INTRODUCTION

Global burden of disease analysis for chronic obstructive pulmonary disease (COPD) in 2019 showed a global prevalence of 10.3% (8.2%–12.8%) in the population between 30–79 years of age.1 Occupational exposure to dust or smoke is a substantial contributor to developing COPD in global settings, resulting in 16.5% (14.6%–18.7%) of disability-adjusted life years (DALYs) due to asthma.1,2 Inhalational exposure to wood dust can occur during the collection and processing of raw wood material and furniture making, thereby resulting in respiratory system-related morbid conditions.3,4 Chemical components in wood dust, such as plicatic acid, monoterpenes, and endotoxin exposure, are implicated in hypersensitization of the pulmonary epithelium. Chronic wood dust exposure results in many respiratory ailments, from allergic manifestations involving the upper respiratory tract to restrictive and obstructive lung pathology, and, in developing cancers involving the sinonasal tract.5–10

While several systematic reviews have established an association between wood dust exposure and upper respiratory tract cancers, there is insufficient pooled evidence to establish a causal link between wood dust and lung function deterioration.3,4,9 Inconclusive results from individual studies based on smaller sample sizes warrant a systematic review of available evidence. This systematic review tests the hypothesis that occupational wood dust exposure is associated with declines in lung function test parameters among wood industry workers.

METHODS

This systematic review was conducted and reported according to Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA 2020).11 The protocol for this review is registered in PROSPERO under CRD42022346220.

We considered adult males and females exposed to wood dust that is generated during the collection, processing, and manufacture of wood products. Studies reporting wood dust exposure, irrespective of geographical settings, type of wood processed, scale of wood industry (small versus large), were considered. Information on exposure to wood dust was retrieved as reported in the studies, regardless of exposure duration. The unexposed population considered for this review comprised the comparator population reported in studies, including administrative staff associated with wood industries at the study site, other industry settings without wood dust, low levels of dust exposure, or community-based controls.

Outcomes considered for this review include lung function parameters such as forced expiratory volume at first second (FEV1), forced vital capacity (FVC), and FEV1/FVC ratio. Also, peak expiratory flow (rate) (PEF/PEFR) measurements from spirometry or peak flowmeter were considered as additional outcomes. Study designs considered for the review include cross-sectional, case-control, and cohort studies. Case studies, case series, narrative reviews, letters to the editor, opinion pieces, and correspondence were excluded.

Literature search was conducted in the MEDLINE database via PubMed and Google Scholar using key words, such as ‘wood dust,’ ‘occupational exposure,’ and ‘pulmonary function tests,’ from the inception until 28 June 2022, followed by a rerun of the literature search on 2 January 2024 for additional articles. Published original research articles retrieved from these databases were imported into the Rayyan software for screening.

Screening of articles

Titles and abstracts were screened in duplicates by two primary reviewers (KMDS and RRN). Studies were screened for eligibility if they included workers employed in occupations generating wood dust and reported performing lung function assessments. A third reviewer (AK) resolved the discrepancies that arose during screening of the title and abstract. Next, the full text of articles was screened to assess exposure setting to wood dust, reporting of at least one lung function parameter, and appropriateness of study design. Records for which full texts were not available for retrieval or were published in languages other than English were excluded. The screening process, along with the reasons for exclusion, are provided in the PRISMA flow diagram (Fig. 1).

FIG 1. PRISMA flow diagram

Export to PPT

Data were extracted from full-text articles screened independently by a team of reviewers (KMDS, RRN, AK, and EM) using a pre-designed data extraction sheet (Google sheet). Following details were extracted: administrative details with name of the first author and year of publication; study details with title, design, setting, participants, controls if any, and sample size; and exposure details such as type and exposure duration, dust concentration measurements, tools for respiratory symptoms and lung function assessment, FEV1, FVC, FEV1/FVC and PEFR among wood workers and comparator groups, along with measures of summary and deviation. Quality of studies was assessed using an adapted version of the Newcastle–Ottawa quality assessment scales for cross-sectional and cohort designs, and were scored as high, medium, and low risk of bias as per the scoring manual.12

