Biomass smoke in Burkina Faso: what is the relationship between particulate matter, carbon monoxide, and kitchen characteristics?

In Burkina Faso where cooking with biomass is very common, little information exists regarding kitchen characteristics and their impact on air pollutant levels. The measurement of air pollutants such as respirable particulate matter (PM10), an important component of biomass smoke that has been linked to adverse health outcomes, can also pose challenges in terms of cost and the type of equipment needed. Carbon monoxide could potentially be a more economical and simpler measure of air pollution. The focus of this study was to first assess the association of kitchen characteristics with measured PM10 and CO levels and second, the relationship of PM10 with CO concentrations, across these different kitchen characteristics in households in Nouna, Burkina Faso. Twenty-four-hour concentrations of PM10 (area) were measured with portable monitors and CO (area and personal) estimated using color dosimeter tubes. Data on kitchen characteristics were collected through surveys. Most households used both wood and charcoal burned in three-stone and charcoal stoves. Mean outdoor kitchen PM10 levels were relatively high (774 mu g/m(3), 95% CI 329-1,218 mu g/m(3)), but lower than indoor concentrations (Satterthwaite t value, -6.14; p < 0.0001). In multivariable analyses, outdoor kitchens were negatively associated with PM10 (OR=0.06, 95 % CI 0.02-0.16, p value < 0.0001) and CO (OR=0.03, 95 % CI 0.01-0.11, p value < 0.0001) concentrations. Strong area PM10 and area CO correlations were found with indoor kitchens (Spearman's r = 0.82, p < 0.0001), indoor stove use (Spearman's r = 0.82, p < 0.0001), and the presence of a smoker in the household (Spearman's r = 0.83, p < 0.0001). Weak correlations between area PM10 and personal CO levels were observed with three-stone (Spearman's r = 0.23, p = 0.008) and improved stoves (Spearman's r = 0.34, p = 0.003). This indicates that the extensive use of biomass fuels and multiple stove types for cooking still produce relatively high levels of exposure, even outdoors, suggesting that both fuel subsidies and stove improvement programs are likely necessary to address this problem. These findings also indicate that area CO color dosimeter tubes could be a useful measure of area PM10 concentrations when levels are influenced by strong emission sources or when used in indoors. The weaker correlation observed between area PM10 and personal CO levels suggests that area exposures are not as useful as proxies for personal exposures, which can vary widely from those recorded by stationary monitors.