HIV and ID Observations

Below is a highly-subjective list of some of the highlights from the Cape Town IAS meeting. I’m sure I missed something — it’s impossible to see everything at these large conferences.  Corrections/additions welcome!

My miss-rate might be particularly high since the international AIDS meetings are appropriately focused on HIV treatment in resource-limited settings (especially Africa) whereas my perspective is as a US treater/researcher.  That said, studies from these countries often have important take-home points for all of HIV medicine.

Apologies over, on to the content:

Immediate therapy increases survival for asymptomatic patients in Haiti (WESY201). This landmark study (CIPRAHT001) randomized asymptomatic patients with CD4 cell counts between 200 and 350 to start treatment immediately with ZDV/3TC + EFV or to wait until CD4 fell to 200.  An independent DSMB stopped the study early after finding 23 deaths in the “standard” therapy arm vs only 6 in the early treatment group.  The difference in the risk of infection-related deaths was particularly striking–  1 vs17 — and the standard therapy group also required greater intensity of lab follow-up and had higher treatment-related toxicity.  The results are a strong challenge to the WHO treatment guidelines (essentially to start rx with cd4 at 200), and can be added to the growing list of studies that find early ART is better than waiting.

Incidentally, the presentation was kind of hidden since apparently the study was submitted after the late-breaker deadline.  Too bad — I hope it gets a larger venue at CROI.

Treating pregnant women with ART for at least 6 months post-partum reduces breast-feeding transmission (WELBB101). Women in Botswana with higher CD4 cell counts were randomized to either ZDV/3TC/ABC or ZDV/3TC + LPV/r, and those with lower counts got standard ZDV/3TC/NVP. The individual regimens are less important than the fact that they continued treatment for 6 months post-partum, and were instructed to breast feed (switching to formula feeding in developing countries might avert some HIV infections, but the overall outcome for the infants is worse — even when clean water can be provided).  The results showed a transmission rate of only 1%, which is the lowest ever reported for a transmission study that includes breastfeeding — and not dissimilar to what is observed in industrialized countries.

Implications?  Seems to me that the myriad byzantine regimens being tested in this setting — including various single-dose NVP strategies, giving the babies prolonged “post-exposure” prophylaxis even though they are not infected, prescribing a “tail” of NRTIs to the moms to reduce resistance with the NVP — can all be replaced just treating the mom with standard antiretroviral therapy.  Of course easier said than done, but since when shouldn’t we try to do what’s best for pregnant women, which in general is the same thing as for non-pregnant women?  Let the controversies ensue, I know they are heated.

ARTEN:  TDV/FTC + NVP is “non-inferior” to TDF/FTC + ATV/r (LBPEB07). In this 600+ randomized phase 4 open label trial, treatment-naive patients could only be enrolled if they were women with cd4 < 250, men < 400 — the thresholds used for hepatic safety according to the NVP product label.  The results between the two NVP treatment arms (once- and twice- daily) and ATV/r arm are largely similar, with a greater number of rash-related treatment discontinuations in the NVP arms, but no Stevens Johnson syndrome or drug related deaths.

Why is this important?  A few small studies of TDF/FTC (or 3TC) plus NVP had early virologic failures with high rates of emerging resistance — a finding not seen in this adequately powered study, raising questions about those earlier reports.  In addition, with resource-limited countries (hopefully) switching away from the use of thymidine analogues (in particular d4T), this TDF/FTC (or 3TC) + NVP combination will undoubtedly get extensive use.  It will be interesting to see whether NVP prescribing in the Western/Northern countries increases as a result.

GSK 572, investigational integrase inhibitor, looks promising (TUAB105). With the relative dearth of drug-development in the HIV field, it was a great to see this presentation of S/GSK1349572 (or “572” to friends), a once-daily integrase inhibitor in development by GSK/Shinogi.  A 10-day dose-finding monotherapy study showed a maximal HIV RNA reduction of 2.5 log with only a 50 mg dose, with no major adverse events, a nicely demonstrated dose-response curve (one of my colleagues called it “beautiful” — we are very nerdy, aren’t we), no selection of drug resistance, and even in vitro activity against raltegravir and elvitegravir-resistant viruses (WEPEA098).  Sure,  everything can look good this early in drug development, but this is a promising start indeed — so much so that we can perhaps forgive the GSK scientist for using the term “unprecedented” half a dozen times during her presentation — about as clear an example of the overlap between science and commercial interests/marketing as one could find!

Phase II studies about to start.

