During the first half of the twentieth century, tuberculosis (TB) rates fell in industrialized countries as living conditions improved; in 1944, cure rates improved dramatically with the development of effective TB chemotherapy following the discovery of streptomycin . Subsequent drug discovery led to the current combination therapies. The HIV epidemic and the associated TB risk have, however, led to a resurgence of TB in countries with high HIV prevalence [2,3]. Immunocompromised persons are more likely to become infected with Mycobacterium tuberculosis, progress to active disease if already infected, relapse if cured, and die of TB if not adequately treated . Sub-Saharan Africa has the highest HIV infection rates, especially eastern and southern Africa [5,6]. In some countries of the former Soviet Union, TB notification rates increased rapidly because of the decline in public health services following the Soviet Union's end, but this trend has been reversed . Both TB and HIV still threaten marginalized people such as drug users, homeless, immigrants, and prisoners [8–11]. Effective TB control falls under the auspices of the Millennium Development Goals; in 2005, the World Health Organization (WHO) declared the TB epidemic in Africa a regional emergency. The UN General Assembly Special Session on AIDS (2006) and the G8 summit (2006) advocated for intensifying TB and HIV collaborative activities [12,13].
The first supragovernmental global program specifically targeted at TB was not officially implemented until the mid-1990s by WHO, under the name of Directly Observed Treatment, Short-course (DOTS). The cornerstone of DOTS is individual monitoring of chemotherapy administered to patients with active pulmonary TB. DOTS has achieved considerable success in many countries (Peru and China), but not in all, especially where HIV prevalence is high [2,14–16].
In the present study, we are concerned with TB control in countries with generalized HIV epidemics [16–19] and, in particular, with the development of shorter drug regimens [20–23]. The medications used today were developed over 40 years ago  and, although success rates can be high when taken as prescribed, treatment regimens for persons with active TB last for 6–8 months, can cause serious side effects and can have negative interactions with HIV drugs . Long treatment duration generates problems at both the clinical and public health levels because of the need to ensure high compliance levels for a long time to minimize treatment default, guarantee high cure rates, and avoid drug resistance [25,26]. The latter is a significant problem as shown by the recent outbreak of extensively drug-resistant TB in South Africa [27,28]. There is considerable interest in the development of shorter drug regimens, both for the individual's benefit and to free scarce resources for other uses [20,26,29–31].
In this analysis, we evaluated the potential health impact of shortening treatment duration, alone or in combination with other advances in TB detection and treatment, compared with present-day control programs. We developed a detailed mathematical and computational framework, elaborating the TB treatment model of  to include four HIV stages . We calibrated our model to TB and HIV data from Kenya (M.S. Sánchez, J. Lloyd-Smith, B.G. Williams, et al., in preparation), a sub-Saharan African country with a high HIV and TB prevalence and a strong disease surveillance program. We investigated the impact of various TB interventions as a function of HIV prevalence, which is important because HIV prevalence varies greatly geographically and among different populations . The benefits provided by each intervention were assessed using both absolute and relative measures, posed, respectively, in terms of total numbers of TB cases and deaths averted or proportional reductions. We considered the following scenarios reflecting enhancements over the status quo:
- reduced treatment duration (several drugs that could eliminate TB more quickly than current drugs are in advanced development stages),
- reduced treatment duration combined with increased case detection (previous studies support the importance of case detection for DOTS success),
- reduced treatment duration combined with increased cure rates (in order to evaluate the relative importance of increasing cure versus detection rates in relation to HIV prevalence), and
- combined implementation of all three enhancements, that is, shorter treatment duration, increased detection, and increased cure [18,20,26,30,33–35].
