of workers do report both skin and respiratory symptoms, though a smaller portion report concurrent skin and respiratory symptoms that are work-related. These symptoms may be due to a single exposure agent or to multiple exposure agents, perhaps even a different exposure in each system; more work is needed to answer these questions. The study of bakery workers and auto body shop workers confirms previous research results that workers with skin symptoms are more likely to report respiratory symptoms. Future studies need to gather detailed exposure information about both skin and airborne exposures into order to better determine the role of allergen and irritants exposures in the development of skin and respiratory symptoms. The study of auto body shop workers also provides support that exposure-response relationships do exist for skin symptoms, both work-related and non-work-related. The potential importance of

136 117 skin exposures in the development of occupational isocyanate-induced asthma is currently being studied by a handful of research groups. Results from these ongoing studies are sure to shed more light on the role of skin exposure in the development of occupational isocyanate-induced asthma. Models constructed to identify predictors of concurrent symptom suggested that subjects with a history of eczema had great odds of reporting concurrent symptoms. This finding supports previous occupational research that showed a relationship between skin disease or skin damage and both the uptake of exposure and the likelihood of reporting respiratory symptoms. The role for skin barrier function in modifying the uptake of skin exposure and potentially contributing to the development of respiratory symptoms should continue to be investigated in occupational studies. After synthesizing the results from this thesis, the proposed framework from Figure 2 has been modified to include the additional factors of co-morbid disease, personal characteristics, and PPE use. These factors, along with the previously included genetic factors, have been added to the framework with dotted lines to indicate potential associations, not necessarily causal relationships.

137 118 Figure 6 Modified Framework for Conceptualizing the Connections Between Skin and Respiratory Symptoms in Occupational Disease.

138 119 Chapter 8 Conclusions This thesis is made up of four studies that explore the connection between skin and airborne exposures as well as between skin and respiratory symptoms and disease. The studies included in this thesis contribute to the collective understanding of the connection between the skin and respiratory systems in terms of exposure, response (symptoms), and exposure-response relationships. Overall, the results of this thesis demonstrate that the skin and respiratory systems are associated more closely in terms of occupational exposure and health outcomes than previously considered. This thesis offers four main findings on the potential connection between the skin and respiratory systems in terms of occupational exposures and health outcomes: 1. Common occupational contact allergens are also capable of causing occupational asthma, but may not be recognized correctly as sensitizers in common reference materials. 2. Exposure-response relationships exist for skin symptoms in auto-body workers, and these relationships deserve further study in all occupations with potential skin exposure. 3. Workers in the workplace, as well as workers with suspected occupational disease in a clinical setting, report concurrent skin and respiratory symptoms. A portion of these symptomatic workers associates both of their concurrent symptoms with their work. 4. Subjects who report a history of eczema, are current smokers, wear a respirator at work and work in large workplaces are more likely to report concurrent skin and respiratory symptoms (both work-related and non-work-related). The relationship between eczema and concurrent symptoms lends support to previous findings that skin barrier function in pathway from exposure to symptoms.

139 120 Chapter 9 Future Directions The results from this thesis have generated many new research questions across a variety of disciplines - exposure assessment, occupational hygiene, disease surveillance and mechanistic research - that could, and should, be the subject of future study. Additionally, there are many research questions that remain to be answered within the data collected for Chapter 6. This data collected as part of the study described in Chapter 6 contains a large amount of data on prevention of occupational exposures. Prevention was not a focus of this thesis but is an important consideration in the study of skin and respiratory outcomes. Future analyses will include a description of personal protective equipment (PPE) use and prevention training in the workplace. The goal will be to identify the predictors of PPE use within this population of workers. The results from Chapter 3 identified seven exposures, which are capable of causing both occupational contact dermatitis and occupational asthma. This list of seven exposures will help identify workplaces and workers for in-depth studies of exposure and response in both the skin and respiratory systems together. Workers exposure to these agents may be at risk of both skin and respiratory symptoms. Smaller studies with carefully planned exposure assessment that includes both skin and airborne exposure measurements will help clarify whether there are significant skin and inhalation exposure hazards for workers. If both exposures exist, this type of data will allow for the investigation of correlation between theses exposures, and exposure-response associations with both skin and respiratory outcomes. These analyses will provide exposure (and exposure scenario) specific insight into whether workers are at risk of skin and/or respiratory outcomes, and the relative importance of each route of exposure with respect to the individual outcomes. Within future studies it will be important to better record the use and protection factors of PPE among participants. The role of PPE in modifying exposure, as well as the potential for PPE use (and PPE protection factors) to differ between the skin and respiratory systems will need to be considered. This task could be undertaken in several ways: self-reported use of PPE (e.g., frequency, duration), direct observation of PPE use in the workplace, or perhaps even biomarkers of exposure that measure the internal dose taking into consideration the effects of PPE.

140 121 The results that demonstrated a lack of agreement across the occupational hygiene reference materials show that there is a need for researchers to better collaborate with occupational hygienists. There is a need to integrate research knowledge with practical field reference documents. This work may not take the form of formal research projects, but could involve improved collaboration of researchers and hygienists in the decision making process around sensitizer and skin notations. This knowledge transfer and exchange between disciplines is crucial for effective and efficient risk communication and the protection of workers. In this age of technology and mobile devices it seems plausible that the information from multiple occupational hygiene or occupational health and safety applications could be merged into a master application which allows workers on-site access to a variety of different documents summarizing relevant health and safety information. A few individual organizations have put their documents into a mobile application format (e.g., NIOSH Handbook, WHIMS) but a merging of information does not appear to have been attempted. Though the overall task would be a large one, it might be possible to undertake the development of a mobile application and linkage of data and for one industry as a pilot/feasibility study. This thesis has highlighted the discrepancy between our understanding of skin exposure-response and our understanding of exposure-response in the respiratory system. There is a basic need to better understand the burden of dermatitis, including occupational dermatitis, in the general population. These estimates could be obtained through the analysis of administrative and workers compensation data for skin disease claims, as has been done successfully for occupational asthma in several Canadian jurisdictions. There is also a need for methodological studies to better determine if, and which, skin symptoms and questionnaire items are predictive of dermatological diagnoses. Studies designed to identify symptoms (or symptoms combinations) that identify cases of allergic contract dermatitis, or that differentiate between irritant and allergic contact dermatitis, would permit the use of questionnaires in place of physician diagnosis in subsequent studies. A study of this nature will require a large data set that includes information on reported skin symptoms as well as physician diagnoses. Within a group that has a history of research standardization, such as the NACDG, this may be feasible.

