Philip J. Devereaux talks
with ScienceWatch.com and answers a few questions
about this month's Fast Breaking Paper in the field of
Article Title: Effects of extended-release
metoprolol succinate inpatients undergoing non-cardiac
surgery (POISE trial): a randomised controlled
trial Authors: Devereaux, PJ, et al.
Year: MAY-JUN 2008
* McMaster Univ, Fac Hlth Sci Clin Epidemiol & Biostat,
Room 2C8,1200 Main St W, Hamilton, ON L8N 3Z5,
* McMaster Univ, Fac Hlth Sci Clin Epidemiol & Biostat,
Hamilton, ON L8N 3Z5, Canada.
(addresses have been truncated)
Why do you think your paper is highly
During the last few decades, substantial advances in noncardiac surgery
(i.e., all surgeries except surgeries performed directly on the heart) have
improved disease treatment and patients' quality of life. As a result, the
number of patients having noncardiac surgery is growing. Worldwide, over
200 million adults annually undergo major noncardiac surgery.
Noncardiac surgery is associated with major cardiovascular complications
(i.e., death due to a cardiovascular cause, nonfatal heart attacks,
nonfatal cardiac arrest, and nonfatal stroke). Worldwide, approximately
3-5.4 million adult patients annually suffer a major perioperative
cardiovascular complication in the first 30 days after surgery. This is
similar to the annual global incidence of new patients acquiring human
immunodeficiency virus (HIV), and identifies major perioperative vascular
complications as a similarly common major public health problem.
"When our methods of data monitoring
detected these cases of fraud, the POISE
Operations Committee, with the support of the
Data Safety and Monitoring Committee (DSMB),
the primary funder of POISE (the Canadian
Institutes of Health Research [CIHR]), and
the ethics committee at McMaster University
decided to exclude all the patients
associated with fraud."
Despite the magnitude of this problem, this is a neglected area; there is
not a single established effective and safe intervention to prevent major
perioperative cardiovascular complications. The striking absence of
prophylactic interventions reflects the paucity of large randomized
controlled trials (RCTs) evaluating perioperative interventions.
Considering this background, there are several reasons why PeriOperative
ISchemic Evaluation (POISE) is so highly cited. First, POISE is the world's
largest randomized controlled trial ever undertaken to assess a potential
prophylactic intervention to prevent major perioperative cardiovascular
events. POISE evaluated a beta-blocker (a commonly used drug to treat high
blood pressure that also keeps a patient's heart rate down) versus a
placebo in 8,351 patients in 190 centers in 23 countries.
Second, for 10 years, guideline committees have recommended giving a
beta-blocker to patients undergoing noncardiac surgery based upon
physiological arguments (e.g., keeping a patient's heart rate down may
protect the heart) and two small randomized trials with substantial
methodological limitations that included a total number of patients less
than 5% of the number of patients included in POISE.
POISE challenges the appropriateness of the current guidelines, in that
POISE demonstrated a perioperative beta-blocker decreases the risk of a
heart attack around the time of surgery, but increases the risk of stroke
and death. Third, these harmful consequences, unanticipated prior to POISE,
highlight the importance and need for large randomized controlled trials in
Does it describe a new discovery, methodology, or
synthesis of knowledge?
POISE describes a new discovery in that, prior to POISE, it was unknown
that starting a patient on a beta-blocker around the time of noncardiac
surgery could increase a patient's risk of death or stroke.
Would you summarize the significance of your paper in
POISE suggests that, for every 1,000 patients undergoing noncardiac
surgery, a beta-blocker would prevent 15 patients from suffering a heart
attack, but it would also result in an excess of eight deaths, with five
patients suffering a stroke.
How did you become involved in this research, and were
there any problems along the way?
Initially, a small group of anesthesiologists, cardiologists, internists,
and surgeons came together and decided to start investing ways to prevent
major cardiovascular complications around the time of noncardiac surgery
because this represents such a large population problem for which there was
very limited research.
We encountered two primary problems in conducting this research. First,
POISE almost did not happen because many physicians (primarily based upon
the influence of the guidelines that advocated patients should get a
beta-blocker around the time of surgery) felt it was unethical to not give
a beta-blocker to a patient having surgery. We had to spend a lot of time
giving talks and meeting with physicians around the world to review the
reality of the perioperative beta-blocker data. We had to point out that
the data was not definitive and that we needed a large RCT to ensure that a
perioperative beta-blocker was beneficial and safe.
Second, we discovered fraud in centers participating in Iran and also with
a study coordinator who participated in Colombia. We employed three methods
of data monitoring in POISE that included: 1. central data consistency
checks that evaluated the consistency of center data as it was submitted;
2. statistical monitoring evaluating data across centers to see if any
centers stood out; and 3. on-site monitoring.
When our methods of data monitoring detected these cases of fraud, the
POISE Operations Committee, with the support of the Data Safety and
Monitoring Committee (DSMB), the primary funder of POISE (the Canadian
Institutes of Health Research [CIHR]), and the ethics committee at McMaster
University decided to exclude all the patients associated with fraud.
Where do you see your research leading in the
We are initiating the POISE-2 pilot and have submitted a grant to fund the
main POISE-2 Trial. POISE-2 will build upon the results in POISE-1 in that
we want to find a way to obtain the benefits we demonstrated with
perioperative beta-blockers (i.e., reduction in heart attacks) but avoid
the harms we demonstrated (i.e., increased risk of death and stroke). The
negative effects demonstrated in POISE-1 appeared to have occurred through
a perioperative beta-blocker causing significant low blood pressure around
the time of surgery.
In POISE-2 we will evaluate a drug (i.e., clonidine, an alpha-2 agonist)
that can help to keep a patient's heart rate down, similar to a
beta-blocker but, based upon some research we have undertaken, may have
less significant reductions in blood pressure. These physiological changes
encourage us that clonidine around the time of noncardiac surgery may
prevent heart attacks without increasing the risk of stroke or death, but
we will require a large randomized controlled trial to compare clonidine
versus a placebo, in order to have confidence in the effect of this drug.
In POISE-2, we will also study the effects of ASA (aspirin) versus placebo
around the time of surgery.
Based on findings in POISE-1, we have also initiated the world's largest
international prospective cohort study evaluating major cardiovascular
complications around the time of noncardiac surgery. This study is called
the VISION Study and will include 40,000 patients. We have recruited over
8,000 patients in the first 18 months and are on track to complete this
study in the next two years. VISION is addressing many questions, including
what proportion of heart attacks physicians may avoid missing around the
time of surgery through monitoring of a simple and relatively inexpensive
blood test, i.e., Troponin T (TnT).
Do you foresee any social or political implications for
POISE has another take-away message that goes beyond perioperative
beta-blockers. Guidelines have recommended perioperative beta-blockers for
over a decade. Even if only 10% of physicians acted on the guideline
recommendations throughout the last decade (several studies suggest that
30% of physicians prescribed a perioperative beta-blocker to their at-risk
patients), 100 million patients would have received a beta-blocker around
the time of surgery.
If the results of POISE are widely applicable, throughout the last decade
800,000 patients would have died prematurely and 500,000 patients would
have suffered a stroke because they were given a beta-blocker around the
time of surgery. This highlights the risk in assuming a perioperative
beta-blocker regimen has benefit without substantial harm, the importance
and need for large randomized trials in the surgical setting, and the risk
in guidelines making recommendations based upon weak evidence.
P. J. Devereaux, M.D., Ph.D.
Department of Clinical Epidemiology and Biostatistics
Joint Member, Department of Medicine (Cardiology)