Do High-Carbohydrate Diets Increase Heart Disease Risk?
For an excellent, in-depth review of diet and heart disease, along with
a full debate on which is best for treating and preventing atherosclerosis,
a very low-fat, near vegetarian diet or a high monounsaturated fat diet,
see 21st Century Heart or call us toll-free
at 800-462-2352. This master kit includes a 20 hour CPE course and excellent
materials to bring the most up-to-date information to your clients,
including Power Point Shows, Handouts, Overheads and Leader Guide. Don't
wait - order now to take advantage of our introductory price.
By James J. Kenney, PhD, RD, FACN
Good through December 2010
Introduction
Clinical Trials Show Regression on Very-Low-Fat
Diets
Replacing SFA with CHO Reduces LDL-C and CAD
But Wouldn't Replacing SFA With UFA Also Prevent
CAD?
Differences in Blood Lipids Due to Genetics
and Diet May Not Be Comparable
A Lower HDL On A High-CHO Diet May Not Be Dangerous
Fasting TG Also Return to Normal over the Long
Run on a High-CHO Diet
How a High-CHO Diet is Fed Impacts Blood Lipids
Do High-CHO Diets Increase Fasting TG and RLP?
Conclusions
References
Post Test for Do High-Carbohydrate Diets Increase Heart Disease Risk
Yet
another study by Stanford University researchers (and others) has been
published which presumably shows that a diet higher in carbohydrate
(CHO) and lower in unsaturated fats alters blood lipids in such a way
that the risk atherosclerosis is increased.[1] This study found a lower HDL-cholesterol
(HDL-C) (39 vs. 44 mg/dl) and higher fasting triglycerides (TG) (206
vs. 113 mg/dl) in 8 healthy subjects fed a high-CHO (25% fat) diet compared
to those same subjects fed a high-fat (45% fat) diet for 2 weeks. This
type of finding is "old hat" for this research group. The
"new" finding was higher postprandial remnant lipoprotein
particles (RLP) on the higher CHO diet compared to the diet higher in
unsaturated fat. All of these changes in blood lipids have been associated
with an increased risk of coronary artery disease (CAD) in epidemiological
studies of Americans. So, the authors concluded, "Given the atherogenic
potential of these changes in lipoprotein metabolism, it seems appropriate
to question the wisdom of recommending that all Americans should replace
dietary saturated fat with CHO." Is such a conclusion justified?
Back
to top
If
high-CHO diets are more atherogenic than diets higher in fat, what are
we to make of several studies that have reported that very-low-fat,
near-vegetarian diets (VLFNV) cause regression of atherosclerosis in
most patients who already have advanced CAD?[2]
[3] [4]
[5] [6]
And if diets higher in CHO do result in a more atherogenic lipoprotein
profile (as the Stanford group suggested), how can we reconcile such
a claim with the results of a 12 year study. This study found that a
VLFNV diet not only greatly reduced deaths from CAD but also markedly
reduced overall mortality in 50 older subjects (all of whom had had
a previous heart attack) compared to the 50 patients in the control
group who maintained a typical high-fat American diet?[7]
Obviously high-CHO diets are not necessarily more atherogenic than diets
higher in saturated fat.
Back
to top
In
healthy people, the benefits of lowering serum cholesterol by reducing
dietary fat and increasing dietary CHO have been well documented in
terms of preventing CAD.[8]
There really can be no rational debate about whether replacing dietary
saturated fatty acids (SFA) with CHO will help to prevent CAD in the
average American. Perhaps there are some people in America that would
do better health-wise on a diet higher in unsaturated fat and lower
in CHO. But until we know for sure this is the case and have cost effective
diagnostic test(s) to determine who these people are, reason dictates
replacing SFA and cholesterol with CHO and fiber is a safe an effective
public health policy to recommend for all Americans for reducing CAD.
More extreme reductions in dietary fat and cholesterol coupled with
an increase in CHO and fiber has been proven to reverse atherosclerosis
and reduce the overall risk of dying in many patients with advanced
CAD. Those who would advocate replacing SFA with unsaturated fatty acids
(UFA) should be aware that there is no such proof that diets higher
in UFA reverse CAD. Therefore, it seems a leap of faith, rather than
science and logic to advocate the replacement of SFA with UFA as more
effective for preventing CAD in most people.