141 122 We also need to further our understanding of skin exposure and exposure-response relationships for skin symptoms. These studies can be undertaken in workplaces with known skin hazards, identified either through the results presented in Chapter 3, or from other surveillance data on common causes of contact dermatitis. These studies of skin exposure should be accompanied by measurements of skin barrier function, either through direct measurement or a questionnaire tool (question or picture based). The incorporation of skin barrier measurement in studies of exposure, biomarkers of exposure and exposure-response will allow for better understanding of how and if the skin barrier modifies exposure-response relationships. Beyond the exposure-response relationships within the skin, there is also a need to better understand the role of skin exposure in the development of respiratory sensitization and disease. As airborne exposure continues to be the focus of research and control strategies, airborne exposures will continue to decrease and the relative importance of skin exposure will increase. This shift is already occurring in the cases of isocyanate and beryllium exposure, and there will likely be other examples in the future. Despite the challenges, these complicated pathways can only be fully addressed in studies where both skin and airborne exposures are measured, and both skin and respiratory outcomes are considered. This design would allow for the relative importance of each route of exposure to be considered; the tendencies for correlation between skin and airborne exposure will complicate this task but it is a challenge that should be overcome. There are also general themes that are highlighted in the limitations of this work. There is a need for well designed, prospective studies that allow for repeated measurements and the study of temporality of the exposure-sensitization-symptom-disease pathway. Though prospective studies are the ideal approach to these problems, careful questionnaire design could also result in rich data that allows for retrospective construction of exposure and response patterns. It seems likely that in order to address the complex relationships between exposure and response in the skin and respiratory symptoms it will be fruitful to focus on studies of one exposure in a more homogenous working population where the variability within and between workers can be accurately assessed.

142 123 The results of this study have contributed to the basic understanding on the relationship between the skin and respiratory systems in terms of exposure and disease, but there is an overwhelming amount of work yet to be done.

143 124 References Akinbami LJ, Moorman JE, Liu X Asthma prevalence, health care use, and mortality: United states, National Health Statistics Reports 32. Alvarez-Cuesta C, Gala Ortiz G, Rodriguez Diaz E, Blanco Barrios S, Galache Osuna C, Raya Aguado C, Pineda F Occupational asthma and IgE-mediated contact dermatitis from sapele wood. Contact Dermatitis 51(2): American Conference of Governmental Industrial Hygienists (ACGIH) TLVs and BEIs threshold limit values for chemical substances and physical agents.. Aprea C, Lunghini L, Banchi B, Peruzzi A, Centi L, Coppi L, Bogi M, Marianelli E, Fantacci M, Catalano P, Benvenuti A, Miligi L, Sciarra G Evaluation of inhaled and cutaneous doses of imidacloprid during stapling ornamental plants in tunnels or greenhouses. J Expo Sci Environ Epidemiol 19(6): Arrandale VH, Liss GM, Tarlo SM, Pratt MD, Sasseville D, Kudla I, Holness DL Occupational contact allergens: Are they also associated with occupational asthma? Am J Ind Med. Arts J, de Koning M, Bloksma N, Kuper C Respiratory allergy to trimellitic anhydride in rats: Concentration-response relationships during elicitation. Inhal Toxicol 16(5): Arts JH, Kuper CF, Spoor SM, Bloksma N Airway morphology and function of rats following dermal sensitization and respiratory challenge with low molecular weight chemicals. Toxicol Appl Pharmacol 152(1): Arts JH, Mommers C, de Heer C Dose-response relationships and threshold levels in skin and respiratory allergy. Crit Rev Toxicol 36(3): Ashley K, Applegate GT, Wise TJ, Fernback JE, Goldcamp MJ Evaluation of a standardized micro-vacuum sampling method for collection of surface dust. J Occup Environ Hyg 4(3):

144 125 Australian Bureau of Statistics. 2011, population by age and sex, regions of Australia (October 18, 2011). Banks DE, Jalloul A Occupational asthma, work-related asthma and reactive airways dysfunction syndrome. Curr Opin Pulm Med 13(2): Barbuzza O, Guarneri F, Galtieri G, Gangemi S, Vaccaro M Protein contact dermatitis and allergic asthma caused by Anisakis simplex. Contact Dermatitis 60(4): Bello D, Redlich CA, Stowe MH, Sparer J, Woskie SR, Streicher RP, Hosgood HD, Liu Y Skin exposure to aliphatic polyisocyanates in the auto body repair and refinishing industry: II. A quantitative assessment. Ann Occup Hyg 52(2): Benfeldt E, Serup J, Menne T Effect of barrier perturbation on cutaneous salicylic acid penetration in human skin: In vivo pharmacokinetics using microdialysis and non-invasive quantification of barrier function. Br J Dermatol 140(4): Bernstein IL, Chan-Yeung M, Malo J, Bernstein DI. 2006a. Asthma in the workplace, and related conditions.. Bernstein IL, Chan-Yeung M, Malo J, Bernstein DI. 2006b. Definition and classification of asthma in the workplace (Chapter 1). In: Bernstein IL, Chan-Yeung M, Malo J, Bernstein DI, editors. 3rd ed. New York: Taylor & Francis. p 1-8. Bjorksten B The role of the gastrointestinal tract in the development of respiratory hypersensitivities. Toxicol Lett 86(2-3): Blanc PD, Toren K How much adult asthma can be attributed to occupational factors? Am J Med 107(6): Boeniger MF The significance of skin exposure. Ann Occup Hyg 47(8): Bos PM, Brouwer DH, Stevenson H, Boogaard PJ, de Kort WL, van Hemmen JJ Proposal for the assessment of quantitative dermal exposure limits in occupational environments: Part 1. Development of a concept to derive a quantitative dermal occupational exposure limit. Occup Environ Med 55(12):

