MRK cirrhotic data and more thoughts on the C-EDGE data set
I tallied all the 12 week data in the cirrhotic arms of the C-EDGE program and you get a very respectable 94%. For TN (mono and co infected combined) it was 98% (103/105), for TE it was 89% in both the rib free and rib containing arm, but adding ribavirin did give a boost in the 16 week TE arms (92 and 100% w/o and w rib), but each arm only had 37 -38 patients so it's not the most robust data set when you chop it up - still curious that ribavirin did nothing in the 12 week arm yet it did in the 16 week arm, although in the ph 2 study presence of ribavirin in general didn't seem to add much (and in fact had lower SVR in some arms of the study). So I don't think it's quite clear how ribavirin plays with grazoprevir/elbisvir. Why the drug combo preforms better in cirrhotic naives versus non-cirrhotics is the larger puzzle to me. It could be that there were more cirrhotics that had GT1b or low viral loads, but i doubt that is the explanation (again if anyone can share the text of the paper woudl be appreciated to know the breakdowns). It could also be that the numbers were too small with large confidence intervals when you look at subgroups and this could be statistical noise. Or there could be something unique going on with this combo compared to what we saw from GILD and ABBV. The first hint that merck's dual therapy behaves a bit differently was from the significant difference in performance based on viral load (something we didn't see with GILD and ABBV with the exception of the much higher 6 million per mL cutoff in the 8 week GILD arm). The only way i can rationalize it biologically is that cirrhotics in theory may have less virus because they have less functional liver and virions are concentrated in hepatocytes. so the very good performance in patients with lower circulating VL may tie in to the apparently good performance in cirrhotics based on this common thread, which may have compensated and then some (at least in naives) for whatever is the cause of the lower historical SVR in cirrhotics.
I think MRK would have been better served performing a more robust ph 3. the numbers for TN for example were half that of ABBV's program. I think the relatively shitty results in GT1a naives, especially the majority who have >2M copies of virus/mL, the better than expected results in TN cirrhotics, etc. all has to be examined in larger patient numbers to really get a good sense of how and where the combo should be used. If i am a GT1a non-cirrhotic pt i want to stay as far away from thsi combo as possible unless I am one of the minority with a low VL - a 5-7% delta in SVR is unacceptable since surely retreatment with an alternative regimen will then be challenging (every combo has at least a PI or NS5A component so resistance to both is a real bitch to have at baseline). If i am a cirrhotic maybe it's the treatment of choice, but hard to make that decision based on 105 patients (even combining the mono and coinfected population). I just think you need more data. The question is will the FDA want more data before they give a blanket approval in GT1 (versus say getting a niche label for renal patients). I think there's a good reason to ask for it
