Air pollution exposure when cooking with electricity compared to gas

We report small-sample evidence from a randomized experiment among a set of urban Ecuadorian households who owned both electric induction and gas stoves. We randomly assigned households to cook only with one stove during a prescribed two-day monitoring period, and then cook only with the other stove in a subsequent two-day period. The order of stove use was randomized, and air pollution was measured during each period. We found that mean 48-hour personal NO2 exposure was 9.9 ppb higher (95% CI, 4.5-15.3) -- a 50% increase over the 48-hour induction mean -- when households were randomized to gas as compared to induction. Mean kitchen area NO2 concentrations were 1 ppb higher (95% CI, 0.4-2.1) (a 6% increase) and mean personal PM2.5 exposure was 11 gm-3 higher (95% CI, -0.1-22.8) (a 44% increase) during study periods when randomized to gas. We use time-resolved cooking and pollution data to illustrate that these differences are driven by LPG cooking, which was associated with a 5.0 ppb increase in 5-minute average NO2 kitchen area concentrations (95% CI, 3.4-6.7) and a 20.8 gm-3 increase in 5-minute average personal PM2.5 exposure (95% CI 8.9-32.6). In contrast, cooking with induction was not associated with changes to short-term NO2 kitchen area concentrations, though it was associated with short-term increased personal PM2.5 exposure (10.8, 95% CI, 5.7-15.9).