145 126 Bremmer SF, Hanifin JM, Simpson EL Clinical detection of ichthyosis vulgaris in an atopic dermatitis clinic: Implications for allergic respiratory disease and prognosis. J Am Acad Dermatol 59(1): Brooks SM, Weiss MA, Bernstein IL Reactive airways dysfunction syndrome (RADS). Persistent asthma syndrome after high level irritant exposures. Chest 88(3): Burney PG, Laitinen LA, Perdrizet S, Huckauf H, Tattersfield AE, Chinn S, Poisson N, Heeren A, Britton JR, Jones T Validity and repeatability of the IUATLD (1984) bronchial symptoms questionnaire: An international comparison. Eur Respir J 2(10): Burstyn I, Ferrari P, Wegh H, Heederik D, Kromhout H Characterizing worker exposure to bitumen during hot mix paving and asphalt mixing operations. AIHA J (Fairfax, VA) 63(3): Cadot P, Tits G, Bussels L, Ceuppens JL Asthma and hand dermatitis to leek. Allergy 56(2): Carstensen O, Rasmussen K, Ponten A, Gruvberger B, Isaksson M, Bruze M The validity of a questionnaire-based epidemiological study of occupational dermatosis. Contact Dermatitis 55(5): Chang FK, Chen ML, Cheng SF, Shih TS, Mao IF. 2007a. Dermal absorption of solvents as a major source of exposure among shipyard spray painters. J Occup Environ Med 49(4): Chang FK, Chen ML, Cheng SF, Shih TS, Mao IF. 2007b. Evaluation of dermal absorption and protective effectiveness of respirators for xylene in spray painters. Int Arch Occup Environ Health 81(2): Chiang WC, Pons L, Kidon MI, Liew WK, Goh A, Wesley Burks A Serological and clinical characteristics of children with peanut sensitization in an Asian community. Pediatr Allergy Immunol 21(2 Pt 2):e Cox-Ganser JM, Rao CY, Park JH, Schumpert JC, Kreiss K Asthma and respiratory symptoms in hospital workers related to dampness and biological contaminants. Indoor Air 19(4):

146 127 Crapo RO, Morris AH, Gardner RM Reference spirometric values using techniques and equipment that meet ATS recommendations. Am Rev Respir Dis 123(6): Cullinan P, Cook A, Jones M, Cannon J, Fitzgerald B, Taylor AJ Clinical responses to ingested fungal alpha-amylase and hemicellulase in persons sensitized to Aspergillus fumigatus? Allergy 52(3): Cullinan P, Cook A, Nieuwenhuijsen MJ, Sandiford C, Tee RD, Venables KM, McDonald JC, Newman Taylor AJ Allergen and dust exposure as determinants of work-related symptoms and sensitization in a cohort of flour-exposed workers; a case-control analysis. Ann Occup Hyg 45(2): Daftarian HS, Lushniak BD, Reh CM, Lewis DM Evaluation of self-reported skin problems among workers exposed to toluene diisocyanate (TDI) at a foam manufacturing facility. J Occup Environ Med 44(12): Day GA, Dufresne A, Stefaniak AB, Schuler CR, Stanton ML, Miller WE, Kent MS, Deubner DC, Kreiss K, Hoover MD Exposure pathway assessment at a copper-beryllium alloy facility. Ann Occup Hyg 51(1): Day GA, Stefaniak AB, Weston A, Tinkle SS Beryllium exposure: Dermal and immunological considerations. Int Arch Occup Environ Health 79(2): de Joode BW, Vermeulen R, Heederik D, van Ginkel K, Kromhout H Evaluation of 2 self-administered questionnaires to ascertain dermatitis among metal workers and its relation with exposure to metalworking fluids. Contact Dermatitis 56(6): De Raeve H, Vandecasteele C, Demedts M, Nemery B Dermal and respiratory sensitization to chromate in a cement floorer. Am J Ind Med 34(2): De Zotti R, Larese F, Fiorito A Asthma and contact urticaria from latex gloves in a hospital nurse. Br J Ind Med 49(8): Diepgen TL, Coenraads PJ The epidemiology of occupational contact dermatitis. Int Arch Occup Environ Health 72(8):

147 128 Doekes G, Douwes J, Wouters I, de Wind S, Houba R, Hollander A Enzyme immunoassays for total and allergen specific IgE in population studies. Occup Environ Med 53(1): Donovan JC, Kudla I, DeKoven JG Rapid development of allergic contact dermatitis from dicyclohexylmethane-4,4'-diisocyanate. Dermatitis 20(4): English JS Current concepts of irritant contact dermatitis. Occup Environ Med 61(8):722-6, 674. Estlander T, Kanerva L, Tupasela O, Keskinen H, Jolanki R Immediate and delayed allergy to nickel with contact urticaria, rhinitis, asthma and contact dermatitis. Clin Exp Allergy 23(4): Fantuzzi G, Righi E, Predieri G, Giacobazzi P, Mastroianni K, Aggazzotti G Prevalence of ocular, respiratory and cutaneous symptoms in indoor swimming pool workers and exposure to disinfection by-products (DBPs). Int J Environ Res Public Health 7(4): Fenske RA, Leffingwell JT, Spear RC. 1986a. A video imaging technique for assessing dermal exposure. I. Instrument design and testing. Am Ind Hyg Assoc J 47(12): Fenske RA, Wong SM, Leffingwell JT, Spear RC. 1986b. A video imaging technique for assessing dermal exposure. II. Fluorescent tracer testing. Am Ind Hyg Assoc J 47(12): Fent KW, Jayaraj K, Ball LM, Nylander-French LA Quantitative monitoring of dermal and inhalation exposure to 1,6-hexamethylene diisocyanate monomer and oligomers. J Environ Monit 10(4): Ferris BG Epidemiology standardization project (American Thoracic Society). Am Rev Respir Dis 118(6): Flack S, Goktepe I, Ball LM, Nylander-French LA Development and application of quantitative methods for monitoring dermal and inhalation exposure to propiconazole. J Environ Monit 10(3):

