Published ahead of print on March 20, 2008, doi:10.1164/rccm.200802-223OC
© 2008 American Thoracic Society doi: 10.1164/rccm.200802-223OC
Pseudoepidemics of Tuberculin Skin Test Conversions in the U.S. Army after Recent Deployments1 Department of Preventive Medicine and Biometrics, Uniformed Services University of the Health Sciences, Bethesda, Maryland; and 2 Army Medical Surveillance Activity, Silver Spring, Maryland Correspondence and requests for reprint should be addressed to James D. Mancuso, M.D., M.P.H., Uniformed Services University of the Health Sciences, Department of Preventive Medicine and Biometrics, 4301 Jones Bridge Road, Bethesda, MD 20814. E-mail: james.mancuso{at}us.army.mil
Rationale: The tuberculin skin test (TST) has many sources of error. These can lead to predominantly false-positive reactions when used in low-risk populations. The U.S. Army deploys to areas considered at high risk for tuberculosis (TB) infection, but often has limited contact with the local population. Objectives: We describe the investigation of eight pseudoepidemics of TST conversions in U.S. Army populations, five of which were associated with overseas deployments. Methods: Outbreak investigations of these pseudoepidemics consisted of several components: evaluation of active and latent TB surveillance data, review of medical records, investigation and interviews of active TB cases and their contacts, evaluation of materials and personnel screening procedures, and placement and reading of repeat skin testing. Measurements and Main Results: Initially reported risk of conversion in the outbreaks ranged from 1.3 to 15%. Repeat testing of converters (positives) found that 30 to 100% were negative on retesting. Several sources of false-positive results were identified in these pseudoepidemics, including variability in reading and administration, product variability, and cross-reactions to nontuberculous mycobacteria. Conclusions: Pseudoepidemics of TST conversions are a common occurrence after U.S. Army deployments and in U.S. Army populations. U.S. Army forces generally have a low risk of TB infection resulting from deployments due to limited exposure to local nationals with active TB, and universal testing in this population has a low positive-predictive value.
Key Words: pseudoepidemic • tuberculin skin test • tuberculosis infection • deployment medicine
Over 100 years ago, the first tuberculin test material was developed by Robert Koch. Although it had little effect as a vaccination against tuberculosis (TB), the purpose for which it was originally designed, it quickly became accepted into clinical and epidemiologic use as a diagnostic tool (1). By the 1930s, tuberculin skin testing had been developed into a screening method to identify those with latent TB infection (LTBI). For the past four decades, treatment of LTBI has been an essential component of TB control in the United States (2). However, due to variable responses to the tuberculin skin test (TST) and the absence of a "gold standard" for the diagnosis of LTBI, the meaning of a "positive" or "negative" skin test is often still uncertain (3, 4). This is because the TST has many sources of error and variability associated with its use. Due to the severity of the consequences of a false-negative result, more attention has typically been paid to factors that influence sensitivity and false-negative results rather than specificity and false-positive results (5). The reasons for false-negative results have been well described, and include host factors, particularly immunosuppression; product-related factors, such as improper storage or handling; and errors in administration, including errors during the injection, reading, or documentation of the test (5, 6). Due to the dramatic decline in the incidence of TB in the United States, effects of false-positive reactions are now increasingly important and notable, particularly in health care and prison settings where testing is common (7–12). Because the predictive value of a test is dependent on the prevalence of disease in a population, an important factor associated with a high proportion of false-positive reactions is a low pretest probability or prevalence of LTBI. This may dramatically reduce the positive-predictive value of the test to below 50% in low-prevalence populations (4, 6), such as the general United States population or the U.S. military. For these reasons, the Centers for Disease Control and Prevention (CDC) recommends a targeted skin approach to skin testing, specifically stating that "screening of low-risk persons is discouraged because it diverts resources from activities of higher priority. In addition, a substantial proportion of tuberculin-test–positive persons from low-risk populations may have false-positive skin tests"(2). There are many factors associated with false-positive TST conversions that are related to the performance characteristics (specificity, in particular) of the test, as summarized in Table 1. These false-positive TSTs affect not only individuals but also result in pseudoepidemics of skin test conversions in populations. Numerous examples of pseudoepidemics of TB skin test conversions have been described for many of the same reasons. Prominent among these is variability in administering the test, including error in the administration (wrong product or dose) (7, 12–14), reading (11), interpretation, and documentation of the TST (15); intra- and intertester variation (16, 17), and prior positives not documented correctly. Another common reason for false-positive TST results is cross-reactivity with nontuberculous mycobacteria (NTM) or bacille Calmette-Guérin vaccine (5). Product-related factors can also result in false-positive TST results, including "hot lots"(8, 18, 19) and variation between manufacturers, in particular the use of Aplisol (Parkedale Pharmaceuticals, Rochester, MI) and the since-withdrawn Sclavo (Sclavo, Wayne, NJ) (8, 9, 20–22). Product-related variability results from hot lots, which have been attributed to particulate matter suspected to be undissolved antigen (10). Other sources of variability include differences between products in the composition of antigens that cross-react to NTM (23). Biologic variability resulting in boosting and reversion is also an important cause of test variability and misclassification (3). These factors lead to a high proportion of false-positive TST results in low-prevalence populations. For this reason, the CDC and other expert advisory panels have recommended a targeted testing approach for over a decade. When any test that is not a gold standard is used in a low-prevalence population, a large proportion of the positives that result will be false positives.
