Rate of Reinfection Tuberculosis after Successful Treatment Is Higher than Rate of New Tuberculosis
http://www.100md.com
美国呼吸和危急护理医学 2005年第6期
Desmond Tutu TB Center, Department of Pediatrics and Child Health
MRC Center for Molecular and Cellular Biology, Department of Medical Biochemistry, Stellenbosch University, Cape Town, South Africa
KNCV Tuberculosis Foundation, The Hague
Department of Infectious Diseases, Tropical Medicine and AIDS, Academic Medical Center, Amsterdam, The Netherlands
International Union against Tuberculosis and Lung Disease (IUATLD), Paris, France
Division of Infectious Diseases, Department of Medicine, McGill University Health Center, Montreal, Quebec, Canada
ABSTRACT
Rationale: In a higheCtuberculosis (TB) incidence area of Cape Town, South Africa, there is a very high rate of unexplained recurrent TB. The incidence of new bacteriologically confirmed disease in the area is 313 per 100,000 individuals. Objective: To estimate the rate of recurrent TB attributable to reinfection after successful treatment. Methods: All patients with reported TB in the area between 1993 and 1998 were followed up to 2001 for disease needing retreatment (recurrences). Patients who were multi-drugeCresistant or who had treatment failure, were transferred, or died during treatment were excluded. Analysis was restricted to patients for whom DNA fingerprinting of their Mycobacterium tuberculosis isolates was obtained. Reinfection TB was defined as a recurrent TB episode in which the strains of the separate episodes differed by more than four bands. Measurements and Main Results: 612 of 897 (68%) patients had a DNA fingerprint available at enrollment. Median duration of follow-up was 5.2 years. Recurrent TB occurred in 108 of 612 (18%) patients, of whom 61 of 447 (14%) experienced recurrence after successful treatment, and 47 of 165 (28%) experience recurrence after default. Of the 108 patients with recurrent TB, 68 (63%) had a DNA fingerprint in the second episode. Among these patients, 24 of 31 (77%) recurrences after successful treatment and 4 of 37 (11%) recurrences after default were attributable to reinfection. The reinfection disease rate after successful treatment was estimated at 2.2 per 100 person-years. Conclusions: The age-adjusted incidence rate of TB attributable to reinfection after successful treatment was four times that of new TB. People who had TB once are at a strongly increased risk of developing TB when reinfected.
Key Words: incidence molecular epidemiology Mycobacterium tuberculosis recurrence survival analysis
Of patients with tuberculosis (TB) who are cured after short-course treatment in trial conditions, up to 7% have recurrent disease needing retreatment within 1 to 2 years (1). With the help of DNA fingerprinting of Mycobacterium tuberculosis, it has been shown that some of these recurrences are not treatment failures but rather represent reinfection with a different strain (2eC4). It is unclear what proportion of recurrent cases is caused by reinfection, because published studies each have fewer than 40 cases with DNA fingerprint in two disease episodes, and methods differ widely (3). In studies with at least eight cases, the proportion of reinfection cases among recurrent cases was reported to be from 0 to 33% in low-incidence areas (5eC10), 12 to 75% in medium-incidence areas (11eC12), and 23 to 75% in high-incidence areas (4, 13eC17). Disease attributable to reinfection is more common in people who are HIV-positive (9, 14, 16eC18).
It is hypothesized that TB with one strain will protect, at least partially, against subsequent reinfection disease with another strain (18eC20). If true, one could expect that the disease rate attributable to reinfection among cured patients would be lower than the incidence rate of new TB in a given population. Only one study reported the rate of reinfection disease after cure, which was 0.35 per 100 person-years (PYRS) in non-HIV infected individuals and 11 of 100 PYRS in patients with HIV (16, 21). However, this study took place in an unusual workplace setting (South African goldmines) with an exceptionally high incidence rate of TB and a high prevalence of silicosis as well as HIV.
The objective of our study was to determine the incidence rate of TB attributable to reinfection among successfully treated patients in an epidemiologic fieldsite in Cape Town, South Africa, with a long duration of follow-up (median, 5 years). A high proportion of reinfections among recurrent cases was previously reported from this setting (4). We have now followed up largely the same and all other patients in the area for 3 additional years, using bacterial DNA fingerprinting to estimate the incidence of reinfection disease.
This study has been presented at conferences in Paris (2003) (22), Prague (2003) (23), Geneva (2004) (24), and Amsterdam (25), and has been reported in the form of abstracts.
METHODS
Setting
The epidemiologic fieldsite of Ravensmead/Uitsig, two adjacent urban communities of Cape Town, has been extensively described (see online supplement) (4, 26eC28). The incidence of new smear- and/or culture-positive disease was on average 313 per 100,000 population per year (1993eC1998). Unfortunately, the HIV status of most patients is unknown. The HIV seroprevalence among women attending antenatal clinics increased from 1.2 to 5.2% between 1994 and 1998. In the study area, the prevalence of HIV infection among new patients with smear-positive TB was 11% in 1999eC2000.
Study Population
All bacteriologically confirmed (by smear or culture) patients with TB who were resident in and reported to the clinics in the fieldsite between 1993 and 1998 were eligible for this retrospective analysis (see online supplement). We included patients with pulmonary TB as well as extrapulmonary TB. We included new patients and patients who had been treated before (retreatment patients). Patients with multidrug resistance in any disease episode were excluded, as well as those who, while on treatment, remained or became smear-positive 5 months or more after starting treatment (treatment failure) and those who were transferred out or died during treatment.
M. tuberculosis Isolates
Isolates of M. tuberculosis were obtained from the majority of patients diagnosed between 1993 and 1998. DNA fingerprinting was determined by IS6110-based restriction fragment length polymorphism (see online supplement for definitions) (29). Strains harboring fewer than five copies of IS6110 copies were also subjected to subtyping with MTB484 (29). The laboratory error rate was 3.4%, as reported previously (4).
Follow-up Period
Patients who had at least one DNA fingerprint were enrolled at the time of the disease episode of their first fingerprint. Patients were followed up until their first bacteriologically confirmed recurrent episode, or until the end of the study period (December 2001). Consequently, the follow-up period varies from 3 to 8.5 years. Recurrent episodes of TB were diagnosed through passive case finding, because there was no active surveillance during the follow-up period. Recurrence and reinfection rates are underestimates because we could not determine mortality or emigration (see online supplement).
Definitions
World Health Organization definitions were used to determine treatment outcome (see online supplement) (30). We defined recurrent disease as a bacteriologically confirmed disease episode needing retreatment after a patient was successfully treated or defaulted during a previous disease episode. Disease attributable to reinfection was defined as a recurrent disease episode with an M. tuberculosis strain different than that found in the disease episode at enrollment, unlikely to be from evolution. To be conservative, we assumed that recurrences with the same strain or a strain with fewer than five bands' difference were attributable to relapse.
