Transcript CURRENT MANAGEMENT OF INFECTED PROSTHETIC …

Weaknesses of major randomized
trials: they are cleverly designed
against coronary surgery. They are
not representative of the daily
practice (highly selected patient
population), but their results are
generalized for the whole
population with CAD.
RANDOMIZATION IN ARTS TRIAL
(Arterial Revascularization Therapy
Study) University Hospital Zürich, AprilDecember 1997 5/986 (0.5 %) patients
CABG
37.6%
Randomized
0.5%
PTCA
61.9%
STUDY ENROLLMENT IN MAJOR
RANDOMIZED STUDIES COMPARING PTCA
TO CABG (91 730 patients)
RITA, ERACI, GABI, EAST, CABRI AND BARI trials
Enrolled
4.9%
Screened
95.1%
It is a well-known fact that patients
recruited for trial have better survival
and better results than those eliminated
from the trial because of some
exclusion criteria.
NIFEDIPINE IN ACUTE MYOCARDIAL
INFARCTION (TRENT TRIAL)
(9292 patients admitted with AMI)
30
26.8
25
20
Mortality
18.2
15
10.2
9.3
10
6.3
6.7
5
0
Excluded
Placebo
Nifedipine
Mortality
with AMI
When analyzing a trial, look
carefully at “Material and
Methods” section: crucial
information explaining the
results might be found there.
Enrollment and Randomization of Patients with Previously Untreated Three-Vessel or Left Main
Coronary Artery Disease in the SYNTAX Trial
2 years, 85
centres:10.6
patients/year
Serruys P et al. N Engl J Med 2009;360:961-972
Average number of patients seen by centres
in Syntax trial is less than 11 patients/year!
If we accept the fact that “all comers”
entered the trial, these centres should have
been closed long ago, according to present
standard of PCI and CABG practice.
Cardiac-Related Medications Given after the Study Procedure
Serruys P et al. N Engl J Med 2009;360:961-972
A study with a radically different drug
treatment protocols in the two
analyzed groups is statistically
invalid.
Difference in outcome in trials funded by for-profit
and not-for-profit sources
JAMA 2006; 295(19):2270-4
100
p for trend < o.oo5
% positive
80
60
67
82
67
57
55
49
54
50
40
35
40
69
66
20
0
ll
A
ls
a
ri
t
li
C
ca
i
n
le
nd
in
o
p
ts
g
ru
D
e
D
ce
i
v
Not for profit
Mixed
For profit
Beware of industry sponsored trials!
Results very often meet the
sponsor’s expectations!
CABG MORTALITY RATES AMONG DIFFERENT
SURGEONS: A PROSPECTIVE STUDY
JAMA 1991;266:803-809
There is a major difference between
swallowing a pill and performing a
complex operation; this fact is not
appreciated by statisticians.
Survival
with CABG
@ 3 years
is 28.4 %
higher!
Hannan E et al. N Engl J Med 2005;352:2174-2183
Non-randomized observational trials
can reach very important results,
which are presently ignored by the
academia schooled in conventional
statistics