Data analysis

Study characteristics, including basic demographic details, type of wood, study tools, allergic/respiratory symptoms, and pulmonary function test measurements, were tabulated. Outcome measures of pulmonary function tests were harmonized as mean and standard deviation (SD) using the Cochrane formulae.13 For full-texts, where mean and SD of outcome measures were provided at the subgroup level, a pooled mean and SD were derived using an online statistical calculator for the exposed and unexposed populations.14 Data were analysed using ‘Stata Statistical Software: Release 14 College Station, TX: StataCorp LP’. Meta-analysis was conducted using a random effects model based on the DerSimonian and Laird approach, and the pooled mean difference (with 95% CI) was reported. Statistical heterogeneity was assessed using the Chi-square test (p<0.05) and I2 statistics. The source of heterogeneity in studies was explored through subgroup analysis of variables such as the type of wood handled by the workers and the risk of bias of the studies.

RESULTS Study characteristics

This systematic review identified 55 primary studies done across various global settings between 1986 and 2023, which included 8635 wood workers and 5367 controls (Table 1).15–69 Of these, 20 (36.4%) were from Asia, followed by 14 (25.5%) from Europe, and 12 (21.8%) from Africa. All, except one study, had a cross-sectional study design, with Bolund et al., adopting a cohort study design.53 Thirty-seven studies men-tioned the type of wood industry from which workers were recruited. Of these, 21 (56.8%) were from saw mills, and 16 (43.2%) from the furniture industry. Among the unexposed population, 19/55 (34.6%) studies included office staff associated with the respective wood industry as their comparator group.

TABLE 1. Characteristics of the studies included in the review (n=55)

