This weekend I did a mini dive into BDSI topical clonidine
program. This is not going to be an extensive post (I still have some questions
in mind), but since results of a Phase 3 trial are supposed to be announced any time soon, I
figured it would be nice to share some thoughts.
BDSI is testing the analgesic effects of a topical clonidine
gel in patients affected by painful diabetic neuropathy. Clonidine is an
alpha2-adrenergic receptor agonist known for its analgesic effects when
administered systemically. Alpha2 receptors are present also on peripheral
nociceptive fibers. Activation of these receptors is known to induce analgesia
in rodent models (Dogrul 2004).
BDSI licensed from Arcion Therapeutics a gel formulation of clonidine for
topical use.
What do we know about clonidine topical use in patients?
There are a couple of older trials in which clonidine was administered as a
transdermal patch (Zeigler 1992;
Byas-Smith 1995).
These trials did not show consistently significant effects of the drug in
reducing pain. However, this is
actually positive for BDSI. In fact, transdermal application was not used to
affect pain locally but as a way to systemically administer the drug. In other
words, the patch was not positioned on the region affected by pain but away
from it. The only possible effect for this patch might have been via an increase
in the blood concentration of clonidine, and likely the concentration was too
low to be too effective centrally.
There are only a handful of papers (a not too quick search
turned two papers plus an abstract) on topical application of clonidine in patients
and they seem encouraging. Mind you, they are all from the Hopkins group that performed
the studies that lead to the clonidine gel. One paper (Davis 1991)
is a small proof of concept (few patients) and shows the potential efficacy of
topical clonidine in sympathetically maintained pain (a form of chronic pain).
The second paper presents the Phase 2 trial results for the
drug currently in Phase 3 (Campbell 2012).
The trial featured 179 patients suffering from painful diabetic neuropathy, 89
in the treatment group and 90 placebo (placebo gel). The treatment was a gel of
clonidine 0.1%. Notably, the authors explain in the methods the rationale for
the dose; previous unpublished data showed that 0.05% was not effective, and 0.2%
was not more effective than 0.1%. As a matter of fact, in an abstract presented
few years earlier (Campbell 2009) 0.2% appeared less effective than 0.1%. While it is always unpleasant not to
see dose-response, the weaker effect of 0.2% might be due to the lower volume of
gel, or, in my opinion, to a stronger tolerance for clonidine. Indeed, tolerance
development (i.e. progressive loss of effectiveness) is a well-known issue with
clonidine, and occurs also with topical administration (Dogrul 2004). After a baseline period, patients entered 12 weeks of treatment. Each day
they had to self-administer the gel (three times/day morning, afternoon and
evening) and self-report the subjective perception of pain using the numerical
pain rating scale (NPRS) once a day. The treatment lasted 12 weeks, and each
week patients were assessed at the clinic and rated on a Brief Pain Inventory
(BPI) scale. A subset of patients underwent a biopsy in the affected region to
quantify nerve fiber density. In addition, all the patients underwent a capsaicin
test. 0.1% capsaicin was applied over a 1 cm diameter area on the affected zone.
Patients were asked to rate the pain evoked by capsaicin. Biopsy and capsaicin
test were used to assess the presence and the function (or hypersensitization)
of nociceptive fibers. I am still trying to figure out the consensus on the
interpretation of this test, but it seems rational to me.
When the entire population was analyzed, no significant
difference (p = 0.07) was seen between clonidine gel and placebo gel. However,
a subgroup analysis based on the capsaicin test revealed a more positive
picture. Indeed, when only patients with a response to capsaicin were analyzed
(i.e., supposedly those with less nerve damage), the treatment appeared
significant. For the group showing capsaicin pain equal or greater than 2, the significance
reached p < 0.01. In other words, the
idea is that only patients with a less extensive damage to nociceptive fibers
are likely to benefit from the topical treatment. In fact, no capsaicin-induced
pain might mean excessive damage to nociceptive fibers (or in alternative less sensitive
fibers).
The latest result lead to the inclusion of the
capsaicin-induced pain test as one of the selection criteria for the ongoing
Phase 3 trial. Only patients with at least a pain rating of 2 in response to
capsaicin were admitted to the Phase 3.
Top line results for the trial were initially guided for EOY2014,
however in the Summer 2014, upon interim analysis, BDSI decided to add 80
patients to the trial. The company stated that they were “encouraged by
the outcome of the interim analysis”, but given double blinding no data, nor statistics were presented .
In summary, what are the encouraging aspects suggesting a
positive readout of P3?
- the rationale for the MOA of topical clonidine is sound
and convincing;
- the results from clonidine trials (particularly the
subgroup analysis) are overall positive;
- the rationale for the subgroup analysis makes sense (but I
need to look more into this);
- the target population can easily be identified using a
simple test (capsaicin reactivity);
How about the caveats and the points of concern? Well, here
they are:
- 0.1 mg clonidine works best than 0.2 mg clonidine, and I
personally never like too much the lack of dose- response data.
- the trial primary endpoint is change in pain as measured
using the NPRS. NPRS records come from daily self-assessment by patients and
this can introduce variability.
- similarly, self-administration of the gel can introduce
more variability;
- while I have no problem with increasing the n after
interim analysis (higher n is a good thing), and while I am not a fan of
reading tea leaves in PR, still interim analysis with no mention of stats might
raise some concern; [PS: Thanks to @BiotechSage and @Biowreck for pointing out that data were not unblinded to management and so company could not unveil any stat - on the basis of this information, this concern carries much less weight now].
To conclude, I lean toward thinking that BDSI has a good shot
at goal with this Phase 3. Definitely, not a slam dunk! Indeed, one
needs to clearly consider the risk of a trial whose primary endpoint measure might
be rather noisy and affected by placebo, and whose previous interim analysis
was defined encouraging, but not disclosed.
