Metformin:
What, Me Worry?
Evan David
Rosen, M.D., Ph.D. Assistant Professor of Medicine,
Harvard Medical School
The
anti-diabetic drug metformin (marketed as Glucophage™
in the U.S.) has had a somewhat checkered past. People
have known about its sugar reducing capabilities for
a long time, and a plant called “goat’s
bane” that makes a metformin-like molecule may
have been used in medieval times as a treatment for
diabetes.
Other biguanide
compounds (as the class of drugs that includes metformin
is called) were used in Europe in the 1930’s,
but for unclear reasons fell out of favor. Metformin
was “rediscovered” in the 1950’s
in the United Kingdom in particular, and has remained
on the market there since that time, enjoying significant
popularity as an oral therapy for type 2 diabetes.
In the
United States, metformin was ignored. Instead, a related
compound called phenformin was licensed and sold.
It proved to be a poor choice, as a significant number
of patients developed a severe derangement of their
blood chemistry called lactic acidosis, half of who
died of the complication. Phenformin was withdrawn
from the U.S. market, and for years sulfonylureas
were the only oral anti-diabetic agent available in
this country. Metformin was, unfortunately, painted
with the same brush as phenformin, and there was little
enthusiasm for using it here because it was assumed
it would suffer from the same propensity to cause
lactic acidosis
.
Nonetheless, metformin continued to be used in Europe,
and studies in the 1980’s showed that it works
by improving insulin resistance rather than by increasing
insulin secretion (as sulfonylureas do). This makes
it a particularly attractive drug for type 2 diabetes,
which is characterized by severe insulin resistance.
Encouraged by the decades-long safety record of metformin
in Europe, the U.S. FDA finally approved it for use
here as well, and it hit the market in 1997. Since
that time, metformin has become the most widely prescribed
oral anti-diabetic agent in this country.
Why did it become so popular so fast? A few reasons,
actually. First, for a few years it was the only insulin-sensitizing
drug available, which is particularly important for
type 2 diabetes (thiazolidinedione drugs like rosiglitazone
and pioglitazone are also insulin sensitizers, and
have since become available as well).
Second,
unlike other anti-diabetic medications like sulfonylureas,
thiazolidinediones, and insulin itself, metformin
is not associated with weight gain. It may, in fact,
have a mild weight-reducing (or at least weight-stabilizing)
effect. This makes it ideal for patients with obesity,
an all too common situation among folks with type
2 diabetes.
Finally,
metformin became immensely popular because it works
well. Reductions in hemoglobin A1c are at least as
good among metformin users as among users of other
anti-diabetic drugs, and important longitudinal studies
like the UKPDS have suggested that metformin might
be even more effective than other agents for obese
patients with diabetes.
The fly
in the ointment for metformin, however, is that it’s
never completely shaken the taint of association with
its evil twin, phenformin. Because of the fear of
lactic acidosis, metformin carries a huge warning
label that discusses the issue, and physicians are
explicitly cautioned not to use the drug in patients
who have secondary conditions that might predispose
them to lactic acidosis, such as renal insufficiency,
liver problems, heart failure, and advanced age. The
number of patients who are thus “ineligible”
for metformin is quite large, and may include up to
50% of those who might otherwise benefit from the
medication.
Furthermore,
patients who take metformin are often switched to
other medications (or worse, simply left untreated)
when they are admitted to the hospital for any reason,
because of the fear that hospital-based factors (such
as IV contrast dyes, other medications, and bleeding)
might temporarily increase the risk of lactic acidosis.
Studies
have shown that lactic acidosis does occur in metformin
users, at a rate of around 2-9 cases per 100,000 person-years.
This is an extremely low risk, but would be reason
enough for all the warnings and concern if the problem
were truly due to the metformin. And here’s
where the connection gets a little sticky. Lactic
acidosis occurs when tissues become “hypoperfused,”
which means they do not get enough oxygen-containing
blood flow, as in heart failure.
Lactic
acidosis can also occur when the lactate molecule,
which is made by all tissues at low levels even in
healthy people, builds up in the blood because it
cannot be excreted by the liver or kidney. As it turns
out, liver, kidney, and heart disease is not all that
uncommon in patients with diabetes. Furthermore, metabolic
adaptations to diabetes per se may also increase lactate
production. One might therefore expect that diabetes,
even in the absence of metformin, should be associated
with lactic acidosis. In fact, that’s exactly
what’s been shown. Lactic acidosis appears in
diabetic patients at around 9-10 cases per 100,000
person-years, even in patients who do not take metformin.
A new study
has just been published which sheds even more doubt
on the whole metformin-lactic acidosis connection.
This study is a “meta-analysis,” which
means it’s a study of other studies, rather
than a direct look at a new group of patients. The
authors compiled 194 papers that looked at metformin
use in type 2 diabetes, and found no examples of either
fatal or non-fatal lactic acidosis in a cumulative
37,000 patient-years. The authors were able to put
a statistical upper bound on the risk of lactic acidosis
in diabetes at 8 cases per 100,000 person-years, which
is not different from that seen in patients who don’t
take metformin.
Now, this
sort of study has significant limitations, but it
clearly suggests that metformin might be a less problematic
therapy than originally anticipated. Interestingly,
many of the studies they analyzed had quite a few
patients with at least one of the contraindications
to metformin mentioned above.
This implies
that (1) the guidelines for avoidance of metformin
are not being rigorously adhered to in the real world,
and (2) despite the presence of such patients, there
were still no cases of lactic acidosis seen.
Because
it is unclear just how many of the patients had heart,
liver, or kidney disease, one cannot start agitating
for the repeal of those contraindications.
It does
point to the need for further analysis, however, that
takes into account the baseline rate of lactic acidosis
in patients with type 2 diabetes, and it opens the
door just a little for the thousands of patients who
are not currently able to use metformin in their glucose
control programs.
Reference:
Archives of Internal Medicine 163:2594 (2003)
Viewpoint is an editorial column that expresses the
opinion of the specific Medical Director, who is solely
responsible for its content. Viewpoint does not represent
the views or opinions of Veritas Medicine and does
not reflect the opinions of other physicians and researchers.
For other
Features from Dr. Rosen, go
to Dr. Rosen’s Archives
