Skip Navigation

High Rates of Recurrence in HIV-Infected and HIV-Uninfected Patients with Tuberculosis

  1. Judith R. Glynn1,
  2. Jill Murray3,4,
  3. Andre Bester5,
  4. Gill Nelson3,4,
  5. Stuart Shearer5 and
  6. Pam Sonnenberg2
  1. 1Department of Epidemiology and Population Health, London School of Hygiene and Tropical Medicine, London, United Kingdom
  2. 2Research Department of Infection and Population Health, University College London, London, United Kingdom
  3. 3National Institute for Occupational Health, National Health Laboratory Service, Johannesberg, South Africa
  4. 4School of Public Health, University of the Witwatersrand, Johannesberg, South Africa
  5. 5Gold Fields Limited, Johannesberg, South Africa
  1. Reprints or correspondence: Prof Judith Glynn, London School of Hygiene and Tropical Medicine, Keppel St, London WC1E 7HT, UK (judith.glynn{at}lshtm.ac.uk).
 
Next Section

Abstract

Background. The rate of recurrent tuberculosis disease due to reinfection, compared with the incidence of new tuberculosis, in those with and without HIV infection is not known.

Methods. In a retrospective cohort study of South African gold miners, men with known dates of seroconversion to HIV (from 1991 to 1997) and HIV-negative men were followed up to 2004. Rates of tuberculosis recurrence >2 years after the first episode were used as a proxy for reinfection disease rates.

Results. Among 342 HIV-positive and 321 HIV-negative men who had had ⩾1 previous episode of tuberculosis, rates of recurrence were 19.7 cases per 100 person-years at risk (PYAR; 95% confidence interval [CI], 16.4–23.7) and 7.7 cases per 100 PYAR (95% CI, 6.1–9.8), respectively. The recurrence rate did not vary by duration of HIV infection. Recurrent pulmonary tuberculosis rates >2 years after the first episode were 24.4 cases per 100 PYAR (95% CI, 17.2–34.8) in HIV-positive men and 4.3 cases per 100 PYAR (95% CI, 2.2–8.3) in HIV-negative men, compared with incidence rates of new pulmonary tuberculosis of 3.7 cases per 100 PYAR (95% CI, 3.3–4.1) in HIV-positive men and 0.75 cases per 100 PYAR (95% CI, 0.67–0.84) in HIV-negative men in the same cohort.

Conclusions. Tuberculosis recurrence rates, likely due to reinfection, were much higher than incidence rates. The findings suggest heterogeneity in susceptibility, implying that a vaccine could still provide useful protection in the population and strengthening the case for secondary preventive therapy.

Recurrent episodes of tuberculosis disease can occur as a result of relapse of the original infection, or following reinfection [1]. In a study in Cape Town, South Africa, the rate of reinfection disease (established by molecular typing) was much higher than the incidence rate of new tuberculosis [2]. This suggests that natural infection is failing to provide protection against reinfection, with worrying implications for vaccine development. However, in that study, although HIV prevalence was thought to be low (11% of new cases), the status was unknown for those with recurrences who might have higher human immunodeficiency virus (HIV) prevalence, so the extent to which this apparent lack of protection is attributable to HIV is unclear.

HIV infection greatly increases the incidence of new and recurrent tuberculosis [3]. For first episodes of disease, the extent of this increase depends on the duration of HIV infection [4] and the degree of immunosuppression [5]. Among HIV-positive patients, the rate of recurrence has been reported to be higher in those patients with lower CD4 counts in some studies [6–8] but not other studies [9, 10] and to be higher in those with symptomatic HIV before tuberculosis [9]. It is not known whether the rate of recurrence varies by duration of HIV infection.

HIV-induced immunosuppression might be expected to increase recurrence by both relapse and reinfection, but when measured directly, HIV infection appears to increase the rate of reinfection disease but not relapse [10, 11]. Relapse and reinfection disease can only be reliably distinguished if molecular typing of strains from the 2 episodes is available. In the absence of paired DNA fingerprints, the timing of the recurrence can give clues to the likely cause, because relapses generally occur soon after the end of treatment, whereas reinfection disease occurs at a more constant rate over time [10–12].

