March 2006
Feature Story
Anne Paxton
Like many medical advances, tight glycemic
control was born from a combination of controlled
trials and uncontrolled observations.
The 2001 study by Greet Van den Berghe, et al
(Intensive insulin therapy in critically ill patients.
N Engl J Med. 2001;345:1359–1367) is the
single core study behind tight glycemic control,
David Baldwin, MD, of Rush University Medical
Center, Chicago, says. Looking at surgery patients
in a large ICU, mostly open-heart surgery
patients, Van den Berghe randomized their
insulin therapy to keep half of them at 110
mg/dL and half at the 160 level, then looked at
a variety of markers for the overall success of
their treatment.
The 2001 study by Greet Van den Berghe, et al
(Intensive insulin therapy in critically ill patients.
N Engl J Med. 2001;345:1359–1367) is the
single core study behind tight glycemic control,
David Baldwin, MD, of Rush University Medical
Center, Chicago, says. Looking at surgery patients
in a large ICU, mostly open-heart surgery
patients, Van den Berghe randomized their
insulin therapy to keep half of them at 110
mg/dL and half at the 160 level, then looked at
a variety of markers for the overall success of
their treatment.
“She found if they stayed in the ICU less than
five days, it didn’t seem to matter whether their
sugars were kept normal or not. But about 40
percent of the patients were sicker and ended up
staying longer than that. That group, the longerstaying
patients, had a dramatic decrease in
overall mortality of about 30 percent, and lower
chance of kidney failure or infection simply by
keeping their blood sugar really normal, super
normal.”
No other study had ever looked at this before,
Dr. Baldwin notes. But several years earlier a
group led by cardiac surgeon Anthony Furnary,
MD, of Providence Health System, Portland,
Ore., had started exploring ways to control infections
in heart surgery patients.
“He was distressed by how often the incisions
through the sternum would get infected in
patients with diabetes since these are very serious
infections. So in 1992 he started a program
of treating the problem with insulin.”
The resulting “Portland Protocol” was a finely
tuned set of orders for intravenous insulin infusions
in hospitalized patients in the ICU and
on the wards.
In the process, Providence got the infection
rate under control, and there is no difference
now between patients with or without diabetes.
According to Providence Health System, Dr.
Furnary’s team was “the first to show that hyperglycemia
was the significant causal factor
for the increased risks of death, infection, and
length of hospital stay in the diabetic patient population,”
and that the elimination of hyperglycemia
with the use of an intravenous insulin
infusion for three perioperative days eradicates
the incremental increases in these complications
that were previously ascribed to the risk factor
“diabetes.”
“We know,” Dr. Baldwin says, “that a lot of
systems in the body which protect you are paralyzed
in many ways by high sugars, so it’s
quite important to keep blood sugar under control
in the critical two or three days after an operation.”
The Portland experience fit very well
with the data from the controlled trial in Belgium,
he notes.
Another large trial, the DIGAMI study (Diabetes
Mellitus, Insulin Glucose Infusion in Acute
Myocardial Infarction) was conducted by several
cardiologists in England in the 1990s. It showed
that using a glucose-insulin infusion to keep
blood sugar below 200 during the acute phase reduced
mortality over 3.5 years. But flaws in the
study’s design and the randomization brought
controversy that “distracted a lot of people from
the basic message,” Dr. Baldwin says, and led to
the study’s results being widely ignored.
The DIGAMI authors took on a second study
in hopes of settling the controversies from the
first and making a bigger impact, and the results
were announced in late 2005. “Unfortunately, the
results were exactly the opposite of what people
hoped,” Dr. Baldwin says. “It was a completely
negative study and showed no difference between
the group with no intervention and the
group with intensive insulin. But the study had
problems, especially that blood glucose wasn’t
well controlled by the participating centers.”
The cardiologists’ interest in the concept was
dampened further, he says.
In the meantime, however, “Lots of other centers,
after the Van den Berghe study was published,
said, ‘That looks great and let’s start implementing
it.’” At Rush, “we used it first in the
surgical ICU starting in 2002, and little by little
we’ve exported it to the rest of the five ICUs we
have throughout our institution.”
To gain the cooperation of surgeons, he focused
on convincing them, a few
at a time, “to let us diabetes specialists
watch over their patients
while they’re in the hospital after
surgery and take care of their blood
sugar issues for them.” Other institutions
have employed various
strategies, he says, such as a topdown
edict from management saying
you have to change your management
of these patients today at
9 AM, or a sudden order that every
patient had to be handled by a diabetes
specialist.
Helping raise awareness along
the way was the American College
of Endocrinology’s Consensus Development
Conference on Inpatient
Diabetes and Metabolic Control in
2003, which provided the first set of
guidelines for tight glycemic control.
“The first time, ever, that the ADA
actually had any recommendations
for inpatient care was in January
2005, in its annual Standards of Care
for Medical Diabetes in America.
And now we have a truly national
set of guidelines for the standard of
care subscribed to by all the mainstream
doctors that specialize in caring
for diabetes patients.” But it
hasn’t translated into a change in
all hospitals because change is so
labor-intensive. “You have to go
change the hearts and minds and
cultures in the entire institution,”
Dr. Baldwin says.
As a result, he estimates, only
five to 10 percent of hospitals in the
United States have adopted tight
glycemic control. “Most medical
schools in the U.S. belong to a collaborative
organization called the
University HealthSystem Consortium,
and in 2005 a joint project was
set up to allow interested hospitals
to have their diabetes hospital care
audited comprehensively to identify
weaknesses and areas for improvements.”
About 37 hospitals signed up for
the audit, and some of them performed
pretty well. “But a lot did
relatively poorly, and even many
of the ones that did well weren’t
exactly stellar.” So, in the adoption
of tight glycemic control, “we have
a long way to go—even in academic
medical centers.”
Anne Paxton is a writer in Seattle. |