So the COSMOS study involves a newer drug in psoriatic arthritis called guselkumab which is one of the IL-23 inhibitors. And we've seen a very strong data for IL-23 inhibitors in psoriasis patients, and that's translated into a license and a lot of use for IL-23 inhibitors in dermatology, but they're still relatively new in rheumatology. And there have been two large phase III trials which resulted in the approval of guselkumab for psoriatic arthritis, but quite a small percentage of the people included in those trials had previously failed a biologic.
And obviously, while it's important to know how well drugs work in biologic-naive patients, it's also really important to know how well they work in people who've already failed a different biologic, because that's an increasing number of patients that we have in our practice. So this study aimed to specifically look at the efficacy of guselkumab in patients with psoriatic arthritis who had previously been treated with a biologic therapy-- so a TNF inhibitor-- and that had an inadequate response either to one or two TNF inhibitors before they joined this trial of guselkumab.
The COSMOS trial is a relatively straightforward randomized-controlled trial, really. So patients were randomized two to one to guselkumab treatments-- 100 milligrams initially every four weeks then every eight weeks-- or to placebo for the first 24 weeks of the trial. And so, we then looked at efficacy at week 24 for those treated with guselkumab versus those treated with placebo.
Obviously, 24 weeks is quite a long time for a placebo arm of a trial, so if patients had no response at all, or a very poor response, at week 16, that was an early escape arm where patients from the placebo group could be switched over to treatment with guselkumab slightly earlier than that at week 16.
The main finding and the primary outcome was ACR-20 and there was a significantly better outcome for guselkumab compared to placebo at week 24 for ACR-20 and also for many of the key secondary outcomes that you would expect. So, ACR-50 and 70 response in terms of function as measured by the hack, and health-related quality of life using SF-36.
And then, because we know this data is particularly strong in skin psoriasis, it's perhaps not surprising that there was a significant benefit for guselkumab compared to placebo in terms of PASI responses-- so PASI 90 and PASI 100-- over the course of the trial.
So I think these response rates are actually pretty similar. Obviously, whenever you have patients who've previously failed a medication, they tend to represent a slightly more resistant group of arthritis patients and often the response rates are very slightly lower. But actually, the response rates here were very good despite the fact that these patients had previously failed a TNF inhibitor. So out at week 24, we saw around 45% of people achieving an ACR-20, which is not too different from the previous phase III trials.
So I think it's great to have additional options for treatments of psoriatic arthritis, and the IL-23 inhibitors are kind of the new kid on the block in terms of treatment for PSA. We know that they're incredibly strong in efficacy in skin disease and there are a number of head-to-head trials in the psoriasis literature showing how good IL-23 inhibitors are, and so it's really useful to have data on musculoskeletal manifestations. So, joints, but also enthesitis and dactylitis as well, to guide us when we're choosing therapy for patients.
So, I think these drugs show good efficacy across all of the peripheral musculoskeletal domains, in addition to the skin. They're obviously particularly strong for skin where they're superior to a number of the other biologics, but show good efficacy in peripheral joints as well. So, I think these represent another good option for treatment of people with psoriatic arthritis, particularly those with more severe skin disease.
I think the biggest question for guselkumab now moving forward is going to be around axial disease. So, we've seen evidence of very strong efficacy in skin disease, of good efficacy in peripheral joints and enthesitis and dactylitis, but we've seen negative trials in ankylosing spondylitis with IL-23 inhibitors. And yet, we've seen good improvement in BASDAI scores in the patients in the PSA trials who were found to have actual disease.
So there's a suggestion from the BASDAI score improvement that these may be showing some efficacy in the spinal aspect of the psoriatic arthritis, but we've got a negative trial in ankylosing spondylitis and we know that there's a lot of commonality between the different forms of spondyloarthritis.
So, I think what we really need to moving forward is whether this is efficacious for axial disease. It's still a good drug for psoriatic arthritis and for skin disease, regardless of the results of that, but obviously, it's important for us to know across all the different domains of psoriatic arthritis. So, what we really need is a specific axial PSA trial with imaging evidence of spinal response so that we have an objective measure of improvement in spinal disease, because we know that all of the questionnaires are often impacted on by improvement in other domains.