UEF//eRepository
DSpace https://erepo.uef.fi
Rinnakkaistallenteet Terveystieteiden tiedekunta
2017
Dietary fatty acids were not independently associated with
lipoprotein subclasses in elderly women
Alaghehband FR
Elsevier BV
info:eu-repo/semantics/article
© Elsevier Inc
CC BY-NC-ND https://creativecommons.org/licenses/by-nc-nd/4.0/
http://dx.doi.org/10.1016/j.nutres.2017.05.014
https://erepo.uef.fi/handle/123456789/2617
Downloaded from University of Eastern Finland's eRepository
1 DIETARY FATTY ACIDS WERE NOT INDEPENDENTLY ASSOCIATED WITH
1
LIPOPROTEIN SUBCLASSES IN ELDERLY WOMEN 2
Fatemeh Ramezan Alaghehbanda, Maria Lankinena, Miika Värrib, Joonas Sirolab,c, Heikki 3
Krӧgerb,c, Arja T. Erkkiläa*
4
aInstitute of Public Health and Clinical Nutrition, University of Eastern Finland, P.O. Box 1627, 5
70211 Kuopio, Finland 6
bKuopio Musculoskeletal Research Unit, University of Eastern Finland, Kuopio, Finland 7
cDepartment of Orthopaedics and Traumatology, Kuopio University Hospital, Kuopio, Finland 8
9
Corresponding author contact information:
10
*Arja Erkkilä, Institute of Public Health and Clinical Nutrition, University of Eastern Finland, 11
P.O. Box 1627, 70211 Kuopio, Finland, e-mail arja.erkkila@uef.fi, telephone +358 40 355 2918 12
2 ABBREVIATIONS
13
apo; apolipoprotein 14
BMI ; body mass index 15
CHD; coronary heart disease 16
CVD ; cardiovascular diseases 17
DHA; docosahexaenoic acid 18
EPA; eicosapentaenoic acid 19
FDR; Benjamini & Hochberg false discovery rates 20
HDL; high-density lipoprotein 21
IDL; intermediate-density lipoprotein 22
LDL; low-density lipoprotein 23
MUFA; monounsaturated fatty acid(s) 24
NMR; nuclear magnetic resonance spectroscopy 25
PUFA; polyunsaturated fatty acid(s) 26
SFA; saturated fatty acid(s) 27
SD; standard deviation 28
VLDL; very low-density lipoprotein 29
30 31 32 33 34 35 36
3 37
38
ABSTRACT 39
40
Dietary fatty acids are known to affect serum lipoproteins. However, little is known about the 41
associations between consumption of dietary fatty acids and lipoprotein subclasses. Hypothesis is 42
that there is an association between dietary fatty acids and lipoprotein subclasses. We aimed to 43
investigate the cross-sectional association of dietary fat intake with subclasses of lipoproteins in 44
elderly women. Altogether 547 women (aged ≥ 65 years), who were part of OSTPRE cohort 45
participated. Dietary intake was assessed by three-day food records, lifestyle and health 46
information through self-administrated questionnaire and lipoprotein subclasses were determined 47
by nuclear magnetic resonance spectroscopy. To analyze the associations between fatty acids and 48
lipoprotein subclasses Pearson and Spearman correlation coefficients and ANCOVA test with 49
adjustment for physical activity, body mass index, age, smoking and intake of lipid lowering 50
drugs were used. There were significant correlations between saturated fatty acids (SFA, % of 51
energy) and concentrations of large, medium and small low-density lipoproteins (LDL), total 52
cholesterol in large, medium and small LDL and phospholipids in large, medium and small LDL 53
after correction for multiple testing. After adjustment for covariates, the higher intake of SFA 54
associated with smaller size of LDL particles (p=0.04, ANCOVA) and lower amount of 55
triglycerides in small very low density lipoproteins (p= 0.046, ANCOVA). However, these 56
associations did not remain significant after correction for multiple testing. In conclusion, high 57
intake of SFA may associate with the size of LDL particles, but the results do not support 58
significant independent associations between dietary fatty acids and lipoprotein subclasses.
