Abstract & Commentary

Antiretroviral initiation during or following treatment of TB

By Dean L. Winslow, MD, FACP, FIDSA, Chief, Division of AIDS Medicine, Santa Clara Valley Medical Center; Clinical Professor, Stanford University School of Medicine.

Synopsis: In this study, 642 patients coinfected with HIV and TB were studied in an open-label, randomized, controlled trial in which patients were assigned to initiating antiretroviral therapy (HAART) early, later during the course of TB treatment, or after the completion of TB treatment. Initiation of antiretroviral therapy during TB treatment improved survival.

Source: Abdool Karim SS, et al. Timing of initiation of antiretroviral drugs during tuberculosis therapy. N Engl J Med. 2010; 362:697-706.

An open-label, randomized, controlled trial of early vs. delayed initiation of HAART was conducted in 642 patients with active TB and HIV infection in Durban, South Africa. Patients were randomized 1:1:1 to one of three study groups. In the "early integrated therapy" group, ARV therapy (consisting of ddI, 3TC, and efavirenz) was initiated within four weeks after starting TB treatment. In the "late integrated therapy" group, HAART was initiated within four weeks after the completion of the intensive phase of TB treatment (initial two months of INH, rifampin, PZA, and EMB in patients with first episode TB and three months of RIPE plus SM for two months in patients with recurrent TB). In the "sequential therapy" group, HAART was initiated within four weeks after completion of the 6-8 month entire course of TB treatment.

The primary endpoint of the study was death. There were 25 deaths in the integrated therapy group (rate 5.4 per 100 person-years) vs. 27 deaths in the sequential therapy group (rate 12.1 per 100 person-years).

Although not quite reaching statistical significance, the combined integrated therapy groups were more likely to be cured, or to successfully complete TB therapy, and were less likely to require interruption of TB treatment. Integrated therapy patients were significantly more likely to achieve HIV RNA < 400 copies/mL at 12 months following randomization than were sequential therapy patients (90% vs. 78%), and to have a better CD4+ lymphocyte count incremental increase (207 vs. 84 cells/uL).


This is an important study which is likely to alter clinical practice in both the developed and developing worlds. Current practice of many TB clinics is to defer antiretroviral therapy in patients with HIV/TB coinfection unless CD4+ count is < 200 cells/uL.

This study provides strong clinical support for not delaying HAART while treating TB. However, one of the limitations of this study is that the somewhat unusual HAART regimen studied (ddI, 3TC, and EFV) does not result in significant drug interactions with the TB medications used. From a practical standpoint, it is most problematic to use ritonavir-boosted protease inhibitor (PI) regimens in patients receiving rifampin. Rifampin is a potent inducer of the cytochrome P450 CyP 3A/4 isoforms, and results in marked reduction of PI serum levels, even in the presence of ritonavir boosting. Rifabutin in significantly reduced doses can be substituted for rifampin but, in combination with ritonavir, it may cause ocular toxicity. Also, concern has recently been raised about frequent subtherapeutic levels of rifabutin seen when used in combination with ARVs.

The bottom line from this study, from my perspective, is that EFV-containing HAART regimens can and should be started early during TB treatment in patients whose baseline HIV genotypes do not show primary NNRTI resistance. However, additional studies should be done to better define optimal dosing of TB regimens and timing of initiation of HAART, with respect to TB treatment in patients who require ritonavir-boosted PI regimens. Although I am not aware of published data to support this practice, another attractive treatment option in patients on TB treatment that require non-EFV-containing HAART regimens would be a combination of NRTIs -- plus the integrase inhibitor raltegravir (due to the paucity of drug interactions associated with raltegravir, which it is not metabolized by the cytochrome P450 system.)