1059
Will a High-Carbohydrate, Low-Fat Diet
Lower Plasma Lipids and Lipoproteins
Without Producing Hypertriglyceridemia?
Daniel Ullmann, William E. Connor, Lauren F. Hatcher,
Sonja L. Connor, and Donna P. Flavell
Downloaded from http://atvb.ahajournals.org/ by guest on June 15, 2017
A sudden increase in dietary carbohydrate invariably increases the plasma levels of very low
density lipoprotein (VLDL) and triglyceride. The present studies were designed to test the
hypothesis that dietary carbohydrate-induced hypertriglyceridemia need not occur. In the first
study we fed gradually increasing amounts of carbohydrate and gradually decreasing amounts
of fat to eight subjects. The usual American diet (40% fat, 45% carbohydrate, and 15% protein)
was followed in sequence by four diets in a phased regimen, the carbohydrate increasing by 5%
of total calories and the fat content decreasing by 5% for each dietary period. In the last dietary
period (phase 4), 20% of the energy was in the form of fat and 65% in the form of carbohydrates;
the cholesterol content was 100 mg/day. Throughout the study, plasma triglyceride and VLDL
triglyceride levels did not change significantly. The plasma total and low density lipoprotein
(LDL) cholesterol levels were greatly reduced, by 15% and 22%, respectively (p=0.004). Plasma
high density lipoprotein (HDL) cholesterol levels decreased concomitantly. In the second study,
after a washout period six of the subjects were initially fed the phase 4 high-carbohydrate diet
for a 10-day period. The plasma triglyceride concentration increased over baseline levels by
47%, and VLDL triglyceride levels increased by 73%. We conclude that although a sudden
increase in dietary carbohydrate increases the plasma triglyceride level, patients gradually
introduced to a high-carbohydrate, low-fat diet may achieve a significant reduction of plasma
total and LDL cholesterol without developing carbohydrate-induced hypertriglyceridemia.
(Arteriosclerosis and Thrombosis 1991;ll:1059-1067)
C
oronary heart disease in Western populations
has surged from a minimal public health
concern at the turn of the century to the
leading cause of mortality.1 Cross-cultural studies
have documented a strong positive association between dietary fat intake and both the prevalence and
incidence of this disease, manifested by myocardial
infarction and sudden death.2 Coronary heart disease
From the Division of Endocrinology, Diabetes, and Clinical
Nutrition of the Department of Medicine, and the Department of
Preventive Medicine, Oregon Health Sciences University, Portland, Ore.
Presented in part (and published in abstract form in Circulation
1988;78[suppl II]:II-73) at the Annual Scientific Sessions of the
American Heart Association, Washington, D.C., 1988.
Supported by grants HL-37940 from the National Heart, Lung,
and Blood Institute, RR-00334 from the General Clinical Research Center of the Division of Research Resources, National
Institutes of Health, and AHA 70172-1 from the American Heart
Association, Oregon Affiliate.
Address for reprints: Dr. Daniel Ullmann, Oregon Health
Sciences University, Division of Endocrinology, Diabetes, and
Clinical Nutrition, Department of Medicine, L-465, 3181 SW Sam
Jackson Park Road, Portland, OR 97201-3098.
Received November 13, 1990; revision accepted April 18, 1991.
in the United States is related to elevated plasma
levels of cholesterol and low density lipoprotein
(LDL), which usually result from consumption of
excessive amounts of dietary saturated fat and cholesterol.3 Accordingly, there has been general agreement in the scientific community that the primary
dietary change required to reduce cardiovascular risk
would be a decrease in the intake of total fat,
saturated fat, and cholesterol.4"8 To replace the fat
calories as fat intake is reduced, it is then necessary
to increase carbohydrate intake because protein intake remains at 15% of total calories. Currently, the
average carbohydrate intake by Americans is 45% of
total calories.9 In contrast, current dietary recommendations for individuals with hyperlipidemia and
even for those with diabetes are diets with carbohydrate contents of as much as 65% of total calories,
with the emphasis on complex carbohydrates from
natural sources.6"8 These recommendations are consistent with the high-carbohydrate and low-fat intakes of populations having a low incidence of coronary heart disease.1011
Because past studies have shown that suddenly
increasing dietary carbohydrate can produce delete-
1060
Arteriosclerosis and Thrombosis
Vol 11, No 4 July/August 1991
TABLE 1. Clinical Characteristics of Study Subjects
Subject
1
2
3
4
5
6
7
8
Mean±SD
Age (yr)/
sex
59/F
44/M
51/M
63/M
36/M
44/M
51/M
60/F
51±9
Body
weight
(kg)
81.2
86.9
88.8
67.2
96.0
98.7
98.7
61.6
84.9±14.1
Body mass
index*
31.3
30.1
28.7
24.2
33.9
32.2
31.9
24.1
29.5 ±3.7
Total
cholesterol
279
232
260
240
140
255
191
226
228±44
Entry lipoprotein values (mg/dl)t
HDL
VLDL
LDL
cholesterol
cholesterol
cholesterol
30
197
59
32
96
108
35
197
39
48
25
167
37
21
72
47
164
48
113
42
45
154
46
44
43±24
147±45
44±8
Total
triglyceride
173
576
182
140
301
264
214
223
254±137
*Body mass index is the weight in kilograms divided by the square of the height in meters.
|To convert cholesterol and triglycerides to mmol/1, divide by 38.7 and 88.5, respectively.
