Respiratory health and breath condensate acidity in sawmill workers
- First Online:
- Cite this article as:
- Ljubičić Ćalušić, A., Varnai, V.M., Čavlović, A.O. et al. Int Arch Occup Environ Health (2013) 86: 815. doi:10.1007/s00420-012-0817-x
- 206 Views
The aim of the study was to evaluate exhaled breath condensate acidity (EBC pH) as a biomarker of airway response to occupational respiratory hazards present in sawmill.
Sixty-one sawmill workers in total (26 from Sawmill 1 and 35 from Sawmill 2) provided EBC samples at the beginning and at the end of the working week. Respiratory symptoms, lung function, bronchodilator test and atopy status were assessed. Occupational environment was checked for the levels of respiratory hazards.
Airborne dust concentrations were below threshold limit value. Endotoxin in Sawmill 1 and Sawmill 2, and moulds in Sawmill 1 were at the levels able to induce inflammatory response in the airways. Mould levels were 2.5 times higher in Sawmill 1 than in Sawmill 2. Compared to Sawmill 2 workers, lower spirometry values, higher prevalence of dry cough and positive bronchodilator test were found in Sawmill 1 workers. Monday EBC pH values did not differ between sawmills, but declined after one working week in Sawmill 1 workers (from 7.88 to 7.49, P = 0.012) and not in Sawmill 2 workers. Similar results were obtained when only respiratory healthy non-smokers were analysed. Monday-to-Friday change of other respiratory parameters was not observed.
The results suggest EBC pH as a biomarker of acute respiratory effects related to occupational exposure to respiratory hazards in sawmills, presumably increased mould levels. The effect was present even at subclinical level, namely in respiratory healthy subjects. The long-term health implications remain unclear and should be evaluated in a follow-up study.
KeywordsAtopyBronchodilator testEBC pHMouldsSmokingWood dust
Exposure to wood dust can cause a variety of respiratory disorders in sawmill workers, including asthma, chronic bronchitis, chronic obstructive pulmonary disease (COPD), hypersensitivity pneumonitis or organic dust toxic syndrome, with atopic and non-atopic inflammation in the airways as predominate underlying pathological mechanism (Jacobsen et al. 2010). Wood dust, bacteria and moulds are considered as major harmful factors related to the above-mentioned disorders. Published data have also shown that prolonged exposure to dust from ‘hardwood’ species, such as oak and beech, is related to increased prevalence of carcinoma of nasal cavities and paranasal sinuses, and IARC listed wood dust as Group 1 carcinogen to humans (IARC 1995).
Early subclinical inflammatory changes in the airways are not easy to detect in clinical practice or epidemiological studies in patients with asthma, COPD, cystic fibrosis, bronchiectasia, pneumonia, sarcoidosis, tuberculosis, reactive airway distress syndrome, and systemic sclerosis with lung involvement (Koutsokera et al. 2008). Routinely applied clinical methods are not sensitive enough (e.g. spirometry), while others are invasive (e.g. bronchoalveolar lavage) or technically demanding (e.g. induced sputum). There is a need for a validated and standardised non-invasive and technically and financially not too demanding method, capable of early detection of airway inflammation. EBC pH was extensively studied as a new marker of inflammatory changes in the airways of above-mentioned respiratory diseases (Koutsokera et al. 2008). EBC pH decrease is believed to be related to endogenous acidification of airways as a consequence of inflammatory process (Ricciardolo et al. 2004), and there are published data that implicate its correlation with severity of inflammation in chronic respiratory disease. (Hunt et al. 2000; Kostikas et al. 2002; Hunt 2006; Hillas et al. 2011; Papaioannou et al. 2011). Due to its availability, low price and relatively good validation status in comparison with other EBC biomarkers, EBC pH is also suggested to be a tool in epidemiological studies, assessing health effects of the occupational exposure to respiratory hazards (Corradi et al. 2010). Several studies addressed this issue so far, evaluating EBC acidity in relation to occupational exposure to grain dust (Do et al. 2008b), welding fumes (Boyce et al. 2006; Fireman et al. 2008; Gube et al. 2010; Hoffmeyer et al. 2011), isocyanates (Ferrazzoni et al. 2009) and asbestos (Chow et al. 2009). As pointed out by Corradi et al. (2010), these studies obtained contradictory results, even in the workers exposed to welding fumes (Boyce et al. 2006; Fireman et al. 2008; Gube et al. 2010). Regarding exposure to organic dust, only one study has been described in the available literature, reporting negative correlation between EBC pH and daily duration of exposure in grain workers (Do et al. 2008b).
