University of California, San Francisco Logo

University of California, San Francisco | About UCSF | Search UCSF | UCSF Medical Center

Early Antiretroviral Therapy in Highly Immunocompromised Patients Coinfected with HIV and Tuberculosis: Much Immune Reconstitution Disease, but Perhaps Fewer Deaths?
HIV InSite's Coverage of the 14th Conference on Retroviruses and Opportunistic Infections
Oliver Bacon, MD, MPH, HIV InSite Medical Editor, Center for HIV Information, University of California San Francisco

Published March 27, 2007
Back to Main Conference Coverage Page



The approach to treatment of persons coinfected with tuberculosis (TB) and HIV is a challenge. Although it is accepted that HIV-infected patients with active TB should receive multidrug antituberculosis therapy (ATT) before starting multidrug antiretroviral therapy (ART), the timing of ART initiation and components of ART in the setting of ATT remain controversial among the various published guidelines.(1,2,3) The possibly high incidence and morbidity and mortality of TB immune reconstitution disease (IRD) seem to justify delaying ART. Intriguing evidence from a large community-based ART program in Cape Town, South Africa, was presented at CROI to support early ART in the setting of coinfected patients with advanced immunosuppression who are starting ATT, despite the high rate of TB-IRD seen in this population.


Abstract: Lawn S, Myer L, Bekker LG, et al. TB-associated Immune Reconstitution Disease: Incidence, Risk Factors, and Effect within an ART Program in Sub-Saharan Africa. Abstract 863.

Description: The Guguletu Community Health Center, in Guguletu-Nyanga District, Capetown, South Africa, is situated in an area of high HIV prevalence (28.5%), with an incidence of TB of 1,400/100,000 per year. Over a 3-year period, a subset of clinic patients was prospectively enrolled in a cohort study at the time of starting ART (consisting of stavudine, lamivudine, and efavirenz). The cohort comprised 160 patients who were retrospectively identified as taking ATT for active TB at the time of enrollment. Cases of TB-IRD were ascertained in retrospect.

Results: Of the 160 participants, 19 (12%) met the case definition for TB-IRD: prevalent TB and ATT before starting ART; response to ATT; deterioration of TB after starting ART; and no evidence of nonadherence to ATT, another opportunistic infection, multiple drug-resistant TB, or a drug adverse effect. 32% of participants starting ART within 2 months of TB diagnosis developed TB-IRD. Clinical manifestations ranged from exacerbation of existing pulmonary symptoms or lymphadenitis (10/19) to the addition of new, largely abdominal, manifestations of disease (7/19), to death from disseminated disease (2/19). Investigators reported that most cases of IRD were self-limiting and that 7 members (4%) of the cohort used secondary health services. In multivariate analysis adjusting for age, sex, baseline CD4 count and viral load, site of TB, and time from TB diagnosis to initiation of ART, risk of TB-IRD significantly increased with lower CD4 count (relative risk: +5.3 per 100 cell/µL drop in CD4 count; p = .018) and decreasing time from TB diagnosis to start of ART. Those starting ART 0-1, 1-2, or 2-3 months after TB diagnosis had a relative risk of TB-IRD that was 70 (p < .001), 11 (p = .007), and 7 (p = .030) times that of patients who started ART >3 months after TB diagnosis, respectively. The authors conclude that the risk of TB-associated IRD in this setting is very high for those with low baseline CD4 cell counts initiating ART early in the course of ATT. However, most cases were self-limiting; overall secondary health care utilization and mortality risk from IRD were low.

Abstract: Lawn S, Myer L, Bekker LG, et al. Early Mortality among Patients with HIV-associated TB in Africa: Implications for the Time to Initiate ART. Abstract 81.

Description: Mortality during the period before ART initiation through 16 weeks of ART was compared among 213 HIV/TB-coinfected patients and their 675 HIV-positive, TB-uninfected peers in a cohort of patients entering the Guguletu community ART program in Capetown, South Africa. Survival analysis was performed on a subcohort of 73 patients whose TB was diagnosed within the program but before initiation of ART.

