Comparative effectiveness of biologic agents for the treatment of psoriasis

Comparative effectiveness of biologic agents for the treatment of psoriasis


Good day, my name is Bruce Strober. I’m Professor and Chair of Dermatology at
the University of Connecticut Health Center. Today I’m going to discuss a paper that came
out in May of 2016 in the Journal of the American Academy of Dermatology that, basically, evaluated
comparative effectiveness of biologic therapy… in a real-world setting though. So, this was a study utilizing a well-known
registry called PSOLAR. PSOLAR is a registry, to be clear, sponsored
by Janssen, and Janssen is the manufacturer of the medication, ustekinumab. So that’s an important caveat. That said, we should discuss this study as
objectively as possible. It was a comparison in the registry (and the
registry is 12,000 patients) between patients on ustekinumab and those on adalimumab, infliximab,
and etanercept; so four different patient populations, with a variety of numbers of
patients in each population. This was real world; these weren’t patients
randomized in a blinded manner. Patients were allocated to the drug of interest
based on clinical decision in a clinic setting and then evaluated within this registry. Now, there were some rules. The rules were: the patient had to have the
drug evaluated as their first biologic they started within the registry—it couldn’t
be the second or third biologic. So if they were on adalimumab and were called
an adalimumab patient, it had to be the first drug they used while on registry. That’s not to say they couldn’t have had a
biologic previously in their life experience. In fact, most did. In fact, half of the adalimumab patients evaluated
were on etanercept previously. Many of the ustekinumab patients were on previous
biologics, so we’re looking at patients who are somewhat mature in their experiences as
psoriasis patients and capturing them in the midst of their life experience having psoriasis
and being treated. So that of course is a complex situation and
does, in my opinion, make this kind of data very different than what you would see in
a clinical trials population—not a pure study. But nevertheless, what we want to see in addition
to clinical trials. Clinical trials are great—very high control,
very clear allocation, really prescripted allocation of patients to various buckets. Here we’re just kind of throwing people in
buckets based on whatever: “I think this patient belongs on ustekinumab for reasons
x, y, and z, and no one’s going to question me, and now that patient’s going to be analyzed
in the registry.” Now, the outcome measures in this study occurred
at two points in time: 6 and 12 months after the patient started the medication of interest,
and the patients were evaluated in really three different ways at 6 and 12 months. Number one: how were they doing on the Dermatology
Life Quality Index —the DLQI, a quality-of-life metric—at 6 and 12 months compared to their
baseline score when they were started on the drug? Also how were they evaluated in percent body
surface area (BSA) relative to baseline? How much did their BSA go down? And finally, Physician’s Global Assessment
of 0 or 1 (PGA 0/1), which is the commonly used metric of success, objective disease
reduction. What we were able to do was analyze about
2000 patients, 2000 out of the 12,000 who were ultimately enrolled in the registry because
there were criteria. You couldn’t be just any patient and be analyzed. It had to be a patient that was followed for
12 months, that their first biologic therapy was the drug of interest, so it narrowed down
the population. And the upshot of the data are in essence,
ustekinumab patients on most metrics did better. For example, with regard to Physician’s
Global Assessment, more patients were PGA 0/1 at 6 months than the other drugs; and
at 12 months, while the PGA differences narrowed between ustekinumab and the other drugs, numerical
superiority was still seen with ustekinumab. Similar statements could be made with body
surface area reduction, so not just the quality of the plaques themselves, which is a PGA
(scale, erythema, induration), but also BSA of involvement—sort of the diffuseness of
the disease—also was better reduced by ustekinumab at 6 and 12 months relative to adalimumab,
etanercept and infliximab, relatively speaking, and statistically better at 6 months. And finally, DLQI really was measurably better
at 6 months. Things sort of evened out at 12 months between
the various drugs. So, all told, ustekinumab, in a real-world
setting, functioned a little better than the other three drugs. The question is, is this real? I will let the reader who uses all these drugs
weigh in, because your experience obviously matters as much as anyone else if you use
a lot of these drugs. But I’ll tell you this: a lot of this is based
on the qualities of the drugs themselves. For example, ustekinumab is given in office
often at an infrequent frequency every 3 months. For that reason alone, better adherence will
be seen with ustekinumab than the other drugs, which are self-injectable, except infliximab,
which is an infusible drug. I kind of want to throw out the infliximab
data—so few patients were in that arm, it’s hard to do the statistics correctly. So three of the drugs (ustekinumab, adalimumab,
and etanercept) are more relevant and they’re more commonly used drugs anyway, so those
comparisons have greater meaning for the audience. But, that said, self-injectable drugs like
adalimumab and etanercept have lower adherence and, for that reason, the real-world efficacy
might truly be lower on that factor alone. That said, dosing variation occurs with all
these drugs. For example, a large percentage of these patients
were not taking etanercept once weekly after the third month, as the label would suggest
they should do in the U.S. Many of them were staying on twice weekly,
so etanercept was given a good chance to succeed based on higher dose. Most of the patients, about 80%, were taking
adalimumab every other week, so it was a fair assessment of adalimumab. In ustekinumab, there were a lot of patients
getting the drug every 8 weeks and not every 12 weeks, and I would bet you a lot of patients
are getting the drug at 90 mg even though their weight would dictate they should be
getting it at 45 mg. So there you see an instance where in real-world
we can kind of cheat “quote-unquote”. We can kind of give drugs in a way that they
weren’t studied in the clinical trials and, therefore, augment their efficacy in the real
world. Now, none of this disqualifies the analysis,
because the analysis is supposed to be real world. The way the drugs are used is the way we should
really be thinking about them with regard to how well they work. We have to remember that each patient is different,
and it was discovered in this analysis that patients with higher weight and higher body
mass index were less likely to do well—that’s not surprising. Disease severity at baseline—if more severe,
the PGA was less likely to improve. All of this was multivariate analysis, so
we tried to control for these issues, but I just want you to keep in mind you may choose
drugs variably based on body weight, disease severity, and other issues such as the presence
of psoriatic arthritis. In my opinion, the presence of psoriatic arthritis,
when really clinically apparent, mandates a TNF inhibitor first. And I will still say that even with the advent
of data that show ustekinumab and IL-17 blockers (the new drugs) work for psoriatic arthritis—I’m
still going to stick with TNF blockers for the time being. And then the final issue is a lot of these
drugs, as I mentioned, were not the first drug the patient had experienced in his or
her lifetime. For example, a lot of the adalimumab patients
had received etanercept in the past, and I would guess a lot of these were etanercept
failures; and an etanercept failure is more likely to be an adalimumab failure. So you’re looking at a population of adalimumab
users who were somewhat set up to fail based on their prior treatment experience, and that
can bias the efficacy data downward for that group of patients. So we’ve got to always caveat real-world
studies while valuing them and, importantly, we need to reinforce data of this nature with
independent analyses from other data sources. Thank you very much.

Leave a Reply

Your email address will not be published. Required fields are marked *