Study population and design

This study was conducted in four paper manufacturing industries in Ethiopia, utilizing a comparative cross-sectional design, from March 2023 to May 2023. We selected the industries based on their size (large-scale industries) and the type of raw materials used. The same four paper mills were included in our previous studies on dust exposure and respiratory symptoms (Tafese et al. 2024a, b). Each of the four paper mills employed between 256 and 559 workers, qualifying them as large-scale industries according to the Eurostat definition (Eurostat 2010), which classifies businesses with more than 250 employees as large-scale.

Description of the industries and the paper-making process

The four mills are described in detail in Tafese et al. (2024a). In brief, all primarily utilized recycled paper as raw material. Two of the mills also used imported pulp in addition to recycled paper, while the other two relied exclusively on recycled paper to produce paper and paper products. Their daily production capacity ranged from 30 to 100 tons.

The paper-making process combines manual labor and mechanized systems, taking place in designated work areas with no control rooms present in any mills. Preparation begins with selecting and sorting raw materials, mainly recycled paper and pulp, which are pulped to break down fibers and remove contaminants. The pulp may be bleached for brightness. The pulp slurry is then placed on a moving wire mesh conveyor, where water drains away, and fibers bond to form a continuous wet sheet. This sheet passes through press rollers to remove moisture and is dried with heated rollers. A sizing treatment may enhance water and ink resistance. In the finishing stage, the dried paper is cut, trimmed, and may be coated or converted into products like packaging materials. Finally, the products undergo quality inspection and are packaged for transport and storage. All four industries were involved in similar production of a wide variety of paper and paper products. This included high-quality white bond paper commonly used for printing, as well as fluting paper, kraft liner, test liner, paper tubes, and cones (Tafese et al. 2024a). They also produced cardboard and corrugated carton boxes, which are essential for packaging and transporting products.

A control group was selected from the water-bottling industry, assuming minimal dust exposure, as previous research in a similar study area reported a mean dust exposure level of 0.33 mg/m2 in water-bottling facilities (Abaya et al. 2018).

Sample size determination and selection procedures

To determine the sample size, we used the mean difference with a 2:1 ratio for exposed and control groups, based on data from a previous study in Germany’s soft tissue paper industry (Kraus et al. 2004), which focused on forced expiratory volume in one second (FEV1) as the primary outcome. The sample size was calculated using the Openepi (http://www.openepi.com/SampleSize/SSMean.htm), incorporating the mean and standard deviation for FEV1 (exposed group: mean = 101% ± 18.1, control group: 107.3% ± 15.8), with a significance level of 0.05 and a power of 90%. Finally, we added 10% to account for non-response rate, resulting in 363 participants: 242 from the exposed group (paper industry workers) were randomly selected from 1405 workers and 121 from the control group (water bottling workers) were also randomly selected from 720 workers to participate in the study. One worker who had previously undergone abdominal surgery and two pregnant female workers were excluded during the data collection period.

Lung function measurements

Lung function tests were carried out per the guidelines for a spirometer (ATS 1995). We used a portable computer-based spirometer SPIROBANK II BASIC-Handheld (MIR 2023). The participants’ standing height and weight were measured using a standard weighing scale. The tests were performed during the day shift work, between 08:00 and 16:00, with the participants in a sitting position. Before conducting the test, we held a brief meeting with employees and supervisors to discuss the purpose of the study. During this meeting, we instructed the employees to refrain from alcohol consumption for at least 4 h, avoid strenuous physical activity within 30 min prior to the test, wear loose-fitting clothing, and avoid consuming a large meal within 2 h prior before the test.

We retained three acceptable manoeuvres with consistent (repeatable) results and selected the best of these for analysis based on the ATS criteria for acceptability: at least 3 good efforts—quick start, no coughing, no early termination (not less than 6 s), smooth and continuous exhalation. Repeatability was assessed by the best two FEV1 and FVC values, which should be within 150 mL of each other. Since there are currently no reference equations for the Ethiopian population to calculate predicted values, only the absolute mean values for lung function were provided in the results. The lung function parameters forced vital capacity (FVC) and forced expiratory volume in one second (FEV1), along with the ratio of FEV1/FVC, were registered. Participants with FEV1/FVC < 0.70 were classified as having airflow limitation (Vogelmeier et al. 2017). We excluded 83 tests among workers exposed to paper dust and 27 from water bottling workers due to invalid spirometry readings.

Participants’ height and weight were measured during the assessment while they were dressed in light clothing and without shoes. Weight was recorded in kilograms using a standard weighing scale, while height was measured in meters with participants standing upright, feet together, eyes level, and looking straight ahead. The weighing scales were calibrated using the automatic zero function, which resets the display to zero upon activation. Body mass index (BMI) was calculated for each participant using the formula weight (kg)/height2 (m2).

Alongside the spirometry tests, data was collected from each worker on socio-demographic factors including age, gender, educational background, and work history. This included weekly working hours (in hours), years of employment in the paper industry, previous experience in other dusty sectors (yes or no), departments and tenure in each category (preparation, paper machine, finishing, and packing), personal history of respiratory illnesses (yes or no), use of biofuels for cooking at home (yes or no), and smoking status (current or former smoker). The initial data was recorded in English and later translated into Afan Oromo and Amharic for fieldwork data collection.

Exposure assessment

In a previous study (Tafese et al. 2024a), we collected 150 personal inhalable paper dust samples from four distinct paper industries, each with four departments, resulting in an overall AM paper dust exposure of 4.5 mg/m3. Specifically, we collected ten personal inhalable paper dust samples from each of the 16 departments using the Institute of Occupational Medicine (IOM) samplers head made by SKC Ltd, running at a flow rate of 2.0 l min−1 (Skaugset et al. 2013). We developed a Job Exposure Matrix (JEM) (Neitzel et al. 2020; Tafese et al. 2024b) using the mean personal inhalable paper dust exposure data from each department: preparation, paper machine, finishing, and packing, across the four industries. Most of the workers had experience in multiple departments during the study period. For each paper worker, we calculated the cumulative dust exposure as the sum of the products of the AM of inhalable paper dust (mg/m3) in each department within the specific factory and the number of years the individual was employed in that department. The AM cumulative paper dust exposure was 30.3 mg year/m3, ranging from 2.5 to 134.1 mg year/m3.

Data management and analysis

Data was analysed using SPSS version 26 (IBM Corp., Armonk, NY, USA). Covariance (ANCOVA) analysis was performed to compare mean lung function parameters between the paper industry workers and control groups, adjusting for age, sex and height. Furthermore, multiple linear regression analyses were conducted to investigate potential relationship between cumulative dust exposure and lung function parameters among the paper dust-exposed workers, adjusting for age, sex, and height. An interaction term between sex and cumulative dust exposure was included to investigate potential differences of this relationship between men and women. While age and cumulative dust showed a high correlation (r = 0.76, p < 0.001), we included age in this regression analysis because both variables influence lung function.