In a large cohort of South African gold miners with known dates of seroconversion to HIV, and a comparison group of HIV-negative miners, we have previously measured incidence of new pulmonary tuberculosis [13]. The overall rates were 0.75 cases per 100 person-years at risk (PYAR) (95% confidence interval [CI], 0.67–0.84) in the HIV-negative men and 3.7 cases per 100 PYAR (95% CI, 3.3–4.1) in the HIV-positive men. Rates increased with age, calendar period, and duration of HIV infection, reaching 1.2 cases per 100 PYAR (95% CI, 0.89–1.7) in the HIV-negative men in 2003–2004 and 10.0 cases per 100 PYAR (95% CI, 6.5–15.5) ⩾10 years after seroconversion to HIV [13]. We now examine recurrence rates in the same cohort during the same time period, by HIV status, and the duration of HIV infection.

Previous SectionNext Section

Methods

We conducted a retrospective cohort study in 4 mines in Gauteng Province, South Africa, using routinely collected data. HIV testing (enzyme-linked immunosorbent assay, with confirmation of positive results) was performed, with counseling and consent, in 2 surveys in the early 1990s, in sexually transmitted disease clinics, and as clinically indicated. Using tests performed from 1991 to 1997, and after excluding tests from medical and tuberculosis wards (as these would be biased toward inclusion of individuals with symptomatic disease), we identified a cohort of individuals with intervals of <3 years between their last negative and first positive HIV test result. They were compared with individuals who were HIV-negative in a survey in 1992–1993, with no later evidence of HIV infection [4, 14]. Approval for the study was received from the ethics committees of the University of Witwatersrand and the London School of Hygiene and Tropical Medicine.

A database is kept of all miners diagnosed with tuberculosis. Medical services are free and provided by the mining company. Until 1995, only pulmonary cases were recorded. All miners are screened by chest X-ray once a year and can self-present at any time. Tuberculosis is diagnosed on the basis of smear, culture, X-ray, and a clinical algorithm. We included pulmonary and extrapulmonary tuberculosis, but disease due to nontuberculous mycobacteria was excluded. Ninety percent of both HIV-positive and HIV-negative cases are bacteriologically confirmed (ie, smear and/or culture) [13].

Patients with tuberculosis from the 4 mines are treated at a single hospital. Both new and recurrent cases receive directly observed therapy, according to national guidelines. For new cases, this comprises 2 months of treatment with isoniazid (H), rifampin (R), pyrazinamide (Z), and ethambutol (E), followed by 4 months with H and R (2HRZE 4HR). The 8-month retreatment regimen adds streptomycin (S) in the initial phase (2HRZES, 1HRZE, and 5HRE). If drug resistance is found, the regimen is altered appropriately [15]. Multidrug resistance was defined as resistance to at least isoniazid and rifampin. Isoniazid prophylaxis was not used during the period of the study. Antiretroviral therapy was introduced in late 2004; December 2004 was taken as the end of the study period.

Episodes of tuberculosis were defined as follows. The start of the episode was taken as the earliest of the culture date, smear date, diagnosis date, admission date, or treatment date. The end of the episode was taken as the date that treatment was stopped. If this was not available, then the date of the third culture (taken toward the end of treatment) was used, or if this was missing, then a date 7 months after starting treatment, or if this was unknown, then a date 8 months after the start of the episode. A recurrence was defined as an episode that started after the end of the previous episode, unless treatment interruption was recorded, in which case it was assumed to be a continuation of the same episode. Outcomes of each episode (such as “cure”) were only recorded from 1995 onward.

Analysis of recurrence rates. We calculated rates of recurrence among individuals who had had an episode of tuberculosis during the study period, using survival analysis and Poisson regression. Individuals entered the analysis at the end of the previous episode and were censored when they developed a new episode of tuberculosis, left the mine, died, or reached the end of the study.

