Sie sind auf Seite 1von 10

Based on true

story

When the conspiracy meets the


code breakers...
Sukit Roongapinun, M.D.
Dept. of Pharmacology
Chiang Mai University

Pharmaco
newsletter
Correspondence:
drsukit@yahoo.com
Is rosiglitazone associated with
cardiovascular disease (CVD)?

Nissen SE, Wolski K. Effect of rosiglitazone on the risk of myocardial


infarction and death from cardiovascular causes. N Engl J Med.
2007;356(24):2457-71.
The Suspicion begins... for MI risk

Study names Odds ratio for MI risk

Many small trials ns

DREAM ns

ADOPT ns

COMBINED ALL
SIG!!!!

•So, some MI risk is hidden somewhere!!!

•?? but where ???


Fishy cardiovascular disease (CVD) death

Study names Odds ratio for CVD death

Many small trials 0.02 SIG!!!!!

DREAM ns

ADOPT ns

COMBINED ALL 0.06 (almost)!!!!

• Many Small trials , sponsor from whom???

• beautiful names = marketing name????

• DREAM, ADOPT ---- who sponsored ?????


2 beauties and a beast....

• OK, think about CVD death...

• TWO beauties DREAM+ADOPT (n=3800) ---> no sig.

• One beast is ‘many small trials (n=6500)’ ---> P=0.02 -->


put rosig to more death.

• When combine beauties and the beast ---> P value =


0.06

• So, addition of 3800 cheerful people to 6500 painful


people cannot make 6500 painful happy. P value still gets
closed to a significant level.
Addition of PLACEBO revealed a secret.
Rosig vs. Drugs below Odds ratio for CVD death
Metformin (M) ns
Sulfonylurea (S) ns
Insulin (I) ns
PLACEBO 0.07 (almost again)
M,S,I + placebo 0.03 (= rosig die easier)

• First, Rosiglitazone gives similar death rate to the metformin, sulfonyl,


or insulin treatment. Seem ok.

• BUT, receiving rosiglitazone seems to die a little easier than receiving


placebo (P=0.07) ... this is frightening!!

• What’s about MSI+placebo vs. Rosig?


MSI = metformin+sulfonylurea+insulin

• OK... MSI are long-time anti-DM therapy... .. should help or at least not hurt!

• So let say, MSI give mortality = 20%

• How about placebo? Placebo should give death rate = 20 % or higher.

• if we pool MSI + placebo, the death rate would be = 20% or higher.

• Now we see, Rosig compared to MSI showed similar death rate (P= ns).

• But when we add placebo (death contributor) to a non-significant MSI, we get P=0.03.

• P=0.03 means the death from ROSIG becomes clearer.

• How could it be? we put more death to comparator side, it turns comparator to survive
(die harder). .... !!!!!
One possible reason

• The data of rosig in those studies of MSI was not completely given.

• The death is hidden. Actually, huge number of death is hidden.

• OK we try to help Rosiglitazone... by putting more death toll (Placebo


= Death Maker) to comparator side.

• The death is even clearer disclosed........ in rosiglitazone.

• Rosiglitazone is really in bad shape at this time.


or is the placebo really lowering the death
rate?

placebo
rosig MSI MSI
rosig
death
death
rate
rate

“Possible excuse is placebo is really


really lighter in death rate than MSI. do
you believe it?”
THE END

of newsletter and rosiglitazone...

Das könnte Ihnen auch gefallen