Exposure to inhalable dust, endotoxin and formaldehyde in factories processing particleboards from eucalyptus trees in Ethiopia

Abstract

Objectives: Eucalyptus trees are used in the particleboard wood industry in Ethiopia. Dust and chemicals from this production may cause respiratory health problems, but the exposure levels have not been studied previously. The aim of this study was to assess workers’ exposure to inhalable dust, endotoxin, and formaldehyde in the two largest particleboard factories in Ethiopia, and compare the results with occupational exposure limits. Methods: A total of 152 inhalable dust and endotoxin samples were collected using a conductive plastic inhalable conical sampler (CIS), in addition to 45 formaldehyde samples using Dräger tubes for collection in two particleboard factories in Ethiopia. Linear mixed models were used to identify exposure determinants. Results: The geometric mean (GM) of inhalable dust exposure was 4.66 mg/m³ and 93% of the samples were above the threshold limit value of 1 mg/m³. For endotoxin exposure GM was 62.2 EU/m³, and 41% of samples were above the recommended occupational limit value of 90 EU/m³. Formaldehyde was added in a blending section of the production line, and the formaldehyde level was highest here (3.5 ppm). The level decreased at the workstations following blending, and 13% of the formaldehyde samples were above the peak exposure limit value of 1.0 ppm. Conclusions: The findings revealed exposure levels higher than recommended for inhalable dust, endotoxin, and formaldehyde in the particleboard factories. A reduction in dust, endotoxin, and formaldehyde exposure levels in these workplaces is recommended. We also recommend provision of proper personal protective equipment.

Introduction

The manufacturing sector in Ethiopia comprises a range of industries, such as wood, metal, food, textile, leather, and construction, and it accounts for 6.9% of the national workforce¹⁾. Particleboards are manufactured primarily from discrete particles, glued with urea formaldehyde resin, and bonded under heat and pressure. Particleboard manufacturing includes raw material procurement or generation, size classification, drying, blending with resin/wax, forming resinated material into a mat, which is hot pressed and cooled before finishing and packing²⁾. Formaldehyde is added to the urea resin and is thus present in the processes of glue mixing, mat forming, pressing, trimming, sanding, and sizing. In Ethiopia, the eucalyptus tree is the main raw material of particleboards. This type of tree is cheap and adaptable to a range of climates and soil types with rapid growth. These factors make eucalyptus trees ideal for plantation in Ethiopia. Eucalyptus is a promising renewable input at a time when other native wood species are diminishing due to deforestation.

Exposure to wood dust, endotoxin, and formaldehyde may cause respiratory health problems^3,4,5,6,7). In a Danish study where 13% of the included workers were employed in particleboard and fiberboard industries, there was an accelerated decline in lung function among the female workers⁸⁾. A study on particleboard manufacturing facilities in Canada indicated that rash, nasal and eye irritation, cough, and annoying odors were the main complaints among workers⁹⁾. A study of Ethiopian particleboard workers show a high prevalence of respiratory symptoms in this group¹⁰⁾. To our knowledge, this is the only study of workers in particleboard production using the eucalyptus tree as a raw material.

In the wood processing industries, like sawmills, furniture production, and carpentry, using both hard woods and soft woods, there is relatively high exposure to organic dust and endotoxin. Endotoxin is a component in the cell walls of gram-negative bacteria, which has been reported to be associated with respiratory health problems in studies from several countries, including Canada, Italy, Tanzania, Sweden, and Thailand^{4,11,12,13,14,15,16)}. A study in Norwegian sawmill reported workers exposure to wood dust, microbial components, and terpenes¹⁷⁾.

High exposure to wood dust has been revealed in medium and small-scale wood industries in Ethiopia producing various furniture products¹⁸⁾. These industries use Austrian Pine, Plywood, Cordia Aficana, Olea Europea-cuspidata, and Pouteria adolfi-federicii as raw material. However, they do not use eucalyptus tree and do not produce particleboard. On top of this, the working situation and the production process is not the same as in Ethiopia in the mentioned industries.

As the particleboard industry is increasing and more and more workers are exposed to dust and chemicals in this production work, there is a need for studies of their work environment. This type of knowledge is important for the development of national work and health policies for this industry.

The aims of this study were to assess exposure to inhalable dust, endotoxin, and formaldehyde in particleboard factories using eucalyptus trees, and compare these with occupational exposure limits with a view to providing needed research and contributing towards preventive measures.

Materials and Methods

Study setting

The study was conducted in the two largest particleboard factories in Ethiopia using eucalyptus, an evergreen hardwood, as a raw material. Factory A, situated in Northern Ethiopia, was established in 2005, has 663 workers, and produces 8–40 mm thick triple layered particleboard with a maximum daily/annual output of 170 m³ and 40,000 m³, respectively. Factory B, located in southern Ethiopia, was established in 2002 and has 249 workers. The main input was 10–40 cm diameter eucalyptus logs with a current consumption rate per annum of 18,167 m³. The factory produces 43,000 m³ of 6–40 mm thick boards annually for local and export markets. The general flow diagram of the particleboard production process is shown in Figure 1.

Fig. 1.

Process-flow diagram of the particleboard production process. MCC1: Machine control center 1; MCC2: Machine control center 2.