Downloaded from http://atvb.ahajournals.org/ by guest on June 15, 2017
rious effects in normal,12 hyperlipidemic,13 and diabetic14 individuals, there is no universal agreement
that the American diet should be higher in carbohydrate. Some undesired effects of a high-carbohydrate
diet include hypertriglyceridemia, postprandial hyperglycemia, hyperinsulinemia, and a decrease of
high density lipoprotein (HDL) levels.15 Because
prospective studies have shown that all four of these
risk factors are associated with an increased incidence of cardiovascular mortality,16"19 it has been
argued that replacement of fat with carbohydrate
may not be totally benign. However, most studies of
the "carbohydrate induction" of hypertriglyceridemia have suddenly increased the amount of dietary
carbohydrate, and most have been quite short in
duration.12-1520 Under different conditions, hypertriglyceridemia after a change to a higher carbohydrate intake may not always occur or persist. Previously, we demonstrated that a marked decrease in
the concentration of plasma cholesterol without hypertriglyceridemia could occur in patients with insulin-dependent diabetes mellitus 1 year after initiation
of a low-fat, high-carbohydrate diet.21 The present
studies were designed to test the hypothesis that
carbohydrate-induced hypertriglyceridemia need not
occur due to a high-carbohydrate diet that was
gradually phased in over time.
Study Protocol
Our null hypothesis was that dietary carbohydrateinduced hypertriglyceridemia would not occur, provided that the carbohydrate content of the diet was
increased gradually as the fat content was decreased.
The studies were designed to test these effects of a
high-carbohydrate, low-fat diet on the plasma concentrations of triglyceride, cholesterol, and lipoproteins. There were two components to the study. In
the first component, the dietary carbohydrate was
phased in gradually as the fat content was reduced.
In the second component, we ascertained that the
patients would actually become hypertriglyceridemic
in response to a short-term high-carbohydrate-diet
challenge without the gradual phased approach.
Methods
Subjects
These studies were conducted at the Clinical Research Center at the Oregon Health Sciences University. Informed consent was obtained in accordance with institutional policy, and the protocol was
approved by the Human Research Committee. The
study group comprised eight healthy nondiabetic
adult individuals (two women and six men). Patients
were recruited from the Lipid Clinic at the Oregon
Health Sciences University. Their ages, weights, body
mass indexes, and plasma lipid levels on entry are
shown in Table 1. Their mean age was 51 years
(range, 36-63 years). Their body weights and body
mass indexes averaged 85 kg and 29.5, respectively,
Study 1: Phased Carbohydrate Diet
In study 1 we measured plasma lipids and lipoproteins when the dietary carbohydrate intake was phased
in gradually as fat was reduced. With eight patients, this
study has a statistical power of 0.80 to detect a plasma
triglyceride change of 75 mg/dl, which reflects an
approximate 1.5 standard deviation shift of the mean,
when a significance value of 0.05 or less is used (twotailed test). All patients were sequentially fed five diets
designed to maintain body weight: the typical American
diet (40% fat, 45% carbohydrate, and 15% protein)
and four diets that were each 5% lower in fat and 5%
higher in carbohydrate than the preceding diet. The
final diet (phase 4) contained 65% of calories as
carbohydrate and 20% of calories as fat. The dietary
which indicated mild overweight. On entry into the
study, while patients 1, 2, 3, 4, 6, and 8 had elevated
plasma cholesterol concentrations (>220 mg/dl), patients 2, 5, 6, 7, and 8 had elevated triglyceride levels
(>200 mg/dl). Thus, on average these patients were
mildly hypertriglyceridemic, with an average plasma
triglyceride level of 254 mg/dl. None had symptoms
of cardiac failure, a history of a recent myocardial
infarction, or disease of the gastrointestinal tract,
kidneys, or endocrine system, and none were taking
any medications known to affect plasma lipids.
Ullmann et al Phased Dietary Carbohydrate and Plasma Triglyceride
TABLE 2.
Composition of the Phased-Carbohydrate Diets*
Phases
Composition
American
diet
(control) 1
2
3
Carbohydrate (% of total calories)
Total
45
50 55 60
17
25 30 35
Complex
Simple sugars (sucrose, glucose,
28
25 25 25
fructose, and galactose)
40
35 30 25
Total fat (% of total calories)
6
15
11
8
Saturated
6
Polyunsaturated
8
8
8
Monounsaturated
19
16 14 11
1
0.4
0.7 1.0 1.3
P to S ratio
Cholesterol (mg/1,000 kcal)
179
143 107 71
26
19 14 10
CSI (per 1,000 kcal)t
4
65
40
25
20
5
8
7
1.6
36
7
Downloaded from http://atvb.ahajournals.org/ by guest on June 15, 2017
P, polyunsaturated; S, saturated.
'Protein is 15% of total calories in each diet.
tCholesterol-saturated fat indexZ3=(1.01xg saturated fat)+
(0.05 mg cholesterol).
periods had a duration of 10 days each, in accordance
with previous evidence that the rise in the plasma
triglyceride level from a high-carbohydrate diet is
largely completed within 10 days.20
Study 2: Carbohydrate Induction
Study 2 was designed to document the physiological response of carbohydrate-induced hypertriglyceridemia that occurs when the amount of dietary
carbohydrate is suddenly increased. The phase 4 diet
(65% carbohydrate, 20% fat, and 15% protein) was
fed acutely to six subjects (subjects 1,2,3,4,7, and 8)
for a 10-day period after a washout period of at least
6 weeks. The values obtained were compared with
the control American diet values, which were also the
bases for comparisons in study 1.
Throughout the study, body weight was measured
daily, with the subject clothed and shoeless, on a beamtype balance scale calibrated with standard weights.
Diets
The research diets were prepared at the Clinical
Research Center and were composed of typical
mixed foods. The diets were eucaloric for each
patient to keep body weight constant. The energy
expenditure for each patient was estimated by the
Boothby-Berkson nomogram,22 incorporating additional calories as needed to adjust for physical activity. The patients were instructed to keep energy
expenditure from physical activity constant to prevent any body weight fluctuations.