Rates
The calculation of recurrence rates, confirmed reinfection rates, and likely reinfection rates is described in the online supplement. We calculated the representativeness of enrolled patients (those with one DNA fingerprint) for all patients, and used only patients with DNA fingerprints to calculate reinfection rates.
To guard against the possibility that some cases of recurrence represented laboratory error, the proportions of recurrent TB attributable to reinfection were recalculated more strictly (i.e., restricted to patients for whom the second episode was associated with two or more sputum samples positive by culture or smear microscopy) (3, 16).
The Kaplan-Meier method was used to construct survival curves, with survival defined as being free of active TB. Cox regression was used to assess if age, sex, smear, location of TB (pulmonary TB or extrapulmonary TB), and outcome were risk factors for relapse and reinfection in successfully treated patients with at least one DNA fingerprint (31).
RESULTS
Description of Patients
A total of 1,093 bacteriologically confirmed patients were diagnosed between 1993 and 1998, of whom 86 had multidrug resistance in any disease episode during this study and were excluded. Of the remaining 1,007 patients, 11 were excluded because of treatment failure, 38 were excluded because they died during treatment, and 61 were excluded because they were transferred during treatment (Figure 1). Of the remaining 897 patients, 836 (93%) had a positive culture; of these patients, 612 (73%) had a DNA fingerprint. These 612 patients were enrolled. Compared with patients without DNA fingerprint data, patients with a DNA fingerprint were less often children younger than 15 years, patients with extrapulmonary TB, and patients who completed treatment without confirmation of cure (Table 1). Recurrence rates in patients with and without DNA fingerprint were similar (Table 1).
Recurrences
The median duration of follow-up was 5.2 years in the enrolled patients. During follow-up, 108 of 612 (18%) of these patients had a recurrence, 61 of 447 (14%) after successful treatment and 47 of 165 (28%) after default. The recurrence rate after successful treatment (2.7/100 PYRS) was lower than that after default (6.5/100 PYRS; hazard ratio, 2.1; 95% confidence interval, 1.5eC2.9). There was no significant difference between the recurrence rates after cure and after treatment completion. The recurrence rate was not significantly different between retreatment (3.4/100 PYRS) and new patients after successful treatment (2.4/100 PYRS; hazard ratio, 0.72; 95% confidence interval, 0.46eC1.13), but it was higher in retreatment than in new patients after default (8.5 vs. 4.6/100 PYRS; hazard ratio, 1.71; 95% confidence interval, 1.01eC2.89).
Reinfections
Of the 108 patients with a recurrence, 68 (63%) had a DNA fingerprint available from their first and second episodes. These 68 were representative of all recurrences with respect to year of diagnosis, age, sex, patient category, and sputum smear (Table 2). However, patients with disease recurrence after default were more likely to have a second DNA fingerprint than patients with a recurrence after successful treatment (Table 2). Among the 68 patients with a DNA fingerprint in their second episode, reinfection occurred in 24 of 31 (77%) of those successfully treated, and in 3 of 37 (8%) of defaulters using IS6110-based typing. Ten of 68 patients showed a strain with five or fewer copies of IS6110 (low copy numbers) in both episodes. These were all among defaulters. Subtyping strains from these 10 patients with MTB484 showed one reinfection, increasing the number of reinfections among defaulters to 4 of 37 (11%; Table 3). Six of 68 patients had a low IS6110 band copy number isolate in one of their episodes, but a strain with more than five bands in the other episode. These were all among the successfully treated patients.
The rates of TB attributable to confirmed and likely reinfection after successful treatment were 1.1 and 2.2 per 100 PYRS, respectively (Table 3). Rates of confirmed reinfection disease were the same in cured patients and in those with treatment completion (hazard ratio, 1.01; 95% confidence interval, 0.38eC2.69). The rate of reinfection disease (2.2/100 PYRS) was approximately seven times the crude incidence rate (313/100,000) and approximately four times the age-adjusted incidence rate of new TB (515/100,000). The disease rate attributable to confirmed reinfection seemed to remain constant for at least 7 years after the end of treatment (Figure 2).
There was no significant difference in rates of confirmed reinfection between those who had previously defaulted and those who had been successfully treated (hazard ratio, 0.54; 95% confidence interval, 0.19eC1.57; Figure 2).
To exclude the possibility that patients with a single positive culture might represent laboratory error, we repeated the analysis limited to patients with two or more positive samples in the second episode. In this restricted analysis, the proportion of recurrences attributable to reinfection remained unchanged: 17 of 21 (81%) after successful treatment and 3 of 23 (13%) after default. Using this subset of cases, confirmed and likely reinfection disease rates after successful treatment hardly changed: they were still 0.8 and 2.2 per 100 PYRS, respectively.
Risk Factors for Reinfection
Age, sex, smear positivity, and being a new or retreatment patient at enrollment were not found to be risk factors for reinfection disease. Among patients successfully treated, no risk factors for recurrence were identified.
DISCUSSION
This study shows that in a high-incidence area the risk of TB attributable to reinfection after successful treatment was approximately 2% per annum. The rate of reinfection disease was approximately seven times the crude incidence rate and approximately four times the age-adjusted incidence rate of new TB. This suggests that individuals who have been successfully treated for TB are at an increased risk of developing TB again, rather than being protected against subsequent episodes after reinfection. Moreover, the disease rate attributable to confirmed reinfection seemed to remain constant for at least 7 years after the end of treatment, consistent with a susceptibility to development of new disease in the face of ongoing risk of reinfection (Figure 2). These results represent an extension of those reported in earlier studies and demonstrate that most TB in this area results from recent infection or reinfection (26). Of considerable concern is the observation that most of this transmission takes place outside the household (27).
The estimated rate of reinfection disease is a minimum estimate, because it is not known how many people have died or moved during the follow-up period. We expect that during a follow-up period of 5 years, approximately 10 to 20% have moved. The reinfection disease rates could also be somewhat underestimated because of reinfection with the same strain, but this effect is not expected to be large because strain diversity in this community is high (28). Because the HIV prevalence is relatively low, we believe that most reinfections cannot be explained by HIV status. Of note, the rate that we estimate lies between what has been observed for HIV-negative and HIV-positive populations in trials of short-course chemotherapy (32, 33). Our relapse rate may be somewhat higher if compliance is lower in the community than in the conditions applied for clinical trials. Conversely, our reinfection rate is likely higher because of a greater annual risk of TB infection in our study area.
