| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Guest Access | Sign In via User Name/Password
|
|
|
|||||||
Guest Access | Sign In via User Name/Password |
|||||||||
Drs. Lauzardo, Hollender, Narita, and Ashkin are from A. G. Holley State Tuberculosis Hospital, the Florida Department of Health. In addition, Drs. Ashkin and Narita are Clinical Assistant Professors, Division of Pulmonary and Critical Care Medicine, University of Miami School of Medicine, and Dr. Lauzardo is an Adjunct Instructor, Division of Pulmonary and Critical Care Medicine, University of Florida School of Medicine.
Correspondence to: David Ashkin, MD, FCCP, A. G. Holley State Hospital, 1199 W Lantana Rd, Lantana, FL 33462; e-mail: David_Ashkin{at}doh.state.fl.us
For some, there is nothing more thrilling than watching an old black-and-white movie in which the object of the film is to solve the mystery. In these movies, people faced with a dilemma pounded the pavement, often until the soles of their shoes revealed holes, searching for the answers to their quandary. For those who do the detective work of tuberculosis (TB) control, this scenario seems very familiar.
Traditional "shoe leather" TB contact investigations consist of interviewing individuals with TB and finding out with whom they have been in contact or what places they have visited. This practice has been the cornerstone of contact investigations for > 40 years and has remained a priority of TB control. Recent reports, however, have suggested that this cornerstone may have cracks, as was suggested by a study by Small et al1 from San Francisco. In their study, it was found that traditional contact investigations only identified 10% of individuals who, later, were linked by molecular analysis. As was apparent in the old movies, the sleuth was only as good as his magnifying glass, and sometimes new tools were necessary.
Enter molecular epidemiology, utilizing techniques such as restriction fragment length polymorphisms (RFLP), better known as DNA fingerprints. With this new tool, public health workers can "dust" the scene of the crime, utilizing the genetic material of TB, looking for those essential fingerprints. These fingerprints can give insight into who and where TB might have spread.
Investigations utilizing RFLP testing have caused us to rethink some TB dogma. Traditionally, it has been taught that most cases of TB were acquired in the distant past and became reactivated when the host’s immune system was vulnerable.2 Some studies1 3 utilizing RFLP have suggested that this may not be true, showing evidence of > 30% of cases in which recent infection was more likely to be responsible. RFLP testing has been used to prove that patients can simultaneously have two different strains causing disease4 or that cross-contamination, either in the laboratory where the specimen was processed5 6 7 or by the instruments used to obtain the specimen (eg, bronchoscopes),8 could cause erroneous results.
However, one of the most important leads that RFLP has provided is a clue to the scene of the crime. RFLP has been instrumental in showing outbreaks of TB in institutions such as jails, hospitals, and prisons or in social settings such as churches, bars, or card games.9 10 11 12 Given the transient nature of the settings and at times the people who frequent these situations, traditional contact investigations have not always revealed these settings as the sites of transmission. Whereas traditional shoe leather public health may have relied more on finding individuals who may have been exposed, RFLP testing has assisted in revealing settings where transmission may be occurring. Once these facilities have been identified, interventions such as environmental controls (eg, ultraviolet lights) or enhanced surveillance have proven to be effective.8 13
In this issue of CHEST, the study by Dobbs et al (see page 1877) is an important contribution to the literature of TB epidemiology, because it describes the utilization of RFLP as an adjunct to TB surveillance in a rural setting. Since 1994, all strains of TB collected by the Alabama State Laboratory (which is required by law for all cultures from active cases) were sent to the University of Alabama at Birmingham for RFLP analysis, surveillance, and storage in an RFLP database. Matching fingerprints were grouped together and tracked over time. For those clusters of cases persisting over a period of time, it was assumed that ongoing transmission was still occurring, and secondary investigation was initiated. An investigation of one such cluster, consisting of 25 isolates, revealed a suspected focus involving the homeless population. On further review, it appeared that the primary sites of transmission were a local jail and two homeless shelters. This outbreak, which was previously unrecognized, involved 10 counties. Due to the recognition of this outbreak, officials initiated targeted screening of the homeless and jail populations.
