A Drolia, P Dewan, P Gupta
A Drolia, P Dewan, P Gupta. Predicting the Severity of Bronchiolitis in a Resource-poor Setting. The Internet Journal of Pediatrics and Neonatology. 2008 Volume 11 Number 1.
Bronchiolitis is a common acute viral infection of infancy clinically characterized by cough, fever, tachypnoea, wheeze, hyperinflation, and chest retraction. 1 Though, a self-limiting condition requiring mainly supportive treatment, admission may be required for close monitoring, hydration, oxygen or bronchodilator therapy. In countries where facilities for high-dependency and intensive care are scarce, and/or are located far from the point of initial assessment, clinical examination can be of special benefit in predicting the severity of bronchiolitis.
Several studies have been conducted in the western hemisphere 23456 and a few in Africa 78 as well, but none in South East Asia, to determine the predictors of severity of bronchiolitis. Previous studies have used a combination of clinical criteria and intensive monitoring like oximetry and blood gas analysis to ascertain the disease severity. Therefore, we designed a study based solely on clinical parameters that could predict the severity of acute bronchiolitis in settings where facilities for radiography, pulse oximetry, blood gas analysis, and hospitalization are not routinely available.
Patients and Methods
This study was conducted prospectively, in the paediatric emergency department of Guru Teg Bahadur Hospital, a tertiary hospital located in Delhi over a period of one year. All children between 2-24 months age, admitted with a clinical diagnosis of bronchiolitis, were included in the study. Clinically, bronchiolitis was defined as the first episode of acute wheezing, tachypnoea, and concomitant signs of viral respiratory illness in a child below 2 years of age. 9 Tachypnoea was defined as per the WHO recommendations, i.e., respiratory rate ≥ 40/min in children ≥ 12 months and respiratory rate ≥ 50/min in children aged 2 months till 12 months. 10 All children included in the study were hospitalized. Children who had already received bronchodilators/steroids prior to admission in our hospital, and those with chronic lung disease were excluded.
A detailed history was obtained and physical examination performed for all children at admission. Respiratory rate and heart rate were measured at admission and again after 10 minutes, while the child was breathing room air without any recent disturbance/feeding/crying. All enrolled cases received standard treatment that included intravenous fluids and humidified oxygen. Participants were re-assessed clinically after 48-72 hours of hospitalization and classified as having moderate or severe bronchiolitis based solely on the improvement in the respiratory rate. Children were classified as having moderate disease if (i) the tachypnoea subsided within 48-72h of hospitalization, or, (ii) there was a decrease in respiratory rate by at least 25% within first 48-72 h of hospitalization. The rest were labeled as having severe bronchiolitis.
Viral isolation and serology
Secretions from the nasopharynx were taken on the day of admission with a mucus collector attached to wall suction. Specimens were stored at 4°C and processed for viral isolation within 4 hours. Viral cultures were done using rhesus monkey kidney, LLC-MK2, HEP-2, and HELA cell cultures. Fluorescent antibody technique (FAT) was used for virus identification in tissue cultures. Serology for respiratory syncytial virus (RSV), parainfluenza virus, influenza A and B and adenovirus was estimated. Serological testing included haemagglutination inhibition for influenza A and B and microneutralization for respiratory syncytial virus (RSV), parainfluenza virus, influenza A and B and adenovirus. A fourfold rise in antibody titers was considered as evidence of infection.
Ethical clearance and consent
Prior approval by the Institutional ethics’ committee and informed written consent from the primary caretakers of all children enrolled in the study were obtained.
The two groups were compared with respect to their historical and physical characteristics, at enrolment. Chi-square was used for categorical variables and t-test for quantitative variables. Univariate analysis was performed to identify the factors (personal, demographic, environmental, or clinical) predicting the severity of bronchiolitis. A multivariate logistic regression model was used to identify the most important factor determining the severity of bronchiolitis (dependent variable). Statistical analysis was done using SPSS software.
Bronchiolitis constituted 1.7% (166 children out of 9500) of all hospital admissions in the pediatrics department. Of these, 100 children (mean age 5.1 ± 4.3 months) fulfilled the inclusion criteria and consented for the study. Viral etiology could be established in 67 cases; Influenza A (H3N2): 35, respiratory syncytial virus: 25, parainfluenza type-1: 10, and adenovirus in 1 child. Four cases had dual infection with influenza A and parainfluenza type-1. Table 1 compares the personal, socio-demographic, environmental, and, clinical parameters in the two groups.
On a univariate analysis, respiratory rate at admission, use of accessory muscles of respiration, presence of diffuse wheeze, crepitations in chest, and grunting, were found to be associated with severe bronchiolitis. Amongst them, respiratory rate at admission was found to be the most important predictor of severe bronchiolitis on logistic regression model using F test at a 0.05 level of significance. A discriminant analysis was then performed and a cut off respiratory rate of ≥ 68/min at presentation was found to have the maximum risk for severe bronchiolitis (sensitivity: 78.3%, specificity: 75.9%, positive predictive value: 77%, negative predictive value: 80.4%). Presence of accessory muscle use had a sensitivity of 85.7% and specificity of 49% in predicting severe bronchiolitis. All the five factors when combined could predict 80% of cases of severe bronchiolitis correctly, whereas respiratory rate ≥68/min at admission could correctly diagnose severe bronchiolitis in 77% of cases.
We found respiratory rate ≥ 68/min and the use of accessory muscles of respiration at presentation to be the most important predictors for severe bronchiolitis. Other clinical criteria found helpful in predicting severe bronchiolitis included the presence of diffuse wheeze, presence of grunting, crepitations in chest, and ill appearance of the infant. Age, exclusive breast-feeding, family history of atopy/asthma, overcrowding, improper drainage and ventilation, exposure to effluents like kerosene, and the presence of animals in home did not affect the severity of bronchiolitis.
As seen in our study, tachypnoea has been previously found to be the best indicator of severity of acute bronchiolitis. 1213 Voets
We did not include pulse oximetry or arterial blood gas estimations for correlating with the severity of bronchiolitis, for the want of advanced diagnostic and intensive care facilities at peripheral health care centers in developing countries. Though hypoxemia has previously been found to be associated with occurence of apnea, duration of viral shedding and tachypnoea in bronchiolitis, 17 it may be an unreliable parameter to ascertain severity of bronchiolitis when used alone. 1819 Mallory
Though, a balanced approach involving a rational combination of clinical signs, laboratory, and radiographic investigations, may provide the best estimate of the severity of bronchiolitis, but tachypnoea coupled with the use of accessory muscles can serve as rapid, simple, and reliable parameters for predicting the severity of bronchiolitis, in a resource-poor setting.