S. No. Author, year Country Numberexposed and type of exposure Wood dust concentration (TD/RD/ID) (in mg/m3) Number unexposed and type of control Questionnaire used Statistically significant allergic and/or respiratory symptoms among wood dust-exposed versus unexposed population 1 Malo et al.,198615 Canada 11 sawmill workers – – – – 2 Holmstrom et al., 199116 Sweden 45 (16 medium density fibre board+29 wood dust group) Mean (range) TD: 1.4 (1–2) 36 controls – – 3 Dahlqvist et al.,199217 Sweden 28 sawmill workers Median TD: 0.26 19 office staff – – 4 Johard et al.,199218 Sweden 19 sawmill workers Median (IQR) TD: 0.25 (0.22–0.34) 25 controls – – 5 Hessel et al.,199519 Canada 94 sawmill workers Mean (range) RD: 1.35 (0.1–2.2) 165 oil field and gas plant workers I U T L D Shortness of breath and wheeze with chest tightness 6 Chea et al.,199620 Australia 45 furniture industry workers – 45 office workers MRC Common cold, running nose, cough and breathlessness 7 Eriksson et al.,199621 Sweden 48 sawmill workers Mean (range) TD: 0.3 (0.1–1.1) – – – 8 Eriksson et al.,199722 Sweden 38 pinewood workers Mean (range) TD: 0.6 (0.1–4.6) – – – 9 Talini et al.,199823 Italy 143 wood workers Mean (SD) TD: 3.6 (2.7) 63 assemblers NHLI Phlegm production 1 0 Mandryk et al.,199924 Australia 168 wood workers Range RD: (0.3–0.7) 30 maintenance workers BMRC Cough, phlegm, chronic bronchitis and nasal block 1 1 Borm et al.,200225 Indonesia 87 wood workers with high exposure Range TD: (5.2–12.3) 572 wood workers in low dust exposure – – 1 2 Milanowski et al., 200226 Poland 48 furniture industry workers – 41 office workers – – 1 3 Schlunssen et al.,200227 Denmark 2033 wood workers Mean (SD) ID: 1.19 (0.86) 474 factory workers with low dust exposure BMRC Night wheeze, daily coughing and common cold symptoms 1 4 Ricciardi et al.,200328 Italy 9 wood workers with occupational asthma – – – – 1 5 Arbak et al.,200429 Turkey 64 furniture industry workers – 62 university students – Cough, shortness of breath, rhinitis and conjunctivitis 1 6 Anyabolu et al.,200530 Nigeria 400 sawmill workers – 200 teaching hospital staff – Cough, sputum and nasal stuffiness/sneezing 1 7 Okwari et al.,200531 Nigeria 221 wood workers Mean (SD) RD: 31.8 (3.4) 200 office staff – Cough, chest pain and nasal irritation 1 8 Ugheoke et al.,200632 Nigeria 150 sawmill workers – 150 automobile technicians MRC Sputum production, cough, breathlessness, wheeze and chest pain 1 9 Baran et al.,S_200933 Poland 70 furniture industry workers Range TD: (0.5–18.2) – – – 2 0 Osman et al.,200934 Turkey 328 furniture industry workers Mean (SD) TD: 2.0 (1.5) 328 store workers – – 2 1 Boskabady et al.,201035 Iran 66 furniture industry workers – 66 community participants Farsi questionnaire Cough, sputum, sneezing and itchy eyes 2 2 Mohan et al.,201336 India 150 carpenters – 150 controls – – 2 3 Sriproed et al.,201337 Thailand 46 sawmill workers Mean (SD) ID: 2.5 (2.3) 22 workers in low dust settings ATS Sneezing and nasal symptoms 2 4 Thetkathuek et al., 201338 Thailand 533 furniture industry workers Mean (range) TD: 4.1 (1.2–11.2) 152 office workers ATS – 25 Usman et al., 201339 India 30 wood workers – 30 shopkeepers Methods respiratory questionnaire – 26 Chatterjee et al., 201440 India 90 wood workers - - - - 27 Danilova et al., 201441 Greece 32 furniture industry - 35 0ffice staff - - 28 Das et al., 201442 Nepal 50 sawmill workers - 100 office staff – – 29 Meo et al., 201443 Pakistan 46 wood workers – 46 salesmen and shopkeepers - - 3 0 Mumuni et al., 201544 Ghana 155 sawmill workers – – Structured respiratory health questionnaire - 3 1 Dunga et al.,201645 Nigeria 200 sawmill workers Range TD: (4.4–22.4) 200 teaching hospital staff – – 3 2 Hussein et al.,201646 Cairo 82 wood workers – 81 controls – – 3 3 Kulkarny et al., 201647 India 50 sawmill workers – 50 college Staff – – 3 4 Tobin et al., 201648 Nigeria 227 sawmill workers Mean (SD) ID: 0.3 (0.1) 277 water bottle company workers BMRC Cough with or without phlegm, breathlessness, wheeze, chest pain and chest tightness 35 Ahmed et al., 201749 Pakistan 120 wood workers – – IRWS – 3 6 Asia et al.,201750 India 60 sawmill workers – 65 controls – Cough and breathlessness 3 7 Badirdast et al.,201751 Iran 100 chipboard workers Geometric mean (SD) ID: 19 (2.0) 50 security guards – – 3 8 Ativie et al., 201852 Nigeria 20 sawmill workers – 20 controls NIOSH – 3 9 Bolund et al.,201853 Denmark 1112 wood workers – 235 workers in low–dust settings – – 40 Omole et al., 201854 Nigeria 102 sawmill workers – 102 college staff – – 41 Ahmed et al., 201955 Pakistan 100 carpenter workers – – IRWS – 4 2 Asgedom et al., 201956 Ethiopia 74 particle board workers Mean (range) ID: 9.2 (0.5–184) 73 water bottle company workers ATS Cough with sputum, phlegm, wheezing, and shortness of breath 43 Das et al.,201957 India 80 carpenters – 70 security guards MRC Chest tightness, phlegm, chronic dyspnoea, cough and nasal irritation 4 4 Fentie et al., 201958 Ethiopia 70 wood workers – 70 shopkeepers – – 45 Jagtap et al., 201959 India 180 sawmill workers – 180 office staff – – 4 6 Paraskevaidou et al., 201960 Greece 23 wood workers – 25 office staff Respiratory Health Survey Questionnaire for Asthma Asthma or rhinitis 4 7 Patil et al., 201961 India 60 sawmill workers – 60 controls – – 48 Asri et al.,202062 Malaysia 41 (17 sawmill and 24 furniture industry workers) Sawmill worker Median (IQR) ID: 2.4x103 (1.1x103–5.8x103); Furniture worker Median (IQR) ID:1.0 (0.3–6.7) 39 administrative staff – – 4 9 Hossein et al.,202063 Iran 276 wood workers – 276 office workers Respiratory symptoms questionnaire Cough, phlegm and chest tightness 5 0 Rasouel et al.,202064 Egypt 100 furniture industry workers Mean (SD) RD: 3.61±2 100 relatives of exposed participants – Rhinitis, cough, expectoration, wheeze, dyspnoea and asthma 5 1 Saldanha et al.,202165 India 35 sawmill workers – 35 community participants – – 5 2 Salman et al.,202166 Bangladesh 50 sawmill workers – 50 community participants – – 53 Taha et al.,202167 Egypt 70 furniture industry workers 70 controls – – 5 4 Ennin et al.,202268 Ghana 86 wood workers Range ID: 0.003–1.02 89 university staff MRC Wheeze, sneeze, cough with or without phlegm, fever, shortness of breath, loss of voice and catarrh 5 5 Mogal et al.,202269 Bangladesh 100 sawmill workers – 100 college Staff ATS and NHLI Breathlessness, cough, sneeze, chest tightness and itching