59
4 60
Key Words: Fatty Acids, Lipoprotein, Lipoprotein Subclasses, Nuclear Magnetic Resonance 61
Spectroscopy, Saturated Fatty Acids 62
63 64
5 1. INTRODUCTION
65 66
Plasma lipid profile is an important factor that exerts a powerful effect on various diseases, 67
especially cardiovascular diseases (CVD). This profile depends on the synthesis of lipids in the 68
body and intake of dietary fat [1]. Therefore, the quality and quantity of dietary fat are very 69
important. There are five different types of lipoproteins with various functions, densities, sizes 70
and compositions and they can be separated into diverse subclasses using methods like nuclear 71
magnetic resonance spectroscopy (NMR), vertical rotor ultracentrifugation, and gradient gel 72
electrophoresis [2-6]. Considering the fact that, the concentration of lipoproteins can be 73
improved by the help of diet, many researches have provided valuable knowledge about the 74
effect of diet on the quantity of lipoproteins in circulation. However, further studies have 75
revealed that the distributions of different subclasses of lipoproteins are extremely important in 76
predicting future diseases. For instance, several studies have reported that the predominance of 77
low-density lipoprotein (LDL) type B (small dense lipoproteins) is related to two or three times 78
increased risk of CVD [5, 7]. One study, considered it even more pathogenic by reporting 79
sevenfold increase in the risk of CVD in subjects with small LDL higher than 2.59 mmol/l [8].
80
Moreover, it is proved that the high-density lipoprotein (HDL) has protective effects on heart 81
diseases, but further studies on different subclasses of HDL showed that the larger HDL is more 82
protective than the smaller HDL [9]. The concentration of HDL2, which are large HDL particles, 83
has had a negative association with progression and severity of coronary lesions. Furthermore, it 84
was discovered that HDL2 is significantly lower in patients who had myocardial infarction [10].
85
Although numerous studies have shown significant effect of intake of various fatty acids on 86
serum lipoprotein concentrations [11], little is known about the association of dietary fatty acids 87
6 and lipoprotein subclasses including their size and particle number. Some associations between 88
subclasses of lipoproteins and intake of various nutrients or foods e.g. carbohydrates, fructose, 89
fatty acids and alcohol have been reported [4, 12-13]. We hypothesized that the intake of dietary 90
fatty acids correlates with lipoprotein subclasses. Our aim was to examine the association 91
between dietary fat intake and serum lipoprotein subclasses in elderly women in a cross-sectional 92
setting.
93 94
2. METHODS AND MATERIALS 95
96
2.1 Participants 97
98
This study was based on the baseline data of Osteoporosis Risk Factor and Prevention- Fracture 99
Prevention Study (OSTPRE-FPS) which was an intervention study that began in 2003 and lasted 100
for 3 years [14]. OSTPRE-FPS was planned to investigate the impact of calcium and vitamin D 101
supplementation on falls and fractures in postmenopausal women older than 65 years. OSTPRE- 102
FPS was accepted by the research ethics committee of Kuopio University Hospital in October 103
2001. The study was registered in Clinical trials.gov by the identification NCT00592917.
104
Moreover, participants completed paper based informed consent form at baseline [14]. The 105
interest of participation in the study was asked by postal enquiries from August to December of 106
2002. Minimum age of 65 years at the end of November 2002 and residing in Kuopio during trial 107
were inclusion criteria of the study. In this study 3432 volunteering women with 63.5% response 108
rate participated (Figure 1) [14]. From this population, 750 participants (375 supplementation 109
7 group and 375 control group) were randomly asked to have laboratory tests and complete 3-day 110
food records at baseline. Dietary food records of 554 participants were valid for analysis [15].
111 112
2.2 Baseline Measurements 113
114
Baseline measurements started in February 2003 and finished in May 2004. Anthropometric 115
measurements were done at baseline. Weight was measured by a digital calibrated scale (Philips, 116
type HF 351/00) and height was determined by calibrated wall meter. In addition, body mass 117
index (BMI) was calculated dividing weight (in kg) by the square of height (in meters) [14].