Back
to top
No
one knows for sure at this point, although changes in blood lipids certainly
suggest that this would be the case. Replacing SFA with UFA does lower
serum total cholesterol (TC) and LDL-C levels about as much as replacing
SFA with CHO. However, no one has shown that a diet high in UFA would
actually cause regression of atherosclerosis in most CAD patients. Indeed,
in one monkey study, a diet high in monounsaturated fatty acids (MUFA)
did improve TC and other blood lipid levels compared to a diet high
in SFA. The changes in blood lipids in the monkeys were in the same
direction as seen in humans fed similar diets. However, despite what
appeared to be "improved blood lipids", atherosclerosis progressed
to a similar degree in the monkeys fed the high-MUFA diet as those fed
the high-SFA diet.[9] So at least in animal models, "improving
blood lipids" does not necessarily slow the progression of atherosclerosis.
The results of this study call into question the wisdom of assuming
that changes in blood lipids in response to dietary changes mean the
same thing as differences in blood lipids observed in individuals who
are all consuming a similar diet. In this latter case, the differences
in blood lipids would be primarily due to genetic factors.
Back
to top
Data
from Framingham and other epidemiological studies have shown that an
increased risk of CAD was correlated with a lower HDL-C and often a
higher-fasting TG level. Few people doubt that higher fasting TG and
lower HDL-C are usually associated with an increased risk of CAD in
people eating a typical high-fat Western diet. Nor does there appear
to be much doubt that higher levels of postprandial RLP are associated
with a more rapid progression of atherosclerosis in people eating high-fat,
Western-style diets. However, there are no studies to show that higher
levels of fasting TG and postprandial RLP and lower HDL-C that result
from switching to a diet higher in CHO and lower in fat actually promotes
atherosclerosis. The Stanford group apparently believes that such changes
observed in blood lipids, during their short-term clinical trials, are
detrimental. But, there is no solid clinical research that demonstrates
that such lipid changes really do promote atherosclerosis and increase
the risk of CAD in people who adhere to high-CHO diets for a prolonged
period of time.
Another
problem with the idea that high-CHO diets could be promote changes in
blood lipids, that increase the risk of CAD, is that it seems to conflict
with most epidemiological cross-cultural observations. Population studies
of people consuming high-CHO diets have shown that CAD is far less common
in those populations than it is in America and other countries where
high-fat diets are the norm. However, in these populations not only
is the intake of dietary CHO higher but the intake of SFA and cholesterol
are much lower and fiber intake is often much higher than they are in
Americans. There may also be differences in activity and other lifestyle
factors that could account for at least some of the differences in CAD
risk between Americans and high-CHO consuming populations.
Back
to top
The
Stanford group and others have suggested that the drop in HDL-C seen
in short-term studies should be expected to increase the risk of CAD
over the long run. There are 2 reasons to be skeptical of such a claim.
First, there is growing evidence that the drop in HDL-C that results
from restricting dietary fat intake does not lead to a permanently lower
HDL-C. This is because replacing high-fat foods with high-CHO foods
usually reduces ad libitum energy intake. This results in weight loss
and a lower body weight is usually associated with an increase in HDL-C.
For example, when a group of hypercholesterolemic men were placed on
an ad libitum VLFNV diet for 3 months their energy intake decreased
and they lost about 16.5 lbs. On this low-fat diet, their LDL-C dropped
from 236 to 139 mg/dl (or -41%) and their TG dropped from 170 to 145
mg/dl (or -15%). But their average HDL-C was essentially unchanged
(36 to 37 mg/dL or +3%).[10] Those who continued to consume
a very-low-fat intake for another 9 months saw their HDL-C continue
to increase. It should be noted that fasting plasma TG levels also fell
on average in this study which is the opposite of what the Stanford
group has repeatedly observed in short-term studies where both the high-fat
and high-CHO diets are fed isocalorically. The results of this and many
other studies make it clear that when a VLFNV diet, that is high in
fiber, is fed ad-libitum to patients at high risk of CAD, that the changes
in blood lipids are usually favorable. Indeed, in many patients such
a diet leads to regression of atherosclerotic plaque.
The
reason a lower HDL-C on a low-fat diet is not necessarily more atherogenic
may be because it has been shown that the fractional clearance rate
of cholesterol is different on a VLF diet than it is on a diet higher
in fat.[11]
This means that it is likely that the amount of cholesterol transported
back to the liver from the arteries may not be impaired on a high-CHO
diet despite a lower HDL-C level. The return of cholesterol from tissues
and blood to the liver is known as reverse cholesterol transport. In
animals this reverse cholesterol transport was not impaired despite
a much lower HDL-C on a high-CHO diet compared to a high-fat diet.[12]
So
while HDL-C does often fall initially when most people first adopt a
VLFNV diet, it is not clear that this lower HDL-C level necessarily
increases the risk of CAD. Furthermore, in most patients, the adoption
of a VLFNV diet will result in weight loss over the long run and this
weight loss will eventually result in HDL-C returning to baseline levels
(or even higher) levels in most patients.
