Tuberculosis (TB) remains one of the most frequent opportunistic infections in people living with HIV. In 2007, an estimated 1.37 million HIV-associated TB cases occurred globally.1 Antiretroviral treatment (ART) has revolutionized the care of HIV-infected individuals with dramatic reductions in morbidity and mortality and a 80% reduction in TB incidence rates among people receiving ART.2-8
Studies across regions have consistently documented high TB incidence in the first months of ART.5,9-14 The high incidence during early phase of ART is, at least in part, the result of unmasking of subclinical TB during the initial rapid restoration of the immune response.15,16 In most cohorts, the incidence of TB declined thereafter but remained higher than the background TB incidence rate in the general population, even after years of receiving ART.5,9,10,12,16 This elevated long-term TB incidence has been shown to be strongly associated with the time spent at CD4 count less than 500 cells/mm3,16.
Understanding the incidence and risk factors of active TB over time in patients on ART is crucial for the design of effective preventive strategies. We analyzed 4 years of data from a large cohort of patients initiating ART in Johannesburg, South Africa, to determine TB incidence by time since ART initiation and assess risk factors separately for early incident TB (first 6 months of ART) and late incident TB (greater than 6 months of ART).
Study Site and Inclusion Criteria
The Themba Lethu Clinical Cohort is an open, prospective clinical cohort of adults initiating ART at a single-site, hospital-based public clinic providing free care in Johannesburg, South Africa. Patients were included in this analysis if they initiated ART between April 1, 2004, and March 31, 2007.
Patient Follow-Up, HIV, and Tuberculosis Care and Treatment
Before ART initiation, patients attended educational and adherence sessions, were assessed by a physician, and screened for the presence of opportunistic infections. After ART initiation, patients were scheduled for monthly pharmacy visits, clinical visits at Month 4 and every 6 months thereafter, and additional visits whenever needed. Laboratory assessments included hemoglobin, CD4 count, and plasma HIV viral load (VL) at baseline and each clinical follow-up visit, except for VL, which was not performed at baseline. Screening for TB was performed routinely at all clinic visits using a questionnaire for cough or fever for greater than 2 weeks, night sweats, weight loss of greater than 10%, or lymphadenopathy.
As per South African guidelines, the majority of patients received a first-line ART regimen of stavudine, lamivudine, and either efavirenz or nevirapine. Second-line ART consisted of azathioprine, didanosine, and ritonavir-boosted lopinavir. TB treatment was provided by the primary healthcare center closest to the patient's home and followed TB control program guidelines.
Patients more than 3 months late for a scheduled visit were traced by phone or home visit and categorized as lost to follow-up if alive and not in care. Vital status was also ascertained by consulting the national death registry.
Definitions of Tuberculosis
TB diagnosis was based on South African TB guidelines in accordance with internationally accepted criteria. Pulmonary TB was diagnosed using smear microscopy or algorithms for smear-negative TB that include chest x-ray, culture, and decision of a physician to prescribe a full course of anti-TB treatment. A case of extrapulmonary TB was defined as a patient in whom a clinician decided to prescribe a full course of anti-TB treatment based on clinical, radiologic, microbiologic, or histopathologic information. Routine culture for Mycobacterium tuberculosis is not recommended by the TB control program.
Incident TB was defined as the first episode of TB after initiation of ART and categorized as early (within 6 months of ART) or late (more than 6 months after ART initiation). Individuals on TB treatment at the time of ART initiation were classified as prevalent TB cases and became at risk of incident TB after completion of treatment for prevalent TB. History of TB was defined as treatment for TB before ART initiation and prevalent TB at the time of ART initiation.
Baseline demographics were characterized using standard descriptive statistics. Anemia was defined as a hemoglobin value below 13.0 g/dL (men), 12.0 g/dL (women), or 11.0 g/dL (pregnant women) down-adjusted by 0.65 g/dL because of altitude.17 Chi square tests were used to compare categorical variables and Wilcoxon rank-sum tests for continuous variables by category.
TB incidence rate was defined as the number of TB cases occurring per 100 person-years observation at risk after ART initiation. Person-time accrued during treatment for prevalent TB was excluded from the denominator when calculating TB incidence rates. Incidence rates were estimated using Poisson regression models with each follow-up interval considered separately. The analysis was further stratified by baseline CD4 cell count. Kaplan-Meier method was used to estimate the probability of active TB while on ART. Censoring of patients occurred at death from any cause, loss to follow-up, or administratively on March 31, 2008.
