Published ahead of print on January 19, 2006, doi:10.1164/rccm.200510-1563OC
© 2006 American Thoracic Society doi: 10.1164/rccm.200510-1563OC
The Scope and Impact of Treatment of Latent Tuberculosis Infection in the United States and CanadaDivision of Infectious Diseases and Center for Health Services Research, Department of Medicine, Vanderbilt University Medical Center, Nashville, Tennessee; Westsat, Rockville, Maryland; Division of Tuberculosis Elimination, Centers for Disease Control and Prevention, Atlanta, Georgia; Denver Public Health and Hospitals Authority, Denver, Colorado; and Department of Epidemiology, Boston University School of Public Health, Boston, Massachusetts Correspondence and requests for reprints should be addressed to Timothy R. Sterling, M.D., A2209 Medical Center North, 1161 21st Avenue South, Nashville, TN. E-mail: timothy.sterling{at}vanderbilt.edu
Rationale: The scope of treatment of latent tuberculosis infection (LTBI) in the United States and Canada is unknown. Identifying the types of clinics that administer such treatment and patients who receive it could guide resource utilization and improve treatment initiation and completion. Objectives: Estimate the number of persons started on LTBI treatment; describe the types of clinics that treat LTBI.
Methods: The Tuberculosis Epidemiologic Studies Consortium, consisting of 19 United States and 2 Canadian sites, conducted a survey among clinics that initiated LTBI treatment for Results: Study catchment areas from the 19 United States sites represented 8.6% of the United States population and 12.7% of all tuberculosis cases in 2000. An estimated 37,857 patients started LTBI treatment during 2002 at 244 clinics surveyed. Of these treatment starts, 29,970 (79%) occurred at general public health clinics; immigrant/refugee clinics (2,409; 6.4%) and correctional/detention facilities (2,325; 6.1%) were the next most common sites. Based on these data, United States tuberculosis case rates, and United States population data, the estimated total number of LTBI treatment starts in the United States was 291,000–433,000. When the 37,145 persons who initiated LTBI treatment in the United States were extrapolated to the entire United States population, with a 5% lifetime risk of tuberculosis without treatment, and 20–60% treatment effectiveness, approximately 4,000–11,000 tuberculosis cases were prevented in the United States. Conclusions: LTBI treatment is initiated among a substantial number of persons in the United States and Canada, primarily in the public sector. Treatment of LTBI can significantly decrease the tuberculosis burden.
Key Words: Mycobacterium tuberculosis • treatment of latent tuberculosis infection • tuberculosis An estimated 11 million persons (95% confidence interval [CI], 9–14 million) in the United States have tuberculosis (TB) infection and are therefore at risk of progressing to TB disease (1). The Centers for Disease Control and Prevention (CDC), the Advisory Council for the Elimination of Tuberculosis, and the Institute of Medicine of the National Academy of Sciences identify treatment of latent TB infection (LTBI), particularly when targeted toward persons at highest risk of developing TB, as one of the major strategies for elimination of TB in the United States (2–4). However, relatively little is known about the scope of treatment of LTBI in the United States and Canada. We report here the findings of the first phase of a multicenter evaluation of the treatment of LTBI in the United States and Canada, conducted by the Tuberculosis Epidemiologic Studies Consortium (TBESC) of the CDC. The primary objectives were to estimate the number of persons started on treatment for LTBI and to describe the types of clinics and medical providers who prescribe LTBI in the United States and Canada, and thereby aid in efforts to improve treatment initiation and completion.
The Division of Tuberculosis Elimination of the CDC established the TBESC in 2001 to strengthen and coordinate TB research. The consortium consists of 21 sites (19 in the United States and 2 in Canada), each of which represents a partnership between an academic institution and a state or local TB control program. For this study, each site selected its patient catchment area based on convenience and the ability to obtain all of the required information. The study sites and their catchment areas are listed in Table 1.