Back
to top
It
appears to take some time for the body to adapt to a higher CHO intake.
The Stanford study, like all studies showing adverse effects on blood
lipids, lasted no more than a few weeks. Figure 1 below shows
what happened to fasting serum TG levels in a group of about 50 postmenopausal
women who were placed on a low-fat, high-CHO diet consisting largely
of whole foods.[13]
Figure 1. Effect of increasing dietary CHO at the
expense of fat on fasting plasma triglyceride levels in postmenopausal
women.
(Adapted
from Parks EJ. Am J Clin Nutr 2000;71:424)
During
the first 4 months of this study, dietary CHO gradually replaced dietary
fat in the diet but subjects were required to consume enough calories
to prevent weight loss. During this time fasting serum TG levels rose
from 151 to 204 mg/dl. The increase in fasting TG levels was less than
that observed by the Stanford study (113 vs 206 mg/dl). Part of the
reason for this greater rise in TG levels observed in the Stanford study
was that Stanford researchers used more refined high-CHO foods (which
raise TG levels more than natural high-CHO foods). Another reason was
that the Stanford study only kept their subjects on the experimental
diets for two weeks. Two weeks is not long enough for the body to fully
adapt to the higher CHO intake. Also, the subjects in this study lost
a little weight during the first 4 months of the study despite the researchers
best attempts to get them to maintain their initial body weights. Weight
loss tends to lower TG levels so even the loss of a few pounds can blunt
the TG raising effects of adopting a high-CHO diet.
During
the next 8 months of this study, the subjects continued to consume the
same high-CHO, low-fat (15% of energy) diet. During this phase, the
researchers no longer tried to control how much their subjects ate or
weighed. During this phase the subjects' calorie intake was ad libitum.
During this second phase of the study, average fasting TG levels gradually
returned to baseline level. Not surprisingly, in this 8-month period
the subjects lost another 4.5 pounds while consuming a self-selected
VLF diet (ad libitum). If most people were taught to consume a VLFNV
diet which consisted largely of natural foods, not only would their
blood lipids improve and their risk of heart disease fall but they would
also lose weight without any need to count calories.
Back
to top
The
new finding by the Stanford researchers was not that HDL-C dropped and
fasting TG increased on the higher CHO diet. They have shown this many
times before and so have other researchers using a similar research
protocol.[14]
[15] The
new finding was higher levels of RLP on the high-CHO diet compared to
the high-UFA diet. However, this finding is also probably simply an
artifact of a flawed experimental design. In general, higher RLP are
seen in people with higher fasting and postprandial TG levels and often
lower HDL-C as well. We have already seen that fasting TG and HDL-C
usually return to baseline levels if the higher CHO is fed ad libitum
and for a long enough time for the body to adapt to the higher CHO intake.
Two
studies have examined the potentially adverse metabolic effects of a
VLF diet (15% fat) compared to a moderate-fat (30% of energy) diet when
the VLF diet was fed either isocalorically with the higher fat diet
or ad libitum. Both studies found that it was only when the VLF diet
was fed at the same calorie level as the higher fat diet that the VLF
diet produced a potentially more atherogenic blood lipid profile.[16] [17]
It seems likely that it is only when subjects are required (by researchers)
to eat past satiety on a higher CHO diet in order to prevent weight
loss that the subjects are likely to experience much higher fasting
TG levels and also higher postprandial TG and RLP.[18] It should be noted that a high-CHO
diet, that is high in sugar and refined white flour and contains a high
calorie density, may not lead to a reduced energy intake and may be
detrimental for some people.[19]
Every
study published by the Stanford group and others which has shown detrimental
effects of high-CHO diets required subjects to consume the same energy
level and/or maintain the same body weight on both a high-fat and a
high-CHO diet. Since a diet consisting of more high-fat foods is generally
more calorie-dense, it would be expected to lead most people to consume
more calories than they would on a diet higher in CHO.[20] This is the main reason the results
of the Stanford group and others have little relevance to the real world
of clinical dietetic practice. This design flaw in their studies would
render their results clinically meaningless even if they should show
in the future that the presumably atherogenic changes in blood lipids
they have observed on diets higher in CHO really do promote atherosclerosis
and/or increase the risk of CAD.