Analyses assessed risk factors for recurrence in HIV-positive and HIV-negative patients and the effect of HIV infection on recurrence. Calendar period, age, years after the first episode, and time from seroconversion to HIV were treated as time-varying covariates, using Lexis expansion. The effect of duration of HIV infection was assessed both as total time from seroconversion and as duration of HIV infection before the first episode. It was hypothesized that individuals who first developed tuberculosis soon after HIV seroconversion might be fast HIV progressors and likely to have a high rate of recurrence, and that individuals with longer time intervals since HIV seroconversion might be more immunosuppressed and therefore more likely to have recurrences. Analyses were stratified by whether the individual had had only 1 or more than 1 previous episode of tuberculosis.

Sensitivity analyses. Because HIV testing was not routine, seroconversions might have been missed, so analyses in HIV-negative miners were repeated, censoring individuals 1 year after their last negative result. However, because tuberculosis may have led to testing, tests done within 1 month of tuberculosis diagnosis were excluded.

Analyses were also repeated, restricted to those episodes due to culture-proven Mycobacterium tuberculosis. For these analyses, follow-up was still censored at recurrences with proven or nonproven tuberculosis, as individuals received treatment from this time.

Estimation of rates of recurrence due to reinfection. Because early recurrences include a high proportion of relapses, analyses were repeated excluding recurrences occurring in the first 2 years after the end of the first episode. This was further restricted to those with reported cure, and with only 1 previous episode of tuberculosis.

Previous SectionNext Section

Results

During the period of the study, 486 HIV-positive and 410 HIV-negative men in the cohort had an episode of tuberculosis. After excluding those whose episodes ended after the end of the study, who died, or who left the mine during treatment, there were 342 HIV-positive and 321 HIV-negative men from the cohort who had an index episode of tuberculosis (pulmonary or extrapulmonary episode) during the study period and were still under observation at the end of the episode, in whom rates of recurrence could be assessed. The end of the episode was defined as the date treatment was completed for the 62% of men included in the analysis.

Recurrence rates in HIV-positive men, compared with HIV-negative men. Overall, recurrences occurred in 115 (34%) of 342 HIV-positive men and 67 (21%) of 321 HIV-negative men. Rates of recurrence are shown in Tables 1 and 2 for HIV-negative and HIV-positive men, respectively.

Figure 1 shows the proportion of patients with recurrences by HIV status by time since the end of the index episode, overall (Figure 1A), and separately for those with 1 previous episode (Figure 1B) or more than 1 previous episode (Figure 1C). The overall rate was higher among those with HIV infection: 19.7 cases per 100 person years at risk (PYAR) versus 7.7 cases per 100 PYAR (hazard ratio [HR], 2.6; 95% CI, 1.9–3.4). This increased slightly after adjusting for current age, number of previous episodes, and duration of follow-up after the index episode (adjusted HR, 2.8; 95% CI, 2.1–3.9). However, there was an interaction with the number of previous episodes of tuberculosis (likelihood ratio test for interaction, P=.02), with HIV infection having much more effect in those with only 1 previous episode (HR, 3.2; 95% CI, 2.2–4.5) than in those with more than 1 previous episode (HR, 1.3; 95% CI, 0.64–2.5).

Figure 1.
View larger version:
Figure 1.

Kaplan Meier failure estimates for recurrences by human immunodeficiency virus (HIV) status and years since the end of the index episode. Recurrences (A) overall (showing numbers of individuals being followed up at each time point), (B) in those with 1 previous episode, and (C) in those with more than 1 previous episode. The lines are shown until there are <10 individuals still being followed up.

Among those with only 1 previous episode of tuberculosis, the effect of HIV hardly changed when adjusted for other factors. Among those with more than 1 previous episode, the estimate increased slightly after adjusting for duration of follow-up (adjusted HR, 1.6; 95% CI, 0.77–3.2) but was not changed by adjusting for other factors.

Restricting to culture-confirmed cases, the rates were slightly lower (Tables 1 and 2), but the results were otherwise similar. HIV infection increased the rate of recurrence more in those with only 1 previous episode (HR, 2.8; 95% CI, 1.8–4.3) than in those with more than 1 previous episode (HR, 1.5; 95% CI, 0.70–3.1; P=.1 test for interaction). Neither of these estimates changed after adjusting for possible confounders.

