Chronic airflow obstruction in Tanzania – a cross-sectional study

Chronic airflow obstruction (CAO), caused by a chronic inflammation of airways in response to exposure to dusts and fumes is a characteristic feature of chronic obstructive pulmonary disease (COPD). CAO leads to a decline in lung function and to subsequent loss of exercise capacity. Cigarette smoking is the major causative factor for CAO in industrialized countries [1], leading to an overall population prevalence between 5% and 25% in Europe, North America, Latin America and Asia [2‐5].

COPD is a disease mainly affecting higher age groups and most studies have so far addressed the population above the age of 40 years. Little is known however on prevalence and risk factors of COPD in developing countries with younger populations, and particularly in sub-Saharan Africa. Recently, the BOLD study investigated prevalences of CAO on a global level and provided data from Latin America and Asia as well as data from two study sites in Africa [2, 3, 6, 7]. Previous studies in sub-Saharan Africa have mostly focussed on risk groups with occupational exposure such as miners [8‐11].

Epidemiology and risk factors of COPD in developing countries likely differ from those in the developed world. While cigarette smoking is less common in developing countries, factors like prolonged exposure to biomass fuels from cooking starting from early childhood, cooking indoors in rooms shared for living and sleeping purposes, high TB prevalence and different age structure have relevant influence on COPD epidemiology [12]. At present, it is unclear which effects this life-long exposure to biomass fuels in Africa has for the development of chronic obstructive pulmonary disease. Moreover, there is little data on chronic obstructive pulmonary disease in young African populations.

To gather data on CAO in Tanzania, we examined patients attending primary healthcare facilities as well as persons accompanying patients using a questionnaire and performed pre- and post-bronchodilator spirometry. Tanzanian Lung Health Study (TLHS) aims to provide first estimates on CAO in Tanzania and to explore risk factors as well as medical and social aspects of CAO, providing baseline data for further studies on a larger scale. Moreover, TLHS aims to give additional information on analysis of spirometric outcomes other than the standard definitions of CAO and their diagnostic value to detect earlier stages of small airways disease in an African setting.

A cross-sectional facility-based study was conducted among attendants, accompanying persons and visitors to the rural Bagamoyo District Hospital (Bagamoyo, Pwani, Tanzania), and the urban Mwananyamala Regional Hospital (Dar es Salaam, Tanzania) between October 2015 and September 2016. Both study centres are located in coastal Tanzania without mining industry or exposure to other region-specific lung hazards. The primary outcome was spirometric evidence of CAO. Secondary outcomes were risk factors and parameters indicating social impact. Participants were eligible for inclusion if they were at least 18 years of age and either attendants of a primary healthcare clinic for any reason other than an acute respiratory problem, persons accompanying patients or visitors to the hospital visiting inpatients. Participants visiting the hospital for other reasons, having medical contraindications against bronchodilator drugs or having any other physical or mental condition limiting the ability to obtain a valid spirometry result were excluded (hospitalisation within <3 months, newly diagnosed heart disease within <3 months, uncontrolled hypertension, resting heart rate > 100/min., uncontrolled diabetes, pregnancy, acute infections or fever, acute pain, smoking during the last 2 h, thoracic or abdominal surgery within <3 months or any other severe physical or mental disease limiting the ability to comply with instructions for spirometry). To allow analysis of outcomes in patients across all relevant age groups, age-stratified sampling was used. Within age-strata, participants were invited randomly using a predefined selection procedure.

A sample size of 600 participants was intended to provide an acceptable prevalence estimate with 2.5% precision for this pilot study assuming a prevalence of 10% of chronic airflow obstruction.

After obtaining informed consent either in writing or by an impartial witness, participants were interviewed by a trained physician or a study nurse to obtain data on lung health and TB history and trained in performing spirometry supported by a video.

After spirometry was performed, patients inhaled 4 times 100 μg of salbutamol under supervision of a nurse using a spacer. Spirometry was repeated after 15 min. Quality criteria for spirometry were applied according to the 2005 ATS/ERS guidelines [13]. Spirometer readings were checked for acceptability and repeatability criteria by the physician performing the test and checked again by a senior pulmonologist. Only spirometer readings meeting all acceptability as well as repeatability criteria (minimum grade A and B) were finally accepted. Spirometry was performed using the NDD EasyONE World Spirometer (Zürich, Switzerland). Spirometric prediction equations of the Global Lung Function Initiative for African Americans were used [14].

Spirometric data was analysed for prevalence of CAO applying the FEV1/FVC < 5th percentile predicted (LLN) method (ATS/ERS spirometric criteria) as well as the GOLD definition with a fixed FEV1/FVC ratio < 0.7. A positive bronchodilator response was defined as an increase of FEV1 and/or FVC >12% and at least 200 ml compared to baseline. To further assess airflow obstruction including earlier stages, the post-bronchodilator forced expiratory flow at 25–75% of forced vital capacity (FEF_25–75) [15] and the FEV1% predicted were used. As pulmonary restriction influences FEF_25–75 and FEV1% predicted, participants with evidence of restriction (FVC < 5th percentile of FVC predicted and FEV1/FVC >0.85, as defined by ATS/ERS criteria [16]) were excluded from this analysis.

Data was analysed using JMP ver. 14. Median FEF_25–75 and FEV1% predicted values were compared among participants exposed vs. unexposed to risk factors for CAO. Parameters were compared between risk groups using the Mann-Whitney-U-test to compare two and the Kruskal-Wallis-Test to compare multiple groups. Χ ₂ test was used to compare categorical outcomes between groups. Differences were considered significant when p was <0.05.

