1 Treatment Outcomes and Long-term Survival in Patients with Extensively Drug-resistant Tuberculosis Doh Hyung Kim 1, Hee Jin Kim 2, Seung-Kyu Park 3,4, Suck-Jun Kong 5, Young Sam Kim 6, Tae-Hyung Kim 7, Eun Kyung Kim 8, Ki Man Lee 9, Sung Soon Lee 10, Jae Seuk Park 1, Won-Jung Koh 11, Chang-Hoon Lee 12, Ji Yeon Kim 2, and Tae Sun Shim 4,13 1 Department of Internal Medicine, Dankook University College of Medicine, Dankook University Hospital, Cheonan, South Korea; 2 Department of Epidemiology, The Korean Institute of Tuberculosis, Seoul, South Korea; 3 Clinical Research Center, National Masan Hospital, Masan, South Korea; 4 Division of Clinical Research, International Tuberculosis Research Center, Masan, South Korea; 5 Department of Thoracic Surgery, Mokpo National Hospital, Mokpo, South Korea; 6 Department of Internal Medicine, Yonsei University College of Medicine, Severance Hospital, Seoul, South Korea; 7 Department of Internal Medicine, Hanyang University College of Medicine, Seoul, South Korea; 8 Division of Pulmonary and Critical Care Medicine, Department of Internal Medicine, College of Medicine, Pochon CHA University, Seongnam, South Korea; 9 Department of Internal Medicine, Chungbuk National University College of Medicine, Chungbuk National University Hospital, Cheongju, South Korea; 10 Department of Internal Medicine, Inje University School of Medicine, Ilsan Paik Hospital, Goyang, South Korea; 11 Division of Pulmonary and Critical Care Medicine, Department of Medicine, Samsung Medical Center, Sungkyunkwan University School of Medicine, Samsung Medical Center, Seoul, South Korea; 12 Division of HIV and TB Control, Department of Disease Prevention, Korea Centers for Disease Control and Prevention, South Korea; 13 Division of Pulmonary and Critical Care Medicine, University of Ulsan College of Medicine, Asan Medical Center, Seoul, South Korea Rationale: The increasing worldwide incidence of extensively drugresistant tuberculosis (XDR-TB) has emerged as a threat to public health and tuberculosis (TB) control. Treatment outcomes have varied among studies, and data on long-term survival are still scarce. Objectives: To retrospectively assess the burden, clinical characteristics, treatment outcomes, and long-term survival rate of patients with XDR-TB in a cohort of patients with HIV-negative multidrugresistant tuberculosis (MDR-TB) in South Korea. Methods: Medical records were reviewed of patients newly diagnosed with or retreated for MDR-TB from 2000 to The cohort was monitored for 3 to 7 years after the initiation of treatment. Initial treatment outcomes and cumulative survival rates were analyzed, and predictors of treatment success and survival were defined. Measurements and Main Results: Of 1,407 patients with MDR-TB 75 (5.3%) had XDR-TB at treatment initiation. The default rate was high (453/1,407; 32%), and patients with XDR-TB had lower treatment success (29.3 vs. 46.2%; P ) and higher all-cause (49.3 vs. 19.4%; P, 0.001) and TB-related disease mortality (41.3 vs. 11.8%; P, 0.001) than other patients with MDR-TB. The presence of XDR-TB significantly affected treatment success (odds ratio, 0.23; 95% confidence interval [CI], ; P ), all-cause mortality (hazards ratio, 3.25; 95% CI, ; P, 0.001), and TB-related mortality (hazards ratio, 4.45; 95% CI, ; P, 0.001) on multivariate analyses. Conclusions: XDR-TB occurred in a substantial proportion of patients with MDR-TB in South Korea, and was the strongest predictor of treatment outcomes and long-term survival in patients with MDR- TB. Adequate TB control policies should be implemented to prevent the further development and spread of drug resistance. Keywords: tuberculosis; extensively drug-resistant tuberculosis; tuberculosis survival rate; treatment efficacy; South Korea (Received in original form January 22, 2008; accepted in final form August 13, 2008) Supported by the International Tuberculosis Research Center (ITRC ). Correspondence and requests for reprints should be addressed to Tae Sun Shim, M.D., Division of Pulmonary and Critical Care Medicine, Department of Internal Medicine, University of Ulsan College of Medicine, Asan Medical Center, Pungnap-dong, Songpa-gu, Seoul, , South Korea. seoul.kr This article has an online supplement, which is accessible from this issue s table of contents at Am J Respir Crit Care Med Vol 178. pp , 2008 Originally Published in Press as DOI: /rccm OC on August 14, 2008 Internet address: AT A GLANCE COMMENTARY Scientific Knowledge on the Subject Although the increasing worldwide incidence of extensively drug-resistant tuberculosis (XDR-TB) has emerged as a threat to public health and tuberculosis control, the results of treatment outcomes have varied among studies and the data on long-term survival of patients of XDR-TB patients are still scarce. What This Study Adds to the Field Our study indicated that XDR-TB was the strongest predictor of poor treatment outcomes and long-term mortality in patients with multidrug-resistant TB. Despite global efforts to control tuberculosis (TB), it remains the leading worldwide cause of death from an infectious agent. An estimated 8 to 9 million new cases develop annually, with 2 million patients dying per year (1). TB control has become more difficult due to HIV coinfection and the rise of multidrug-resistant TB (MDR-TB) (2 4). Also, extensively drug-resistant TB (XDR-TB), reported in 49 countries as of May 2008 (5), has emerged as a threat to public health (3, 6, 7). XDR-TB was originally defined as MDR-TB with bacillary resistance to three or more of the six classes of second-line drugs (6). In October 2006, the definition was revised by the World Health Organization (WHO) as MDR-TB with bacillary resistance to any fluoroquinolones and at least one second-line injectable drug (amikacin, capreomycin, or kanamycin) (7), emphasizing the important roles of fluoroquinolones and aminoglycoside/polypeptide for MDR-TB treatment. The proportions of MDR-TB isolates that were XDR-TB according to the original definition were as follows: 15% in South Korea, 4% in the United States, and 19% in Latvia (6). XDR-TB, based on the revised definition, occurred in 3% of 1,665 patients with MDR-TB in the United States from 1993 to 2006 (8). According to recent South Korean surveys, 1 23% of patients with MDR-TB had XDR-TB, that is, MDR-TB with