The aim of the study was to evaluate EBC pH as a biomarker of airway response to occupational respiratory hazards in a sawmill. EBC pH was measured before and after one-week exposure to wood dust, on Monday before the work shift and on Friday after the work shift. Standard parameters of respiratory health, that is, respiratory symptoms, lung function and bronchial hyperreactivity were also assessed, as well as atopy status. Occupational environment was checked for the levels of organic dust components, namely dust, moulds and endotoxin.
Subjects and study protocol
The study was designed as a cross-sectional study in accordance with Declaration of Helsinki and was approved by the Ethical Committee of the Institute for Medical Research and Occupational Health, Zagreb. All subjects employed at working places with exposure to wood dust were invited to participate. Recruitment was carried out on a voluntary basis after an invitation given by their employers. All participants signed the informed consent form and they were free to leave the study at any time.
The study involved sawmill workers from two sawmills (Sawmill 1 and Sawmill 2) located near Zagreb, Croatia, who reported ≥20 h per week exposure to wood dust. They all worked from Monday to Friday, 8 h per day. Response rates were 92 % for Sawmill 1 and 86 % for Sawmill 2. Exclusion criteria were acute respiratory disease (e.g. common cold) during the last 2 weeks, gastric disturbances (e.g. gastro-oesophageal reflux) and medical conditions that contraindicated scheduled clinical tests or affected reliability of test results. The absence of gastric disturbances was determined by a medical interview. Gastro-oesophageal reflux disease was ruled out if following questions were negatively answered: ‘Have you ever experienced “heartburn” or regurgitation symptoms? Have you ever used medications for these symptoms?’ Out of 65 workers willing to participate in the study, three were excluded due to acute upper respiratory illness and one due to gastro-oesophageal reflux disease. Finally, 61 sawmill workers were included in the study, 26 from Sawmill 1 and 35 from Sawmill 2.
On Monday morning (before the work shift), between 5 and 7 a.m., following procedures were undertaken in precise order: questionnaire, venous blood sampling, EBC sampling, spirometry, bronchodilator test and skin prick testing. On Friday (after the work shift), between 1 and 3 p.m., EBC sampling, spirometry and bronchodilator test were repeated.
Both sawmills process yearly around 20,000 m3 of non-debarked logs. Considering the wood species, Sawmill 1 processes both hardwood (2/3 of annual production) and softwood species (1/3 of annual production), while Sawmill 2 processes only hardwood species. Hardwood species include beech (Fagus silvatica L.) and oak (Quercus robur L.) in Sawmill 1, and beech, oak and ash (Fraxinus angustifolia) in Sawmill 2. Softwood species includes fir (Abiesalba Mill.) and spruce (Picea abies). Hardwood and softwood species are processed simultaneously at different working stations in the same hall of Sawmill 1. Both sawmills have suction systems for wood dust near the saws, and cleaning is performed manually at the end of the week.