The five experimental diets were fed sequentially
for 10 days each in study 1. Only one diet (phase 4)
was fed in study 2. The composition of the five
different eucaloric diets is presented in Table 2. The
1061
initial diet was planned to simulate the composition
of the usual American diet.9 It provided 40% of the
total energy as fat (15% saturated, 6% polyunsaturated, and 19% monounsaturated fatty acids) and
45% as carbohydrate (17% complex and 28% simple
carbohydrates). The four subsequent diets had a
gradual increase in carbohydrate content to 65% in
phase 4, the last dietary period. The phase 4 diet
contributed 20% of the energy as fat (5% saturated,
8% polyunsaturated, and 7% monounsaturated fatty
acids) and 65% of energy as carbohydrate (40%
complex and 25% simple carbohydrate). The diets
ranged from 179 to 36 mg cholesterol per 1,000 kcal,
respectively. The phase 4 diet was thus low in fat,
saturated fat, and cholesterol. The association between coronary atherosclerosis and dietary cholesterol and saturated fat can be denoted by the cholesterol-saturated fat index,23 which was 26 per 1,000
kcal for the American diet and only 7 for the phase 4
diet. Protein intake remained constant at 15% of
energy. No alcohol was permitted, and coffee and tea
were limited to a constant four cups per day.
During the feeding studies the sources of fat were
commercial margarines, oils, meats, and dairy products. The composition of dietary fats was adjusted by
changing the types and amounts of margarines and
oils used, while saturated fat sources (beef, cheese,
whole milk, and chocolate) were greatly reduced over
time. Dietary cholesterol was altered by decreasing
the amount of meat and poultry provided daily, as
well as by eliminating egg yolk and butterfat in milk
and cheese. Dietary complex carbohydrates along
with fiber were increased by using whole grains
rather than refined grain products, as well as by
including a larger number of servings of fruits, vegetables, and legumes. We calculated that the total
fiber increased from 17.5 to 50.9 g/day, and the
soluble fiber increased from 5.8 to 14.0 g/day for a
3,200-kcal diet. Of note was the fact that the simple
sugar content of the diet was held constant.
All diets were adequate in all essential nutrients in
accordance with the National Academy of Sciences
Recommended Dietary Allowances.24 The food
sources were purchased in bulk to minimize any
variability in composition. The absolute amounts of
cholesterol, dietary fiber, sodium, and potassium
were calorically related.
Biochemical Analysis
Blood samples were collected into tubes containing 0.1% EDTA, after a 14-hour fast, on days 4, 7,
and 10 when each dietary period was completed. The
plasma specimens were analyzed for total cholesterol, triglyceride, and lipoprotein concentrations.
The laboratory procedures for the plasma lipid and
lipoprotein determinations were in compliance with
the standardization and surveillance programs of the
Centers for Disease Control Laboratory in Atlanta,
Ga., according to procedures established by the Lipid
Research Clinics of the National Heart, Lung, and
Blood Institute.25 Plasma total cholesterol and tri-
1062
Arteriosclerosis and Thrombosis
Vol 11, No 4 July/August 1991
Downloaded from http://atvb.ahajournals.org/ by guest on June 15, 2017
glyceride concentrations were measured fluorometrically with an AutoAnalyzer II (Technicon Instruments Corp., Tarrytown, N.Y.). Plasma HDL
cholesterol was measured enzymatically after precipitation of the apolipoprotein (apo) B-containing
lipoproteins by heparin-MnCl2. The very low density lipoprotein (VLDL) fraction was separated by
ultracentrifugal flotation {d< 1.006 g/ml) to yield a
supranatant containing VLDL. LDL cholesterol
was determined by difference, that is, cholesterol in
the infranatant fraction after ultracentrifugation
(HDL plus LDL) minus HDL cholesterol equals
LDL cholesterol.
In study 1 in addition to lipid and lipoprotein
analyses, plasma apos A-I, B, C-III, and E were also
assessed in fasting blood samples collected at baseline and at the end of phase 4. The quantitative
determinations of apolipoproteins were performed
by electroimmunoassay procedures at the Oklahoma
Medical Research Foundation.26
Total Cholesterol
p <0.0001
Statistical Analysis
Means and standard errors of the mean were
calculated for all variables at the end of each of the
10-day dietary periods. Baseline plasma lipid and
lipoprotein values were compared with the mean of
the lipid and lipoprotein values in each dietary
period by repeated-measures analysis of variance
(ANOVA) calculated according to the general linear model procedure of SAS.27 When the ANOVA
was significant (/><0.05), contrasts were used to
compare differences between the control and the
different dietary periods of interest with the use of
Bonferroni's corrected paired / tests.27 We report
both the probability values for the ANOVA as well
as the corrected probability values for comparisons
with the control diet. Exact probability values are
reported. Apolipoprotein data analyses were performed by comparing the apolipoprotein values at
baseline with their respective levels at the completion of the study, or phase 4, by use of two-tailed
paired t tests. Two-tailed paired t tests were also
used to analyze the effects of the high-carbohydrate
diet during the acute challenge.
Results
Study 1
All patients consumed only the food provided to
them. Adherence to the dietary intervention was
excellent, and the high-carbohydrate, low-fat diets
did not cause side effects in any patient. Body weights
did not change during the different dietary periods.
At entry, these patients had an average body weight
of 84.9 ±14 kg. On completion of the study, their
average body weight remained unchanged (85.0±14
kg). Individual responses to each dietary period for
levels of plasma cholesterol, LDL cholesterol, HDL
cholesterol, triglyceride, and VLDL triglyceride are
presented in Figures 1 and 2. Mean levels of plasma
lipids and lipoproteins and the ratios of total and
Control
%CHO 45
FIGURE 1. Line plots showing influence of the phased
high-carbohydrate (CHO) diet on plasma concentration of
total cholesterol, low density lipoprotein (LDL) cholesterol,
and high density lipoprotein (HDL) cholesterol (all in mmol/l
and mg/dl) in eight subjects. Each data point represents values
from individual subjects obtained on the 10th day of each
dietary phase. Probability values denote overall significance by
repeated-measures analysis of variance.
