HGSI >>post from google ths am
n many ways, this situation reminds me of DNDN. DNDN went back and
did a retrospective analysis and found a large DISTINCT AND WELL KNOWN
subgroup that responded to their vaccine. The analysts dismissed this
a cherry picking data, but they were right to be suspicious since many
biotechs have been guilty of wishful thinking. But DNDN had just
missed the endpoint,,,,so it was not much of a leap to see that the
followup study probably would be good.
HGSI
Lupus can mainly be thought of as seropositive and seronegative types.
It would have be logical to prospectively look at these separately
from the beginning.
This first thing I wanted to know was how seropositive vs negative
patients responded. These subsets of patients are long believed to be
different diseases which progress and respond to treatment
differently.
Before the results of the phase II study came out, I was talking with
other biotech friends and wondering why they did not have the
seropositive patients designated as a prospective population,,,so HGSI
could claim statistical significant if it came out positive.
This is not your typical retrospective analysis,,,,an excerpt from a
news story below,,,
>>>"The latest study includes only patients who have certain immune-system antibodies, which may indicate more disease activity. These patient made up about 75% of the Phase II study. Furthermore, the trial's endpoint uses a combination of several disease-activity measures that have never been used in a study before.
A retrospective analysis of the failed Phase II study shows it would
have succeeded using the new endpoint, which gives Benlysta an edge
that previous therapies didn't have, Labinger said.
After working with the FDA, Human Genome obtained a special protocol
assessment, which is an agreement with the agency on a study's design
that meets regulatory requirements for a new drug application and
could lead to a drug's approval.
Although Birchenough acknowledges that the endpoint may be legitimate,
he argues that there have been "few, if any" cases where a
retrospective subset analysis of a failed Phase II trial led to
subsequent success in a late-stage trial. (WRONG, WRONG,
WRONG,,,although "usually true, I also usually wonder why some
companies go to phase III when the data for the subset is so
weak....But there are also plenty of drugs that had a great response
in a large subset that succeeded).
Researchers have long struggled with the design of lupus clinical
trials because almost every patient has different symptoms and,
accordingly, gets different treatments including anti-inflammatory
drugs, steroids, anti-malarials and immunosuppressants."<<<<
This is not a tiny subpopulation,, first of all,,,75% (72% in HGSI
later press release) of 449 patients were seropositive.
second,,,significance was p<0.01.
I think the only question was lack of dose response,,,but if the low
dose was fully effective,,,then a higher dose will not work any better
(typical of most drugs). But just as important,,,,ALL 3 DOSES SHOWED
ACTIVITY OVER PLACEBO.
AI
