Enhancement of Treatment Completion for Latent Tuberculosis Infection With 4 Months of Rifampin*

CHEST Enhancement of Treatment Completion for Latent Tuberculosis Infection With 4 Months of Rifampin* Alfred Lardizabal, MD; Marian Passannante, PhD; Faysal Kojakali, MPH; Christopher Hayden, BA; and Lee B. Reichman, MD, MPH, FCCP Original Research TUBERCULOSIS Background: Isoniazid is the standard medication used to treat latent tuberculosis infection (LTBI). The lengthy treatment with isoniazid, its perceived hepatotoxicity, and the increasing influx of foreign-born persons from countries with a higher prevalence of isoniazid resistance have compromised this regimen. In 2000, the Centers for Disease Control and Prevention guidelines recommended 4 months of rifampin (4R) as an acceptable alternative regimen to 9 months of isoniazid (9H). In a county chest clinic in northern New Jersey, a self-administered 9H regimen for patients with LTBI was generally prescribed until the year 2002. After recognizing poor completion rates, LTBI treatment was shifted predominantly to the alternative 4R regimen. Methods: Medical records of patients placed on LTBI treatment during 2000 (predominantly a 9H regimen) and 2003 (predominantly a 4R regimen) were reviewed. A total of 474 patients were included in the study. 2, Fishers exact, two-sample t, and Wilcoxon rank-sum tests and logistic regression were used to analyze the data. The main outcome variable was treatment completion. Results: A total of 80.5% of patients receiving 4R and 53.1% receiving 9H completed treatment (p < 0.0001); 34.7% of patients receiving 9H were unavailable for follow-up, compared to 12.6% receiving 4R (p <0.0001). Fewer drug reactions were observed in the group receiving 4R compared to the group receiving 9H (3.1% vs 5.8%), although this was not statistically significant. Logistic regression analysis identified treatment regimen as a significant predictor for treatment completion (odds ratio [OR], 5.1; 95% confidence interval [CI], 3.3 to 8.1). Employment was negatively associated with treatment completion in the same model (OR, 0.54; 95% CI, 0.34 to 0.94). Conclusions: Patients receiving 4R were significantly more likely to complete therapy than those receiving 9H. (CHEST 2006; 130:1712 1717) Key words: isoniazid; latent tuberculosis infection treatment; rifampin Abbreviations: ATS American Thoracic Society; CDC Centers of Disease Control and Prevention; CI confidence interval; IDSA Infectious Disease Society of America; LTBI latent tuberculosis infection; OR odds ratio; TB tuberculosis; 2RZ 2 months of rifampin and pyrazinamide; 4R 4 months of rifampin; 6H 6 months of isoniazid; 9H 9 months of isoniazid In 2004, a total of 14,511 cases of tuberculosis (TB) in the United States were reported to the Centers of Disease Control and Prevention (CDC), the lowest number since national reporting was instituted in *From the Global Tuberculosis Institute (Drs. Lardizabal and Reichman, and Mr. Hayden), School of Public Health (Mr. Kojakali), and the Department of Preventive Medicine and Community Health (Dr. Passannante), New Jersey Medical School, Newark, NJ. Drs. Lardizabal and Reichman are supported by the Regional Tuberculosis Training and Medical Consultation Center grant, Centers for Disease Control and Prevention. 1953. 1 While maintaining efforts to effectively manage TB cases and contacts, the CDC and the Institutes of Medicine have called for increased treat- None of the authors have any conflicts of interest to disclose. Manuscript received June 5, 2006; revision accepted September 2, 2006. Reproduction of this article is prohibited without written permission from the American College of Chest Physicians (www.chestjournal. org/misc/reprints.shtml). Correspondence to: Alfred A. Lardizabal, MD, 225 Warren St, Second Floor, Newark, NJ 07103; e-mail: lardizaa@umdnj.edu DOI: 10.1378/chest.130.6.1712 1712 Original Research

ment of high-risk persons with latent TB infection (LTBI) in order to accelerate the decline toward TB elimination. 2,3 Isoniazid has been the mainstay of LTBI treatment for 40 years. Unfortunately, the recommended lengthy (9 months) treatment with isonia- For editorial comment see page 1638 zid, 2 its reputation for toxicity, 4 and the increasing influx of foreign-born persons from countries with a high prevalence of isoniazid resistance 5 have compromised this regimen as an effective tool. Indeed, it has been recognized that many persons fail to complete the full course of LTBI treatment and remain at risk for TB. 2,6 Nationwide, even among infected contacts of sputum-positive TB patients in 2003, only 73% were placed on LTBI treatment; of these, only 59% completed therapy, resulting in a Dear Colleague letter from the CDC to call attention to this situation. 