LDL cholesterol to HDL cholesterol are shown in
Table 3.
Compared with the values during consumption of
the control lower-carbohydrate diet (45%), the
plasma levels of triglyceride did not change significantly during the high-carbohydrate diet at any time
point (Table 3). The even more sensitive VLDL
triglyceride levels also remained unchanged. The
plasma triglyceride levels were 213 mg/dl during the
Ullmann et al Phased Dietary Carbohydrate and Plasma Triglyceride
drate diet further reduced the plasma total cholesterol from 232 to 198 mg/dl, a 15% decline
(p=0.0036). Significant changes occurred in LDL
cholesterol levels in all dietary periods compared
with baseline. LDL cholesterol decreased from 161
to 144 mg/dl (phase 1) to 141 mg/dl (phase 2), to 134
mg/dl (phase 3), and to 126 mg/dl at the end of the
study, a 22% reduction (p=0.001).
HDL cholesterol decreased from 43 to 36 mg/dl at
the completion of the study, or phase 4 (16%).
However, after adjustment for post hoc comparisons,
this decrease was not statistically significant. No
significant difference occurred in the ratios of total or
LDL cholesterol to HDL cholesterol.
Although the high-carbohydrate diet did not produce significant differences in plasma concentrations
of apolipoproteins (Table 4) between the control and
the final dietary period, the apolipoprotein concentrations of A-I, A-II, and B tended to decrease.
Those decreases were concurrent with the significant
declines in HDL and LDL cholesterol levels.
Triglycerides
NS
Downloaded from http://atvb.ahajournals.org/ by guest on June 15, 2017
Control
%CHO 45
I
50
II
55
60
IV
65
FIGURE 2. Line plots showing effects of the high-carbohydrate (CHO) low-fat diet on the plasma total and very low
density lipoprotein triglyceride (VLDL-TG) concentrations in
eight subjects (all in mmolll and mg/dl). NS, nonsignificant
differences.
control diet and 230 mg/dl after the phase 4 highcarbohydrate diets (p=0.70). VLDL triglyceride reflected a similar pattern: 147 mg/dl while patients
were on the control diet and 157 mg/dl after the
high-carbohydrate diet (p=0,92). Furthermore, no
significant differences were noted in the plasma
levels of VLDL cholesterol.
When the plasma cholesterol levels during phases
1 and 2 were compared with the levels during the
low-carbohydrate period (45%), no significant
change was noted for the first two dietary periods.
However, by the end of phase 3 (60% carbohydrate),
plasma cholesterol levels decreased from 232 to 209
mg/dl (p=0.04), and phase 4 of the high-carbohyp<0.05
400
Triglycerides
VLDL-TG
T
1063
300
200
100
J 0
Percent
CHO
FIGURE 3. Bar graph showing response to a high-carbohydrate (CHO) intake (acute challenge) on the plasma triglyceride (•) and very low density lipoprotein triglyceride (VLDLTG, ®) concentrations (in mmolll and mg/dl) in six subjects
fed for a 10-day period.
Study 2
The effects of a sudden increase in dietary carbohydrate for each patient are depicted in Table 5 and
Figure 3. Body weights did not change during this
dietary period. During the acute challenge of the
high-carbohydrate diet (phase 4) to six of the subjects
(subjects 1, 2, 3, 4, 7, and 8), total triglyceride
concentration increased significantly, from 204 to 296
mg/dl, a 47% increase. VLDL triglyceride levels
increased concurrently, from 137 to 215 mg/dl, a 73%
increase.
Although total cholesterol concentrations had
not attained a steady state, we observed that while
the VLDL cholesterol increased by 46% 10 days
after this high-carbohydrate challenge, the LDL
cholesterol decreased from 174 to 143 mg/dl, an
18% decline (/?<0.0001), indicating a shift in the
distribution of the plasma cholesterol from LDL to
VLDL. HDL cholesterol remained unchanged at
the end of this 10-day period.
Discussion
Carbohydrate-induced hypertriglyceridemia was
prevented in eight male and female subjects by
gradually increasing the dietary carbohydrate intake
over time. Typically, when the amount of dietary
carbohydrate is suddenly increased, hypertriglyceridemia invariably results within a few days,12"15-20-28 a
fact also documented in the acute challenge (study 2)
by the 65% carbohydrate, 20% fat diet of our study.
This result indicated the susceptibility of our subjects
to the phenomenon of carbohydrate-induced hypertriglyceridemia, a physiological response: the higher
the initial plasma triglyceride concentration, the
greater the hypertriglyceridemia that ensued.20 The
mechanism of this effect is enhanced synthesis of
triglyceride and VLDL by the liver.29 In addition, a
high-carbohydrate diet produces both greater num-
1064
Arteriosclerosis and Thrombosis
Vol 11, No 4
July/August 1991
TABLE 3.
Plasma Lipid and Lipoprotein Levels* After Different Dietary Phases of the High-Carbohydrate Diets
Phases
2
3
4
45
50
55
60
65
232±15
40+9
223 + 15
43±7
144+16§
41±3
216±15
41 ±6
141 + 16||
44+4
5.1±0.4
209±14t
41±5
0.0001
0.9280
0.7042
Percent of calories
as carbohydrate
Cholesterol
Total
VLDL'
LDL*
HDL*
Total/HDL
LDL/HDL
161±17
43±4
5.8±0.7
Triglyceride
Total
VLDL
p values for
ANOVA
Control
1
4.1±0.7
5.6±0.5
3.6±0.4
213±38
147±39
232 ±37
169±42
3.3±0.3
5.1±0.2
3.2±0.3
198±13*
42±7
126±15#
36±3
5.6±0.4
3.6±0.4
237±30
171 ±37
230±37
171±46
230±35
157±37
134±14H
42±3
0.0001
0.0190
0.2445
0.0888
0.9280
Downloaded from http://atvb.ahajournals.org/ by guest on June 15, 2017
VLDL, very low density lipoprotein; LDL, low density lipoprotein; HDL, high density lipoprotein.