Figure 2.
View larger version:
Figure 2.

Recurrence in human immunodeficiency virus (HIV)-positive men after a first episode of tuberculosis by (A) total time since seroconversion to HIV and (B) duration of HIV infection before the first episode.

Table 1.
View larger version:
Table 1.

Recurrence Rates in 321 HIV-Negative Men and Hazard Ratios for Recurrence

Table 2.
View larger version:
Table 2.

Recurrence Rates in 342 Human Immunodeficiency Virus (HIV)-Positive Men and Hazard Ratios for Recurrence

Risk factors for recurrence in HIV-negative men. All available risk factors are shown in Table 1. Among the HIV-negative men, the only factors associated with recurrence in the univariate analysis were isoniazid resistance in the index episode (HR, 2.8; 95% CI, 1.6–5.1), and number of previous episodes of tuberculosis, with a higher rate in those with more than 1 previous episode than in those with only 1 episode (HR, 2.4; 1.5–4.0). There was an interaction between isoniazid resistance and number of previous episodes (likelihood ratio test, P=.02), with the increased risk associated with isoniazid resistance only being seen in those with only 1 previous episode (HR, 4.6; 95% CI, 1.9–11.2) and not in those with more than 1 previous episode (HR, 0.95; 95% CI, 0.40–2.3). None of the other factors shown in Table 1 were associated with recurrence in those with only 1 previous episode of tuberculosis, and no factors were significantly associated with recurrence in those with more than 1 previous episode of tuberculosis. The results restricted to those with culture-proven episodes were very similar (data not shown).

Among the HIV-negative men with recurrences, 43% had negative tests available at the time of the recurrence or later, 18 of 37 with recurrences within 2 years, and 11 of 30 with later recurrences. Restricting to the HIV-negative men censored at 1 year after the last negative test (excluding tests done around the time of tuberculosis), there were 19 recurrences in 253 PYAR (rate, 7.5 cases per 100 PYAR; 95% CI, 4.8–11.8). Estimates of rates by each of the available risk factors were similar to those in the whole data set (Table 1), but confidence intervals were wide because of the small numbers. However, in this restricted data set, recurrence was less common more than 2 years after the index episode than within the first 2 years (HR, 0.37; 95% CI, 0.11–1.3).

Risk factors for recurrence among HIV-positive men. Among the HIV-positive men, the number of previous episodes of tuberculosis and drug resistance in the index episode made little difference to recurrence rates (Table 2). The only factor associated with recurrence in the univariate analysis was time since the first episode, with higher rates after the first 2 years (HR, 1.7; 95% CI, 1.2–2.4). No other associations were apparent in the multivariate analysis, and the association with time since the first episode was not changed by adjusting for other factors.

Analyses were repeated separately for those with only 1 or more than 1 previous episode of tuberculosis. Among those with more than 1 episode, there were no associations with recurrence, but numbers were small (13 recurrences in total). Among those with 1 previous episode, the only factor associated with recurrence was the time since the first episode, with higher rates after the first 2 years (HR, 1.6; 95% CI, 1.1–2.4).

There was little evidence that either total time since seroconversion or duration of HIV infection before the index episode influenced the recurrence rate, either overall (Table 2; Figure 2) or when restricting to those with only 1 previous episode, or when additionally restricting to the period more than 2 years after the first episode. The HR for recurrence after the first episode of tuberculosis per year increase in total time since seroconversion was 1.05 (95% CI, 0.97–1.13), and this was similar when restricted to the period more than 2 years after the first episode (HR, 1.06; 95% CI, 0.91–1.22). Equivalent figures for the HR for recurrence per year increase in duration of HIV before the first episode of tuberculosis were 1.01 (95% CI, 0.93–1.09) for the whole period and 1.10 (95% CI, 0.97–1.24) for the period more than 2 years after the first episode.

