Prevalence of Chronic Obstructive Pulmonary Disease (COPD) in Sri Lankan adults: results of a cross sectional study Authors: W. A. D. L. Amarasiri, A. R. Wickremasinghe, W. Gunasinghe, A. Sadikeen, A. Fernando, D. Medagedara, B. P. R. Perera, Kirthi Gunasekera Published: 2026 ABSTRACT BackgroundThis study assessed the prevalence of COPD, its associated risk factors and co-morbidities in Sri Lankan adults.MethodsA cross-sectional study was conducted in 3 rural and 4 urban areas in 7 out of 9 provinces of Sri Lanka using a randomly selected sample of Sri Lankan adults above 18 years. The Burden of Lung Disease (BOLD) survey tool was used with pre- and post-bronchodilator spirometry. Chronic Obstructive Pulmonary Disease (COPD) was classified according to the Global Initiative on Obstructive Lung Disease (GOLD) criteria. The association of COPD with known risk factors: smoking, passive smoking, dusty job, use of firewood; and comorbidities; diabetes mellitus, past history of tuberculosis, hypertension, heart disease, stroke and lung cancer was assessed.ResultsOf 1985 subjects (45.1% male) assessed, the overall prevalence of COPD was 7.3% (95% CI 6.1%-8.4%) and the prevalence among participants above 40 years of age was 10.5% (95% CI 8.9% -12.3%). Among those with COPD, 38.9% was in stage 1, 45.8% in stage 2, 10.4% in stage 3 and 4.9% in stage 4. The prevalence of COPD was significantly higher in males (10.4%) than in females (4.7%). Of the COPD patients, 54.8% were non-smokers. COPD was significantly associated with sex, smoking and exposure to dust among the over-40-year age group. Other known risk factors such as exposure to biomass fumes and a past history of tuberculosis were not significant risk factors in this sample.ConclusionsThe prevalence of COPD in Sri Lankan adults is similar to global figures, is more common in males and the majority had less severe forms of the disease. A considerable proportion of COPD patients were females and non-smokers. Interventions focusing on targeted educational campaigns for risk groups and having facilities for spirometry at primary care level while decreasing the incidence of COPD is likely to improve early detection of the condition.Supplementary InformationThe online version contains supplementary material available at 10.1186/s12890-026-04390-0. Background Chronic obstructive pulmonary disease (COPD) is a preventable chronic inflammatory airway disease, a major cause of morbidity and mortality globally [1, 2]. It is characterized by the persistent limitation of airflow, usually progressive and not fully reversible [3]. There was an estimated COPD disease burden of over 200 million people worldwide, over 3 million deaths, and 74.4 million disability adjusted life years (DALYs) in 2019 [4, 5]. COPD is usually not diagnosed at the early stages until it impairs the quality of life due to the poor utilization of spirometry in primary healthcare settings and the silent nature of the disease itself. Therefore, this condition is under-recognized by doctors and under-reported in individuals [6, 7]. The World Health Organization (WHO) estimates that the total number of deaths globally due to COPD will increase by > 30% in the next 10 years; by 2030, COPD would become the fourth leading cause of death worldwide [8], primarily due to continued exposure to COPD risk factors and population ageing [9]. The burden of COPD in Asia is currently greater than that in high-income countries, both in terms of the total number of deaths and the burden of disease, as measured in years of life lost and years spent living with disability [10]. It is estimated that 90% of COPD-related deaths are reported in low- and middle-income countries [1, 11, 12]. A recent systematic review of the prevalence of COPD in South Asia reported that the estimated pooled prevalence of COPD was 11.1% (95% CI: 7.4–14.8%) using the Global Initiative for Chronic Obstructive Lung Disease (GOLD) fixed criteria, and 8.0% (95% CI: 5.6–10.4%) using the lower limit of normal criteria, with the highest prevalence in north India (19.4%) and Bangladesh (13.5%), and in men [13]. According to the 2017 Global Burden of Disease study, COPD is the most common cause of premature deaths among chronic respiratory diseases in South Asia mainly linked to tobacco exposure and indoor and outdoor air pollution [13, 14]. The prevalence of COPD in Sri Lanka has not been reported in detail previously although COPD is a significant burden on the healthcare system of the country. The prevalence of asthma and its symptoms in Sri Lankan adults [15] and COPD in non-smoking Sri Lankan adults [16] have already been published. The purpose of this study was to determine the prevalence of COPD among Sri Lankan adults and country-specific risk factors of COPD and other related respiratory diseases. Methods The data for this study were drawn from the cross-sectional survey conducted in June and July 2013 under the Global Burden of Obstructive Lung Disease (BOLD) initiative involving Sri Lankan adults. This study was approved by the Ethics Review Committee of the Faculty of