Isoniazid and ARVs reduce TB risk by 90% in two South African clinics

Isoniazid preventive therapy (IPT) combined with antiretroviral therapy (ART) reduced the risk of developing active tuberculosis (TB) in people with HIV by almost 90% compared with no treatment, a South African study has shown. Use of both interventions had a significantly greater impact on TB incidence than either intervention used alone, according to findings reported in the March 13^th edition of AIDS.

The researchers suggest that the data from this prospective cohort study supports the need for the provision of IPT together with the roll-out of ARVs as a cheap and cost-effective strategy against TB.

TB remains the leading opportunistic infection and cause of death for people infected with HIV in resource-poor countries. The World Health Organization recommends IPT for people infected with HIV in high-prevalence countries, but few countries have this policy in place. Isoniazid is taken once a day, either in a six-month course of treatment, or indefinitely in some settings, and can be initiated before an individual qualifies for antiretroviral therapy.

Results of a recent study in Brazil showed that while IPT and ART used alone are effective against TB, the combination of the two offers significantly greater protection. In sub-Saharan Africa, where the incidence of TB is considerably higher, data of this kind are lacking.

From 1 June 2003 until 31 December 2007 a total of 3868 HIV-infected adults receiving primary HIV care at two clinics, one in a large urban setting, the Perinatal HIV Research Unit (PHRU) in Johannesburg, and Tintswalo Hospital in a remote rural area of Mpumalanga province, were followed.

Both sites offered IPT for six months. At PHRU it was given following a positive tuberculin skin test (TST), whereas at Tintswalo no test was given. A government decision to stop providing isoniazid tablets meant that few people were on IPT.

Government guidelines required the provision of ART on diagnosis with WHO stage 3 (pulmonary tuberculosis) or 4 (extra-pulmonary tuberculosis) disease or CD4 counts of <200 cells/mm³ and began at the PHRU in 2004 and from 2005 at Tintswalo.

A diagnosis of tuberculosis was the primary outcome irrespective of where the diagnosis took place. The primary exposures of interest were receipt of IPT and/or ART.

Follow-up time was divided into 4 exposure categories:

• Patients having received neither IPT nor ART
• Patients who received IPT but not ART
• Patients who received ART without prior IPT; and
• Patients on ART who had received IPT and then ART.

Follow-up time began with the patient’s baseline cohort visit and ended at the first incidence of tuberculosis or at the last recorded clinic visit.

Of the total 3868 of HIV-infected adults with a minimum of two data-collection visits, 282 were excluded from the analysis because of missing CD4 counts and a further 838 due to a history of tuberculosis or recent diagnosis. Of the remaining 2778 patients, 80% were female and 76% from the urban setting.

The overall tuberculosis incidence rate was 6.2/100 person-years. Factors that put patients at an increased risk for tuberculosis included being male, from a rural area, over 30 years of age and having a CD4 count of <100.

Tuberculosis incidence for treatment-naive patients was 7.1/100 person-years (95% CI 6.2 to 8.2) and in comparison decreased by 27% in those who only got IPT (IR=5.2/100 person-years; IRR=0.73; 95% CI 0.44), and by 35% in those receiving only ART (IR=4.6/100 person-years; IRR=0.65; 95% CI 0.46 to 0.91) and by 85% (IR=1.1/100 person-years; IRR=0.15; CI 0.004 to 0.85) in those who had received IPT before getting ART.

In the final adjusted model similar trends to the incidence rate ratios emerged. While IPT alone did not result in a statistically significant reduction in risk compared to treatment-naive patients after adjustment for CD4 counts and other variables, living in a rural area and a CD4 count of <100 were linked with an increased risk.

Despite concerns raised about the failure to detect active TB in people given isoniazid preventive therapy in high-burden settings, selection for isoniazid resistance as a result of single-drug prophylaxis has not been an issue in studies to date.

Several limitations are noted by the authors. As with other studies, adherence to IPT is an issue, with only 59% of patients receiving 6 months of IPT.

While the sample of patients receiving both IPT and ART is relatively small, the authors believe that since follow-up in all groups was at least one year the results are convincing.

The observational nature of the study raises the possibility for bias.

Adjustment for possible confounders eliminates the suggestion that stronger patients survived long enough to receive IPT and then ART according to the authors.

The authors conclude that their findings support widespread implementation of isoniazid preventive therapy prior to antiretroviral therapy in high-burden countries. They suggest that IPT be given prior to ART when CD4 counts are above 300 cells/mm³ to reduce the risk of early diagnosis with TB following ART initiation.