118
Furthermore, lifestyle and health information such as intake of dietary supplements and 119
medications, physical activity were asked through self-administrated postal questionnaire [14].
120
Participants were asked if they had diagnosed diabetes treated with insulin, oral diabetic 121
medications or diet.
122 123
2.3 Lipoprotein Subclasses 124
125
The blood samples of participants were taken after overnight fast. The lipoprotein subclasses 126
were assessed by the NMR spectroscopy method in native serum. Bruker AVANCE 3 127
spectrometer, which works at 500.36 MHz was used for measuring the NMR information. The 128
methodology has been explained previously in detail [16, 17]. In this study, we used data only 129
from LIPO window, which includes concentrations of lipoprotein particles, including 130
chylomicrons, extremely large very low-density lipoprotein (VLDL) (≥75 nm), very large 131
VLDL (64.0 nm), large VLDL (53.6 nm), medium VLDL (44.5 nm), small VLDL (36.8), very 132
8 small VLDL (31.3 nm), intermediate-density lipoprotein (IDL) (28.6 nm), large LDL (25.5 nm), 133
medium LDL (23.0 nm), small LDL (18.7 nm), very large HDL (14.3 nm), large HDL (12.1 nm), 134
medium HDL (10.9 nm) and small HDL (8.7 nm). Moreover, amounts of phospholipid, total 135
cholesterol, cholesterol ester, free cholesterol and triglyceride in mentioned lipoprotein 136
subclasses and apolipoprotein (apo) B were measured. Furthermore, the mean diameter of 137
lipoprotein subclasses was measured by weighting the subclass diameter with their related 138
particle concentration. IDL particles were considered in LDL measure [18].
139 140
2.4 Dietary Intake 141
142
Participants were asked to fill out 3-day food records at baseline. The instruction of filling 143
records and food record were sent to the women, and the subjects were asked to bring the 144
completed form on the research visit. It was demanded to fill the records for 3 successive days (2 145
weekdays and 1 weekend day). Regarding fat intake, questions were asked about the quality of 146
fat consumed on bread, in cooking, and baking. Furthermore, the fish consumption frequency 147
was asked in the self-administered questionnaire that included all of the lifestyle habits, diseases 148
and medications. Nutritionist also asked about the unclear parts of the diaries through phone 149
calls. Finally, the intakes of nutrients were calculated by Nutrica program version 2.5 (Finnish 150
social insurance institute, Turku. Finland) [15].
151 152
2.5 Statistical Analyses 153
154
9 The power analysis was done based on incidence of fractures for the intervention study [14].
155
There was no a priori power analysis to calculate the size of the subsample of 750 women 156
randomly selected from the 3,432 women at the baseline.
157 158
Data were analyzed using IBM SPSS statistics software (version 21). Means ± standard deviation 159
(SD), and valid percentages are reported. The Kolmogorov-Smirnov test was used to assess 160
normality, and the skewed variables were transformed using log 10 transformation. Transformed 161
variables were used for the analysis, but back transformed means are reported. The physical 162
activity was categorized into two categories, including two times or less per week and more than 163
three times per week. In addition, smoking status was categorized into current smoker and not 164
current smoker. The Pearson correlation coefficients were calculated between normally 165
distributed dietary factors and lipoprotein subclasses and Spearman correlation coefficients were 166
calculated between non-normally distributed dietary factors and lipoprotein subclasses.
167
Frequency of fish intake was classified into three categories, including no use or once per month, 168
0.5-1 times per week and two or more times per week. Kruskall-Wallis test was used to test the 169
differences in distribution of VLDL and HDL subclasses among the categories of fish intake.
