Tuberculosis (TB) is a leading cause of death for HIV-infected individuals, accounting for up to 11% of AIDS-related mortality worldwide [1,2]. Case fatality rates for TB in HIV-infected patients are extremely high (up to 40%) in sub-Saharan Africa . Close linkage of HIV and TB care services is fundamental for successful management of both epidemics [3–5]. Recent initiatives to expand access to antiretroviral therapy (ART) in Africa are likely to reduce significantly the incidence and mortality of TB in HIV-infected individuals in sub-Saharan Africa. However few data exist that will allow an estimation of this effect .
Studies of ART in African settings have found higher rates of all-cause mortality than in industrialized countries, largely attributed to the late stage at which most patients begin therapy [7,8]. However, it is not known what proportion of this mortality is a result of TB, whether improved TB treatment would modify these high mortality rates, and how TB-related immune reconstitution inflammatory syndrome (IRIS) influences mortality .
To improve understanding of the burden of disease attributable to TB in patients initiating ART in an area with a high prevalence of TB, a nested cohort study was conducted in rural Uganda as part of a randomized clinical trial of three different clinical monitoring strategies. The objective was to understand the clinical and sociodemographic factors associated with having TB at baseline, the incidence of TB over the first 1.5 years of ART, and treatment outcomes for individuals being treated for both HIV and TB. The study also sought to derive an estimated effectiveness of ART in reducing TB incidence and TB-associated mortality.
Setting and study participants
The Home-based AIDS Care Project (HBAC) is a clinical trial of three different monitoring strategies for patients receiving ART in Tororo district, a rural area in eastern Uganda. The trial compares the efficacy of three monitoring regimens: (a) clinical monitoring plus quarterly CD4 T lymphocyte counting and viral load testing, (b) clinical monitoring plus quarterly CD4 cell counting alone, and (c) clinical monitoring alone. Registered clients of The AIDS Support Organization (TASO), a local HIV/AIDS care and support organization in Tororo and Busia districts, were invited to be screened for ART eligibility. The study included participants from a prior diarrhea prevention and co-trimoxazole study, described elsewhere , as well as newly recruited clients. Preventive therapy for TB using isoniazid is not generally available in Uganda, as the Ugandan Ministry of Health does not have a policy recommending its use. Prior to HBAC, TASO clients did not have access to ART in Tororo. HBAC has received ethical approval from the Institutional Review Boards of the Uganda Virus Research Institute, the US Centers for Disease Control and Prevention (CDC) and the University of California, San Francisco. Enrollment began in May 2003.
All study subjects were screened for ART eligibility through clinical and laboratory assessments. Those with CD4 cell counts < 250 cells/μl or who had symptomatic HIV infection [World Health Organization (WHO) stages III or IV, excluding pulmonary TB] were offered ART, with nevirapine, stavudine and lamivudine as the standard regimen. Participants received drug delivery and monitoring by trained lay field officers through weekly home visits and referrals as needed to see clinicians and counsellors at the HBAC clinic; they also had the option of seeking acute care at the HBAC clinic at any time.
In addition, clients were screened for active TB by history taking and clinical examination. Diagnosis and management of TB cases followed the guidelines of the National TB and Leprosy Programme of the Ugandan Ministry of Health. Patients with symptoms of pulmonary TB were required to have two positive sputum smears results for acid-fast bacilli to be diagnosed with TB. Those without two positive smears underwent chest radiography and were given a 2-week course of broad-spectrum antibiotics. If they were still symptomatic after this and their chest radiographs were compatible with TB, they were then diagnosed with smear-negative pulmonary TB. The diagnosis of extrapulmonary TB was based on symptomatology and clinical presentation. The HBAC Medical Evaluation Committee discussed all smear-negative and extrapulmonary TB diagnoses prior to initiating patients on TB therapy. Participants diagnosed with TB were provided with home-based TB treatment. Patients who were ART eligible but were diagnosed with TB completed the first 2 months of TB treatment (when rifampin is used) prior to initiating ART, unless they had CD4 cell cell counts < 50 cells/μl or were severely ill, in which case they were offered ART combination using efavirenz, rather than nevirapine. Participants who were diagnosed with TB while on ART were changed to efavirenz-based ART and were maintained on ART unless they were severely symptomatic from their TB, in which case they were offered a treatment interruption of 1–4 weeks. All ART-eligible participants and HIV-infected household members were prescribed daily co-trimoxazole therapy.
