Reducing Ongoing Transmission of Tuberculosis

Restriction fragment length polymorphism (RFLP) analysis to identify specific Mycobacterium tuberculosis strains, in combination with epidemiologic investigation, has shattered old dogmas and yielded new insights into the transmission dynamics of tuberculosis. In several cities in the United States, RFLP studies and epidemiologic analysis have shown that 19% to 54% of tuberculosis cases probably result from recent infection.^{1 - 4} Effective tuberculosis control measures should reduce this percentage. Yet, in this issue of THE JOURNAL, Bishai and colleagues⁵ suggest that recent transmission accounted for 32% of tuberculosis cases in Baltimore, Md, where an excellent tuberculosis control program has used community-based directly observed therapy (DOT) since 1981.

How can these results be explained? First, during the 5 years preceding the study period, 33% of patients diagnosed as having tuberculosis in Baltimore were not in the DOT program,⁶ and suboptimal therapy for these patients may have contributed to continued disease transmission. Second, limiting transmission of tuberculosis requires a multifaceted program that includes timely identification of cases, isolation of contagious cases, aggressive contact investigation, and preventive therapy for persons infected with M tuberculosis. It is difficult to assess these aspects of a tuberculosis control program, and suboptimal performance in these areas also may have contributed to ongoing tuberculosis transmission.

Despite these qualifications, the most likely explanation for continuing transmission of tuberculosis in Baltimore and in many other locations is that transmission of most infections preceded the initiation of antituberculosis therapy. Most patients with tuberculosis are socially disadvantaged and often delay seeking medical care. Asch et al⁷ recently reported that 30% of symptomatic patients with tuberculosis did not obtain medical attention for more than 30 days after symptom onset. In addition, substantial transmission of tuberculosis probably occurs to casual contacts in urban areas where tuberculosis is prevalent. In support of this view, most patients in Baltimore infected with the same M tuberculosis strain did not have close contact but were geographically clustered. Furthermore, extensive transmission of tuberculosis to casual contacts has been documented in workplaces, homeless shelters, bars, and other social settings.^{4 ,8 - 10} In 26% of cases where 2 patients with tuberculosis live in the same household, these patients are infected with different strains of M tuberculosis, indicating that infection was acquired outside the household.¹¹

To reduce transmission of tuberculosis, it is critical first to identify high-risk populations and sites where tuberculosis is transmitted. Because the transmission dynamics of tuberculosis vary widely geographically, identification of high-risk populations and tuberculosis transmission sites depends on local public health authorities, who must use their knowledge of tuberculosis epidemiology, perhaps with the aid of RFLP analysis. For example, RFLP analysis and epidemiologic investigation showed that 41% of tuberculosis cases in central Los Angeles, Calif, were due to extensive transmission of only 2 strains at 3 homeless shelters.⁴ With the availability of a rapid RFLP method that has the potential for use in clinical laboratories,¹² it may soon be possible to obtain RFLP results within 48 hours of obtaining a positive result from an acid-fast smear or culture analysis. Rapid RFLP analysis has the potential to detect tuberculosis outbreaks quickly and before they become widespread, allow intensive epidemiologic investigation of patients infected with the same M tuberculosis strain, and facilitate institution of public health interventions. Commitment by scientific and public health agencies is needed to fully develop rapid RFLP methods for clinical use and to improve communication between laboratory facilities and public health authorities.

If high-risk populations and tuberculosis transmission sites can be identified, 1 approach to reduce disease transmission is to identify persons with tuberculosis before they seek medical attention. Case finding can be done by tuberculin skin testing, chest radiography, or acid-fast smear and mycobacterial culture of sputum. Tuberculin skin testing is problematic because it requires 2 visits and because negative test results are common in patients diagnosed as having tuberculosis. Sputum analysis allows rapid detection of patients who have a positive acid-fast smear result. However, 50% of patients with tuberculosis have negative smear results,^{13 - 14} and these patients may be difficult to locate when culture results become available. Chest radiography may be a useful approach because it permits rapid identification of most patients with infectious tuberculosis. Layton et al¹⁵ found that screening chest radiographs of 3643 persons in a New York jail identified 15 unsuspected cases of tuberculosis, or 412 cases per 100,000 persons. Screening chest radiographs of 9877 homeless individuals in Los Angeles yielded 42 tuberculosis cases, or 425 cases per 100,000 persons (Paul T. Davidson, MD, written communication, September 1998). These rates are more than 50 times that in the US population and screening is likely to be cost-effective in well-defined high-risk populations, although formal studies have not been done.

In addition to case finding, installation of upper-room germicidal irradiation at suspected tuberculosis transmission sites also may be useful, particularly at congregate sites such as homeless shelters.¹⁶ Early studies showed that air piped from rooms housing patients with tuberculosis was rendered noninfectious for guinea pigs by germicidal irradiation.¹⁷ However, this approach remains controversial because controlled data on its efficacy in clinical settings are lacking.

The majority of tuberculosis cases in the United States result from reactivation of remote infection and are avoidable only by administration of preventive therapy to persons infected with M tuberculosis. Screening for tuberculosis infection with tuberculin skin testing is recommended in high-risk populations such as foreign-born persons from countries with high incidence of tuberculosis and medically underserved low-income persons.¹⁸ However, these recommendations may not be followed because local public health departments likely have inadequate resources with which to screen large populations and to ensure completion of preventive therapy. The feasibility of screening may be improved by identification of smaller, more well-defined high-risk populations. For example, Cegielski et al¹⁹ found that identification of high-risk neighborhoods by precise mapping of tuberculosis cases, combined with door-to-door tuberculin skin testing, yielded a high percentage of persons who were candidates for preventive therapy.

The study by Bishai and colleagues⁵ demonstrates that significant transmission of tuberculosis continues despite the widespread use of DOT. Limiting transmission of tuberculosis hinges on local efforts to identify high-risk populations and tuberculosis transmission sites, as well as the political will to expend resources for this purpose. Identification of high-risk groups depends on local epidemiologic information, supplemented by RFLP analysis, preferably using rapid RFLP methods as these become available. Once high-risk populations are identified, case finding based on screening chest radiographs, use of upper-room germicidal irradiation at tuberculosis transmission sites, and targeted screening for tuberculosis infection, followed by preventive therapy, should help reduce ongoing transmission of tuberculosis, a necessary prelude to eliminating this disease from the United States.