7 In the American Thoracic Society (ATS), CDC, and Infectious Disease Society of America (IDSA) official joint statement published in 2000, 6 months of isoniazid (6H), 4 months of rifampin (4R), and 2 months of rifampin and pyrazinamide (2RZ) were recommended as acceptable alternatives to the preferred 9 months of isoniazid (9H). 2 After the recommendations were published, because of severe toxicity and mortality among patients receiving 2RZ, recommendations were revised to indicate that this regimen should not generally be offered. 8 A clinical review of LTBI by Jasmer and colleagues 9 discusses the importance of diagnosing and treating LTBI; the authors mention that adverse reactions to rifampin are rare. However, the discussion about its role in the treatment of LTBI is brief. In an opinion piece, 10 we have offered several reasons for considering the expanded use of the 4R regimen in selected populations, based on the likelihood of high efficacy, its low toxicity, improved adherence, higher levels of isoniazid resistance among certain immigrant groups, as well as cost and site-specific considerations. In this present communication, we present our experience with this regimen, suggesting that it might go a long way in ameliorate the dismal LTBI adherence rates cited by the CDC. 7 Materials and Methods At a county chest clinic in central New Jersey, 9H was routinely prescribed for all LTBI patients until the year 2002. Due to poor adherence and completion rates, we began discussing other treatment options with patients, considering the length and complexity of the regimens, possible adverse effects, and potential drug interactions as recommended in the 2000 ATS/CDC/ IDSA guidelines. 2 Discussions of treatment options with each patient were made by the same physician (A.A.L.) with each patient, emphasizing 9H as the standard treatment and 4R as an acceptable alternative. During these discussions, we found that patients overwhelmingly requested the shorter regimen. By the end of 2002, 4R was the predominant regimen used to treat LTBI, unless contraindicated. Since the clinical staff and follow-up procedures have remained constant since 2000, we decided to review medical records of patients referred for LTBI evaluation during 2000 and 2003 in an effort to determine adherence with and acceptability of 4R compared with 9H. These two specific years were chosen because they were close enough not to present changes in staffing or clinic administration and far enough apart to represent distinct time periods when the vast majority of patients received only one of the treatments. Following published LTBI guidelines, 2 patients with LTBI during both study periods who had an increased risk for progression to disease were recommended to start treatment regardless of age. Baseline liver function tests were performed only if the patients had increased risk for hepatotoxicity (HIV infection, history of hepatitis or chronic liver disease, excess alcohol use, pregnancy). For both cohorts, per usual clinic practice, the nurses monitored each patient for adverse reactions and adherence during each face-to-face visit before the prescription refill each month. Adherence was judged by pill counts in returned bottles plus extensive discussions with the nurse and case manager to assess compliance and the fact that the patients continued to present themselves for the next dose. Blood work was ordered when the patient presented with complaints during these visits. This is an historical prospective review of patients treated for LTBI at a county chest clinic. An index file has always been maintained at the clinic of all patients referred for LTBI evaluation. For our study, patients referred during 2000 and 2003 were identified and their medical records were reviewed. Selected data items were abstracted onto a form, including demographic information, TB risk factors, and treatmentrelated variables such as adverse reactions and completion of therapy. For patients prescribed 9H, completion of therapy was defined as receiving enough medications to complete 270 doses within 12 months, although patients completing a minimum of 180 doses (6H) were also counted as completed. For patients prescribed 4R, patients completed therapy after receiving 120 doses within 6 months. Approval from the University of Medicine and Dentistry of New Jersey institutional review board including a waiver for informed consent was obtained prior to the initiation of this study. Data Analysis The collected data were entered into a password-protected computer and analyzed using the JMP statistical software package (SAS Institute; Cary, NC) to compare patient characteristics and treatment outcomes with each drug regimen. Significance tests were conducted at the 0.05 level. The main outcome measured was treatment completion. This was measured as a dichotomous outcome. The secondary study outcome was adverse drug reactions requiring cessation of treatment. 