If laboratory cross-contamination or administrative errors occur, this would lead to an overestimate of the proportion of recurrences attributed to reinfection. In our study, we believe that these errors did not play a major role for two reasons. First, the proportion of reinfections among recurrences was much higher after successful treatment (77%) than after default (11%). There is no reason why recurrences after successful treatment would be more often involved in laboratory or administrative errors than recurrences after default. Second, laboratory cross-contamination and administrative errors are most likely to explain reinfection based on isolated positive cultures. However, when excluding episodes with isolated positive cultures from the analysis, the proportion of recurrences attributed to reinfection remained similar. Furthermore, the rate of reinfection TB after successful treatment was still twice the incidence rate of new TB.
The main limitation of this study was the low percentage of enrolled patients who had DNA fingerprints available: 68% of all bacteriologically confirmed patients and 63% of patients during the recurrence. This missing data cause a twofold difference between confirmed and likely reinfection disease rate. However, we do not believe this missing data could have biased our results because the patients with DNA fingerprints were representative for those without, except for patients with lower bacterial load (< 15 years of age and patients with extrapulmonary TB). Patients after default were more likely to have a DNA fingerprint at recurrence than patients after successful treatment, perhaps because of more intensive follow-up by the health system. This should not affect the proportion of recurrences that are attributable to reinfection because the availability of a DNA fingerprint did not depend on whether disease was attributable to reinfection or relapse. However, some patients may have had a dual infection with M. tuberculosis, of which one strain was cultured during the first disease episode and the other during the recurrent episode (34, 35). By DNA fingerprinting, it is impossible to distinguish between reinfection with a new strain and dual infection followed by reactivation of that same strain.
A possible bias in our study is that patients reporting for their first episode may be more likely to report for their second, thus leading to an overestimate of rate ratio of reinfection disease and new disease. Furthermore, because case detection is passive, microcommunities with undiagnosed disease may be missed. However, to explain a rate ratio of four, this would require that all recurrent patients are diagnosed and only 25% of new patients are diagnosed. The fieldsite has two clinics in an area of less than 4 km2, and the case detection rate is estimated at over 50%. Therefore, we do not consider selection bias a sufficient explanation for our results.
We conclude that people who have been treated successfully for TB are at higher risk of developing TB from reinfection than the general population. This suggests that a subgroup of individuals is intrinsically vulnerable to TB. Further study is needed to find out whether the high risk of reinfection disease is attributable to a high risk of reinfection or a high risk of breakdown to disease. This is currently difficult because there is no test that differentiates between an old infection (e.g., after healed disease) and a recent infection. Possible risk factors for progression to disease should be studied in ex-patients. These may include socioeconomic factors, genetic risk factors (36, 37), lung damage caused by the previous episode, and perhaps smoking and drug abuse. Regardless, this result challenges the hypothesis that, in immunocompetent persons, infection with one strain of M. tuberculosis protects against disease attributable to subsequent reinfection with another strain (18eC20). The failure of natural disease to protect against reinfection disease at a later point may partially explain the relative ineffectiveness of vaccination with bacillus Calmette-Gueerin (19). For national TB control programs in areas with a high infection risk, patients who have been successfully treated for TB should be made aware of their high risk of recurrent disease, and contact tracing should get more attention.
Acknowledgments
The authors thank the nurses and patients of the Ravensmead and Uitsig clinics and Dr. I. Toms, Director City Health, Cape Town. They thank Dr. R. Gie for dedicated reading of all chest x-rays and Ms. V. Chihota and Mr. S. Ndabambi and the TB research team for their valuable assistance.
This article has an online supplement, which is accessible from this issue's table of contents at www.atsjournals.org
We dedicate this paper to the memory of Madalene Richardson, who sadly passed away while this manuscript was under review.
REFERENCES
El-Sadr WM, Perlman DC, Denning E, Matts JP, Cohn DL. A review of efficacy studies of 6-month short-course therapy for tuberculosis among patients infected with human immunodeficiency virus: differences in study outcomes. Clin Infect Dis 2001;32:623eC632.
Fine PEM, Small PM. Exogenous reinfection in tuberculosis. N Engl J Med 1999;341:1226eC1227.
Lambert ML, Hasker E, Van Deun A, Roberfroid D, Boelaert M, Van der Stuyft P. Recurrence in tuberculosis: relapse or reinfection Lancet Infect Dis 2003;3:282eC287.
Van Rie A, Warren R, Richardson M, Victor TC, Gie RP, Enarson DA, Beyers N, van Helden PD. Exogenous reinfection as a cause of recurrent tuberculosis after curative treatment. N Engl J Med 1999;341:1174eC1179.
Bandera A, Gori A, Catozzi L, Degli Esposti A, Marchetti G, Molteni C, Ferrario G, Codecasa L, Penati V, Matteelli A, et al. Molecular epidemiology study of exogenous reinfection in an area with a low incidence of tuberculosis. J Clin Microbiol 2001;39:2213eC2218.
De Boer AS, van Soolingen D, Borgdorff MW. Recurrent tuberculosis due to exogenous reinfection. N Engl J Med 2000;342:1050eC1051.
de Viedma DG, Marin M, Hernangomez S, Diaz M, Ruiz Serrano MJ, Alcala L, Bouza E. Tuberculosis recurrences: reinfection plays a role in a population whose clinical/epidemiological characteristics do not favor reinfection. Arch Intern Med 2002;162:1873eC1879.
Sudre P, Pfyffer GE, Bodmer T, Prod'hom G, Furrer H, Bassetti S, Bernasconi E, Matter L, Telenti A, Strassle A, et al. Molecular epidemiology of tuberculosis among HIV-infected persons in Switzerland: a countrywide 9-year cohort study. Swiss HIV Cohort Study. Infection 1999;27:323eC330.
Benator D, Bhattacharya M, Bozeman L, Burman W, Cantazaro A, Chaisson R, Gordin F, Horsburgh CR, Horton J, Khan A, et al., for the Tuberculosis Trials Consortium. Rifapentine and isoniazid once a week versus rifampicin and isoniazid twice a week for treatment of drug-susceptible pulmonary tuberculosis in HIV-negative patients: a randomised clinical trial. Lancet 2002;360:528eC534.
Jasmer RM, Bozeman L, Schwartzman K, Cave MD, Saukkonen JJ, Metchock B, Khan A, Burman WJ, and the Tuberculosis Trials Consortium. Recurrent tuberculosis in the United States and Canada: relapse or reinfection Am J Respir Crit Care Med 2004;170:1360eC1366.
