Household ventilation and tuberculosis transmission in Kampala, Uganda

To test the feasibility of measuring household ventilation and evaluate whether ventilation is associated with tuberculosis (TB) in household contacts in Kampala, Uganda.

Adults with pulmonary TB and their household contacts received home visits to ascertain social and structural household characteristics. Ventilation was measured in air changes per hour (ACH) in each room by raising carbon dioxide (CO₂) levels using dry ice, removing the dry ice, and measuring changes in the natural log of CO₂ (lnCO2) over time. Ventilation was compared in homes with and without co-prevalent TB.

Members of 61 of 66 (92%) households approached were enrolled. Households averaged 5.4 residents/home, with a median of one room/home. Twelve homes (20%) reported co-prevalent TB in household contacts. Median ventilation for all rooms was 14 ACH (interquartile range [IQR] 10-18). Median ventilation was 12 vs. 15 ACH in index cases' sleeping rooms in households with vs. those without co-prevalent TB (P = 0.12). Among smear-positive indexes not infected by the human immunodeficiency virus (HIV), median ventilation was 11 vs. 17 ACH in index cases' sleeping rooms in homes with vs. those without co-prevalent TB (P = 0.1).

Our findings provide evidence that a simple CO₂ decay method used to measure ventilation in clinical settings can be adapted to homes, adding a novel tool and a neglected variable, ventilation, to the study of household TB transmission.