HIV and ID Observations

From NIAID, an important clinical trial has been stopped early:

The Phase IV study … was evaluating whether HIV-infected participants hospitalized with cryptococcal meningitis (CM) but not yet taking antiretroviral therapy (ART) would improve their chances of survival if they began ART while receiving CM treatment as inpatients compared with the standard practice of beginning ART as outpatients, approximately five weeks after receiving CM treatment. Two reviews of the COAT trial’s safety and effectiveness data last month by an independent data and safety monitoring board (DSMB) found substantially higher mortality rates among the 87 participants who received early ART compared with the 87 participants who received delayed HIV treatment.

More details on the study design can be found here, and we’ll undoubtedly hear and read additional details on the results in an upcoming meeting and when the paper is published. But since it was stopped after only 174 (out of a planned 500) enrolled, the difference between the two strategies must have been really huge. [See edit below.]

Until then, we can postulate that ART-induced immune response — IRIS — triggered inflammation, which could have worsened intracranial pressure and other manifestations of cryptococcal disease. Inflammation is just bad news in meningitis. This could explain why crypto and perhaps TB meningitis seem to be the exceptions to the rule that early ART is beneficial with acute OIs — a strategy confirmed now in several clinical trials (ACTG 5164 and these three TB studies: SAPIT, STRIDE, and CAMELIA).

Importantly, the study confirms the results of this other cryptococcal meningitis study (which amazingly was led by one of our ID fellows before she started fellowship — how’s that for impressive?), but differs in at least two important ways: First, initial therapy was with the standard-of-care amphotericin B and not fluconazole; and second, the ART was efavirenz- not nevirapine-based. Both of these address at least some of the concerns about the earlier study’s generalizability.  A5164 did not find that early ART was harmful in crypto meningitis; however, the number of study subjects with cryptococcosis was relatively small (35 out of 282).

Though the COAT trial was conducted in Uganda and South Africa, the results have immediate applicability to practice world-wide. Early ART in cryptococcal meningitis is clearly harmful, and clinicians should delay starting ART for at least 5-6 weeks in their patients with the disease.

And the study reminds us more broadly that critical studies on the optimal management of HIV-related OIs will necessarily need to come from outside the United States and Western Europe.

[Edit:  David Boulware, the PI of the COAT trial, kindly provided a bit more data on the study results. He noted that the “absolute difference in 6-month survival is approx. 15% between the two arms (with ongoing follow up still occurring through October 2012)”, which is perhaps not as big an effect as I had anticipated. However, the message remains the same – early ART after cryptococcal meningitis decreases survival, and should be avoided. He also noted that linkage to outpatient HIV care remains a very important consideration for delayed ART.]