Invasive Pneumococcal Disease Clusters Disproportionally Impact Persons Experiencing Homelessness, Injecting Drug Users, and the Western United States

[[[ Background: ]]] Invasive pneumococcal disease (IPD) isolates forming genomic clusters can reflect rapid disease transmission between vulnerable individuals. [[[ Methods: ]]] We performed whole genome sequencing of 2820 IPD isolates recovered during 2019 through Centers for Disease Control and Prevention’s Active Bacterial Core surveillance to provide strain information (serotypes, resistance, genotypes), and 2778 of these genomes were analyzed to detect highly related genomic clusters. [[[ Results: ]]] Isolates from persons experiencing homelessness (PEH) were more often within genomic clusters than those from persons not experiencing homelessness (PNEH) (105/198 [53.0%] vs 592/2551 [23.2%]; P < .001). The 4 western sites accounted for 33.4% (929/2778) of isolates subjected to cluster analysis yet accounted for 48.7% (343/705) of clustering isolates (P < .001) and 75.8% (150/198) of isolates recovered from PEH (P < .001). Serotypes most frequent among PEH were (in rank order) 12F, 4, 3, 9N, 8, 20, and 22F, all of which were among the 10 serotypes exhibiting the highest proportions of clustering isolates among all cases. These serotypes accounted for 44.9% (1265/2820) of all IPD cases and are included within available vaccines. [[[ Conclusions: ]]] We identified serotype-specific and geographic differences in IPD transmission. We show the vulnerability of PEH within different regions to rapidly spreading IPD transmission networks representing several pneumococcal serotypes included in available vaccines.