2 1076 AMERICAN JOURNAL OF RESPIRATORY AND CRITICAL CARE MEDICINE VOL bacillary resistance to both ofloxacin and kanamycin (9 11). The different results seem to be due to the heterogeneity of patients clinical characteristics and severity of illness, according to institutions responsible for TB treatment. In South Korea, the majority of patients with MDR-TB are treated in (1) the two national TB hospitals (as a public sector), (2) the 76 university or tertiary-care hospitals (as a private sector), or (3) the nine Korean National Tuberculosis Association (KNTA) chest clinics (between public and private sectors) (12). An investigation of the status of XDR-TB in South Korea should be conducted in a representative sample of the patients of these three categories of institutions. Treatment outcomes of XDR-TB have varied among studies (3, 6, 10, 12 16). One European study showed a 39% success rate, which is lower than the treatment success rate of MDR-TB with bacillary resistance to all first-line drugs (15). Another study showed the risk of death to be five times higher in patients with XDR-TB than in patients with MDR-TB (16). The effect of XDR-TB on long-term survival has not been determined. Recently, we performed a large survey to determine the status of MDR-TB and XDR-TB in South Korea, and we assessed the effect of XDR-TB on treatment outcomes and long-term survival of MDR-TB. Preliminary results were presented at the Fourth South Korean Swiss Biomedical Symposium in 2007 (17). METHODS Study Participants and Data Collection Patients newly diagnosed with, or retreated for, MDR-TB from January 2000 to December 2002 were retrospectively evaluated. A total of 1,407 patients with culture-proven MDR-TB were enrolled from all national TB hospitals, all KNTA chest clinics, and eight randomly selected university hospitals near Seoul. Medical records were reviewed for patients demographics, TB treatment history, comorbidities, acid-fast bacilli (AFB) culture and drug susceptibility test (DST) results, chest radiographs, and treatment modalities and outcomes. On the basis of the WHO guidelines, patients were registered into one of three categories: (1) no history of previous TB treatment, (2) history of TB treatment with first-line drugs only, and (3) history of TB treatment with second-line drugs (19), with the history of treatment defined as 1 month or more. All study patients were monitored for 3 to 7 years after treatment began. The most recent information on vital and health status was obtained from the death registry of the Korea National Statistical Office as of December 2005 and the Internet-based Korea TB Surveillance System as of March 2006 (19). This study was approved by the Institutional Review Board of the Asan Medical Center and other institutions. AFB Cultures and Drug Susceptibility Tests Sputum specimens were examined microscopically with Ziehl-Neelsen staining, and cultured on egg-based Ogawa medium. DSTs for isoniazid, rifampicin, ethambutol, pyrazinamide, streptomycin, kanamycin, cycloserine, p-aminosalicylic acid, prothionamide, and ofloxacin TABLE 1. PATIENT CHARACTERISTICS XDR-TB Other MDR-TB Total Characteristics (n 5 75) (n 5 1,332) (n 5 1,407) P Value Age, yr (16 80) (13 89) (13 89) Male sex 53 (70.7) 986 (74.0) 1,039 (73.8) BMI* ( ) ( ) ( ) Institutions National TB hospital (n 5 360) 40/360 (11.1) 320/360 (88.9) 360 (100.0) University hospital (n 5 211) 15/211 (7.1) 196/211 (92.9) 211 (100.0) KNTA chest clinic (n 5 836) 20/836 (2.4) 816/836 (97.6) 836 (100.0) Previous history of TB treatment (n 5 1,359) (n 5 73) (n 5 1,286) (n 5 1,359),0.001 No history of previous TB treatment 14 (19.2) 371 (28.8) 385 (28.3) History of TB treatment with first-line drugs only 33 (45.2) 758 (58.9) 791 (58.2) History of TB treatment with second-line drugs 26 (35.6) 157 (12.2) 183 (13.5) >2 Previous TB treatments (n 5 1,335) 31/65 (47.7) 339/1,270 (26.7) 370/1,335 (27.7),0.001 Family history of TB (n 5 542) 18/47 (38.3) 179/495 (36.2) 197/542 (36.3) Family history of MDR-TB (n 5 219) 0/20 (0.0) 15/199 (7.5) 15/219 (6.8) Underlying diseases Diabetes mellitus 14 (18.7) 225 (16.9) 239 (17.0) Chronic liver disease 1 (1.3) 25 (1.9) 26 (1.8) Malignancy 1 (1.3) 12 (0.9) 13 (0.9) Other 4 (5.4) 36 (2.7) 40 (2.8) HIV seropositive (n 5 66) 0/6 (0.0) 1/60 (1.7) 1/66 (1.5) Extrapulmonary TB 1 (1.3) 52 (3.9) 53 (3.8) Positive AFB smear at treatment initiation 52 (69.3) 906 (68.0) 958 (68.1) Radiologic severity (n 5 1,251) (n 5 67) (n 5 1,184) (n 5 1,251) Minimal 2 (3.0) 93 (7.9) 95 (7.6) Moderately advanced 31 (46.3) 713 (60.2) 744 (59.5) Far advanced 34 (50.7) 378 (31.9) 412 (32.9) Cavitary disease 37 (49.3) 562 (42.2) 599 (42.6) Bilateral disease (n 5 1,190) 51/67 (76.1) 829/1,123 (73.8) 880/1,190 (73.9) Admission at treatment initiation (n 5 1,402) 45/74 (60.8) 372/1,328 (28.0) 417/1,402 (29.7),0.001 Definition of abbreviations: AFB 5 acid-fast bacilli; BMI 5 body mass index; KNTA 5 Korean National Tuberculosis Association; MDR 5 multidrug-resistant; TB 5 tuberculosis; XDR 5 extensively drug-resistant. Values are number (%) or mean 6 SD and (range). * BMI is calculated as weight in kilograms divided by the square of height in meters. Statistical comparison of data was performed using Fisher s exact test. Other included chronic lung disease (n 5 9), chronic renal disease (n 5 4), hypo- or hyperthyroidism (n 5 5), connective tissue disease (n 5 3), psychiatric disease (n 5 5), malnutrition, and postgastrectomy state (n 5 14).