Restricting to those with culture-confirmed episodes, the results were similar, although the association with duration of follow-up was reduced (HR, 1.3; 95% CI, 0.82–2.0). There were no associations between duration of HIV infection and recurrence.

Comparison of incidence rates and recurrence rates. Incidence rates, as estimated elsewhere [13], and recurrence rates from this study are shown in Table 3. Recurrence rates were much higher than incidence rates for both HIV-negative and HIV-positive individuals. After minimizing the likely contribution of relapse (by restricting to those with only 1 previous episode, excluding the first 2 years, and including only those with recorded cure) and restricting to pulmonary tuberculosis to allow comparison with the earlier data, rates of recurrence were still ∼3 times that of incident tuberculosis in both HIV-negative and HIV-positive individuals.

Table 3.
View larger version:
Table 3.

Rates of New Tuberculosis (as Estimated Elsewhere [13]) and Recurrent Tuberculosis (as Estimated in This Study), by Human Immunodeficiency Virus (HIV) Status

Previous SectionNext Section

Discussion

The rates of recurrent tuberculosis in this population were very high in both HIV-positive and HIV-negative men, at 19.7 and 7.7 cases per 100 PYAR, respectively. In a previous meta-analysis, among individuals receiving 5–6 months of rifampin-containing regimens, recurrence rates were estimated at 2.0 cases per 100 PYAR for HIV-negative individuals and 4.2 cases per 100 PYAR for HIV-positive individuals over a median of 34 months, with a tuberculosis incidence of 250 cases per 100,000 PYAR [3]. The authors estimated that there would be an additional 0.14 cases per 100 PYAR for each increase in tuberculosis incidence of 100 cases per 100,000 PYAR to allow for more reinfection disease. Incidence of tuberculosis was very high among the miners, rising from 800 to more than 2000 new cases per 100,000 PYAR over the period of the study [13]. Using an incidence of 2000 cases per 100,000 PYAR, recurrence rates of 4.5 cases per 100 PYAR in the HIV-negative men and 6.7 cases per 100 PYAR in the HIV-positive men might be expected. However, this implicitly assumes that the contribution of reinfection disease is the same for HIV-positive and HIV-negative individuals, which we know not to be the case [10, 11].

For the HIV-negative men, our rates were higher than the estimate, but not much higher when restricted to those for whom cure was recorded (6.1 cases per 100 PYAR). It is possible that some seroconversions in the HIV-negative cohort were missed, but nearly half of those with recurrences were known still to be negative at the time of recurrence, and censoring the analysis at 1 year after the last negative test gave similar results, although based on small numbers. Susceptibility due to silica exposure is likely to have contributed to the high rates [16].

For the HIV-positive men, our rates (19.7 cases per 100 PYAR) were much higher than the estimate. This is consistent with an important role played by reinfection disease, which may have been underestimated in the model. A fall in recurrence rates after the first 2 years would be expected, reflecting decreasing relapse rates with time [3]. This was not seen for either the HIV-negative or HIV-positive men; in fact, the rate increased slightly in the HIV-positive men. This suggests that reinfection disease predominates among the HIV-positive men, as we have also shown directly [10]. Among the HIV-negative men, the expected fall may have been masked by unknown seroconversions. When censored 1 year after the last negative test, recurrence rates in the HIV-negative men did decrease over time.

Because the high rate of recurrence in the HIV-positive men was probably largely due to reinfection disease, it would be expected that the rates should increase with calendar year (since the background incidence increased over time) and by duration of HIV infection (since this has been shown to be strongly associated with the risk of the first episode of tuberculosis). Neither effect was found.

It seems as though having had an episode of tuberculosis is such a sensitive marker of susceptibility to future infection and/ or disease that other factors become relatively unimportant. This is also apparent in the HIV-negative men. As shown in Table 3, rates of recurrence likely to be due to reinfection were >3 times higher than the incidence of new pulmonary tuberculosis during this period. Furthermore, HIV-negative men with more than 1 previous episode had recurrence rates approaching those of the HIV-positive men, and the timing of the recurrences (Figure 1C) did not suggest that they were all relapses.

