Atherosclerosis and Lipoproteins |
From the Division of Human Nutrition and Epidemiology (N.M.d.R., M.B.K.), Wageningen University, Wageningen, the Netherlands; Julius Center for Patient Oriented Research (M.L.B.), University Medical Center, Utrecht, the Netherlands; and The Wageningen Center for Food Sciences (M.B.K.), Wageningen, the Netherlands.
Correspondence to Dr Martijn B. Katan, Division of Human Nutrition and Epidemiology, Wageningen University, Bomenweg 2, 6703 HD Wageningen, Netherlands.
| Abstract |
|---|
|
|
|---|
Key Words: lipoproteins HDL trans fatty acids endothelium arteriosclerosis
| Introduction |
|---|
|
|
|---|
We investigated whether the intake of trans fat would indeed increase the risk of CHD more than the intake of saturated fat by comparing the effects of these fats on endothelial function, a surrogate cardiovascular end point.14 15 16 We assessed endothelial function as flow-mediated vasodilation (FMD) of the brachial artery, because this is a noninvasive measurement that correlates well with known risk factors17 18 19 20 21 22 and other markers of CHD.23 24 25 Moreover, 2 longitudinal studies show an association between FMD in the past with future CHD events.26 27 The diets were given for a minimum of 3 weeks, a time period long enough to establish changes in serum lipids28 and FMD.21 We hypothesized that FMD would be lower after the diet rich in trans fat than after the diet rich in saturated fat because of the expected difference in serum HDL-C.
| Methods |
|---|
|
|
|---|
Study Design
We provided 2 controlled diets for 4 weeks, each in a
randomized crossover design. The diets consisted of conventional food
items supplemented with special margarines and were given in a 28-day
menu cycle. On Mondays through Fridays, subjects came to our dining
room and ate a hot meal under our supervision. All other foods (bread;
margarine; meat and/or cheese; honey, jam, or sprinkles; fruit; milk
and/or yogurt) were packed for consumption at home, as was food for the
weekends.
Habitual energy intake of the subjects was estimated from a food-frequency questionnaire. We designed menus for 14 levels of energy intake, ranging from 7 to 20 MJ/d, and allocated the subjects to an intake level close to their habitual energy intake. We provided 90% of energy (en%); all this food was weighed out for each subject. We measured body weight twice a week; if body weight changed >1 kg, subjects were switched to a different energy intake level. The remaining 10 en% had to be chosen from a list of low-fat food items and recorded in a diary. Subjects received the diets for 21 to 32 days (mean 27.5 days).
Diets
The experimental diets differed in margarine only
(Table 1
). The composition of the diets was calculated by
using food composition
tables3 29 30
and checked by collecting duplicates of all meals
(Table 2
). The analyzed values were similar to the
calculated composition.
|
|
The margarine in the diet rich in TFAs (Trans-diet) was a blend of 70 parts partially hydrogenated soy oil, containing 44% trans-C18:1 (Goudas Glorie, Van Dijk Foods), 14 parts vegetable oil containing 63% linoleic acid and 23% oleic acid (Becel, Unilever), and 16 parts water. The margarine in the diet rich in saturated fat (Sat-diet) was a blend of 60 parts palm kernel fat (Loders Croklaan) and 40 parts commercially available margarine made from a blend of vegetable oils and solid vegetable fats (Blue Band, Van den Bergh BV). Both margarines were produced at NIZO Food Research. The margarines were used as shortenings in bread and cookies, in sauce and gravy, and as a spread. They supplied 62% of fat in the diets; the remaining 38% was mainly derived from meat, cheese and other dairy products, eggs, and salad dressings.
Blood Lipids
We took blood samples after an overnight fast on 2
separate days after day 19 of each diet. All 4 blood samples of each
subject were analyzed in duplicate within 1 run. Total
cholesterol and triglycerides
(Cholesterol Flex and
Triglycerides Flex reagent
cartridge, Dade Behring) and HDL-C
(Liquid
N-geneous HDL-C assay, Instruchemie BV) were
measured, and LDL cholesterol was calculated with the
Friedewald formula. The coefficient of variation of 64 duplicate
measurements was 0.4% for total cholesterol, 1.5% for
triglycerides, and 1.1% for HDL-C.
