Low correlation between household carbon monoxide and particulate matter concentrations from biomass-related pollution in three resource-poor settings

Household air pollution from the burning of biomass fuels is recognized as the third greatest contributor to the global burden of disease. Incomplete combustion of biomass fuels releases a complex mixture of carbon monoxide (CO), particulate matter (PM) and other toxins into the household environment. Some investigators have used indoor CO concentrations as a reliable surrogate of indoor PM concentrations; however, the assumption that indoor CO concentration is a reasonable proxy of indoor PM concentration has been a subject of controversy. We sought to describe the relationship between indoor PM_2.5 and CO concentrations in 128 households across three resource-poor settings in Peru, Nepal, and Kenya. We simultaneously collected minute-to-minute PM_2.5 and CO concentrations within a meter of the open-fire stove for approximately 24h using the EasyLog-USB-CO data logger (Lascar Electronics, Erie, PA) and the personal DataRAM-1000AN (Thermo Fisher Scientific Inc., Waltham, MA), respectively. We also collected information regarding household construction characteristics, and cooking practices of the primary cook. Average 24h indoor PM_2.5 and CO concentrations ranged between 615 and 1440μg/m³, and between 9.1 and 35.1ppm, respectively. Minute-to-minute indoor PM_2.5 concentrations were in a safe range (<25μg/m³) between 17% and 65% of the time, and exceeded 1000μg/m³ between 8% and 21% of the time, whereas indoor CO concentrations were in a safe range (<7ppm) between 46% and 79% of the time and exceeded 50ppm between 4%, and 20% of the time. Overall correlations between indoor PM_2.5 and CO concentrations were low to moderate (Spearmanρ between 0.59 and 0.83). There was also poor agreement and evidence of proportional bias between observed indoor PM_2.5 concentrations vs. those estimated based on indoor CO concentrations, with greater discordance at lower concentrations. Our analysis does not support the notion that indoor CO concentration is a surrogate marker for indoor PM_2.5 concentration across all settings. Both are important markers of household air pollution with different health and environmental implications and should therefore be independently measured.