Sputum samples were collected from patients who started TB therapy and incubated for Mycobacterium tuberculosis culture at the laboratories of the Joint Clinical Research Collaboration in Kampala. An independent consultant radiologist later reviewed pre- and post-treatment chest radiographs for those diagnosed with pulmonary TB. Sputum screening for TB was offered and clinic visits were recommended for clients complaining of a cough lasting > 3 weeks or other TB-associated symptoms. Physicians responsible for patients in the two study arms that included laboratory testing received these results on a quarterly basis. All physicians received monthly weights on all patients. Monitoring or diagnostic procedures for TB did not differ between study arms of the clinical trial.
Study physicians at Tororo District Hospital collected clinical information during screening using standardized instruments. All laboratory testing results were transmitted electronically from the CDC laboratory in Entebbe to clinicians in Tororo and into the HBAC database. Field workers completed weekly client monitoring forms that included information on client symptoms, problems with taking medication, or other information that might impact participant health. No attempt was made to define specific causes of death for the purpose of this analysis, but symptoms and clinical conditions present in deceased and living subjects were reported. Clinical and questionnaire data were double-entered using Epi Info 2004 (CDC, Atlanta, Georgia, USA).
Participants taking TB treatment or diagnosed with TB at ART-eligibility screening were compared with those without TB using multivariate logistic regression analysis to examine which baseline factors were independently associated with having TB. Incidence rates of TB diagnoses and mortality rates were calculated using Kaplan–Meier methods and Cox proportional hazards modeling was used to examine associations between baseline variables and TB incidence, and between baseline variables, variables associated with TB diagnosis, and treatment, and mortality. The effect of IRIS on TB-associated mortality was approximated by examining associations of the time at which TB was diagnosed in relation to ART start date, with diagnoses occurring within the first 3 months of ART initiation used as a proxy measure for IRIS.
The expected number of TB cases was calculated by applying the incidence rate of 5.28/100 person-years from an observational study of HIV-infected individuals with WHO clinical stages III or IV or CD4 cell counts < 250 cells/μl in rural Masaka district Uganda from 1990 to 2006 (L. VanderPaal, personal communication) to the study population, multiplied by the number of person-years of follow-up time from the current study. The expected number of TB deaths in follow-up was calculated by applying a mortality rate of 36% to the expected number of incident TB cases above. This is a published 1-year mortality rate for HIV-infected individuals with CD4 cell cell count < 200 cells/μl diagnosed with active TB at Mulago Hospital in Kampala . Finally the estimated effectiveness of ART in reducing TB incidence and associated mortality was calculated using the formula: 1 – (observed/expected) both for the total follow-up time and for time periods that excluded the first 6 months of ART, when clinical responses to therapy are still incomplete. All statistical analyses were conducted in SAS version 9.0 (SAS Institute, Cary, North Carolina, USA).
From 1 May 2003 until 30 June 2005, 1044 patients initiated ART. The median baseline CD4 T lymphocyte count was 127 cells/μl and the median follow-up time was 1.4 years. In total, 75 (7.2%) participants had active TB at baseline; of these, 37 were already receiving TB treatment and 38 were diagnosed with TB during screening for ART eligibility. In the latter group, 23(61%) subjects initiated ART with efavirenz within 2 weeks of starting TB treatment, rather than delay ART initiation because of low baseline CD4 T cell counts. A total of 13 (17%) were classified as having smear-positive pulmonary TB, 32 (43%) were classified as having smear-negative pulmonary TB and 2 (3%) were classified as having extra-pulmonary TB. A further 28 (37%) were not classified because of missing data, largely within the group that were diagnosed with TB prior to study initiation. The median time from the time of first sputum smear collection to start of TB therapy was 7.5 days for smear-positive pulmonary TB subjects and 9.0 days with smear-negative pulmonary TB (P = 0.182). One subject died with an incomplete TB assessment. Only one subject reported previously receiving isoniazid preventive therapy for TB, but 961 (92%) individuals reported daily co-trimoxazole use at baseline. Risk factors associated with TB at baseline included having a body mass index (BMI) ≤ 18 [adjusted odds ratio (AOR), 4.95; 95% confidence interval (CI), 2.95–8.31], having a prior history of TB treatment (adjusted OR, 3.09; 95% CI, 1.85–5.15), and being a participant in a previous co-trimoxazole/safe water vessel study (adjusted OR, 1.94; 95% CI, 1.15–3.27). Younger age was marginally associated with a reduced risk of TB (adjusted OR, 0.97 per year; P = 0.06) (Table 1).