Our compartmental TB model includes 42 TB infection and treatment categories; flows among these categories reflect the most important processes determining TB incidence, prevalence, and mortality. We included five HIV clinically recognized categories (uninfected and WHO stages I–IV ) and calibrated our model using TB and HIV data from Kenya spanning 1980–2004 [5,7]. The most recent official HIV trends reported for Kenya  show prevalence peaking at 10% in 1997 and dropping to 7% in 2005. As discussed in another study (M.S. Sánchez, J. Lloyd-Smith, B.G. Williams, et al., in preparation), given the adequate fit of the model to Kenyan TB-HIV data up to 1997, the analyses we present here used the parameter set with the best fit to this time period. The supporting document provides a detailed explanation of the TB–HIV category structure, model formulation, and parameter values used.
HIV prevalence levels
To generalize our findings to other HIV-affected settings, we evaluated the benefits obtained by implementing the alternative TB control measures described below under five different HIV prevalence levels, that is, 3, 7, 15, 20, and 35%. We introduced the alternative scenarios in 2006 and projected the data forward for 25 years, until 2030 (Fig. S2). During this period, we assumed HIV prevalence remained approximately constant for the five different levels considered. For each HIV setting, HIV trends prior to 2006 were scaled to those reported in another study . As such, the HIV epidemic with a 35% prevalence in 2006 peaked at approximately 51% in 1997–1998. We investigated TB dynamics at such an elevated HIV prevalence because similar levels have been reported in certain populations .
In order to determine the most effective TB control policies in different HIV settings, we compared how HIV impacts TB trends under programs that emphasize alternative control measures in addition to reduced treatment duration . The scenarios considered are discussed below.
Baseline scenario: The first scenario followed present-day Kenyan conditions, with treatment under DOTS lasting 6 months and that under non-DOTS programs lasting 8 months. Default rates in the second and the third month were assumed to be, respectively, three and two times more than the default rates in the first month. Patients in the fourth to sixth month of DOTS and those in the fourth to eighth month for non-DOTS defaulted at the same rate as those in the first month. This pattern follows observed default rates [29,37,38].
- Reduced treatment duration. We reduced treatment duration to 2 months in both DOTS and non-DOTS regimens and for smear-positive and smear-negative TB cases. We assumed there was a corresponding reduction in the rate at which patients stopped treatment before completing their regimen, such that default rates in the second month were equal to those in the first month. In this scenario, the probability of cure at the end of 2 months was assumed to be equal to the probability of cure at the end of treatment (6 or 8 months) in the status quo.
- Reduced treatment duration and increased detection rate. We reran scenario 1 but doubled the case detection rate across all HIV categories.
- Reduced treatment duration and increased cure rate. We reran scenario 1 but now directed into the completely cured category 80% of patients who were previously susceptible to relapse (i.e., transiently recovered) upon completion of both DOTS and non-DOTS treatment across all five HIV categories.
- Reduced treatment duration, increased detection rate, and increased cure rate. We reran scenario 1 with both scenario 2 and 3 modifications (i.e., interventions in all three scenarios combined).
Benefits were measured as ‘new TB cases’ and ‘TB deaths’ avoided under each alternative scenario (1–4) as compared with the baseline. We considered both the absolute number of cases avoided:
Under this definition, a relative benefit of three corresponds to a three-fold decrease in new TB cases under the alternative as compared with the baseline scenario. The same reasoning was applied to obtain absolute and relative benefits in terms of TB deaths avoided.
In our simulations of present-day TB control conditions, TB incidence and mortality rates decrease throughout the period examined, assuming an HIV prevalence of 7% in the period 2006–2030 (Fig. 1). Shortening treatment duration from 6 to 2 months could additionally reduce TB incidence and deaths by approximately 15–20%, respectively, over the next 25 years. Reducing treatment duration and increasing detection or cure rates provide further benefits, and, if all three interventions were implemented together, it should be possible to reduce incidence and deaths by 80% during the projection period. Under our model formulation, treatment programs were shortened instantaneously from 6 to 2 months. This generated an initial adjustment period due to the momentarily greater influx of persons completing treatment and the corresponding short-term rise in the number of people eligible to relapse. We observed a temporary increase in TB incidence under those scenarios that did not include an increase in cure rate (i.e., scenarios 1 and 2; Fig. 1a). The brief increase displayed by TB mortality under all scenarios was very small in magnitude (Fig. 1b).