Medicine, University of Kelaniya, Sri Lanka (ERC No P5/01/2011). All study participants provided written informed consent. Sampling plan and recruitment The sampling methodology was reviewed and approved by the BOLD coordinating centre. Sampling was done randomly from non-institutionalized individuals living within the administrative boundaries of the selected geographical areas using the recent voting lists. The sampling plan and recruitment method have been described in detail elsewhere [15, 16]. Houses were sampled within administrative boundaries and one individual from each household was randomly selected. A summary of the recruitment is given in Additional File 1. The flow chart of the study design is given in Additional File 2. Questionnaires The Sinhala and Tamil versions of the interviewer-administered BOLD core questionnaire [17] was used to obtain information about respiratory symptoms (cough, sputum, wheezing, shortness of breath), exposure to potential risk factors, occupation, respiratory diagnoses (asthma, emphysema, COPD, chronic bronchitis, etc.), co-morbidities, healthcare utilization, medication use, activity limitation, and health status without any modifications. The questionnaire included sections from the ATS/DLD Respiratory Symptoms Questionnaire developed by the American Thoracic Society (ATS) for the Division of Lung Diseases (DLD) [18] and the questionnaires used in the European Community Respiratory Health Study [19], the National Research Council of Italy (CNR) study [20], the Obstructive Lung Disease in Northern Sweden (OLIN) study [21] and Short Form-12 (SF-12) to assess the overall health status [22]. Definitions A geographical area administered by a municipal or urban council as per the local government system in Sri Lanka was defined as an urban area. The estate sector was defined as plantations which are 20 acres or more in extent and with ten or more resident labourers. A rural area was defined as an area not categorised as an urban or estate sector area [23]. Data collection Invited participants who presented themselves at mobile clinics were explained the objectives of the study and informed written consent was obtained. Interviews were carried out by trained staff. Patients were excluded if they had a history of prior lung surgery including lung volume reduction, uncontrolled cancer, or radiation therapy to the chest, mediastinum, or breast. Patients who refused salbutamol administration and pregnant women were excluded. Physical examination All participants were examined by a physician and given necessary advice if required. Standing height and the weight of each participant were recorded. Height was measured using a wall-mounted stadiometer. Body mass index (BMI) was calculated as the patient’s weight in kilograms divided by the square of the height in meters. The WHO BMI cut-off points of 18·5–24·9 kg/m2 (normal range), > 25 (overweight), 25–29·9 kg/m2 (pre-obese), > 30 kg/m2 (obese) were used [24]. Spirometry The diagnosis and severity of COPD were assessed on the Global Initiative on Obstructive Lung Disease (GOLD) criteria based on spirometry. Spirometry testing was performed in all subjects using the NDD EasyOne™ Spirometer, as prescribed by the BOLD protocol, using the “American Thoracic Society/European Respiratory Society (ATS/ ERS) standards”. Spirometry was conducted by a staff member trained under the instructions of the BOLD core team. Spirometry was performed with participants seated, using disposable mouthpieces having antibacterial filters to reduce infection risk. Participants were advised not to take any bronchodilators 6–12 h before the testing. Salbutamol 200mcg was administered via a metered dose inhaler and spacer. Two readings were recorded, 15 min apart, one reading before and one reading following the administration. During the assessment the participants were observed for signs of dizziness or other problems and the staff were instructed to stop the manoeuvre immediately in the event of an adverse outcome. All spirometry data were reviewed by experts and quality control was assured according to the BOLD study protocol [25]. Spirometry measurements The highest reading of the forced expiratory volume during the first second (FEV1) and the forced vital capacity from the accepted study trials were obtained as the measurements [25]. Individuals with FEV1/FVC ratio less than 0.70 following bronchodilator administration were classified as individuals having COPD based on the GOLD spirometry criteria. Among them, persons with FEV1 > 80% of predicted were classified as GOLD stage 1: FEV1 50–80% of predicted as GOLD stage 2: FEV1 30–50% of predicted as GOLD stage 3: and FEV1 < 30% of predicted as GOLD stage 4. All spirometry data were sent electronically to the Coordinating Centre for review of quality and accuracy. Participants were excluded if they were not within the age range of the study or had unsatisfactory spirometry results as spirometry was used to classify whether a participant had COPD or not. Data management and analysis Individual questionnaire data were double entered