Description of the work process

The production lines are similar in the two factories, having two separate sections. The first is open-air with shades, and houses chipping, flaking, silo, drier/boiler, Machine Control Center 1, and silo. The solid eucalyptus log is chipped with a maximum chip length of 30 mm and downsized further by the flaker machine. Wood chips are temporary stored in a silo. The drier removes moisture from the wood chips using generated hot gas. The chipping and flaking process is controlled through Machine Control Center 1.

The second section is a closed, large hall located inside the factory. Manufacturing activities here include blending/chemical, forming, pressing, trimming, sanding, and sizing sections arranged consecutively. In the glue kitchen of the chemical section, urea-formaldehyde powder is mixed manually with water to prepare formaldehyde solution and is further mixed/blended with wood chips. The mat is formed from the wood mix and is pressed under high hydraulic pressure (250 bar) and temperature (175–190°C) to convert the material into solid board. The press was partly enclosed with iron sheets in Factory A, but not in Factory B. The solid board is cooled for some time before it is trimmed, sized and sent for sanding to attain uniform board thickness and a smooth surface. It can be further resized according to the customer’s specification. Workers are present in the entire section, and no facemasks are used.

In addition, there are other miscellaneous factory workers: cleaners who sweep the floor on the production line, quality control workers who monitor the moisture content of chips and the bond strength of the produced particleboard, and maintenance workers who repair machines and monitor the overall mechanical process. The production process is controlled through Machine Control Center 2, found inside the big hall. Neither of the factories has mechanical ventilation systems or dust collecting hoods.

Personal sampling of dust and endotoxin

The study population were the production workers in the 10 working sections: chipping (n=83), flaking (n=29), chemical/blending (n=21), forming (n=28), pressing and trimming (n=21), sanding (n=27), sizing (n=12), cleaners (n=27), quality control (n=18), and maintenance (n=34). We assigned the workers by section into assumed similar exposure groups. The sample size for measurement of inhalable dust, which is appropriate for aerosols like wood dust that usually deposit in the extra thoracic airways¹⁹⁾, was 10 to 20 measurements per observed group (i.e., two repeated measurements from 5 to 10 randomly selected persons) in each of the work sections²⁰⁾. Totally, 76 workers were selected for repeated sampling of inhalable dust (n=152) with 10 days sampling for factory A and 8 days for factory B. In addition, one field blank sample was taken per day (n=18). The persons were selected by consensus between the researcher and supervisors. Selection was based on willingness to participate, work convenience, and representativeness for the working section.

Inhalable dust samples in the workers’ breathing zone were taken using a conductive plastic inhalable conical sampler (CIS; JS Holdings, Stevenage, UK)²¹⁾ mounted with a 37-mm glass-fiber (GFA) filter (Whatman International Ltd, Maidstone, UK) and connected to a Side Kick Casella (SKC) pump using the CIS samplers recommended air flowrate of 3.5 l/min²²⁾. On each sampling day, we registered the production downtime and the volume of board produced. During the field visit in the first phase of sampling, we observed high dust exposure due to the nature of the tasks. Thus, we decided not to take full shift measurements (8 hours), to avoid problems with the pumps during the sampling. We judge this to be acceptable, as we judge the 2–4 hours to be representative for the whole working day. The mean sampling time was 196 minutes (range 120–240). The air flow rate was measured before, during, and after each measurement. Collected samples were securely packed and transported to the laboratory. All samples were analyzed for both dust and endotoxin at Aarhus University, Denmark. Filters were weighted in a room with controlled climatic conditions (22ºC, 45% relative humidity; desiccation ≥24 h) using an analytical balance with 0.1 µg readability (Mettler-Toledo Ltd, Greifensee, Switzerland) and dust concentration was estimated in mg/m³. The gravimetric analysis followed an approved and standardized protocol as stated elsewhere²³⁾.

For analysis of endotoxin, the glass-fiber filters were extracted in 5 mL pyrogen free water with 0.05% (v/v) Tween-20 by orbital shaking (300 rpm) at room temperature for 60 minutes and centrifuging (1000 g) for 15 minutes. The supernatant was stored at −80°C until used for the endotoxin assay. The supernatant was analyzed for endotoxin using the Kinetic Amoebocyte Lysate test (Kinetic-QCL endotoxin kit, BioWhittaker, Walkersville, MA, USA) at Aarhus University, Denmark^23,24) following a standardized protocol and procedure. The endotoxin results were expressed in EU/m³. Of the 18 blank filters, three of the blanks were deviating strongly from the other blank filter value, possibly because of contamination during handling, and were therefore discarded.