Figures 2–5 show that HIV prevalence had a strong influence on the efficacy of TB control programs, which varied between HIV-uninfected and HIV-infected persons. In these plots, HIV prevalence peaked in 1997 and approached a lower steady state in 2005 (Fig. S2), with trends scaled to those reported in another study . We observed qualitatively similar results in response to the different TB control strategies considered under different calibration schemes, that is, calibrating to previously reported HIV numbers in Kenya depicting a monotonically increasing HIV epidemic that reached 14% prevalence in 2003 , and using the same parameter values obtained in our calibration as that used in another study (M.S. Sánchez, J. Lloyd-Smith, B.G. Williams, et al., in preparation) (Table S2) but assuming HIV prevalence increased monotonically until a 7% HIV prevalence was reached. The numerical values plotted in Figs. 2–5 are presented in Table S4. We further discuss our observations on the benefits derived from increasing detection and/or cure without reducing treatment duration in the supporting document.
When measured in terms of total numbers of cases and deaths averted, the benefits obtained by all four alternative TB control options rose considerably with increasing HIV prevalence. The only exception occurred at the extreme HIV prevalence of 35% when only reducing treatment duration, in which the number of new cases slightly increased. This result was evident when considering the whole population (Fig. 2a), but was due to an increase in only the HIV-infected subpopulation (Fig. 2c). A reduction in treatment duration always decreased the number of TB cases in the HIV-uninfected subpopulation (Fig. 2b). The absolute number of deaths avoided increased with HIV prevalence for all control scenarios (Fig. 3).
Relative benefits, measured as the proportional reductions in numbers of TB cases and deaths, exhibited markedly different patterns from those of the absolute benefits. In the scenario in which only treatment time is reduced, relative reductions in incidence and mortality decreased slightly as HIV prevalence increased (Figs 4a and 5a). As for the absolute benefits, this effect culminated at the 35% HIV prevalence, with a slightly greater number of cases in the whole population (Fig. 4a) and in the people infected with HIV (Fig. 4c). For the people not infected with HIV, reducing treatment duration always provided a net increase in TB cases and deaths avoided (Figs 4b and 5b). The reduction in treatment duration combined with increased detection and/or cure provided a different pattern regarding the relative reductions in cases and deaths: benefits initially increased until an HIV prevalence of approximately 15%, after which they gradually decreased, both for the whole population and when partitioning the population into HIV uninfected and infected (Figs 4 and 5). On a relative scale, the improved TB control strategies provided greater benefits in terms of deaths avoided than in cases avoided.
Comparing benefits of the alternative scenarios according to HIV status
In terms of absolute cases and deaths avoided, benefits were proportionally greater among the people infected with HIV than among those not infected with HIV for the same HIV prevalence. This pattern was most noticeable at higher HIV prevalence levels (Figs 2 and 3). Again, the only exception arose for the 35% HIV prevalence, because the model predicted a greater number of HIV-infected persons would become infected with TB when only reduced treatment duration is implemented (Fig. 2c). In contrast, the relative benefits for all alternative TB control programs considered were consistently smaller for people infected with HIV than those not infected with HIV at the same HIV prevalence (Figs 4 and 5).
Comparing the benefits of increasing detection in comparison with cure
When comparing the benefits obtained from combining reduced treatment duration with either increased detection or cure, results varied depending on HIV prevalence and on whether we considered the people not infected with HIV together with, or separately from, those infected with HIV. The qualitative impact of HIV on TB control already became evident when investigating avoided cases under the different scenarios in the whole population Figs 2a and 4a: for an HIV prevalence less than 15%, increasing detection provided slightly greater benefits, but for higher HIV levels, increasing cure became progressively more effective. Detection always avoided a greater number of deaths when considering the whole population, but the relative gains were minimal at 35% HIV prevalence (Figs 3a and 5a).