Brachial Artery Measurements
All subjects had an overnight fast of at least 12
hours before the measurements. We measured FMD of the brachial artery
as described by Celermajer et
al22 and Sorensen et
al.31 We used the diameter
of the artery at rest and at maximum vasodilation to calculate the
percentage increase or FMD. All measurements were performed at end
diastole by the use of the R wave of the ECG. The
ultrasound images were made by 1 technician with a 7.5-MHz linear array
transducer of an Ultramark 9 HDI duplex scanner.
All images were stored on super-VHS videotapes for offline
analysis.
Subjects were made to lie down in a temperature-controlled room (range 20°C to 23°C) with the right arm in 2 arm support cushions. An inflatable cuff was placed around the lower arm. The transducer was held in position at the site of the antecubital crease with a specially developed transducer arm holder (method developed by R. Meijers group, Vascular Imaging Center, The Julius Center for Patient Oriented Research UMC, Utrecht, the Netherlands).
We first obtained an optimal 2D B-mode ultrasound image of the brachial artery at rest and recorded 3 images to measure the diameter. We then inflated the cuff to 250 mm Hg and kept this pressure constant for 5 minutes to induce ischemia in the forearm and hand. After 5 minutes, the cuff was deflated. The image of the brachial artery was optimized, and changes in the diameter of the artery were recorded during the next 5 minutes. Every 15 seconds, a frozen image was stored on videotape. At the end of the second feeding period, we also measured endothelium-independent vasodilation after a sublingual dose of 400 µg of nitroglycerin.
One reader who was blinded to the treatment read all the images at the Vascular Imaging Center of the University Medical Center in Utrecht. The reader rated the quality of the images from class 1 (perfect) to class 4 (unfit for use). All 32 subjects were measured twice on both diets, so we had 4 measurements per subject. Of these 128 measurements, 24 were rated as perfect, 71 as fair, 26 as marginal, and 2 as unfit. Five measurements were missing. We used only measurements that were rated perfect or fair, which left us with 29 subjects for whom we had observations on both diets. At a mean FMD of 5.3% of the resting diameter, the SD within subjects was 2.6% points, so the corresponding coefficient of variation was 49%. The largest difference in a duplicate FMD measurement was 18.2% points (FMD 2.6% and 20.8% of the resting diameter); the smallest difference was 0.16% points (measurements were 7.2% and 7.4% of the resting diameter). The coefficient of variation of the resting and maximum diameter was 8%.
Statistical Analysis
We averaged the duplicate measurements in each
dietary period and tested for order effects by ANOVA, with diet and
order as main effects in the
model.32 Because the order
of the 2 diets did not significantly contribute to the model, we then
calculated for each subject the difference between treatments. We
tested whether these differences were significantly different from zero
by the Student t test for
paired samples. We give 2-sided 95% CIs for the
differences.
| Results |
|---|
|
|
|---|
Body Weight
During the 4-week feeding periods, body weight remained
basically stable, with mean decreases of 0.4 kg during the Trans-diet
and 0.6 kg during the Sat-diet
(P=0.43 for difference in
change between diets).
Blood Lipids
Serum HDL-C decreased from 1.87±0.46 mmol/L
(73.1±17.8 mg/dL) on the diet rich in saturated fats to
1.49±0.33 mmol/L (56.5±12.8 mg/dL) on the diet rich in
trans fats
(Table 3
). The decrease was 0.39 mmol/L (95% CI
-0.50 to -0.28), or 21%. Serum LDL cholesterol and
triglycerides remained stable. The order of the 2 diets
hardly affected the change in HDL-C: the mean change was
0.35±0.25 mmol/L in subjects who went from the Trans-diet to the
Sat-diet and 0.43±0.32 mmol/L in the subjects who received the
diets in the reverse order.
|
Brachial Artery Measurements
The diameter of the brachial artery at rest was
4.02±0.70 mm on the Sat-diet and 4.08±0.73 mm on the
Trans-diet. The maximum diameter was 4.33±0.80 mm on the Sat-diet
and 4.19±0.73 mm on the Trans-diet. FMD was 6.2±3.0% on the
Sat-diet and 4.4±2.3% on the Trans-diet
(P=0.015). Thus, FMD was 1.8%
(95% CI -3.2 to -0.4), or 29% lower on the Trans-diet than on the
Sat-diet
(Figure
).