in web-based forms and transmitted via internet directly to the Coordinating Centre, where a variety of edit checks was conducted. All completed forms were stored at the Central Chest Clinic, Colombo, Sri Lanka. Data were entered into a password protected computer. For the prevalence estimates and other confidence interval calculations, only the data from the participants who had completed the questionnaire and had acceptable post-bronchodilator spirometry data were used. Demographic information and health profiles of the participants were reported using counts and frequencies. Between groups comparisons were done using t-test for proportions or independent sample t-test. COPD risk factors and the prevalence of other respiratory symptoms among COPD patients were analysed using the chi square statistic, stratified by the age group. The impact of each risk factor on COPD was evaluated using logistic regression analyses for the < 40- and > 40-year-old age groups, separately. P-values < 0.05 were considered statistically significant. All statistical analyses were performed using SPSS version 22.0 software [26]. Ethics statement Approval was obtained from the Ethics Review Committee of the Faculty of Medicine, University of Kelaniya, Sri Lanka (Ref No. P5/01/2011). Informed written consent was obtained from all participants. Confidentiality of data was maintained. If participants had a medical problem, they were advised or referred to the nearest chest clinic which provided healthcare services free-of-charge. Patient and public involvement The study was conducted according to the BOLD protocol and there was no patient involvement during the planning stage. The protocol was explained to the participants prior to data collection; participants were given access to their screening results and were directed to medical care, if necessary. Results A total of 1985 participants were included in the analysis. The socio-demographic profile of participants is given in Table 1. Females had significantly higher BMIs, but lower prevalence of COPD than in males; none of the females had ever smoked. Table 1Participants’ profile (n=1985) Males N=895 (45.1%) Females N=1090 (54.9 %) Test statistic P value Age in years (mean, SD)44.7 (14.9)45.5 (13.5)t = (-1.279)ap = 0.201Age category (n (%)) 18-39 years357 (39.9)389 (35.7)\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\chi_{1}^{2}$$\end{document}=3.696,p=0.055 40 and above538 (60.1)701 (64.3)Body mass index in kg/m2 (mean, SD)22.9 (4.2)24.6 (5.1)t = (-8.435)ap < 0.001Body mass index category (n (%)) Underweight 136 (15.2)113 (10.4)\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\chi_{3}^{2}$$\end{document}=63.474, p < 0.001 Normal weight for height488 (54.5)481 (44.1) Pre-obese222 (24.8)342 (31.4) Obese49 (5.5)154 (14.1)Educational level (n (%))* Primary School176 (20.9)199 (20.0)\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\chi_{3}^{2}$$\end{document}=0.917,p=0.821c Middle School429 (51.0)509 (51.2) High School191 (22.7)237 (23.8) College or University30 (3.5)29 (2.9) None or unknown16 (1.9)Smoking status (n (%))* Ever smokers341 (40.5)00 (0.0)\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\chi_{2}^{2}$$\end{document}=498.001p<0.001 Never smokers502 (59.5)993 (100)Duration of smoking in years (mean (sd))23.2 (14.1)--Health Profile (n (%)) Tuberculosis07 (0.8)08 (0.8)t=0.123b, p=0.90 Diabetes72 (8.5)100 (10.1)t=-0.89b, p=0.37 Hypertension113 (13.4)161 (16.2)t=-1.37b, p=0.16 Heart Diseases36 (4.3)32 (3.2)t=1.32b, p=0.18Doctor diagnosed emphysema, asthma, chronic bronchitis or COPD41 (4.6)73 (6.7)t=-2.01b, p=0.04Spirometry-diagnosed COPD (n (%))93 (10.4)51 (4.7)t=4.88b, p<0.001aindependent sample t-test comparing two meansbt-test comparing two proportionsceducational level was categorized into 4 groups as no education, primary education, middle school and/or high school education, and college and/or university education. For the chi square test, no education or unknown group was excluded*excluding persons with missing values Table 1Participants’ profile (n=1985) Males N=895 (45.1%) Females N=1090 (54.9 %) Test statistic P value Age in years (mean, SD)44.7 (14.9)45.5 (13.5)t = (-1.279)ap = 0.201Age category (n (%)) 18-39 years357 (39.9)389 (35.7)\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\chi_{1}^{2}$$\end{document}=3.696,p=0.055 40 and above538 (60.1)701 (64.3)Body mass index in kg/m2 (mean, SD)22.9 (4.2)24.6 (5.1)t = (-8.435)ap < 0.001Body mass index category (n (%)) Underweight 136 (15.2)113 (10.4)\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\chi_{3}^{2}$$\end{document}=63.474, p < 0.001 Normal weight for height488 (54.5)481 (44.1) Pre-obese222 (24.8)342 (31.4) Obese49 (5.5)154 (14.1)Educational level (n (%))* Primary School176 (20.9)199 (20.0)\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\chi_{3}^{2}$$\end{document}=0.917,p=0.821c Middle School429 (51.0)509 (51.2) High School191 (22.7)237 (23.8) College or University30 (3.5)29 (2.9) None or unknown16 (1.9)Smoking status (n (%))* Ever smokers341 (40.5)00 (0.0)\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\chi_{2}^{2}$$\end{document}=498.001p<0.001 Never smokers502 (59.5)993 (100)Duration of smoking in years (mean (sd))23.2 (14.1)--Health Profile (n (%)) Tuberculosis07 (0.8)08 (0.8)t=0.123b, p=0.90 Diabetes72 (8.5)100 (10.1)t=-0.89b, p=0.37 Hypertension113 (13.4)161 (16.2)t=-1.37b, p=0.16 Heart Diseases36 (4.3)32 (3.2)t=1.32b, p=0.18Doctor diagnosed emphysema, asthma, chronic bronchitis or COPD41 (4.6)73 (6.7)t=-2.01b, p=0.04Spirometry-diagnosed COPD (n (%))93 (10.4)51 (4.7)t=4.88b, p<0.001aindependent sample t-test comparing two meansbt-test comparing two proportionsceducational level was categorized into 4 groups as no education, primary education, middle school and/or high school education, and college and/or university education. For the chi square test, no education or unknown group was excluded*excluding persons with missing values Participants’ profile (n=1985) t = (-1.279)a p = 0.201 \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\chi_{1}^{2}$$\end{document}=3.696, p=0.055 t = (-8.435)a p < 0.001 \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\chi_{3}^{2}$$\end{document}=63.474, p < 0.001 \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\chi_{3}^{2}$$\end{document}=0.917, p=0.821c \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\chi_{2}^{2}$$\end{document}=498.001 p<0.001 aindependent sample t-test comparing two means bt-test comparing two proportions ceducational level was categorized into 4 groups as no education, primary education, middle school and/or high school education, and college and/or university education. For the chi square test, no education or unknown group was excluded *excluding persons with missing values Prevalence of COPD The overall prevalence of COPD was 7.3% (95% CI 6.1%-8.4%), the prevalence being higher in males (10.4% vs. 4.7%) (Table 1). The prevalence in the above 40-year age group was 10.5% (95% CI 8.7%-12.2%), whereas in the younger population it was 1.9% (95% CI 0.9%-2.8%) (Table 2). Among all the participants diagnosed with COPD using the standard spirometry criteria according to the GOLD classification, most were in stage 2 (45.8%), followed by stage 1 (38.9%). This pattern was consistent in the above-40 age group. However, in the below-40 age group, a higher number of patients were from stage 1 (Fig. 1). Table 2Risk factors of COPDVariableAge < 40(n = 746)Age ≥ 40(n = 1239)No COPD(n = 732)COPD1(n = 14)Test statisticP valueNo COPD(n = 1109)COPD1(n = 130)Test statisticP valueSex (n (%)) Male352 (98.6)05 (1.4)\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\chi_{1}^{2}$$\end{document}= 0.843, p = 0.359450 (83.6)88 (16.4)\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\chi_{1}^{2}$$\end{document}=34.8, p < 0.001 Female380 (97.7)09 (2.3)659 (94.0)42 (6.0)Smoking status (n (%))* Ever-smokers103 (98.1)02 (1.9)Fisher’s exact: p = 1.000186 (78.5)51 (21.5)\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\chi_{1}^{2}$$\end{document}= 40.3, p < 0.001 Never smokers569 (98.1)11 (1.9)847 (92.6)68 (7.4)Duration of smoking in years (mean (sd))10.3 (5.6)13.6 (6.1)t103=-0.824, p = 0.41226.1 (12.3)39.0 (10.1)T234=-6.806, p < 0.001Duration of dust exposure in years (mean (sd))1.76 (4.2)2.69 (6.2)T683=-0.782, p = 0.4345.99 (10.9)10.61 (14.0)T1150 =-4.228, p < 0.0011Diagnosis based on spirometry (please refer methods for classification of COPD based on spirometry)*excluding persons with missing values Table 2Risk factors of COPDVariableAge < 40(n = 746)Age ≥ 40(n = 1239)No COPD(n = 732)COPD1(n = 14)Test statisticP valueNo COPD(n = 1109)COPD1(n = 130)Test statisticP valueSex (n (%)) Male352 (98.6)05 (1.4)\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\chi_{1}^{2}$$\end{document}= 0.843, p = 0.359450 (83.6)88 (16.4)\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\chi_{1}^{2}$$\end{document}=34.8, p < 0.001 Female380 (97.7)09 (2.3)659 (94.0)42 (6.0)Smoking status (n (%))* Ever-smokers103 (98.1)02 (1.9)Fisher’s exact: p = 1.000186 (78.5)51 (21.5)\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\chi_{1}^{2}$$\end{document}= 40.3, p < 0.001 Never smokers569 (98.1)11 (1.9)847 (92.6)68 (7.4)Duration of smoking in years (mean (sd))10.3 (5.6)13.6 (6.1)t103=-0.824, p = 0.41226.1 (12.3)39.0 (10.1)T234=-6.806, p < 0.001Duration of dust exposure in years (mean (sd))1.76 (4.2)2.69 (6.2)T683=-0.782, p = 0.4345.99 (10.9)10.61 (14.0)T1150 =-4.228, p < 0.0011Diagnosis based on spirometry (please refer methods for classification of COPD based on spirometry)*excluding persons with missing values Risk factors of COPD 1Diagnosis based on spirometry (please refer methods for classification of COPD based on spirometry) *excluding persons with missing values Fig. 1Disease severity among COPD patients diagnosed by spirometry by sex and age category (n = 144) Disease severity among COPD patients diagnosed by spirometry by sex and age category (n = 144) Among spirometry-diagnosed COPD patients, a higher proportion was over 40 years of age (90.3% vs. 9.7%). The percentages of males and older persons were significantly higher among spirometry-diagnosed COPD patients than in