We estimated hazard ratios and 95% confidence intervals (CIs) for risk factors for incident TB (0-3 months, 4-6 months, and greater than 6 months) using multivariate Cox proportional hazards regression. Analysis of risk factors for early incident TB included baseline body mass index (BMI), CD4 count, and anemia. Risk factors for late incident TB included time-updated BMI, CD4 count, anemia, and VL.
All data were entered in a clinic data management system, TherapyEdge-HIV (Associated Biological Systems, Johannesburg, South Africa). Analysis was performed using SAS Version 9.1.3 (SAS, Cary, NC).
The study was approved by the Institutional Review Boards of the University of the Witwatersrand, South Africa, and the University of North Carolina at Chapel Hill, Chapel Hill, NC.
Patient Population, Baseline Characteristics, and Follow-Up
A total of 7536 people initiated ART between April 1, 2004, and March 31, 2008, contributing 161,000 person-months of observation with a median follow-up of 21.4 months. During follow-up, 496 patients (6.6%) died and 1036 patients (13.7%) were lost to follow-up. A total of 1556 patients (20.6%) were on TB treatment at the time of ART initiation (prevalent TB). Five hundred one patients developed TB on ART (incident TB), of which 284 cases occurred in the first 6 months of ART (early incident TB) and 217 after the first 6 months of ART (late incident TB). Of the 501 patients with incident TB, 55 also had a history of prevalent TB with a median time of 8 months (interquartile range, 3-16) between the end of prevalent TB episode and diagnosis of incident TB.
Patients with prevalent TB had the most advanced disease at time of ART initiation: 42% had CD4 less than 50 cells/mm3, 32% BMI less than 18.5 kg/m2, and 72% had anemia. Those with early incident TB also had advanced HIV disease, whereas those who developed late incident TB had baseline characteristics similar to those who never experienced TB (Table 1).
Tuberculosis Incidence Rates
The cumulative probability of developing TB on ART was 5.7% (95% CI, 5.2-6.3) in the first year of ART, 7.6% (95% CI, 6.9-8.3) by the end of the second year, 9% (95% CI, 8.2-9.8) by the end of the third year, and 10.1.% (95% CI, 8.9-11.5) by the end of the fourth year of ART.
The overall incidence rate of TB on ART was 4.2 (3.8-4.5) per 100 person-years (py). The TB incidence rate was the highest in the first 90 days of ART (13.9 per 100 py; 95% CI, 11.9-15.8), dropped by 50% in the third to sixth months of ART (6.3 per 100 py; 95% CI, 5.1-7.9), and further decreased with increasing time on ART to 1.6 per 100 py (95% CI, 1.2-2.3) in the third year of ART, after which the incidence remained stable (Fig. 1; Table 2).
The rate of early incident TB was strongly determined by baseline CD4 count (Fig. 1). TB incidence rates in the first 90 days was 21.7 per 100 py (95% CI, 17.5-26.7) among those with baseline CD4 count below 50 cells/mm3 compared with 12.8 (95% CI, 9.1-18.0), 9.5 (95% CI, 6.4-16.2), and 10.2 (95% CI, 6.4-16.2) per 100 py for those with baseline CD4 count 50 to 100, 101 to 200, and 200 or more, respectively (test for linear trend P < 0.0001).
Risk Factors for Early Incident Tuberculosis
Severity of immune suppression at time of ART initiation was associated with the risk of developing active TB in the first 6 months of ART (Table 3). Compared with those presenting with a CD4 count greater than 200 cells/mm3, patients with low CD4 count (less than 50 cells/mm3) were almost twice as likely to develop TB (adjusted hazard ratio [aHR], 1.68; 95% CI, 1.03-2.74), those with CD4 count between 50 and 100 tended to have a higher hazard (aHR, 1.31; 95% CI, 0.78-2.20), whereas those with CD4 count between 100 and 200 were not at increased risk (aHR, 1.08; 95% CI, 0.66-1.77). Poor nutritional status at the time of presentation for ART was also associated with risk of early incident TB with increasing TB rates as BMI decreased: aHR 1.69 (95% CI, 1.18-2.42) and 2.03 (95% CI, 1.34-3.07) for those with BMI 18.5 to 24.9 kg/m2 and less than 18.5 kg/m2, respectively, when compared to BMI greater than 25 kg/m2. The presence of anemia was also independently associated with early incident TB (aHR, 2.54; 95% CI, 1.92-3.37). A history of TB treatment and being employed at the time of ART initiation reduced the risk of early incident TB. Results were similar but less precise when evaluated separately for 0- to 3-month and 4- to 6-month categories (data not shown).