The order of the 2 diets hardly affected the results: 15 subjects went
from an FMD of 4.8% after the Trans-diet to 6.4% after the Sat-diet,
whereas 14 other subjects went from 5.9% after the Sat-diet to 4.2%
after the Trans-diet.
|
All subjects showed vasodilation after nitroglycerin (range 4.4% to 20.8%). Diet had no effect on nitroglycerin-mediated vasodilation, which was 14.3±3.4% on the Trans-diet and 13.4±5.3% on the Sat-diet (unpaired t test, P=0.64).
A decrease in HDL-C went together with a decrease in FMD in 18 of 29 subjects. The correlation between changes in HDL-C and FMD was positive (r=0.12, 95% CI -0.26 to 0.46) but not significant (P=0.55).
| Discussion |
|---|
|
|
|---|
HDL-C, Other Dietary Factors, and
Endothelial Function
There is some evidence that changes in HDL-C
concentration could change endothelial function. First,
higher serum HDL-C is associated with better
endothelial
function.24 33 34
This might be due to the proposed antioxidant capacity of
HDL,35 which might prevent
oxidation of LDL and therefore prevent adverse effects of oxidatively
modified LDL on endothelial function. We know of no
other interventions aimed at HDL, but other antioxidants, such as
vitamin C,36 37
were shown to improve FMD. Second, there is ample evidence that
reductions in other known risk factors, such as LDL
cholesterol21 25 38
or homocysteine,39 improve
FMD, suggesting that changes in HDL-C could have similar effects. The
fact that we did not find a significant correlation between changes in
HDL-C and FMD does not rule out a causal relation, because the data
were too scarce to correct for possible confounding variables, such
as sex and age. On the other hand, a significant correlation would be
no proof of a causal relation.
Other factors in the diets might account for the effect on
FMD. As shown in
Table 2
, there was a small difference in linoleic acid
between the 2 diets, and studies with rats show that TFAs have stronger
effects at low intakes of linoleic
acid.40 Although this might
apply to humans, those rat studies were performed at very high intakes
of TFAs (20 en%), and the adverse effects could be counteracted with a
linoleic acid intake as low as 2 en%. Thus, the 4.1 en% provided by
linoleic acid in our 9.2 en% Trans-diet was not low compared with
percentages in the rat studies. Also, we think that the difference in
linoleic acid between the Sat-diet and Trans-diet was too small to
fully explain the effects seen on FMD. Another factor is vitamin E; the
different fat mixtures likely differed by 10 to 20 mg/100 g. However,
studies that showed an effect of vitamin E on
FMD41 used much higher
doses, and even at these high doses, most studies did not show an
effect.42 43 44
Last, FMD is impaired in
diabetes,45 and if TFAs and
saturated fatty acids have different effects on insulin
metabolism, this could have biased the results. However, it
is unlikely that fasting serum insulin was different between the 2
diets.46
We do not know of studies that compared long-term effects of different fats on FMD. Postprandial effects of saturated and cis-monounsaturated fats seem to be similar; they all appear to impair FMD compared with preprandial values or compared with low-fat control meals.36 47 48 However, some of these studies36 47 are flawed because the low-fat meals had a higher vitamin C content than the fat-enriched meals, which might have improved FMD.49 We know of no short-term effects of TFAs on FMD.
Study Limitations
We used a crossover design to eliminate variation due
to differences between subjects. The order of the 2 diets was balanced
and randomized per subject to eliminate bias due to a systematic drift
of the outcome variables over
time.32 Although we did not
include a washout period, we did not find a significant order effect on
any of the blood lipoproteins or for FMD.
