Genotyping indicates marked heterogeneity of tuberculosis transmission in the United States, 2009-2018

Abstract

Heterogeneity in the number of secondary tuberculosis (TB) cases per source case, the effective reproductive number, R, is important in modelling prevention strategies' impact on incidence. We estimated mean R (Rm) and calculate the dispersion parameter of this distribution, k, using surveillance and genotyping data for U.S. cases during 2009-2018. We modelled transmission assuming cases in a cluster have matching genotypes and share characteristics related to geography, temporal proximity (i.e. serial interval) and time since U.S. arrival among non-U.S.-born persons. Complete data were available for 55 330/85 958 cases. Varying the serial interval and geographic proximity used to derive clusters, we consistently estimated Rm<1.0 and k < 0.08; the low value of k indicates a small number of source cases produce a disproportionate number of secondary cases. U.S. TB reproductive number has a highly skewed distribution, indicating a minority of source cases disproportionately contribute to transmission.