Formaldehyde measurements

For formaldehyde exposure, we applied a “worst-case” strategy using Dräger tubes²⁵⁾. During the morning shift, measurements were taken in seven workstations (chipping, glue kitchen, blending, forming, pressing, trimming, sanding) in Factory A and from eight workstations (chipping, glue kitchen, blending, forming, pressing, trimming, sanding, sizing) in factory B. One measurement was taken in each workstation at a height of 1.5 m on an operator work site on 3 different days after urea-formaldehyde had been mixed with water in the glue kitchen. Thus, the total number of measurements of formaldehyde was 45 (i.e., 21 from Factory A and 24 from Factory B). Selection of measuring ranges of the Dräger tubes, either 0.2 to 5 ppm or 2 to 40 ppm, was based on the assumed concentration of formaldehyde. The glue kitchen and the blending section were assumed to have relatively high concentrations; consequently, for some of the measurements, we used the 2 to 40 ppm measuring range of the tubes initially (i.e., five strokes with the hand pump). The measuring range of 0.2 to 5 ppm (i.e., 10 strokes for 90 seconds) was used for the samples in the other workstations²⁵⁾. The non-detectable values for the Dräger tubes with a measuring range of 0.2 to 5 ppm was set to be less than the minimum reading of the Dräger tube (i.e., <0.2 ppm). The formaldehyde concentrations were presented using median and range due to the few samples taken.

Occupational exposure limits

For comparison purposes, we refer to the American Conference of Governmental Industrial Hygienists (ACGIH) Threshold Limit Value (TLV) for inhalable wood dust (8 h TLV=1 mg/m³)²⁶⁾. For endotoxin, we have used the Dutch health-based recommended occupational exposure limit of 90 EU/m³ as a reference value²⁷⁾. For formaldehyde, we refer to the Norwegian peak exposure limit value of 1 ppm²⁸⁾.

Data analysis

Statistical analysis was performed using IBM SPSS version 25 for Windows (IBM Corp, Armonk, NY, USA). The exposure data were not normally distributed, so they were natural log (ln)-transformed before analysis. Data were presented using measure of central tendency (arithmetic and geometric mean) and variation (range and geometric standard deviation) across the workstations. Correlation between the volume of particleboard produced, downtime, inhalable dust, and endotoxin level was calculated using the Pearson correlation test. Initially, linear mixed models were used to explore differences in mean inhalable dust and endotoxin exposure between the factories (A, B) and among the workstations. Two separate mixed effects model were subsequently used to identify significant determinants and variance components for dust and endotoxin levels, respectively. Worker was entered as random effect and significant variables/determinants in preliminary analysis (p<0.05); time with no production (downtime), volume of particleboard produced, study site or factory (A, B) and workstations were entered as fixed effects. Press trimming was used as reference workstation in the models. Determinants were retained in the models when significant (p<0.05).

Ethical approval

The study protocol was approved by the Regional Committee for Medical and Health Research Ethics, West Norway on June 2, 2016 with IRB ref: IRB00006245 and by the Ethiopian Ministry of Science and Technology on October 7, 2016 with Ref. No. 3.10/148/2016. Study participants were informed about the purpose of the study and consent from study participants and factory management was assured before data collection began.

Results

In total, 152 filter samples from 76 workers were collected for assessing inhalable dust and endotoxin levels. Ten samples were excluded before the analysis by the laboratory due to visible loose dust on the filter. There was an interruption of production in the factories (hereafter called downtime): Factory A had an average of 4.3 hours (range, 0.7–5.9) and Factory B had 3.2 (range, 0–4.4) hours of downtime during the sampling days. Factory A had sporadic downtime during all 10 days of sampling and Factory B had 7 days of downtime in 8 days of sampling. The total volume of particleboard produced during the sampling days was 258.6 m³ (range, 8.78–48.82) and 90.55 m³ (range, 0–18.31) for Factory A and Factory B, respectively.

The overall geometric mean (GM) of inhalable dust exposure (n=142) was 4.66 mg/m³. Personal exposure to inhalable dust was statistically different (p<0.001) between Factory B (GM=8.67 mg/m³) and Factory A (2.83 mg/m³) (Table 1). In the initial analysis, inhalable dust exposure levels in chipping, flaking, forming, and sizing were significantly different from the reference workstation (press trimming).

Table 1. Personal inhalable dust and endotoxin levels in Ethiopian particleboard factories by factory and workstation

