ABSTRACT
BACKGROUND: DNA fingerprinting establishes the genetic relatedness of Mycobacterium tuberculosis isolates and has become a powerful tool in tuberculosis epidemiology. OBJECTIVE: To use DNA fingerprinting to assess the efficacy of current tuberculosis infection-control practices. DESIGN: Retrospective molecular and descriptive epidemiologic study. SETTING: A 700-bed urban public hospital that follows the Centers for Disease Control and Prevention (CDC) guidelines for tuberculosis infection control. PATIENTS: 183 patients who had positive cultures for M. tuberculosis from 1 April 1995 to 31 March 1996. RESULTS: 173 of 183 M. tuberculosis isolates from the study period underwent DNA fingerprinting. Fingerprinting revealed that five isolates represented false-positive cultures and that 91 (54%) of the remaining 168 isolates were in 15 DNA fingerprinting clusters, which ranged in size from 2 to 29 isolates. Risk factors for clustering were birth in the United States, African-American ethnicity, homelessness, substance abuse, and male sex. Retrospective epidemiologic analysis of inpatient and outpatient visits by the 91 patients who had clustered isolates revealed only one possible instance of patient-to-patient transmission. CONCLUSIONS: The DNA fingerprinting of all M. tuberculosis isolates from a 1-year period revealed one possible instance of nosocomial transmission and five false-positive M. tuberculosis cultures. However, these results did not lead to changes in infection-control practices or in clinical care. The study findings do not support the use of DNA fingerprinting for nosocomial tuberculosis surveillance, but they suggest that compliance with the CDC tuberculosis infection-control guidelines may control patient-to-patient transmission in high-risk urban hospitals.