Let’s talk Sarepta. And FDA approval, because you can’t bring up that company without immediately starting a regulatory affairs argument. I was not happy when their initial exon-skipping therapy (Exondys, eteplirsen) for Duchenne muscular dystrophy was approved in 2016, because I thought that the efficacy data were simply not strong enough for such approval. Last summer, another exon-skipping application from the company (golodirsen) was rejected by the FDA with a Complete Response Letter (CRL). The agency does not make those letters public at the time, despite many people wanting this to happen, so we’re dependent on whatever the receiving company wants to say about them. At the time, Sarepta’s management said that the rejection was due to concerns about kidney tox and possible risk of infection.
But last month, the FDA suddenly approved golodirsen (known as Vyondys 53). This took everyone by surprise, and there has been no explanation for the change of opinion on the agency’s part. As the Wikipedia article linked above correctly states, “No clinical benefit of using golodirsen has been established”, and honestly, I didn’t think we approved too many drugs without establishing that. But here we are.
The story gets even more complicated as of today, because the FDA has actually posted the CRL for all to see. It’s quite the document. It points out that the surrogate marker the company was using (truncated dystrophin) changed by a very small amount, which would lead one to think that the clinical benefit would be minimal. That’s even if you accept that the marker is a valid one, which is an arguable thesis all by itself. There was, as noted, no correlation between the physical decline of the subjects in the study and the levels of the truncated protein.
The golodirsen package is unavoidably tangled up with the previous eteplirsen story. The CRL details a number of serious infection events (sepsis, death) associated with eteplirsen since its approval. Duchenne patients are at risk for this sort of thing, since they’re taking high doses of corticosteroids. The letter notes furthermore that the golodirsen package manages not to mention any of these despite being a very similar oligonucleotide which will be administered in the same manner:
Some may characterize these infections as device-related rather than drug-related and downplay their importance. This is a weak argument: these individuals would have had no reason for implantation of a central intravenous infusion port other than to receive eteplirsen; many of these infections would not have occurred if the patients had not been receiving eteplirsen.
If these devices are necessary to deliver the drug, then these infections must be construed to be drug-related. Although the aforementioned information was obtained from the eteplirsen postmarketing experience, such information is directly applicable to the potential marketing of golodirsen, which too would be administered through implanted infusion ports in many patients.
As for the renal toxicity, it’s a concern with antisense oligos in general, and it appears that golodirsen just barely make it through preclinical tox studies because of this very issue. It’s excreted unchanged, so any decline in renal function is going to increase exposure in the kidney and make things worse. Which will increase exposure even more, which will. . .you get the idea. The potential for rapid (and likely fatal) kidney tox is always going to be present. And there is no accepted way to monitor renal function in Duchenne patients – the reduced muscle mass makes creatine levels an inappropriate marker.
Now we come to that whole “eteplirsen postmarketing experience” part. The letter points out that no severe infections were seen during eteplirsen’s clinical trials, but it has been a real problem since drug approval. Golodirsen should be in the same situation for that risk, and its kidney tox risk is clearly greater when you compare the two sets of data. So what did that mean for its proposed approval?
When eteplirsen was granted accelerated approval in 2016, you were required to perform a postmarketing study to verify eteplirsen’s clinical benefit. Patients were to be randomized to the approved dosage of eteplirsen (30 mg/kg weekly) or to a dose 7-fold higher, e.g., 30 mg/kg daily. We thought that the considerable increase in dose was acceptable because we were not aware of any significant toxicity related to eteplirsen. The same cannot be said of golodirsen, where nonclinical data have demonstrated considerable risk of renal toxicity.
In summary, therefore, at the time of eteplirsen’s approval, there were no adverse events of note in the clinical data, and no nonclinical concerns. Although we recognized that the clinical database was limited, the drug had no known toxicities, which enabled us to require a confirmatory trial to study a dose that was 7-fold higher than the to-be-marketed dose. The risk of serious and life-threatening infections, including sepsis, was not known in 2016 when eteplirsen gained accelerated approval. In essence, the small potential benefit of eteplirsen was weighed against essentially zero risk.
Here’s the problem: that postmarketing study has never taken place. In fact, it hasn’t even been started. (Edit: looks like they’ve finally gotten it off the ground since the CRL was issued). The CRL says that because of this failure, we know nothing more now than we did three years ago about whether eteplirsen does any good at all. What we do know, though, is that it’s much riskier than the agency thought, and we also know that golodirsen is even riskier (and with, as mentioned, extremely thin evidence of any benefit). The letter ends with a long, detailed list of what the FDA expected Sarepta to do to remedy the deficiencies in the drug’s approval package.
And then a few months later, they said “Ah, never mind” and just approved it anyway. No explanation (edit: unless you count this response, which I do not find convincing). So those are the drugs that Sarepta has on the market: an initial therapy that was not shown to work in the clinic and has shown significant risks since it was brought to market. And a follow-up which likewise has not been shown to work in the clinic and can already be expected to have ever greater risks once marketed. What the hell is going on? The idea is that the FDA evaluates things on the basis of safety and efficacy, but it can apparently throw both of those out the window when it feels like it? This is infuriating, and it damages things in all directions: not only the health of suffering Duchenne patients, but also the integrity of the whole regulatory process and the reputation of the FDA. You don’t just get those things back easily, you know. That’s what we’re going to be learning for years to come in our current situation, and I’m not just talking about drug approvals. Fairness, competence, impartiality, integrity, and the reputation for having them: they don’t just reappear when you decide later on that you’d rather have them back.
Edit: I meant to add more about the response letter from Peter Stein of the FDA linked above. It requests that Sarepta continue with a confirmatory study for golodirsen, which should read out in 2023, and that the company commit to withdrawing the drug from the market should the results of that study not support its use. In the abstract, I support that idea. But why didn’t the agency hold the company to that sort of agreement with eteplirsen? Why approve yet another therapy in this class, with similarly weak data? Will the FDA commit to re-evaluating its recent tendency to overrule its own scientists if these drugs injure more Duchenne patients in the next three years without doing them any good?