Molecular genetic typing (or genotyping) of Mycobacterium tuberculosis strains has revolutionized the field of tuberculosis (TB) research, prevention, and control (1–3). The subtypes characterized by molecular genetic typing methods provide a greater power and ability to differentiate strains than previous methods, such as comparisons of patterns of drug resistance or phage-typing (4,5). When molecular genotyping technology is applied to outbreaks or unusual clusters of disease, persons sharing M. tuberculosis strains can be identified, which can lead to important clues about the pattern and dynamics of transmission.

Methods of molecular genotyping have been increasingly applied to the epidemiology of TB. Because of its natural history, the transmission of M. tuberculosis is difficult to study; M. tuberculosis is spread by airborne droplets of respiratory secretions expelled by an infectious person to a susceptible host, who may or may not be known to the source (6). The bacterium can remain latent as an asymptomatic infection for years, and the source of such infections can be difficult to ascertain. Thus, the places and persons involved in a chain of transmission may be puzzling to identify or exclude. Molecular typing of M. tuberculosis adds important pieces to the construction of such a transmission puzzle; persons who harbor the same strain of M. tuberculosis are likely to have shared that strain in a common chain of transmission; by contrast, persons who are infected by unique and distinct strains were probably infected by means of a different exposure. Among its applications, genotyping has served to elucidate the poorly understood role of relapses and exogenous reinfection of persons with recurrent TB after cure. Several reports have relied on DNA genotyping to describe and document the occurrence of exogenous reinfection with distinct strains of M. tuberculosis as the cause of TB following successful treatment (7–9).

The usefulness of molecular typing was also confirmed in several epidemiologic investigations of HIV-associated multidrug-resistant TB outbreaks in hospices, hospitals, and prisons during the late 1980s and early 1990s and provided compelling evidence of institutional transmission (10–16). Unique features of these outbreaks included prolonged infectiousness of patients who were not recognized to harbor multidrug-resistant TB until months after their TB was diagnosed and the relatively rapid progression from latent infection to active TB disease in persons with HIV-associated immunosuppression. These data were also used to state the need to implement effective interventions to halt such outbreaks (17,18).

In addition to the use of DNA genotyping during outbreak investigations, the technology has been applied as a complementary tool to conventional methods in TB control (19–22). In two of the earliest studies conducted in the United States, one in San Francisco and one in New York City, the authors assumed that M. tuberculosis isolates with matching DNA fingerprints were epidemiologically related and represented recent transmission of M. tuberculosis among the patients involved (i.e., within 2 years before diagnosis) (19–20). In these studies, 30% to 40% of the patients had M. tuberculosis isolates with DNA fingerprint patterns that matched at least one other isolate. This finding led the authors to conclude that as many as 40% of TB cases in these two cities were the result of recent transmission and that TB control practices in San Francisco and New York were not effectively decreasing M. tuberculosis transmission. The observations in these two reports were useful and innovative. However, the findings could not be generalized to other geographic areas in the United States because the study populations were exclusively urban residents, a large proportion were HIV infected, and detailed epidemiologic information was incomplete for either study, which limited the ability to corroborate actual contacts or exposures among the TB patients studied.

A separate study evaluated the use of DNA genotyping in TB patients from a large, rural population in the state of Arkansas (21). Analysis of M. tuberculosis isolates from TB patients for a 2-year period (1992–1993) found that more than half the isolates matched at least one other patient isolate. Epidemiologic investigation of the patients with matching M. tuberculosis isolates (i.e., clustered cases) revealed 24 persons who had documented latent TB infection or active TB many years in the past but which produced disease with matching isolates at the time of the study. In five of these patients, a remote epidemiologic connection (i.e., common exposures to a person with TB) was identified that occurred 20–25 years earlier. The authors also reported that some TB patients had isolates showing identical specific patterns, yet their geographic, social, and medical histories were so disparate that transmission among them was highly unlikely. Further investigation by genotyping with an independent method revealed that the capacity of the IS6110 method to differentiate strains was roughly proportional to the number of bands present in the original fingerprint pattern. Thus, isolates with a low number of bands (e.g., fewer than five) required a second method for appropriate differentiation. These findings suggested the need for additional assessments and evaluations of emerging assumptions in the interpretation of DNA genotyping of M. tuberculosis.

The fruitful use of DNA genotyping to confirm and refine our understanding of M. tuberculosis transmission in outbreaks provided the major impetus to evaluate the use of this technology in other settings and to determine its broader application as a tool in TB prevention and control. Specifically, we sought to assess the usefulness of this technology in searching for unidentified outbreaks, identifying risk factors for TB at the population level, and identifying and monitoring laboratory cross-contamination (23). Consequently, the Centers for Disease Control and Prevention (CDC) funded six laboratories in 1993 to develop regional databases of DNA genotype fingerprint patterns and undertake regional molecular epidemiologic studies. This genotyping network was expanded in 1996 to include sentinel surveillance sampling. Ultimately, the National Tuberculosis Genotyping and Surveillance Network comprised CDC, seven laboratories, and seven sentinel surveillance sites in the United States.

Top