Additional insight was gained when comparing benefits obtained from increased detection in comparison with increased cure according to HIV status. In the whole population, improved cure rate out-performed improved detection in terms of reducing TB incidence for HIV prevalences of 15% or higher. In the HIV-uninfected people, this shift was observed only at 35% HIV prevalence (Figs 2b and 4b), whereas among the HIV infected people, it held at all HIV levels (Figs 2c and 4c). A greater number of deaths were avoided by decreasing treatment duration combined with increased detection, as compared with increased cure, for all HIV levels in the HIV-uninfected people (Figs 3b and 5b) and HIV-infected people for HIV prevalences less than 35% (Figs 3c and 5c). At the extreme HIV prevalence of 35%, the deaths avoided in the HIV-infected people by the two TB control policies, although comparable, begin to favor an increased cure rate.
Our investigation into the impact of HIV on TB control shows that, as expected, the number of new TB cases and deaths increases substantially as HIV prevalence increases. More effective measures, therefore, have the potential to provide correspondingly greater absolute benefits at higher HIV levels [2,3,30,33]. The relative benefits of improved TB control, however, either become progressively smaller as HIV prevalence increases when only implementing reduced treatment duration, or increase with rising HIV prevalence up to 15%, then decrease at higher prevalence levels for the three remaining scenarios (reduced treatment time combined with increased detection, increased cure, or all concurrently). We do stress that, because of the great uncertainties regarding TB–HIV codynamics [4,14,39,40], we focused our analysis on understanding interactions and potential qualitative outcomes rather than obtaining precise numerical predictions. In particular, we consider cure rates close to those officially reported for Kenya (supporting document ); substantially lower population-level cure rates may influence the response of the TB epidemic to the interventions considered, highlighting the critical need for thorough data collection and evaluation on appropriate managerial scales.
TB control measures are less effective among the HIV-infected TB patients. Although in absolute terms benefits are proportionally greater among the HIV-infected patients for a given HIV prevalence (a pattern most noticeable at higher levels, with the exception of decreasing treatment duration at 35% HIV prevalence), relative benefits are always smaller. We therefore observed substantially different epidemiological responses to TB control in the HIV-uninfected and HIV-infected patients, even when assuming the two groups mixed randomly (and hence TB transmission was not biased by contact patterns).
Extremely high HIV prevalence appears to reverse the positive epidemiological impact of reducing TB treatment duration: the absolute number of new TB cases and deaths always decreases when changing from a 6-month to a 2-month regimen, except at 35% HIV prevalence when the number of new TB cases increases over the 25-year period considered. This pattern occurs in the HIV-infected patients but not in the HIV-uninfected patients. Moreover, because of the overwhelming proportion of HIV-infected patients, it translates into an increased number of new cases when considering the whole population. This outcome is due to immunocompromised persons being very susceptible to TB infection, reinfection, progression, and relapse following treatment completion. In effect, extended TB treatment can act as a prophylactic and transmission blocker for this population, such that in our model there is an epidemiological benefit to longer treatment duration when HIV prevalence reaches these levels. In this regard, TB relapse rates in TB–HIV-coinfected patients may be reduced when TB treatment is extended beyond the routinely recommended treatment duration with current drugs . In any case, we need to bear in mind that our evaluation does not account for the potentially toxic effects of extended chemotherapy, negative interactions with HIV treatments, or the additional burden on public health systems of longer TB treatment regimes .
Considering benefits both in absolute and relative terms provides complementary information regarding the potential impact of different TB interventions. Considering first the benefits among HIV-uninfected patients, we find that absolute benefits rise with increasing HIV prevalence, because of the greater TB burden present in the population at higher HIV levels; in contrast, relative benefits rise as HIV prevalence increases to 15% but then decline at higher HIV prevalence levels because of diminishing per capita efficiency of TB control measures. Comparisons of absolute and relative benefits in the HIV-infected patients yield further insights: the absolute number of cases avoided when increasing cure versus detection is substantially greater at higher HIV levels among the HIV-infected patients; conversely, increasing cure provides similar relative benefits in terms of cases avoided in comparison to increasing detection at all HIV levels. Therefore, although knowledge of absolute measures is essential, relative numbers can provide further critical insights into relevance to cost-effective analyses , particularly when differences in absolute terms are small.