170
The differences in lipoprotein subclasses among the tertiles of SFA (% of energy) were analyzed 171
by ANCOVA test with adjustment for physical activity, BMI, age, smoking and intake of lipid 172
lowering drugs. However, unadjusted means are reported. The Bonferroni test was conducted to 173
test the pairwise differences. Benjamini & Hochberg false discovery rates (FDR) were calculated 174
using R Project for Statistical Computing version 3.0.2 to adjust results for multiple comparisons 175
[19]. The level of significance was set to FDR p value less than 0.05 for all of the tests, and two- 176
tailed p values were reported.
177
10 178
3. RESULTS 179
180
Clinical and biochemical characteristics of participants are shown in Table 1. Altogether, 5.4%
181
were current smokers and the means (±SD) BMI of the participants was 28.77± 4.73 kg/m2. 182
Hypertension was reported by 41.1% of the participants and 9.3% had diabetes. Coronary heart 183
disease (CHD) was reported by 16.6% of the participants. In addition, around 25% of the women 184
used lipid lowering drugs.
185 186
Table 2 shows the dietary intakes of nutrients among participants. Approximately 18% of energy 187
was derived from protein, 50% of energy was supplied by carbohydrate, and 31% percent of 188
energy was derived from total fat. Participants obtained on average 12.2 % of energy from 189
saturated fatty acids (SFA), 9.8% of energy from monounsaturated fatty acids (MUFA) and 5.1%
190
of energy from polyunsaturated fatty acids (PUFA). On average, they consumed 0.12 g/day 191
eicosapentaenoic acid (EPA) and 0.28 g/day docosahexaenoic acid (DHA). Main contributors to 192
SFA intake were fat spreads, milk and dairy products, meat and meat products. Fat spreads, oil 193
meat and meat products were the main foods contributing to MUFA intake. In addition, PUFA 194
was mainly derived from consumption of fat spreads, oil, cereal products and fish.
195 196
The results of correlation analyses between dietary fat intake and lipoprotein subclasses are 197
reported in Table 3. SFA (% of energy) had positive correlations with concentrations of large, 198
medium and small LDL, total cholesterol in large, medium and small LDL and phospholipids in 199
large, medium and small LDL after correction for multiple testing. Smaller mean diameter of 200
11 LDL significantly related to higher intake of MUFA and PUFA, however, these correlations lost 201
significance after correction for multiple testing. Intake of PUFA had significant positive 202
correlations with triglycerides in medium VLDL and small VLDL, concentration of small VLDL 203
and mean diameter of VLDL, however, again correlations were not significant after correction 204
for multiple comparisons. In addition, palmitic, linoleic, linolenic and arachidonic acids were not 205
significantly correlated with lipoprotein subclasses (data not shown). Higher intake of EPA 206
significantly correlated with higher total cholesterol in very small VLDL (r=0.107, p=0.013), 207
similarly intake of DHA had positive correlation with total cholesterol in very small VLDL 208
(r=0.095, p=0.027). In addition, EPA had negative correlation with the mean diameter of VLDL 209
(r=-0.086, p=0.043). Correlations between lipoprotein subclasses and EPA and DHA were not 210
significant after correction for multiple testing. There were no significant differences in the 211
VLDL and HDL subclasses among the categories of fish intake (data not shown). Moreover, a 212
negative correlation between apo B concentration and the mean size of LDL was found (r =-0.21, 213
p<0.001).
214 215
Although MUFA, PUFA, EPA and DHA had some significant correlations with different 216
subclasses of lipoprotein, the strongest correlations were observed between SFA and lipoprotein 217
subclasses. In addition, subclasses of VLDL and LDL had the higher amounts of significant 218
correlations with dietary fatty acids in comparison to subclasses of HDL. Therefore, tertiles of 219
SFA intake (% of energy) were made to evaluate the relation of SFA and subclasses of VLDL 220
and LDL. The results of these ANCOVA tests with adjustment for physical activity, intake of 221
lipid lowering drugs, smoking, age and BMI are shown in Table 4. After the adjustments, two 222
nominally significant and not linear relations between the intake of SFA and subclasses of LDL 223
12 and VLDL were found. The higher intake of SFA associated with smaller size of LDL particles 224
(p=0.040) (Figure 2). Post hoc analysis showed that there was a difference in the mean diameter 225
of LDL between the middle and the highest tertile categories of SFA (p= 0.035). In addition, 226
higher intake of SFA was related to the lower amount of triglyceride in small VLDL (p=0.046).