148 129 Franko J, Jackson LG, Hubbs A, Kashon M, Meade BJ, Anderson SE Evaluation of furfuryl alcohol sensitization potential following dermal and pulmonary exposure: Enhancement of airway responsiveness. Toxicol Sci. Frick M, Bjorkner B, Hamnerius N, Zimerson E Allergic contact dermatitis from dicyclohexylmethane-4,4'-diisocyanate. Contact Dermatitis 48(6): Friis L, Norback D, Edling C Self-reported asthma and respiratory symptoms in sewage workers. Journal of Occupational Health 41(2): Frost G, Darnton A, Harding AH The effect of smoking on the risk of lung cancer mortality for asbestos workers in Great Britain ( ). Ann Occup Hyg 55(3): Fujita H, Ogawa M, Endo Y A case of occupational bronchial asthma and contact dermatitis caused by ortho-phthalaldehyde exposure in a medical worker. J Occup Health 48(6): Gautrin D, Bernstein IL, Brooks SM, Henneberger PK Ch. 25 reactive airways dysfunction syndrome and irritant induced asthma. In: Bernstein IL, Bernstein DI, Chan-Yeung M, Malo J, editors. Asthma in the workplace. 2nd ed. New York: M. Dekker. Goe SK, Henneberger PK, Reilly MJ, Rosenman KD, Schill DP, Valiante D, Flattery J, Harrison R, Reinisch F, Tumpowsky C, Filios MS A descriptive study of work aggravated asthma. Occup Environ Med 61(6): Guarneri F, Barbuzza O, Vaccaro M, Galtieri G Allergic contact dermatitis and asthma caused by limonene in a labourer handling citrus fruits. Contact Dermatitis 58(5): Hannaford-Turner K, Elder D, Sim MR, Abramson MJ, Johnson AR, Yates DH Surveillance of Australian workplace Based Respiratory Events (SABRE) in New South Wales. Occup Med (Lond). Hannu T, Frilander H, Kauppi P, Kuuliala O, Alanko K IgE-mediated occupational asthma from epoxy resin. Int Arch Allergy Immunol 148(1): Hastie T Generalized additive models. London: Chapman and Hall..

149 130 Heldal KK, Madso L, Huser PO, Eduard W Exposure, symptoms and airway inflammation among sewage workers. Ann Agric Environ Med 17(2): Herrick CA, Xu L, Wisnewski AV, Das J, Redlich CA, Bottomly K A novel mouse model of diisocyanate-induced asthma showing allergic-type inflammation in the lung after inhaled antigen challenge. J Allergy Clin Immunol 109(5): Higgins IT Tobacco smoking, respiratory symptoms, and ventilatory capacity; studies in random samples of the population. Br Med J 1(5118): Hill AB The environment and disease: Association or causation? Proc R Soc Med 58: Hino R, Nishio D, Kabashima K, Tokura Y Percutaneous penetration via hand eczema is the major accelerating factor for systemic absorption of toluene and xylene during car spray painting. Contact Dermatitis 58(2): Holgate ST The airway epithelium is central to the pathogenesis of asthma. Allergol Int 57(1):1-10. Holness DL, Broder I, Corey PN, Booth N, Mozzon D, Nazar MA, Guirguis S Respiratory variables and exposure-effect relationships in isocyanate-exposed workers. J Occup Med 26(6): Holness DL, Nethercott JR Health status of funeral service workers exposed to formaldehyde. Arch Environ Health 44(4): Holness DL, Purdham JT, Nethercott JR Acute and chronic respiratory effects of occupational exposure to ammonia. Am Ind Hyg Assoc J 50(12): Holness DL, Tarlo SM, Sussman G, Nethercott JR Exposure characteristics and cutaneous problems in operating room staff. Contact Dermatitis 32(6): Holter G, Irgens A, Nyfors A, Aasen TB, Florvaag E, Overa KB, Elsayed S, Naerheim J Self-reported skin and respiratory symptoms related to latex exposure among 5,087 hospital employees in Norway. Dermatology 205(1):28-31.

150 131 Hughson GW, Cherrie JW Comparison of measured dermal dust exposures with predicted exposures given by the EASE expert system. Ann Occup Hyg 49(2): Hughson GW, Galea KS, Heim KE Characterization and assessment of dermal and inhalable nickel exposures in nickel production and primary user industries. Ann Occup Hyg 54(1):8-22. Huusom AJ, Agner T, Backer V, Ebbehoj N, Jacobsen P Skin and respiratory disorders following the identification of disaster victims in Thailand. Forensic Sci Med Pathol. Igea JM, Fernandez M, Quirce S, de la Hoz B, Diez Gomez ML Green bean hypersensitivity: An occupational allergy in a homemaker. J Allergy Clin Immunol 94(1): Industry Canada Canadian industry statistics - glossary of terms.. Isaksson M, Bruze M Allergic contact dermatitis in response to budesonide reactivated by inhalation of the allergen. J Am Acad Dermatol 46(6): Jaakkola MS, Suuronen K, Luukkonen R, Jarvela M, Tuomi T, Alanko K, Makela EA, Jolanki R Respiratory symptoms and conditions related to occupational exposures in machine shops. Scand J Work Environ Health 35(1): Jacobs JH, Meijster T, Meijer E, Suarthana E, Heederik D Wheat allergen exposure and the prevalence of work-related sensitization and allergy in bakery workers. Allergy 63(12): Jacobs MC, Rycroft RJ Contact dermatitis and asthma from sodium metabisulfite in a photographic technician. Contact Dermatitis 33(1): Jakasa I, de Jongh CM, Verberk MM, Bos JD, Kezic S Percutaneous penetration of sodium lauryl sulphate is increased in uninvolved skin of patients with atopic dermatitis compared with control subjects. Br J Dermatol 155(1): Jenkins MA, Clarke JR, Carlin JB, Robertson CF, Hopper JL, Dalton MF, Holst DP, Choi K, Giles GG Validation of questionnaire and bronchial hyperresponsiveness against respiratory physician assessment in the diagnosis of asthma. Int J Epidemiol 25(3):

