Previous immune damage and response to HIV treatment associated with TB after starting antiretrovirals

HIV-positive individuals in Europe and North America are considerably more likely to develop tuberculosis (TB) than the general population, even when taking potent antiretroviral therapy, according to a large study published in the December 15^th edition of Clinical Infectious Diseases. Investigators from the Antiretroviral Cohort Collaboration found that the level of immune suppression before anti-HIV therapy was started was predictive of the risk of developing tuberculosis, as was the six-month immunologic and virologic response to HIV treatment.

An accompanying editorial states that the study’s findings have important implications for antiretroviral treatment programmes in low-income countries.

There are approximately 500,000 cases of tuberculosis annually in HIV-positive individuals. Fewer cases of tuberculosis are seen in individuals who take effective antiretroviral therapy than those who do not, however little is known about the risk of developing of tuberculosis after antiretroviral treatment is started and if either the level of immune damage before HIV drugs are started or the response to such treatment affects the risk of tuberculosis.

Investigators from twelve large cohorts of HIV-positive patients in Europe and North America looked at the incidence of new tuberculosis cases in HIV-positive individuals who had never taken anti-HIV drugs before. Patients who had ever been diagnosed with AIDS were excluded from the study. Information on tuberculosis cases during the first three years of antiretroviral therapy were obtained. The investigators gathered information on the patients’ sex, mode of HIV infection, CD4 cell count and viral load when anti-HIV drugs were started, type of anti-HIV regimen used, and year in which antiretroviral therapy was started. Data were also gathered on increases in CD4 cell count and fall in viral load after six months of antiretroviral therapy. Median CD4 cell count at the time HIV treatment was started was 280 cells/mmsup>3 and median viral load was a little under 60,000 copies/ml.

A total of just under 37,000 patient years of follow-up were available for analysis. During this time, 173 patients developed tuberculosis, an incidence of five cases per 1000 patient years, approximately five times the incidence that would be seen in the HIV-negative population in Europe and the United States.

Using multivariate analysis, the investigators found that the incidence of tuberculosis was more frequent in individuals who started HIV therapy with a low CD4 cell count (below 200 cells/mm³, incidence seven per 1000 patient years versus above 350 cells/mm³, incidence two per 1000 patient years), a high viral load (above 100,000 copies/ml, incidence six per 1000 patient years versus below 100,000 copies/ml six per 1000 patient years).

Analysis also revealed that the risk of developing tuberculosis during the first three years of HIV therapy differed according to HIV risk group. Incidence was lower amongst men (four cases per 1000 patient years) than women (six cases per 1000 patient years), and lower amongst gay men (two cases per 1000 patient years) than injecting drug users (five cases per 1000 patient years) or individuals who had heterosexual sex as their HIV risk activity (seven cases per 1000 patient years).

Attention was then turned to the six-month response to antiretroviral therapy and the subsequent risk of tuberculosis. Six months after starting HIV drugs, the median increase in CD4 cell count was 107 cells/mm³ and 67% of patients has achieved a viral load below 400 copies/ml. A total of 88 cases of tuberculosis were seen after the initial six-month period. There were significantly more likely to occur in patients who had either a low CD4 cell count when anti-HIV treatment was first started (p = 0.009), or had experienced only modest increases in their CD4 cell count during the first six months of HIV therapy (p = 0.07). The investigators also found that patients who had a viral load above 400 copies/ml after six months of antiretrovirals were at greater risk of tuberculosis in the future (p = 0.003). Once again, risk group was found to be a significant factor, with gay men having the lowest risk of tuberculosis and injecting drug users the highest (p = 0.006). The year in which anti-HIV therapy was started also emerged as significant factor, with patients who started treatment after 2001 having a higher risk of tuberculosis after six months of HIV therapy than those who started HIV treatment between 1996 and 2000 (p = 0.003).

“Those at higher risk of tuberculosis are those with poorer immunological or virological responses to HAART and those who starting HAART with more-advanced immuno-suppression”, comment the investigators.

The investigators note that the incidence of tuberculosis was particularly high – 13 per 1000 patient years – during the first three months of HIV therapy. This rate approached that seen for untreated HIV-infected individuals and may be due either to late diagnosis of HIV or immune reconstitution.

Given that tuberculosis still occurred with increased frequency during long-term anti-HIV therapy the investigators speculate, “complete restoration of cellular immunity against tuberculosis may not be possible.” The investigators suggest that the lower incidence of tuberculosis seen in gay men reflected different background levels of tuberculosis seen in different ethnic and social groups.

The investigators conclude, “the use of HAART appears to progressively decrease, but not to abrogate, tuberculosis risk. If this is the case…then HAART may have only a limited role in tuberculosis-control measures in countries with a high burden of tuberculosis and HIV infection.” This conclusion is echoed in the accompanying editorial, which notes that in many resource-limited settings anti-HIV therapy is not initiated until individuals have also experienced severe immune damage, and according to this study, would therefore remain at high risk of tuberculosis despite antiretroviral treatment.