Risk Factors for Late Incident Tuberculosis
Among individuals who did not experience TB in the first 6 months of ART, baseline characteristics were not associated with risk of developing late incident TB, except for a history of TB (Table 3). Factors associated with poor response to ART were strongly associated with an increased risk of late incident TB. Current undernutrition (BMI less than 18.5 kg/m2) and below-normal BMI (18.5-24.9 kg/m2) increased the hazard of TB (aHR, 4.07; 95% CI, 2.53-6.57 and aHR, 1.71; 95% CI, 1.20-2.44, respectively). The hazard of late incident TB was also associated with current CD4 counts less than 350, current anemia, and VL greater than 10,000 copies/mL.
The large sample size of more than 7500 individuals residing in a high TB burden area allowed us to estimate TB incidence rates stratified by CD4 count and to quantify risk factors for TB separately for the early and later phases of ART. Among people attending a South African inner-city ARV clinic, we observed a 10% 4-year risk of developing TB on ART compared with the 10% lifetime risk in HIV uninfected individuals.18 This burden is in addition to a 20% prevalence of TB at the time of presentation for ART. These results highlight that TB will remain a major challenge to the health of people living with HIV in sub-Saharan Africa, even in the era of increasing access to ART.
Similar to a study by Lawn et al in Cape Town, South Africa,16 we observed extremely high incidence rates in the first months of ART with a point estimate of 21.7 (95% CI, 17.5-26.7) cases per 100 py in the first 3 months of ART among those with baseline CD4 counts below 50 cells/mm3. In contrast to the study in Cape Town, we found that high rates of early incident TB were not restricted to those with baseline CD4 cell counts less than 200 cells/mm3. Instead, we observed significantly increased rates of TB in the first 3 months of ART across all CD4 strata. Person characteristics other than CD4 count independently associated with early incident TB were anemia, nutritional status, and lack of employment. A log-linear association between risk of TB and BMI in the range 18.5 and 30 kg/m2 was recently demonstrated in HIV-uninfected individuals.19 Our study confirms this associated in patients receiving ART and extends the association to BMI below 18.5 kg/m2. Similar to our findings, anemia at the time of ART initiation was associated with a threefold increase in TB risk in Abidjan, but the association did not reach statistical significance in that study.20 The World Health Organization recently recommended replacing stavudine by zidovudine or tenofovir in first-line ART regimens.21 Because anemia is a frequent side effect of zidovudine, it could be important to monitor the effect of this increased use of zidovudine on the risk of incident TB.
TB incidence rates dropped dramatically during the first year of highly active ART and stabilized thereafter. This trend may be related to the rapid positive response to ART in our population with 90% of people suppressing virus replication by 6 months of ART and dramatic restoration of immune function in the first year of ART (increase of mean CD4 count from 111 to 289 cells/mm3) followed by smaller gains in CD4 cell numbers thereafter.17 The observed TB rate of 1.6 to 2.2 per 100 py in Years 2 to 4 of ART is lower than the 3.2 to 4.5 per 100 py incidence rate in Years 2 to 4 of ART in Cape Town, South Africa16 but significantly higher than the 0.18 per 100 py TB incidence rate documented after the first year of ART in Burkina Faso.13 The risk stabilized at a rate approximately four times that of the estimated background TB rate in the South African studies (1.03 and 0.5 per 100 py for Western Cape and Gauteng province, respectively; Source: C. Idema, Department of Health). In contrast, the long-term TB incidence rates on ART in Burkina Faso dropped to that of the estimated background rate in the general population. These finding suggest that, although the relative impact of ART is consistent across settings (70%-90% reduction),2-8 the level at which the TB incidence rate stabilizes may be more variable and related to the background force of infection in the community.