*A11 values are in mg/dl and are mean±SEM.
Bonferroni corrected paired t tests: t/>=0.0400, 45% vs. 60% carbohydrate values; $/>=0.0036, 45% versus 65% carbohydrate values;
§/?=0.0108, 45% vs. 50% carbohydrate values; ||p=0.0376,45% vs. 55% carbohydrate values; 11^=0.0048,45% vs. 60% carbohydrate values;
#p=0.0004, 45% vs. 65% carbohydrate values.
bers of VLDL particles and larger particles that are
relatively richer in triglyceride content.29'30
Using a diet high in carbohydrate content, Melish
and colleagues29 reported that the VLDL triglyceride
synthetic rate increased; there was, however, no
increase in VLDL apo B production rate, and thus,
increased rates of secretion of VLDL triglyceride in
the setting of constant VLDL apo B secretion must
result in nascent VLDL particles with a high ratio of
triglyceride to apo B. We routinely used a standard
high-carbohydrate diet to induce hypertriglyceridemia. In one of our published studies,20 plasma
triglyceride increased even more (85%) in response
to a high-carbohydrate formula diet. This response
was completely blocked by the concurrent ingestion
of co-3 fatty acids from fish, which inhibit the hepatic
TABLE 4. Apolipoprotein Changes After a High-Carbohydrate
Diet
Variable
Apolipoprotein
Apolipoprotein
Apolipoprotein
Apolipoprotein
Apolipoprotein
A-I
A-II
B
C-III
E
Entry
65% CHO diet
139±6.0
57+2.5
90±10.5
14±2.1
13±2.1
124±13.4
49±4.5
76+8.1
14+2.0
13±1.4
Values are in mg/dl and are mean±SEM.
CHO, carbohydrate.
synthesis of triglyceride.31 In our present study, while
the acute feeding of a high-carbohydrate diet produced hypertriglyceridemia, the gradual phased-carbohydrate approach presumably allowed metabolic
adaptation to occur without the usual stimulus of a
high-carbohydrate intake to enhance VLDL and
triglyceride synthesis.29
In a historical context, Ahrens et al28 published
their classic article in 1961 demonstrating that subjects with elevated plasma triglyceride levels developed further increases when acutely fed a highcarbohydrate diet. They introduced the concept of
"carbohydrate-induced lipemia" and postulated that
this was a common phenomenon, "especially in the
areas of the world distinguished by caloric abundance
and obesity." Antonis and Bersohn32 showed that
changing from a low- to a high-carbohydrate diet
caused a rise in fasting serum triglyceride concentration, regardless of the type of fat consumed in the
preceding regimen. The difference in carbohydrate
content between the two diets corresponded to 25%
of the total caloric intake. The average concentration
of serum triglyceride reached a maximum, about
double the starting value, 3-5 weeks after the dietary
change. After 32 weeks on the high-carbohydrate
diet, most of the subjects had triglyceride levels
similar to initial values. A substantial degree of
TABLE 5. Effects of a High-Carbohydrate Diet on Plasma Lipid and Lipoprotein Levels in Study 2 (Acute Challenge)
Cholesterol (mg/dl)
Variable
Control diet
65% CHO diet
Percent A
P
Triglycerides (mg/dl)
Total
VLDL
LDL
HDL
Total
VLDL
LDL
HDL
243 ±12
233 ±13
-4
NS
39±11
50±12
+46
0.02
174±15
143±13
-18
0.0001
43±6
41±3
-2
NS
204±48
296±75
+47
0.0001
137±49
215+78
+73
0.01
57±11
10±l
7±2
67±6
+29
NS
-34
NS
Values are in mg/dl and are mean±SEM.
VLDL, very low density lipoprotein; LDL, low density lipoprotein; HDL, high density lipoprotein; CHO, carbohydrate; A, change.
Ullmann et al Phased Dietary Carbohydrate and Plasma Triglyceride
Downloaded from http://atvb.ahajournals.org/ by guest on June 15, 2017
variability in the individual responses was observed;
some of the subjects showed only moderate increases
in triglyceride concentration, while others developed
marked hypertriglyceridemia. This study supported
the concept that carbohydrate-induced hypertriglyceridemia may ultimately be resolved in some subjects.
However, Reaven and colleagues14 found that the
elevated plasma triglyceride from a high-carbohydrate diet did not resolve after 6 weeks.
Fasting plasma triglyceride levels in free-living
populations accustomed to receiving 65% or more of
their calories from carbohydrates are not profoundly
higher than those seen in Western populations.
Tarahumara Indians of Mexico consume 75% of
their calories as carbohydrate from corn and
beans33-34 and have mean plasma triglyceride levels of
150 mg/dl and very low plasma cholesterol levels of
133 mg/dl. It is apparent from these studies that diets
nutritionally adequate in every respect may contain a
very high carbohydrate content and yet result in
triglyceride levels similar to those observed in Western populations, but with greatly lower plasma cholesterol levels as in the Tarahumaras.
Another significant advantage of the low-fat, highcarbohydrate diet used in this study would be a great
reduction in postprandial triglyceride-rich lipoproteins, which are not measured in fasting plasma;
these include chylomicrons and their remnants. Such
postprandial particles are certainly considered
atherogenic35 and should be avoided whenever possible. Our phase 4 diet with an energy intake of 2,800
kcal and 20% fat content would contain only 50 g fat.