Air samples for total and respirable dust and for endotoxin measurements were sampled during the summer 2007 (Sawmill 1) and summer 2008 (Sawmill 2), and airborne fungi were collected over a 1-year period (end of April 2008–beginning of February 2009) at two-month interval at Sawmill 1 and Sawmill 2. Locations for sampling in sawmills were chosen near sawing places. Air temperature (°C), relative humidity (%) and airflow rate (m/s) were determined with a Kestrel® 4000 Pocket Weather Tracker (Nielsen-Kellerman Co, USA) next to air sampling places during each sampling procedure. Air samples for total and respirable dust and for endotoxin measurements were sampled during one working shift (8 h) on glass fibre filter paper (Whatman QM-A, USA) with stationary samplers Casella (Bedford, UK). The air flow was 2 L/min (EN ISO 10882-1:2001). Air filters were stored at −20 °C until analysed. Filters were weighed before and after sampling in a controlled laboratory at air temperature of 22 °C and relative humidity of 45 % (±5 %) using gravimetric method (Hauptverband der gewerblichen Berufsgenossenschaften, Zentralstelle für Unfallverhütung und Arbeitsmedezin, ZH 1/120.41) with microscale Mettler-Toledo MX-5 (Switzerland).
Air samples for determining mould concentrations were collected with Merck MAS-100 (Merck KgaA, Darmstadt, Germany) device onto commercial Sabouraud medium enriched with 50 mg/mL streptomycin and 20,000 IU penicillin (Biolife, Milan, Italy). The impaction velocity of the sampler was approximately 10.8 m/s, and airflow rate was 100 L/min. A volume of 50 L (30 s) was chosen for sampling. Medium was then incubated for 4–7 days at 25 ± 2 °C. Grown mould colonies were counted at the time intervals of 3, 5 and 7 days, and expressed as colony-forming units per m3 (CFU/m3) as described in detail previously (Šegvić Klarić et al. 2012).
Endotoxin was analysed using the end point Limulus amoebocyte lysate method (LAL) with the commercial kit for endotoxin assay (Endochrome, Charles River Endosafe, USA), as previously described (Varnai et al. 2004). Sterile and pyrogen-free glassware and microplates (Greiner Labortechnik, GmbH, Germany) were used. Each sample was analysed in duplicate. Optical density was read at 405 nm on a Personal Laboratory microplate analyser (Iason, Graz, Austria). Endotoxin concentrations were read from the standard curve and expressed in endotoxin units per m3 (EU/m3).
Questionnaire and body mass index
A questionnaire based on ‘Organic dust questionnaire’ (Rylander et al. 1990) was completed by a physician interviewing each subject. The questionnaire recorded reports on the occurrence of following symptoms in the past year not related to acute respiratory illness: symptoms of rhinitis, asthma and/or chronic bronchitis, dry cough, phlegm (not compatible with criteria for chronic bronchitis) and general symptoms. Rhinitic symptoms included rhinorrhea, nasal itching and nasal obstruction (not related to common cold). Asthmatic symptoms included wheezing and/or dyspnoea. Chronic bronchitis was defined as an everyday cough with phlegm during at least 3 months per year, for two consecutive years. General symptoms include fever and joint pain.
Smoking habit was expressed as dichotomous variable (smokers and non-smokers) and as smoking index for current smokers (number of cigarettes smoked per day multiplied by number of smoking years). Subjects who never smoked and ex-smokers were designated as non-smokers. All ex-smokers quit smoking at least 1 year before enrolment in this study. Seniority was expressed as years of employment at the present workplace, namely Sawmill 1 or Sawmill 2. Education was defined in two categories: unqualified workers—primary school (≤8 years of education) and qualified workers—secondary school (8 + 4 years of education). Body mass index (BMI) was calculated from height and weight according to the formula: weight [kg]/height [m]2. Subjects with BMI ≥30 kg/m2 were categorised as obese.
Forced vital capacity (FVC), forced expiratory volume in 1 s (FEV1), FEV1/FVC ratio, forced expiratory flow at 50 % of FVC (FEF50), forced expiratory flow between 25 and 75 % of FVC (FEF25–75), and forced expiratory flow at 75 % of FVC (FEF75) were determined by a standard method according to the referent values CECA II (Quanjer et al. 1993) using the spirometer Pneumoscreen II (Jaeger, Wurtzburg, Germany). At least three measurements were recorded per subject, and the best value was used for analysis. Ventilatory function parameters were expressed and analysed as a percentage of referent values and as the number of subjects with ventilatory function parameters below lower limit of normal (LLN, calculated for each subject as a predicted value minus 1.645 residual standard deviation (RSD) for FVC, FEV1 and FEV1/FVC ratio (Pellegrino et al. 2005). RSD values of 0.61 for FVC, 0.51 for FEV1 and 7.17 for FEV1/FVC ratio were taken from CECA II prediction equations.
