High Incidence of Hospital Admissions With Multidrug-Resistant and Extensively Drug-Resistant Tuberculosis Among South African Health Care Workers

Background: Nosocomial transmission has been described in extensively drug-resistant tuberculosis (XDR-TB) and HIV co-infected patients in South Africa. However, little is known about the rates of drug-resistant tuberculosis among health care workers in countries with high tuberculosis and HIV burden. Objective: To estimate rates of multidrug-resistant tuberculosis (MDR-TB) and XDR-TB hospitalizations among health care workers in KwaZulu-Natal, South Africa. Design: Retrospective study of patients with drug-resistant tuberculosis who were admitted from 2003 to 2008 for the initiation of drug-resistant tuberculosis therapy. Setting: A public tuberculosis referral hospital in KwaZulu-Natal, South Africa. Participants: 231 health care workers and 4151 non-health care workers admitted for initiation of MDR-TB or XDR-TB treatment. Measurements: Hospital admission rates and hospital admission incidence rate ratios. Results: Estimated incidence of MDR-TB hospitalization was 64.8 per 100 000 health care workers versus 11.9 per 100 000 non-health care workers (incidence rate ratio, 5.46 [95% CI, 4.75 to 6.28]). Estimated incidence of XDR-TB hospitalizations was 7.2 per 100 000 health care workers versus 1.1 per 100 000 non-health care workers (incidence rate ratio, 6.69 [CI, 4.38 to 10.20]). A higher percentage of health care workers than non-health care workers with MDR-TB or XDR-TB were women (78% vs. 47%; P < 0.001), and health care workers were less likely to report previous tuberculosis treatment (41% vs. 92%; P < 0.001). HIV infection did not differ between health care workers and nonhealth care workers (55% vs. 57%); however, among HIV-infected patients, a higher percentage of health care workers were receiving antiretroviral medications (63% vs. 47%; P < 0.001). Limitation: The study had an observational retrospective design, is subject to referral bias, and had no information on type of health care work or duration of occupational exposure to tuberculosis. Conclusion: Health care workers in this HIV-endemic area were substantially more likely to be hospitalized with either MDR-TB or XDR-TB than were non-health care workers. The increased risk may be explained by occupational exposure, underlining the urgent need for tuberculosis infection-control programs.