Back
to top
A
crossover design study that compared the effects of a VLF diet to a
diet higher in fat on blood lipids may help put the Stanford study in
better perspective. In this study, a VLF (15%), high-CHO diet was compared
to a 30% fat diet. In this study the VLF diet was first fed isocalorically
with the more moderate-fat diet (as was done in all the Stanford studies).
However in this study, the same high-CHO diet was also fed ad libitum
to the same subjects.16
When
a VLF (15% of energy) diet was fed isocalorically with a moderate fat
diet, the fasting TG levels were much higher (188 vs. 115 mg/dl) on
the higher CHO diet than on the moderate fat diet. Just as the Stanford
researchers observed in their most recent study, the results of this
study also showed that postprandial TG (and presumably RLP) were also
much higher on the higher CHO diet than on the diet higher in fat (see
Figure. 2 below). HDL-C was also lower (42 vs. 35 mg/dl) in this study
on the higher CHO diet just as it was in the Stanford study. LDL-C
was somewhat higher on the VLF (134 vs. 128 mg/dl) than the moderate
fat diet when both diets were fed isocalorically. However, when the
VLF, high-CHO diet was fed ad libitum, the LDL-C was now lower (119
vs. 128 mg/dl) than on the 30% fat diet. Remarkably, this lower LDL
occured despite a much higher PUFA content (11.2% vs. 2.5% energy) and
polyunsaturated to saturated fat (P/S) ratio (1.6 vs. 0.5) on the 30%
fat diet compared to the VLF diet. The P/S ratio in the Stanford study
was also higher on their higher fat diet than on their higher CHO diet.
And while the fasting TG levels were still a little higher (130 vs 115
mg/dl) on the VLF diet when fed ad libitum compared to the moderate
fat diet, the postprandial TG level was already considerably lower on
the VLF diet compared to the 30% fat diet.
Figure 2. Effect of an AHA-Style Diet and a VLF
Diet (Fed either Ad Libitum or Isocalorically with the AHA-Style Diet)
on Serum TG Levels
As we have seen, an increased
fasting and postprandial TG level may be associated with more potentially
atherogenic RLP. As Figure 2 above clearly shows, even when fasting TG
levels are somewhat higher on a higher CHO diet, they may still be much
lower during most of the day. Because most people spend most of the day
in the postprandial state, it seems clinically more relevant to study
the impact of dietary changes on postprandial blood lipids rather than
just fasting blood lipid levels. This study clearly shows that a diet
higher in CHO can result in lower serum TG levels over a 2 year period
even though fasting levels may be somewhat higher. This is because the
postprandial rise in TG levels (and presumably RLP) will be much less
with an ad libitum VLF diet than it would be on an ad libitum diet with
added fats and oils. It seems likely that potentially atherogenic RLP
would also be lower on a VLF diet fed ad libitum but these were not measured
in this study. It should be clear that clinicians can't assume the higher
risk of CVD, often associated with higher fasting and postprandial TG
levels (in people consuming high-fat Western-style diets), would be comparable
to the risk of CVD with a similar fasting TG level in people consuming
a VLF diet. While it seems likely that RLP would fall along with fasting
and postprandial TG levels on a low-fat diet fed ad libitum over the long-term
this remains to be proven. This is an area that deserves more attention
from researchers.
Back
to top
It
appears that the presumably adverse impact of high-CHO diets on blood
lipids is limited primarily to short-term clinical trials in which the
research subjects' calorie intake is artificially manipulated (ie, controlled
by researchers) and the high-CHO diet is composed largely of sugar and
other refined CHOs with little fiber. This is the Stanford model. The
flaw in their experimental design was all explained to the Stanford
group in a letter to the editor.[19] Unfortunately,
Dr. Reaven's (point man for the Stanford group) replied, "In an
effort to make results meaningful, we maintained energy intake and output
constant throughout the study." Clearly he missed my main criticism
of their experimental design.[21] Apparently Dr. Reaven and other
academic researchers just don't get it. These and other researchers
should compare a high-CHO diet consisting largely of natural foods high
in fiber, like fruits, vegetables, whole grains, and beans, with a similar
diet to which olive oil or other unsaturated oils are added. They should
also allow their research subjects to consume both diets ad libitum
(rather than imposing artificial controls on how much people eat). Then
they would likely find that a VLFNV diet does not produce adverse changes
in blood lipids that increase the risk of CAD as they imply. Indeed,
any changes in blood lipids that such a diet causes must be viewed as
favorable simply because such a diet has been proven to reverse atherosclerosis.