non-COPD individuals (Table 2). Among the above-40 sample, higher proportion of males had COPD while in the below 40 group, there was no difference between COPD prevalence based on sex; however, the prevalence of COPD of varying degrees of severity was significantly higher among males 40 years and above than in females. None of the females had stage 4 COPD. Spirometry-diagnosed COPD patients had a higher proportion of ever-smokers (15.5% vs. 5.3%), the difference being significant among the above-40-year age-group only (p < 0.001). Among ever-smokers, the duration of smoking was significantly higher in COPD patients (38.0 years vs. 20.5 years, p < 0.001). COPD patients were exposed longer to dusty environments than non-COPD participants (Table 2). The prevalences of all other respiratory symptoms were significantly higher in spirometry-diagnosed COPD patients with coughing, phlegm and wheezing being prominent (Table 3). The proportions diagnosed with asthma, emphysema and chronic bronchitis was higher among COPD patients above-40 years (Table 3). 125 participants, i.e. 7.2% of the total sample, who were diagnosed with COPD have never been diagnosed with the condition in any of the prior visits to a healthcare professional. COPD was previously diagnosed in 11 (0.6%) non-COPD participants based on spirometry; 7 participants who were previously diagnosed with COPD had evidence of COPD based on spirometry. Table 3Prevalence of other respiratory symptoms among spirometry-diagnosed COPD and non-COPD categoriesCondition*Age < 40Age > 40No COPDN = 732 (98.1%)n (%)COPD1N = 14 (1.9%)n (%)Test statisticP valueNo COPDN = 1109 (89.5%)n (%)COPD1N = 130 (10.5%)n (%)Test statistic2P valueCoughs without having a cold (n (%))141 (21.0)04 (30.8)Fisher’s exact test: p value = 0.489281 (27.2)55 (46.2)\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\chi_{1}^{2}$$\end{document}=18.6, p<0.001Phlegm without having a cold (n (%))215 (32.0)07(53.8)Fisher’s exact test: p value = 0.132358 (34.7)66 (55.5)\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\chi_{1}^{2}$$\end{document}=19.8, p < 0.001Coughs most days at least 3 months/year (n (%))18 (2.7)02(15.4)Fisher’s exact test: p value = 0.05255 (5.3)20 (16.8)\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\chi_{1}^{2}$$\end{document}=23.1, p < 0.001Phlegm most days at least 3 months/year (n (%))50 (7.4)05 (38.5)Fisher’s exact test: p value = 0.002105 (10.2)24 (20.2)\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\chi_{1}^{2}$$\end{document}=10.7, p = 0.001Wheezing during the last 12 months (n (%))132 (19.6)04 (30.8)Fisher’s exact test: p value = 0.302276 (26.7)68 (57.1)\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\chi_{1}^{2}$$\end{document}=47.1, p < 0.001Wheezing and shortness of breath during the last 12 months (n (%))98 (14.6)04(30.8)Fisher’s exact test: p value = 0.114188 (18.2)49 (41.2)\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\chi_{1}^{2}$$\end{document}=34.4, p < 0.001Diagnosed with emphysema (n (%))12 (1.8)01 (7.7)Fisher’s exact test: p value = 0.22322 (2.1)07 (5.9)Fisher’s exact test: p = 0.024Diagnosed with asthma (n (%))15 (2.2)01 (7.7)Fisher’s exact test: p value = 0.26751 (4.9)18(15.1)\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\chi_{1}^{2}$$\end{document}=19.6, p < 0.001Current asthma (n (%))09 (1.3)01 (7.7)Fisher’s exact test: p value = 0.17633 (3.2)14 (11.8)Fisher’s exact test: p < 0.001Diagnosed with Chronic Bronchitis (n (%))04 (0.6)00 (0.0)Fisher’s exact test: p value = 1.00013 (1.3)09 (7.6)Fisher’s exact test: p < 0.001Current chronic bronchitis (n (%))02 (0.3)000 (00)Fisher’s exact test: p value = 1.00009 (0.9)07 (5.9)Fisher’s exact test: p = 0.001Diagnosed with COPD (n (%))02 (0.3)00 (0.0)Fisher’s exact test: p value = 1.00009 (0.9)07 (5.9)Fisher’s exact test: p = 0.001Past history of Tuberculosis (n (%))06 (0.9)00 (00)Fisher’s exact test: p value = 1.00006 (0.6)03 (2.5)Fisher’s exact test: p = 0.0571Diagnosis based on spirometry2Chi square statistic was used when the expected value for < 20% of the cells exceeded 5. The Fisher’s exact test was used when the expected value for > 20% of cells exceeded 5*Excluding persons with missing values Table 3Prevalence of other respiratory symptoms among spirometry-diagnosed COPD and non-COPD categoriesCondition*Age < 40Age > 40No COPDN = 732 (98.1%)n (%)COPD1N = 14 (1.9%)n (%)Test statisticP valueNo COPDN = 1109 (89.5%)n (%)COPD1N = 130 (10.5%)n (%)Test statistic2P valueCoughs without having a cold (n (%))141 (21.0)04 (30.8)Fisher’s exact test: p value = 0.489281 (27.2)55 (46.2)\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\chi_{1}^{2}$$\end{document}=18.6, p<0.001Phlegm without having a cold (n (%))215 (32.0)07(53.8)Fisher’s exact test: p value = 0.132358 (34.7)66 (55.5)\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\chi_{1}^{2}$$\end{document}=19.8, p < 0.001Coughs most days at least 3 months/year (n (%))18 (2.7)02(15.4)Fisher’s exact test: p value = 0.05255 (5.3)20 (16.8)\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\chi_{1}^{2}$$\end{document}=23.1, p < 0.001Phlegm most days