This would clearly generate only half as much postprandial triglyceridemia as would a diet containing
100 g fat, the typical American norm. Even a high-fat
diet of monounsaturated fat, which might lower LDL
levels, would still generate large amounts of postprandial particles. Fish oil, however, did inhibit postprandial lipemia.36
In addition to preventing hypertriglyceridemia in
study 1, the concomitant gradual reduction of dietary
saturated fat and cholesterol greatly lowered the
total and LDL cholesterol levels. On average, plasma
total and LDL cholesterol levels were 15% and 22%
lower, respectively, after the high-carbohydrate diet
when compared with their control baseline diet values. Furthermore, significant changes in mean
plasma cholesterol levels were seen even at the
completion of phase 3 (60% carbohydrate, 25% fat,
and 15% protein). The dietary composition of this
phase is similar to the recent recommendation by the
National Cholesterol Education Program Expert
Panel on Dietary Therapy of Hypercholesterolemia.6
The Bonferroni adjusted t tests showed that LDL
concentrations actually decreased significantly for
each phase of the study. Because emphasis has
recently been placed on the particular atherogenicity
of LDL, an apo B-100-rich lipoprotein, the decreases
of LDL levels obtained in this study are especially
pertinent. The major benefit of the high-carbohydrate diet was that it was possible to greatly reduce
1065
total and saturated fat as well as dietary cholesterol.
The biochemical mechanism whereby dietary cholesterol and saturated fat elevate plasma LDL levels is
through their effect on the LDL receptor in the liver.
These dietary factors downregulate LDL receptor
activity by decreasing the synthesis of LDL receptor
A convenient way to express the action of dietary
cholesterol and saturated fat on the LDL receptor
would be through a single number, the cholesterolsaturated fat index developed from regression equations.23 The cholesterol-saturated fat index is 26 per
1,000 kcal for the American diet and only 7 for the
phase 4, 20% fat diet. Consequently, the high-carbohydrate, low-fat diet would permit the expression of
maximal LDL receptor activity for the removal of
LDL from the plasma.39 Synthesis of cholesterol in
the liver is probably also reduced with less dietary
saturated fat.40
We cannot exclude the possibility that some mild
plasma total and LDL cholesterol lowering seen with
the high-carbohydrate diet may have resulted from the
soluble fiber, saponins, and other substances in plants,
but we judge these effects not to have been great.41
With regard to dietary fiber effects on plasma triglyceride levels, a high-fiber diet did not lower triglyceride
levels when body weight was maintained.42
The effects of the high-carbohydrate diet on
plasma LDL cholesterol during the acute challenge
reflect a downward shift, although the effect of the
high-carbohydrate diet may not have been fully appreciated because this dietary period had a duration
of 10 days, which is insufficient to attain stability in
cholesterol and LDL cholesterol levels. However,
stability for plasma triglyceride and VLDL probably
occurred because of their shorter turnover time.
Even for plasma LDL and total cholesterol, it should
be indicated that the phased approach (study 1) did
take 40 days, and relative stability of their values had
occurred by day 30 of the study.
With the high-carbohydrate diet in study 1, the
plasma level of HDL cholesterol had decreased 16%
by the end of the study. This is an expected response
to a low-fat diet; other metabolic studies have shown
similar responses.153243 It appears that the decrease
in HDL cholesterol is due in part to the enhanced
catabolic rate for apo A-I. Furthermore, in a recent
report Katan and associates44 described the effects in
populations consuming a high-carbohydrate diet over
a lifetime. They observed lower concentrations of
HDL cholesterol, and these lower concentrations
were not associated with an increased risk of coronary heart disease because the diets were low in fat
and cholesterol content and the resultant plasma
LDL cholesterol levels were also low.44 This concept
is further supported by the absence of an inverse
association between HDL cholesterol concentration
and the risk of coronary heart disease among Tarahumara Indians,33 Masai,4S vegetarians,46 and others
following low-fat, high-carbohydrate diets.11
1066
Arteriosclerosis and Thrombosis
Vol 11, No 4 July/August 1991
Downloaded from http://atvb.ahajournals.org/ by guest on June 15, 2017
The HDL changes in study 1 were significant in that
a low HDL cholesterol level is commonly regarded as
an important risk factor for the development of coronary heart disease.19 Low HDL concentrations have
been noted in all population studies in which the
dietary fat content is 20% or less of the total caloric
intake. A classic population that has undergone many
metabolic studies is the Tarahumara Indians of Chihuahua Province in Mexico.34 Their plasma total cholesterol levels are 133 mg/dl, LDL cholesterol, 87
mg/dl, and HDL cholesterol concentrations, 25 mg/dl.
These HDL levels would certainly be considered
atherogenic in the American population. Yet, coronary
heart disease in this population as well as in other
populations consuming low-fat diets is exceptionally
rare because the LDL level is also low. It is inappropriate to assume that diet-induced decreases in HDL
cholesterol concentrations carry the same burden of
atherosclerosis as do low HDL cholesterol levels in
high-fat-diet populations.
In summary, our findings indicate that although a
sudden increase in dietary carbohydrate increases
the levels of triglyceride, human subjects gradually
fed a high-carbohydrate diet that ultimately contains
65% carbohydrate, 20% fat, 15% protein, and 100 mg
cholesterol had a profound lowering of plasma total
and LDL cholesterol levels without developing hypertriglyceridemia at any time during the 7 weeks of
the study. For the future, the effects of diets high in
carbohydrate and low in fat need critical evaluation
in patients with non-insulin-dependent diabetes mellitus and hypertriglyceridemia. These patients are
especially susceptible to coronary disease and should
have maximal lowering of plasma LDL cholesterol
levels. We suggest that the phased-carbohydrate approach may prevent the usual hypertriglyceridemic
effect in these adult-onset diabetic patients. The data
of the present study certainly suggest the benefits of
the high-carbohydrate diet in human subjects with
mild hypertriglyceridemia, who ordinarily would be
the most susceptible to dietary induction of further
hypertriglyceridemia.