We were interested only in differences between the 2 test
diets but not in changes from the habitual diet; therefore, no baseline
data were collected. We can only speculate on changes in blood
lipoproteins and FMD from baseline. Both experimental diets differed in
fat content from habitual diets: the amount of TFAs in the Trans-diet
was
23 g/d, which is 5-fold higher than the estimated 4.8 g/d for
men and 3.8 g/d for women in the
Netherlands.3 The amount of
saturated fat in the Sat-diet was 58 g/d, which is also higher than the
habitual intake of 42 g/d for men and 32 g/d for women in the
Netherlands. Because of the low habitual intake of TFAs, replacing them
all by saturated fatty acids would probably hardly improve
endothelial function. Conversely, our findings imply
that replacing all saturated fatty acids by TFAs could impair FMD and
should therefore be discouraged.
The inclusion of women in the present study may have increased the variation in FMD response, because changes in serum estradiol concentrations affect FMD.50 However, we minimized this variation with 4-week study periods, the length of a menstrual cycle. Compared with the men, the women appeared to respond stronger to the diets, with a 2.3 percentage-unit (95% CI 0.4 to 4.2) smaller FMD on the Trans-diet than on the Sat-diet; in the men, the difference was 0.8 percentage units (95% CI -1.3 to 3.0). However, the number of men was small (n=10); therefore, the present study was not powered to test for sex differences. Further studies with larger numbers of men and women are needed to test for differences in response.
Repeatability of the FMD Measurement
We found a mean FMD of 5.3%. This is somewhat lower
than values for healthy volunteers reported by
others,16 36 50
but differences in methodology (eg, the position of the inflatable
cuff)51 could account for
this. The variability in FMD was high; we found a coefficient of
variation of 49%. This is comparable with the variability found in
some studies52 53
but higher than values reported by
others.23 31 50 54 55
However, in most studies it is unclear how the values for variability
have been calculated.
In conclusion, we showed that replacement of saturated fatty acids by TFAs in the diet lowered serum HDL-C and impaired FMD. This suggests that TFAs increase the risk of CHD more than the intake of saturated fats, with similar effects on LDL cholesterol. Further studies are needed to verify whether decreases in HDL-C indeed impair endothelial function and thereby explain the increased risk of CHD at high intakes of trans fats.
| Acknowledgments |
|---|
Received December 27, 2000; accepted March 30, 2001.
| References |
|---|
|
|
|---|
lfl G, Maurer G, Weidinger F.
Long-term prognostic value of low-mediated vasodilation in the brachial
artery of patients with angina pectoris: results of a 5-year follow-up
study. Circulation.
1999;100(suppl I):I-48. Abstract.
This article has been cited by other articles:
![]() |
S. S. Bassuk and J. E. Manson Lifestyle and Risk of Cardiovascular Disease and Type 2 Diabetes in Women: A Review of the Epidemiologic Evidence American Journal of Lifestyle Medicine, June 1, 2008; 2(3): 191 - 213. [Abstract] [PDF] |
||||
![]() |
M.D KONTOGIANNI, A ZAMPELAS, and C TSIGOS Nutrition and Inflammatory Load Ann. N.Y. Acad. Sci., November 1, 2006; 1083(1): 214 - 238. [Abstract] [Full Text] [PDF] |
||||
![]() |
