October 16, 1996

A Multi-institutional Outbreak of Highly Drug-Resistant TuberculosisEpidemiology and Clinical Outcomes

Author Affiliations

From the New York City Department of Health, Bureau of Tuberculosis Control (Drs Frieden, Maw, Fujiwara, and Brudney and Mss Sherman and Nivin); the Division of Tuberculosis Elimination, Centers for Disease Control and Prevention, Atlanta, Ga (Drs Frieden and Fujiwara); the Division of AIDS, STD, and TB Laboratory Research, National Center for Infectious Disease, Atlanta, Ga (Dr Crawford); St Clare's Hospital, New York, NY (Dr Sharp); Lincoln Medical and Mental Health Center, Bronx, NY; (Dr Hewlett); Columbia Presbyterian Medical Center, New York City (Dr Brudney); Howard Hughes Medical Institution, Albert Einstein College of Medicine, Bronx, NY (Dr Alland); and Public Health Research Institute, New York, NY (Dr Kreiswirth). Dr Sharp is now with the Beth Israel Medical Center, New York, NY.

JAMA. 1996;276(15):1229-1235. doi:10.1001/jama.1996.03540150031027

Objective.  —To investigate a multi-institutional outbreak of highly resistant tuberculosis and evaluate patient outcome.

Design.  —Epidemiologic investigation of every tuberculosis case reported in New York City.

Setting.  —Patients cared for at all public and nonpublic institutions from January 1, 1990, to August 1, 1993 (43 months).

Patients.  —We reviewed medical and public health records and conducted clinical, epidemiologic, drug susceptibility, and restriction fragment length polymorphism (RFLP) analyses. A case was defined as tuberculosis in a patient with an isolate resistant to isoniazid, rifampin, ethambutol hydrochloride, and streptomycin (and rifabutin, if sensitivity testing included it), and, if RFLP testing was done, a pattern identical to or closely related to strain W.

Main Outcome Measures.  —Patient survival and the conversion of sputum cultures from positive to negative.

Results.  —Of the 357 patients who met the case definition, 267 had identical or nearly identical RFLP patterns; isolates from the other 90 patients were not available for RFLP testing. Among these 267 patients, 86% were human immunodeficiency virus (HIV)-infected, 7% were HIV-negative, and 7% had unknown HIV status. All-cause mortality was 83%. Epidemiologic linkages were identified for 70% of patients, of whom 96% likely had nosocomially acquired disease at 11 hospitals. Survival was prolonged among patients who recieved medications to which their isolate was susceptible, especially capreomycin sulfate, and among patients with a CD4+ T-lymphocyte count greater than 0.200×109/L (200/μL). Treatment with isoniazid and a fluoroquinolone antibiotic was also independently associated with longer survival.

Conclusions.  —This outbreak accounted for nearly one fourth of the cases of multidrug-resistant tuberculosis in the United States during a 43-month period. Most patients had nosocomially acquired disease, were infected with HIV, and unless promptly and appropriately treated, died rapidly. With appropriate directly observed treatment, especially combinations including an injectable medication, even severely immunocompromised patients had culture conversion and prolonged, tuberculosis-free survival.