2 and Fisher exact tests were used for comparisons of categorical variables. Two-sample t tests and Wilcoxon rank-sum tests were used for comparison of continuous and nonnormally distributed data, respectively. Logistic regression was used to assess the independent contribution of independent variables on treatment completion. www.chestjournal.org CHEST / 130 / 6/ DECEMBER, 2006 1713

Results Five hundred fifty-four patients referred for LTBI evaluation were identified. Patients with class 3 TB (current disease; n 4) and class 4 TB (old healed, untreated disease; n 16) were excluded from the study, as were 15 patients for whom LTBI treatment was not recommended. Patients with class 4 TB were excluded because these patients are started with multiple drugs until the diagnosis is confirmed with negative culture results. Of the 519 patients started on LTBI treatment, 22 who were prescribed specifically 6H were excluded so as not to confound the analysis comparing 9H and 4R, as were 23 patients who had moved to other areas (n 23) to continue treatment. The remaining 474 LTBI patients were included in the study. Table 1 characterizes our study population during the two time periods, 2000 (n 195) and 2003 (n 279). For both time periods, the majority of patients evaluated ( 90%) were foreign born, although the 2003 population was significantly more likely to be somewhat older, Hispanic, employed, and residing in the United States for 5 years. Approximately 99% of LTBI patients in the 2000 period were prescribed isoniazid, and 93% of LTBI patients in 2003 were prescribed rifampin. Table 1 Description of the Study Population, 2000 and 2003* Demographic and Risk Factors Year 2000 Year 2003 p Value Age group, yr 1 24 171 (87.7) 177 (63.4) 0.0001 2 25 34 13 (6.7) 35 (12.5) 35 44 7 (3.6) 42 (15.1) 45 4 (2.1) 25 (9.0) Gender Male 111 (56.9) 146 (52.3) 0.35 F Female 84 (43.1) 133 (47.7) Race White 22 (11.8) 18 (7.1) 0.04 2 Black 13 (7.0) 22 (8.7) Hispanic 90 (48.1) 150 (59.5) Other 62 (33.2) 62 (24.6) Employment Employed 24 (12.3) 94 (33.7) 0.0001 F Unemployed 171 (87.7) 185 (66.3) Country of origin United States 12 (6.2) 28 (10.1) 0.14 F Foreign 183 (93.9) 249 (89.9) Lived in United States US born 12 (6.2) 28 (10.1) 0.0002 2 5 yr 161 (82.6) 182 (65.7) 5 yr 22 (11.3) 67 (24.2) Regimen 4R 1 (0.5) 260 (93.2) 0.0001 F 9H 194 (99.5) 19 (6.8) *Data are presented as No. (%). F Fisher exact test. Potential predictors of treatment completion are listed in Table 2. The number of risk factors variable was created by summing up the number of medical (ie, close contact, immunosuppression, skin test conversion, injection drug use, chronic medical conditions, silicosis, malnutrition, organ transplant history) and population risks (ie, prison/jail inmate or employee, long-term facility resident/employee, healthcare employee, homeless shelter resident/employee, foreign born in the United States 5 years, migrant worker) for development of TB identified for each patient, with a possible high score of 4 and a possible low score of 0. The total risk scores ranged from 0 to 3, with 76% having exactly one risk factor. Age was analyzed as a categorical variable with the two older groups (35 to 44 years and 45 years) separated to see if adverse drug reactions in these group might have an impact on treatment completion. The data in Table 2 illustrate that only two variables were significantly associated with treatment completion: treatment regimen and years in the United States; 80.5% of those receiving the 4R regimen completed treatment vs only 53.1% of those Table 2 Bivariate Analysis of Predictors of Treatment Completion* Demographic and Risk Factors Completed Treatment Not Completed p Value Age group, yr 1 24 244 (70.1) 104 (29.9) 0.22 2 25 34 32 (66.7) 16 (33.3) 35 44 27 (55.1) 22 (44.9) 45 20 (69.0) 9 (31.0) Gender Male 173 (67.3) 84 (32.7) 0.69 F Female 150 (69.1) 67 (30.9) Race White 27 (67.5) 13 (32.5) 0.87 2 Black 26 (74.3) 9 (25.7) Hispanic 163 (67.9) 77 (32.1) Other 87 (70.2) 37 (29.8) Employment Employed 72 (61.0) 46 (39.0) 0.07 F Not employed 251 (70.5) 105 (29.5) Country of origin United States 25 (62.5) 15 (37.5) 0.48 F Foreign 298 (69.0) 134 (31.0) Length of time in the United States US born 25 (62.5) 15 (37.5) 0.04 2 5 yr 52 (58.4) 37 (41.6) 5 yr 246 (71.7) 97 (28.3) TB risk factors 1.0 0.48 1.0 0.62 0.67 W Regimen 4R 210 (80.5) 51 (19.5) 0.0001 F 9H 113 (53.15) 100 (47.0) *Data are presented as No. (% of row) unless otherwise indicated. W Wilcoxon signed rank. See Table 1 for expansion of abbreviation. 1714 Original Research