The magnitude of the impact of HIV on TB control will vary across geographic areas and social groups, but will be greatest in settings of high HIV prevalence, including countries (sub-Saharan Africa ), regions or communities within countries (Carletonville South African gold mining community in South Africa [42,43]), and high-risk groups (the homeless, incarcerated, and drug users in industrialized nations [8,10,11], military personnel ). Our study further indicates that if HIV treatment is effective in making the treated HIV-infected patients behave epidemiologically as if they were not infected with HIV, then such treatment can provide valuable secondary benefits in TB control both on an absolute and on a per capita basis. However, if the mortality of treated HIV-infected persons is reduced by antiretrovirals (ARVs) but their susceptibility to disease remains higher than that of the HIV-uninfected persons, the overall TB burden in the population will likely increase [4,6,45–48]. Further detailed investigations are required to fully evaluate the public health impact of shortening TB treatment duration, in particular regarding the rollout of antiretroviral drugs, and the generation and spread of multidrug-resistant and extensively drug-resistant tuberculosis [6,49–52]. Moreover, results also indicate greater benefits may be obtained by interventions targeted at communities according to their TB and HIV risk [11,53–56]. For instance, we observed that the benefits of increasing detection versus cure varied according to HIV levels, depending on how HIV interacts with the intervention. As the proportion of HIV-infected persons increased, increasing cure gradually became more beneficial compared with increasing detection; this pattern was most apparent in the HIV-infected fraction. A targeted approach may be particularly fruitful for those groups with extremely high HIV prevalence compared with the general population, as evidenced by the exceptional results obtained at the 35% HIV prevalence level.
Our study shows how reducing treatment duration, alone or in combination with other TB control strategies, can provide enormous benefits at high HIV prevalence, although both absolute and relative benefits vary substantially according to the proportion of the population that is HIV infected. Challenges arise not only because in absolute terms the total amount of infected people and deaths increases dramatically with increasing HIV prevalence [2,3,30,33] but also because the relative efficacy of TB control measures display a nonlinear pattern whereby they become less effective on a per capita basis at HIV prevalences higher than 15%. Therefore, a qualitative shift in TB dynamics occurs as HIV levels increase – the population's response to TB control becomes more in consonance with the response of the HIV-infected fraction ( and references therein) up to the point where the benefits of reducing treatment duration may even be reversed at extreme HIV levels. Moreover, the benefits of increasing detection versus cure decrease as HIV prevalence increases. Our results indicate that TB control policies need to consider HIV levels because the efficacy of different interventions varies substantially with HIV prevalence [58,59].
This research was supported by NIH-NIDA (M.S.S. and W.M.G.), the Global Alliance for TB Drug Development (W.M.G., J.A.S.), the James S. McDonnell Foundation 21st Century Science Initiative (W.M.G.), SACEMA, the South African Center for Epidemiological Modeling and Analysis (W.M.G.), and the Center for Infectious Disease Dynamics, Pennsylvania State University (J.L.S.). We thank three anonymous reviewers for their insightful suggestions, together with David Bangsberg, Jason Barbour, Frank Cobelens, Judy Hahn, Philip Hopewell, Gwynne Oosterbaan, Perry de Valpine, Suzanne Verver, and members of Wayne Getz's and Phil Hopewell's laboratories for their help and orientation. We would also like to thank the Kenyan public health authorities and the persons participating in the studies for their invaluable data collection.
1. Fujiwara PI, Munsiff SS, Simone PM. Treatment of tuberculosis
. In: Reichman LB, Hershfield ES, editors. Tuberculosis. A comprehensive international approach
. New York: Marcel Dekker; 2000. pp. 401–446.