Variable	Inhalable dust (mg/m3)					%† >TLV	Endotoxin (EU/m3)			%† >OEL	ME/MD
Workstation	Nw	Ns	AM	Range	GM(GSD)		AM	Range	GM (GSD)
Factory A
Chipping	5	10	5.73	1.39 – 9.94	4.76(1.97)	100	692.6	82.7 – 1748.1	492.7(2.5)	90	98.6
Flaking	4	8	10.21	2.09 – 36.32	6.82(2.51)	100	386.9	64.9 – 1442.1	357.8(3.2)	75	68.2
Chemical/blending	5	10	3.10	0.68 – 7.16	2.41(2.12)	90	26.8	1.5 – 74.5	16.1(3.4)	0	9.7
Forming	3	6	3.09	2.01 – 4.47	2.97(1.35)	100	43.1	29.7 – 51.2	41.7(1.3)	0	16.7
Press-Trimming	6	12	1.67	0.67 – 4.89	1.43(1.72)	83	9.5	0.9 – 29.7	6.1(2.9)	0	4.3
Sanding	2	4	1.31	0.47 – 3.32	0.93(2.51)	25	45.44	1.4 – 129.3	15.3(7.3)	25	34.7
Quality control	3	6	1.16	0.80 – 2.01	1.09(1.40)	50	10.6	2.9 – 35.9	7.2(2.4)	0	5.7
Maintenance	5	10	4.06	0.69 – 7.92	3.19(2.23)	90	129.3	12.1 – 343.2	75.2(3.3)	40	19.7
Cleaner	6	12	7.20	1.32 – 29.38	4.62(2.61)	100	233.9	3.7 – 1077.8	78.3(4.9)	41	14.5
Total	39	78	4.46	0.47 – 36.32	2.83(2.53)	87	212.9	0.9 – 1748.1	46.5(6.8)	32	17.6
Factory B
Chipping	3	6	17.95	6.41 – 29.59	15.49(1.80)	100	233.9	3.7 – 1077.8	219.2(3. 1)	83	60.9
Flaking	6	12	9.57	4.9 – 26.5	8.25(1.70)	100	200.6	47.5 – 643.3	135.6(2.3)	58	14.4
Chemical/blending	2	4	20.67	13.01 – 28.73	19.11(1.55)	100	202.4	87.6 – 308.2	170.7(1.9)	75	10.1
Forming	2	6	70.93	6.81 – 183.97	36.6(4.22)	100	283.1	190.5 – 503.0	259.8(1.5)	100	4.7
Press-Trimming	4	8	5.75	2.48 – 12.56	4.76(1.88)	100	59.4	3.2 – 107.6	40.9(3.2)	37	14.4
Sanding	3	6	6.01	1.85 – 10.21	4.85(2.10)	100	83.2	3.5 – 160.1	47.5(4.3)	50	12.1
Sizing	6	12	10.70	2.19 – 23.33	8.67(2.03)	100	64. 7	3.4 – 147.2	45.1(2.8)	33	5.2
Maintenance	5	10	14.53	1.12 – 68.82	6.62(3.49)	100	996.7	21.1 – 9202.2	108.8(5.5)	50	16.5
Cleaner	1	2	4.89	3.35 – 6.44	4.62(1.58)	100	43.3	32.2 – 54.5	41.7(1.4)	0	8.8
Total	32	64	14.91	1.12 – 183.97	8.67(2.59)	100	285.4	3.2 – 9202.2	88.2(3.6)	53	11.0
Factory A and B	71	142	9.17	0.47 – 183.97	4.66(2.94)	93	245.6	0.9 – 9202.2	62.2(5.4)	41	13.3

ME/MD, The median of the ratio of endotoxin level to inhalable dust level based on individual measurements; Nw, Number of workers; Ns, Number of samples; OEL, Dutch health-based recommended occupational exposure limit (OEL) for endotoxin (90 EU/m³); TLV, American Conference of Governmental Industrial Hygienists (ACGIH) Threshold Limit Value (TLV) for inhalable wood dust (1 mg/m³); %^†: percent of samples above TLV or OEL and the denominator is number of samples in each factory or workstation.

Endotoxin exposure (n=142) had a GM of 62.2 EU/m³. There was no statistically significant difference (p=0.069) in endotoxin exposure between Factory A (GM=46.53 EU/m³) and Factory B (GM=88.23 EU/m³) (Table 1). In the initial analysis, endotoxin exposure levels in chipping, flaking, forming, sizing, maintenance, and cleaning were significantly different from the reference workstation (press trimming).

Of the 142 samples 93% exceeded 1 mg/m³, the TLV set by ACGIH and the percent varies between factories (range 87–100%) (Table 1). Additionally, 41% of samples also exceeded 90 EU/m³, the Dutch health-based recommended occupational exposure limit and the percent varies between factories (range, 32–53%) (Table 1).

Endotoxin level correlated significantly with inhalable dust level (r=0.66, p<0.001) (Figure 2). Downtime during sampling correlated with inhalable dust level (r=−0.34, p<0.001). The volume of particleboard produced correlated with inhalable dust (r=−0.18, p=0.029). We a priori expected a higher particleboard production (volume of particleboard produced) to result in an increased concentration of the inhalable dust, but in fact our results show the opposite. The reason could be the production rate (volume of particleboard produced) was consistently higher in Factory A than in Factory B, since Factory A had the lowest dust exposure irrespective of the higher volume of particleboard recorded. Downtime and volume of particleboard produced were not significantly correlated (0.024, p=0.78).

Fig. 2.

Correlation between samples of inhalable dust (mg/m³) and endotoxin levels (EU/m³) in two particleboard factories plotted on logarithmic axes (r=0.66, p<0.001, n=142).

The between worker variance was higher than within worker variance both for inhalable dust (within- to between-worker variation ratio, 0.57) and endotoxin (ratio, 0.73) (Table 2). Factory (A, B) and downtime explained 27.0% of the total variability in inhalable dust level (11.6% of the within worker and 36.0% of the between worker variability). The exposure model predicts that a 1-hour increase in downtime will decrease exposure to inhalable dust by 15%. The workstations were not statistically significant determinants in the final model when other determinants were taken into account.

