The World Turned Upside Down: The Second Low-Carbohydrate Revolution (13 page)

Brehm's experiment is far from the
best
but it is representative
of the type of experiment, the low-carb diet does better but there is
big
spread in the values.

Figure
5-6 A
. Meta-analysis of comparisons of weight
loss on low-fat and
low-carbohydrate diets. Data from reference
[46]
B
.
Responses to
different lipids in Zucker rats.

In 2006 a meta-analysis, that is, a
re-examination of previous
studies, was performed by Nordmann,
et
al
. Meta-analysis is a weak, possibly useless
method. The idea is to
average previous studies but most of us think that averaging errors
makes
things worse, not better. On the other hand, a meta-analysis usually
does give
you a chance to see the results from several different studies.
Figure
5-6 A
shows
the conclusion,
that low-carbohydrate diets lead to better weight loss. In interpreting
the
graph, it is important to understand the when the error bars
(horizontal lines
in this figure) cross the line at 1.0, it means that there is no
effect, either
way. Most of the studies found the low-carbohydrate diet to be more
effective
at six months but, again, no difference at one year. The reason things
got
worse after six months is that the experimenters let them get worse –
they
didn't know how to keep everybody on track. When it's your diet, you
won't let
that happen.

Several other studies have been done
since
and with the same
conclusion. They present the same picture: the shorter ones turn out
better but
it almost always goes one way: you almost never see a study showing
that
low-fat is better. Nutritionists tend to consider that a draw means
low-fat
wins, and when low-carbohydrate wins the results are simply ignored or,
as in
Foster's study, "more work needs to be done." That's why we have the
recommendations that we do and why we have the nutritional mess.

There is always the objection that
studies
that depend on diet
records, as most of these do, are prone to error and that the people in
the
study are mis-reporting what they eat. Errors in reporting have been
documented
but they are not completely inaccurate – usually they are about 80 %
accurate.
All experiments have error, however. It is only a question of how you
deal with
the error. For inaccuracy in dietary reporting to account for the
difference
between diets it would be necessary for subjects in the low-fat group
to
under-report what they ate and for the low-carbohydrate people to
over
-report
what they
ate, or both. These are certainly possible but, again, from a practical
standpoint it might be good to be on a diet where you think you ate
more than
you actually did.

In nutrition, animal studies have
obvious
limitations – it is
considered a departure from normality when we "eat like a bird," or
"like a
pig" – but experimentally, such studies have an advantage because the
subjects
eat what you tell them to eat. The experimenter can control food
intake;
Figure 5-6 B
shows
that weight gain in a species of rats depends on how much they eat, of
course.
But there is also a difference in whether their source of oil is butter
or soy
oil; butter is less fattening for rats in general but the obese ones
really go
for it. You can't generalize the specific result to humans but it shows
that
there is nothing strange about weight gain being dependent on the exact
composition of the diet. Finally, biochemistry tells us that there is
no
receptor for calories, that is, no way for cells to sense how many
calories are
coming in, only how much of each type of macronutrient. The available
chemical
energy will ultimately show up but there is no reason to think that
there will
be a direct relation between calories and weight gain or loss. Calories
do
count but not in a simple way.

Fat – the failure to
accept failure

The decline in the percentage of
dietary
fat in the past thirty
years was describe in the
last chapter
.
The reduction in
percentage
is
partly due
to the increased intake of total calories but, at least for men, the
absolute
amount of total and saturated fat has also gone down. The fact that a
decrease
in dietary fat has been accompanied by overeating and noticeable
increases in
diabetes and metabolic syndrome stands as an inescapable indictment of
the
low-fat doctrine. And, while survival has increased due to treatment,
the
incidence of heart disease has not changed substantially, at least in
the
United States. In other countries where it has, it is more likely due
to
reduction in smoking. It is widely said that association does not
predict
causality – does not
necessarily
predict causality is more accurate – but a
lack
of
association is strong evidence for a
lack
of
causality.

Animal Models.

To be fair to the low-fat doctrine,
it is
easy to be misled.
Animal studies are critical in biological science and it seems that
mice,
especially the mice bred for laboratory work, will get fat on high fat
diets
even without any carbohydrate
[47-49]
.
How is that
possible? People don't usually get fat on high fat diets, or at least,
don't
over-consume fat to such a high degree on high fat diets, unless the
diet is
also high in carbohydrate. Mice aren't men but they are a generally
good
metabolic model for many things so this is a drastic difference.
Carbohydrate
is key in human metabolism. Not necessarily so in rodent metabolism
where a
high fat diet can bring on obesity, diabetes and cardiovascular disease
even in
the absence of carbohydrate. We don't yet have a theory to encompass
the
differences. If our understanding of the catalytic role of insulin is
correct,
then it may well be that mice (who normally live on high carbohydrate
diets)
may maintain a functionally high level of insulin all the time, so the
bias
toward an anabolic state that occurs with high carbohydrates in humans
is
always 'on' in mice.

Whatever the explanation, it may be
hard
to recognize that the
model system that we have used so extensively might have most impact
because it
is
different
from the way people respond. In this case, animal models might offer a
clue to
what people do
because
humans do not behave like the animals.

The real problem is that we have not
faced
the practical,
experimental tests in humans. We have large expensive clinical trials
with a
consistent and reliable outcome: there is no effect of dietary fat on
obesity,
cardiovascular disease or just about anything else. The unwillingness
to face
these failures makes this a remarkable phenomenon in the history of
medicine –
that it persists in a period of sophisticated science and technology
makes it
nearly unbelievable.