3 Kim, Kim, Park, et al.: Long-term Survival of XDR-TB 1077 were performed at all six participating laboratories. (See the online supplement for details.) Enviomycin/capreomycin susceptibility was determined at the Korean Institute of Tuberculosis (KIT), which is a Supranational TB Reference Laboratory, and one private laboratory. The proportion method was used for DST at the KIT, and pyrazinamide susceptibility was determined using the pyrazinamidase test (9). Definitions of Study Measures XDR-TB. XDR-TB is defined as MDR-TB with bacillary resistance to both (1) ofloxacin and (2) one of the second-line injectable drugs (kanamycin, capreomycin, or enviomycin). Treatment outcomes. We used the six treatment outcome categories recommended by the WHO (cure, treatment completion, transfer out, default, death, and treatment failure) (18) in addition to short-term treatment completion. The duration of adequate treatment was defined as 18 months or more and 12 months or more after culture conversion. The definition of short-term treatment completion was applied to patients who met all of the following criteria: (1) inadequate treatment duration but duration of more than 6 months, (2) more than three consecutive negative cultures before treatment completion, and (3) treatment completion by a doctor based on favorable treatment response. Treatment success was defined as the sum of cure, treatment completion, and short-term treatment completion. As recommended by the WHO, all treatment outcomes were based on AFB culture results (18). Radiographic severity. Radiographic severity was estimated by using the recommendations of the National Tuberculosis Association of the United States (20). Statistical Analysis Chi-square tests, Fisher s exact tests, Student s t tests, and Mann-Whitney U tests were used for univariate analyses, and logistic regression and Cox regression were used for multivariate analyses. Cumulative survival was compared by using the Kaplan-Meier method with the log-rank test. SPSS software (SPSS, Inc., Chicago, IL), version 13.0, was used for all analyses, and P, 0.05 indicated statistical significance. RESULTS Patient Demographics Of 1,407 patients with MDR-TB, 75 (5.3%) were defined as having XDR-TB; these patients included 40 from national TB hospitals (11.1%, 40/360), 20 from KNTA chest clinics (2.4%, 20/836), and 15 from private university hospitals (7.1%, 15/211). The mean age was 42.9 years (range, 13 89), the male to female ratio was 2.8:1, and mean body mass index (BMI) was 19.2 kg/m 2 (range, ). HIV-ELISA tests were conducted on the 66 (4.7%) patients with clinically suspected MDR-TB, and one patient was seropositive. Comorbidities were found in 300 of 1,407 (21.3%) patients, with diabetes mellitus being the most common (17.0%). Extrapulmonary TB was observed in 53 (3.8%) patients, with the pleura being the most frequently involved site (45.3%, 24/53). Positive AFB smears at treatment initiation were observed in 958 (68.1%) patients. When compared with other patients with MDR-TB, the mean age was significantly higher in patients with XDR-TB ( vs yr; P ), as was history of previous TB treatment with second-line drugs (35.6 vs. 12.2%; P, 0.001), the percentage with two or more previous TB treatments (47.7 vs. 26.7%; P, 0.001), the percentage with advanced disease (50.7 vs. 31.9%; P, 0.004), and the percentage admitted at treatment initiation (60.8 vs. 28.0%; P, 0.001) (Table 1). Drug Resistance Rate DSTs for enviomycin and capreomycin were performed in 1,002 and in 25 patients, respectively. The resistance rates in all drugs were significantly higher in patients with XDR-TB than in other patients with MDR-TB (Table 2). TABLE 2. DRUG RESISTANCE RATE AT TREATMENT INITIATION XDR-TB Other MDR-TB Total TB Drugs (n 5 75) (n 5 1,332) (n 5 1,407) P Value Isoniazid 75 (100.0) 1,332 (100.0) 1,407 (100.0) Rifampicin 75 (100.0) 1,332 (100.0) 1,407 (100.0) Ethambutol 60 (80.0) 708 (53.2) 768 (54.6),0.001 Pyrazinamide 37 (49.3) 445 (33.4) 482 (34.3) Streptomycin 48 (64.0) 284 (21.3) 332 (23.6),0.001 Kanamycin 72 (96.0) 115 (8.6) 186 (13.3),0.001 Ofloxacin 75 (100.0) 159 (11.9) 234 (16.6),0.001 Cycloserine* 34 (45.3) 58 (4.4) 92 (6.5),0.001 p-aminosalicylic acid 48 (64.0) 226 (17.0) 274 (19.5),0.001 Prothionamide 49 (65.3) 190 (14.3) 239 (17.0),0.001 Enviomycin 14/30 (46.7) 8/972 (0.8) 22/1,002 (2.2),0.001 (n 5 1,002)* Capreomycin (n 5 25)* 5/6 (83.3) 4/19 (21.1) 9/25 (36.0) Definition of abbreviations: MDR 5 multidrug-resistant; TB 5 tuberculosis; XDR 5 extensively drug-resistant. Values are number (%). * Statistical comparison of data was performed using Fisher s exact test. Enviomycin and capreomycin susceptibility tests were conducted at the laboratory of the Korean Institute of Tuberculosis and one private laboratory. Treatment Modalities All patients were treated by individualized regimens on the basis of DST results and history of previous TB drug use. Directly observed therapy was performed only on patients admitted in national TB hospitals. Patients with XDR-TB received the same median number of drugs as did other patients with MDR-TB during previous treatment courses (4 [range, 0 16] vs. 4 [range, 0 15]; P ), whereas the median number of potentially effective drugs at MDR-TB treatment initiation was significantly lower in patients with XDR-TB than in other patients with MDR-TB (1 [range, 0 5] vs. 4 [range, 0 7]; P, 0.001). A significantly lower proportion of patients with XDR- TB than those with MDR-TB were treated with at least three TABLE 3. TREATMENT MODALITIES XDR-TB Other MDR-TB Total Treatment Modalities (n 5 75) (n 5 1,332) (n 5 1,407) P-value No. of TB drugs used before* 4 (0 16) 4 (0 15) 4 (0 16) No. of TB drugs used at