Table 2. Linear mixed effect model of ln-transformed inhalable dust (mg/m³) and endotoxin (EU/m³) using all exposure measurements (n=142) from two particleboard factories in Ethiopia

Determinants	Description	Inhalable dust exposure (mg/m3)			Endotoxin exposure (EU/m3)
		Random effect model β (SE)	Mixed-effect model β (SE)	p	Random effect model β (SE)	Mixed-effect model β (SE)	p
	Intercept	1.54 (0.11)	1.70 (0.26)	<0.001	4.13 (0.17)	3.08 (0.20)	<0.001
Workstation related determinants
Chipping	0=No, 1=Yes					2.81 (0.44)	<0.001
Flaking	0=No, 1=Yes					2.21 (0.41)	<0.001
Forming	0=No, 1=Yes					1.37 (0.54)	0.014
Maintenance	0=No, 1=Yes					1.41 (0.41)	0.001
Cleaner	0=No, 1=Yes					1.18 (0.47)	0.015
Factory	0=Factory A 1=Factory B		0.94 (0.20)	<0.001
Downtime			−0.15 (0.05)	0.005
wwσ2		0.43 (0.07)	0.38 (0.06)		1.20 (0.20)	1.20 (0.20)
bwσ2		0.75 (0.16)	0.48 (0.11)		1.65 (0.39)	0.68 (0.24)
Variance explained by the fixed effects
wwσ2			11.6%			0%
bwσ2			36.0%			58.8%
Total			27.0%			34.0%

β, regression coefficients; bwσ², between-worker variance; SE, standard error of the regression coefficient; p, p-value; wwσ², within-worker variance.

In the mixed-effects model for endotoxin exposure, five workstations (chipping, flaking, forming, maintenance, and cleaning) explained 34.0% of the total variability (58.8% of the between-worker variability).

Formaldehyde was detected in all workstations along the production line except in the first and last station (chipping and sizing) (Table 3). The highest median concentration was found in blending (3.5 ppm) followed by glue kitchen (1 ppm) with decreasing concentration further down the production line. Of the 45 formaldehyde samples, 13% exceeded the Norwegian peak exposure limit of 1 ppm.

Table 3. Stationary formaldehyde level (ppm) in particleboard factories by factory and workstation

Variable	Ns	N<0.2 ppm	Median (Range) (ppm)
Factory A	21	7	0.5 (<0.2 – 4)
Factory B	24	13	<0.2 (<0.2 – 5)
Total	45	20	0.5 (<0.2 – 5)
Workstation based levels
Chipping	6	6	<0.2
Glue kitchen	6	0	1.0 (0.5 – 1)
Blending	6	0	3.5 (0.5 – 5)
Forming	6	1	0.5 (<0.2 – 3)
Pressing	6	2	0.5 (<0.2 – 1)
Trimming	6	3	<0.5 (<0.2 – 1)
Sanding	6	5	<0.2 (<0.2 – 0.5)
Sizing	3	3	<0.2

N, number of non-detectable samples (i.e., <0.2 ppm: 0.2 ppm is the lowest reading value of color tubes; Ns, number of samples.

Discussion

The workers in the two biggest particleboard factories in Ethiopia were exposed to a geometric mean of 4.66 mg/m³ of inhalable dust and 62.2 EU/m³ of endotoxins. The highest median air concentration of formaldehyde was found in blending (3.5 ppm), with decreasing concentration further down the production line. From the collected samples 93%, 41%, and 13% of samples exceeded the occupational exposure limits for inhalable dust (1 mg/m³), endotoxin (90 EU/m³), and formaldehyde (1 ppm), respectively.

Exposure to inhalable dust in our study was higher than reported in small scale wood industries in Tanzania (3.3 mg/m³)¹⁴⁾. This difference could partly be due to the outdoor location of the Tanzanian wood industries that provides better general ventilation compared to our study that was done mostly in a closed area. In addition, differences in size of the factories and in the manufacturing, process might also have contributed to variation in dust exposures between the studies.

A study in joinery and furniture factories in the Netherlands reported lower inhalable dust exposure (2.1 mg/m³)²⁹⁾ compared to our study. Also, studies in large- and medium-sized industrial sawmills in Norway¹⁷⁾ and in furniture industries in Denmark³⁰⁾ recorded lower dust levels (0.72 mg/m³ and 0.6 mg/m³, respectively). These differences could be due to the technological advancement found in Norway, the Netherlands, and Denmark to control dust emission using mechanical ventilation as compared to our study, where such control measures were not present.

A study done in the member states of the European Union estimated the highest dust exposure levels to occur in construction and furniture industries³¹⁾. In Ethiopian small- and medium-scale wood industries¹⁸⁾, the geometric mean dust exposure was 6.82 mg/m³, which is comparable with our study from large-scale factories. However, to our knowledge there are no previous studies on inhalable dust exposure in particleboard factories in Ethiopia.

Inhalable dust exposure varied by workstation. The highest inhalable dust exposure was recorded in the workstations of sizing, forming, flaking, and chipping and is probably linked to the dust-emitting, woodcutting processes. Dust exposure was lowest in quality control. This might be because these workers mainly stayed in the laboratory. However, all workstations had dust levels above the TLV of 1 mg/m³ set by ACGIH for wood dust²⁶⁾. A study done in Danish furniture factories supports our finding that work tasks are significant determinants of wood dust exposure³⁰⁾.