2. Nunn P, Williams B, Floyd K, Dye C, Elzinga G, Raviglione M. Tuberculosis
control in the era of HIV. Nat Rev Immunol 2005; 5:819–826.
3. Corbett EL, Watt CJ, Walker N, Maher D, Williams BG, Raviglione MC, Dye C. The growing burden of tuberculosis
: global trends and interactions with the HIV epidemic. Arch Intern Med 2003; 163:1009–1021.
4. Williams BG, Dye C. Antiretroviral drugs for tuberculosis
control in the era of HIV/AIDS. Science 2003; 301:1535–1537.
5. UNAIDS. UNAIDS annual report; 2006 report on the global AIDS epidemic: a UNAIDS 10th anniversary special edition
. Geneva: WHO; 2006.
6. Corbett EL, Marston B, Churchyard GJ, De Cock KM. Tuberculosis
in sub-Saharan Africa: opportunities, challenges, and change in the era of antiretroviral treatment. Lancet 2006; 367:926–937.
7. World Health Organization. Global tuberculosis control: surveillance, planning, financing
. WHO Report 2006. Geneva: WHO; 2006.
8. Fernández de la Hoz K, Inigo J, Fernández-Martín JI, Arce A, Alonso Sanz M, Gomez-Pintado P, et al
. The influence of HIV infection and imprisonment on dissemination of Mycobacterium tuberculosis
in a large Spanish city. Int J Tuberc Lung Dis 2001; 5:696–702.
9. Schneider E, Laserson KF, Wells CD, Moore M. Tuberculosis
along the United States–Mexico border, 1993–2001. Rev Panam Salud Publica 2004; 16:23–34.
10. Moss AR, Hahn JA, Tulsky JP, Daley CL, Small PM, Hopewell PC. Tuberculosis
in the homeless: a prospective study. Am J Respir Crit Care Med 2000; 162:460–464.
11. Daley CL, Hahn JA, Moss AR, Hopewell PC, Schecter GF. Incidence of tuberculosis
in injection drug users in San Francisco: impact of anergy. Am J Respir Crit Care Med 1998; 157:19–22.
12. Raviglione MC. The global plan to stop TB, 2006–2015. Int J Tuberc Lung Dis 2006; 10:238–239.
13. Reid A, Scano F, Getahun H, Williams B, Dye C, Nunn P, et al
. Towards universal access to HIV prevention, treatment, care, and support: the role of tuberculosis
/HIV collaboration. Lancet Infect Dis 2006; 6:483–495.
14. Dye C, Watt CJ, Bleed DM, Hosseini SM, Raviglione MC. Evolution of tuberculosis
control and prospects for reducing tuberculosis
incidence, prevalence, and deaths globally. JAMA 2005; 293:2767–2775.
15. Dye C, Maher D, Weil D, Espinal M, Raviglione M. Targets for global tuberculosis
control. Int J Tuberc Lung Dis 2006; 10:460–462.
16. Maher D, Harries A, Getahun H. Tuberculosis
and HIV interaction in sub-Saharan Africa: impact on patients and programmes; implications for policies. Trop Med Int Health 2005; 10:734–742.
17. Bates I, Fenton C, Gruber J, Lalloo D, Lara AM, Squire SB, et al
. Vulnerability to malaria, tuberculosis
, and HIV/AIDS infection and disease. Part II: Determinants operating at environmental and institutional level. Lancet Infect Dis 2004; 4:368–375.
18. Onyebujoh P, Rodriguez W, Mwaba P. Priorities in tuberculosis
research. Lancet 2006; 367:940–942.
19. Dowdy DW, Chaisson RE, Moulton LH, Dorman SE. The potential impact of enhanced diagnostic techniques for tuberculosis
driven by HIV: a mathematical model. AIDS 2006; 20:751–762.