Acknowledgments
We are grateful to the nursing and dietary staff of
the Clinical Research Center for patient care and
assistance in this study and to Petar Alaupovic,
Oklahoma Medical Research Foundation, for determination of apolipoproteins. We are also grateful to
the patients who participated in this study for their
enthusiasm, interest, and cooperation. The authors
wish to thank Gary Miranda for his helpful comments
in the preparation of the manuscript, Jann Poersch
for typing the manuscript, CharEll Melfi for assistance in designing the figures, and Gary Sexton and
David Wilson for assistance in the statistical analysis
for this project.
References
1. Higgins M, Thorn T: Trends in coronary heart disease in the
United States. Int J Epidemiol 1989;18:S58-S66
2. Keys A: Seven Countries: A Multivariate Analysis of Death and
Coronary Heart Disease. Cambridge, Mass, Harvard University
Press, 1980
3. Stamler J: Population studies: A global review, in Levy RE,
Rifkind BM, Dennis BH, Ernst N (eds): Nutrition, Lipids, and
Coronary Disease. New York, Raven Press, Publishers, 1979,
pp 25-88
4. Dietary guidelines for healthy American adults: A statement
for physicians and health professionals by the Nutrition Committee, American Heart Association. Circulation 1988;77:
721A-724A
5. US Dept of Health and Human Services: The Surgeon General's Report on Nutrition and Health. Washington, DC, US
Department of Health and Human Services, Public Health
Service, US Government Printing Office, 1988
6. The Expert Panel: Report of the National Cholesterol Education Program Expert Panel on detection, evaluation, and
treatment of high blood cholesterol in adults. Arch Intern Med
1988;148:36-69
7. American Diabetes Association: Special Report: Principles of
nutrition and dietary recommendations for individuals with
diabetes mellitus. Diabetes Care 1979;2:520-523
8. Canadian Diabetes Association: Special report committee
guidelines for the nutritional management of diabetes mellitus: A special report from the Canadian Diabetes Association.
J Can DietAssoc 1981;42:110-118
9. Little JA, Graves K, Suchindran CM: Customary diet, anthropometry, and dyslipoproteinemia in selected North American
populations: The Lipid Research Clinics Program Prevalence
Study. Circulation 1986;73(suppl I):I-80-I-90
10. Connor WE, Connor SL: The key role of nutritional factors in
the prevention of coronary heart disease. Prev Med 1972;1:
49-83
11. Ullmann D, Phillips RL, Beeson LW, Dewey HG, Brin BN,
Kuzma JW, Mathews CP, Hirst AE: Cause-specific mortality
among physicians with differing lifestyles. JAMA 1991;265:
2352-2359
12. Reaven GM, Olefsky JM: Increased plasma glucose and
insulin response to high carbohydrate feeding in normal
subjects. / Clin Endocrinol Metab 1974;38:151-154
13. Ginsburg H, Olefsky JM, Kimmerling G: Induction of hypertriglyceridemia by a low-fat diet. / Clin Endocrinol Metab
1976;42:729-735
14. Coulston AM, Hollenbeck CB, Swislocki AL, Reaven GM:
Persistence of hypertriglyceridemic effect of low-fat highcarbohydrate diets in NIDDM patients. Diabetes Care 1989;
12:94-101
15. Gonen B, Patsch W, Kuisk I, Schonfeld G: The effect of
short-term feeding of a high carbohydrate diet on HDL
subclasses in normal subjects. Metabolism 1981;30:1125-1129
16. Austin MA: Plasma triglyceride as a risk factor for coronary
heart disease: The epidemiologic evidence and beyond. Am J
Epidemiol 1989;129:249-259
17. Pyorala K: Relationship of glucose tolerance and plasma
insulin to the incidence of coronary heart disease: Results
from two population studies in Finland. Diabetes Care 1979;2:
131-141
18. Ducimetiere P, Eschwege E, Papoz L, Richard JL, Claude JR,
Rosselin G: Relationship of plasma insulin levels to the
incidence of myocardial infarction and coronary disease mortality in a middle aged population. Diabetologia 1980;19:
205-210
19. Abbott RD, Wilson PW, Kannel WB, Castelli WP: High
density lipoprotein cholesterol, total cholesterol screening,
and myocardial infarction: The Framingham Study. Arteriosclerosis 1988;8:207-211
20. Harris WS, Connor WE, Inkeles SB, Illingworth DR: Dietary
omega-3 fatty acids prevent carbohydrate-induced hypertriglyceridemia. Metabolism 1984;33:1016-1019
21. Stone DB, Connor WE: The prolonged effects of a low
cholesterol, high carbohydrate diet upon the serum lipids in
diabetic patients. Diabetes 1963;12:127-132
Ullmann et al
Phased Dietary Carbohydrate and Plasma Triglyceride
Downloaded from http://atvb.ahajournals.org/ by guest on June 15, 2017
22. Jolliffee N, Alpert E: "The Performance Index" as a method
for estimating effectiveness of reducing regimens. Postgrad
Med 1950;9:106-115
23. Connor SL, Gustafson JR, Artaud-Wild SM, Classick-Kohn
CJ, Connor WE: The cholesterol-saturated fat index for
coronary prevention: Background, use, and a comprehensive
table of foods. J Am Diet Assoc 1989;89:807-816
24. The National Research Council: Recommended Dietary
Allowances. Washington, DC, National Academy of Sciences,
1980
25. Lipid Research Clinics Program: Manual of Laboratory Operations, Lipid and Lipoprotein Analysis, 2nd ed. Washington,
DC, Dept of Health and Human Services publications (NIH),
1982
26. Alaupovic P, Curry MD, McConathy WJ: Quantitative determination of human plasma apolipoproteins by electroimmunoassays, in Carlson LA, Paoletti IR, Weber G (eds): International Conference on Atherosclerosis. New York, Raven Press,