A total of 53 subjects were diagnosed new TB events over a median of 1.4 years of follow up, for a cumulative incidence rate of 3.90/100 person-years. Of these, 12 (23%) were classified as smear-positive pulmonary TB (pulmonary TB), 30 (57%) were classified as smear-negative pulmonary TB and 9 (17%) were classified as extra-pulmonary TB. Two subjects (4%) were not classified because of missing data. Incidence was highest within the first 6 months of initiating ART (7.47/100 person-years) and declined to 2.18/100 person-years from 7 to 18 months after ART initiation (Fig. 1 and Table 2). Incidence in months 1 to 6 was significantly higher than in months 12 to 18 [relative hazard (RH), 4.02; 95% CI 1.78–9.06], but there was no significant difference in TB incidence between months 7–12 and months 13–18 (P = 0.582). TB incidence was associated with a BMI ≤ 18 at baseline (RH, 2.80; 95% CI, 1.59–4.92) and marginally associated with a prior history of TB treatment (RH, 1.74; P = 0.07) (Table 3). A total of 35% of those with pulmonary TB were culture positive. Seventy-nine before and after TB treatment radiographs were reviewed by a consultant radiologist and 70% of them displayed significant improvements from baseline.
Cumulative mortality for those with TB at baseline or follow-up was 17.9/100 person-years, compared with 3.80/100 person-years for those without TB (P < 0.001; log rank test; Fig. 2). Patients with TB accounted for 31 of 88 (35%) all deaths in the first 18 months of the HBAC study. Mortality in those diagnosed with TB was associated with low baseline CD4 cell counts (RH, 0.99 per 1 cell/μl increase; P = 0.03) and marginally associated with a BMI ≤ 18 (RH, 2.04; P = 0.10) and increasing age (RH, 1.04 per year; P = 0.11) (Table 4). A total of 23 participants (43% of incident TB cases) were diagnosed with TB within the first 3 months of initiating ART, but these subjects did not experience an excess mortality risk (RH, 0.86; 95% CI, 0.31–2.41) compared with subjects who were diagnosed later.
The expected number of incident TB cases in the study population in follow-up without ART was calculated to be 72. The observed number of incident cases, 53, represented a 26% reduction in TB incidence. The expected number of TB-associated deaths in follow-up without ART was calculated to be 26, in comparison with the 15 observed. This represented a 42% reduction in TB associated- mortality. Excluding the first 6 months of ART from the comparison increased the estimated effectiveness of ART to a 61% reduction in incidence and a 52% reduction in mortality over expected values.
TB was a common illness among HIV-infected individuals initiating ART in rural Uganda and was associated with 35% of all mortality during follow-up. While 43% of incident TB was diagnosed within the first 3 months after initiating ART, these diagnoses did not impart a greater RH for mortality than TB diagnosed at baseline or later in the course of ART. IRIS, defined here as TB that was diagnosed in the first 3 months of ART, may result in an increased clinical presentation of TB in patients starting ART in rural Uganda, but it does not appear to cause excess mortality compared with people who develop TB later in the course of ART.
