Effect of acute respiratory infections in infancy on pulmonary function test at 3 years of age: a prospective birth cohort study

Kumar, Prawin; Mukherjee, Aparna; Randev, Shivani; Medigeshi, Guruprasad R; Jat, Kana Ram; Kapil, Arti; Lodha, Rakesh; Kabra, Sushil Kumar

Parents were advised to visit the hospital whenever their child developed symptoms suggestive of ARI like fever, cold, cough, fast or difficulty in breathing, noisy breathing or chest retraction. Health status of enrolled children was also enquired telephonically at 2-4 weeks interval. If parents expressed difficulty in bringing the child to hospital, the study nurse visited their home and collected the required sample. LRTI was diagnosed during ARI in the presence of fast breathing for age as defined by WHO (respiratory rate:<2 months: ≥60/min; 2-12 months: ≥50/ min), and upper respiratory tract infections (URTIs) in the absence of fast breathing for age. 27 Wheezing during these ARI episodes were documented by a paediatrician during the hospital visit. They were treated symptomatically during these episodes, and an antibiotic (Amoxicillin) was only used in case of LRTI. Children were hospitalised if they had inadequate oral intake, not responding to supportive measures, stayed far-away from medical facility and in case of parental Open access Quantitative variables were summarised using mean and SD if normally distributed; for skewed distribution, median (IQR) were used. IPFT indices at 3 years of age were compared between children who had ARI episodes during infancy and those who did not, using Wilcoxon rank-sum test. The signed-rank test was used to compare the change in IPFT indices from baseline to 3 years of age in the two groups. For multivariate analysis, separate linear regression models were built using the indices that were significantly associated with ARI in bivariate analysis as the dependent variables. Gender, family history of atopy, presence of smoking in the household, weight and height at 3 years of age and the tidal volume at birth were added as covariates for each regression model. The difference was considered statistically significant at the p-value of <0.05.