Similarly, for the HIV-positive cohort, the incidence of late recurrences was much higher than the incidence of new pulmonary tuberculosis and was twice as high as the incidence of new pulmonary tuberculosis in those infected for >10 years.

This analysis was based on data on tuberculosis kept routinely by the mine hospital in a custom-built database. We used the decisions recorded in this database to define episodes of tuberculosis, including some that were not bacteriologically confirmed. However, the results were very similar when restricted to culture-proven episodes. Another potential limitation is that men were lost to follow-up when they left the mine. Sicker men may be more likely to leave, which would underestimate the recurrence rate.

The very high incidence of both new disease and of recurrence reflect the mine environment, but our results are consistent with those from a general population in Cape Town, in which rates of recurrence due to reinfection were estimated at 4 times the rate of new disease [2]. In our study, we could only infer the contribution of reinfection; in the Cape Town study, HIV status was unknown, but the results complement each other. Together, they suggest that whatever makes the individual susceptible to tuberculosis disease initially continues to act as a major risk factor for recurrence. This could be, for example, genetic factors, or underlying lung disease, including, in our study, silica exposure, alcohol consumption, or the immuno-suppression that led to the first episode. In this mine population it is unlikely to be variation in exposure to infection or socio-economic differences.

These findings have 2 important implications. Although recurrence due to reinfection implies that immunity owing to prior natural disease can fail, if it is heterogeneity in susceptibility of individuals that leads to higher rates of reinfection disease rather than new disease, then a (hypothetical) vaccine could still provide useful protection in much of the population. There may be a subset of the population that is particularly susceptible for whatever reason, but a vaccine could still be useful for the rest. The apparent failure of natural infection to provide any protection would be seen because the first episode had selected particularly susceptible individuals. And, if the first episode selects those at high risk, then this strengthens the case for secondary preventive therapy [6].

Previous SectionNext Section

Footnotes

  • Potential conflicts of interest: S Shearer and A Bester were employed by Gold Fields Ltd, but mine management had no input or influence on the research process or the findings. All other authors report no potential conflict.

  • Financial support: Colt Foundation, United Kingdom.

