Prana Biotechnology finally announced the results from their Huntington’s Phase II trial today, and depending on who one reads, one might come away with the impression of great success–or abject failure. We think the truth is in between, but is closer to the former than the latter, and here is why:
It was a 109 pt trial (just 104 completers), which means one must look beyond the p=.05 criterion for establishing an effect. PBT2 did hit statistical significance on a measure of front area/executive functioning–based on Trailmaking B–and showed a trend towards improvement on a functional capacity measure. A tiny fMRI imaging substudy (six patients–four on PBT2) showed evidence of slowed atrophy in those areas where HD tends to produce the most degeneration. It is worth noting that the specific effect upon executive functioning measured by Trailmaking B (which is a simple but useful neuropsych screening measure) was also seen in the previous Alzheimer’s trial–which argues against the findings being fluky, and raises some intriguing questions about why the mechanism might be particularly important in that neuroanatomical area.
No detail regarding other cognitive measures was provided, which likely means that no trends were observed. Even so, it is in executive functioning that HD tends to display its more prominent impact on cognition, thus it is where one might expect to see an impact upon decline over a relatively short period (six months). It is worth noting that a 2011 Lancet Neurology report found that imaging was the best measure of disease progression–and this was over two years–and that progression as measured by other criteria is highly variable, which makes a six month timeframe all-too-short.
This brings us to the headline from Adam Feuerstein of Street.com: “Prana Bio Huntington’s Disease Drug Fails Key Efficacy Hurdles.” In general, we like Feuerstein’s work a lot–he punctures the balloons of PR hype with great accuracy and zeal. But not here, where his zeal remains, but his aim was off. He zeroed in on the lack of statistical significance for broad spectrum assessments of cognition and motor function and deemed the trial a failure–dismissing the signs of improvement as essentially irrelevant.
The problem with his conclusions is that they ignore some basic principles of assessing change in neurodegeneration: The duration of the trial, which here was quite short; the size of the sample, which here was very small; and the difference between prespecified and datamined divergences. Alzheimer’s studies tend to now cover at least twelve months, more often eighteen months of time, and enroll thousands of patients–even in trials where diagnosis is certain, like Lilly’s revamped solanezumab Phase III design. Had Prana’s study had that kind of size and duration, and yielded these results, it would be justifiably deemed a failure. It did not, and its results must be considered more akin to a Phase IIa trial, a pilot study that points to possible utility, albeit not establishing it. The positive findings were where previous data would suggest they would be found, not a random walk through the statistical thickets.
What can be said unequivocally here? That Prana’s financial constraints prevented them from running a more definitive study that enrolled enough people, covered enough time, and utilized sufficient measures (especially in the imaging component) to provide something approximating certainty. But we believe, particularly in concert with a similar cognitive effect shown in an entirely different neurodegenerative disorder, that there is a signal here worth pursuing, and that while PBT2 has a long road to proof ahead of it, it has earned that opportunity with some preliminary indicators of benefit.