Biotech Values

[iwfal]

XOMA -

1) They described the powering assumptions of the Behcets trial - some pretty aggressive assumptions with an assumed HR=0.3 and designed power of 90% (which can be reverse engineered to say that the event count is just about 30 patients and a measured HR of about .45 i the worst HR that will be stat sig.). (Note: in a similarly designed infliximab vs adalumumab trial for general uveitis the HR was around this value - so it is possible. But aggressive all the same.)

Just some further context (but with the caveat that the randomized data is VERY sparse, with small trials and with very varied endpoints): the benefit of adalimumab vs infliximab is somewhere in the neighborhood of about OR/HR=0.3 to 0.4. And in addition the benefit of infliximab vs etanercept appears to be in the same range. And, finally, the benefit of etanercept has been shown to be essentially nil wrt uveitis.

Thus, by tenuous extension the benefit of adalimumab is in the neighborhood of OR/HR~0.1. (Note - although the extension is definitely 'tenuous', it is somewhat supported by the fact that in these kinds of inflammatory indications it is clear that infliximab generally provides benefit, but adalimumab provides substantively more.)

Random thoughts: my reaction to the design HR of 0.3 was based upon oncology (where most HRs exist - spaces outside onc more typically use responder analyses etc). But given the above data it isn't unreasonable. That said, still makes me uncomfortable.