[Skip to Navigation]
October 8, 1997

Patterns of Tuberculosis Transmission in Central Los Angeles

Author Affiliations

From the Division of Infectious Diseases, Departments of Medicine (Drs Barnes and Jones and Ms Otaya) and Preventive Medicine (Dr Preston-Martin), University of Southern California School of Medicine, County of Los Angeles Department of Health Services Tuberculosis Control Program (Ms Knowles), Los Angeles County Public Health Laboratory (Dr Harvey), Los Angeles; Statology (Dr Pogoda), Truckee, Calif; and the Departments of Pathology (Drs Yang and Eisenach) and Anatomy (Dr Cave), University of Arkansas for Medical Sciences and Medical Research Service, McClellan Memorial Veterans Hospital, Little Rock, Ark.

JAMA. 1997;278(14):1159-1163. doi:10.1001/jama.1997.03550140051039

Context.  —Recent studies suggest that many tuberculosis cases in urban areas result from recent transmission. Delineation of the epidemiologic links between patients is important to optimize strategies to reduce tuberculosis transmission.

Objective.  —To identify epidemiologic links among recently infected urban patients with tuberculosis.

Design.  —Prospective evaluation of patients with tuberculosis.

Setting.  —Central Los Angeles, Calif.

Patients.  —A total of 162 patients who had culture-proven tuberculosis.

Interventions.  —Patients were prospectively interviewed to identify their contacts and whereabouts. The IS6110-based and pTBN12-based restriction fragment length polymorphism analyses were performed on Mycobacterium tuberculosis isolates. Patients whose isolates had identical or closely related restriction fragment length polymorphism patterns were considered a cluster. Unconditional logistic regression was used to identify independent predictors of clustering.

Main Outcome Measures.  —Relationship of clinical and epidemiologic variables to clustering.

Results.  —A total of 96 (59%) of 162 patients were in 8 clusters. Only 2 of the 96 clustered patients named others in the cluster as contacts. The degree of homelessness was an independent predictor of clustering. Compared with non-clustered patients, patients in 6 clusters were significantly more likely to have spent time at 3 shelters and other locations when at least 1 patient in the cluster was contagious, and these locations were independent predictors of clustering. Among nonhomeless persons, clustered patients were significantly more likely than non-clustered patients to have used daytime services at 3 shelters.

Conclusions.  —(1) Traditional contact investigation does not reliably identify patients infected with the same M tuberculosis strain, and (2) locations at which the homeless congregate are important sites of tuberculosis transmission for homeless and nonhomeless persons. Measures that reduce tuberculosis transmission should be based on locations rather than on personal contacts.