Active Tuberculosis Does Not Always Imply the Initiation of Highly Active Antiretroviral Therapy in HIV-Infected Patients

Abstract

To the Editor: Infection with HIV is an important risk factor for tuberculosis (TB). One-half million cases of TB attributable to HIV infection occur worldwide each year.1 Since the advent of highly active antiretroviral therapy (HAART), the morbidity and mortality of HIV-infected patients have declined.2 However, concomitant TB treatment and HAART are hampered by a high pill burden, complex drug interactions, treatment-related adverse events, and paradoxic reactions.3,4 Studies that have assessed the effect of HAART in the course of TB and HIV coinfection are few.5-8 All these retrospective studies suggest that coinfected patients with a CD4 count <100 cells/mm3 would probably benefit from early initiation of HAART. Unlike other HIV-associated opportunistic infections, TB may occur at high CD4 cell counts. The indication for HAART in these particular cases has not been assessed, and current antiretroviral (ARV) treatment guidelines are based on expert opinion.9 We evaluated retrospectively 93 episodes of TB diagnosed in 87 HIV-infected patients from January 2000 to February 2004 at Bichat Hospital in an attempt to assess the clinical aspects and outcome of TB, according to the level of immunosuppression at TB diagnosis and to the prescription of HAART in the course of TB treatment. Of our 87 patients, 58 (67%) originated from sub-Saharan Africa. At baseline, the CD4 count was ≥250 cells/mm3 in 25 cases (27%; group 1) and <250 cells/mm3 in 68 cases (73%; group 2). The mean plasma viral load (log) was 4.68 ± 1.58. The clinical presentation, treatment, and outcome of TB cases are summarized in Table 1. TB treatment was completed in 61 (66%) of 93 cases. The standard 6-month course was applied in 9 of 61 cases. An extended duration of TB treatment was prescribed in 52 cases, mainly because of extrapulmonary disease. A total of 22 adverse events leading to a change in TB therapy were recorded in 20 cases (22%). Adverse events led to interruption of TB treatment in 12 cases (13%). Toxicity of TB treatment was observed in 14 cases (22%) with concomitant HAART and in 6 cases (21%) without concomitant HAART. The occurrence of TB treatment toxicity led to HAART interruption in 3 cases. In these 3 cases, HAART was initiated <1 month after TB treatment initiation.TABLE 1: Presentation, Treatment, and Outcome of TB CasesHAART was prescribed in 64 cases, including initiation of a first HAART regimen in 36 cases. In 29 cases, patients did not receive ARV therapy during treatment of TB. In group 1, HAART was not initiated in 11 naive patients and was interrupted in 5 patients previously exposed to HAART, although not immunosuppressed. In group 2, HAART was not initiated in 1 patient who died 1 month after TB diagnosis, in 6 patients who were lost to follow-up, and in 2 patients who interrupted TB treatment prematurely and experienced early TB relapse. In 4 cases, observance difficulties prevented initiation of ARV therapy. In the case of concomitant ARV and TB treatment, 12 (19%) of 64 patients described side effects motivating interruption or change of HAART. Eleven patients developed an immune reconstitution inflammatory syndrome (IRIS). In 9 cases, patients belonged to group 2. Ten patients were receiving ARV treatment. Overall, 19 patients (20%) were lost to follow-up. During TB therapy, 7 patients died. Six (86%) of 7 deaths occurred in group 2, with a CD4 count <100 cells/mm3 at the time of TB diagnosis. Six patients experienced TB relapse in 7 cases. In only 1 case did TB relapse occur despite well-conducted TB therapy of 12 months. During the study period, a further AIDS-defining event (ADE) occurred in 20 cases. The CD4 count was <100 cells/mm3 in 16 cases. In 60 cases, patients completed treatment and achieved TB cure. In group 1, the cure rate was 56%, whatever the prescription of HAART. None of the 16 patients who did not start HAART experienced a new ADE or a significant decrease in CD4 count (456 ± 141 cells/ mm3 vs. 493 ± 186 cells/mm3). In group 2, the cure rate was higher when patients received HAART (76% vs. 31% with and without HAART, respectively; P = 0.003). In this study, patients mainly originated from endemic countries for TB and HIV. This could explain the high proportion of patients who were not deeply immunosuppressed at the time of TB diagnosis. Overall, 19 patients (20%) were lost to follow-up. Because of this high proportion of patients lost to follow-up, only 66% of TB treatments were successfully completed. Leonard et al10 found that the major factor improving the prognosis of patients coinfected with TB and HIV from 1991 to 2000 was improved follow-up during TB treatment, leading to an increase in survival, especially between 1991 and 1994 before the advent of HAART. Nevertheless, there is evidence to support the benefit of HAART in reducing TB mortality in HIV-infected patients. Several retrospective studies showed an improvement in the survival of HIV-infected patients with TB in the era of HAART.11,12 The decision regarding when to initiate ARV therapy in patients with coinfection must balance the risk of HIV disease progression with the potential risk of drug toxicities. In our cohort, concomitant HAART significantly improved survival in patients with a CD4 count <250 cells/mm3. Patients with a CD4 count <100 cells/mm3 had a high risk of clinical events. These data are concordant with previously published studies, suggesting the high benefit of early HAART prescription in these subjects.5-8 The benefit of HAART in patients with a high CD4 cell count at TB diagnosis has not been clearly demonstrated. In our cohort, patients with a CD4 count ≥250 cells/mm3 experienced a similar cure rate, whatever the prescription of HAART. After 12 months, none of the 16 patients remaining free of HAART experienced a new ADE or a significant decrease in CD4 cell count. Although the initiation of HAART is not systematically recommended in asymptomatic patients with a CD4 count ≥250 cells/mm3, it is recommended in case of TB coinfection, whatever the level of immunosuppression.9,13 By contrast, our data suggest that TB should not always imply HAART initiation in this particular subgroup of subjects. In our cohort, the incidence of TB treatment side effects was not influenced by concomitant HAART. In case of early initiation of HAART, however, the occurrence of side effects often led to discontinuation of TB and ARV treatment. Previous cohort studies have shown various rates of TB treatment adverse events (12% to 54%).6,8,14,15 When only adverse events leading to discontinuation of TB treatment are considered, this incidence is the same in all studies (10% to 15%).8,14,15 An interruption or change in HAART occurred in 19% of our cases. Prospective studies evaluating HAART in HIV-infected patients with TB registered HAART adverse effects in 9% to 20% of the cases.16,17 Use of new ARV agents and better knowledge of drug interactions may explain the lower rate of side effects. In our cohort, an IRIS was diagnosed in 12% of cases. Almost all cases were observed in patients with a CD4 count <250 cells/mm3 at baseline and receiving concomitant TB and ARV treatment. The estimated incidence of IRIS in TB- and HIV-coinfected patients ranges from 8% to 36%.4,6,7 In conclusion, our study confirms that the prescription of HAART to HIV-infected patients with active TB at a low baseline CD4 cell count decreases the incidence of death and ADEs without increasing the incidence of adverse events related to TB treatment. Conversely, patients with a CD4 count >250 cells/mm3 may remain free of HAART, although TB is considered to be an ADE. Agnès Meybeck, MD* Claude Fortin, MD* Raymond Ruimy, PhD† Véronique Joly, MD* Patrick Yeni, MD* *Service de Maladies Tropicales et Infectieuses Hôpital Bichat-Claude Bernard Paris, France †Service de Bactériologie Hôpital Bichat-Claude Bernard Paris, France