current treatment initiation* Total TB drugs used 5 (2 7) 5 (2 9) 5 (2 9) Active drugs 2 (0 5) 4 (0 9) 4 (0 9),0.001 Drugs with unknown 0 (0 2) 0 (0 6) 0 (0 6) activity Previously unused and 1 (0 5) 3 (0 7) 3 (0 7),0.001 active drugs Potentially effective 1 (0 5) 4 (0 7) 4 (0 7),0.001 drugs No. of patients, active 21 (28.0) 906 (68.0) 927 (65.9),0.001 drugs >3 No. of patients, previously 14 (18.7) 746 (56.0) 760 (54.0),0.001 unused and active drugs >3 No. of patients, potentially 16 (21.3) 1,001 (75.2) 1,017 (72.3),0.001 effective drugs >3 No. of patients, surgical 3 (4.0) 57 (4.3) 60 (4.3) resection x For definition of abbreviations, see Table 2. Values are median (range) or number (%). * Mann-Whitney U test was used for comparison of the number of drugs used. Drugs with unknown activity included drugs with unclear TB efficacy (clofazimine and amoxicillin/clavulanate) and TB drugs with unknown DST results (amikacin, moxifloxacin, and rifabutin). Potentially effective drugs included either previously unused and active drugs or previously unused and activity-unknown drugs. x Statistical comparison of data was performed using Fisher s exact test.
4 1078 AMERICAN JOURNAL OF RESPIRATORY AND CRITICAL CARE MEDICINE VOL TABLE 4. TREATMENT OUTCOMES XDR-TB Other MDR-TB Total Treatment Outcomes (n 5 75) (n 5 1,332) (n 5 1,407) P Value Cure 9 (12.0) 416 (31.2) 425 (30.2),0.001 Treatment completion 4 (5.3) 89 (6.7) 93 (6.6) Short-term treatment 9 (12.0) 110 (8.3) 119 (8.5) completion Failure 12 (16.0) 53 (4.0) 65 (4.6),0.001 Transfer out 9 (12.0) 99 (7.4) 108 (7.7) Default 12 (16.0) 441 (33.1) 453 (32.2) Death 20 (26.7) 124 (9.3) 144 (10.2),0.001 Treatment success 22 (29.3) 615 (46.2) 637 (45.3) Relapse (n 5 637)* 3/22 (13.6) 54/615 (8.8) 57/637 (8.9) For definition of abbreviations, see Table 2. Values are number (%). * Statistical comparison of data was performed using Fisher s exact test. active drugs (P, 0.001). Surgical resection was performed on 60 (4.3%) patients in the entire cohort, without a statistically significant difference in the rates between patients with XDR- TB and those with MDR-TB (P ) (Table 3). Treatment Outcomes and Predictors of Treatment Success Treatment success on initial treatment was observed in 637 (45.3%) of 1,407 patients, a cure was determined in 425 (30.2%), treatment completion in 93 (6.6%), and short-term treatment completion in 119 (8.5%). Treatment success was more common in other patients with MDR-TB than in patients with XDR-TB (46.2 vs. 29.3%; P ), whereas treatment failure (16.0 vs. 4.0%; P, 0.001) and death (26.7 vs. 9.3%; P, 0.001) were more common in patients with XDR-TB. The default rate was 32.2% (453/1,407) in total, and it was about one-half as high in patients with XDR-TB than in other patients with MDR-TB (16.0 vs. 33.1%; P ). Among the 453 defaulted patients, treatment was reinitiated in 96 (21.2%) patients during the study period, not reinitiated in 210 (46.4%) patients, and further information could not be obtained for 139 (30.7%) patients. Among the 96 retreated patients, 8 (8.3%) eventually completed treatment, 81 (84.4%) defaulted again, 4 (4.2%) transferred out, and 1 (1.0%) failed. Another 2 (2.1%) patients were retreated in other hospitals, but their outcome information could not be obtained. Relapse occurred in 57 (8.9%) of 637 patients who attained treatment success, including 8.2% (35/425) of the cured patients, 7.5% (7/93) of the patients who completed treatment, and 12.6% (15/119) of the patients who completed short-term treatment (P ). The relapse rate was higher in patients with XDR-TB (13.6%) than in other patients with MDR-TB (8.8%), but the difference was not statistically significant (P ) (Table 4). When the outcomes were compared between patients treated medically and those treated medically and surgically, adjunctive surgical treatment imparted a higher treatment success rate in other patients with MDR-TB (68.3%, 41/60, vs. 44.2%, 596/1,347; P, 0.001) and a higher but not statistically significant rate in patients with XDR-TB (66.7%, 2/3, vs. 27.8%, 20/72; P ). In multivariate logistic regression analysis, treatment success was inversely associated with a BMI less than 18.5 kg/m 2 (odds ratio [OR], 0.46; 95% confidence interval [CI], ; P, 0.001), two or more previous TB treatments (OR, 0.49; 95% CI, , P ), and XDR-TB (OR, 0.23; 95% CI, ; P ). Although XDR-TB was the strongest negative predictor of treatment success, surgical treatment was the only positive predictor of treatment success (OR, 3.87; 95% CI, ; P ) (Table 5). Survival Analysis and Predictors of Mortality Death was the initial treatment outcome in 144 patients, and 151 additional deaths occurred during the 3 to 7 years of TABLE 5. PREDICTORS OF TREATMENT SUCCESS Univariate Analysis Multivariate Analysis Variables Treatment Success (n 5 637) Others* (n 5 770) OR OR Age.40 yr (56.5) (48.4) ( ) Female sex (27.9) (24.7) ( ) BMI,18.5 kg/m , ,0.001 (9.4) (16.1) ( ) ( ) Previous TB treatment with 50/ / ,0.001 second-line drugs (n 5 1,359) (8.1) (17.9) ( ) >2 Previous TB treatments (n 5 1,335) 141/ / (23.2) (31.5) ( ) ( ) Far advanced disease (n 5 1,251) 146/ / ,0.001 (25.9) (38.7) ( ) Cavitary disease (44.9) (40.6) ( ) Surgical treatment , (6.4) (2.5) ( ) ( ) No. of TB drugs used before ,0.001 ( ) No. of potentially effective TB drugs ,0.001 ( ) Cycloserine resistance 34 (5.3) 58 (7.5) ( ) XDR-TB (3.5) (6.9) ( ) ( ) Definition of abbreviations: BMI 5 body mass index; CI 5 confidence interval; OR 5 odds ratio; TB 5 tuberculosis; XDR 5 extensively drug-resistant. Values are number (%) or mean 6 SD. * Include failed, defaulted, and transferred-out patients and patients who died as treatment outcomes. OR was for treatment success.