Among individuals who did not experience early incident TB, the risk of late incident TB was associated with response to treatment. Updated CD4 count was strongly associated with TB risk with a fourfold increased hazard among people with CD4 count below 50 cells/mm3 compared with those with CD4 count above 350 cells/mm3, similar to the report by Lawn et al.16 The impact of current poor nutritional status was of the same magnitude as the effect of low CD4 count with a fourfold higher hazard of TB. Lack of viral suppression (VL greater than 10,000 copies/mL) was associated with a 2.5-fold increased adjusted hazard. Previous reports on viral suppression and risk of developing TB have been conflicting. The US and European Antiretroviral Therapy Cohort Collaboration observed an increased risk of TB (aHR, 2.21) in patients who did not achieve viral suppression (defined as VL less than 400 copies/mL) after 6 months of ART,9 whereas updated VL (greater than 400 copies/mL) was not associated with incident TB in Cape Town.16
Analyzing routine data from a large, public ART clinic in a high-burden country resulted in a large number of patients on ART and TB cases observed and allowed precise estimates of incidence and hazard rates. The reliance on routine procedures, however, resulted in limitations. Patients are mainly diagnosed with TB at primary healthcare clinics, resulting in the lack of bacteriologic confirmation of TB and the possibility of misclassification of TB status. Causes of death are unknown in most cases, resulting in possible underestimation of TB rates.
Results of this study generated evidence to support, develop, and refine strategies to reduce the burden of TB among people receiving ART (Table 4). First, a higher population median CD4 count at ART initiation will significantly reduce the risk of incident TB, reiterating the importance of initiating of ART earlier, as recommended by the World Health Organization.21 Second, prevention of early incident TB will require a shift to a more sensitive intensified case-finding strategy. The standard symptom screen (cough or fever for greater than 2 weeks, night sweats, weight loss of greater than 10%, chest pain, hemoptysis) has been shown to lack sensitivity to detect culture-positive TB in HIV-infected individuals, and smear microscopy has a sensitivity of just 14% compared with liquid culture in individuals presenting for ART.22 Furthermore, it has been shown that subclinical TB occurs in a substantial proportion of HIV-infected individuals.23-25 Although the use of a more sensitive definition of a TB suspect (for example, cough or fever of any duration) is highly likely to result in the detection of more cases, research is needed to determine the most optimal, feasible, and cost-effective TB screening strategy. Third, there is growing evidence that preventive therapy administered before or along with ART has an additive effect on reduction in TB incidence.26-29 In 2010, the South African Department of Health recommended isoniazid preventive therapy for patients on ART in whom active TB has been excluded, but this recommendation was conditional because current evidence is limited to retrospective cohort studies, not randomized controlled trials. To determine the optimal strategy for preventive therapy among individuals on ART, a number of questions will need to be addressed concurrently, including whether the intervention should be universal or targeted to those in whom infection with M. tuberculosis is documented by tuberculin skin test or interferon-γ release assay, whether people with a history of TB should be targeted, what the optimal timing of preventive therapy is in relation to initiation of ART, what the optimal drug regimen is, especially in areas with high background isoniazid resistance, and the optimal duration of preventive therapy in people on ART. Fourth, as more patients are on ART for longer periods of time, adherence to ART to prevent treatment failure and timely introduction of a second-line regimen will become increasingly important in controlling TB among people on ART. Finally, other interventions that could reduce the rate of incident TB include nutritional supplements, treatment of anemia, and poverty reduction.
In conclusion, TB remains a major challenge to the health of people living with HIV in sub-Saharan Africa, even among those receiving ART. Severity of HIV disease is the strongest risk factor for prevalent and early incident TB and response to ART the strongest risk factor for late incident TB. Early initiation of ART will significantly reduce TB incidence among people on ART, but additional interventions such as screening for TB using highly sensitive tools, preventive therapy, nutritional interventions, anemia, and poverty reduction may be needed to further reduce the burden of TB among people on ART.
We gratefully acknowledge the dedicated staff of the Themba Lethu Clinic and all clinic patients for allowing us to use their clinic data for research purposes.
1. World Health Organization. Global Tuberculosis Control 2009. Epidemiology, Strategy, Financing
. Geneva: WHO; 2009.