Caminero JA, Pena MJ, Campos-Herrero MI, Rodriguez JC, Afonso O, Martin C, Pavon JM, Torres MJ, Burgos M, Cabrera P, et al. Exogenous reinfection with tuberculosis on a European island with a moderate incidence of disease. Am J Respir Crit Care Med 2001;163:717eC720.
Das S, Chan SL, Allen BW, Mitchison DA, Lowrie DB. Application of DNA fingerprinting with IS986 to sequential mycobacterial isolates obtained from pulmonary tuberculosis patients in Hong Kong before, during and after short-course chemotherapy. Tuber Lung Dis 1993;74:47eC51.
Das S, Paramasivan CN, Lowrie DB, Prabhakar R, Narayanan PR. IS6110 restriction fragment length polymorphism typing of clinical isolates of Mycobacterium tuberculosis from patients with pulmonary tuberculosis in Madras, South India. Tuber Lung Dis 1995;76:550eC554.
Fitzpatrick LK, Okwera A, Mugerwa R, Ridzon R, Ehiner J, Onorato I. An investigation of suspected exogenous reinfection in tuberculosis patients in Kampala, Uganda. Int J Tuberc Lung Dis 2002;6:550eC552.
Kruuner A, Pehme L, Ghebremichael S, Koivula T, Hoffner SE, Mikelsaar M. Use of molecular techniques to distinguish between treatment failure and exogenous reinfection with Mycobacterium tuberculosis. Clin Infect Dis 2002;35:146eC155.
Sonnenberg P, Murray J, Glynn JR, Shearer S, Kambashi B, Godfrey-Faussett P. HIV-1 and recurrence, relapse, and reinfection of tuberculosis after cure: a cohort study in South African mineworkers. Lancet 2001;358:1687eC1693.
Glynn JR, Yates MD, Crampin AC, Ngwira BM, Mwaungulu FD, Black GF, Chaguluka SD, Mwafulirwa DT, Floyd S, Murphy C, et al. DNA fingerprint changes in tuberculosis: reinfection, evolution, or laboratory error J Infect Dis 2004;190:1158eC1166.
Barnes PF, Cave MD. Molecular epidemiology of tuberculosis. N Engl J Med 2003;349:1149eC1156.
Bjartveit K. Olaf Scheel and Johannes Heimbeck: their contribution to understanding the pathogenesis and prevention of tuberculosis. Int J Tuberc Lung Dis 2003;7:306eC311.
Vynnycky E, Fine PE. The natural history of tuberculosis: the implications of age-dependent risks of disease and the role of reinfection. Epidemiol Infect 1997;119:183eC201.
Sonnenberg P, Godfrey-Faussett P, Glynn JR, Shearer S, Murray J. Authors reply on ‘HIV-1 and tuberculosis infection’. Lancet 2002;359:1619eC1620.
Verver S, Warren RM, Borgdorff MW, Beyers N, Richardson M, van Helden P. Reinfection and relapse in a high incidence suburb. Int J Tuberc Lung Dis 2003;7:S210.
Verver S, Warren RM, Borgdorff MW, Beyers N, Richardson M, van Helden P. Reinfection and relapse of tuberculosis in a high incidence urban community. Presented at the third meeting of Concerted Action Project: New Generation Genetic Markers and Techniques for the Epidemiology of Tuberculosis, Prague, Czech Republic, 2003.
Verver S, Warren RM, Beyers N, Richardson M, van der Spuy GD, Borgdorff MW, Enarson DA, Behr MA, van Helden P. The rate of reinfection tuberculosis after successful treatment is higher than the rate of new tuberculosis. Presented at the annual TSRU conference in Geneva, Switzerland, 2004.
Verver S, Warren RM, Beyers N, Richardson M, van der Spuy GD, Borgdorff MW, Enarson DA, Behr MA, van Helden P. The rate of reinfection tuberculosis after successful treatment is higher than the tuberculosis incidence. Presented at the symposium Research for Better Health in Amsterdam, The Netherlands, 2004.
Verver S, Warren RM, Munch Z, Vynnycky E, Van Helden PD, Richardson M, Van der Spuy GD, Enarson DA, Borgdorff MW, Behr MA, et al. Transmission of tuberculosis in a high incidence urban community in South Africa. Int J Epidemiol 2004;33:351eC357.
Verver S, Warren RM, Munch Z, Richardson M, van der Spuy GD, Borgdorff MW, Behr MA, Beyers N, van Helden PD. Proportion of tuberculosis transmission that takes place in households in a high-incidence area. Lancet 2004;363:212eC214.
Warren RM, Richardson M, van der Spuy GD, Victor T, Sampson S, Beyers N, van Helden PD. DNA fingerprinting and molecular epidemiology of tuberculosis: use and interpretation in an epidemic setting. Electrophoresis 1999;20:1807eC1812.
Van Embden JDA, Cave MD, Crawford JT, Dale JW, Eisenach KD, Gicquel B, Hermans P, Martin C, McAdam R, Shinnick TM, et al. Strain identification of Mycobacterium tuberculosis by DNA fingerprinting: recommendations for a standardized methodology. J Clin Microbiol 1993;31:406eC409.
World Health Organization, International Union against Tuberculosis and Lung Disease, Royal Netherlands Tuberculosis Association. Revised international definitions in tuberculosis control. Int J Tuberc Lung Dis 2001;5:213eC215.
Kleinbaum DG. Survival analysis: a self-learning text (statistics in the health sciences). New York: Springer; 1996.
Korenromp EL, Scano F, Williams BG, Dye C, Nunn P. Effects of human immundeficiency virus infection on recurrence of tuberculosis after rifampicin based treatment: an analytical review. Clin Infect Dis 2003;37:101eC112.
Perriens JH, St Louis ME, Mukadi YB, Brown C, Prignot J, Pouthier F, Portaels F, Willame JC, Mandala JK, Kaboto M, et al. Pulmonary tuberculosis in HIV-infected patients in Zaire: a controlled trial of treatment for either 6 or 12 months. N Engl J Med 1995;332:779eC784.
Warren RM, Victor TC, Streicher EM, Richardson M, Beyers N, Gey Van Pittius NC, Van Helden PD. Patients with active tuberculosis often have different strains in the same sputum specimen. Am J Respir Crit Care Med 2004;169:610eC614.
Yeh RW, Hopewell PC, Daley CL. Simultaneous infection with two strains of Mycobacterium tuberculosis identified by restriction fragment length polymorphism analysis. Int J Tuberc Lung Dis 1999;3:537eC539.
Comstock G. Tuberculosis in twins: a re-analysis of the Prophit survey. Am Rev Respir Dis 1978;117:621eC624.