5 Kim, Kim, Park, et al.: Long-term Survival of XDR-TB 1079 follow-up. These 295 deaths included the following: (1) 37 (49.3%) of 75 patients with XDR-TB, with 31 (41.3%) being TB-related deaths; and (2) 258 (19.4%) of 1,332 other patients with MDR-TB, with 157 (11.8%) being TB-related deaths. Both all-cause and TB-related mortality were approximately three to four times more likely in patients with XDR-TB than in patients with MDR-TB; they were about two to three times as likely in patients older than 40 years, in patients with a BMI below 18.5 kg/m 2, and in patients with advanced disease (Tables 6 and 7). XDR-TB was the strongest predictor of both all-cause and TB-related mortality, and survival curves showed higher cumulative mortality in patients with XDR-TB than in other patients with MDR-TB (P, 0.001) (Figure 1). Even if the survival analysis was restricted to younger patients (age <40 yr), XDR- TB and lower BMI were statistically significant predictors: the hazards ratio of XDR-TB was 3.82 (95% CI, ; P ) for all-cause mortality and 5.60 (95% CI, ; P ) for TB-related mortality, and the hazards ratio of a lower BMI was 2.26 (95% CI, ; P ) for all-cause mortality and 3.51 (95% CI, ; P ) for TBrelated mortality. (See the online supplement for details.) DISCUSSION Although this study was retrospective, it is the largest report that we know of that compares patients with XDR-TB with other patients with MDR-TB to determine the impact of XDR- TB on treatment outcomes and long-term survival in mostly HIV-negative patients with MDR-TB. Our study largely adopted the WHO-recommended definitions of XDR-TB, case registration, and treatment outcomes (7, 18). The overall treatment success rate was less than 50%, and the treatment success rate in patients with XDR-TB was only 29%. In addition, the high default rate (32.2%) indicates that the TB control program in South Korea should be reevaluated. We found that XDR-TB was the strongest predictor of treatment failure and long-term mortality in a substantial proportion (5.3%) of patients with MDR-TB. In South Korea, the proportion of MDR-TB among new cases of TB increased from 1.6% in 1994 to 2.7% in 2004 (21), which is a major threat to public health. Since the U.S. Centers for Disease Control and Prevention and the WHO reported that 15% of South Korean patients with MDR-TB had a diagnosis of XDR-TB, on the basis of the original definition (6), three new reports have estimated the burden of XDR-TB in South Korea. Their estimates range from 1.4 to 23% (9 11) on the basis of limited data from individual treatment institutions. In contrast, the patients in our study were recruited from all national TB hospitals, all KNTA chest clinics, and a fraction of university hospitals, and were estimated to include one-third to one-half of all patients with MDR-TB whose disease was newly diagnosed or who had been retreated in South Korea between 2000 and Thus, we believe that our study results are closer to the real status of patients with MDR-TB and XDR-TB in South Korea. The overall treatment success rate (45.3%) of patients with MDR-TB in our study was similar to that in another South Korean report (48.2%) (9), but lower than the 51 77% rates in other reports (10, 13 15, 22 24). Also, the 29.3% treatment TABLE 6. PREDICTORS OF ALL-CAUSE MORTALITY Univariate Analysis Multivariate Analysis Variables Patients (n 5 295) HR* HR* Age.40 yr 211 (71.5) 2.47, ( ) ( ) Male sex 234 (79.3) ( ) BMI,18.5 kg/m 2 73 (24.7) 2.42, ,0.001 ( ) ( ) Previous TB treatment with 75/183 (41.0) 2.54,0.001 second-line drugs (n 5 183) ( ) >2 Previous TB treatments (n 5 370) 105/370 (28.4) 1.75,0.001 ( ) Far advanced disease (n 5 412) 158/412 (38.3) 3.31, ( ) ( ) Positive AFB smear at treatment initiation 212 (71.9) ( ) No. of TB drugs used before ,0.001 ( ) No. of potentially effective TB drugs ,0.001 ( ) Surgical treatment 7 (2.4) ( ) XDR-TB (n 5 75) 37/75 (49.3) 2.82, ,0.001 ( ) ( ) Streptomycin resistance 83 (28.1) ( ) Prothionamide resistance 80 (27.1) 1.92,0.001 ( ) Cycloserine resistance 34 (11.5) 2.11,0.001 ( ) p-aminosalicylic acid resistance 79 (26.8) 1.62,0.001 ( ) Definition of abbreviations: AFB 5 acid-fast bacilli; BMI 5 body mass index; CI 5 confidence interval; HR 5 hazard ratio; TB 5 tuberculosis; XDR 5 extensively drug-resistant. Values are number (%) or mean 6 SD. * HR for all-cause mortality.