151 132 Kanerva L, Estlander T, Keskinen H, Jolanki R Occupational allergic airborne contact dermatitis and delayed bronchial asthma from epoxy resin revealed by bronchial provocation test. Eur J Dermatol 10(6): Kanerva L, Jolanki R, Tupasela O, Halmepuro L, Keskinen H, Estlander T, Sysilampi ML Immediate and delayed allergy from epoxy resins based on diglycidyl ether of bisphenol A. Scand J Work Environ Health 17(3): Kanerva L, Keskinen H, Autio P, Estlander T, Tuppurainen M, Jolanki R Occupational respiratory and skin sensitization caused by polyfunctional aziridine hardener. Clin Exp Allergy 25(5): Karjalainen A, Kurppa K, Virtanen S, Keskinen H, Nordman H Incidence of occupational asthma by occupation and industry in Finland. Am J Ind Med 37(5): Kimber I, Basketter DA, Gerberick GF, Dearman RJ Allergic contact dermatitis. Int Immunopharmacol 2(2-3): Klink KJ, Meade BJ Dermal exposure to 3-amino-5-mercapto-1,2,4-triazole (AMT) induces sensitization and airway hyperreactivity in BALB/c mice. Toxicol Sci 75(1): Krautheim AB, Jermann TH, Bircher AJ Chlorhexidine anaphylaxis: Case report and review of the literature. Contact Dermatitis 50(3): Kujala VM, Reijula KE Glove-induced dermal and respiratory symptoms among health care workers in one Finnish hospital. Am J Ind Med 28(1): Kuper CF, Radonjic M, van Triel J, Stierum R, de Groot RJ, Arts JH Oxazolone (OXA) is a respiratory allergen in brown Norway rats. Toxicology 290(1): Kuper CF, Stierum RH, Boorsma A, Schijf MA, Prinsen M, Bruijntjes JP, Bloksma N, Arts JH The contact allergen dinitrochlorobenzene (DNCB) and respiratory allergy in the Th2- prone brown Norway rat. Toxicology 246(2-3): Lachapelle J Historical aspects. In: Frosch PJ, Menné T, Rycroft RJG, editors. Textbook of contact dermatitis. 2nd ed. New York: Springer-Verlag.

152 133 Lakwijk N, Van Strien RT, Doekes G, Brunekreef B, Gerritsen J Validation of a screening questionnaire for atopy with serum IgE tests in a population of pregnant Dutch women. Clin Exp Allergy 28(4): Lastbom L, Boman A, Camner P, Ryrfeldt A Increased airway responsiveness after skin sensitisation to 3-carene, studied in isolated guinea pig lungs. Toxicology 147(3): Lastbom L, Boman A, Camner P, Ryrfeldt A Does airway responsiveness increase after skin sensitisation to 3-carene: A study in isolated guinea pig lungs. Toxicology 125(1): Lee HK, Alarie Y, Karol MH Induction of formaldehyde sensitivity in guinea pigs. Toxicol Appl Pharmacol 75(1): Leffler CT, Milton DK Occupational asthma and contact dermatitis in a spray painter after introduction of an aziridine cross-linker. Environ Health Perspect 107(7): Lehto M, Haapakoski R, Wolff H, Majuri ML, Makela MJ, Leino M, Reunala T, Turjanmaa K, Palosuo T, Alenius H Cutaneous, but not airway, latex exposure induces allergic lung inflammation and airway hyperreactivity in mice. J Invest Dermatol 125(5): Liljelind I, Norberg C, Egelrud L, Westberg H, Eriksson K, Nylander-French LA Dermal and inhalation exposure to methylene bisphenyl isocyanate (MDI) in iron foundry workers. Ann Occup Hyg 54(1): Lillienberg L, Andersson EM, Jarvholm B, Toren K Respiratory symptoms and exposureresponse relations in workers exposed to metalworking fluid aerosols. Ann Occup Hyg 54(4): Lindgren T, Andersson K, Dammstrom BG, Norback D Ocular, nasal, dermal and general symptoms among commercial airline crews. International Archives of Occupational & Environmental Health 75(7): Lockman LE Case report: Allergic contact dermatitis and new-onset asthma. Chromium exposure during leather tanning. Can Fam Physician 48:

153 134 Lombardo LJ, Balmes JR Occupational asthma: A review. Environ Health Perspect 108 Suppl 4: Lundgren L, Skare L, Liden C Measuring dust on skin with a small vacuuming sampler--a comparison with other sampling techniques. Ann Occup Hyg 50(1): Lushniak BD Occupational contact dermatitis. Dermatol Ther 17(3): Lynde CB, Obadia M, Liss GM, Ribeiro M, Holness DL, Tarlo SM Cutaneous and respiratory symptoms among professional cleaners. Occup Med (Lond) 59(4): Maestrelli P, Boschetto P, Fabbri LM, Mapp CE Mechanisms of occupational asthma. J Allergy Clin Immunol 123(3): Malo JL, Chan-Yeung M Agents causing occupational asthma. J Allergy Clin Immunol 123(3): Mapp CE Agents, old and new, causing occupational asthma. Occup Environ Med 58(5):354-60, 290. Mapp CE, Boschetto P, Maestrelli P, Fabbri LM Occupational asthma. Am J Respir Crit Care Med 172(3): Marks JG, Elsner P, DeLeo VA Allergic and irritant contact dermatitis. In: Contact and occupational dermatology. St. Louis: Mosby Year Book. Martin PE, Matheson MC, Gurrin L, Burgess JA, Osborne N, Lowe AJ, Morrison S, Meszaros D, Giles GG, Abramson MJ, Walters EH, Allen KJ, Dharmage SC Childhood eczema and rhinitis predict atopic but not non-atopic adult asthma: A prospective cohort study over 4 decades. J Allergy Clin Immunol 127(6): e1. Mathias CGT Occupational dermatoses. In: Dickerson OB, Horvath EP, Zenz C, editors. Occupational medicine. 3rd ed. ed. St. Louis: Mosby. McClean MD, Rinehart RD, Ngo L, Eisen EA, Kelsey KT, Herrick RF Inhalation and dermal exposure among asphalt paving workers. Ann Occup Hyg 48(8):

154 135 McDonald JC, Beck MH, Chen Y, Cherry NM Incidence by occupation and industry of work-related skin diseases in the united kingdom, Occup Med (Lond) 56(6): McDonald JC, Chen Y, Zekveld C, Cherry NM Incidence by occupation and industry of acute work related respiratory diseases in the UK, Occup Environ Med 62(12): Medical Research Council on the Aetiology of Chronic Bronchitis Standardised questionnaire on respiratory symptoms. Br Med J 2:1665. Meding B, Barregard L Validity of self-reports of hand eczema. Contact Dermatitis 45(2): Meijster T, Tielemans E, de Pater N, Heederik D Modelling exposure in flour processing sectors in the Netherlands: A baseline measurement in the context of an intervention program. Ann Occup Hyg 51(3): Moulin P, Magnan A, Lehucher-Michel MP Occupational allergic contact dermatitis and asthma due to a single low molecular weight agent. J Occup Health 51(1): National Institute for Occupational Safety and Health NIOSH pocket guide to chemical hazards. DHHS (NIOSH) Publication No National Institutes of Health Haz-Map: Occupational exposures to hazardous agents.. Nemery B, Lenaerts L Exposure to methylene diphenyl diisocyanate in coal mines. Lancet 341(8840):318. Nemoto-Hasebe I, Akiyama M, Nomura T, Sandilands A, McLean WI, Shimizu H Clinical severity correlates with impaired barrier in filaggrin-related eczema. J Invest Dermatol. Nethercott JR, Holness DL Occupational dermatitis in food handlers and bakers. J Am Acad Dermatol 21(3 Pt 1): Nethercott JR, Holness DL Health status of a group of sewage treatment workers in Toronto, Canada. Am Ind Hyg Assoc J 49(7):