Several studies used tools including the International Union against Tuberculosis and Lung Disease, Medical Research Council, and British Medical Research Council questionnaire to understand allergic and respiratory symptoms among workers. Qualitative synthesis of 20 studies in the review reported a significant increase in the frequency of allergic or respiratory symptoms among wood workers compared with the unexposed population. Respiratory symptoms reported included cough, breathlessness, phlegm production, and chest tightness, while allergic symptoms included wheeze, asthma, rhinitis, conjunctivitis, nasal stuffiness, and itching of skin.

Lung function assessment

FEV1. We considered 16 studies that reported FEV1 in percentage predicted for meta-analysis using a random effects model and estimated a pooled mean difference of –8.57 (95% CI: –12.07 to –5.08) percentage. Wood dust-exposed workers had a significant decline of FEV1 of about 9% compared to controls. A high heterogeneity was observed between the studies (I2=95.26%). Subgroup analysis showed that including studies of varying risk of bias could partly explain the heterogeneity (Tables 2, S1; Figs. 2, S1).18,23–25,29,34,35,38,41,46,47,60,64,67,68

TABLE 2. Meta-analysis of lung function measures

Lung function No. of studies included Pooled mean difference (95% CI) FEV1 L/minute 22 –0.37 (–0.49 to –0.26) FEV1 percentage predicted 16 –8.57 (–12.07 to –5.08) FVC L/minute 2 3 –0.42 (–0.58 to –0.25) FVC percentage predicted 14 –7.43 (–11.35 to –3.51) FEV1/FVC 2 9 –2.13 (–4.76 to 0.50) PEFR L/minute 1 6 –109.4 (–135.34 to –83.47) PEFR percentage predicted 9 –13.43 (–25.14 to –1.72)

FIG 2. Forest plot of forced expiratory volume at first second (FEV1) in percentage across studies among wood workers

Export to PPT

Similarly, 22 studies had complete information on FEV1 in L/minute among wood workers and the comparator group (Table 2). Meta-analysis of 22 studies using a random-effects model showed a pooled mean difference of FEV1 of –0.37 (95% CI: –0.49 to –0.26) L/minute. This indicates that workers exposed to wood dust had a significant reduction in FEV1 by 0.37 L/minute compared to controls. A high heterogeneity was observed between the studies (I2=96.87%). Subgroup analysis showed that varying risk of bias between included studies could partly explain the heterogeneity (Tables 2, S1; Figs. S2a, S2b).