6 1080 AMERICAN JOURNAL OF RESPIRATORY AND CRITICAL CARE MEDICINE VOL TABLE 7. PREDICTORS OF TUBERCULOSIS-RELATED MORTALITY Univariate Analysis Multivariate Analysis Variables Patients (n 5 188) HR* HR* Age.40 yr 121 (64.4) 1.77, ( ) ( ) Male sex 144 (76.6) ( ) BMI,18.5 kg/m 2 55 (29.3) 3.06, ,0.001 ( ) ( ) Previous TB treatment with 59/183 (32.2) 3.45,0.001 second-line drugs (n 5 183) ( ) >2 Previous TB treatments (n 5 370) 78/370 (21.1) 2.35,0.001 ( ) Far advanced disease (n 5 412) 110/412 (26.7) 3.89, ( ) ( ) Cavitary disease 92 (48.9) ( ) Positive AFB smear at treatment initiation 138 (73.4) ( ) No. of TB drugs used before ,0.001 ( ) No. of potentially effective TB drugs ,0.001 ( ) XDR-TB (n 5 75) 31/75 (41.3) 3.85, ,0.001 ( ) ( ) Streptomycin resistance 61 (32.4) ( ) Prothionamide resistance 64 (34.0) 2.65,0.001 ( ) Cycloserine resistance 27 (14.4) 2.75,0.001 ( ) p-aminosalicylic acid resistance 59 (31.4) 2.02,0.001 ( ) For definition of abbreviations, see Table 6. Values are number (%) or mean 6 SD. * HR for tuberculosis-related mortality. success rate of patients with XDR-TB in our study was much lower than the 39 54% rates in European and other South Korean studies (10, 15). In multivariate analysis, XDR-TB was the strongest negative predictor of treatment success; in combination with the high default rate, XDR-TB probably contributes to the low rate of treatment success in patients with MDR- TB in South Korea. Adjunctive surgical resection and fluoroquinolone use are reported to be associated with improved outcomes in patients with MDR-TB. Chan and colleagues (13) reported that 92% of 108 patients with MDR-TB who underwent surgical resection had an initial favorable response. Surgery was associated with a nearly fivefold increase in the odds of initial favorable outcome and increased long-term survival. Although only 60 (4.3%) patients in our study had surgical resection, we determined through multivariate analysis that it is the only positive predictor of treatment success in XDR-TB treatment, with treatment success rates greater than 65% in patients with XDR-TB and other patients with MDR-TB. Considering the difficulty in managing patients with MDR-TB and XDR-TB, surgical resection should always be considered an important option. Unfortunately, the number of surgeons who are experienced in MDR-TB surgery is decreasing. In one report, the complication rate for MDR-TB surgery was 22.8% and the treatment failure rate was 27.8% (25), which suggests that the proper selection of patients is also important to lung resection results. Because almost all our study patients had received fluoroquinolones or injectable drugs at least once during their course of treatment, outcomes could not be compared between those patients who did and those who did not use fluoroquinolones or injectable drugs. Information about the long-term mortality of XDR-TB and its effect on patients with MDR-TB is still scarce. In one European study, patients with XDR-TB had a fivefold higher risk for death and needed longer hospitalization and duration of treatment (16). However, the data were derived from only 11 patients with XDR-TB among the 124 patients with MDR-TB, and patient follow-up was relatively short. Our study had 75 patients with XDR-TB who were monitored for 3 to 7 years, and approximately 50% died during follow-up. The 50% mortality rate is not too different from the 5-year mortality rate in South Indian patients with TB who were not treated with TB drugs (26), which confirms the poor prognosis of XDR-TB despite treatment. Patients with XDR-TB had 3.3- and 4.5-fold greater risks than other patients with MDR-TB for all-cause and TBrelated mortality, respectively, and XDR-TB was the strongest predictor of long-term mortality in patients with MDR-TB, even in patients younger than 40 years, in multivariate analysis. Obese and overweight (BMI >25 kg/m 2 ) individuals have been reported to be at significantly lower risk of developing active TB than normal-weight individuals (BMI, kg/m 2 ) (27), and a cholesterol-rich diet has been reported to accelerate bacteriologic sterilization in pulmonary TB (28). On the basis of the findings in our study and others, we believe that nutritional status is another important predictor of treatment outcomes and long-term survival of patients with MDR-TB. Treatment outcomes of patients with XDR-TB seem to be negatively affected by HIV coinfection (29, 30). In KwaZulu
7 Kim, Kim, Park, et al.: Long-term Survival of XDR-TB 1081 Figure 1. Kaplan-Meier survival analysis estimates for (A) all-cause and (B) tuberculosis (TB)-related mortality between patients with extensively drug-resistant TB and other patients with multidrug-resistant TB. Patients with extensively drug-resistant TB had significantly shorter median survival time for both all-cause (62 vs. 84 mo; P, 0.001) and TB-related mortalities (63 vs. 87 mo; P, 0.001). Dashed line indicates other multidrug-resistant TB; solid line indicates extensively drugresistant TB. Natal, South Africa, 52 of 53 HIV-positive patients with XDR- TB died within a median of 16 days after being tested for TB (3, 29). South Korea has been a low HIV-burden country (31, 32), with a reported HIV prevalence below 0.1% in adults aged years (33). According to the Korea Centers for Disease Control and Prevention, only 77 of 152,887 patients diagnosed with TB between 2001 and 2005 were confirmed to have HIV during a mean follow-up of 2.6 years (32). HIV-ELISA tests were conducted on 66 (4.7%) patients in our study, with 1 patient being HIV positive and the other 1,341 patients not suspected of being infected. Considering the HIV