While we could not directly measure the impact of ART on TB incidence and mortality in our study, we estimate that the use of ART resulted in a 61% reduction in TB incidence and a 52% reduction in TB-associated mortality, after the first 6 months of ART. Previous studies conducted in Rio de Janeiro, Brazil  and Cape Town, South Africa  have estimated an 80% reduction of TB incidence for HIV-infected adults associated with the use of ART. These differing results may be because the incidence rate we applied to calculate the expected number of TB cases was derived from another rural district in Uganda, where the community prevalence and incidence may differ from those in Tororo. The background rate of TB in the study from South Africa was nearly double our estimated incidence for TB without ART, although the incidence rates for TB while taking ART were similar (3.70/100 person-years in Tororo versus 2.4/100 person-years in Cape Town), Furthermore, median baseline CD4 T cell counts in the South African study were much higher than those in our study (254 compared with 128 cells/μl). It is feasible that ART may have more effectiveness in preventing TB for subjects with higher CD4 cell counts. We are confident that the expected mortality rate of 36% that we applied to our estimates was reasonably accurate as it was similar to the 40% mortality associated with untreated HIV-infected patients with CD4 cell counts < 200 cells/μl in the previous safe water vessel/co-trimoxazole study  conducted in Tororo.
Our study demonstrated that TB incidence declines dramatically after subjects have been treated with ART for 6 months; the incidence rate from 7 to 18 months (2.18/100 person-years) was significantly lower than that in the first 6 months of therapy (7.47/100 person-years). This suggests that previous studies, which did not analyze the first 6 or 12 months of ART separately, may have underestimated the true decline in incident TB associated with ART. This was also the conclusion reached from a recent update of the South African study .
The observation that 7% of subjects were receiving TB treatment or were diagnosed with TB while being screened for ART eligibility highlights the importance of having a close linkage between TB and ART care services. The ongoing high mortality associated with TB diagnosis is a concern for patients initiating ART in areas of high TB prevalence. Much of this mortality may be linked to the late stage at which many subjects in this study initiated ART. As well, TB prevalence, incidence, and mortality were all associated with having a BMI ≤ 18. Among people eligible for ART who already have low CD4 cell counts, a low BMI may provide additional evidence of advanced immunosuppression and potential early mortality. Subjects who were recruited from previous studies conducted in Tororo were more likely to be diagnosed with TB at baseline. This is probably because these individuals receiving better medical care prior to enrollment in the current study, which made them less likely to have undiagnosed TB at baseline when compared with other study subjects.
TB is likely to remain a concern for patients initiating therapy even at higher CD4 counts, as immune responses to TB do not appear to be completely restored by ART, at least in the short to medium term . The systematic screening program for TB that was included as part of screening for ART eligibility in our study may have mitigated some of the TB-associated mortality; however, more research is needed to develop clinical interventions to reduce this mortality rate even further. The provision of isoniazid prophylaxis to all subjects initiating ART has been suggested as an intervention that could be of benefit , but it has yet to be evaluated among people initiating ART in Africa.
It is possible that some overdiagnosis of TB occurred, as our diagnostic methods were restricted to sputum smears, radiographs, and clinical judgment. Mycobacterial culture, which is not commonly used in rural African settings for diagnostic purposes, was only available for those who had already been diagnosed with TB. A study of HIV-infected patients hospitalized for acid-fast bacilli smear-negative respiratory infections in an urban Ugandan hospital found that 38% had Pneumocystistis jiroveci pneumonia and only 24% had TB . However, while only 35% of our study subjects with pulmonary TB had positive sputum cultures, approximately 70% showed significant improvements in their radiographs from baseline on TB treatment. Furthermore, we found no difference in risk for mortality associated with sputum culture results, suggesting that the poor yield of sputum culture was more likely related to difficulties in sputum collection, storage, and transport than because of true absence of mycobacteria.
This study was limited by the lack of directly comparable data on TB incidence and mortality from ART-eligible individuals not receiving ART in Uganda. However, obvious ethical considerations prevented us from collecting such data, and historical comparisons from the study area were not possible as previous studies had not collected such information. Additionally, it is possible that some of our analyses, such as those comparing mortality associated with TB diagnoses within 3 months of initiating ART or mortality associated with type of TB, lacked statistical power to show differences. However, this study has the largest number of incident and prevalent TB cases reported from a developing country setting and, therefore, represents a marked improvement in statistical power from those previously published.
TB remains an important cause of illness and death for HIV-infected individuals receiving ART in Africa. Close linkage of TB and HIV diagnostic and treatment services will be necessary to reduce the burden of both diseases. Further research is needed to define better clinical strategies to reduce TB-associated mortality, especially in the first few months of ART.