20. Andries K, Verhasselt P, Guillemont J, Göhlmann HW, Neefs JM, Winkler H, et al
. A diarylquinoline drug active on the ATP synthase of Mycobacterium tuberculosis
. Science 2005; 307:223–227.
21. Connolly LE, Edelstein PH, Ramakrishnan L. Why is long-term therapy required to cure tuberculosis
? PLoS Med 2007; 4:e120.
22. Williams KJ, Duncan K. Current strategies for identifying and validating targets for new treatment-shortening drugs for TB. Curr Mol Med 2007; 7:297–307.
23. Spigelman MK. New tuberculosis
therapeutics: a growing pipeline. J Infect Dis 2007; 196(Suppl 1):S28–S34.
24. de Jong BC, Israelski DM, Corbett EL, Small PM. Clinical management of tuberculosis
in the context of HIV infection. Annu Rev Med 2004; 55:283–301.
25. Shargie EB, Lindtjorn B. Determinants of treatment adherence among smear-positive pulmonary tuberculosis
patients in southern Ethiopia. PLoS Med 2007; 4:e37.
26. Spigelman M, Gillespie S. Tuberculosis
drug development pipeline: progress and hope. Lancet 2006; 367:945–947.
27. Cohen J. Infectious disease. Extensively drug-resistant TB gets foothold in South Africa. Science 2006; 313:1554.
28. Lawn SD, Wilkinson R. Extensively drug resistant tuberculosis
. BMJ 2006; 333:559–560.
29. Combs DL, O'Brien RJ, Geiter LJ. USPHS Tuberculosis
Short-Course Chemotherapy Trial 21: effectiveness, toxicity, and acceptability. The report of final results. Ann Intern Med 1990; 112:397–406.
30. Salomon JA, Lloyd-Smith JO, Getz WM, Resch S, Sánchez MS, Porco TC, Borqdoff MW, et al
. Prospects for advancing tuberculosis
control efforts through novel therapies. PLoS Med 2006; 3:e273.
31. Ginsberg AM, Spigelman M. Challenges in tuberculosis
drug research and development. Nat Med 2007; 13:290–294.
32. Morgan D, Mahe C, Mayanja B, Okongo JM, Lubega R, Whitworth JAG. HIV-1 infection in rural Africa: is there a difference in median time to AIDS and survival compared with that in industrialized countries? AIDS 2002; 16:597–603.
33. Currie CS, Williams BG, Cheng RC, Dye C. Tuberculosis
epidemics driven by HIV: is prevention better than cure? AIDS 2003; 17:2501–2508.
34. Dye C, Garnett GP, Sleeman K, Williams BG. Prospects for worldwide tuberculosis
control under the WHO DOTS strategy. Directly observed short-course therapy. Lancet 1998; 352:1886–1891.
35. Currie CSM, Floyd K, Williams BG, Dye C. Cost, affordability and cost-effectiveness of strategies to control tuberculosis
in countries with high HIV prevalence. BMC Public Health 2005; 5:130.
36. Cheluget B, Baltazar G, Orege P, Ibrahim M, Marum LH, Stover J. Evidence for population level declines in adult HIV prevalence in Kenya. Sex Transm Infect 2006; 82:I21–I26.
37. Chee CB, Boudville IC, Chan SP, Zee YK, Wang YT. Patient and disease characteristics, and outcome of treatment defaulters from the Singapore TB control unit: a one-year retrospective survey. Int J Tuberc Lung Dis 2000; 4:496–503.
38. Chan-Yeung M, Noertjojo K, Leung CC, Chan SL, Tam CM. Prevalence and predictors of default from tuberculosis
treatment in Hong Kong. Hong Kong Med J 2003; 9:263–268.
39. Day JH, Grant AD, Fielding KL, Morris L, Moloi V, Charalambous S, et al
. Does tuberculosis
increase HIV load? J Infect Dis 2004; 190:1677–1684.