situation in South Korea, the treatment outcomes in our study largely reflect the experience of HIV-negative patients with MDR- TB. Although the outcomes are poor, they are better than those of HIV-positive patients in other countries with MDR-TB and XDR-TB. To date, six categories have been defined for MDR-TB treatment outcomes (18, 34). These definitions are designed for national TB-control program management and evaluation rather than for individual clinical care and may need to be modified in diverse clinical settings. In our study, an arbitrarily added outcome category, short-term treatment completion, included a nonnegligible proportion (8.5%) of study patients. Without this category, these patients would have been assigned to the default outcome because their treatment duration and AFB culture results did not fit the criteria for cure or treatment completion. We believe that patients who completed short-term treatment patients should be distinguished from defaulted patients because the definition of short-term treatment completion was applied to patients for whom the treatment was completed by doctors and was based on favorable clinical and microbiologic responses. In our study, the relapse rates were 8.2, 7.5, and 12.6% in cured patients, patients who completed treatment, and patients who completed shortterm treatment, respectively, but the rate did not differ significantly (P ). The clinical courses of patients with MDR- TB who completed treatment are variable and there are no proven best treatment regimens: the time to culture conversion is widely variable, and MDR-TB treatment is occasionally completed by short-course, first-line drug treatment only (35). Thus, more studies are needed to correctly define the treatment outcomes for patients with MDR-TB. If the category of shortterm treatment completion had not been used in this study, the treatment success rates would be 36.8% in patients with MDR- TB and 17.3% in those with XDR-TB, and the default rates would be 40.7% in patients with MDR-TB and 28.0% in those with XDR-TB. This study has several limitations that are inherent to retrospective studies. First, most information was obtained from medical records, which limited our choice of predictors of treatment outcomes and survival. For example, radiographic severity was estimated by the findings on simple chest X-rays alone, and all TB drug information could not be obtained from some patients, especially in the remote history of their treatment. Only some study patients had undergone lung resectional surgery, and surgical indications were not clearly defined, although localized disease with adequate lung function was the general indication for resectional surgery for patients with MDR-TB. If surgical resection had been used more often to manage patients with MDR-TB, it might have been a more important predictor of treatment success and long-term survival. Second, the definition of XDR used in this study was not fully consistent with the WHO-recommended definition. DSTs for amikacin and other fluoroquinolones except ofloxacin were not performed, and capreomycin susceptibility testing was performed in only a minority of study patients. Thus, the actual burden of XDR-TB might be underestimated in this study. Third, the quality of DSTs was not strictly controlled, and double-checking of DST was not performed. Although more than 70% of DSTs were performed at a Supranational TB Reference Laboratory, in which the internal and external quality control programs are adequate, the other tests were not quality controlled as recommended by the WHO. The other laboratories had their own internal and external quality-control programs, such as an external quality-control program of the College of American Pathologists or irregular interlaboratory proficiency testing with the KIT. Fourth, the data presented in this study were not derived from nationwide representative sampling. However, our study patients were recruited from all the institutions that are primarily responsible for treatment of MDR-TB in South Korea, so the results are likely to be closer to the real situation of MDR-TB and XDR-TB in South Korea than are previous results. Fifth, too many patients (32.2%) were categorized as having the default outcome to get enough information to assess actual treatment outcomes, especially in terms of treatment failure and relapse. Hence, we tried to analyze the second treatment outcomes of these defaulted patients. However, most of the initially defaulted patients
8 1082 AMERICAN JOURNAL OF RESPIRATORY AND CRITICAL CARE MEDICINE VOL defaulted again, and only a minor proportion (1.8%, 8/453) of patients completed treatment. Despite these limitations, to our knowledge, this study is the largest one on MDR-TB and XDR-TB. Present in a substantial proportion of HIV-negative patients with MDR-TB in South Korea, XDR-TB is the strongest negative predictor of treatment success and long-term survival. This study confirms the importance of establishing adequate TB control programs to reduce defaulted patients. Effective TB control practices should be implemented to prevent the development and spread of MDR-TB and XDR-TB in South Korea. 