at least 3 months/year (n (%))50 (7.4)05 (38.5)Fisher’s exact test: p value = 0.002105 (10.2)24 (20.2)\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\chi_{1}^{2}$$\end{document}=10.7, p = 0.001Wheezing during the last 12 months (n (%))132 (19.6)04 (30.8)Fisher’s exact test: p value = 0.302276 (26.7)68 (57.1)\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\chi_{1}^{2}$$\end{document}=47.1, p < 0.001Wheezing and shortness of breath during the last 12 months (n (%))98 (14.6)04(30.8)Fisher’s exact test: p value = 0.114188 (18.2)49 (41.2)\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\chi_{1}^{2}$$\end{document}=34.4, p < 0.001Diagnosed with emphysema (n (%))12 (1.8)01 (7.7)Fisher’s exact test: p value = 0.22322 (2.1)07 (5.9)Fisher’s exact test: p = 0.024Diagnosed with asthma (n (%))15 (2.2)01 (7.7)Fisher’s exact test: p value = 0.26751 (4.9)18(15.1)\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\chi_{1}^{2}$$\end{document}=19.6, p < 0.001Current asthma (n (%))09 (1.3)01 (7.7)Fisher’s exact test: p value = 0.17633 (3.2)14 (11.8)Fisher’s exact test: p < 0.001Diagnosed with Chronic Bronchitis (n (%))04 (0.6)00 (0.0)Fisher’s exact test: p value = 1.00013 (1.3)09 (7.6)Fisher’s exact test: p < 0.001Current chronic bronchitis (n (%))02 (0.3)000 (00)Fisher’s exact test: p value = 1.00009 (0.9)07 (5.9)Fisher’s exact test: p = 0.001Diagnosed with COPD (n (%))02 (0.3)00 (0.0)Fisher’s exact test: p value = 1.00009 (0.9)07 (5.9)Fisher’s exact test: p = 0.001Past history of Tuberculosis (n (%))06 (0.9)00 (00)Fisher’s exact test: p value = 1.00006 (0.6)03 (2.5)Fisher’s exact test: p = 0.0571Diagnosis based on spirometry2Chi square statistic was used when the expected value for < 20% of the cells exceeded 5. The Fisher’s exact test was used when the expected value for > 20% of cells exceeded 5*Excluding persons with missing values Prevalence of other respiratory symptoms among spirometry-diagnosed COPD and non-COPD categories 1Diagnosis based on spirometry 2Chi square statistic was used when the expected value for < 20% of the cells exceeded 5. The Fisher’s exact test was used when the expected value for > 20% of cells exceeded 5 *Excluding persons with missing values A logistic regression analysis was performed to ascertain the effects of sex, job characteristics (whether the job involves working in a dusty environment), fuel type used for cooking, smoking status and a past history of tuberculosis on the likelihood of participants being diagnosed with COPD separately for the 2 age groups. All variables except sex and a history of TB were independently associated with COPD among the > 40-year-old age groups; none of the variables were significant in the multivariable model among the < 40-year-old age group. The multiple logistic regression model of the > 40-year age group with sex, smoking status and job characteristic was statistically significant (χ23 = 94.210, p < 0.001) (Table 4). The model explained 8.1% (Nagelkerke R2) of the variance in COPD and correctly classified 89.7% of cases. Ever-smokers had a 2 (95% CI 1.286–3.413) times higher risk of developing COPD after controlling for sex and dusty environment (p = 0.003). Working in a dusty environment increased the likelihood of developing COPD 1.569 (95% CI 1.033–2.381) times after controlling for smoking status and sex (p = 0.034) (Table 4). No significant interactions were present in the model. Cooking fuel type and past history of TB were not associated with COPD after controlling for sex, smoking status and job characteristic. Table 4Association between COPD and risk factors Variable Age <40 Age > 40 Unadjusted Model Adjusted Model (for sex, smoking status and job characteristics) Unadjusted Model Adjusted Model (for sex, smoking status and job characteristics) OR (95% confidence interval of OR) P value OR (95% confidence interval of OR) P value OR (95% confidence interval of OR) P value OR (95% confidence interval of OR) P value Sex Male0.600 (0.198–1.797)0.3643.068 (2.084–4.517)< 0.0011.754 (1.045–2.947)0.034 Female1.0001.0001.000Job characteristic Dusty1.399 (0.426–4.613)0.5802.319 (1.581–3.401)< 0.0011.569(1.033–2.381)0.034 No1.0001.0001.000Cooking fuel type Wood/crop residue use for more than 6 months4.584 (0.980-21.313)0.0531.053 (0.704–1.576)0.700 Clean fuel1.0001.000Smoking status Ever smoked1.004 (0.220–4.606)0.9953.415 (2.298–5.076)< 0.0012.095 (1.286–3.413)0.003 Never smoked1.0001.0001.000Past history of Tuberculosis (TB) Yes0.0000.9994.427 (1.093–17.936)0.037 No1.0001.000 Table 4Association between COPD and risk factors Variable Age <40 Age > 40 Unadjusted Model Adjusted Model (for sex, smoking status and job characteristics) Unadjusted Model Adjusted Model (for sex, smoking status and job characteristics) OR (95% confidence interval of OR) P value OR (95% confidence interval of OR) P value OR (95% confidence interval of OR) P value OR (95% confidence interval of OR) P value Sex Male0.600 (0.198–1.797)0.3643.068 (2.084–4.517)< 0.0011.754 (1.045–2.947)0.034 Female1.0001.0001.000Job characteristic Dusty1.399 (0.426–4.613)0.5802.319 (1.581–3.401)< 