Stead WW. Genetics and resistance to tuberculosis. Ann Intern Med 1992;116:937eC941.(Suzanne Verver, Robin M. )
MRC Center for Molecular and Cellular Biology, Department of Medical Biochemistry, Stellenbosch University, Cape Town, South Africa
KNCV Tuberculosis Foundation, The Hague
Department of Infectious Diseases, Tropical Medicine and AIDS, Academic Medical Center, Amsterdam, The Netherlands
International Union against Tuberculosis and Lung Disease (IUATLD), Paris, France
Division of Infectious Diseases, Department of Medicine, McGill University Health Center, Montreal, Quebec, Canada
ABSTRACT
Rationale: In a higheCtuberculosis (TB) incidence area of Cape Town, South Africa, there is a very high rate of unexplained recurrent TB. The incidence of new bacteriologically confirmed disease in the area is 313 per 100,000 individuals. Objective: To estimate the rate of recurrent TB attributable to reinfection after successful treatment. Methods: All patients with reported TB in the area between 1993 and 1998 were followed up to 2001 for disease needing retreatment (recurrences). Patients who were multi-drugeCresistant or who had treatment failure, were transferred, or died during treatment were excluded. Analysis was restricted to patients for whom DNA fingerprinting of their Mycobacterium tuberculosis isolates was obtained. Reinfection TB was defined as a recurrent TB episode in which the strains of the separate episodes differed by more than four bands. Measurements and Main Results: 612 of 897 (68%) patients had a DNA fingerprint available at enrollment. Median duration of follow-up was 5.2 years. Recurrent TB occurred in 108 of 612 (18%) patients, of whom 61 of 447 (14%) experienced recurrence after successful treatment, and 47 of 165 (28%) experience recurrence after default. Of the 108 patients with recurrent TB, 68 (63%) had a DNA fingerprint in the second episode. Among these patients, 24 of 31 (77%) recurrences after successful treatment and 4 of 37 (11%) recurrences after default were attributable to reinfection. The reinfection disease rate after successful treatment was estimated at 2.2 per 100 person-years. Conclusions: The age-adjusted incidence rate of TB attributable to reinfection after successful treatment was four times that of new TB. People who had TB once are at a strongly increased risk of developing TB when reinfected.
Key Words: incidence molecular epidemiology Mycobacterium tuberculosis recurrence survival analysis
Of patients with tuberculosis (TB) who are cured after short-course treatment in trial conditions, up to 7% have recurrent disease needing retreatment within 1 to 2 years (1). With the help of DNA fingerprinting of Mycobacterium tuberculosis, it has been shown that some of these recurrences are not treatment failures but rather represent reinfection with a different strain (2eC4). It is unclear what proportion of recurrent cases is caused by reinfection, because published studies each have fewer than 40 cases with DNA fingerprint in two disease episodes, and methods differ widely (3). In studies with at least eight cases, the proportion of reinfection cases among recurrent cases was reported to be from 0 to 33% in low-incidence areas (5eC10), 12 to 75% in medium-incidence areas (11eC12), and 23 to 75% in high-incidence areas (4, 13eC17). Disease attributable to reinfection is more common in people who are HIV-positive (9, 14, 16eC18).
It is hypothesized that TB with one strain will protect, at least partially, against subsequent reinfection disease with another strain (18eC20). If true, one could expect that the disease rate attributable to reinfection among cured patients would be lower than the incidence rate of new TB in a given population. Only one study reported the rate of reinfection disease after cure, which was 0.35 per 100 person-years (PYRS) in non-HIV infected individuals and 11 of 100 PYRS in patients with HIV (16, 21). However, this study took place in an unusual workplace setting (South African goldmines) with an exceptionally high incidence rate of TB and a high prevalence of silicosis as well as HIV.
The objective of our study was to determine the incidence rate of TB attributable to reinfection among successfully treated patients in an epidemiologic fieldsite in Cape Town, South Africa, with a long duration of follow-up (median, 5 years). A high proportion of reinfections among recurrent cases was previously reported from this setting (4). We have now followed up largely the same and all other patients in the area for 3 additional years, using bacterial DNA fingerprinting to estimate the incidence of reinfection disease.
This study has been presented at conferences in Paris (2003) (22), Prague (2003) (23), Geneva (2004) (24), and Amsterdam (25), and has been reported in the form of abstracts.
METHODS
Setting
The epidemiologic fieldsite of Ravensmead/Uitsig, two adjacent urban communities of Cape Town, has been extensively described (see online supplement) (4, 26eC28). The incidence of new smear- and/or culture-positive disease was on average 313 per 100,000 population per year (1993eC1998). Unfortunately, the HIV status of most patients is unknown. The HIV seroprevalence among women attending antenatal clinics increased from 1.2 to 5.2% between 1994 and 1998. In the study area, the prevalence of HIV infection among new patients with smear-positive TB was 11% in 1999eC2000.
Study Population
All bacteriologically confirmed (by smear or culture) patients with TB who were resident in and reported to the clinics in the fieldsite between 1993 and 1998 were eligible for this retrospective analysis (see online supplement). We included patients with pulmonary TB as well as extrapulmonary TB. We included new patients and patients who had been treated before (retreatment patients). Patients with multidrug resistance in any disease episode were excluded, as well as those who, while on treatment, remained or became smear-positive 5 months or more after starting treatment (treatment failure) and those who were transferred out or died during treatment.
M. tuberculosis Isolates
Isolates of M. tuberculosis were obtained from the majority of patients diagnosed between 1993 and 1998. DNA fingerprinting was determined by IS6110-based restriction fragment length polymorphism (see online supplement for definitions) (29). Strains harboring fewer than five copies of IS6110 copies were also subjected to subtyping with MTB484 (29). The laboratory error rate was 3.4%, as reported previously (4).
Follow-up Period
Patients who had at least one DNA fingerprint were enrolled at the time of the disease episode of their first fingerprint. Patients were followed up until their first bacteriologically confirmed recurrent episode, or until the end of the study period (December 2001). Consequently, the follow-up period varies from 3 to 8.5 years. Recurrent episodes of TB were diagnosed through passive case finding, because there was no active surveillance during the follow-up period. Recurrence and reinfection rates are underestimates because we could not determine mortality or emigration (see online supplement).
Definitions
World Health Organization definitions were used to determine treatment outcome (see online supplement) (30). We defined recurrent disease as a bacteriologically confirmed disease episode needing retreatment after a patient was successfully treated or defaulted during a previous disease episode. Disease attributable to reinfection was defined as a recurrent disease episode with an M. tuberculosis strain different than that found in the disease episode at enrollment, unlikely to be from evolution. To be conservative, we assumed that recurrences with the same strain or a strain with fewer than five bands' difference were attributable to relapse.