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. References 1. Frieden TR, Sterling TR, Munsiff SS, Watt CJ, Dye C. Tuberculosis. Lancet 2003;362: AzizMA,WrightA,LaszloA,DeMuynckA,PortaelsF,VanDeunA, Wells C, Nunn P, Blanc L, Raviglione M. Epidemiology of antituberculosis drug resistance (The Global Project on Anti-tuberculosis Drug Resistance Surveillance): an updated analysis. Lancet 2006;368: Gandhi NR, Moll A, Sturm AW, Pawinski R, Govender T, Lalloo U, Zeller K, Andrews J, Friedland G. Extensively drug-resistant tuberculosis as a cause of death in patients co-infected with tuberculosis and HIV in a rural area of South Africa. Lancet 2006;368: Zignol M, Hosseini MS, Wright A, Weezenbeek CL, Nunn P, Watt CJ, Williams BG, Dye C. Global incidence of multidrug-resistant tuberculosis. J Infect Dis 2006;194: World Health Organization. Global map and information on XDR-TB: countries with XDR-TB confirmed cases as of May 2008 [Internet] [accessed 2008 Aug 4]. Available from: challenges/xdr/xdrmap_may08_en.pdf 6. Centers for Disease Control and Prevention. Emergence of Mycobacterium tuberculosis with extensive resistance to second-line drugs worldwide, MMWR Morb Mortal Wkly Rep 2006;55: World Health Organization. Extensively drug-resistant tuberculosis (XDR-TB): recommendations for prevention and control. Wkly Epidemiol Rec 2006;81: Centers for Disease Control and Prevention. Extensively drug-resistant tuberculosis United States, MMWR Morb Mortal Wkly Rep 2007;56: Bai GH, Park YK, Choi YW, Bai JI, Kim HJ, Chang CL, Lee JK, Kim SJ. Trend of anti-tuberculosis drug resistance in Korea, Int J Tuberc Lung Dis 2007;11: Kim HR, Hwang SS, Kim HJ, Lee SM, Yoo CG, Kim YW, Han SK, Shim YS, Yim JJ. Impact of extensive drug resistance on treatment outcomes in non-hiv-infected patients with multidrug-resistant tuberculosis. Clin Infect Dis 2007;45: Jeon CY, Hwang SH, Min JH, Prevots DR, Goldfeder LC, Lee H, Eum SY, Jeon DS, Kang HS, Kim JH, et al. Extensively drug-resistant tuberculosis in South Korea: risk factors and treatment outcomes among patients at a tertiary referral hospital. Clin Infect Dis 2008;46: Seung KJ, Bai GH, Kim SJ, Lew WJ, Park SK, Kim JY. The treatment of tuberculosis in South Korea. Int J Tuberc Lung Dis 2003;7: Chan ED, Laurel V, Strand MJ, Chan JF, Huynh ML, Goble M, Iseman MD. Treatment and outcome analysis of 205 patients with multidrugresistant tuberculosis. Am J Respir Crit Care Med 2004;169: Chiang CY, Enarson DA, Yu MC, Bai KJ, Huang RM, Hsu CJ, Suo J, Lin TP. Outcome of pulmonary multidrug-resistant tuberculosis: a 6-yr follow-up study. Eur Respir J 2006;28: Migliori GB, Besozzi G, Girardi E, Kliiman K, Lange C, Toungoussova OS, Ferrara G, Cirillo DM, Gori A, Matteelli A, et al. Clinical and operational value of the extensively drug-resistant tuberculosis definition. Eur Respir J 2007;30: Migliori GB, Ortmann J, Girardi E, Besozzi G, Lange C, Cirillo DM, Ferrarese M, De Iaco G, Gori A, Raviglione MC. Extensively drugresistant tuberculosis, Italy and Germany. Emerg Infect Dis 2007;13: Shim TS. Current situation of MDR-TB in Korea [abstract]. Presented at the fourth Korean-Swiss Biomedical Symposium June 6, 2007, Masan, South Korea: World Health Organization. Guidelines for the programmatic management of drug-resistant tuberculosis. Geneva, Switzerland: World Health Organization; Publication No. WHO/HTM/TB/ Lew WJ, Lee EG, Bai JY, Kim HJ, Bai GH, Ahn DI, Lee JK, Kim SJ. An Internet-based surveillance system for tuberculosis in Korea. Int J Tuberc Lung Dis 2006;10: National Tuberculosis Association. Diagnostic standards and classification of tuberculosis. New York: The Association; Kim HJ, Hong YP, Kim SJ, Lew WJ, Lee EG. Ambulatory treatment of multidrug-resistant pulmonary tuberculosis patients at a chest clinic. Int J Tuberc Lung Dis 2001;5: Goble M, Iseman MD, Madsen LA, Waite D, Ackerson L, Horsburgh CR Jr. Treatment of 171 patients with pulmonary tuberculosis resistant to isoniazid and rifampin. N Engl J Med 1993;328: Nathanson E, Lambregts-van Weezenbeek C, Rich ML, Gupta R, Bayona J, Blondal K, Caminero JA, Cegielski JP, Danilovits M, Espinal MA, et al. Multidrug-resistant tuberculosis management in resource-limited settings. Emerg Infect Dis 2006;12: Tahaoglu K, Torun T, Sevim T, Atac G, Kir A, Karasulu L, Ozmen I, Kapakli N. The treatment of multidrug-resistant tuberculosis in Turkey. N Engl J Med 2001;345: Kim HJ, Kang CH, Kim YT, Sung SW, Kim JH, Lee SM, Yoo CG, Lee CT, Kim YW, Han SK, et al. Prognostic factors for surgical resection in patients with multidrug-resistant tuberculosis. Eur Respir J 2006;28: National Tuberculosis Institute, Bangalore. Tuberculosis in a rural population of South India: a five-year epidemiological study. Bull World Health Organ 1974;51: Leung CC, Lam TH, Chan WM, Yew WW, Ho KS, Leung G, Law WS, Tam CM, Chan CK, Chang KC. Lower risk of tuberculosis in obesity. Arch Intern Med 2007;167: Perez-Guzman C, Vargas MH, Quinonez F, Bazavilvazo N, Aguilar A. A cholesterol-rich diet accelerates bacteriologic sterilization in pulmonary tuberculosis. Chest 2005;127: Koenig R. Drug-resistant tuberculosis: in South Africa, XDR TB and HIV prove a deadly combination. Science 2008;319: Chaisson RE, Martinson NA. Tuberculosis in Africa: combating an HIV-driven crisis. N Engl J Med 2008;358: Oh MD, Choe K. Epidemiology of HIV infection in the Republic of Korea. J Korean Med Sci 1999;14: Lee CH, Hwang JY, Bae JY, Kim HJ, Bae KH, Nam JK, Park BH, Kim HH. HIV/tuberculosis co-infection in Korea [abstract]. Tuberc Respir Dis 2007;61: Choi BY. Perspectives of polices on HIV/AIDS and tuberculosis control in Korea. Korean J Epidemiol 2006;28: Laserson KF, Thorpe LE, Leimane V, Weyer K, Mitnick CD, Riekstina V, Zarovska E, Rich ML, Fraser HS, Alarcon E, et al. Speaking the same language: treatment outcome definitions for multidrug-resistant tuberculosis. Int J Tuberc Lung Dis 2005;9: Espinal MA, Kim SJ, Suarez PG, Kam KM, Khomenko AG, Migliori GB, Baez J, Kochi A, Dye C, Raviglione MC. Standard short-course chemotherapy for drug-resistant tuberculosis: treatment outcomes in 6 countries. JAMA 2000;283: