A new take on developing drugs for chronic pain
Chronic pain is a tough area for drug development — the preclinical models are lousy, the endpoints are messy, and the “placebo effect” is a nuisance. So even if Vertex had done everything right, it’s not entirely surprising their NaV1.8 inhibitor suzetrigine (VX-548) whiffed in a phase 2 study in sciatica in late 2024.
Instead of rehashing Vertex’s results, though, I want to discuss data on another company’s chronic pain candidate that were released a week before Vertex’s.
Like suzetrigine, Lilly’s TRP1A antagonist (LY3526318) also failed in a phase 2 trial in chronic pain — but the way Lilly designed, analyzed, and acted upon the results of their study shows a possible path forward for successfully making new medicines in this indication, and other companies (including Vertex) should take notice.
So let’s dig in, and see what we can learn about how to — and not to — develop a chronic pain drug.
* * *
First, some background. Lilly’s pain trial was run under a “chronic pain master protocol” (CPMP) they developed with input from FDA under its Complex Innovative Trial Design Meeting Program. The general term “master protocols” encompasses three types of studies — umbrella, basket, and platform trials — and Lilly’s CPMP used all three elements. (This short review in NEJM by FDA’s Janet Woodcock and Lisa LaVange is a great explainer with super illustrations; highly recommended.)
The “umbrella” part is that Lilly is testing a bunch of drugs in parallel with different mechanisms. All the agents share the same trial logistics (sites, recruitment, etc.), and they’re all compared to a shared control arm.
The “basket” aspect is that Lilly is testing drugs in several related indications (more on this below) under the same overarching protocol.
The “platform” aspect is that Lilly can add and remove trial arms over time as new agents become Phase 2-ready and ones in the trial finish generating data.
Several benefits of master protocol trials are fairly general across disease areas. For example, they’re more efficient operationally (in terms of site vetting, set up, IRB approval, etc.). They also reduce the total number of patients needed because of the shared control arm. And also because of the shared control arm, they may be more attractive for prospective patients, because the odds of receiving an experimental drug (vs. placebo) are higher.
These sorts of trials aren’t new, but they seem to be gaining popularity as companies get more aware of their benefits and more experienced with designing and running them, and the FDA continues to signal its enthusiasm. Some of the best-known examples in biopharma are I-SPY 2 (neoadjuvant breast cancer) and HEALEY ALS. More recently, an academic team kicked off a master protocol trial for progressive supranuclear palsy, and Sanofi initiated one for two types of kidney diseases, focal segmental glomerulosclerosis and minimal change disease. I think master protocol studies could be particularly useful in rare diseases, and wrote a whitepaper on this topic (see my contribution (#7) to “The Case for Clinical Trial Abundance” by the Institute for Progress).
* * *
A master protocol study has specific benefits in chronic pain, particularly for a firm like Lilly that’s committed to this area:
First is the “umbrella” aspect of testing many drugs in parallel. In chronic pain R&D there’s no shortage of biological hypotheses, but the preclinical models for target validation are, to put it mildly, terrible. That means if you want to succeed, you need to efficiently run many phase 2a “proof of concept trials” that can give you a good sense of whether it’s worth moving into larger studies. Thus, the master protocol is great fit for a company like Lilly that is committed to developing drugs in this area and has many plausible ideas, because they can test a medley of mechanisms in the same trial.
Second is the ability to interrogate a “basket” of related indications. There are several different clinical scenarios one might want to test in chronic pain, and it’s not obvious which drug will work best in which of them. Lilly’s CPMP tests each drug in knee osteoarthritis, diabetic peripheral neuropathy, and chronic low back pain, which reflect different underlying pain pathologies (nociceptive, neuropathic, and mixed). Instead of taking expensive rifle-shots on specific chronic pain subtypes, the master protocol study lets Lilly hedge its bets and let the results speak for themselves on whether a particular target is effective in a particular setting.
And finally, there’s the “platform”. Lilly appears to be devoted to chronic pain R&D for the long-term and is exploring many different mechanisms. But there’s almost no chance all of the potential therapies would be phase 2-ready at the same time. Under the master protocol, once a new molecule makes it through phase 1, Lilly can add an arm to test it under the CPMP. And when testing of a drug candidate in the CPMP is complete, Lilly can seamlessly remove that arm without affecting the rest of the study.
* * *
Separate from the master protocol, another notable and important feature of Lilly’s study is that, as their TRP1A antagonist experience demonstrates, they are using it appropriately to kill ineffective drugs.
A big reason to run a phase 2 study is to help you avoid wasting money on a costly phase 3 program that has little to no chance of succeeding, so you need to have a clear idea of what results would lead you to continue development versus killing the program. That clarity is even more important in an area like chronic pain, where the phase 3 failure rate is so high and it’s often hard to distinguish a real signal from the so-called “placebo effect”.
Lilly’s CPMP trial is designed the right way to answer the experiment: it randomizes patients to drug vs. placebo and compares their final pain scores head-to-head using a pre-set threshold for defining success. When Lilly found that the post-treatment effect of their TRP1A antagonist was no better than that of a sugar pill, they didn’t beat around the bush and look for ways to spin the data (as many shady biotechs would’ve done). Instead, they transparently reported the drug’s failure and abandoned it in this indication, while continuing to test other agents in the study and work on advancing earlier-stage chronic pain compounds toward Phase 2.
Vertex, in contrast, made different choices. Instead of directly comparing post-treatment patients on suzetrigine to those on placebo, they opted to use a comparison of pre- vs. post-treatment pain scores of patients on their experimental drug as their main outcome. This sort of “pre/post” analysis is common, but many statisticians have pointed out that results from within-arm comparisons like this are virtually impossible to interpret, and given that the study had a control arm, it’s puzzling why they would have chosen to analyze the data this way. Although Vertex’s claim that suzetrigine “met the primary endpoint” is technically true, a quick look at the placebo-treated patients shows the drug probably had no effect on chronic pain reduction. (Apparently that’s not stopping Vertex from advancing suzetrigine into phase 3 in this indication, however.)
* * *
Chronic pain is a particularly tough area for drug R&D, plagued by apparently promising targets that fall over in pivotal trials. But if anyone’s going to crack it, I’d put my money on a company with an approach like Lilly’s. Their CPMP study uses the key advantages of umbrella, basket, and platform trials to efficiently and robustly test whether drug candidates work in this indication, and Lilly seems to be using stringent pre-set criteria to ruthlessly cull the ones that don’t make the grade. In that context, the failure of Lilly’s TRP1A antagonist was a “good failure,” because it accomplished its objective: Lilly will now avoid spending more money on an ineffective therapy, and redirect those resources toward more promising ones. This is a great example of the science-driven decision making it’s going to take to bring new chronic pain drugs to market.
Photo by Pietro Jeng via Pexels.