Chronic airflow obstruction is an important cause of morbidity and mortality, and has major impact on the lives of those affected. Data from this pilot study in coastal Tanzania indicate a relevant level of CAO among urban and rural inhabitants. CAO can be found across all age groups and those affected may suffer from limitations in their ability to work and in quality of life.

This study showed that approximately one out of twenty inhabitants of coastal Tanzania has CAO according to both definitions currently used. Whereas the ATS/ERS definition using the LLN method showed only a moderate increase of prevalence with age, the GOLD definition showed a six-fold higher prevalence of CAO in the highest vs. the youngest age group. As the GOLD definition uses a fixed FEV1/FVC ratio of 0.7 as cut-off to define CAO, older age groups are overrepresented as the FEV1/FVC ratio declines physiologically with increasing age [17]. There is a persisting controversy on which classification to use; the LLN method indicating CAO is better suited for research purposes as it controls for ageing of the lung whereas the GOLD spirometric criteria are preferred by many clinicians due to their simplicity. The results from this study however underscore the concerns regarding over-diagnosis of elderly patients using the GOLD definition as well as potential under-diagnosis in younger patients, and demonstrate this effect in an African setting. Future epidemiological studies in Africa should therefore report results using both definitions until a universally accepted method has been developed. Moreover, results should include prevalences stratified by age groups, as exposure to biomass fuel smoke starting from birth or even in utero [12] may lead to an earlier onset and higher prevalence of CAO in young African populations.

Few studies have reported prevalence of CAO in general populations from sub-Saharan Africa not exposed to special risk factors (e.g. gold miners). The reports available so far do not give consistent results across African regions; the prevalences found in this study were lower than in a recent study from Nigeria (5.1%–10% (LLN)) [6], Malawi (14% (GOLD)) [18] or from Uganda (12.4% (LLN), 16.2% (GOLD)) [19], but higher than in Cameroon (2.4% (LLN), 0.5% (GOLD)) [20]. Consistent across all studies, the majority of patients had either mild or moderate severity of obstruction. There is still no sufficient data from large population-level spirometry-based and quality-assured studies in sub-Saharan Africa to give a good CAO prevalence estimate in general or specific to African regions; the data presented in this study add relevant baseline data for planning studies on a larger scale and underscore the importance of reporting data in different age groups.

Risk factors for CAO likely differ from other regions of the world due to different living circumstances, living standards and lifestyle of populations. Smoking [1] and exposure to particulate matter [21] in the air are the predominant risk factors for CAO in industrialized countries. Data from China, Latin America and Turkey show a moderately increased risk of CAO after exposure to smoke from biomass fuels, with a stronger effect for firewood than for charcoal [22]. Following these principles, indoor air pollution is generally assumed to be a major risk factor in Sub-Saharan Africa [23]. Four recent studies however did not demonstrate a clear association between indoor use of biomass fuel with CAO [6, 7, 19, 20]. Chronic airflow obstruction may begin early in life and investigation of first signs may be missed by applying the ATS/ERS or GOLD criteria only – the predominant method of analysis used in studies from Africa. Large population based epidemiological studies have however demonstrated excellent sensitivity for early detection of airway pathology using continuous analysis of FEV1 and in particular FEF_25–75 [21, 24, 25] as well as FEV1 as a percentage of FVC [21]. In an exploratory approach, we therefore used FEF_25–75 as well as FEV1% predicted to detect more subtle changes in airway pathology. Analysis of FEF_25–75 seemed to have a higher sensitivity and better correlation with severity of symptoms and physical limitations than FEV1% predicted. Both parameters indicated differences among participants exposed and unexposed to risk factors for CAO typical for an African setting and indicated their potential significance for causing social impact. We therefore recommend that in larger studies on CAO in Africa, results should be reported both using the standard CAO definitions, but also using analysis of continuous spirometric values such as FEF_25–75 and FEV1% predicted.

Recent meta-analyses and studies on a global scale were able to demonstrate that active tuberculosis as an infection causing chronic pulmonary inflammation over a prolonged period of time is a major factor contributing to CAO [26‐28]. In this study, a history of active TB infection was the most relevant risk factor associated with a lower FEF_25–75 as well as a lower FEV1% predicted. The observed effect of TB infection is consistent with other recently published studies from local African populations [6, 18, 29] and underscores the significance of TB for development of chronic lung disease in Africa.

This study had limitations. In order to examine CAO in a sufficient number of subjects in each age class in this study, age-stratified sampling was used. The population sample examined consisted of a mixed population composed of hospital visitors and patients. Therefore, particularly the prevalence of CAO should not necessarily be regarded as representative for the general population. An unexpectedly high number of study participants was unable to produce a spirometry meeting ATS/ERS quality criteria despite intensive and standardized pre-spirometry training including a training video two times. In previous studies in developing country settings in Latin America and Africa [6, 19, 30], observed spirometry failure rates were lower except for one study from Malawi with a similar rate of exclusions [7]. Local cultural and religious factors in this community causing people not to expose themselves in a medical examination requiring their participation with maximum muscular force used, together with a generally low level of education probably have contributed to this high failure rate. Main findings in this study, particularly with regard to risk factors for CAO need to be confirmed in larger population-based studies.

In conclusion, this study demonstrated a relevant burden of CAO in a Tanzanian population and showed major differences in prevalence as well as age-specific prevalence between the two standard definitions of CAO used. Although most individuals with evidence of CAO had only mild or moderate obstruction, CAO likely has relevant social and social and economic impact for those affected. TB may have more impact on development of chronic airway obstruction than smoking in Africa. Analysis of FEV1 and FEF_25–75 in epidemiologic studies on CAO may give important information on risk factors and impact in addition to using standard definitions for CAO alone.