FVC. A random-effects model was used to pool the 14 studies that reported FVC in percentage predicted, showing a significant mean difference of FVC percentage of –7.43 (95% CI –11.35 to –3.51). The wood dust-exposed group showed a statistically significant reduction in FVC percentage compared to unexposed workers. High heterogeneity was observed between the studies (I2=94.73%), and subgroup analysis showed the risk of bias of the studies considered as a probable cause of heterogeneity. Twenty-three studies included in a random-effects meta-analysis showed a pooled mean difference in FVC of –0.42 (95% CI –0.58 to –0.25) L/ minute. The wood dust-exposed group showed a statistically significant reduction in FVC by 0.42 L/minute compared to the unexposed population. A high heterogeneity was noted between the studies, with I2 of 98.76%. Subgroup analysis showed that including studies of varying risk of bias could partly explain the heterogeneity. (Tables 2, S1; Figs. 3, S3a, S3b, S4).23–25,29,34,35,38,41,46,60,63,64,67,68

FIG 3. Forest plot of forced vital capacity (FVC) in percentage across studies among wood workers

Export to PPT

FEV1/FVC. Twenty-nine studies reported the FEV1/FVC across wood-exposed and unexposed populations (Table 2). However, meta-analysis using random effects model showed that there was no statistically significant reduction of FEV1/ FVC (mean difference –2.13; 95% CI: –4.76 to 0.50) among wood workers compared to the unexposed population (Tables 2 and S1; Figs. S5a and S5b). Subgroup analysis by wood dust type showed that sawmill workers had a significant decline in FEV1/FVC ratio by 5.4 units. Given that sawmill workers also had a decline in FEV1, it is plausible that they were more likely to have an obstructive lung function pattern (Figs. S1b, S5b). In contrast, furniture industry workers did not present with reduced FEV1/FVC in the subgroup analysis. By factoring in this observation along with the presence of significant FVC reduction among furniture workers, it is inferred that furniture workers tend to be more likely to present a restrictive pattern (Figs. S5b, S3a, S3b).

Peak expiratory flow rate (PEFR)

Meta-analysis of 9 studies that reported predicted PEFR percentage showed that wood dust-exposed workers had a reduction in predicted PEFR percentage by 13.43 compared with the unexposed population (pooled mean difference: –13.43; 95% CI –25.14 to –1.72). Subgroup analysis revealed that the risk of bias of the included studies could partially explain the heterogeneity in the meta-analysis estimate. Sixteen studies reported PEFR values in L/minute among wood-exposed and comparator groups. Meta-analysis of 16 studies showed a pooled mean difference of –109.4 (95% CI –135.34 to –83.47) L/minute, indicating a reduction in PEFR by 109.4 L/minute among the wood dust-exposed group compared to the control. High heterogeneity (I2=98.63%) was noted between the studies, and subgroup analysis did not identify any potential variable that resulted in heterogeneity. (Tables 2, S2; Figs. 4, S6a, S6b, S7).

FIG 4. Forest plot of PEFR in percentage across studies among wood workers.

Export to PPT

Risk of bias

Risk of bias assessment for the 55 studies included in this review showed that most (47, 85.46%) were at high risk of bias. Fewer studies were rated as medium and low risk of bias, including 5 (9.09%) and 3 (5.46%), respectively (Tables S1 and S2).

Publication bias

Funnel plots of various outcomes related to lung function tests and peak expiratory flow rate showed asymmetry, suggesting publication bias among the studies included in this meta-analysis (Figs. S8a–S8f).

DISCUSSION

This systematic review, conducted across studies involving wood workers from diverse settings worldwide, demonstrated a significant reduction in lung function parameters compared with the unexposed population. We found a statistically significant reduction in FEV1, FVC, and PEFR, indicating that lung function impairment among wood workers could be of both obstructive and restrictive patterns. Our meta-analysis showed high heterogeneity among the studies, likely reflecting variability in population, dust exposure, and methods of outcome assessment. Most of the studies in the systematic review were at high risk of bias. Subgroup analysis, including risk of bias and the type of wood handled by the workers, could partly explain the heterogeneity in the meta-analysis estimates.