155 136 Nettis E, Assennato G, Ferrannini A, Tursi A Type I allergy to natural rubber latex and type IV allergy to rubber chemicals in health care workers with glove-related skin symptoms. Clin Exp Allergy 32(3): Nielsen JB Percutaneous penetration through slightly damaged skin. Arch Dermatol Res 296(12): Paggiaro PL, Loi AM, Toma G Bronchial asthma and dermatitis due to spiramycin in a chick breeder. Clin Allergy 9(6): Pal TM, de Wilde NS, van Beurden MM, Coenraads PJ, Bruynzeel DP Notification of occupational skin diseases by dermatologists in the Netherlands. Occup Med (Lond) 59(1): Pauluhn J Brown Norway rat asthma model of diphenylmethane-4,4'-diisocyanate (MDI): Analysis of the elicitation dose-response relationship. Toxicol Sci 104(2): Pekkanen J, Sunyer J, Anto JM, Burney P, European Community Respiratory Health Study Operational definitions of asthma in studies on its aetiology. Eur Respir J 26(1): Pelissier S, Chaboillez S, Teolis L, Lemiere C Outcome of subjects diagnosed with occupational asthma and work-aggravated asthma after removal from exposure. J Occup Environ Med 48(7): Petsonk EL, Wang ML, Lewis DM, Siegel PD, Husberg BJ Asthma-like symptoms in wood product plant workers exposed to methylene diphenyl diisocyanate. Chest 118(4): Pratt MD, Belsito DV, DeLeo VA, Fowler Jr. JF, Fransway AF, Maibach HI, Marks JG, Mathias CG, Rietschel RL, Sasseville D, Sherertz EF, Storrs FJ, Taylor JS, Zug K North American Contact Dermatitis Group patch-test results, study period. Dermatitis 15(4): Pronk A, Preller L, Raulf-Heimsoth M, Jonkers IC, Lammers JW, Wouters IM, Doekes G, Wisnewski AV, Heederik D Respiratory symptoms, sensitization, and exposure response relationships in spray painters exposed to isocyanates. Am J Respir Crit Care Med 176(11):

156 137 Pronk A, Tielemans E, Skarping G, Bobeldijk I, VAN Hemmen J, Heederik D, Preller L. 2006a. Inhalation exposure to isocyanates of car body repair shop workers and industrial spray painters. Ann Occup Hyg 50(1):1-14. Pronk A, Yu F, Vlaanderen J, Tielemans E, Preller L, Bobeldijk I, Deddens JA, Latza U, Baur X, Heederik D. 2006b. Dermal, inhalation, and internal exposure to 1,6-HDI and its oligomers in car body repair shop workers and industrial spray painters. Occup Environ Med 63(9): Randolph BW, Lallooug, Gouws E, Colvin MS An evaluation of the respiratory health status of automotive spray-painters exposed to paints containing hexamethylene di-isocyanates in the greater Durban area. S Afr Med J 87(3): Rattray NJ, Botham PA, Hext PM, Woodcock DR, Fielding I, Dearman RJ, Kimber I Induction of respiratory hypersensitivity to diphenylmethane-4,4'-diisocyanate (MDI) in guinea pigs. influence of route of exposure. Toxicology 88(1-3): Redlich CA Skin exposure and asthma: Is there a connection? Proc Am Thorac Soc 7(2): Redlich CA, Herrick CA Lung/skin connections in occupational lung disease. Curr Opin Allergy Clin Immunol 8(2): Rietschel RL, Mathias CG, Fowler Jr. JF, Pratt M, Taylor JS, Sherertz EF, Marks Jr. JG, Belsito DV, Storrs FJ, Maibach HI, Fransway AF, Deleo VA, North American Contact Dermatitis G Relationship of occupation to contact dermatitis: Evaluation in patients tested from 1998 to Am J Contact Dermat 13(4): Riihimäki H, Kurppa K, Karjalainen A, Palo L, Jolanki R, Keskinen H, Mäkinen I, Saalo A, Kauppinen T Occupational diseases in Finland in 2002: New cases of occupational diseases reported to the Finnish register of occupational disease.. SAS Institute Inc SAS v.9.2. Version 9.2. Sass-Kortsak AM, Holness DL, Pilger CW, Nethercott JR Wood dust and formaldehyde exposures in the cabinet-making industry. Am Ind Hyg Assoc J 47(12):

157 138 Schneider T, Vermeulen R, Brouwer DH, Cherrie JW, Kromhout H, Fogh CL Conceptual model for assessment of dermal exposure. Occup Environ Med 56(11): Schnuch A, Lessmann H, Schulz KH, Becker D, Diepgen TL, Drexler H, Erdmann S, Fartasch M, Greim H, Kricke-Helling P, Merget R, Merk H, Nowak D, Rothe A, Stropp G, Uter W, Wallenstein G When should a substance be designated as sensitizing for the skin ('sh') or for the airways ('sa')? Hum Exp Toxicol 21(8): Sistek D, Wickens K, Amstrong R, D'Souza W, Town I, Crane J Predictive value of respiratory symptoms and bronchial hyperresponsiveness to diagnose asthma in New Zealand. Respir Med 100(12): Smit HA, Coenraads PJ A retrospective cohort study on the incidence of hand dermatitis in nurses. Int Arch Occup Environ Health 64(8): Smit HA, Coenraads PJ, Lavrijsen AP, Nater JP Evaluation of a self-administered questionnaire on hand dermatitis. Contact Dermatitis 26(1): Smit HA, van Rijssen A, Vandenbroucke JP, Coenraads PJ Susceptibility to and incidence of hand dermatitis in a cohort of apprentice hairdressers and nurses. Scand J Work Environ Health 20(2): Smit LA, Heederik D, Doekes G, Blom C, van Zweden I, Wouters IM Exposure-response analysis of allergy and respiratory symptoms in endotoxin-exposed adults. Eur Respir J 31(6): Sripaiboonkij P, Phanprasit W, Jaakkola MS. 2009a. Respiratory and skin effects of exposure to wood dust from the rubber tree Hevea brasiliensis. Occup Environ Med 66(7): Sripaiboonkij P, Sripaiboonkij N, Phanprasit W, Jaakkola MS. 2009b. Respiratory and skin health among glass microfiber production workers: A cross-sectional study. Environ Health 8:36. Statistics Canada National occupational classification - statistics (NOC-S) Steiner MF, Dick FD, Scaife AR, Semple S, Paudyal P, Ayres JG High prevalence of skin symptoms among bakery workers. Occup Med (Lond) 61(4):