The authors would like to thank the field officers, counselors and clinical staff who care for patients in the HBAC project, the informatics team at CDC-Uganda who compiled the data for analysis, and the participants in the HBAC project. We also thank Dr Lieve VanderPaal of the British Medical Research Council in Entebbe for providing us with TB incidence rates from Masaka, Uganda. We also acknowledge the support of the Ugandan Ministry of Health and TASO.
1. Corbett EL, Watt CJ, Walker N, Maher D, Williams BG, Raviglione MC, et al
. The growing burden of tuberculosis: global trends and interactions with the HIV epidemic. Arch Int Med 2003; 163:1009–1021.
2. Holmes CB, Losina E, Walensky RP, Yazdanpanah Y, Freedberg KA. Review of human immunodeficiency virus type 1-related opportunistic infections in sub-Saharan Africa. Clin Infect Dis 2003; 36:652–662.
3. Abdool-Karim SS, Abdool-Karim Q, Friedland G, Lalloo U, El Sadr WM. Implementing antiretroviral therapy in resource-constrained settings: opportunities and challenges in integrating HIV and tuberculosis care. AIDS 2004; 18:975–979.
4. Williams BG, Dye C. Antiretroviral drugs for tuberculosis control in the era of HIV/AIDS. Science 2003; 301:1535–1537.
5. Elzinga G, Raviglione MC, Maher D. Scale up: meeting targets in global tuberculosis control. Lancet 2004; 363:814–819.
6. Badri M, Wilson D, Wood R. Effect of highly active antiretroviral therapy on incidence of tuberculosis in south Africa: a cohort study. Lancet 2002; 359:2059–2064.
7. Laurent C, Diakhate N, Gueye NF, Toure MA, Sow PS, Faye MA, et al
. The Senegalese government's highly active antiretroviral therapy initiative: an 18-month follow-up study. AIDS 2002; 16:1363–1370.
8. Coetzee D, Hildebrand K, Boulle A, Maartens G, Louis F, Labatala V, et al
. Outcomes after two years of providing antiretroviral treatment in Khayelitsha, South Africa. AIDS 2004; 18:887–895.
9. Lawn SD, Bekker LG, Miller RF. Immune reconstitution disease associated with mycobacterial infections in HIV-infected individuals receiving antiretrovirals. Lancet Infect Dis 2005; 5:361–373.
10. Mermin J, Lule J, Ekwaru JP, Malamba S, Downing R, Ransom R, et al
. Effect of cotrimoxazole prophylaxis on morbidity, mortality, CD4 cell count, and HIV viral load among persons with HIV in rural Uganda
. Lancet 2004; 364:1428–1434.
11. Whalen CC, Nsubuga P, Okwera A, Johnson JL, Hom DL, Michael NL, et al
. Impact of pulmonary tuberculosis on survival of HIV-infected adults: a prospective epidemiologic study in Uganda
. AIDS 2000; 14:1219–1228.
12. Santoro-Lopes G, de Pinho AMF, Harrison LH, Schechter M. Reduced risk of tuberculosis among Brazilian patients with advanced human immunodeficiency virus infection treated with highly active antiretroviral therapy. Clin Infect Dis 2002; 34:543–546.
13. 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 2005; 19:2109–2116.
14. Lawn SD, Bekker LG, Wood R. How effectively does HAART restore immune responses to Mycobacterium tuberculosis? Implications for tuberculosis control. AIDS 2005; 19:1113–1124.
15. Grant A, Fielding K. Charalambous S. Stenson A. Whetham J. Puso T. et al.Risk factors for early mortality among HIV-infected individuals starting antiretroviral therapy in South Africa.Third International AIDS Society Conference on HIV Pathogenesis and Treatment.
Rio de Janeiro, July 2005 [abstract MoPe11.6C25].
16. Worodria W, Okot-Nwang M, Yoo SD, Aisu T. Causes of lower respiratory infection in HIV-infected Ugandan adults who are sputum AFB smear-negative. Int J Tuberc Lung Dis 2003; 7:117–123.