Results: In the larger cohort, patients coinfected with TB were 1.8 times as likely to die as those without TB (39.7 deaths per 100 person-years vs 21.7 deaths per 100 person-years; p = .003). In multivariate analysis adjusting for age, sex, viral load and CD4 count, TB (none vs inactive vs active), and World Health Organization (WHO) stage of HIV (stage 4 vs stages 1-3), only baseline WHO stage 4 HIV disease (adjusted hazard ratio [AHR]: 2.94; 95% confidence interval [CI]: 1.80-4.82) and baseline CD4 count <100 cells/µL (AHR: 2.85; 95% CI: 1.52-5.34) were independently associated with mortality.

In the survival analysis of 73 patients diagnosed with active TB upon entry to the program, the median time from TB diagnosis to ART was 42 days, and 48 patients (66%) had started ART by the time of analysis. There were no significant differences in median CD4 count or proportion with stage 4 disease between those who had started ART and those who had not. Mortality was significantly higher in patients waiting for ART: Of the 14 deaths, 10 (71%) occurred in group waiting for ART and 4 occurred in those taking ART (p < .01). Two of the deaths on ART were thought to be associated with disseminated TB-IRD, and occurred in patients who had started ART 28 and 47 days post-TB diagnosis and were not given steroids (determination of TB-IRD was made retrospectively, during the data analysis, rather than in real time).


These two studies by Lawn et al examine data from the same large, community-based ART program in an area of high TB prevalence. They demonstrate the very high risk of encountering TB-IRD when starting ART in patients with active TB disease and advanced immunosuppression: Overall, roughly one third of patients starting ART within 2 months of a TB diagnosis developed TB-IRD. The association between proximity of ART initiation to TB diagnosis and development of TB-IRD shown in the multivariate analysis has the characteristics of a dose-response effect. Despite 2 IRD-related deaths, however, the risk of death associated with waiting to start ART appears greater. Although mortality was significantly higher among those coinfected with HIV and TB vs those infected with HIV alone (odds ratio: 1.82), advanced immunosuppression, rather than TB, remained an independent risk factor for mortality in multivariate analysis. Moreover, survival analysis among the 74 participants whose TB was diagnosed during the ART program showed that early mortality was vastly higher among those waiting to start ART (10/14) than among those who had already started ART (4/14), even when corrected for baseline patient or disease characteristics. Two of the four deaths on ART were thought to be related to TB-IRD. Although IRD could be said to account for 50% of the mortality among patients on ART, it is important to note that the diagnosis of IRD was made at the time of data analysis, rather than when patients were being seen, and measures to ameliorate IRD therefore had not been taken. The mortality data come from a relatively small portion of a clinic cohort, and are not from randomized study participants. Firm recommendations regarding early initiation of ART in coinfected, highly immunosuppressed patients therefore await data from randomized controlled trials. Indeed, given the strong association between TB-IRD and proximity of ART, the incidence of IRD would be expected to rise with early treatment among the highly immunosuppressed. It is tempting to speculate, based on the second study, however, that this would be outweighed by a fall in overall mortality.


  1. U.S. Department of Health and Human Services. Guidelines for the Use of Antiretroviral Agents in HIV-1-Infected Adults and Adolescents. October 10, 2006.
  2. Pozniak AL, Miller RF, Lipman MC, et al. BHIVA treatment guidelines for tuberculosis (TB)/HIV infection 2005. HIV Med. 2005 Jul;6 Suppl 2:62-83.
  3. World Health Organization. Antiretroviral Therapy for HIV Infection in Adults and Adolescents in Resource-Limited Settings: Towards Universal Access. Recommendations for a Public Health Approach. 2006 Revision. World Health Organization: Geneva; 2006.