  • Received July 20, 2009.
  • Accepted October 6, 2009.
Previous Section
 

References

  1. 1.↵
    1. Lambert ML,
    2. Hasker E,
    3. Van Deun A,
    4. Roberfroid D,
    5. Boelaert M,
    6. Van der Stuyft P
    . Recurrence in tuberculosis: relapse or reinfection? Lancet Infect Dis 2003;3:282-287.
    CrossRefMedlineWeb of Science
  2. 2.↵
    1. Verver S,
    2. Warren RM,
    3. Beyers N,
    4. et al
    . Rate of reinfection tuberculosis after successful treatment is higher than rate of new tuberculosis. Am J Respir Crit Care Med 2005;171:1430-1435.
    Abstract/FREE Full Text
  3. 3.↵
    1. Korenromp EL,
    2. Scano F,
    3. Williams BG,
    4. Dye C,
    5. Nunn P
    . Effects of human immunodeficiency virus infection on recurrence of tuberculosis after rifampin-based treatment: an analytical review. Clin Infect Dis 2003;37:101-112.
    Abstract/FREE Full Text
  4. 4.↵
    1. Sonnenberg P,
    2. Glynn JR,
    3. Fielding K,
    4. Murray J,
    5. Godfrey-Faussett P,
    6. Shearer S
    . How soon after infection with HIV does the risk of tuberculosis start to increase? a retrospective cohort study in South African gold miners. J Infect Dis 2005;191:150-158.
    Abstract/FREE Full Text
  5. 5.↵
    1. van der Sande MA,
    2. Schim van der Loeff MF,
    3. Bennett RC,
    4. et al
    . Incidence of tuberculosis and survival after its diagnosis in patients infected with HIV-1 and HIV-2. AIDS 2004;18:1933-1941.
    CrossRefMedlineWeb of Science
  6. 6.↵
    1. Churchyard GJ,
    2. Fielding K,
    3. Charalambous S,
    4. et al
    . Efficacy of secondary isoniazid preventive therapy among HIV-infected Southern Africans: time to change policy? AIDS 2003;17:2063-2070.
    CrossRefMedlineWeb of Science
  7. 7.
    1. Vernon A,
    2. Burman W,
    3. Benator D,
    4. Khan A,
    5. Bozeman L
    . Acquired rifamycin monoresistance in patients with HIV-related tuberculosis treated with once-weekly rifapentine and isoniazid. Tuberculosis Trials Consortium. Lancet 1999;353:1843-1847.
    CrossRefMedlineWeb of Science
  8. 8.↵
    1. Charalambous S,
    2. Grant AD,
    3. Moloi V,
    4. et al
    . Contribution of reinfection to recurrent tuberculosis in South African gold miners. Int J Tuberc Lung Dis 2008;12:942-948.
    MedlineWeb of Science
  9. 9.↵
    1. Fitzgerald DW,
    2. Desvarieux M,
    3. Severe P,
    4. Joseph P,
    5. Johnson WD Jr.,
    6. Pape JW
    . Effect of post-treatment isoniazid on prevention of recurrent tuberculosis in HIV-1-infected individuals: a randomised trial. Lancet 2000;356:1470-1474.
    CrossRefMedlineWeb of Science
  10. 10.↵
    1. Sonnenberg P,
    2. Murray J,
    3. Glynn JR,
    4. Shearer S,
    5. Kambashi B,
    6. Godfrey-Faussett P
    . HIV-1 and recurrence relapse and reinfection of tuberculosis after cure: a cohort study in South African mineworkers. Lancet 2001;358:1687-1693.
    CrossRefMedlineWeb of Science
  11. 11.↵
    1. Crampin AC,
    2. Mwaungulu JN,
    3. Mwaungulu FD,
    4. et al
    . Recurrent tuberculosis: relapse or reinfection? The effect of HIV in a general population cohort in Malawi. AIDS 2010;24:417-426.
    CrossRefMedlineWeb of Science
  12. 12.↵
    1. Lambert ML,
    2. Hasker E,
    3. van Deun A,
    4. Roberfroid D,
    5. van der Stuyft P
    . Rifampicin and recurrence of tuberculosis among patients infected with HIV. Clin Infect Dis 2003;37:1719-1720.
    FREE Full Text
  13. 13.↵
    1. Glynn JR,
    2. Murray J,
    3. Bester A,
    4. Nelson G,
    5. Shearer S,
    6. Sonnenberg P
    . Effects of duration of HIV infection and secondary tuberculosis transmission on tuberculosis incidence in the South African gold mines. AIDS 2008;22:1859-1867.
    CrossRefMedlineWeb of Science
  14. 14.↵
    1. Sonnenberg P,
    2. Glynn JR,
    3. Fielding K,
    4. Murray J,
    5. Godfrey Fausett P,
    6. Shearer S
    . HIV and pulmonary tuberculosis: the impact goes beyond those infected with HIV. AIDS 2004;18:657-662.
    CrossRefMedlineWeb of Science
  15. 15.↵
    1. Sonnenberg P,
    2. Murray J,
    3. Shearer S,
    4. Glynn JR,
    5. Kambashi B,
    6. Godfrey-Faussett P
    . Tuberculosis treatment failure and drug resistance—same strain or reinfection? Trans R Soc Trop Med Hyg 2000;94:603-607.
    CrossRefMedlineWeb of Science
  16. 16.↵
    1. Cowie RL
    . Silicotuberculosis: long-term outcome after short-course chemotherapy. Tuber Lung Dis 1995;76:39-42.
    MedlineWeb of Science
| Table of Contents

This Article

  1. J Infect Dis. (2010) 201 (5): 704-711. doi: 10.1086/650529
  1. AbstractFree
  2. » Full Text (HTML)Free
  3. Full Text (PDF)Free

Classifications

Share

  1. Email this article

Navigate This Article

Current Issue

  1. March 1, 2012 205 (5)
  1. Journal of Infectious Diseases
  1. Alert me to new issues

Most