Indoor Air Pollution and Childhood Pneumonia: Impact of Reduced Exposure on Severity and Clinical Features

Abstract

ISEE-583 Objective: In a RCT of reduced indoor air pollution (IAP) and child acute lower respiratory illness (ALRI)—RESPIRE-Guatemala—we previously reported the intervention, a chimney wood-stove, was associated with 10% and 40% (P<0.05) reductions in fieldworker (FW) assessed ALRI and severe-ALRI respectively, and 15% and 40% (P=0.05) reductions in all physician-diagnosed pneumonia and severe (hypoxic) non-viral pneumonia, respectively. We use detailed clinical information to investigate 2 issues: the impact on clinical features of pneumonia, including severity the validity of FW classified severe ALRI. Materials and Methods: 518 households with a pregnant women or child <4 months were randomized to intervention or continuing with the traditional open fire, and followed until the child was 18 months old. ALRI cases were ascertained by (i) weekly household visits by trained FW, then (ii) referral to a study physician working in a local community center who carried out clinical examination and pulse oximetry, and (on pneumonia cases) a RSV test and chest x-ray. Results: Among all physician-diagnosed pneumonias, fewer intervention cases had crepitations (P=0.03), a finding seen mainly when less severe and RSV+ve. These cases also had longer duration of illness, and of cough and wheeze. However, severe (hypoxic) cases showed no differences for clinical features, median oxygen saturation (SaO2), or CXR findings. Among FW-assessed ALRI, compared with non-severe cases, severe cases exhibited significantly higher prevalence of key clinical physician-assessed signs of severity, including chest wall indrawing, nasal flaring, central cyanosis, accessory muscle use, and crepitations, and lower median SaO2. Conclusions: The intervention reduced incidence (primary prevention), but did not modify clinical features (secondary prevention), of severe physician-diagnosed cases. There was some indication of reduced severity among non-hypoxic cases, particularly RSV+ve. The reduced incidence of FW-assessed severe cases provides corroborative evidence, due to the strong consistency with physician's diagnostic criteria.