Portal de Periódicos da CAPES

Impact of a Chimney Wood Stove on Risk of Pneumonia in Children Aged Less Than 18 Months in Rural Guatemala: Results from a Randomized, Controlled Trial

2006; Lippincott Williams & Wilkins; Volume: 17; Issue: Suppl Linguagem: Inglês

10.1097/00001648-200611001-00079

1531-5487

Kirk R. Smith, Nigel Bruce, Martin Weber, Alan Hubbard, Alisa Jenny, Mukesh Dherani, Rudinio Acevedo, Byron Arana,

Energy and Environment Impacts

MS3-03 Background: Observational studies have reported a 2-fold risk of childhood ALRI associated with indoor air pollution (IAP) from household solid fuel but are at risk for confounding, nonstandardized outcome assessment and lack of direct exposure assessment. We report here results of first-ever RCT testing impact of reduced IAP exposure on childhood pneumonia. Setting: This study was conducted in 518 children aged <18 months taking part in the RESPIRE–Guatemala trial. Objectives: The objectives of this study were 1) to determine impact of an improved wood stove (with chimney) on ALRI incidence in children <18 months in comparison with traditional open fire; and 2) to quantify change in children's exposure to air pollution in relation to ALRI risk. Methods: After extensive pilot work, a study area was selected in the western Guatemalan highlands where randomized introduction of chimney stoves was conducted successfully (see associated abstract). Case finding was by weekly active surveillance (see associated abstract). These methods allow analysis of a range of outcomes from very sensitive but nonspecific (field worker assessed) through moderately specific (MD diagnosis) to highly specific defined according to criteria of respiratory syncytial virus (RSV: positive or negative); hypoxia measured by pulse oximetry; and chest x-rays (WHO-defined end point pneumonia). Child exposure was assessed by mean 48-hour CO measured 3 times monthly with detailed assessment of indoor PM and CO in a subsample for modeling child PM exposure. Risk estimates are obtained by generalized estimating equations to allow for nonindependence of repeat episodes. Findings: Fieldworkers (FWs) identified 735 cases with ALRI signs after 597.6 child-years of follow up (incidence of 1.23/child per year). There were 265 cases of MD-diagnosed ALRI (incidence 0.44 episodes/child per year). Preliminary intention-to-treat analysis found rate ratio for intervention versus control of 0.84 (confidence interval [CI]: 0.73–0.97; P = 0.02) for FW-defined ALRI and 0.85 (CI: 0.67–1.08) for MD-diagnosed ALRI. There was almost 40% reduction between groups over time for RSV negative cases with hypoxia (SpO2 2 standard deviations below mean for well children at study altitude), with an odds ratio (OR) of 0.62 (CI: 0.36–1.05; P = 0.07), a potentially important finding because these cases are 1) likely to be bacterial and 2) have more severe disease and hence the highest mortality. In contrast, there was no impact on hypoxic RSV-positive cases (OR: 1.01 CI: 0.60–1.73). The overall child exposure reduction based on modeling of 48-hour CO measurements was approximately 44%. Discussion: Despite modest exposure reduction, there were consistent reductions in ALRI between groups defined by FW and MD. If confirmed, the observed 40% reduction in severe RSV-negative cases strongly supports the case for prioritizing IAP control measures.