Rates
The calculation of recurrence rates, confirmed reinfection rates, and likely reinfection rates is described in the online supplement. We calculated the representativeness of enrolled patients (those with one DNA fingerprint) for all patients, and used only patients with DNA fingerprints to calculate reinfection rates.
To guard against the possibility that some cases of recurrence represented laboratory error, the proportions of recurrent TB attributable to reinfection were recalculated more strictly (i.e., restricted to patients for whom the second episode was associated with two or more sputum samples positive by culture or smear microscopy) (3, 16).
The Kaplan-Meier method was used to construct survival curves, with survival defined as being free of active TB. Cox regression was used to assess if age, sex, smear, location of TB (pulmonary TB or extrapulmonary TB), and outcome were risk factors for relapse and reinfection in successfully treated patients with at least one DNA fingerprint (31).
RESULTS
Description of Patients
A total of 1,093 bacteriologically confirmed patients were diagnosed between 1993 and 1998, of whom 86 had multidrug resistance in any disease episode during this study and were excluded. Of the remaining 1,007 patients, 11 were excluded because of treatment failure, 38 were excluded because they died during treatment, and 61 were excluded because they were transferred during treatment (Figure 1). Of the remaining 897 patients, 836 (93%) had a positive culture; of these patients, 612 (73%) had a DNA fingerprint. These 612 patients were enrolled. Compared with patients without DNA fingerprint data, patients with a DNA fingerprint were less often children younger than 15 years, patients with extrapulmonary TB, and patients who completed treatment without confirmation of cure (Table 1). Recurrence rates in patients with and without DNA fingerprint were similar (Table 1).
Recurrences
The median duration of follow-up was 5.2 years in the enrolled patients. During follow-up, 108 of 612 (18%) of these patients had a recurrence, 61 of 447 (14%) after successful treatment and 47 of 165 (28%) after default. The recurrence rate after successful treatment (2.7/100 PYRS) was lower than that after default (6.5/100 PYRS; hazard ratio, 2.1; 95% confidence interval, 1.5eC2.9). There was no significant difference between the recurrence rates after cure and after treatment completion. The recurrence rate was not significantly different between retreatment (3.4/100 PYRS) and new patients after successful treatment (2.4/100 PYRS; hazard ratio, 0.72; 95% confidence interval, 0.46eC1.13), but it was higher in retreatment than in new patients after default (8.5 vs. 4.6/100 PYRS; hazard ratio, 1.71; 95% confidence interval, 1.01eC2.89).
Reinfections
Of the 108 patients with a recurrence, 68 (63%) had a DNA fingerprint available from their first and second episodes. These 68 were representative of all recurrences with respect to year of diagnosis, age, sex, patient category, and sputum smear (Table 2). However, patients with disease recurrence after default were more likely to have a second DNA fingerprint than patients with a recurrence after successful treatment (Table 2). Among the 68 patients with a DNA fingerprint in their second episode, reinfection occurred in 24 of 31 (77%) of those successfully treated, and in 3 of 37 (8%) of defaulters using IS6110-based typing. Ten of 68 patients showed a strain with five or fewer copies of IS6110 (low copy numbers) in both episodes. These were all among defaulters. Subtyping strains from these 10 patients with MTB484 showed one reinfection, increasing the number of reinfections among defaulters to 4 of 37 (11%; Table 3). Six of 68 patients had a low IS6110 band copy number isolate in one of their episodes, but a strain with more than five bands in the other episode. These were all among the successfully treated patients.
The rates of TB attributable to confirmed and likely reinfection after successful treatment were 1.1 and 2.2 per 100 PYRS, respectively (Table 3). Rates of confirmed reinfection disease were the same in cured patients and in those with treatment completion (hazard ratio, 1.01; 95% confidence interval, 0.38eC2.69). The rate of reinfection disease (2.2/100 PYRS) was approximately seven times the crude incidence rate (313/100,000) and approximately four times the age-adjusted incidence rate of new TB (515/100,000). The disease rate attributable to confirmed reinfection seemed to remain constant for at least 7 years after the end of treatment (Figure 2).
There was no significant difference in rates of confirmed reinfection between those who had previously defaulted and those who had been successfully treated (hazard ratio, 0.54; 95% confidence interval, 0.19eC1.57; Figure 2).
To exclude the possibility that patients with a single positive culture might represent laboratory error, we repeated the analysis limited to patients with two or more positive samples in the second episode. In this restricted analysis, the proportion of recurrences attributable to reinfection remained unchanged: 17 of 21 (81%) after successful treatment and 3 of 23 (13%) after default. Using this subset of cases, confirmed and likely reinfection disease rates after successful treatment hardly changed: they were still 0.8 and 2.2 per 100 PYRS, respectively.
Risk Factors for Reinfection
Age, sex, smear positivity, and being a new or retreatment patient at enrollment were not found to be risk factors for reinfection disease. Among patients successfully treated, no risk factors for recurrence were identified.
DISCUSSION
This study shows that in a high-incidence area the risk of TB attributable to reinfection after successful treatment was approximately 2% per annum. The rate of reinfection disease was approximately seven times the crude incidence rate and approximately four times the age-adjusted incidence rate of new TB. This suggests that individuals who have been successfully treated for TB are at an increased risk of developing TB again, rather than being protected against subsequent episodes after reinfection. Moreover, the disease rate attributable to confirmed reinfection seemed to remain constant for at least 7 years after the end of treatment, consistent with a susceptibility to development of new disease in the face of ongoing risk of reinfection (Figure 2). These results represent an extension of those reported in earlier studies and demonstrate that most TB in this area results from recent infection or reinfection (26). Of considerable concern is the observation that most of this transmission takes place outside the household (27).
The estimated rate of reinfection disease is a minimum estimate, because it is not known how many people have died or moved during the follow-up period. We expect that during a follow-up period of 5 years, approximately 10 to 20% have moved. The reinfection disease rates could also be somewhat underestimated because of reinfection with the same strain, but this effect is not expected to be large because strain diversity in this community is high (28). Because the HIV prevalence is relatively low, we believe that most reinfections cannot be explained by HIV status. Of note, the rate that we estimate lies between what has been observed for HIV-negative and HIV-positive populations in trials of short-course chemotherapy (32, 33). Our relapse rate may be somewhat higher if compliance is lower in the community than in the conditions applied for clinical trials. Conversely, our reinfection rate is likely higher because of a greater annual risk of TB infection in our study area.