Active TB has long been a concern among deployed U.S. military forces, primarily due to activation of TB acquired in the United States (24–27). Concerns about TB exposure have been raised in recent deployments to Iraq and Afghanistan, which are reported to have among the highest rates of active TB in the world (28). More than 500,000 TSTs are performed in the U.S. Army each year, including testing both before and twice after deployment (29). Universal TB skin testing is currently performed in deploying Army forces because their risk of TB infection is considered to be analogous to that of long-term travelers, although the value of routine TST screening of civilian travelers has also been questioned (30). Testing after recent deployments to these endemic and hyperendemic areas has occasionally resulted in large numbers of U.S. Army service members with TST conversions and massive efforts aimed at preventing active TB. However, the actual contact with local nationals in deployed settings is often very restricted, with many U.S. service members having limited or no contact with local nationals outside U.S. Army installations. Thus, the resultant risk of TB infection is probably much lower than in most long-term travelers. Furthermore, the Army has seen a decrease in active TB cases since the 1980s that parallels the decreases seen in the United States population (31–33). The testing of this predominantly low-risk population results in a low positive-predictive value of the test. This leads to both a greater proportion and absolute number of false positives among the positive test results. From reviews of available records and published reports, we describe eight pseudoepidemics of TST conversions in Army populations, five of which were associated with overseas deployments. The investigation of these pseudoepidemics included repeat testing and a description of factors the led to false-positive skin test results. These findings have relevance for recent and ongoing overseas military deployments in suggesting ways of improving TB prevention and control efforts among deployed U.S. military forces.
The U.S. Army Center for Health Promotion and Preventive Medicine (USACHPPM) conducts surveillance, provides technical assistance, and maintains technical oversight of local and regional preventive medicine and public health activities throughout the U.S. Army. The outbreaks were initially identified and reported to the USACHPPM by medical personnel in the deployed or nondeployed locations, and assistance in investigating the outbreaks was requested. Outbreak investigations of clusters of TST conversions consisted of several components: (1) surveillance data were obtained for both active and latent TB cases to estimate the existence, nature, timing, magnitude, and scope of potential outbreaks; (2) medical records were reviewed to document past test results, prior diagnosis or treatment of TB, risk factors, and signs or symptoms of active disease; (3) cases of TST conversions were interviewed for identification of contacts with TB and other risk factors; (4) materials and personnel screening procedures were evaluated; and (5) repeat skin tests were independently placed and read. Since 2003, U.S. Army policy has included testing all soldiers with a TST before deployment, testing again at time of redeployment, and testing a third time 3 to 6 months after returning from deployment (29). In practice, prior to 2003, the predeployment TST was not always performed. All soldiers should have had at least one test documented before deployment, since all U.S. military services test all recruits upon entry. However, the documentation of the TST has been highly variable at initial entry and predeployment, and results are often not available at time of redeployment. All repeat testing was done with Tubersol (Aventis-Pasteur, Swiftwater, PA) only, due to concerns of product variability with Aplisol. A positive TST was defined as an induration of 10 mm or more in diameter, and a conversion was defined as an increase in diameter of 10 mm or more within a 2-year period.
Example Pseudoepidemic Investigation—Afghanistan, 2005 The USACHPPM was called to Afghanistan in 2005 to investigate a high reported risk of conversion in an aviation unit. This unit reported that 30 of 198 soldiers (15%) had TST conversions during deployment. No active TB cases were identified despite active surveillance among local and third-country nationals and military forces on the installation. Using strict quality-control measures and exclusive use of Tubersol, the unit was retested. Of 16 of the previously identified positives, 13 (81%) had a negative repeat TST. The revised estimate of conversion risk in the unit was 16 of 374 (4.3%). This was considered a conservative estimate of risk, because 3 of the 16 remaining converters reported previous similar (albeit undocumented) reactions to the TST, and 3 were foreign born. This investigation demonstrated that false positives were associated with variability in the reading of the TST, with improper documentation of previous TST status, and with the use of Aplisol.
Other Pseudoepidemics of TST Conversions in U.S. Army Forces, 1983–2005
Figure 1 compares the risk of TST conversion in documented pseudoepidemics after recent Army deployments estimated before and after investigation with long-term civilian travelers (39) and Peace Corps volunteers (40). Most military operations entail less exposure to local nationals (and thus to TB) than is experienced by Peace Corps volunteers or other long-term civilian travelers. Therefore, the finding of a higher risk of TST conversion in Army units further suggests that many of the remaining conversions seen in the deployed Army populations were false positives.
