Risk Factors for Tuberculosis among HIV-infected Patients Receiving Antiretroviral Treatment

Abstract

Although tuberculosis (TB) incidence among HIV-infected cohorts receiving highly active antiretroviral treatment (HAART) is reduced by 70–90%, it remains substantially higher than among HIV-noninfected patients (1, 2). Strategies to reduce the burden ofTBduringHAARTare needed.We therefore readwith interest the article by Seyler and colleagues, which reported risk factors for TB among patients receiving HAART in Abidjan (3). Using multivariate analysis, Seyler and colleagues found that a past history of TB was the only significant risk factor for TB during HAART (3). The 31 patients in the cohort with a past history of TB had all been declared cured following 6 mo or more of rifampicin-containing anti-TB treatment, and yet six developed TB during HAART over a median of 26 mo observation. This represents an unusually high recurrence rate (4) even without taking into account the major protective effect of HAART. Of note, however, 33% of the incident cases were smearand culture-negative; potential misdiagnoses of active TB among patients with symptomatic posttuberculous lung disease may have confounded past history of TB as a risk factor. Advanced immunodeficiency might be expected to be associated with an ongoing increased risk of TB during HAART, at least until a satisfactory degree of immune restoration has been achieved. We have studied a large cohort in Cape Town (described in Reference 2) that includes patients with a diverse range of CD4 cell counts and clinical stages. Analysis reveals that the ongoing risk of TB during HAART is much greater among those with low nadir CD4 cell counts and advanced clinical stage of disease; past history of TB is not a risk factor. The disparities with the findings of Seyler and colleagues may reflect differences in diagnostic criteria for TB and the fact that the restricted enrollment criteria for the Abidjan cohort resulted in inclusion of few patients with stage 1 or 2 disease or CD4 cell counts greater than 200 cells/ l; the resulting cohort composition diminishes the power to assess CD4 count and clinical stage as risk factors. This is a problem common to analysis of data from many cohorts receiving HAART in Africa where often only those with advanced disease are treated. If it is confirmed that the risk of TB during HAART is associated with advanced pretreatment immunodeficiency, the risk might be decreased by earlier initiation of HAART. In addition, those patients with advanced immunodeficiency may derive greatest benefit from adjunctive treatments such as isoniazid prophylaxis.