J. A Nettleton, L. M Steffen, E. J Mayer-Davis, N. S Jenny, R. Jiang, D. M Herrington, and D. R Jacobs Jr Dietary patterns are associated with biochemical markers of inflammation and endothelial activation in the Multi-Ethnic Study of Atherosclerosis (MESA) Am. J. Clinical Nutrition, June 1, 2006; 83(6): 1369 - 1379. [Abstract] [Full Text] [PDF] |
||||
![]() |
D. Mozaffarian, M. B. Katan, A. Ascherio, M. J. Stampfer, and W. C. Willett Trans Fatty Acids and Cardiovascular Disease N. Engl. J. Med., April 13, 2006; 354(15): 1601 - 1613. [Full Text] [PDF] |
||||
![]() |
A. D. Mooradian, M. J. Haas, and N. C. W. Wong The Effect of Select Nutrients on Serum High-Density Lipoprotein Cholesterol and Apolipoprotein A-I Levels Endocr. Rev., February 1, 2006; 27(1): 2 - 16. [Abstract] [Full Text] [PDF] |
||||
![]() |
J. B. Keogh, J. A. Grieger, M. Noakes, and P. M. Clifton Flow-Mediated Dilatation Is Impaired by a High-Saturated Fat Diet but Not by a High-Carbohydrate Diet Arterioscler. Thromb. Vasc. Biol., June 1, 2005; 25(6): 1274 - 1279. [Abstract] [Full Text] [PDF] |
||||
![]() |
K. Oh, F. B. Hu, J. E. Manson, M. J. Stampfer, and W. C. Willett Dietary Fat Intake and Risk of Coronary Heart Disease in Women: 20 Years of Follow-up of the Nurses' Health Study Am. J. Epidemiol., April 1, 2005; 161(7): 672 - 679. [Abstract] [Full Text] [PDF] |
||||
![]() |
E. Lopez-Garcia, M. B. Schulze, J. B. Meigs, J. E. Manson, N. Rifai, M. J. Stampfer, W. C. Willett, and F. B. Hu Consumption of Trans Fatty Acids Is Related to Plasma Biomarkers of Inflammation and Endothelial Dysfunction J. Nutr., March 1, 2005; 135(3): 562 - 566. [Abstract] [Full Text] [PDF] |
||||
![]() |
M. L. Bots, J. Westerink, T. J. Rabelink, and E. J.P. de Koning Assessment of flow-mediated vasodilatation (FMD) of the brachial artery: effects of technical aspects of the FMD measurement on the FMD response Eur. Heart J., February 2, 2005; 26(4): 363 - 368. [Abstract] [Full Text] [PDF] |
||||
![]() |
S. Kreijkamp-Kaspers, L. Kok, M. L Bots, D. E Grobbee, J. W Lampe, and Y. T van der Schouw Randomized controlled trial of the effects of soy protein containing isoflavones on vascular function in postmenopausal women Am. J. Clinical Nutrition, January 1, 2005; 81(1): 189 - 195. [Abstract] [Full Text] [PDF] |
||||
![]() |
W. M. R. Broekmans, I. A. A. Klopping-Ketelaars, J. A. Weststrate, L. B. M. Tijburg, G. van Poppel, A. A. Vink, T. T.J.M Berendschot, M. L. Bots, W. A. M. Castenmiller, and A. F. M. Kardinaal Decreased Carotenoid Concentrations Due to Dietary Sucrose Polyesters Do Not Affect Possible Markers of Disease Risk in Humans J. Nutr., March 1, 2003; 133(3): 720 - 726. [Abstract] [Full Text] [PDF] |
||||
![]() |
C. A Francois, S. L Connor, L. C Bolewicz, and W. E Connor Supplementing lactating women with flaxseed oil does not increase docosahexaenoic acid in their milk Am. J. Clinical Nutrition, January 1, 2003; 77(1): 226 - 233. [Abstract] [Full Text] [PDF] |
||||
![]() |
N. Dashti, Q. Feng, M. R. Freeman, M. Gandhi, and F. A. Franklin Trans Polyunsaturated Fatty Acids Have More Adverse Effects than Saturated Fatty Acids on the Concentration and Composition of Lipoproteins Secreted by Human Hepatoma HepG2 Cells J. Nutr., September 1, 2002; 132(9): 2651 - 2659. [Abstract] [Full Text] [PDF] |
||||
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
ATVB Home | Subscriptions | Archives | Feedback | Authors | Help | AHA Journals Home | Search Copyright © 2001 American Heart Association, Inc. All rights reserved. Unauthorized use prohibited. |