© 2004 American Thoracic Society
Tuberculosis Contact InvestigationsPlease Don't Fail Me NowUniversity of California, San Francisco San Francisco, California Molecular genotyping has revolutionized our ability to track strains of Mycobacterium tuberculosis as they spread through a community. Studies of tuberculosis that use molecular typing techniques in combination with standard epidemiologic investigations have elucidated both suspected and unsuspected transmission and estimated the proportion of cases due to recent infection (as opposed to reactivation) (1–7). Molecular epidemiology studies have also identified some of the strengths and weaknesses of our current tuberculosis control and prevention strategies including the investigation of contacts to infectious cases. Contact investigation, or tracing, is one of the top priorities of tuberculosis control programs, ranking second behind case finding and treatment of active cases (8). Molecular epidemiology studies, however, have demonstrated that only 5 to 10% of clustered cases are identified through contact investigations (2, 5, 6). These studies have questioned the ability of contact investigations to identify cases in our community and raised the possibility of casual transmission, that is, transmission occurring outside of the usual household, work, and social settings. In this issue of the Journal (pp. 806–810), van Deutekom and colleagues (9) describe the molecular epidemiology of tuberculosis in the Netherlands using IS6110-based restriction fragment length polymorphism analysis (RFLP) along with an extensive standardized questionnaire that focused on establishing epidemiologic linkages. Some patients, including those who were found to be part of a cluster, underwent a second interview to identify potential epidemiologic linkages with others in the cluster. Cases were considered clustered if their isolate of M. tuberculosis was identical to another patient's isolate during the preceding two years. Based on the information obtained through contact tracing, interviews, and genotyping, the authors attempted to answer two important questions. Do clustered cases represent recent infection? Can clustered cases be identified earlier through contact tracing? The answer to the first question has important implications for all molecular epidemiology studies. The basic assumption in molecular epidemiology studies is that two persons whose isolates have identical genotyping patterns are connected either directly or indirectly in a chain of transmission. Individuals whose isolates share the same genotyping pattern are referred to as "clustered" and when these cases occur over a relatively short period of time, they are thought to represent recent infection with rapid progression to disease. Numerous studies have used clustering as a proxy for recent infection (1–7) while others have reported that clustering does not represent recent infection, particularly in low incidence rural areas (10). van Deutekom and colleagues (9) report that 20% of the 481 patients in North Holland fell into 43 different clusters. Of the clustered cases, 86% had epidemiologic links suggesting that recent transmission had occurred. Thus, in this study, clustering was indicative of recent infection. The study design used by the investigators is admirable and adds support to the study findings. To fully understand the significance of clustering as a marker for recent infection, however, similar studies will need to be performed in different settings (e.g., low incidence, high incidence, urban, and rural areas) using a similar study design. The second question, "Can tuberculosis cases be identified earlier?", has implications for all tuberculosis control programs. van Deutekom and colleagues (9) divided the patients into one of five "transmission groups": clear epidemiologic links, confirmed by genotyping and contact tracing (24%); clear epidemiologic links, confirmed by genotyping and second interview—but not through contact tracing (6%); initially unclear epidemiologic links that became likely after genotyping and second interview (55%); no epidemiologic links but genotyping indicated clustering (14%); and patients who were part of a different cluster than expected (1%). For those in group 1, contact tracing was successful in that it identified secondary cases associated with the index case. Patients in group 2, however, represent failed contact investigations and the reasons for these failures are known to all who work in tuberculosis control: the contact was not mentioned by the source case (n = 3), refusal to participate in the investigation (n = 3), or simply program error (n = 4). Combining groups 1 and 2 would suggest that the best contact investigations could have done was to identify approximately 30% of the clustered cases. Perhaps the most important findings were in group 3, those whose epidemiologic links were only discovered after genotyping information became available and a second interview was performed. This group of patients represented 63% of those who had evidence for recent transmission. The authors noted, as have others (11), that much of the transmission that occurred in this group occurred in settings outside of the traditional concentric circle approach used to perform contact investigations and among groups that frequently share social settings or living facilities and who are often reticent to provide names of contacts. Therefore, the answer to the second question was that there were few opportunities to identify recently infected cases earlier. The study by van Deutokom and colleagues (9) provides several lessons for tuberculosis control in general and, molecular epidemiology studies in particular. First, clustering of tuberculosis cases in settings like North Holland usually represents recent infection. Intensive epidemiologic investigations, however, must be combined with genotyping to understand transmission dynamics. Second, contact tracing identifies only a small proportion of cases that are part of a cluster. It is important to point out, however, that contact investigations continue to be one of the highest yield methods of active case finding, identifying tuberculosis in 1–2% of contacts (8, 12). Moreover, contact investigations find individuals with latent infection who represent preventable cases. Third, and perhaps most important, most clustered cases that are eventually determined to be linked epidemiologically provide limited opportunity for earlier intervention and often result from "casual" contact. Although the authors view this last point as a limitation of contact investigations, the study may have identified a potential solution. By using genotyping data to prompt a second interview the authors were able to identify potential circumstances and places where transmission occurred. They did so by using genotyping techniques that require cultures of M. tuberculosis and thus valuable time was lost waiting for the organism to grow and for genotyping results. As real-time, amplification-based genotyping techniques become available, we will be able to more rapidly identify persons who need a second interview and, thus, more quickly detect secondary cases, preventable cases, and sites of transmission. Clearly, the concentric circle method may not address all contact investigation needs and there are considerable limitations with this approach. Contact tracing, however, is not failing us in our attempts to find active cases; it just needs a little help. By using rapid genotyping methods and targeted second interviews, contact investigations will likely become an increasingly important element of tuberculosis control and prevention activities in the future. FOOTNOTES Conflict of Interest Statement: C.L.D. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript. REFERENCES
This article has been cited by other articles:
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