If laboratory cross-contamination or administrative errors occur, this would lead to an overestimate of the proportion of recurrences attributed to reinfection. In our study, we believe that these errors did not play a major role for two reasons. First, the proportion of reinfections among recurrences was much higher after successful treatment (77%) than after default (11%). There is no reason why recurrences after successful treatment would be more often involved in laboratory or administrative errors than recurrences after default. Second, laboratory cross-contamination and administrative errors are most likely to explain reinfection based on isolated positive cultures. However, when excluding episodes with isolated positive cultures from the analysis, the proportion of recurrences attributed to reinfection remained similar. Furthermore, the rate of reinfection TB after successful treatment was still twice the incidence rate of new TB.
The main limitation of this study was the low percentage of enrolled patients who had DNA fingerprints available: 68% of all bacteriologically confirmed patients and 63% of patients during the recurrence. This missing data cause a twofold difference between confirmed and likely reinfection disease rate. However, we do not believe this missing data could have biased our results because the patients with DNA fingerprints were representative for those without, except for patients with lower bacterial load (< 15 years of age and patients with extrapulmonary TB). Patients after default were more likely to have a DNA fingerprint at recurrence than patients after successful treatment, perhaps because of more intensive follow-up by the health system. This should not affect the proportion of recurrences that are attributable to reinfection because the availability of a DNA fingerprint did not depend on whether disease was attributable to reinfection or relapse. However, some patients may have had a dual infection with M. tuberculosis, of which one strain was cultured during the first disease episode and the other during the recurrent episode (34, 35). By DNA fingerprinting, it is impossible to distinguish between reinfection with a new strain and dual infection followed by reactivation of that same strain.
A possible bias in our study is that patients reporting for their first episode may be more likely to report for their second, thus leading to an overestimate of rate ratio of reinfection disease and new disease. Furthermore, because case detection is passive, microcommunities with undiagnosed disease may be missed. However, to explain a rate ratio of four, this would require that all recurrent patients are diagnosed and only 25% of new patients are diagnosed. The fieldsite has two clinics in an area of less than 4 km2, and the case detection rate is estimated at over 50%. Therefore, we do not consider selection bias a sufficient explanation for our results.
We conclude that people who have been treated successfully for TB are at higher risk of developing TB from reinfection than the general population. This suggests that a subgroup of individuals is intrinsically vulnerable to TB. Further study is needed to find out whether the high risk of reinfection disease is attributable to a high risk of reinfection or a high risk of breakdown to disease. This is currently difficult because there is no test that differentiates between an old infection (e.g., after healed disease) and a recent infection. Possible risk factors for progression to disease should be studied in ex-patients. These may include socioeconomic factors, genetic risk factors (36, 37), lung damage caused by the previous episode, and perhaps smoking and drug abuse. Regardless, this result challenges the hypothesis that, in immunocompetent persons, infection with one strain of M. tuberculosis protects against disease attributable to subsequent reinfection with another strain (18eC20). The failure of natural disease to protect against reinfection disease at a later point may partially explain the relative ineffectiveness of vaccination with bacillus Calmette-Gueerin (19). For national TB control programs in areas with a high infection risk, patients who have been successfully treated for TB should be made aware of their high risk of recurrent disease, and contact tracing should get more attention.
Acknowledgments
The authors thank the nurses and patients of the Ravensmead and Uitsig clinics and Dr. I. Toms, Director City Health, Cape Town. They thank Dr. R. Gie for dedicated reading of all chest x-rays and Ms. V. Chihota and Mr. S. Ndabambi and the TB research team for their valuable assistance.
This article has an online supplement, which is accessible from this issue's table of contents at www.atsjournals.org
We dedicate this paper to the memory of Madalene Richardson, who sadly passed away while this manuscript was under review.
REFERENCES
El-Sadr WM, Perlman DC, Denning E, Matts JP, Cohn DL. A review of efficacy studies of 6-month short-course therapy for tuberculosis among patients infected with human immunodeficiency virus: differences in study outcomes. Clin Infect Dis 2001;32:623eC632.
Fine PEM, Small PM. Exogenous reinfection in tuberculosis. N Engl J Med 1999;341:1226eC1227.
Lambert ML, Hasker E, Van Deun A, Roberfroid D, Boelaert M, Van der Stuyft P. Recurrence in tuberculosis: relapse or reinfection Lancet Infect Dis 2003;3:282eC287.
Van Rie A, Warren R, Richardson M, Victor TC, Gie RP, Enarson DA, Beyers N, van Helden PD. Exogenous reinfection as a cause of recurrent tuberculosis after curative treatment. N Engl J Med 1999;341:1174eC1179.
Bandera A, Gori A, Catozzi L, Degli Esposti A, Marchetti G, Molteni C, Ferrario G, Codecasa L, Penati V, Matteelli A, et al. Molecular epidemiology study of exogenous reinfection in an area with a low incidence of tuberculosis. J Clin Microbiol 2001;39:2213eC2218.
De Boer AS, van Soolingen D, Borgdorff MW. Recurrent tuberculosis due to exogenous reinfection. N Engl J Med 2000;342:1050eC1051.
de Viedma DG, Marin M, Hernangomez S, Diaz M, Ruiz Serrano MJ, Alcala L, Bouza E. Tuberculosis recurrences: reinfection plays a role in a population whose clinical/epidemiological characteristics do not favor reinfection. Arch Intern Med 2002;162:1873eC1879.
Sudre P, Pfyffer GE, Bodmer T, Prod'hom G, Furrer H, Bassetti S, Bernasconi E, Matter L, Telenti A, Strassle A, et al. Molecular epidemiology of tuberculosis among HIV-infected persons in Switzerland: a countrywide 9-year cohort study. Swiss HIV Cohort Study. Infection 1999;27:323eC330.
Benator D, Bhattacharya M, Bozeman L, Burman W, Cantazaro A, Chaisson R, Gordin F, Horsburgh CR, Horton J, Khan A, et al., for the Tuberculosis Trials Consortium. Rifapentine and isoniazid once a week versus rifampicin and isoniazid twice a week for treatment of drug-susceptible pulmonary tuberculosis in HIV-negative patients: a randomised clinical trial. Lancet 2002;360:528eC534.
Jasmer RM, Bozeman L, Schwartzman K, Cave MD, Saukkonen JJ, Metchock B, Khan A, Burman WJ, and the Tuberculosis Trials Consortium. Recurrent tuberculosis in the United States and Canada: relapse or reinfection Am J Respir Crit Care Med 2004;170:1360eC1366.
Caminero JA, Pena MJ, Campos-Herrero MI, Rodriguez JC, Afonso O, Martin C, Pavon JM, Torres MJ, Burgos M, Cabrera P, et al. Exogenous reinfection with tuberculosis on a European island with a moderate incidence of disease. Am J Respir Crit Care Med 2001;163:717eC720.