The geometric mean endotoxin exposure in our study was 62.2 EU/m³, with a wide range (0.9 to 9,202 EU/m³). The endotoxin exposure in our study is lower than the findings from small-scale wood industries in Tanzania (geometric mean, 91 EU/m³)¹⁴⁾. A study done in sawmills in Norway also showed a lower geometric mean of endotoxin exposure (17 EU/m³)¹⁷⁾. The arithmetic mean for endotoxin exposure of particleboard workers in different workstations in our study varied from 10.6 to 564.8 EU/m³, which is different to the results of a study with a mean range of 16.15 to 1,974.0 EU/m³ in different working sections in the fiberboard industry³²⁾, and from <0.125 to 217.4 EU/m³ in chipboard factories in Poland³²⁾.

The mean endotoxin level was lower than the Dutch committee’s recommended occupational exposure limit of 90 EU/m³. However, in the first two workstations of the production process, chipping and flaking, the mean exceeded this level. Our finding is in line with a study done in Poland, where a higher endotoxin exposure level was found during the initial stage of the production and then sharply decreased during the subsequent production stages³²⁾. High moisture content in the protective bark of the eucalyptus tree in chipping and flaking at the beginning of the production process might constitute an optimal environment for endotoxin producing gram-negative bacteria. After removal of the bark by flaking, the endotoxin per mg of inhalable dust in further processes was reduced. Chipping and flaking were performed outdoors and did not take place in the indoors as the rest of the working section, but they still have shades that protects the process from rain and sun. The impact of endotoxin from the outdoor environment might thus not be significant.

Although the endotoxin content per mg dust decreased from the first to the later stages in the production process, the endotoxin exposure correlated significantly with the inhalable dust level, and was considerably stronger (r=0.66) compared to a Tanzanian study (r=0.44)¹⁴⁾. This might be due to difference in study setting (i.e., type of industry, sampling season, and storage conditions for the raw material wood).

Factory (A, B) and downtime explained 36.0% of the between-worker variance inhalable dust, which is a lower proportion compared to earlier studies in the wood industry, for example in Denmark, which were able to explain 42.0% of the between-worker variability for inhalable dust³⁰⁾. Downtime explained a relatively small part of the within-worker variance (11.6%) in inhalable dust. Since none of the workers changed factory from day to day, factory could contribute to partly explain between workers variability only.

Workstations explained 58.7% between-worker variance for endotoxin. Chipping and flaking had the highest impact on endotoxin exposure. This is presumably due to the high content of endotoxin per mg inhalable dust in the first stages of the production process. It seems reasonable that factory and workstation were unable to explain any within-worker variance, since none of the workers changed workstations from day to day and the downtime of production was not included in the final mixed models.

Formaldehyde was detected in all selected workstations, except the first (outdoors) and last workstation (indoors). The highest concentration was recorded in blending, where the prepared formaldehyde solution and wood chips are mixed together. The measurements were considered as “worst-case” measurements. Thus, they do not indicate full shift levels of formaldehyde in the factory. Full-shift measurement of formaldehyde using the United States National Institute of Occupational Safety and Health method (NIOSH Method 2016) could have been used. However, such samples would only have 34 days of stability in cold storage (5°C) after sampling³³⁾, and it was not practical to apply this approach in our setting, because the exposure sampling was planned to take a minimum of 2 months. We are not aware of any other published data on such worst-case measurements of formaldehyde in particleboard industries. Full-shift measurements of formaldehyde in wood processing industries ranged from 0.5 to 1.52 ppm in Iran³⁴⁾ and from 0.02 to 2.2 ppm in Turkey³⁵⁾, while mean area concentration of formaldehyde varied from 0.03 mg/m³ to 0.31 mg/m³ in Finland³⁶⁾. In an American study, short term formaldehyde level varied from 0.28 to 3.48 ppm⁵⁾, and a geometric mean of 0.06 ppm was recorded in New Zealand³⁷⁾. The variation might be due to a difference in sampling (stationary vs. personal) duration (peak vs. full shift and short term), the measuring equipment (indicator tubes vs. sorbent tubes).

None of the factories had mechanical ventilation systems or dust-collecting hoods. On top of this, the particleboard workers did not use proper PPE during work, the details of which are published in a separate paper³⁸⁾. The absence of mechanical ventilation systems and proper PPE in this working environment with high wood dust, endotoxin, and formaldehyde exposure may cause respiratory health problems for the workers in the future. As previously published, there is a high prevalence of respiratory symptoms for the workers in these factories¹⁰⁾.

To our knowledge, this is the first study performed in large-scale particleboard manufacturing industries in Africa that use eucalyptus wood as a raw material. Strengths of the study include calibration of the pump air flow rate at 1-hour intervals, quality of the measurement equipment, the sample size, and several days of sampling with repeated measurements. The sampling time was 4 hours in factory A and it was reduced from 4 to 2 hours in factory B due to visible loose dust on the filters. Since the production did not change systematically during the day, we have considered the 2 to 4 hours sampling time to be representative for the workers full shift exposure.

The selection of persons from each working section was not totally random, but it is unlikely that it has influenced the representativeness of the study. Participants were selected in consultation with the supervisors in each workstation and in the presence of the researcher. Together with the supervisor, the individual participants were clearly informed the aim of the study, how to perform, and its benefit. The workers were all positive towards the study and highly cooperative. We understand that the workers behavior may challenge the validity of the collected data in many observational studies, but in the present study the researcher visited each study participant every hour to check the flow rate of the sampling pump and to monitor any unusual activities. No manipulation of the instruments or activities was seen. As a result, we do not expect the workers behavior to have manipulated the sampling results.