After initial spirometry testing described above, bronchodilation was performed with two breaths of salbutamol aerosol (400 μg), and 20 min afterwards, spirometry was repeated. Test was considered positive if FEV1 increased by ≥12 % and by ≥0.2 L compared to the initial value (Pellegrino et al. 2005).
Serum total IgE antibodies were measured from venous blood samples by the ELISA method (IASON, Austria) (Aberer and Kranke 2002). IgE levels were expressed in k IU/L, and values of >150 k IU/L were considered as elevated.
Skin prick testing
In all subjects, skin prick testing (SPT) was performed by standard method (Dreborg and Frew 1993) with a panel of 12 common inhalatory allergens: grass pollen mixture, birch, hazel, weed (Ambrosia elatior, Artemisia vulgaris) pollens, mites (Dermatophagoides pteronyssinus, Dermatophagoides farinae and Lepidoglyphus destructor), cat, dog and moulds (Cladosporium herbarum,Alternaria alternata) (Allergopharma, Germany). Workers from Sawmill 1 were also tested with commercial allergen preparations of beech (Fagus sylvatica), spruce (P. abies) and oak (Quercus robur), and workers from Sawmill 2 with commercial allergen preparations of beech, oak and ash (F. angustifolia) (Institute of Immunology, Zagreb, Croatia). SPT included testing with positive control solution (10 mg/mL of histamine hydrochloride) and negative control solution (buffer). Skin reaction (wheal) was evaluated after 15 min. The mean skin reaction (mean wheal diameter) was calculated according to the formula (D + d)/2, where D represents the largest longitudinal diameter and d its mid-point orthogonal diameter in mm. The result was considered positive if mean wheal diameter was larger than the negative control by more than 3 mm. Atopic status was defined as a positive reaction to at least one tested common inhalatory allergen in the skin prick test.
EBC sampling and pH measurement
EBC sampling was performed according to the recommendations of ATS/ERS Task Force on Exhaled Breath Condensate (Horvath et al. 2005). Subjects provided samples breathing tidally into a commercial condenser (EcoScreen; Jaeger, Germany) for 10 min through a mouthpiece and a two-way non-rebreathing valve that also prevented saliva contamination due to integrated saliva trap. Spoken recommendations were given to all subjects regarding the time frame of 1 h prior to EBC sampling in which they should not smoke, eat and drink anything else except non-sparkling water. The condensate was collected into a Teflon-coated tube, which was disinfected and rinsed with distilled water and wiped before sampling. Before sampling, subjects rinsed the mouth with tap water. All subjects wore nose clips. pH was measured within 8 months from sampling (samples were kept at −20 °C). After being taken out from the ice, samples were left at room temperature for thawing and measured after 10-min argonisation (bubbling through the sample at a flow rate of 350 mL/min). pH meter MP 220 (Mettler Toledo, Switzerland) with standard glass micro-electrode InLab Micro Pro (Mettler Toledo, Switzerland) (range of 0.00–14.00, accuracy ±0.01) was used, calibrated with standard buffers (Mettler Toledo, Switzerland) at two points (pH 7.00 and pH 4.01).
Statistical analysis was performed using software Stata/SE 11.1 for Windows (StataCorp LP, USA). The results were considered statistically significant at value of P < 0.05. Differences in environmental parameters of two sawmills, which were non-normally distributed, were analysed by Mann–Whitney U test. Univariate analysis of variables related to workers’ characteristics included Student’s t test for independent normally distributed data (age, spirometry data, EBC pH weekly change), Mann–Whitney U test for independent non-normally distributed data (smoking index, years of smoking, cigarettes smoked per day, seniority, BMI, total serum IgE, EBC pH), and Pearson’s χ2 test or Fisher’s exact test (if any of the expected values in contingency table was ≤5) for independent categorical data (smoking habit, educational level, urban/rural lifestyle, prevalence of respiratory symptoms, below-normal spirometry values, positive bronchodilator test, elevated IgE, positive SPT). Differences between repeated measurements (Monday and Friday values) were tested by Student’s t test for paired data (normally distributed variables), Wilcoxon matched-pairs signed-ranks test (non-normally distributed variables) and McNemar’s χ2 test (categorical variables). Spearman’s correlation was used to analyse the association of EBC pH with age, seniority, spirometry values, total serum IgE and smoking index. The association of EBC pH weekly change with spirometry parameters was tested by Pearson’s correlation and with smoking index by Spearman’s correlation.