40. Cruciani M, Malena M, Bosco O, Gatti G, Serpelloni G. The impact of human immunodeficiency virus type 1 on infectiousness of tuberculosis
: a meta-analysis. Clin Infect Dis 2001; 33:1922–1930.
41. Nahid P, Gonzalez LC, Rudoy I, de Jong BC, Unger A, Kawamura LM, et al
. Treatment outcomes of patients with HIV and tuberculosis
. Am J Respir Crit Care Med 2007; 175:1199–1206.
42. Sonnenberg P, Glynn JR, Fielding K, Murray J, Godfrey-Faussett P, 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.
43. Corbett EL, Charalambous S, Moloi VM, Fielding K, Grant AD, Dye C, et al
. Human immunodeficiency virus and the prevalence of undiagnosed tuberculosis
in African gold miners. Am J Respir Crit Care Med 2004; 170:673–679.
44. Feldbaum H, Lee K, Patel P. The national security implications of HIV/AIDS. PLoS Med 2006; 3:774–778.
45. Girardi E, Sabin CA, d'Arminio Monforte A, Hoqq B, Phillips AM, Gill MJ, et al
. Incidence of tuberculosis
among HIV-infected patients receiving highly active antiretroviral therapy in Europe and North America. Clin Infect Dis 2005; 41:1772–1782.
46. Godfrey-Faussett P, Ayles H. Can we control tuberculosis
in high HIV prevalence settings? Tuberculosis
47. Lawn SD, Myer L, Bekker LG, Wood R. Burden of tuberculosis
in an antiretroviral treatment programme in sub-Saharan Africa: impact on treatment outcomes and implications for tuberculosis
control. AIDS 2006; 20:1605–1612.
48. Wood R. The case for integrating tuberculosis
and HIV treatment services in South Africa. J Infect Dis 2007; 196:S497–S499.
49. Blower SM, Chou T. Modeling the emergence of the ‘hot zones’: tuberculosis
and the amplification dynamics of drug resistance. Nat Med 2004; 10:1111–1116.
50. Dye C, Williams BG. Criteria for the control of drug-resistant tuberculosis
. Proc Natl Acad Sci U S A 2000; 97:8180–8185.
51. Williams BG, Granich R, Chauhan LS, Dharmshaktu NS, Dye C. The impact of HIV/AIDS on the control of tuberculosis
in India. Proc Natl Acad Sci U S A 2005; 102:9619–9624.
52. Williams BG, Korenromp EL, Gouws E, Schmid GP, Auvert B, Dye C. HIV infection, antiretroviral therapy, and CD4+ cell count distributions in African populations. J Infect Dis 2006; 194:1450–1458.
53. De Cock KM, Chaisson RE. Will DOTS do it? A reappraisal of tuberculosis
control in countries with high rates of HIV infection. Int J Tuberc Lung Dis 1999; 3:457–465.
54. Lawn SD, Bekker LG, Middelkoop K, Myer L, Wood R. Impact of HIV infection on the epidemiology of tuberculosis
in a peri-urban community in South Africa: the need for age-specific interventions. Clin Infect Dis 2006; 42:1040–1047.
55. Borgdorff MW, Nagelkerke NJD, de Haas PEW, van Soolingen D. Transmission of Mycobacterium tuberculosis
depending on the age and sex of source cases. Am J Epidemiol 2001; 154:934–943.
56. Gouws E, White PJ, Stover J, Brown T. Short term estimates of adult HIV incidence by mode of transmission: Kenya and Thailand as examples. Sex Transm Infect 2006; 82(Suppl 3):iii51–iii55.
57. Lawn SD, Wood R. The epidemic of HIV-associated tuberculosis
in sub-Saharan Africa: does this also impact non-HIV-infected individuals? AIDS 2006; 20:1787–1788.
58. Mandavilli A. News feature: a clash of cultures. Nat Med 2007; 13:268–270.
59. Nunn P, Reid A, De Cock KM. Tuberculosis
and HIV infection: the global setting. J Infect Dis 2007; 196(Suppl 1):S5–S14.