The study targeted large wood processing industries assumed to be representative of the large-scale particleboard manufacturing using eucalyptus trees in a developing country. However, it might be difficult to generalize the results to small- or medium-scale and less formal wood manufacturing industries, for which the situation could be different.

Conclusions

This study indicates that the geometric mean inhalable dust exposure in particleboard factories in Ethiopia was above 1 mg/m³, the TLV set by ACGIH, and the level varied by workstation. The geometric mean endotoxin level was lower than the 90 EU/m³ health-based Dutch occupational exposure limit. However, the level exceeded this limit in chipping and flaking workstations. The median formaldehyde concentration was highest in the blending workstation. A reduction in dust, endotoxin, and formaldehyde exposure levels in these workplaces is recommended. We also recommend proper personal protective equipment, as this type of protective equipment was not used in the factories.

Acknowledgments

We would like to acknowledge the Norwegian Educational Loan Fund (Lånekassen) and NORHED Project cooperation between University of Bergen (Norway), Addis Ababa University (Ethiopia) and Muhimbili University of Health and Allied Sciences (Tanzania) for financial support. The funding bodies have no role in in the design of the study, collection, analysis, and interpretation of data and in writing the manuscript. We also extend our appreciation to the factory management and production workers for their kind cooperation and facilitation during the study.

Funding

This study was funded by Norwegian Educational Loan Fund (Lånekassen) and NORHED Project.

Authors’ contributions

A.A.A, M.B., V.S. and B.E.M. conceived the ideas; A.A.A collected the data; A.A.A, M.B., V.S. and B.E.M. analyzed the data and wrote the manuscript

Conflict of interest

The Authors declare that there are no conflict of interest.