Multivariate analysis was used to test the correlation between dry cough, spirometry values and EBC pH weekly change (as dependent variables) and employment at Sawmill 1 versus Sawmill 2, adjusted for potential confounders. Multiple logistic regression analysis with dry cough as dependent variable was adjusted for age, seniority, FVC, smoking status and positive SPT. Multiple linear regression models with spirometry parameters (FVC, FEV1, FEF50, FEF25–75 and FEF75) as dependent variables were adjusted for age, seniority, educational level, smoking status and presence of chronic lower airway symptoms (asthma and/or chronic bronchitis). Multiple linear regression model with EBC pH weekly change as dependent variable was adjusted for age, educational level, smoking status, presence of asthma and/or chronic bronchitis symptoms, positive SPT and baseline (Monday) EBC pH values. Seniority and spirometry parameters (FVC or FEV1) could not have been included as predictors in the models with EBC pH weekly change as dependent variable due to high collinearity with other predictors.
Total and respirable dust, endotoxin and moulds in the air of analysed sawmills
Wood dust (mg/m3)
General characteristics of study subjects
General characteristic of study subjects
Number of subjects, n
Age (years), arithmetic mean ± SD
41 ± 13
31 ± 7
Smokers, n (%)
Cigarettes per day, median (IQ range)
Years of smoking, median (IQ range)
Smoking index, median (IQ range)
Seniority (years), median (IQ range)
Urban/rural residence, n (%)
Education level, n (%)
≤8 years of education
12 years of education
BMI (kg/m2), median (IQ range)
Obese (BMI ≥30 kg/m2), n (%)
Respiratory and general symptoms, markers of atopy and lung function
Prevalence of respiratory symptoms and objective atopy markers in sawmill workers
Cough with phlegma
The prevalence of elevated IgE was the same (17 %) in both sawmills. The prevalence of workers with positive SPT reactions to at least one common inhalatory allergen also did not differ significantly between sawmills. In both sawmills, the most prevalent was sensitisation to dust mites (D. pteronyssinus and D. farinae), storage mite L. destructor and weed Artemisia vulgaris. There were no positive SPT reactions to tested wood allergens (oak, beech, ash and spruce).
In Sawmill 1 compared to Sawmill 2 workers, significantly lower spirometry values were found. FVC and FEV1 values measured on Monday were lower for 17 % (96 vs. 113 % for FVC and 99 vs. 116 % for FEV1, respectively), and the difference remained significant when adjusted for potential confounders, that is, age, seniority, educational level, smoking status and presence of chronic lower airway symptoms). Four subjects in Sawmill 1 (15 %) and only one subject in Sawmill 2 (3 %) had FVC and/or FEV1 values below LLN. One subject from Sawmill 2 had FEV1/FVC ratio below LLN. There were no Monday-to-Friday differences in spirometry values in either sawmill.
Higher prevalence of positive bronchodilator test was observed in workers from Sawmill 1 compared to Sawmill 2 workers on Monday (19 vs. 3 %, respectively, P = 0.083) and on Friday (23 vs. 3 %, respectively, P = 0.035). There were no Monday-to-Friday differences in the prevalence of positive bronchodilator test in either sawmill.
