Mycobacterium tuberculosis Transmission in Healthcare Settings: Is It Influenced by Coinfection with Human Immunodeficiency Virus?

Abstract

Infection with Mycobacterium tuberculosis is a recognized occupational risk in healthcare environments, and guidelines exist for preventing its transmission in these settings.1 In the United States, several recent outbreaks of tuberculosis, including outbreaks of multidrug-resistant strains of M tuberculosis, have heightened concerns about nosocomial transmission and served as a reminder of the need to fully implement measures to minimize the risk of tuberculosis infection within healthcare facilities.2-7 In each instance, coinfection with human immunodeficiency virus (HIV) and M tuberculosis contributed to the spread of the outbreak by accelerating the progression from tuberculosis infection to active disease. However, in none of the outbreaks did the investigators find evidence suggesting that HIV-infected patients with tuberculosis are more likely to transmit tuberculosis than non-HIV-infected tuberculosis patients. In this issue, Di Perri and colleagues report that the rate of clinically active tuberculosis in healthcare workers caring for HIV-infected patients with tuberculosis was significantly higher than the rate in healthcare workers caring for non-HIV-infected patients with tuberculosis.8 At first glance, the data appear to support the authors' argument. However, under more careful scrutiny, at least five factors potentially confound the scientific argument presented by Di Perri and colleagues. First, no information is provided about the HIV infection status or other medical conditions of the healthcare workers who developed active tuberculosis. In the United States, HV-infected healthcare workers commonly volunteer to care for HIV-infected patients. These healthcare workers would be at greatly increased risk of developing active tuberculosis if exposed to and infected with M tuberculosis. Thus, the apparent difference could be due to host factors causing an increased risk of progression to active disease, rather than to an increased risk of becoming infected. Second, the best measure of recent infection with tuberculosis in immunocompetent healthcare workers is documentation of tuberculin skin-test conversions following exposure to patients with infectious tuberculosis. Di Perri and colleagues state that, because of routine BCG vaccination of healthcare workers at employment, tuberculin skin-test evaluations in the five hospitals are unreliable. However, no clinical data are presented to clearly demonstrate that the tuberculosis diagnosed in the nine healthcare workers represented primary tuberculosis after recent infection versus reactivation of latent tuberculosis. Therefore, reactivation of latent tuberculosis remains a possibility, especially for healthcare workers originating from areas with high rates of tuberculosis. Third, no laboratory data are presented to suggest that both patients and healthcare workers were infected with the same strain of M tuberculosis. Recent epidemiologic observations of possible nosocomial tuberculosis transmission have been confirmed by matching patterns of DNA-fingerprints (restriction fragment-length polymorphism) of M tuberculosis strains obtained from patients and healthcare workers.4,5,7 Fourth, since the acid-fast bacilli isolation precaution procedures and facilities are not described in detail, it cannot be determined whether differences in these could account for the observed differences in the incidence of active tuberculosis among healthcare workers in the three infectious diseases wards and the two pneumology wards. When describing the settings where HIV-infected tuberculosis patients are hospitalized, the authors state that masks are routinely worn by members of hospital staff. However, these facial masks were either usually or rarely worn in the pulmonary and tuberculosis wards housing non-HY-infected tuberculosis patients. Ventilation control precautions for acid-fast bacilli isolation are considered much more important than masks (ie, particulate respirators) in preventing tuberculosis transmission.1 In many of the