2. Badri M, Wilson D, Wood R. Effect of highly active antiretroviral therapy on incidence of tuberculosis in South Africa
: a cohort study. Lancet
3. Jones JL, Hanson DL, Dworkin MS, et al. HIV-associated tuberculosis in the era of highly active antiretroviral therapy. The Adult/Adolescent Spectrum of HIV Disease Group. Int J Tuberc Lung Dis
4. Kirk O, Gatell JM, Mocroft A, et al. Infections with Mycobacterium tuberculosis
and Mycobacterium avium
among HIV-infected patients after the introduction of highly active antiretroviral therapy. EuroSIDA Study Group JD. Am J Respir Crit Care Med
5. Moore D, Liechty C, Ekwaru P, et al. Prevalence, incidence and mortality associated with tuberculosis in HIV-infected patients initiating antiretroviral therapy in rural Uganda. AIDS
6. Santoro-Lopes G, de Pinho AM, Harrison LH, et al. Reduced risk of tuberculosis among Brazilian patients with advanced human immunodeficiency virus infection treated with highly active antiretroviral therapy. Clin Infect Dis
7. Girardi E, Antonucci G, Vanacore P, et al. Impact of combination antiretroviral therapy on the risk of tuberculosis among persons with HIV infection. AIDS
8. Miranda A, Morgan M, Jamal L, et al. Impact of antiretroviral therapy on the incidence of tuberculosis: the Brazilian experience, 1995-2001. PLoS One
9. Girardi E, Sabin CA, d'Arminio Monforte A, et al. Incidence of tuberculosis among HIV-infected patients receiving highly active antiretroviral therapy in Europe and North America. Clin Infect Dis
10. Lawn SD, Badri M, Wood R. Tuberculosis among HIV-infected patients receiving HAART: long term incidence and risk factors in a South African cohort. AIDS
11. Bonnet MM, Pinoges LL, Varaine FF, et al. Tuberculosis after HAART initiation in HIV-positive patients from five countries with a high tuberculosis burden. AIDS
12. Brinkhof MW, Egger M, Boulle A, et al. Tuberculosis after initiation of antiretroviral therapy in low-income and high-income countries. Clin Infect Dis
13. Dembele M, Saleri N, Carvalho AC, et al. Incidence of tuberculosis after HAART initiation in a cohort of HIV-positive patients in Burkina Faso. Int J Tuberc Lung Dis
14. Rajasekaran S, Raja K, Jeyaseelan L, et al. Post-HAART tuberculosis in adults and adolescents with HIV in India: incidence, clinical and immunological profile. Indian J Tuberc
15. Lawn SD, Wilkinson RJ, Lipman MC, et al. Immune reconstitution and ‘unmasking’ of tuberculosis during antiretroviral therapy. Am J Respir Crit Care Med
16. Lawn SD, Myer L, Edwards D, et al. Short-term and long-term risk of tuberculosis associated with CD4 cell recovery during antiretroviral therapy in South Africa
17. Sanne IM, Westreich D, Macphail AP, et al. Long term outcomes of antiretroviral therapy in a large HIV/AIDS care clinic in urban South Africa
: a prospective cohort study. J Int AIDS Soc
18. Styblo K. The relation ship between the risk of tuberculosis infection and the risk of developing infectious tuberculosis. Bull Int Union Tuberc Lung Dis
19. Lonnroth K, Williams BG, Cegielski P, et al. A consistent log-linear relationship between tuberculosis incidence and body mass index. Int J Epidemiol
20. Seyler C, Toure S, Messou E, et al. Risk factors for active tuberculosis after antiretroviral treatment initiation in Abidjan. Am J Respir Crit Care Med
21. World Health Organization. Rapid Advice: Antiretroviral Therapy for HIV Infection in Adults and Adolescents
. Geneva: WHO; 2009.
22. Lawn SD, Edwards DJ, Kranzer K, et al. Urine lipoarabinomannan assay for tuberculosis screening before antiretroviral therapy diagnostic yield and association with immune reconstitution disease. AIDS
23. Mtei L, Matee M, Herfort O, et al. High rates of clinical and subclinical tuberculosis among HIV-infected ambulatory subjects in Tanzania. Clin Infect Dis
24. Corbett EL, Bandason T, Cheung YB, et al. Epidemiology of tuberculosis in a high HIV prevalence population provided with enhanced diagnosis of symptomatic disease. PLoS Med
25. Wood R, Middelkoop K, Myer L, et al. Undiagnosed tuberculosis in a community with high HIV prevalence: implications for tuberculosis control. Am J Respir Crit Care Med
26. Golub JE, Pronyk P, Mohapi L, et al. Isoniazid preventive therapy, HAART and tuberculosis risk in HIV-infected adults in South Africa
: a prospective cohort. AIDS
27. Golub JE, Saraceni V, Cavalcante SC, et al. The impact of antiretroviral therapy and isoniazid preventive therapy on tuberculosis incidence in HIV-infected patients in Rio de Janeiro, Brazil. AIDS
28. Innes C. Effectiveness of isoniazid preventive therapy in reducing mortality in patients on ART
. 17th Conference on Retroviruses and Opportunistic Infections; San Francisco; 2010.
29. Samandari T. Randomized, placebo-controlled trial of 6 vs 36 months isoniazid TB preventive therapy for HIV infected adults in Botswana
. 17th Conference on Retroviruses and Opportunistic Infections; San Francisco; 2010.