High-fat diets have not been shown to regress atherosclerosis and are
usually associated with its progression.
Blood
lipids are simply the messenger about what is going on in the artery
wall. The message that patients are looking for from their health professionals
is how does one prevent or reverse atherosclerosis and CAD. The Stanford
researchers appear to be so preoccupied with the messenger (short-term
changes in fasting blood lipids which don't always get the message straight)
that they have lost sight of the message. New medical technology may
soon make it possible to accurately measure the health and function
of arteries. Research should now be shifting from what is happening
to fasting blood lipids to what is happening to the arteries themselves.
Back to top
References
[1]
Abbasi F, McLaughlin T, Lamendola C, et al. High carbohydrate diets,
triglyceride-rich lipoproteins, and coronary heart disease risk. Am
J Cardiol 2000;85:45-8
[2] Rutledge JC, Hyson DA, Garduno D, et
al. Lifestyle modification program in management of patients with
coronary artery disease: the clinical experience in a tertiary care
hospital. J Cardiopulm Rehabil 1999;19:226-34
[3] Ornish D, Brown SE, Scherwitz LW, et
al. Can lifestyle changes reverse coronary artery disease? Lancet
1990;336:129-33
[4] Ornish D. Serum lipids after a low-fat
diet. JAMA 1998;279:1345-6
[5] Ornish D, Sherwitz LW, Billings JH,
et al. Intensive life style changes for reversal of coronary heart
disease. JAMA 1998;280:2001-7.
[6] Schueler G, Hambrecht R, Schrief G,
et al. Regular physical exercise and a low-fat diet. effects on progression
of coronary artery disease. Circulation 1992;86:1-11
[7] Morrison L. Diet and coronary atherosclerosis.
JAMA 1960; 173:884-8
[8] Castelli WP. The triglyceride issue:
a view from Framingham. Am Heart J 1986;112:432-40
[9] Rudel LL, Parks JS, Sawyer JK. Compared
with dietary monounsaturated and saturated fat, polyunsaturated fat
protects African green monkeys from coronary artery atherosclerosis.
Arterioscler Thrombo Vasc Biol 1995;15:2101-10
[10] Thuesen L, Henriksen LB, Engby B.
One-year experience with a low-fat, low-cholesterol diet in patients
with coronary heart disease. Am J Clin Nutr 1986;44:212-9
[11] Brinton EA, Eisenberg S, Breslow
JL. A low-fat diet decreases high density lipoprotein(HDL) cholesterol
levels by decreasing HDL apoprotein transport rates. J Clin Invest
1990;85:144-51
[12] Woolett LA, Kearney DM, Spady DK.
Diet modification alters plasma cholesterol concentrations but not
the transport of HDL cholesteryl esters to the liver in hamsters.
J Lipid Res 1997;38:2289-302
[13] Kasim-Karakas SE, Lane E, Almario
R, et al. Effects of dietary fat restriction on particle size of plasma
lipoproteins in postmenopausal women. Metabolism 1997;46:431-6
[14] Garg A, Bantle JP, Henry RR, et al.
Effects of varying carbohydrate content of diet in patients with non-insulin-dependent
diabetes mellitus. JAMA 1994;271:1421-8
[15] Jeppensen J, Schaaf P, Jones C, et
al. Effects of low-fat, high-carbohydrate diets on risk factors for
heart disease in postmenopausal women. Am J Clin Nutr 1997;55:1027-33
[16] Lichtenstein AH, Ausman LM, Carrasco
W, et al. Short-term consumption of a low-fat diet beneficially affects
plasma lipid concentrations only when accompanied by weight loss.
Arterioscler Thromb 1994; 14:1751-60
[17] Schaefer EJ, Lichtenstein AH, Lamon-Fava
S, et al. Body weight and low-density lipoprotein changes after consumption
of a low-fat ad libitum diet. JAMA 1995; 274:1450-5
[18] Parks EJ, Hellerstein MK. Carbohydrate-induced
hypertriacylglycerolemia: historical perspective and review of biological
mechanisms. Am J Clin Nutr 2000;71:412-33
[19] Kenney JJ. Low-fat, high-carbohydrate
diets and risk for ischemic heart disease. Am J Clin Nutr 1997;66:1293
[21] Reaven G. Reply to J Kenney. Am
J Clin Nutr 1997;66:1294
Please note: It is your responsibility to report your CPE hours to ADA and your state licensing department. Proof of completion will be mailed to you as soon as we grade your test. Please retain it for reporting and your records.