Analysis showed no difference in EBC pH between workers from two sawmills on Monday. Monday EBC pH values were not correlated with age, seniority, spirometry values, total serum IgE and BMI or with smoking index. Nine workers that reported asthmatic and/or chronic bronchitis symptoms had significantly lower EBC pH values compared to workers without these symptoms (median, range) (7.65, 7.38–8.05 vs. 7.96, 7.41–8.43, respectively, z = 2.15, P = 0.032).
The results obtained by univariate analysis were confirmed by multiple regression analysis in which relationship between Monday-to-Friday EBC pH change and occupational exposure (Sawmill 1 vs. Sawmill 2) was adjusted for age, educational level, current smoking status, presence of asthma and/or chronic bronchitis symptoms, positive SPT and baseline (Monday) EBC pH values (Pmodel < 0.001, adjusted R2 = 0.314). The analysis showed correlation between EBC pH change and occupational exposure, that is, workers at Sawmill 1 had 0.29 units greater weekly decline in EBC pH than Sawmill 2 workers (95 % CI 0.08–0.49, β = 0.36, P = 0.009). EBC pH change was negatively correlated with baseline EBC pH values (β = −0.561, P < 0.001), namely the higher the EBC pH values recorded on Monday, the greater the weekly decline in EBC pH. Other predictors were not found to be correlated with EBC pH change. EBC pH change was not correlated with spirometry parameters (Pearson’s correlation), seniority and smoking index (Spearman’s correlation).
The aim of this study was to evaluate EBC pH as a biomarker of airway response to occupational respiratory hazards present in two sawmills (Sawmill 1 and Sawmill 2) located near Zagreb, Croatia. Significant decrease in EBC pH after weekly wood dust exposure was observed, but only in workers from Sawmill 1.
Environmental checking in both sawmills showed the presence of hazardous constituents of wood dust at levels above known thresholds for respiratory health effects, specifically airway inflammation. Average total and respirable dust concentrations were lower than 1 mg/m3, and thus well below Croatian and European limits of exposure (Kos et al. 2004; Kauppinen et al. 2006). However, airborne endotoxin and mould levels were higher than thresholds related to inflammatory changes in the airways (100 EU/m3 and 10,000 CFU/m3, respectively) (Rylander and Carvalheiro 2006; Heida et al. 1995), with significant difference between sawmills regarding mould levels. While all measured endotoxin values in both sawmills were higher than 100 EU/m3, mould concentrations were significantly higher (2.5 times) in Sawmill 1 compared to Sawmill 2, and hazardous mould concentrations (>10,000 CFU/m3) were observed only in this sawmill (20 % of samples in Sawmill 1). The second important difference between sawmills was the processing of different wood species. Softwood species (fir and spruce), which are considered as wood species with higher inflammatory potency in airways compared to hardwood species (Demers et al. 1997; Hessel et al. 1995; Douwes et al. 2001), were processed only in Sawmill 1.
Another issue was the difference in age, educational status and smoking history between workers from two sawmills. Workers from Sawmill 1 were on average 10 years older and had about four times higher proportion of non-qualified workers than workers from Sawmill 2. Sawmill 1 workers had approximately 2 times higher smoking index in comparison with workers from Sawmill 2, primarily because they were older and had longer smoking history. Keeping in mind differences in occupational exposure and individual characteristics of the tested workers, this study revealed higher frequencies of dry cough, lower spirometry values and higher frequencies of positive bronchodilator test in workers from Sawmill 1 in comparison with Sawmill 2 workers. These results suggest higher level of airway inflammation in Sawmill 1 workers that worsens after weekly exposure to wood dust, which is supported by Monday-to-Friday decrease of EBC pH. Here, we would like to point out that our study design permits us only to hypothesise that observed EBC pH change is due to inflammation-induced airway acidification, since direct evaluation of inflammatory process (e.g. identification of inflammatory cells in induced sputum) was not performed.
Interestingly, EBC pH measured on Monday did not reveal differences between workers from two sawmills, although other lung parameters indicated better respiratory health status in Sawmill 2 workers. On the other hand, only EBC pH was able to detect airways’ response to weekly exposure to occupational respiratory hazards, while spirometry and bronchodilator test were unable to detect any changes during this short time period. The results also showed that neither baseline EBC pH values (measured on Monday) nor EBC pH weekly changes were correlated with spirometry values and bronchodilator response, which seemed to better reflect long term than acute respiratory effects of hazardous exposures in the present study.