"Unbelievable" is the key word. That
a
science is incomplete or
has flaws is what one expects but that the whole of the establishment
opinion
on diet-heart is totally meaningless, is hard to understand. How could
they
keep doing the same experiment over and over without success? How could
that
be? How could they get away with it? Why would they
want
to
get away with it? New trials
continue to show nothing. Well, not nothing. They clearly show that
low-fat is
ineffective for weight loss or just about everything else.

In science, excluding a theory is
always
stronger than showing
consistency. In this case, if the fat-cholesterol-heart story were as
they say,
as salient and inescapable a risk as they make it out to be, then none
of these
big studies should fail. Not one. In fact, almost all fail. An
occasional one
shows an effect. There has been an increasing admission that high
carbohydrate
is not a good thing. The admission usually comes with a qualifier:
"especially
refined sugar" or "refined starch", but no study has directly compared
"refined"
and "unrefined" – high-GI and low-GI are not measures of refinement and
they
are, in any case weak predictors. The drastic increase in total
carbohydrate
that has accompanied the obesity epidemic is its most salient feature.
You may
choose to ignore it if you don't like it, but it is there.

So, does this mean we can add more
fat? Is
fat still bad? It's
probably fair to say that most people think that, in some way, fat it
still
bad. It will turn out that the role of fat in the body is itself
controlled, directly
or indirectly through hormones, by carbohydrates. Deleterious effects
of lipid
metabolism are under the control of carbohydrates. Perhaps, most
surprising,
the biochemistry shows that it is the fat (more precisely the fatty
acids) in
your
blood
that
are the problem but they are more likely to come from dietary
carbohydrate than
from dietary fat.

Are carbohydrates
fattening?

"I don't understand. I went
to this conference and
they had a buffet every night and I really pigged out on roast beef and
lobster
but I didn't gain any weight"  – Author's brother.

 Nobody ever says
that after going on a cruise where
they pigged out on pasta and didn't gain any weight. In the faucet
analogy, in
the area of weight gain and loss, fat can be packaged into fat cells.
Insulin
opens the faucet for fat storage but it shuts down the faucet for fat
oxidation. At this point, though, you might ask whether all this
matters.
Doesn't it all even out in the end? Isn't it just calories in, calories
out,
or, as they always say in the news releases "a calorie is a calorie?"
And,
don't the laws of thermodynamics tell us that?

It is hard to tell the extent to
which you can lose more
weight calorie-for-calorie by changing the composition of the diet. One
clue,
though, is that when an experiment shows that one macronutrient is more
inefficient than another (wastes calories as heat), it is usually the
low-carbohydrate arm that is less fattening. Critics say that the
results are
due to inaccurate reporting of food intake and that it is just about
calories.
Low-carb diets, they say, simply reduce total energy intake. As in the
Brehm
study, food frequency records can have substantial error and, again, if
that is
why the low-carb group always wins in these face-offs, low-carb
participants would
have to be over-reporting what they ate or low-fat comparisons groups
would
have to be under-reporting what they ate, or both. Again, there might
be a real
benefit of being on a diet where you think you ate more than you did.
And if
you thought that reducing calories was "just," you wouldn't be reading
this
book.

And thermodynamics does not predict
that "a calorie is a
calorie." Most people who quote the "laws of thermodynamics"– they
usually mean
just the first law – have never studied thermodynamics and simply don't
know
what it is about. The essential feature of thermodynamics rests, not
with the
first law which is about energy conservation, but rather with the
second law
which says that all (real) processes are inefficient. Energy is
dissipated. The
variable efficiency (the extent to which energy is wasted as heat) of
fat,
protein and carbohydrate is well known but, in the medical literature,
ignored
at will. Often, however, total calories may turn out to be the
controlling
variable but insofar as it is independent of macronutrient composition,
it is
because of the homeostatic (stabilizing) mechanisms of biological
systems, not
because of thermodynamics. Thermodynamics is my special interest and
we'll come
back to it in
Chapter
14
. Promise.
First, look further at some of the underlying chemistry.

"The Atkins diet is a
high calorie starvation
diet."

The quotation is from George Cahill,
one of the pioneers in
the study of metabolism and the response to starvation. The idea is
that the
reduction in blood glucose and insulin and the increase in glucagon
that
accompanies reductions in carbohydrate intake resembles the changes
that are
associated with total reduction in calories. In starvation, insulin
goes down
and glucagon goes up and fat oxidation increases and at some point
ketone
bodies are generated.

In 1992, Klein & Wolfe
[50]
carried out a defining experiment. They had subjects go without food
for three
days and then, after a period of rest, they went through another three
days of
starvation. This time, however, the people in the experiment received
intravenous injection of a lipid emulsion that was designed to meet
their
resting energy requirements. Klein & Wolfe measured several
physiologic
parameters in both experiments and as shown in Table 5-1, there was not
a great
difference between the two tests despite the very large difference in
energy
intake. The levels of free fatty acids (FFA) were the same and, in both
cases,
ketone bodies were reasonably high. The presence of fat, if anything
reduced
the rates of fat oxidation. The conclusion from the study was that
"these
results demonstrate that restriction of dietary carbohydrate, not the
general
absence of energy intake itself, is responsible for initiation of the
metabolic
response to short-term fasting." The statement is undoubtedly something
of an
exaggeration, but the experiment brings out one of the major themes in
diet and
metabolism: carbohydrate is a controlling element while dietary fat
plays a
relatively passive role. This is not to say that circulating fat, body
fat or
fatty acids do not play a role in metabolism but assuming that dietary
fat
equates to body lipids is wrong, way wrong. This is a major theme in
this book:
"you are what you eat" is not a good principle.