Cholesterol drug question
If low is good, is lower best?

Peter Battershill and Alan Cassels
Special to the Sun

June 27, 2005

'Cholesterol drug best in large doses," a recent
front-page headline
proclaims.

The text goes on to drive home the message that a
recent
cholesterol-lowering study has found "the strongest
evidence yet that
driving cholesterol down to very low levels offers
additional protection
from heart attacks and strokes."

Before you run to the doctor for a larger dose of
cholesterol-lowering drug,
you need to know some of the finer details that lay
behind those breathless
headlines.

The first thing to ask when reading any report of a
new study is this: "Was
this study looking at people like me?" A crucial
question, because the
people involved in the study may look a lot different
from those who end up
taking the drug in the real world, as it did in this
case.

This study compared patients who were taking two
different doses -- 10 mg
vs. 80 mg -- of atorvastatin (trade name Lipitor), the
world's most popular
cholesterol-lowering drug. It started off with 18,469
eligible participants,
but then the culling began.

To be eligible for this study, you needed to have been
diagnosed with heart
disease, the population who would use Lipitor in the
real world. About 3,005
were excluded right off the bat, probably because they
had other health
conditions.

We're not really sure what those were because the
report, published in the
New England Journal of Medicine, wouldn't elaborate on
why those people were
excluded. Obviously, the researchers wouldn't want
other diseases clouding
their results. That's fair enough, isn't it?

But next there's a bit of sleight of hand. The
remaining 15,464 participants
took 10 mg of atorvastatin for eight weeks to see if
they "responded."

If they didn't attain a certain lipid profile (i.e. an
LDL level below 130
mg/dl) they were dropped from the study. This got rid
of 4,634 real
participants. Another 827 people taking the low-dose
atorvastatin
experienced adverse events or left the study for other
reasons, leaving the
researchers with 10,003 -- a bit more than half of the
original group.

Selectively removing people from clinical studies
introduces a significant
bias as those real-world patients are eliminated. The
study results can't
tell you anything about the drug's effect on a real
cross-section of the
population that includes those who have more than one
illness or who don't
respond to the studied medication at all.

Finally, the individuals in this "enriched sample"
were randomly assigned to
take 10 mg or 80 mg of atorvastatin for the next five
years.

So, after five years of study, what were the results?

The rate of cardiovascular events (heart attacks and
strokes) in the low
dose group was 10.9 per cent: That is, 109 out of
1,000 of low-dose drug
takers had an event during the trial. In the higher
dose group, 8.7 per cent
or 87 in 1,000 had an event. The difference between
these two groups?
Twenty-two people out of a thousand, or 2.2 per cent.

Another way to describe this finding is that a doctor
would have to treat 45
pre-selected, screened patients for five years with
the higher dose drug to
prevent one cardiovascular event.

The kicker, though, is there was no difference in
overall survival; it
didn't matter which group you were in because the
death rate in both was the
same. The lower rate of cardiovascular deaths in the
80 mg group was offset
by an increase in other kinds of deaths for the people
in the 10 mg group!

The researchers bizarrely conclude that a "substantial
clinical benefit" is
derived from preventing cardiovascular disease with
higher-dose
atorvastatin. Go figure.

What is the advantage of reducing cardiovascular
deaths if it is
counterbalanced by an increase in other causes of
death? Does it make sense
to take a drug to change which disease you'll die
from?

Another important bit of information is knowing who
exactly were the primary
authors of this study. No surprises here, these
doctor/researchers are
affiliated with universities and cardiac institutes.

Yet, as their published conflict-of-interest
statements indicate, 10 of 11
principal authors have financial ties with Pfizer,
Lipitor's maker.

This is tantamount to allowing marketing to masquerade
as science.

Dr. Peter Battershill is a general practitioner in
Victoria. Alan Cassels is
a drug policy researcher with the University of
Victoria.