EBC pH Monday’s values in workers from both sawmills were in the range of values reported for respiratory healthy adult population (Varnai et al. 2009). As mentioned above, EBC pH was not related to age, which was shown by other authors as well (Paget-Brown et al. 2006; Brooks et al. 2006; Cruz et al. 2009). It was lower in subjects reporting asthma-like or chronic bronchitis symptoms in the past year, but was not related to atopy status, BMI, smoking status or cumulative cigarette smoke exposure (smoking index) in current smokers. The atopy status, expressed as the prevalence of positive SPT reactions to common inhalatory allergens, was similar in both sawmills and close to the prevalence found in a population-based study in adult male Croatian subjects (27 %) (Macan et al. 2007). The lack of correlation with BMI could be explained by low number of obese subjects in the present study (five workers), namely the influence of nutritional status on EBC pH in the published literature, in terms of EBC acidification, was found for obese subjects (Carpagnano et al. 2008; Do et al. 2008b). Lower EBC pH values in current smokers compared to non-smoking subjects have been repeatedly described (Do et al. 2008b; Koczulla et al. 2010; Ljubičić Ćalušić et al. 2011; MacNee et al. 2011), and it has been found that cumulative smoking dose is inversely correlated with EBC pH (Do et al. 2008b). Nevertheless, some studies (Bloemen et al. 2009; Hauswirth et al. 2008; Do et al. 2008a; Yeh et al. 2008), including the present one, failed to confirm these findings. Based on the results of study of Koczulla et al. (2010) and our previous study (Ljubičić Ćalušić et al. 2011), we hypothesise that too short a period of smoking abstinence before EBC sampling could be one of the possible reasons for contradictory findings regarding the influence of smoking on EBC pH. In our experimental study, we showed that EBC pH values obtained from 15 respiratory healthy smoking volunteers after overnight tobacco abstinence were significantly lower compared to healthy non-smokers, probably as a result of tobacco-induced chronic inflammation in the airways (Ljubičić Ćalušić et al. 2011). EBC pH values, however, increased after smoking one cigarette and reached the values found in non-smokers. This effect lasted for 2-h observation period. Mild elevation of EBC pH shortly after smoking is not easy to explain. It could be related to both immunological and non-immunological mechanisms, such as a change in dissociation constant of buffers in the airway lining fluid induced by smoking-related raise of airway temperature, ammonia present in tobacco smoke, dilution changes of EBC induced by physical and/or chemical factors of cigarette smoke and acute immunosuppressive effects of inhaled tobacco smoke. These potential mechanisms are extensively discussed in our previous publication (Ljubičić Ćalušić et al. 2011). In the present study, it proved unattainable to require overnight abstinence from smoking, and the majority of workers (approximately 75 %) smoked within less than 2 h before EBC sampling, both on Monday and Friday. Although acute effects of smoking were not consistently reported in the published literature (Papaioannou et al. 2010), we took a precautionary approach and assumed that in our study, potential group acute effects of recent smoking could mask its chronic effects, so they could not be adequately evaluated in the present study.
