Abstract
Both epidemiologic and strain-related factors may contribute to large clusters of tuberculosis patients.
Tuberculosis patients with identical strains of
Mycobacterium tuberculosis
are described as clustered. Cluster size may depend on patient or strain characteristics. In a 7-year population-based study of tuberculosis in Karonga District, Malawi, clusters were defined by using IS
6110
restriction fragment length polymorphism, excluding patterns with <5 bands. Spoligotyping was used to compare strains with an international database. Among 682 clustered patients, cluster size ranged from 2 to 37. Male patients, young adults, and town residents were over-represented in large clusters. Cluster size was not associated with HIV status or death from tuberculosis. Spoligotypes from 9 (90%) of 10 large cluster strains were identical or very similar (1 spacer different) to common spoligotypes found elsewhere, compared with 37 (66%) of 56 of those from nonclustered patients (p = 0.3). Large clusters were associated with factors likely to be related to social mixing, but spoligotypes of common strains in this setting were also common types elsewhere, consistent with strain differences in transmissibility.