Das S, Chan SL, Allen BW, Mitchison DA, Lowrie DB. Application of DNA fingerprinting with IS986 to sequential mycobacterial isolates obtained from pulmonary tuberculosis patients in Hong Kong before, during and after short-course chemotherapy. Tuber Lung Dis 1993;74:47eC51.
Das S, Paramasivan CN, Lowrie DB, Prabhakar R, Narayanan PR. IS6110 restriction fragment length polymorphism typing of clinical isolates of Mycobacterium tuberculosis from patients with pulmonary tuberculosis in Madras, South India. Tuber Lung Dis 1995;76:550eC554.
Fitzpatrick LK, Okwera A, Mugerwa R, Ridzon R, Ehiner J, Onorato I. An investigation of suspected exogenous reinfection in tuberculosis patients in Kampala, Uganda. Int J Tuberc Lung Dis 2002;6:550eC552.
Kruuner A, Pehme L, Ghebremichael S, Koivula T, Hoffner SE, Mikelsaar M. Use of molecular techniques to distinguish between treatment failure and exogenous reinfection with Mycobacterium tuberculosis. Clin Infect Dis 2002;35:146eC155.
Sonnenberg P, Murray J, Glynn JR, Shearer S, Kambashi B, Godfrey-Faussett P. HIV-1 and recurrence, relapse, and reinfection of tuberculosis after cure: a cohort study in South African mineworkers. Lancet 2001;358:1687eC1693.
Glynn JR, Yates MD, Crampin AC, Ngwira BM, Mwaungulu FD, Black GF, Chaguluka SD, Mwafulirwa DT, Floyd S, Murphy C, et al. DNA fingerprint changes in tuberculosis: reinfection, evolution, or laboratory error J Infect Dis 2004;190:1158eC1166.
Barnes PF, Cave MD. Molecular epidemiology of tuberculosis. N Engl J Med 2003;349:1149eC1156.
Bjartveit K. Olaf Scheel and Johannes Heimbeck: their contribution to understanding the pathogenesis and prevention of tuberculosis. Int J Tuberc Lung Dis 2003;7:306eC311.
Vynnycky E, Fine PE. The natural history of tuberculosis: the implications of age-dependent risks of disease and the role of reinfection. Epidemiol Infect 1997;119:183eC201.
Sonnenberg P, Godfrey-Faussett P, Glynn JR, Shearer S, Murray J. Authors reply on ‘HIV-1 and tuberculosis infection’. Lancet 2002;359:1619eC1620.
Verver S, Warren RM, Borgdorff MW, Beyers N, Richardson M, van Helden P. Reinfection and relapse in a high incidence suburb. Int J Tuberc Lung Dis 2003;7:S210.
Verver S, Warren RM, Borgdorff MW, Beyers N, Richardson M, van Helden P. Reinfection and relapse of tuberculosis in a high incidence urban community. Presented at the third meeting of Concerted Action Project: New Generation Genetic Markers and Techniques for the Epidemiology of Tuberculosis, Prague, Czech Republic, 2003.
Verver S, Warren RM, Beyers N, Richardson M, van der Spuy GD, Borgdorff MW, Enarson DA, Behr MA, van Helden P. The rate of reinfection tuberculosis after successful treatment is higher than the rate of new tuberculosis. Presented at the annual TSRU conference in Geneva, Switzerland, 2004.
Verver S, Warren RM, Beyers N, Richardson M, van der Spuy GD, Borgdorff MW, Enarson DA, Behr MA, van Helden P. The rate of reinfection tuberculosis after successful treatment is higher than the tuberculosis incidence. Presented at the symposium Research for Better Health in Amsterdam, The Netherlands, 2004.
Verver S, Warren RM, Munch Z, Vynnycky E, Van Helden PD, Richardson M, Van der Spuy GD, Enarson DA, Borgdorff MW, Behr MA, et al. Transmission of tuberculosis in a high incidence urban community in South Africa. Int J Epidemiol 2004;33:351eC357.
Verver S, Warren RM, Munch Z, Richardson M, van der Spuy GD, Borgdorff MW, Behr MA, Beyers N, van Helden PD. Proportion of tuberculosis transmission that takes place in households in a high-incidence area. Lancet 2004;363:212eC214.
Warren RM, Richardson M, van der Spuy GD, Victor T, Sampson S, Beyers N, van Helden PD. DNA fingerprinting and molecular epidemiology of tuberculosis: use and interpretation in an epidemic setting. Electrophoresis 1999;20:1807eC1812.
Van Embden JDA, Cave MD, Crawford JT, Dale JW, Eisenach KD, Gicquel B, Hermans P, Martin C, McAdam R, Shinnick TM, et al. Strain identification of Mycobacterium tuberculosis by DNA fingerprinting: recommendations for a standardized methodology. J Clin Microbiol 1993;31:406eC409.
World Health Organization, International Union against Tuberculosis and Lung Disease, Royal Netherlands Tuberculosis Association. Revised international definitions in tuberculosis control. Int J Tuberc Lung Dis 2001;5:213eC215.
Kleinbaum DG. Survival analysis: a self-learning text (statistics in the health sciences). New York: Springer; 1996.
Korenromp EL, Scano F, Williams BG, Dye C, Nunn P. Effects of human immundeficiency virus infection on recurrence of tuberculosis after rifampicin based treatment: an analytical review. Clin Infect Dis 2003;37:101eC112.
Perriens JH, St Louis ME, Mukadi YB, Brown C, Prignot J, Pouthier F, Portaels F, Willame JC, Mandala JK, Kaboto M, et al. Pulmonary tuberculosis in HIV-infected patients in Zaire: a controlled trial of treatment for either 6 or 12 months. N Engl J Med 1995;332:779eC784.
Warren RM, Victor TC, Streicher EM, Richardson M, Beyers N, Gey Van Pittius NC, Van Helden PD. Patients with active tuberculosis often have different strains in the same sputum specimen. Am J Respir Crit Care Med 2004;169:610eC614.
Yeh RW, Hopewell PC, Daley CL. Simultaneous infection with two strains of Mycobacterium tuberculosis identified by restriction fragment length polymorphism analysis. Int J Tuberc Lung Dis 1999;3:537eC539.
Comstock G. Tuberculosis in twins: a re-analysis of the Prophit survey. Am Rev Respir Dis 1978;117:621eC624.
Stead WW. Genetics and resistance to tuberculosis. Ann Intern Med 1992;116:937eC941.(Suzanne Verver, Robin M. )