Epidemiology of asthma and associated factors in an urban Pakistani population: adult asthma study-Karachi

This population based cross-sectional survey was conducted in Karachi, Pakistan, from May 2014 to August 2015. A megacity, and the largest urban center and economic hub of Pakistan, Karachi is estimated to have a population of over 20 million [12]. The city inhabits a mix of various regional and national ethno-linguistic groups while Urdu is the most commonly spoken language [13]. There are an estimated 702 slums in the city harboring 40–61% of the population [13]. The city is administratively divided into six districts; South, East, West, Central, Malir and Korangi; in addition, there are six military cantonments [12].

Multi-stage cluster sampling was used for selection of clusters (primary sampling units) and households (secondary sampling units) in Karachi. Using the sampling frame for clusters or enumeration blocks available from the Pakistan Bureau of Statistics (PBS) (2003), 75 clusters, out of 9400, were randomly selected. Line-listing was carried out within each cluster and 40 households per cluster (out of 250–300) were selected randomly. All eligible participants (adults aged ≥18 years and those living in Karachi in the same household for at least six months) from the selected households were recruited in the study.

The required sample size was calculated to be 1284 participants (using the Open Epi version 2.3) based on prevalence of asthma assessed through spirometry 2% to 3.6% [14], or self-reported 2% to 6.5% [15], keeping bound on the error 1.5% and 2% respectively, confidence level of 95%, a design effect of 2 and inflated by 10% to adjust for non-response and unacceptable quality of spirometry.

Interviews were conducted in the local language by trained field staff using a structured questionnaire that was developed by adapting already published questionnaires for identifying possible risk factors of asthma [16]. It included questions related to the socio-demographic and economic variables (age, gender, socio-economic and educational status); indoor air pollution and ventilation (exposure to environmental tobacco smoke, housing structure, ventilation in house, kitchen characteristics and cooking habits) and other risk factors for asthma (presence of carpet at home, use of air-conditioning or incense and mosquito coils, recent paint or new furniture being brought at home and presence of pets, birds or animals at home). Questions regarding respiratory symptoms and illnesses (cough, sputum, wheeze, shortness of breath and any pre-existing respiratory conditions) including asthma, family history of asthma and other respiratory diseases, smoking habits, and occupation were added from the American Thoracic Society (ATS-DLD-78A) questionnaire, which has been validated in Pakistan [17, 18]. Questionnaire was translated into Urdu and then back translated to English to check the accuracy of translation and pre-tested before use.

Lung function was assessed through spirometry by trained field staff using Vitalograph Alpha spirometer (Vitalograph New Alpha 6000; Vitalograph Ltd., Buckingham, England) according to the ATS guidelines [19, 20]. Participants were explained about the procedure and spirometry was performed in a sitting position with nose clip attached. Forced Vital Capacity (FVC), Forced Expiratory Volume in first second (FEV₁) and FEV₁/FVC were recorded in liters. Post-bronchodilator reversibility in FEV₁ was assessed by administering Salbutamol (200 μg) through a 500-mL spacer device and repeating the test after 15 min. Three manoeuvers were performed and acceptable readings were recorded for both pre and post-bronchodilator response. Participants reporting eye, heart, lung, chest or abdominal surgery in the past 6 months, heart attack in the past 3 months, those with pulse greater than 120 beats per minute, recent respiratory infections (tuberculosis or pneumonia) and those who were pregnant at the time of data collection were excluded from performing spirometry. Anthropometric measurements including height and weight were also taken.