It must be emphasized that the authors went to significant lengths to ensure that those cases that were grouped together in the cluster were truly linked. The cluster was a 12-band cluster. Secondary typing was performed and was identical among the 25 specimens, and a review of the registries from the 10 counties, utilizing staff members from the local health department, was performed. In addition, the individuals involved in the cluster of cases were from two high-risk groups for TB transmission (ie, the homeless and the incarcerated). The evidence provided gives a strong indication that these cases are part of a previously undetected outbreak.
The slow transmission of the disease, spreading for > 5 years and involving numerous counties, made the detection of a possible link by traditional techniques difficult. In addition, 14 of the 25 patients were homeless at some point. Traditional epidemiologic techniques have proven to function poorly in this population of individuals who are often transient and stay temporarily in congregate living conditions.14 This outbreak was subsequently shown to have spread across county and state lines.
Just like those movies from the 1930s, which were often set in urban areas, many of the early investigations involving RFLPs focused on urban settings where TB in the United States was most prevalent. Transmission patterns in urban settings often were assumed to be similar to those of rural settings and, thereby, applicable to rural settings.
Although the study by Dobbs et al points out the usefulness of molecular epidemiology in a rural setting with endemic TB, a word of caution would be prudent. The relatedness of strains and the cases that they produce are dependent on identical RFLP strains and population factors. The age of the population, the rate of previous infection, the risk of progressing to active disease within the population, and the stability of the population are all factors that can influence the predictive value for epidemiologic associations among identical RFLP strains. In a study performed in Arkansas, a rural state with demographics that are similar to those of Alabama, Braden et al15 found that most clustered cases were not epidemiologically linked. They concluded that RFLP clustering may have a low predictive value for recent epidemiologic connections in stable rural populations.
In contrast, when a number of cases occur over a short period of time in a rural setting, an assumption might be made that the cases are in fact connected, as was recently seen in a small outbreak (five active TB cases in < 6 months) in an isolated, rural town in Florida. However, all of the isolates of Mycobacterium tuberculosis were proven to be different strains by RFLP (M. Lauzardo, MD; unpublished data). Without the knowledge that these cases were not, in fact, connected, valuable resources might have been wasted looking for an index case or a source of transmission.
However, as we have learned from other new technologies, the utilization and interpretation of results gathered from molecular epidemiology must be better delineated and the exact role that such tools will play must be better defined. There are studies that suggest that DNA patterns may change over time or "drift,"16 thus blurring the ability of this technique to accurately assess linkage. In addition, this technique is costly and technically demanding to perform, interpret and, later, compare. We need to ask (and answer) certain of the following questions: (1) should RFLP be utilized universally, or its use limited, and, if so, how; (2) who should store the data and who should have access to that data in order to look for and examine clusters that extend past county, state, or national lines; (3) what is the most accurate (or practical) way to define a cluster;(4) who does the surveillance work looking for such clusters; (5) who does the epidemiologic legwork investigating the cluster; and (6) who is responsible for implementing the necessary measures to halt further transmission in sites where TB is found to occur.
Importantly, it must be realized that RFLP studies alone are of little value and must be used in conjunction with traditional shoe leather epidemiology to discover the source and site of transmission and to take steps to interrupt further transmission. As we have witnessed in the past with decreasing disease rates, the trend has been to reduce funding for TB control activities. This practice not only serves to cripple the ability of public health to continue to contain TB, but also has helped to allow the disease to emerge stronger and as a greater threat to public health. Thus, it is important that we ensure that resources are available to maintain the essential "soul" (or "sole") of shoe leather public health and not throw the shoe away when a hole appears, but instead reinforce and maintain it, utilizing new tools and innovative approaches that can allow public health to continue to beat those pavements searching for cases of TB.
RFLP analysis, as an adjunct to conventional shoe leather epidemiology, is a great tool for providing clues about transmission in both rural and urban settings. Future investigators need to bear in mind the limitations of this technology, particularly in rural settings with endemic disease, where transmission may have occurred years, even decades, ago. Nonetheless, as in the final scene of "Casablanca" with the engines roaring, public health and RFLPs may be at "the beginning of a beautiful friendship." And, just as in those great movies from the 1930s, at the end of the story we are often left with more questions than answers.
References
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