Publishers, 1978, pp 109-115
27. SAS Institute Inc: SAS User's Guide: Statistics, Version 5 ed.
Cary, NC, SAS Institute Inc, 1985, pp 433-506
28. Ahrens EH, Hirsch J, Oette K, Farquhar JW, Stein Y:
Carbohydrate-induced and fat-induced lipemia. Trans Assoc
Am Phys 1961;74:134-146
29. Melish J, Le NA, Ginsberg H, Steinberg D, Brown V: Dissociation of apoprotein B and triglyceride production in verylow-density lipoproteins. Am J Physiol 1980;239(Endocrinol
Metab 2):E354-E362
30. Ruderman ND, Johns A, Krauss RM, Shafrir E: A biochemical and morphologic study of very low density lipoproteins in
carbohydrate-induced hypertriglyceridemia. J Clin Invest 1971;
50:1355-1368
31. Benner KG, Sasaki A, Gowen DR, Weaver A, Connor WE:
The differential effect of eicosapentaenoic acid and oleic acid
on lipid synthesis and VLDL secretion in rabbit hepatocytes.
Lipids 1990;25:534-540
32. Antonis A, Bersohn I: The influence of diet on serum triglycerides in South African white and Bantu prisoners. Lancet
1961;l:3-9
33. Connor WE, Cerqueria MT, Connor RW, Wallace RB, Malinow MR, Casdorph HR: The plasma lipids, lipoproteins, and
diet of the Tarahumara Indians of Mexico. Am J Clin Nutr
1978;31:1131-1142
34. Cerqueria MT, Fry MM, Connor WE: The food and nutrient
intakes of the Tarahumara Indians of Mexico. Am I Clin Nutr
1979;32:905-915
1067
35. Weintraub MS, Eisenberg S, Breslow JL: Different patterns of
postprandial lipoprotein metabolism in normal, type Ila, type
III, and type IV hyperlipoproteinemic individuals. / Clin Invest
1987;79:1110-1119
36. Harris WS, Connor WE, Alam N, Illingworth DR: The
reduction of postprandial triglyceridemia in humans by dietary
n-3 fatty adds. J Lipid Res 1988;29:1451-1460
37. Fox JC, McGill HC Jr, Carey KD, Getz GS: In vivo regulation
of hepatic LDL receptor mRNA in the baboon: Differential
effects of saturated and unsaturated fat. JBiol Chem 1989;262:
7014-7020
38. Sorci-Thomas M, Wilson MD, Johnson FL, Williams DL,
Rudel LL: Studies on the expression of genes encoding
apolipoproteins B100 and B48 and the low density lipoprotein
receptor in nonhuman primates: Comparison of dietary fat
and cholesterol. / Biol Chem 1989;264:9039-9045
39. Brown MS, Goldstein JL: A receptor mediated pathway for
cholesterol homeostasis. Science 1986;232:34-47
40. Spady DK, Dietschy JM: Interaction of dietary cholesterol and
triglycerides in the regulation of hepatic low-density lipoprotein transport in the hamster. / Clin Invest 1988;81:300-309
41. Connor WE: Dietary fiber: Nostrum or critical nutrient? N
EnglJMed 1990;322:193-195
42. Raymond TL, Connor WE, Lin DS, Fry MM, Connor SL: The
interactions of dietary fibers and cholesterol upon the plasma
lipids and lipoproteins, the sterol balance, and bowel function
in human subjects. / Clin Invest 1977;60:1429-1431
43. Brinton EA, Eisenberg S, Breslow JL: A low-fat diet decreases
high-density lipoprotein (HDL) cholesterol levels by decreasing HDL apolipoprotein transport rates. / Clin Invest 1990;85:
144-151
44. Knuiman JT, West CE, Katan MB, Hautvast GAJ: Total
cholesterol and high density lipoprotein cholesterol levels in
populations differing in fat and carbohydrate intake. Arteriosclerosis 1987;7:612-619
45. Robinson P, Williams P: High density lipoprotein cholesterol
in the Masai of East Africa: A cautionary note (letter). BrMed
J 1979;1:1249
46. Sacks FM, Castelli WP, Donner A, Kass EH: Plasma lipids
and lipoproteins in vegetarians and controls. N Engl J Med
1975;292:1148-1151
KEY WORDS • carbohydrate-induced hypertriglyceridemia •
plasma cholesterol • triglycerides • low density lipoproteins •
very low density lipoproteins • low-fat diet • low dietary
cholesterol • low saturated fat
Downloaded from http://atvb.ahajournals.org/ by guest on June 15, 2017
Will a high-carbohydrate, low-fat diet lower plasma lipids and lipoproteins without
producing hypertriglyceridemia?
D Ullmann, W E Connor, L F Hatcher, S L Connor and D P Flavell
Arterioscler Thromb Vasc Biol. 1991;11:1059-1067
doi: 10.1161/01.ATV.11.4.1059
Arteriosclerosis, Thrombosis, and Vascular Biology is published by the American Heart Association, 7272 Greenville
Avenue, Dallas, TX 75231
Copyright © 1991 American Heart Association, Inc. All rights reserved.
Print ISSN: 1079-5642. Online ISSN: 1524-4636
The online version of this article, along with updated information and services, is located on the
World Wide Web at:
http://atvb.ahajournals.org/content/11/4/1059
Permissions: Requests for permissions to reproduce figures, tables, or portions of articles originally published in
Arteriosclerosis, Thrombosis, and Vascular Biology can be obtained via RightsLink, a service of the Copyright
Clearance Center, not the Editorial Office. Once the online version of the published article for which permission
is being requested is located, click Request Permissions in the middle column of the Web page under Services.
Further information about this process is available in the Permissions and Rights Question and Answerdocument.
Reprints: Information about reprints can be found online at:
http://www.lww.com/reprints
Subscriptions: Information about subscribing to Arteriosclerosis, Thrombosis, and Vascular Biology is online
at:
http://atvb.ahajournals.org//subscriptions/