0.0011.569(1.033–2.381)0.034 No1.0001.0001.000Cooking fuel type Wood/crop residue use for more than 6 months4.584 (0.980-21.313)0.0531.053 (0.704–1.576)0.700 Clean fuel1.0001.000Smoking status Ever smoked1.004 (0.220–4.606)0.9953.415 (2.298–5.076)< 0.0012.095 (1.286–3.413)0.003 Never smoked1.0001.0001.000Past history of Tuberculosis (TB) Yes0.0000.9994.427 (1.093–17.936)0.037 No1.0001.000 Association between COPD and risk factors Discussion This study was the Sri Lankan arm of the BOLD initiative and the first assessment of COPD prevalence in Sri Lanka. The overall prevalence of COPD in adults was 7.3% (95% CI 6.1%-8.4%), with the majority having mild to moderately severe disease. The prevalence of COPD based on spirometry-evidence was higher in males and in the above-40 age group. A systematic review conducted using studies published between 2004 and 2015 estimated the global COPD prevalence of COPD to be 12.16% (95% CI = 10.91%-13.40%) [27]. The Sri Lankan estimates are lower but consistent with the global values. In this sample, the majority of COPD patients had stage 1 or 2 severity. A similar pattern is seen in the global prevalence where the most common stage of irreversible airflow obstruction was stage 2, with a prevalence of 50.46% (95% CI 44.59% − 56.33%), followed by stage 1, with a prevalence of 35.21% (95% CI 26.70%- 44.23%) [28]. COPD stages 3 and 4 are less prevalent. In our sample, only 7.3% were diagnosed to have COPD based on spirometry. However, a high percentage of participants who did not have a significant obstructive deficit answered positively to questions which are suggestive of asthma, emphysema or bronchitis. Some have even been diagnosed to have such conditions by a healthcare professional. Mild COPD is often treated as asthma and air flow obstruction is a late presentation. The demarcation between asthma and COPD is not very clear due to limitations of spirometry. In this study, COPD was previously diagnosed in 11 (0.6%) non-COPD participants based on spirometry while seven participants who were previously diagnosed with COPD had evidence of COPD based on spirometry. It is possible that at least a small proportion of participants who have been diagnosed to have asthma, emphysema and/chronic bronchitis are at the initial stage of the disease. Moreover, the self-reported physician diagnosis of COPD relied on the knowledge of the terms COPD, chronic bronchitis and emphysema by physicians [29]. It is concerning that 7.2% of the sample who were never diagnosed to have COPD by a doctor had air flow reduction and COPD confirmed through spirometry. It shows the magnitude of the problem that goes unnoticed. This could possibly be due to spirometry not being used at primary care level and physicians relying mostly on other clinical features to establish a diagnosis. Smoking is one of the most important risk factors of COPD which was confirmed by a recent meta-analysis done on COPD. Studies show that the prevalence of COPD among ever-smokers is two times higher than that of non-smokers [27, 30]. Consistently, COPD prevalence was significantly higher in ever-smoker males in this sample and the adjusted odds ratio of developing COPD was 2.095 (95% CI: 1.286–3.413) than in non-smokers. The ever-smoker prevalence among males in this sample was 40%; none of the females had ever smoked. The overall prevalence of smoking in Sri Lanka in 2013 was 34.5% with age-wise prevalences being 36.7% for the above 40-year age group, 35.9% for the 25-39-year age group and 26.9% for the 15-24-age group. In 2018, among male smokers 7.2.% smoked daily [31]. In 2015, the prevalence of smoking among males was 43% whereas among females it was 3% [32]. Among participants < 40 years of age, COPD was not significantly independently associated with any of the risk factors considered i.e., sex, smoking status, job characteristics, use of cooking fuel type and a past history of tuberculosis. These are known confounders of COPD; the associations we report here have adjusted for these potential confounding factors. The wide ranges of the 95% confidence intervals of the unadjusted odds ratios are probably due to the small size in this group. Even though a past history of tuberculosis is an identified risk factor of COPD [33], in our sample the association was marginally significant in the over-40-year age group only. This could probably be due to the small sample size of 15 with a history of tuberculosis. It has been shown that COPD is a heterogeneous and multifactorial chronic condition which is preventable [34]. Individuals are exposed to different types of particulate matter during different stages of life and their effects are different. The impact of dusty environment on the development of COPD has been demonstrated. In our study, there was no association between biomass use and COPD. As smokers were predominantly males, exposure to biomass fumes and other life style factors is unlikely to have influenced development of COPD in this sample. Indoor air pollution was shown to be a significant risk factor of COPD among noon-smoking females [16]. A recent