CROI: HIV treatment should not be delayed in TB patients with low CD4 counts, South African study shows

Keith Alcorn, Virginia Differding
Published: 28 February 2007


Although many authorities recommend a window period of four to eight weeks within which to stabilise patients on TB medications before starting ART, in this population, these findings narrow that window period to within four weeks of TB diagnosis.

Two deaths were attributed to immune reconstitution inflammatory disease versus ten deaths in those patients who did not receive ART or in whom ART was deferred longer than four weeks, regardless of baseline patient or disease characteristics.

Although clinical management challenges are involved in initiating ART soon after initiating TB treatment, in patients with CD4 cell counts below 100 cells/mm3, the benefit conclusively outweighed the risk, Stephen Lawn concluded.

He noted that the median delay between TB diagnosis and beginning ART was 42 days, but in patients diagnosed in TB clinics and then referred to the ART programme, the delay was often greater than three months.

People with HIV who are diagnosed with tuberculosis (TB) have a 50% probability of death if they wait three months to begin HIV treatment, a large review of South African HIV-positive patients shows. In contrast, few who began antiretroviral treatment soon after TB treatment died as a result of immune reconstitution inflammatory syndrome.

“Doctors in South Africa routinely cite concerns about IRIS as a reason for delaying treatment. These findings blow that view out of the water,” said Dr Stephen Lawn of the Desmond Tutu HIV Centre in Cape Town, speaking on Tuesday at the Fourteenth Conference on Retroviruses and Opportunistic Infections in Los Angeles. He was presenting results from a study of 888 HIV-positive patients enrolled for antiretroviral treatment.

In the prospective South African study, the risk of mortality was independently associated with a CD4 cell count below 100 cells/mm3 or baseline WHO stage 4 disease, and not with the presence of TB/HIV coinfection alone.

The study looked at the risk of death in people enrolled into an HIV treatment programme, comparing the risk between people with TB coinfection at baseline or within the first year of joining the programme, and those who remained free of TB.

Twenty-six per cent had TB on entry to the programme, and over one-third of all participants had overlapping treatment for TB and HIV in the first year.

Mortality among patients with TB coinfection (n=213) was 1.82-fold higher than in those not coinfected (n=675); however, multivariate analysis did not demonstrate an independent association between TB coinfection and mortality.

In the entire cohort (n=888), mortality was independently associated with CD4 count <100 cells/mm3 (AHR = 2.85, 95%CI 1.52 to 5.34) and baseline WHO stage 4 AIDS classification (AHR = 2.94, 95%CI 1.80 to 4.82).

Sixty-six per cent of the deaths within the first three months of enrolment occurred in TB patients who did not receive HIV treatment.

Patients were followed in the time before any ARV or TB treatment began and up to 16 weeks following ART initiation. Among 73 coinfected patients, 48 began ARV therapy between four to ten weeks after their TB diagnosis.

Ten patients died before ART was initiated; seven of those patients died within a month of beginning TB treatment.

Only four deaths occurred in patients receiving ART. Immune reconstitution inflammatory syndrome was the cause of two of those deaths.


Lawn S et al. Early mortality among patients with HIV-associated TB in Africa: implications for the time to initiate ART. Fourteenth Conference on Retroviruses and Opportunistic Infections, Los Angeles, abstract 81, 2007.