receiving 9H (p 0.0001). Foreign-born persons in the United States for 5 years were significantly more likely to complete treatment (71.7%) than US-born persons (62.5%) or foreign-born persons in the United States for 5 years (58.3%) [p 0.039]. Fewer employed persons completed therapy (61.0%) compared to those who were unemployed (71.0%), but this was only marginally significant (children and students were counted as unemployed). Three independent variables were included in the logistic regression model (Table 3): treatment regimen, employment status, and time in the United States ( 5 years or 5 years). After controlling for the other independent variables, the logistic regression model identified a significant association between receiving the 4R regimen and treatment completion (p 0.0001), with an odds ratio (OR) of 5.1 (95% confidence interval [CI], 3.3 to 8.1). Employment was negatively associated with treatment completion (p 0.03). Time in the United States was entered into the model as dummy variables with US-born patients as the comparison group. Being in the United States for 5 years compared with US-born patients was marginally associated with treatment completion (p 0.09). Foreign-born patients who had lived in the United States for 5 years were not more likely to complete treatment when compared to US-born individuals with LTBI (p 0.88). Table 4 lists the various reasons for cessation of therapy for each treatment regimen. More patients receiving 9H discontinued therapy due to complaints of side effects and drug reactions (13 of 213 patients; 6.1%) compared with patients receiving 4R (8 of 261 patients; 3.1%), although the difference was not statistically significant (p 0.12, Fisher exact test). Three cases of drug-induced hepatitis were diagnosed after patients presented with complaints of side effects in the 9H group and no cases in the 4R group. Overall, adverse events requiring cessation of therapy occurred in 21 of 474 patients. In order to see if there was an association between treatment regimen and adverse reaction that might vary by age group, a Cochran-Mantel-Haenszel 2 test was conducted. This test did not identify any significant Table 4 Reason for Treatment Cessation* Reasons 4R (n 261) 9H (n 213) Treatment completed 210 (80.5) 113 (53.1) Patient stopped against advice 10 (3.8) 11 (5.2) Rash/allergy 4 (1.5) 5 (2.4) Headache/flu-like symptoms 3 (1.2) 2 (0.94) Hepatitis 0 3 (1.4) Abdominal discomfort 1 (0.38) 3 (1.41) Unavailable 33 (12.6) 74 (34.7) Diagnosis of active TB 0 1 (0.47) Other medical advice 0 1 (0.47) *Data are presented as No. (% of column). association within each of four age groupings: 24 years, 25 to 34 years, 35 to 44 years, and 45 years (p 0. 06). The major reason for treatment cessation was having the patient be unavailable for follow-up during treatment (n 107; 22.6% overall). Significantly more patients were unavailable in the 9H group (34.7%) compared with the 4R group (12.6%) [p 0.0001, Fisher exact test]. For the 9H cohort, most patients were unavailable for follow-up between months 1 and 3. For the 4R group, most patients were unavailable after the first month, signifying that there was really no intention to complete therapy after complying with requirements to rule out TB disease. Discussion Using logistic regression analysis, this study showed that a significantly larger proportion of patients receiving the 4R regimen completed LTBI treatment than patients receiving 9H (OR, 5.1; 95% CI, 3.3 to 8.1). The proportion unavailable for follow-up during treatment was nearly three times higher among those receiving 9H compared with those receiving 4R and would likely have been higher had we not allowed those who completed 6 months of INH to be counted as completed. Being employed was negatively associated with treatment completion. This is consistent with our observation that a patient s need to hold on to a job is often more important than keeping an appointment with the Table 3 Logistic Regression Analysis: Treatment Completion Parameters Estimate SE 2 p Value OR Lower 95% CI Upper 95% CI Intercept 0.3139 0.3830 0.67 0.41 0.54 0.31 0.94 Employed 0.6099 0.2794 4.76 0.03 Foreign born 5 yr vs US born 0.6549 0.3834 2.92 0.09 1.92 0.89 4.06 Foreign born 5 yr vs US born 0.0639 0.4040 0.02 0.88 0.94 0.40 2.19 Rifampin vs isoniazid 1.6283 0.2330 48.85 0.0001 5.10 3.26 8.14 www.chestjournal.org CHEST / 130 / 6/ DECEMBER, 2006 1715