systematic review and meta-analysis revealed that the prevalence of COPD is higher in males [28]. We have observed a similar trend with a prevalence of 10.4% (95% CI = 8.4%-12.5%) among males compared to 4.7% (95% CI = 3.5%-6.1%) among females. However, this difference was significant only among the over-40-year age group. The prevalence of COPD among females is increasing globally making it an important factor to consider [35]. In a large scale cross sectional study of the five most populated South Asian countries, the prevalence rates of asthma were 5.2%, 6.3%, 4.2%, 3.7%, and 5.3% in Bangladesh, India, Nepal, Pakistan, and Sri Lanka, respectively; and that of dyspnea was 7.4%, 7.2%, 11.3%, 3.4%, and 6%, respectively; and of chronic cough was 14.4%, 11%, 15.3%, 10%, and 5.9%, respectively [36]. The sampling strategy ensuring diversity of the population from heterogenous categories from seven different provinces and the use of objective measures for the diagnosis of COPD is a strength of the study. The sampling strategy used ensured that a representative sample from the urban, rural and plantation sectors were included; the urban and rural sectors were representative of different parts of the country. We acknowledge the fact that the clusters used in the sampling procedure may have resulted in some homogeneity within clusters, a major limitation of the study, especially where the sample for each province was selected from a single cluster; the sample would have been more representative of the province if participants were selected from more than one cluster. However, as the sample size included in the analysis was in excess of the minimum sample size, we believe that any effects on the results would have been mitigated. We diagnosed and classified COPD patients according to the internationally accepted GOLD guidelines that have also been adopted by the BOLD study. We paid much attention to quality control of spirometry in accordance with the BOLD protocol which has been widely used internationally. Another main limitation was the small sample size when compared to the other sites of the BOLD study. The study sample, though not a random one, was selected deliberately to provide adequate representation from different socio-demographic contexts in Sri Lanka. Rural and urban areas were deliberately oversampled to provide more reliable estimates as advised in the original BOLD study methodology [25]. We oversampled participants from the Colombo district to compensate for the small sample size from the Galle district; both Colombo and Galle districts are located along the southwestern coast of Sri Lanka, about 120 kms apart, having similar urban to semiurban characteristics and precipitation patterns. We surmise that these common characteristics make the two sampling sites similar in terms of environmental conditions and would have had little or no impact on our study findings including spirometry assessments. We did not calculate a cumulative index to estimate lifetime exposure to dust and smoking, as these exposures are likely to have changed over time; it is also likely that changes in exposure may have occurred due to symptoms of the disease that affected participants may have experienced. The results we report are based on data collected in 2013. Since then, no large-scale representative studies have been conducted in Sri Lanka to report changes over time; we acknowledge this as a limitation of the study. Conclusions The outcome of the BOLD study has important recommendations for the diagnosis and management of COPD in Sri Lanka. The finding that most of the participants with confirmed airflow limitation consistent with a COPD diagnosis have not been diagnosed earlier suggests that hidden burden of COPD in the community is large. It also emphasizes the need to make spirometry available at primary care settings. Primary care physicians should be updated on classification of COPD and use of spirometry for diagnosis of COPD. Targeted educational programmes for risk groups is likely to decrease the incidence of COPD. The findings will be useful in development of management protocols for people living with COPD. Supplementary Information Supplementary Material 1. Supplemental Figure 1: Participant recruitment. *Additional sampling was done in Colombo to compensate the small sample size of Galle district. Galle and Colombo districts are urban areas located along the southwestern coast of Sri Lanka sharing common environmental conditions. A sample from Badulla district was included to represent the plantation sector. Supplementary Material 1. Supplemental Figure 1: Participant recruitment. *Additional sampling was done in Colombo to compensate the small sample size of Galle district. Galle and Colombo districts are urban areas located along the southwestern coast of Sri Lanka sharing common environmental conditions. A sample from Badulla district was included to represent the plantation sector. Supplementary Material 2. Supplemental Figure 2: Flow chart of study design. Supplementary Material 2. Supplemental Figure 2: Flow chart of study design.