Three distinct definitions for asthma were used; including ‘self-reported, physician-diagnosed asthma’ based on information from the questionnaire; ‘reversibility in FEV₁’, assessed on presence of post-bronchodilator reversibility ≥200 ml in FEV₁; ‘respiratory symptom(s) and reversibility in FEV₁’, based on post-bronchodilator reversibility ≥200 ml in FEV₁ and one or more respiratory symptoms, or self-reported asthma [10]. ‘Acute cough or phlegm’ was defined as cough or phlegm as much as 4 to 6 times a day in a week and/or first thing in morning and/or at all during the rest of the day or at night. ‘Chronic cough or phlegm’ was defined as cough or phlegm for at least 3 consecutive months a year, for at least 2 years. ‘Chronic wheeze’ was defined as whistling sounds from chest (with or without cold), for at least 2 years. Shortness of breath was defined according to the Medical Research Council breathlessness scale which represent a spectrum of respiratory disability based on severity ranging from grade 1 to grade 5 [21]. Socio-economic status was defined using the proxy indicator of average monthly household income, which included income of all members living in the same house as well as additional earnings based on any business or other investment. ‘Ever smoker’ was defined as smoking more than 20 pack of cigarettes in lifetime or more than one cigarette a day for one year. ‘Pack years of smoking’ was defined as the number of cigarettes smoked per day divided by 20 and multiplied by the number of years that the person smoked. The variable ‘ethnicity’ was defined and categorized on the basis of five commonly spoken languages in Pakistan where minor variants and less commonly spoken dialects were merged with the more commonly spoken ones. For ‘educational level’ those who never attended school or did not know how to read or write were considered as illiterate while those who had been to school were categorized as literate. For the variable ‘type of cluster’; planned areas included those with permanent housing structure, sufficient living place, access to safe water and adequate sanitation system, while unplanned areas were densely populated areas of substandard housing, characterized by poverty, unsanitary and inferior living conditions and social disorganization [22]. ‘Animal or birds inside the house’ included both pets as well as animals kept as livestock. Type of kitchen was categorized based on its location, either inside or outside the house. Type of household was categorized based on construction of the household; pakka house refers to brick dwelling or concrete, Kaccha-pakka house refers to a mix of mud or thatched and brick dwelling. ‘Exposure to environmental tobacco smoke’ was defined as being exposed to cigarette smoke anywhere inside the house. Body mass index was defined according to WHO criteria for Asian population and categorized as: underweighted, < 18.5 kg/m²; normal, 18.5–23 kg/m²; overweight and obese, ≥ 23 kg/m² [23]. ‘Current employment status’ was categorized as currently employed or self-employed and unemployed (included students, housewives, those currently not working anywhere or retired). The International Standard Classification of Occupations (ISCO) categories were grouped into three i.e. not working, high and low skilled blue collar workers (involved in manual work), high and low skilled white collar workers (involved in desk work) [24].

Data were entered twice and validated through Epi Data version 3.1 while analysis was done using SPSS version 19.0. Frequencies and proportions were calculated for categorical variables including the covariates listed above, as well as the respiratory symptoms and outcome variables based on the three definitions of asthma. Chi-square test was done to assess distribution of participants according to asthma outcome status. The age of the participant was categorized considering ten year interval from age 18 years and older; three categories were used where those aged 38 years or older were merged together. The variable pack-years of smoking was re-categorized to adjust for small cell count into three categories for multivariate analysis. Multi-collinearity was assessed using Eta, Phi and Cramer tests between covariates (cut off value > 0.5) and high collinearity was identified between gender and cooking status, as well as gender and occupational history. Interaction was checked between family history of asthma and history of allergies but it was not significant. The continuous variable of socio-economic status was categorized into tertiles and association was checked with outcome variables in univariate and multivariate models. Univariate logistic regression analysis was done to calculate unadjusted odds ratio of factors associated with two outcome variables; ‘reversibility in EFV₁’ and ‘respiratory symptoms and reversibility in EFV₁’. For the outcome defined as ‘self-reported asthma’, adjusted analysis could not be performed due to small proportion of asthmatics. The variables found to be significant in univariate models (p-value < 0.25) and those with biological plausibility were assessed further in the multivariate models. For covariates with more than two categories the cut-off of p-value < 0.25 was used for any one of the categories. Multivariate logistic regression was carried out to estimate the adjusted odds ratios for associated factors of asthma.

Ethical approval for the study was taken from Ethical Review Committee of Aga Khan University (2311-CHS-ERC-12). Prior to the interview, written informed consent was obtained from each respondent. The purpose and nature of the study was explained and spirometry procedure was demonstrated. Participants identified with any abnormality on spirometry were provided the report and were counselled and referred for further work-up and management.