158 139 Svensson A, Lindberg M, Meding B, Sundberg K, Stenberg B Self-reported hand eczema: Symptom-based reports do not increase the validity of diagnosis. Br J Dermatol 147(2): Tarkowski M, Vanoirbeek JA, Vanhooren HM, De Vooght V, Mercier CM, Ceuppens J, Nemery B, Hoet PH Immunological determinants of ventilatory changes induced in mice by dermal sensitization and respiratory challenge with toluene diisocyanate. Am J Physiol Lung Cell Mol Physiol 292(1):L Tarlo SM, Balmes J, Balkissoon R, Beach J, Beckett W, Bernstein D, Blanc PD, Brooks SM, Cowl CT, Daroowalla F, Harber P, Lemiere C, Liss GM, Pacheco KA, Redlich CA, Rowe B, Heitzer J Diagnosis and management of work-related asthma: American college of chest physicians consensus statement. Chest 134(3 Suppl):1S-41S. Tarlo SM, Boulet LP, Cartier A, Cockcroft D, Cote J, Hargreave FE, Holness L, Liss G, Malo JL, Chan-Yeung M Canadian Thoracic Society guidelines for occupational asthma. Can Respir J 5(4): Tarlo SM, Malo JL, ATS/ERS An ATS/ERS report: 100 key questions and needs in occupational asthma. Eur Respir J 27(3): Tinkle SS, Antonini JM, Rich BA, Roberts JR, Salmen R, DePree K, Adkins EJ Skin as a route of exposure and sensitization in chronic beryllium disease. Environ Health Perspect 111(9): To T, Tarlo SM, McLimont S, Haines T, Holness DL, Lougheed MD, Liss GM, Cicutto L Feasibility of a provincial voluntary reporting system for work-related asthma in Ontario. Can Respir J 18(5): Toren K, Blanc PD Asthma caused by occupational exposures is common - A systematic analysis of estimates of the population-attributable fraction. BMC Pulm Med 9(1):7. Toren K, Brisman J, Olin AC, Blanc PD Asthma on the job: Work-related factors in newonset asthma and in exacerbations of pre-existing asthma. Respir Med 94(6): Tsakirakis A, Kasiotis KM, Arapaki N, Charistou A, Tsatsakis A, Glass CR, Machera K Determination of operator exposure levels to insecticide during bait applications in olive trees:

159 140 Study of coverall performance and duration of application. International Journal of Hygiene & Environmental Health 214(1): Tupker RA, De Monchy JG, Coenraads PJ, Homan A, van der Meer JB Induction of atopic dermatitis by inhalation of house dust mite. J Allergy Clin Immunol 97(5): Turner S, Carder M, van Tongeren M, McNamee R, Lines S, Hussey L, Bolton A, Beck MH, Wilkinson M, Agius R The incidence of occupational skin disease as reported to the health and occupation reporting (THOR) network between 2002 and Br J Dermatol 157(4): UK Health and Safety Executive, Asthmagen? Critical assessment of the evidence for agents implicated in occupational asthma.. United States Department of Labor Bureau of Labor Statistics International comparisons of annual labor force statistics, adjusted to U.S. concepts, 10 countries, Uter W, Nohle M, Randerath B, Schwanitz HJ Occupational contact urticaria and latephase bronchial asthma caused by compositae pollen in a florist. Am J Contact Dermat 12(3): Uter W, Ramsch C, Aberer W, Ayala F, Balato A, Beliauskiene A, Fortina AB, Bircher A, Brasch J, Chowdhury MM, Coenraads PJ, Schuttelaar ML, Cooper S, Corradin MT, Elsner P, English JS, Fartasch M, Mahler V, Frosch PJ, Fuchs T, Gawkrodger DJ, Gimenez-Arnau AM, Green CM, Horne HL, Jolanki R, King CM, Krecisz B, Kiec-Swierczynska M, Ormerod AD, Orton DI, Peserico A, Rantanen T, Rustemeyer T, Sansom JE, Simon D, Statham BN, Wilkinson M, Schnuch A The European baseline series in 10 European countries, 2005/ Results of the European Surveillance System on Contact Allergies (ESSCA). Contact Derm 61(1): Valdivieso R, Subiza J, Varela-Losada S, Subiza JL, Narganes MJ, Martinez-Cocera C, Cabrera M Bronchial asthma, rhinoconjunctivitis, and contact dermatitis caused by onion. J Allergy Clin Immunol 94(5):

160 141 Valks R, Conde-Salazar L, Barrantes OL Occupational allergic contact urticaria and asthma from diphenylmethane-4,4'-diisocyanate. Contact Dermatitis 49(3): van der Lende R, Orie NG The MRC-ECCS questionnaire on respiratory symptoms (use in epidemiology). Scand J Respir Dis 53(4): van Kampen V, Merget R, Baur X Occupational airway sensitizers: An overview on the respective literature. Am J Ind Med 38(2): Van Nimmen NF, Poels KL, Veulemans HA Identification of exposure pathways for opioid narcotic analgesics in pharmaceutical production workers. Ann Occup Hyg 50(7): van Triel JJ, Arts JH, Muijser H, Kuper CF Allergic inflammation in the upper respiratory tract of the rat upon repeated inhalation exposure to the contact allergen dinitrochlorobenzene (DNCB). Toxicology 269(1): Vandenplas O, Ghezzo H, Munoz X, Moscato G, Perfetti L, Lemiere C, Labrecque M, L'Archeveque J, Malo JL What are the questionnaire items most useful in identifying subjects with occupational asthma? Eur Respir J 26(6): Vandenplas O, Hereng MP, Heymans J, Huaux F, Lilet-Leclercq C, Dezfoulian B, Grand JL, Thimpont J Respiratory and skin hypersensitivity reactions caused by a peptide coupling reagent. Occup Environ Med 65(10): Vanoirbeek JA, Tarkowski M, Ceuppens JL, Verbeken EK, Nemery B, Hoet PH Respiratory response to toluene diisocyanate depends on prior frequency and concentration of dermal sensitization in mice. Toxicol Sci 80(2): Vanoirbeek JA, Tarkowski M, Vanhooren HM, De Vooght V, Nemery B, Hoet PH Validation of a mouse model of chemical-induced asthma using trimellitic anhydride, a respiratory sensitizer, and dinitrochlorobenzene, a dermal sensitizer. J Allergy Clin Immunol 117(5):