Each study site was asked to provide a list of clinics and practices in their catchment area that provided LTBI treatment to at least 10 persons in 2002. Strategies to identify these clinics included the following: (1) use of state LTBI reporting and registry; (2) accessing pharmacy records to identify clinics dispensing or prescribing LTBI treatment; (3) asking providers known to treat LTBI to provide names of other clinicians who also treat LTBI; (4) surveying providers by contacting medical societies and posting queries on their websites or newsletters; (5) conducting Internet searches for LTBI treatment or isoniazid; and (6) contacting state and federal facilities, civil surgeons, and health care facilities (e.g., hospitals, nursing homes, other long-term care facilities) that may have LTBI treatment programs. Each site used the strategies that were appropriate and convenient to their site. After clinics were identified and contacted, sites were instructed to collect the following information: clinic name and type (i.e., general public health, HIV, pediatric, immigrant or refugee, private practice, correctional and detention facility, homeless, migrant farm worker, substance abuse), estimated total number of patients treated for LTBI in 2002, and estimated percentage of treated patients who were in the following tuberculosis risk groups: HIV-infected, incarcerated, homeless, migrant farm worker, drug users, Native American/Aboriginals, TB contacts, recently arrived foreign-born persons from countries with the greatest contribution of foreign-born TB burden (i.e., Mexico, the Philippines, Vietnam, India, China, Haiti), and other foreign-born individuals. Patients could have been classified into more than one category (e.g., both HIV-infected and drug user). However, after this information was collected, clinics were counted in only one stratum to avoid counting the patient more than once. The National Center for HIV, STD, and TB Prevention of the CDC determined that the study was a program evaluation and not human subject research as defined by federal regulation. Estimates of the population in the United States catchment areas were obtained from the 2000 U.S. census; the number of United States tuberculosis cases in 2000 was obtained from CDC surveillance data (5). Estimates of the population in Canadian catchment areas were obtained from the 2001 Statistics Canada website; the numbers of Canadian cases in 2000 were obtained from site reports and the Tuberculosis in Canada 2000 report (6, 7). Estimates of the number of persons initiating LTBI treatment throughout the United States were obtained by multiplying the reported number of persons treated for LTBI in the United States study catchment areas by two methods: (1) the ratio of the United States population/population at United States study catchment areas and (2) the ratio of the number of United States TB cases/TB cases at United States study catchment areas. The estimated standard deviation of these estimates was calculated.
As shown in Table 1, the catchment area for all study sites had a total population of 27,073,654. The patient catchment areas for the United States sites, which included both urban and rural jurisdictions, had a total population in 2000 of 24,143,398, or 8.6% of the overall United States population of 281,421,906 (2000 U.S. census). Of the 2,439 TB cases at all study sites, the United States sites reported 2,087 cases of TB in 2000, or 12.7% of the United States total of 16,377 (5). The number of TB cases, LTBI treatment starts, and corresponding rates are also provided in Table 1. A comparison of the demographic characteristics of the population in the United States catchment areas versus the entire United States population is shown in Table 2.
Table 3 summarizes the data provided by the participating TBESC sites. A total of 358 clinics that treated patients for LTBI were identified, of which 244 (68%) provided data on the number of patients with LTBI treated; these clinics were generally not university-based. The total number of patients with LTBI treated by the 244 clinics in 2002 was 37,857. The majority of patients with LTBI (79%) were seen in general public health clinics, which accounted for 52% (126/244) of the reporting clinics. Immigrant/refugee clinics accounted for 6.4% and jail or prison clinics for 6.1% of all patients treated. HIV, pediatric, and substance abuse clinics accounted for less than 1% of patients treated. About 2.1% of the patients started treatment in private practice clinics; however, only 48 of 121 (40%) private practice clinics identified reported the number of patients with LTBI treated there. Of the 37,857 patients who were estimated to have started treatment of LTBI, 37,145 were in the United States and 712 were in Canada.
Two hundred of the clinics provided detailed data on the percentage of patients in each risk group, reporting on 34,644 patients with LTBI (91.5% of the patients reported at all clinics surveyed). As shown in Table 4, foreign-born individuals accounted for 56.2% of persons receiving treatment; 13.5% were close contacts of TB cases, and 12.3% were incarcerated. In Table 5, the proportion of patients seen at each type of clinic with specific TB risk factors is provided. For example, of the 27,339 patients treated in public health clinics and for whom such data were available, 55.2% were foreign born and 14.9% were close contacts of TB cases. Of the 365 patients treated at clinics that serve the homeless, 57.8% were foreign born and 69% were close contacts.