clinic. None of the demographic factors of the two groups were significant predictors of treatment completion. No significant difference was observed in the occurrence of adverse drug reactions in the two groups. For both regimens, none of the adverse reactions were serious or life threatening, although a smaller proportion of those receiving 4R had drug reactions. The results of this study are consistent with a randomized trial 11 that reported significantly better treatment completion with 4R (86%) compared to 9H (62%), at a significantly lower cost. This study also reported fewer adverse drug reactions in the 4R group, although this was not statistically significant. The principal limitation of our study is that it was not a randomized or blinded trial and that our control group (9H) was from an earlier time period. We recognize that since this was an historical prospective study with historical control subjects, potential unmeasured confounding variables may have affected the results. However, since the treating physician and nurse case manager and nursing staff were the same for both study periods, it is unlikely that the significant differences in completion rates could be ascribed to the caregivers and their management approach. In addition, there were no changes in clinic visit procedures during these time periods. Finally, since the difference in adherence rates were so highly significant, to levels not often if ever seen in self-administered isoniazid therapy for LTBI, we believed that despite the limitations the results were important enough to report now without waiting to conduct a randomized clinical trial. Much of the concern regarding the use of the 4R regimen for treatment of LTBI centers on the paucity of data regarding its efficacy. In an experimental model using mice, Lecoeur and coworkers 12 suggested that 2RZ and 3R were both more effective than 6H. Likewise, the efficacy of rifampin containing preventive chemotherapy (rifampin, rifampinisoniazid, rifampin-pyrazinamide, rifampin-isoniazid-pyrazinamide) was demonstrated by Dhillon et al 13 using Cornell model mice. Only one randomized clinical trial 14 in human subjects directly compared the efficacy of rifampin alone against treatment with isoniazid in the treatment of LTBI in patients with silicosis. The 3 months of rifampin regimen had an efficacy of 63%, compared to 48% for the 6H arm. 14 Other nonrandomized clinical studies 15,16 have suggested the efficacy of rifampin (using a 6-month regimen of rifampin) in preventing TB disease among patients with LTBI who had been exposed to isoniazid-resistant TB cases. Although the American Academy of Pediatrics also recommends 6 months of rifampin in the treatment of children with LTBI who have been exposed to a patient with isoniazidresistant TB, 17 this regimen is not recommended in the current ATS/CDC/IDSA guidelines, 2 and there have been no studies to document the efficacy of this regimen compared with 4R. None of the patients treated during these time periods have been referred back to the clinic to be assessed for development of active disease. A review of the state TB registry is possible and may point to the efficacy of treatment for each group. However, the usefulness of this approach may be limited due to the mobility of the population and the relatively small sample size. A limitation of rifampin for the treatment of LTBI is its potential for drug/drug interactions with drugs such as oral contraceptives, warfarin, and methadone. In our review, most of our patients did not have this problem; but when occasionally encountered, isoniazid was used. The possibility of rifampin staining body fluids, contact lenses, and ocular implants was also discussed with patients. Treatment of LTBI with rifampin is not advised in patients with HIV infection particularly because establishing a diagnosis of TB disease may be sometimes be difficult, increasing the risk for the development of rifampin-resistant disease because of a larger bacterial burden. Additionally, it is well known that rifampin interacts with many antiretrovirals used for HIV. This is an important limitation that could be approached by using isoniazid in any HIV-positive patient starting treatment for LTBI. Risk assessment for HIV infection and counseling for HIV testing should be encouraged. Treatment of LTBI addresses a small bacterial population with monotherapy in order to eliminate them before they multiply and cause TB disease. When starting any patient on treatment with rifampin, as is done for all patients starting treatment for LTBI, the presence of TB disease must be