Factors associated with asthma among under-fives in Mulago hospital, Kampala Uganda: A cross sectional study

Background: Asthma is the most common chronic childhood illness, with rapidly increasing prevalence in low-income countries. Among young children, asthma is often under-diagnosed.We investigated the factors associated with asthma among under-fives presenting with acute respiratory symptoms at Mulago hospital, Uganda.Methods: A hospital-based cross sectional study of 614 children with cough and/or difficult breathing, and fast breathing, was conducted between August 2011 and June 2012. A questionnaire focusing on clinical history of the child was administered to the caretakers. A physical examination and, laboratory and radiological investigations were done. Asthma was defined according to GINA (Global Initiative for Asthma) guidelines which were modified by excluding the symptom of ” chest tightness”, spirometry/peak expiratory flow measurements and by adding chest x-ray findings to distinguish asthma from pneumonia. A panel of three paediatricians reviewed the participants’ case reports and, guided by the study definitions, made a diagnosis of asthma or other. Multivariable logistic regression analysis was done to determine factors independently associated with asthma.Results: Of the 614 children, 128 (20.8%) had asthma, 125 (20.4%) bronchiolitis, 167 (27.2%) bacterial pneumonia only, 163 (26.5%) viral pneumonia while 31 (5.1%) had other diagnoses including pulmonary tuberculosis. The majority (71.1%) of children with asthma were aged ≥ 12 months. Factors associated with asthma included maternal asthma (AOR 2.4, 95% CI 1.2, 4.6), a history of allergy in the patient (AOR 2.6, 95% CI 1.2, 5.4,), use of gas for cooking (AOR 3.8, 95% CI 1.2, 13.3), prematurity (AOR 9.3, 95% CI 1.2, 83.3) and high level of education of caretaker (AOR 9.1, 95% CI 1.1, 72.8).Conclusion: Maternal asthma, a history of allergy in the patient, use of gas for cooking, prematurity and high level of education of caretaker were significantly associated with asthma. There is need for studies to explore the role of the above factors in development and exacerbation of childhood asthma to provide information that can be used to design strategies for asthma prevention and control. © 2013 Nantanda et al.; licensee BioMed Central Ltd. We conducted a cross-sectional study among children aged 2 to 59 months presenting at the emergency paediatric unit of Mulago hospital Kampala between August 2011 and July 2012. Mulago hospital is a national referral hospital. It also acts as a district hospital serving an urban and peri-urban catchment population of about two million people. The paediatric emergency unit attends to children aged 1 day to 12 years. The average daily attendance is 80 children, 75% of whom are aged 2 to 59 months. An estimated 25% of the children present with cough and/or difficulty in breathing. The hospital was selected as the study site because of its ability to handle laboratory and radiological investigations for diagnosis of asthma and pneumonia, facilities that are not readily available in rural Ugandan hospitals. The study was approved by the Higher Degrees, Ethics and Research Committee (HDREC) at Makerere University College of Health Sciences and the Uganda National Council of Science and Technology. Informed written consent was obtained from the caretakers of the participants. The process of questionnaire development was led by one of us (MSO) in Denmark. Through a literature search, research team debates and expert opinion, study definitions and concepts for asthma were discussed. A qualitative study, focusing on caretakers of children with asthma, was undertaken to identify relevant items to be included in the questionnaire. The interviews were aimed at understanding the presentation and progression of the disease from the caretaker’s perspective. A preliminary questionnaire was hence developed consisting of items focusing on acute symptoms, past medical history of the child and medicine use during previous illnesses and, family history of asthma. The items were then tested for language, understanding and relevance through focus group discussions with caretakers. A new questionnaire was developed and this was further tested twice in the same way and a final version was generated, which was then tested for internal and external validity. It was then translated into English. The questionnaire was then adapted for use in Uganda by one of us (RN). It was translated to Luganda, the language commonly used in central Uganda, where Mulago hospital is located. It was then back-translated into English. Both the Luganda and English versions were pre-tested on a sample of 35 mothers to check for understanding of the questionnaire items and time taken to administer the questionnaire. Any necessary changes were made and a final questionnaire was developed. We enrolled children aged 2 to 59 months who presented at the paediatric emergency unit of Mulago hospital with cough and/or difficulty in breathing plus fast breathing, and whose caretakers gave informed written consent. The definition of fast breathing was based on WHO criteria [16] Children with heart conditions, or cardiac failure secondary to severe anaemia, based on the caretaker’s history, physical examination findings and medical records, were excluded. All potential participants were triaged and those with ‘severe classification’ according to the WHO guidelines [17] were given urgent care before proceeding with the consent process. Children with wheezing were nebulised with salbutamol solution using an ultrasonic nebulizer according to