Abstract
BACKGROUND: Transmission of multidrug-resistant tuberculosis (MDRTB) requires spatial proximity between infectious cases and susceptible persons. We assess activity space overlap among MDRTB cases and community controls to identify potential areas of transmission. METHODS: We enrolled 35 MDRTB cases and 64 TB-free community controls in Lima, Peru. Cases were whole genome sequenced and strain clustering was used as a proxy for transmission. GPS data were gathered from participants over seven days. Kernel density estimation methods were used to construct activity spaces from GPS locations and the utilization distribution overlap index (UDOI) was used to quantify activity space overlap. RESULTS: Activity spaces of controls (median = 35.6 km(2), IQR = 25.1-54) were larger than cases (median = 21.3 km(2), IQR = 17.9-48.6) (P = 0.02). Activity space overlap was greatest among genetically clustered cases (mean UDOI = 0.63, sd = 0.67) and lowest between cases and controls (mean UDOI = 0.13, sd = 0.28). UDOI was positively associated with genetic similarity of MDRTB strains between case pairs (P < 0.001). The odds of two cases being genetically clustered increased by 22% per 0.10 increase in UDOI (OR = 1.22, CI = 1.09-1.36, P < 0.001). CONCLUSIONS: Activity space overlap is associated with MDRTB clustering. MDRTB transmission may be occurring in small, overlapping activity spaces in community settings. GPS studies may be useful in identifying new areas of MDRTB transmission.