Both environmental and workers’ characteristics could have been responsible for the observed difference in EBC pH change after weekly occupational exposure to respiratory hazards in workers from two sawmills. As mentioned above, workers from two sawmills were dissimilar regarding age, smoking history, educational level and respiratory health status. To evaluate relative contribution of these factors, the analysis of association of EBC pH weekly change and occupational exposure (employment at Sawmill 1 vs. Sawmill 2) was adjusted for potential confounders. It was found that the only parameters significantly influencing EBC pH weekly decline were occupational exposure (Sawmill 1 vs. Sawmill 2) and baseline EBC pH values. Comparing Monday and Friday EBC pH values in respiratory healthy non-smokers from each sawmill, significant drop in EBC pH was again observed only in Sawmill 1 workers. A weekly decline of 0.49 logarithmic units found in these subjects was greater than the difference in Monday EBC pH values between workers with asthmatic and/or chronic bronchitis symptoms and workers without these symptoms (7.96 vs. 7.65, i.e. 0.31 units). The results suggest that differences in occupational exposure are the main reason for different weekly response of EBC pH. Since the average wood dust concentrations in the air were lower than the official limit of exposure in both sawmills, the difference in types of processed wood could not by easily related to the observed effect on EBC pH. On the other hand, significantly higher mould concentrations in Sawmill 1 compared to Sawmill 2, reaching the levels hazardous for respiratory system, are the potential cause of EBC pH drop after one-week occupational exposure in Sawmill 1. Irritative symptoms, such as dry cough, could also be related to this exposure. Described results also suggest that EBC pH in this study was responsive to acute, short-term effects of occupational exposure and not to its long-term effects. EBC pH was not correlated with seniority and seemed to return to more alkaline, physiological level after two-day weekly rest period. Similar results were described in grain workers in which EBC pH was negatively correlated with daily exposure to grain dust but not to the duration of employment in the grain industry (Do et al. 2008b).
The present study has several limitations. Short duration of employment at the present workplace (3.5 years on average) limits the evaluation of long-term effects of occupational exposure on respiratory health. Additionally, the impact of previous employment(s) on studied respiratory and atopic parameters is difficult to assess, especially in rural population which is exposed to various respiratory hazards from early age (Von Mutius et al. 2000; Braun-Fahrlander et al. 2002; Radon 2006). Environmental monitoring in sawmills was performed by stationary samplers during one season (summer) for total and respirable dust and for endotoxin measurements, and airborne fungi were collected over a 1-year period (end of April 2008–beginning of February 2009) at 2-month interval at Sawmill 1 and Sawmill 2. Personal sampling, preferably on multiple occasions during other seasons, would improve the assessment of exposure–effect relationship. This is especially true for Sawmill 1 where both hardwood and softwood species were processed simultaneously in the same hall. A rather small size of study groups (26 and 35 workers in Sawmill 1 and Sawmill 2, respectively) limits the interpretation of results, especially of multivariate analysis. Nevertheless, univariate and multivariate analyses of factors influencing Monday-to-Friday change in EBC pH yielded similar results, supporting the final conclusions on the effect of occupational exposure to respiratory hazards in the sawmill on EBC pH. Assumed acute effect of smoking on EBC pH is discussed above. It is possible that smoking shortly before EBC sampling masked not only chronic effects of smoking on EBC pH, but of occupational exposure as well. This potential issue was overcome by adjusting the analysis for current smoking status and by repeating the analysis in a subsample of healthy non-smoking workers.
EBC pH was so far rarely evaluated as a biomarker of airways’ response to occupational exposure to respiratory hazards. Studies were focused mostly on welders, but provided contradictory results mainly due to the differences in methodology, tested subjects and occupational exposure (welding fumes with different metal content) (Boyce et al. 2006; Fireman et al. 2008; Gube et al. 2010). However, study of Fireman et al. (2008) showed a greater proportion of neutrophils in induced sputum of welders with lower values of EBC pH compared to welders with higher EBC pH values (71 vs. 42 %, respectively), supporting the role of EBC pH as a biomarker of airway inflammation. The study of Do et al. in grain workers (Do et al. 2008b), as well as the present study in sawmill workers, support the role of EBC pH as a biomarker of airway's response to occupational exposure to organic dust (presumably increased mould levels) in a sawmill. The results also suggest EBC pH is a biomarker of acute rather than long-term respiratory effects of occupational exposure, which can be detected even at subclinical level, namely in respiratory healthy workers. The long-term implications of this reversible, short-term EBC pH change on respiratory health remain unclear and should be evaluated in follow-up studies.
The study was done within the scientific project No. 022-0222411-2410, financially supported by the Ministry of Science, Education and Sports of Republic of Croatia. Project was approved by authorized Ethical Committees.
Conflict of interest
The authors declare that they have no conflict of interest.