References

• 1.  Central Statistical Agency. Key findings on the 2013 national labour force survey. Addis Ababa; 2013.
• 2.  Environmental Protection Agency. Particleboard manufacturing. In: Wood Products Industry. Washington, DC: Environmental Protection Agency; 2002: 29.
• 3.   Liebers  V,  Brüning  T,  Raulf-Heimsoth  M. Occupational endotoxin-exposure and possible health effects on humans. Am J Ind Med. 2006; 49(6): 474-491.
• 4.   Sriproed  S,  Osiri  P,  Sujirarat  D, et al. Respiratory effects among rubberwood furniture factory workers in Thailand. Arch Environ Occup Health. 2013; 68(2): 87-94.
• 5.   Malaka  T,  Kodama  AM. Respiratory health of plywood workers occupationally exposed to formaldehyde. Arch Environ Health. 1990; 45(5): 288-294.
• 6.   Tomioka  K,  Kumagai  S,  Kameda  M,  Kataoka  Y. A case of occupational asthma induced by falcata wood (Albizia falcataria). J Occup Health. 2006; 48(5): 392-395.
• 7.   Færden  K,  Lund  MB,  Mogens Aaløkken  T, et al. Hypersensitivity pneumonitis in a cluster of sawmill workers: a 10-year follow-up of exposure, symptoms, and lung function. Int J Occup Environ Health. 2014; 20(2): 167-173.
• 8.   Jacobsen  G,  Schlünssen  V,  Schaumburg  I,  Taudorf  E,  Sigsgaard  T. Longitudinal lung function decline and wood dust exposure in the furniture industry. Eur Respir J. 2008; 31(2): 334-342.
• 9.   Saary  MJ,  House  RA,  Holness  DL. Dermatitis in a particleboard manufacturing facility. Contact Dermat. 2001; 44(6): 325-330.
• 10.   Asgedom  AA,  Bråtveit  M,  Moen  BE. High prevalence of respiratory symptoms among particleboard workers in Ethiopia: A Cross-sectional study. Int J Environ Res Public Health. 2019; 16(12): 2158.
• 11.   Duchaine  C,  Mériaux  A,  Thorne  PS,  Cormier  Y. Assessment of particulates and bioaerosols in eastern Canadian sawmills. AIHAJ - American Industrial Hygiene Association. 2000; 61(5): 727-732.
• 12.   Gioffrè  A,  Marramao  A,  Iannò  A. Airborne microorganisms, endotoxin, and dust concentration in wood factories in Italy. Ann Occup Hyg. 2012; 56(2): 161-169.
• 13.   Rongo  LM,  Besselink  A,  Douwes  J, et al. Respiratory symptoms and dust exposure among male workers in small-scale wood industries in Tanzania. J Occup Environ Med. 2002; 44(12): 1153-1160.
• 14.   Rongo  LM,  Msamanga  GI,  Burstyn  I,  Barten  F,  Dolmans  WM,  Heederik  D. Exposure to wood dust and endotoxin in small-scale wood industries in Tanzania. J Expo Anal Environ Epidemiol. 2004; 14(7): 544-550.
• 15.   Rylander  R,  Michel  O. Organic dust induced inflammation—role of atopy and TLR-4 and CD14 gene polymorphisms. Am J Ind Med. 2005; 48(4): 302-307.
• 16.   Scarselli  A,  Binazzi  A,  Ferrante  P,  Marinaccio  A. Occupational exposure levels to wood dust in Italy, 1996-2006. Occup Environ Med. 2008; 65(8): 567-574.
• 17.   Straumfors  A,  Olsen  R,  Daae  HL, et al. Exposure to wood dust, microbial components, and terpenes in the Norwegian sawmill industry. Ann Work Expo Health. 2018; 62(6): 674-688.
• 18.   Ayalew  E,  Gebre  Y,  De Wael  K. A survey of occupational exposure to inhalable wood dust among workers in small- and medium-scale wood-processing enterprises in Ethiopia. Ann Occup Hyg. 2015; 59(2): 253-257.
• 19.   Gardiner  K,  Harrington  JM. Occupational Hygiene. 3rd ed. Massachusetts, USA: Blackwell Publishing, Inc; 2005.
• 20.   Rappaport  SM,  Kupper  L. Quantitative Exposure Assessment. California, USA: Stephen Rappaport El Cerrito; 2008.
• 21.   Tatum  VL,  Ray  AE,  Rovell-Rixx  DC. The performance of personal inhalable dust samplers in wood-products industry facilities. Appl Occup Environ Hyg. 2001; 16(7): 763-769.
• 22.  Health and Safety Executive. General methods for sampling and gravimetric analysis of respirable, thoracic and inhalable aerosols. UK: HSE; 2014: 13.
• 23.   Basinas  I,  Sigsgaard  T,  Heederik  D, et al. Exposure to inhalable dust and endotoxin among Danish livestock farmers: results from the SUS cohort study. J Environ Monit. 2012; 14(2): 604-614.
• 24.   Thilsing  T,  Madsen  AM,  Basinas  I,  Schlünssen  V,  Tendal  K,  Bælum  J. Dust, endotoxin, fungi, and bacteria exposure as determined by work task, season, and type of plant in a flower greenhouse. Ann Occup Hyg. 2015; 59(2): 142-157.
• 25.  Dräger Safety AG & Co. KGaA. Dräger-Tubes & CMS-Handbook 16th edition. In: Soil, Water, and Air Investigations as well as Technical Gas Analysis. 16 ed.2011: https://www.draeger.com/Library/Content/tubeshandbook_br_9092086_en.pdf.
• 26.  American Conference of Governmental Industrial Hygienists. Threshold Limit Values and Biological Exposure Indices. In. Ohio, USA: ACGIH; 2019.
• 27.  Dutch expert Committee on Occupational Safety. Endotoxins: Health-based recommended occupational exposure limit. Health Council of the Netherlands. In. The Hague, Netherlands. 2010: 100.
• 28.  Ministry of Labor and Social Affairs. Regulations on measures and limit values. In. List of limit values for contaminants in the working atmosphere. Norway. 2011: 77.
• 29.   Scheeper  B,  Kromhout  H,  Boleij  JSM. Wood-dust exposure during wood-working processes. Ann Occup Hyg. 1995; 39(2): 141-154.
• 30.   Schlünssen  V,  Jacobsen  G,  Erlandsen  M,  Mikkelsen  AB,  Schaumburg  I,  Sigsgaard  T. Determinants of wood dust exposure in the Danish furniture industry—results from two cross-sectional studies 6 years apart. Ann Occup Hyg. 2008; 52(4): 227-238.
• 31.   Kauppinen  T,  Vincent  R,  Liukkonen  T, et al. Occupational exposure to inhalable wood dust in the member states of the European Union. Ann Occup Hyg. 2006; 50(6): 549-561.
• 32.   Dutkiewicz  J,  Olenchock  S,  Krysińska-Traczyk  E,  Skórska  C,  Sitkowska  J,  Prazmo  Z. Exposure to airborne microorganisms in fiberboard and chipboard factories. Ann Agric Environ Med. 2001; 8(2): 191-199.
• 33.  National Institute of Occupational Safety and Health. NIOSH Manual of Analytical Methods. In. Method 2016: Formaldehyde. Washington, DC. 2003: 7.
• 34.   Jafari  MJ,  Rahimi  A,  Omidi  L,  Behzadi  MH,  Rajabi  MH. Occupational Exposure and Health Impairments of Formaldehyde on Employees of a Wood Industry. Health Promot Perspect. 2016; 5(4): 296-303.
• 35.   Vaizoğlu  SA,  Aycan  S,  Akin  L, et al. Determination of formaldehyde levels in 100 furniture workshops in Ankara. Tohoku J Exp Med. 2005; 207(2): 157-163.
• 36.   Mäkinen  M,  Kalliokoski  P,  Kangas  J. Assessment of total exposure to phenol-formaldehyde resin glue in plywood manufacturing. Int Arch Occup Environ Health. 1999; 72(5): 309-314.
• 37.   Fransman  W,  McLean  D,  Douwes  J,  Demers  PA,  Leung  V,  Pearce  N. Respiratory symptoms and occupational exposures in New Zealand plywood mill workers. Ann Occup Hyg. 2003; 47(4): 287-295.
• 38.   Asgedom  AA,  Bråtveit  M,  Moen  BE. Knowledge, attitude and practice related to chemical hazards and personal protective equipment among particleboard workers in Ethiopia: a cross-sectional study. BMC Public Health. 2019; 19(1): 440.