Persons of foreign birth constituted a substantial proportion of LTBI treatment patients at every type of clinic. Persons of Mexican birth comprised more than 25% of patients with LTBI in correctional and detention facility clinics and clinics serving the homeless, and approximately 15% of patients at drug or alcohol clinics. Although almost 30% of patients with LTBI in private practice were foreign born, only 6.0% were of Mexican birth. Two methods were used to estimate the total number of persons treated for LTBI in the United States in 2002 based on our data. First, based on the population represented by the United States study catchment areas (24,143,398) and the total United States population in 2000 (281,421,906—from the 2000 U.S. census), it was estimated that the 37,145 persons treated for LTBI in the study's catchment areas would correspond to approximately 433,000 persons treated throughout the United States; the estimated standard deviation of this estimate is 279,000. Second, based on the number of TB cases in the United States catchment areas (2,087) and throughout the United States (16,377) in 2000, it was estimated that there were 291,000 persons treated throughout the United States; the estimated standard deviation of this estimate is 188,000. Because only two Canadian sites participated in this study, extrapolations were not made to estimate LTBI treatment throughout Canada. The risk of developing TB varies according to patient risk group (e.g., close contact of TB case, underlying diabetes mellitus, or malignancy). However, based on the published literature and adjusted for the age distribution of the United States population in 2000, the overall lifetime risk of TB progressing from LTBI in the United States is 5.1% (95% CI, 3.1–8.4%) (8). Among the 37,145 persons who were treated for LTBI in the United States catchment areas in this study, 1,894 cases of TB would have occurred without treatment. Given the 69% (6-mo course) to 93% (12-mo course) efficacy of isoniazid in preventing TB (4, 9), and the 30 to 64% completion rate of isoniazid in clinic settings (4, 8, 10, 11), the effectiveness of isoniazid (which takes into account efficacy and adherence) would be approximately 20 to 60%. Thus, approximately 400 to 1,100 cases of TB were prevented among the reported LTBI treatment patients. Because the United States component of the study population represented approximately 10% of the United States population and United States TB burden, the number of TB cases prevented by LTBI treatment throughout the United States was approximately 10-fold greater, or 4,000 to 11,000 cases.
This is the largest study conducted to date to describe a representative sample of the overall population initiating treatment of LTBI in the United States and Canada. Study catchment areas in the United States accounted for 8.6% of the United States population and 12.7% of United States tuberculosis cases. Among the entire United States and Canadian study population, there were 37,857 persons treated for LTBI during 2002. Of note, almost 80% of these patients were treated at public health clinics. The other two most common sites at which patients received treatment were refugee or immigrant clinics and correctional and detention facility clinics. Together, these three clinic settings accounted for 92% of all patients reported to have received LTBI treatment. Although persons receiving treatment in a private practice setting accounted for only 2.1% of the study population, it should be noted that study sites were asked to identify only clinics that treated at least 10 persons/yr. Because there could be many clinics and small practices that treated only a few patients with LTBI each year, the proportion of persons receiving treatment of LTBI in the private practice setting was likely underestimated. In addition, only 40% of private practice clinics identified responded to queries from study sites. Of note, only 0.1% of persons treated for LTBI were in drug or alcohol treatment centers. This patient group has a high prevalence of LTBI, though is not necessarily at increased risk of progressing to disease. Because treatment of LTBI can be incorporated into substance abuse treatment relatively easily, such as methadone maintenance, efforts should be made to increase LTBI treatment in this setting. It is notable that 56% of persons treated for LTBI were foreign born. This is consistent with the epidemiology of TB in the United States and Canada, in which more than 50% of cases of active disease are among foreign-born individuals (12). The proportion of foreign-born persons with LTBI who had arrived in the United States and Canada within the previous 5 yr and were therefore at the highest risk of progressing to active disease (4) was not determined in this study. The two methods used to estimate the number of persons in the United States that initiated treatment of LTBI in 2002 resulted in two different estimates: 433,000, if based on population in the patient catchment areas, and 291,000, if based on TB cases. Of the two methods, the estimate based on the number of TB cases is probably more reliable, because the basis for such ratio estimates is correlation, and the geographic correlation between persons with LTBI and TB cases is probably stronger than the correlation between LTBI and the general population. Several limitations of this study should be noted. First, the catchment areas were not randomly selected but rather were convenience samples at sites that participate in a TB research consortium. In addition, the United States