Pseudoepidemics of TST conversions are a common occurrence after Army deployments and in Army populations. This report describes several of these pseudoepidemics, which were reported independently by U.S. Army units from 1983 to 2005. False-positive TST conversions may result from many factors, as demonstrated by the pseudoepidemics described in this report. These included the following: error and variability in skin test administration, reading, interpretation, and documentation; product variability, particularly the use of Aplisol; and cross-reactivity to NTM. Most of the pseudoepidemics reported here included several of these factors, emphasizing the difficulties with such a large, complex program in a deployed setting. These pseudoepidemics cause confusion among medical providers, public health personnel, and in the Army and civilian communities at large. The testing of this predominantly low-risk population leads to false-positive results in individuals and pseudoepidemics of false-positive TST conversions in Army populations. Due to the variability and error inherent to the test and the testing procedures in practice, the CDC currently recommends against testing low-risk populations. Most of the pseudoepidemics of TST conversions reported here were the result of universal (rather than targeted) testing in a low-risk Army population, leading to a significant number of false positives. Further problems arise due to the magnitude of the testing program, with over 500,000 TSTs administered each year before and after deployment. This program may result in quality being sacrificed for quantity, and the inability to identify or target those truly at increased risk for the infection. False-positive TST conversions may be minimized in several ways. The most important of these in an Army population is testing only those at higher risk for the infection (i.e., performing targeted testing). Other important ways include reducing sources of error in TST administration, reading, and documentation throughout the service member's career. Proper training, supervision, and oversight of medical personnel involved in the TB control program are critical to accurate assessment of conversion status. This includes all phases of the skin testing process, from TST placement, reading, and documentation, to interpretation of results and follow-up care. Due to the problems associated with use of Aplisol, the exclusive use of Tubersol is recommended in low-risk populations such as the U.S. military. Documentation of manufacturer and lot number is warranted in all situations to facilitate better follow-up care as well as for investigation of outbreaks, if necessary (10). Finally, cross-reactivity to NTM should be considered in Army populations. Army service members may be at higher risk for sensitization to NTM because of the presence of a large proportion of Army bases in the southeast United States, the potential for increased exposure to NTM overseas, and from soil contact during training and deployment (41–43). Furthermore, there is little evidence that universal skin testing is more effective than targeted testing in identifying or preventing cases of active TB. In fact, there is evidence that the focus on universal testing may detract from the identification of active TB, as demonstrated by the outbreak on the USS Wasp in 1998 (44). In this outbreak, the sailor who was the index case had a delayed diagnosis of active TB due to prior testing that was first positive, then negative on retesting and thus not treated. Because the sailor was untreated while symptomatic for 3 months, the delay in diagnosis resulted in 21 cases of active TB. This emphasizes the need for constant vigilance and consideration of the diagnosis of active TB in military and other populations.
The IFN- There are several limitations to this report. The most important of these is the lack of certainty regarding previous skin test results, due to variability in prior skin test administration, reading, and documentation. Because many of these pseudoepidemics included several factors associated with false-positive results, it is difficult to assess the independent contributions of these factors. Because most of these outbreaks occurred in the deployed setting, there were limitations in the interventions and investigations that were possible due to operational, resource, and time constraints. In particular, complete medical records are not available during deployment; they are typically kept at home station. This probably resulted in the misclassification of prior (predeployment) TST status as negative, suggesting falsely elevated subsequent estimates of residual conversion risk. Repeat testing results may not have been more accurate than the original tests, and the assumption that they were may have resulted in falsely lowered estimates of conversion risk after investigation. Despite these limitations, the consistent reporting of pseudoepidemics in Army populations with similar associated factors over a large time interval suggests that the findings are valid. The absence of active TB cases resulting from any of these pseudoepidemics further supports this conclusion. U.S. Army forces generally have a low risk of TB infection resulting from deployments, and moving to targeted rather than universal testing of this population should be considered. Targeted testing has been the recommended strategy for the prevention of LTBI in the United States since 2000 (2). The U.S. Air Force implemented a targeted program for redeploying air personnel using a questionnaire in 2005 (50), but it has not been formally evaluated and was not adopted by the other services. Future efforts in TB prevention in the U.S. Army should include targeted testing for LTBI, with increased emphasis on proper quality-control measures and follow-up procedures for those identified as high risk. Early identification of cases of active TB has also been shown to be critical in preventing TB in military populations (44). Regardless of the testing strategy used, any testing program must be designed and conducted using strategies to minimize false positives and future pseudoepidemics. Finally, close attention should be paid to the surveillance and control of infectious diseases among local nationals and detainees working or residing in contact with U.S. service members in deployed locations.
The opinions or assertions contained herein are the private views of the authors and are not to be construed as official, or as reflecting true views of the Department of the Army or the Department of Defense. Originally Published in Press as DOI: 10.1164/rccm.200802-223OC on March 20, 2008 Conflict of Interest Statement: None of the authors has a financial relationship with a commercial entity that has an interest in the subject of this manuscript. Received in original form February 5, 2008; accepted in final form March 17, 2008
This article has been cited by other articles:
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
TOP
ABSTRACT
AT A GLANCE COMMENTARY
release assay (IGRA) may have utility in Army populations, due to its reportedly increased specificity for LTBI (
