HIV and ID Observations

Tons of interesting stuff at this year’s combined ICAAC/IDSA meeting, most of it in non-HIV related Infectious Diseases.  In aggregate, literally hundreds of posters, presentations, and symposia on MRSA, C diff, osteomyelitis, complicated UTIs, hospital-acquired pneumonia, antibiotic resistance … It’s a great meeting to catch up on general ID, and the literature review sessions alone are practically worth the price of admission.

But there are almost always a few major HIV-related studies presented as well, and this year was no exception.  These two understandably got the most attention:

1. Early antiretroviral therapy increases survival [Kitahata H-896b].  The NA-ACCORD study compared all-cause mortality among more than 8,000 patients with HIV followed since 1996.  Compared with those starting therapy with a CD4 cell count between 200 and 350, patients starting with CD4s between 350 and 500 had a 70% reduction in the risk of death.  Wow.  [Addendum:  Our astute editor at AIDS Clinical Care points out that this is really a 43% reduction, as the relative hazard for mortality for the group deferring is 1.7.  Still wow.]  Usual caveats on the limitations of cohort studies apply, but this was a very well done study with a huge sample size; its conclusions were further bolstered by the observation that virologic suppression rates did not differ between the two groups — implying comparable levels of medication adherence.  Stay tuned for a similar analysis of those starting with a CD4 > 500 cells — whispers that we’ll see this info at next year’s CROI in Montreal.   Will this be the final word on the “when to start treatment” question we’ve been debating now for two decades?  Should pretty much every patients with HIV be on therapy?  What will happen to the planned “START” randomized trial?  I sense unless something truly unanticipated happens with drug toxicity, we’re going to be starting a lot of asymptomatic patients on treatment over the next few years.
2. Raltegravir vs efavirenz as initial therapy [Lennox H-896a].  Can something be as good as efavirenz (essentially our current gold standard)  for initial therapy?  Apparently yes — raltegravir was “non-inferior*” to efavirenz when combined with TDF/FTC in a large phase III, double-blind study:  Virologic suppression rates at 48 weeks were 86% for RAL compared to 82% for EFV.  There were also lower rates of drug-related averse events in the raltegravir arm, with protocol-specified CNS toxicity occurring in 18% of efavirenz and 10% of raltegravir-treated subjects.  We’ve been using the 2NRTI + NNRTI or PI  approach as initial therapy for a long time — here at last is a new approach. 

And for the record, I’m a big fan of this combined ICAAC/IDSA meeting, but it will be separate for at least the next two years at least.  Oh well.

*The use of the term “non-inferior” always sounds like pedantic statistics-ese to me.  86% “non-inferior” to 82%?  But it does mean something specific — best described to me as “not too much worse than”, with “too much worse” generally defined as within 10-12%.  But the lower-limit of the 95% confidence interval can’t be below this 10-12%, and that’s where peculiar statements such as “86% is noninferior to 82%” come from.