the hospital protocol [18] and the response noted. Participants were enrolled from 8.00.am to 10.00.pm on weekdays. After enrolment, a questionnaire (Additional file 1) was administered by the nurse. A physical examination was performed by the doctor. For all participants, we measured the peripheral oxygen saturation (SaO2) in room air. Children with SaO2 less than 92% were given oxygen by mask or nasal prongs. Six mill-litres of venous blood were drawn from the cubital vein or dorsum of the hand using a BD™ blood collection set in three aliquots; for blood culture, white cell count, and serum C-reactive protein (CRP) titres. A peripheral blood smear for malaria parasites was also done. A specimen of nasopharyngeal epithelium was collected for identification of Respiratory Syncytial Virus (RSV) according to the manufacturer’s instructions (BD Diagnostics, Becton, Dickinson and Company, Maryland USA). All specimens were delivered to the laboratory within six hours of collection. A posterior-anterior chest x-ray was taken for each of the study participants within 48 hours of enrolment. Total and differential white blood cell counts were determined using Coulter counter method (Beckman Coulter Inc. Z™ series). CRP titres were analyzed using CRP (Human) ELISA kit-ABNOVA™, Taiwan, according to the manufacturer’s instructions. Blood culture was done using the Bactec method (Becton, Dickson and Company Maryland USA) and positive samples were further analyzed for the bacterial species using a Gram stain. Drug susceptibility tests were done using the Disc diffusion method [19]. For malaria diagnosis, a peripheral blood smear was prepared using Leishman’s stain. Identification of Respiratory Synctial Virus (RSV) from nasal epithelium was done using Direct Fluorescence Antibody (DFA) technique (Light Diagnostics™ USA). The x-rays were interpreted by two independent radiologists who were blinded to the clinical and laboratory findings of the participants. Radiographic end-points included consolidation, collapse, alveolar and interstitial infiltrates, pleural effusion, hyper-inflation and normal. X-rays with discordant results from the two radiologists were interpreted by a third reader and the result taken as final if there was concordance between the third reader and any of the primary readers. There are no diagnostic gold standards for asthma among under-fives. In this study, we modified the GINA (Global Initiative for Asthma) guidelines for diagnosis of asthma [20] as follows: In the history; we excluded “recurrent chest tightness” as a symptom because it is not easily expressed by children less than five years [21,22]. We also we excluded peak expiratory flow measurements because children less than five years are not able to perform this test effectively [23]. Furthermore, we included chest x-rays to help distinguish asthma from pneumonia. Pneumonia is common in Uganda and, in under-fives, has a presentation similar to that of acute asthma [13,24]. The case definition of bronchiolitis was based on South African guidelines [14] for diagnosis, management and prevention of acute viral bronchiolitis. The details of the study definitions are provided in Table 1. Case definitions for asthma, bronchiolitis and pneumonia A panel of experts comprising paediatricians with experience in pulmonology and infectious diseases reviewed the participants’ case records. The experts had no access to the participants; hence the diagnoses were made post hoc. Each expert studied the case record of the participant, and guided by the study definitions, made a diagnosis, which was then discussed by all the panellists. A diagnosis of asthma or of some other condition such as pneumonia was made following agreement of all or two of the panellists. Where there was discordance between all the three panellists, the case records were subjected to a further discussion until a diagnosis was agreed upon. One of us (RN) took the minutes during the proceedings but did not participate in the discussions. To describe factors associated with asthma among young children with cough and/or difficult breathing, a minimum sample size of 308 was calculated. We assumed two-sided significance level of 95%, power of 80%, proportion of children with asthma and who had a family history of asthma to be 52%, and Odds’ ratio of 2.5, based on a study of asthma in preschoolers by Haby and colleagues [7]. However, this was part of a larger study involving 614 children and all were included in the analysis. Data was double-entered into Epidata version 3.0 and exported to Stata version 12.0 (Stata Corp, College station Texas, USA) for analysis. To determine factors independently associated with asthma, multivariable analysis was done. A logistic regression model was built by including all factors with a p value less than 0.2 at bivariate analysis. Adjusted Odd’s ratios were computed to adjust for confounding. Multi-colinearlity and interaction of the predictor variables was checked until we obtained the best fitting model. Cohen’s kappa was used to measure the degree of agreement between the primary radiologists. A p value of ≤0.05 was considered statistically significant. We also performed logistic regression analysis for factors associated with bronchiolitis and compared them to those associated with asthma. Results are summarized as frequencies, proportions, figures and tables as appropriate. In this study, children from the capital city, municipalities and town councils in Uganda were collectively referred to as coming from “urban setting” and the rest from “rural setting”. This was adapted from the Uganda Demographic and Health Survey [24].