ART as Prevention Gets Top Billing (MOPL101). In one of the opening plenaries, Reuben Granich from the WHO presented his elegant model of how universal testing with immediate treatment — “test and treat”  — might end the epidemic in high prevalence settings.  (See Lancet for published paper.)  It’s a terrific talk, and the concept has some merit — especially with the accumulating data supporting the marked reduction in risk of transmission with suppressive antiretroviral therapy.  However, execution will be difficult (understatement of the year), and one model of this strategy applied to Washington D.C. (with its 3% prevalence) found that if you go through all the steps — test offered, test accepted, person tested linked to care, person linked successfully treated — the actual effect on epidemic may be limited.

Abacavir and CV disease — again (MOAB202, TUPEB175). After Dominique Costagliola opened this session by apparently retracting her CROI presentation since it didn’t adjust for IDU — at least I think this is what she did — a VA group analyzed the relationship between ABC and CV disease in their large database, and didn’t find any. Importantly, they controlled for renal disease, a factor known to be associated with CVD AND non-use of tenofovir, hence a potential “chaneller” to abacavir use which would therefore confound analyses of the relationship between ABC and heart disease.  I’m amazed (and embarrassed) I didn’t think of this before (was focusing more on typical cardiac risk factors such as smoking, HTN, etc) — could this confounder be the explanation we’ve all been waiting for?

Not so simple — it turns out that the DAD group did look at renal disease in a follow-up to their original published study (Lancet September 6 2008), and they found the ABC-heart disease relationship persisted even after adjusting for serum creatinine.

Meanwhile, a group from Montreal looked at CV outcomes going back to 1985 (a long time ago — ZDV not available until 1987), with results very similar to the DAD study.

What do we definitively know about the relationship between ABC and CV disease?  Aside from the fact that some studies show it and others don’t, not much.  (Warning:  baseball analogy to follow.)  With all this back and forth on the issue, I’m reminded of commentaries about steroid use in baseball that purport to “clear” someone’s name, only then to increase suspicion by the very mention of that person:  “I’m not saying Albert Pujols (David Ortiz, Raul Ibanez, you choose) is using steroids, but …”

ARIES — Is it safe to drop the ritonavir? (WELBB103). In the ARIES study, 515 treatment-naive patients initially received ABC/3TC + ATV/r; at week 36, once the HIV RNA was < 50, they were then randomized to continue the current regimen or drop the ritonavir and increase the dose of ATV to 400 mg qd.  Forty-eight weeks after this randomization, virologic outcomes were non-inferior (actually slightly favored the non-ritonavir arm, interestingly), and triglycerides improved somewhat with the switch as well.  The relevance of this study largely depends on the take-home message from the item above, but at the very least it gives us a nice option for “ritonophobes” after virologic suppression.

Boosted PI monotherapy again — this time with darunavir (TUAB106LB, WELBB102). Two studies — MONET (with a hard “T”, it’s Spanish) and MONOI (pronounced like “mon-oy”, with “oy” as in “oy vey”) — looked at this de-intensification strategy after virologic suppression.  Some of the differences in the two studies are important — for example, MONET used once-daily DRV/r, MONOI twice daily, and only the latter required a lead-in period actually on DRV/r before the switch.

But the take home message is similar to what we saw from OK04, the largest of the LPV/r prospective de-intensification studies:  the strategy works for most people, and indeed may be “non-inferior” to continuing the NRTIs.  However, it’s not offering much in the way of reduced toxicity, and there appears to be a higher rate of low-level viremia (50-400 cop/ml range) in those who are on the boosted PI alone.  Plus, there’s a lingering concern about CNS viral replication, as two patients in the MONOI study had detectable CSF HIV RNA, with at least one of them having symptomatic meningitis.

Bottom line:  biggest benefit is cost — predicted here by my colleague Bruce Schackman — and for now this isn’t enough to make this standard practice.

ALTAIR — the end of NRTI-only regimens (LBPEB09) Around 300 treatment-naive patients were randomized to receive TDF/FTC + EFV, or ATV/r, or the NRTI-only regimen of TDF/FTC + ABC + ZDV.  (That’s 4 nukes, for those keeping score at home.)  At this point, it’s probably of little surprise that the NRTI-only regimen was inferior (though through an interesting statistical quirk, was “non-inferior” in the primary endpoint– don’t ask).  It was worse from the perspective of virologic failure, metabolics, tolerability, and (almost) serious adverse effects.

Are we done with this NRTI-only strategy yet?  I think so.