Hospital outbreak of multidrug-resistant Mycobacterium tuberculosis infections. Factors in transmission to staff and HIV-infected patients.

OBJECTIVE: To describe transmission of multidrug-resistant (MDR) Mycobacterium tuberculosis infection among patients and health care workers (HCWs) in a ward and clinic for human immunodeficiency virus (HIV)-infected patients in a hospital, four studies were conducted.
METHODS: Case patients and control patients were persons who had been treated in the HIV ward or clinic, whose clinical course was consistent with tuberculosis and who had at least one positive culture for M tuberculosis between January 1, 1988, and January 31, 1990, resistant to at least isoniazid and rifampin (case patients), or whose isolates were susceptible to all drugs tested (control patients). In the first study, case patients and control patients were compared to identify risk factors for MDR tuberculosis. In the second study, inpatient and outpatient days of MDR tuberculosis case patients were compared to determine whether acid-fast bacillus (AFB) smear-positivity or aerosolized pentamidine use was associated with higher numbers of subsequent MDR tuberculosis cases among exposed patients. In the third study, restriction fragment length polymorphism analysis was performed on available MDR and sensitive M tuberculosis isolates. In the fourth study, skin test conversion rates among HCWs in the HIV ward and clinic were compared with those of HCWs in another ward, and the strength of the associations between skin test conversions among HCWs on the HIV ward and the number of person-days that AFB smear-positive case patients and control patients were on this ward was estimated.
RESULTS: Case patients were more likely than control patients to have been exposed on the HIV ward or clinic to an AFB smear-positive case patient (P less than .001). Inpatient and outpatient days of MDR tuberculosis case patients were associated with more subsequent cases of MDR tuberculosis if exposing case patients were smear-positive or if they received aerosolized pentamidine (P less than or equal to .01). Of 13 MDR isolates, all had one of two restriction fragment length polymorphism patterns; 10 sensitive isolates had restriction fragment length polymorphism patterns that were different from each other. The HCW skin test conversion rate was higher on the HIV ward and clinic than on the comparison ward (P less than .01). The risk of occupational acquisition of infection increased in direct proportion to the number of person-days that AFB smear-positive case patients were on the HIV ward (r = .75; P = .005), but did not increase in proportion to the number of person-days that AFB smear-positive control patients were there (r = -.36; P = NS). After isolation measures for AFB smear-positive tuberculosis patients were improved, MDR tuberculosis cases decreased to seven of 214 tuberculosis patients.
CONCLUSIONS: Nosocomial transmission of MDR M tuberculosis infection to patients and HCWs occurred on the HIV ward and clinic. Infectiousness of MDR tuberculosis case patients was associated with AFB sputum-smear positivity. Case patients with MDR tuberculosis created a greater risk of skin test conversion for HCWs on the HIV ward than drug-susceptible control patients.