stringently ruled out to avoid the development of drug resistance. The key to treatment of LTBI in any case with any antibiotic is the correct diagnosis. The Tuberculosis Clinical Trials Consortium, funded by the CDC, Department of Health and Human Services is conducting a clinical trial comparing the efficacy of once-weekly directly observed dose of isoniazid and rifapentine for 3 months with that of self-administered 9H for the treatment of LTBI, but it will take several more years to reach a conclusion for this study. Shorter treatments for LTBI are on the distant horizon. However at this point, we believe that the use of 4R should be strongly considered as an important alternative to 9H, particularly in situations in which completion of treatment may be a problem or the likelihood of infection with isoniazid-resistant Mycobacterium tuberculosis is high. 1716 Original Research

Treatment of LTBI is an important component of the US TB-elimination strategy. The results of this study demonstrated a very high level of patient acceptance and adherence with a 4R regimen. 4R is already a recommended alternative regimen to 9H. 2 We believe our data suggest it should be more widely adopted as an alternate treatment regimen for LTBI. ACKNOWLEDGMENT: The authors thank the staff of the Middlesex County, New Jersey tuberculosis clinic, particularly Patty Woods, Marybeth Caruso, and Debbie Apostolis. References 1 Centers for Disease Control and Prevention. Trends in tuberculosis: United States, 2004. MMWR Morb Mortal Wkly Rep 2005; 54:245 250 2 American Thoracic Society, Centers for Disease Control and Prevention, and Infectious Diseases Society of America. Targeted tuberculin testing and treatment of latent tuberculosis infection. Am J Respir Crit Care Med 2000; 161:S221 S247 3 Institute of Medicine. Ending neglect: the elimination of tuberculosis in the United States. In: Geiter L, ed. Committee on the Elimination of Tuberculosis in the United States, Division of Health Promotion and Disease Prevention, Institute of Medicine. Washington, DC: National Academy Press, 2000 4 Jordan TJ, Lewitt EM, Montgomery RL, et al. Isoniazid as preventive therapy in HIV infected intravenous drug abusers. JAMA 1991; 265:2987 2991 5 International Union Against Tuberculosis and Lung Disease. Anti-tuberculosis drug resistance in the world: report No. 3. Geneva, Switzerland: World Health Organization, 2004 6 LoBue P, Moser K. Use of isoniazid for latent tuberculosis infection in a public health clinic. Am J Respir Crit Care Med 2003; 168:443 447 7 Dear colleague letter from Dr. Kenneth Castro, Director, Division of Tuberculosis Elimination. Atlanta, GA: Centers for Disease Control and Prevention, 2006 8 Centers for Disease Control and Prevention. Update: adverse event data and revised American Thoracic Society/CDC recommendations against the use of rifampin and pyrazinamide for treatment of latent tuberculosis infection United States, 2003. MMWR Morb Mortal Wkly Rep 2003; 52:735 739 9 Jasmer RM, Nahid P, Hopewell PC. Latent tuberculosis infection. N Engl J Med 2003; 347:1860 1866 10 Reichman LB, Lardizabal A, Hayden CH. Considering the role of four months of rifampin in the treatment of latent tuberculosis infection. Am J Respir Crit Care Med 2004; 170:832 835 11 Menzies RI, Dion M, Rabinovitch B, et al. Treatment completion and costs of a randomized trial of rifampin for 4 months versus isoniazid for 9 months. Am J Respir Crit Care Med 2004; 170:445 449 12 Lecoeur HF, Truffot-Pernot C, Grosset JH. Experimental short course chemotherapy with rifampin and pyrazinamide. Am Rev Respir Dis 1989; 140:1189 1193 13 Dhillon J, Dickinson JM, Sole K, et al. Preventive chemotherapy of tuberculosis in Cornell model mice with combinations of rifampin, isoniazid, and pyrazinamide. Antimicrob Agents Chemother 1996; 40:552 555 14 Hong Kong Chest Service, Tuberculosis Research Centre, Madras, and BMJ Research Council. 1992. A double-blind placebo-controlled clinical trial of three antituberculosis chemoprophylaxis regimens in patients with silicosis in Hong Kong. Am Rev Respir Dis 1992; 145:36 41 15 Polesky A, Farber HW, Gottlieb DJ, et al. Rifampin preventive therapy for tuberculosis in Boston s homeless. Am J Respir Crit Care Med 1996; 154:1473 1477 16 Villarino M, Ridzon R, Weismuller P, et al. Rifampin preventive therapy for tuberculosis infection. Am J Respir Crit Care Med 1997; 155:1735 1738 17 American Academy of Pediatrics. Tuberculosis. In: Pickering LK, ed. Red Book: 2003; Report of the Committee on Infectious Diseases, 26th ed. Elk Grove Village, IL: American Academy of Pediatrics, 2003; 652 www.chestjournal.org CHEST / 130 / 6/ DECEMBER, 2006 1717