A recent systematic review has pooled evidence from longitudinal studies related to lung function decline following exposure to mineral or biological dust in occupational settings.70 This analysis has shown that workers classified as ‘ever’ being exposed, as well as those with ‘cumulative’ exposure to dust, present with a significant decline in FEV1 estimates.70 This is similar to our review that reported FEV1 reduction in the wood dust-exposed group compared to the unexposed population, while measured in both L/minute and percentages. Considering these findings, another systematic review highlighted a greater risk of about 1.5 times to develop asthma following wood dust exposure compared to the unexposed population and provided supportive evidence for a dose-response relationship for the association between dust exposure and risk of asthma.3

Similar to wood dust, occupational exposure to quartz (silica) dust has been implicated in developing obstructive pattern in lung function with a reduction in pooled FEV1 percentage predicted by 4.62 points in a systematic review of cross-sectional studies.71 While our review observed a similar reduction in FEV1 to that of the review on quartz exposure, it is important to note that this review did not find a significant reduction in FEV1/FVC ratio.71 Despite both the reviews synthesizing cross-sectional evidence, the lack of reduction in FEV1/FVC ratio in our review could be influenced by differing exposure characteristics across studies, because of including participants from the sawmill and furniture industries. Subgroup analyses in this review showed that sawmill workers were more likely to have an obstructive pattern (reduced FEV1 and FEV1/FVC) compared to furniture workers, who were more likely to have a restrictive pattern (reduced FVC alone). The lack of significance in the overall FEV1/FVC meta-analysis could be due to the pooling of studies that included sawmill as well as furniture industry workers together in a single meta-analysis, where the individual effect of each of these groups tends towards the opposite direction, leaving the summary estimate around the line of no effect.

Our review identified a significant reduction in FVC in wood-exposed workers compared to controls, indicating a restrictive lung function impairment. A similar restrictive pathology due to occupational exposure to wood dust has been reported in a systematic review, which found that wood dust exposure was associated with a 62% higher risk of being diagnosed with idiopathic pulmonary fibrosis compared to controls.72

We also identified a significant reduction in PEFR measurements among wood workers compared to controls. This could probably reflect airflow obstruction due to the irritant effect of inhaled wood dust, resulting in an asthma presentation. This finding is supported by a systematic review by Zhang et al, where exposure to organic dust, such as wood, resulted in 1.62 times higher chances of presenting with asthma compared to controls.73

This systematic review has several limitations as it synthesized evidence from observational study designs that are prone to bias, although these are ideal for measuring the association between risk factors and disease. Of the potential confounders that could affect the association between wood dust exposure and lung function in our meta-analysis, the type of wood dust and the risk of bias of included studies were considered in the subgroup meta-analysis. Insufficient information on the duration of exposure to wood dust and smoking habits precluded inclusion of these variables in the subgroup meta-analysis. This review identified most studies as cross-sectional, except for a single cohort study from Denmark. This review identifies a potential research gap in understanding the natural progression of lung disease following wood dust exposure through cohort studies.

Given substantial evidence linking wood dust exposure to lung function impairment, adopting workplace dust-limiting measures, such as substitutions and improved ventilation to replace air contaminated with dust, is important. In India, early identification of workers with lung impairment through periodic medical examinations in organized industrial settings, as well as bringing workers under the umbrella of medical insurance coverage under the Employee Insurance Scheme, must be prioritized.

Conclusion

We found a deterioration in lung function measurements such as FEV1, FVC and PEFR among wood dust-exposed workers, pointing towards obstructive as well as restrictive lung function patterns. Sawmill workers are likely to have an obstructive lung function pattern, whereas furniture industry workers present with a restrictive pattern. This evidence emphasises the need to adopt preventive measures, such as workplace engineering modifications and promoting the use of personal protective equipment among wood workers, along with periodic screening of workers to reduce health impacts from wood exposure.