1. Home>
2. Specialties>
3. HIV/AIDS Clinical Care>
4. Meeting Report

When Should ART Be Initiated in Patients with OIs?

Data presented at the 2009 Retrovirus Conference bring us closer to a definitive answer.

The most beneficial time to initiate antiretroviral therapy (ART) in patients with opportunistic infections (OIs) is still debated. This issue arises frequently in the U.S. because HIV infection is often not identified until a first OI arises, but it has even greater relative importance in resource-limited settings, where late presentation with tuberculosis or cryptococcal meningitis is common. Now, data on this topic are available from several randomized, controlled trials (including SAPiT, which is described elsewhere in this report).

In Zimbabwe, Azure Makadzange and colleagues randomized 54 HIV-infected patients with newly diagnosed cryptococcal meningitis to either start nevirapine/d4T/3TC within 72 hours or delay ART for 10 weeks [Abstract 36cLB]. All patients immediately received oral fluconazole (800 mg daily) for 10 weeks. During 2 years of follow-up, 62% of the patients died. Surprisingly, the early-ART group had a significantly higher mortality rate than did the delayed-ART arm (82% vs. 37%) and also a significantly shorter median duration of survival (5 weeks vs. 39 weeks). Notably, in both groups, deaths occurred early, within the first 4 weeks of treatment, which suggests that immune reconstitution inflammatory syndrome (IRIS) might have been responsible. This hypothesis is logical, given that an increased inflammatory response within the rigid cranium could result in increased intracranial pressure and herniation. However, data collection was incomplete, cerebrospinal fluid pressures were not measured, and no specific treatment was used when IRIS was suspected. Although the results are not fully explained, the study does raise concerns about initiating ART early in patients with cryptococcal meningitis.

Some of the concern about the dangers of IRIS with early ART might be allayed if treatment with anti-inflammatory agents could be proven beneficial. Graeme Meintjes and colleagues began to provide such proof with a randomized, controlled trial in South Africa [Abstract 34]. A total of 110 patients being treated for both pulmonary TB and HIV infection who met the recently established case definition for IRIS were randomized in a double-blind fashion to receive either prednisone (1.5 mg/kg for 2 weeks, followed by 0.75 mg/kg for 2 weeks) or placebo. During 12 weeks of follow-up, the prednisone group had significantly fewer cumulative days of hospitalization and outpatient procedures (the combined endpoint) than did the placebo group. In addition, symptom, radiographic, and Karnofsky scores all improved significantly in the prednisone group at weeks 2 and 4, compared with the placebo group. Importantly, the two groups had similar rates of death, of other infections, and of events that could be considered prednisone side effects. Overall, these results suggest that prednisone treatment of IRIS among patients with TB is safe and reduces the need for medical interventions. These findings, along with those from SAPiT [Abdool Karim S et al. Abstract 36a] and ACTG 5164 [Grant P et al. Abstract 775], should encourage earlier initiation of ART in patients with TB and perhaps other OIs, with the possible exception of cryptococcal meningitis.

— Gerald H. Friedland, MD

Published in AIDS Clinical Care March 9, 2009