161 142 Van-Wendel-de-Joode B, Brouwer DH, Vermeulen R, Van Hemmen JJ, Heederik D, Kromhout H DREAM: A method for semi-quantitative dermal exposure assessment. Ann Occup Hyg 47(1): Vermeulen R, Bos RP, Pertijs J, Kromhout H Exposure related mutagens in urine of rubber workers associated with inhalable particulate and dermal exposure. Occup Environ Med 60(2): Vermeulen R, Kromhout H, Bruynzeel DP, de Boer EM Ascertainment of hand dermatitis using a symptom-based questionnaire; applicability in an industrial population. Contact Dermatitis 42(4): Warshaw EM, Schram SE, Maibach HI, Belsito DV, Marks JG,Jr, Fowler JF,Jr, Rietschel RL, Taylor JS, Mathias CG, DeLeo VA, Zug KA, Sasseville D, Storrs FJ, Pratt MD Occupation-related contact dermatitis in North American health care workers referred for patch testing: Cross-sectional data, 1998 to Dermatitis 19(5): Wisnewski AV Developments in laboratory diagnostics for isocyanate asthma. Current Opinion in Allergy and Clinical Immunology 7(2): Woolhiser MR, Munson AE, Meade BJ Immunological responses of mice following administration of natural rubber latex proteins by different routes of exposure. Toxicol Sci 55(2): Woolhiser MR, Munson AE, Meade BJ Role of sensitization routes in the development of type I hypersensitivity to natural rubber latex in mice. Am J Ind Med Suppl 1: Zhang XD, Fedan JS, Lewis DM, Siegel PD Asthma-like biphasic airway responses in brown Norway rats sensitized by dermal exposure to dry trimellitic anhydride powder. J Allergy Clin Immunol 113(2): Zhang XD, Hubbs AF, Siegel PD Changes in asthma-like responses after extended removal from exposure to trimellitic anhydride in the brown Norway rat model. Clin Exp Allergy 39(11):

162 143 Zug KA, Warshaw EM, Fowler JF,Jr, Maibach HI, Belsito DL, Pratt MD, Sasseville D, Storrs FJ, Taylor JS, Mathias CG, Deleo VA, Rietschel RL Patch-test results of the North American Contact Dermatitis Group Dermatitis 20(3):

163 Appendices 144

164 145 Appendix 1: Supplemental Figures for Chapter 5 Skin Symptoms in Bakery and Auto Body Shop Workers: Associations with Exposure and Respiratory Symptoms

165 146 Figure 7 Auto Body Shop Workers Associations Between Average Isocyanate Exposure and Respiratory Symptoms, Shown In Smoothed Plots Stratified by Atopy. Data rugs indicate the distribution of observations by exposure level. (a) Asthma-like symptoms in atopic subjects (linear: NS; spline: NS), (b) Work-related chest tightness in atopic subjects (linear: NS; spline: df=3, p<0.05), (c) Asthma-like symptoms in non-atopic subjects (linear: p<0.05; spline: NS), (d) Work-related chest tightness in non-atopic subjects (linear: p<0.05; spline: df=3, p<0.05).

166 147 Figure 8 Bakery Workers Associations Between Average Wheat Exposure and Respiratory Symptoms, Shown in Smoothed Plots Stratified by Atopy. Data rugs indicate the distribution of observations by exposure level. (a) Asthma-like symptoms in atopic subjects (linear: NS; spline: NS), (b) Work-related chest tightness in atopic subjects (linear: NS; spline: NS), (c) Asthma-like symptoms in non-atopic subjects (linear: NS; spline: NS), (d) Work-related chest tightness in non-atopic subjects (linear: NS; spline: NS).

167 148 Table 23 Results of Generalized Linear Models Describing the Simple Relationship Between Exposure, Respiratory Symptoms, Atopy and Specific IgE. Each Reported Prevalence Ratio (PR) was Estimated From a Separate Model. Models Adjusted for Age and Sex. Symptom Models Additionally Adjusted for Smoking. (WR=Work-related) Independent Variable Dependant Variable PR (95% CI) Auto Body Repair Workers (n=473): Average Isocyanate Exposure Asthma-like Symptoms 1.12 ( ) (µg-nco*m -3 ) WR Chest Tightness 1.71 ( ) Atopy 0.83 ( ) HDI-Specific IgE 10.0 (1.6-72) Atopy Asthma-like Symptoms 1.40 ( ) WR Chest Tightness 2.61 ( ) HDI-Specific IgE Asthma-like Symptoms 1.13 ( ) WR Chest Tightness 4.89 (1.3-18) Bakery Workers (n=723): Average Wheat Exposure (µg*m -3 ) Asthma-like Symptoms 0.89 ( ) WR Chest Tightness 0.92 ( ) Atopy 0.91 ( ) Wheat-Specific IgE 1.12 ( ) Atopy Asthma-like Symptoms 2.67 ( ) WR Chest Tightness 4.11 (1.4-12) Wheat -Specific IgE Asthma-like Symptoms 2.6 ( ) WR Chest Tightness 15.7 (5.5-45)

168 149 Appendix 2: Interviewer-Administered Questionnaire for Chapter 6 Skin and Respiratory Symptoms Among Workers with Suspected Work-Related Disease

169 Work-related skin and respiratory symptoms and disease Part 1. Demographic Information 1. Study ID 2. Name (first) (mi) (last) 3. Address 4. City 5. Province 6. Postal Code 7. Interviewer Initials 8. Interview Date MM DD YY 9. Birth Date 10. Place of Birth 11. Sex Male Female Last updated 05/12/2009