study sites represented only approximately 10% of the entire United States population and TB case burden. Although this could affect the generalizability of the results, the study sites represented a reasonable cross-section of the United States: urban and rural, as well as high- and low-TB burden areas. Extrapolations for Canada are more limited because only two Canadian sites participated. Second, the extent to which data reported by clinics to the study sites were estimates versus actual rates is unknown. However, the clinics that reported data in this study are part of a more intensive follow-up study in which a random sample of the clinics was selected for detailed chart reviews. In the sampled clinics, the charts of all persons treated for LTBI were identified; a subset of all identified charts was then selected for data abstraction. We have validated the data reported by 25 clinics via this process. These validation data suggest a slight tendency for clinics to overreport LTBI cases. The average number of LTBI cases reported in the data used in this article was 281.6, whereas the average number found in onsite visits was 246.2. The average discrepancy between reported and validated cases was 35.4, or approximately 14% of the validated average cases; the median discrepancy was 24.0, about 10% of the average validated cases. Third, the extent to which all study sites applied the suggested strategies to identify clinics that provided LTBI treatment is unknown. There was no verification that sites identified all locations that treated at least 10 persons for LTBI. We may have underestimated the total number of persons treated for LTBI because we did not adjust for sites that did not respond to the survey. However, the degree of underestimation was likely low because the clinics that did not respond were small and presumably did not treat many patients. Fourth, sites were asked to list clinics that offer LTBI treatment; however, it is uncertain whether the clinics contacted had a record-keeping system that could identify all persons who received LTBI treatment. Finally, estimates regarding the number of TB cases prevented were based on the assumption that isoniazid is 69 to 93% efficacious and completed by 30 to 64% of persons who initiate it. Although the isoniazid completion rate was based on a summary of available data from clinic and operational settings (and not clinical trials, which have higher completion rates) (4, 8, 10, 11), even lower completion rates have been reported in some field studies (13, 14). Lower isoniazid completion rates would result in lower effectiveness and therefore fewer cases of TB prevented. In addition, if the risk of developing TB without treatment was less than the 5% estimate used, this would also result in fewer TB cases prevented by LTBI treatment. This study describes for the first time the scope of treatment of LTBI in the United States and Canada. The overwhelming majority of persons who initiated treatment were treated in the public health sector. Treatment of LTBI was initiated among a large number of persons in the United States and Canada, and likely prevented a substantial number of TB cases. Treatment of LTBI can have a significant impact on decreasing the burden of TB.
Protocol Team: C. Robert Horsburgh, Jr., M.D., M.U.S. Protocol Chair, Boston University School of Public Health (Boston, MA); Stefan Goldberg, M.D., Primary CDC Contact, Co-Project Officer, Division of TB Elimination, NCHSTP/CDC (Atlanta, GA); Robin Shrestha-Kuwahara, M.P.H., Co-Project Officer, Division of TB Elimination, NCHSTP/CDC (Atlanta, GA); Jim Bethel, Ph.D., Westat (Rockville, MD); Henry M. Blumberg, M.D., Emory University School of Medicine (Atlanta, GA); Paul Colson, Ph.D., Charles P. Felton National TB Center at Harlem Hospital (New York, NY); Carol Dukes-Hamilton, M.D., Research Triangle Institute, NC Department of Health and Human Services (Durham, NC); Eduard Eduardo, M.P.H., Westat (Atlanta, GA); Yael Hirsch-Moverman, M.P.H., Charles P. Felton National TB Center at Harlem Hospital (New York, NY); Stephen E. Hughes, Ph.D., New York State Department of Health (Albany, NY); Heather Joseph, M.P.H., Division of TB Elimination, NCHSTP/CDC (Atlanta, GA); Rick O'Brien, M.D., Foundation for Innovative New Diagnostics (Geneva, Switzerland); Barbara Roche, B.Sc.N., M.P.H., University of Manitoba (Winnipeg, Manitoba, Canada); Nandini Selvam, M.P.H., University of Medicine and Dentistry of New Jersey (Newark, NJ); Timothy R. Sterling, M.D., Vanderbilt University Medical Center (Nashville, TN); Paul Weinfurter, M.P.H., Westat (Atlanta, GA); Lourdes Yun, M.D., M.P.H., Denver Public Health and Hospitals Authority (Denver, CO). Drs. Sterling and Bethel had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. Representatives of the sponsor (CDC) participated in the design and conduct of the study; collection, management, and interpretation of the data; and preparation, review, and approval of the manuscript.
This study was supported by the Tuberculosis Epidemiologic Studies Consortium, Centers for Disease Control and Prevention. Originally Published in Press as DOI: 10.1164/rccm.200510-1563OC on January 19, 2006 Conflict of Interest Statement: None of the authors have a financial relationship with a commercial entity that has an interest in the subject of this manuscript. Received in original form October 4, 2005; accepted in final form January 19, 2006
This article has been cited by other articles:
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
TOP
ABSTRACT
METHODS
10 patients in 2002. 