227
There was a difference in the mean triglycerides in small VLDL between the lowest and middle 228
tertiles of SFA (p= 0.039). Although the mentioned differences were found, none of them 229
remained statistically significant after correction for multiple comparison tests.
230 231
4. DISCUSSION 232
233
This was the first study that showed non-linear associations between the intake of SFA and 234
subclasses of LDL and VLDL defined by the NMR method. We found that the mean diameter of 235
LDL was lower in the highest tertile category of SFA in comparison to the middle tertile of SFA 236
intake. In addition, women in the lowest tertile of SFA intake had higher mean triglycerides in 237
small VLDL than women in the middle tertile of SFA. It should be noted, that the mentioned 238
non-linear and negative correlation was found after adjustment for BMI, age, physical activity, 239
lipid lowering drugs and smoking in postmenopausal women. However, none of the mentioned 240
results remained significant after correction for multiple testing. Therefore, we reject our 241
hypothesis even though the data suggest that SFA intake could be associated with smaller size of 242
LDL.
243 244
The result of our study between intake of SFA and size of LDL particles is consistent with an 245
interventional study involving ten mildly hypercholesterolemic men. The crossover intervention 246
13 started with the consumption of hamburger rich in SFA (MUFA/ SFA= 0.95) for 5 weeks, then 247
followed by usual diet for 3 weeks. Afterwards, participants consumed hamburger rich in MUFA 248
(MUFA/ SFA= 1.31) for another 5 weeks. The result of mentioned study showed that the intake 249
of hamburger rich in SFA decreased the LDL particle diameter [21]. Furthermore, in another 250
interventional study, Smith and his colleagues reported that the intake of hamburger rich in SFA 251
(MUFA/ SFA= 0.83-0.96) increased the apo B/ LDL cholesterol ratio [22]. Since according to 252
the result of present study and another study conducted by Mora and colleagues, the apo B 253
negatively correlated with the mean size of LDL [23], higher apo B/ LDL ratio might be related 254
to decreased size and increased density of LDL. Moreover, our results are supported by another 255
interventional study, where low carbohydrate and high SFA diet (15% of energy from SFA) 256
decreased the mean diameter of LDL compared with low carbohydrate, low SFA diet (8% of 257
energy from SFA), but this reduction was not significant [24].
258 259
Based on a review which was done by Schwab and Uusitupa partially replacing SFA by PUFA 260
decreases the risk of CHD [11]. In addition, another review mentioned that in most of the studies 261
aimed to find the effects of SFA intake on risk of CHD, a positive correlation between intake of 262
SFA and risk of CHD or other coronary events was found [25]. Numerous studies have reported 263
that the predominance of small and dense LDL is related to two or three times increased risk of 264
CVD [5, 7]. It is hypothesized that the intake of SFA might increase the risk of CHD through 265
decreasing the mean diameter of LDL.
266 267
On the contrary, in an interventional study Dreon and his colleagues reported that intake of SFA 268
increased the mean diameter of LDL particles [26]. In this study, healthy men consumed either a 269
14 high fat (46% of energy from fat, 18% of energy from SFA, and 39 % of energy from
270
carbohydrate) or a low fat diet (24% of energy from fat, 6% of energy from SFA, and 59% of 271
energy from carbohydrate) in a crossover design. Although the intake of MUFA and PUFA were 272
similar in both diets, the intake of carbohydrate differed greatly and carbohydrate was replaced 273
with SFA in the low fat diet. It is reported that dietary carbohydrate did not have a linear 274
correlation with plasma lipoprotein, but it might have a non-linear correlation with lipoprotein 275
particles. In addition, based on a review article, recent studies demonstrated that high intake of 276
carbohydrate decreased the size of LDL [5]. Therefore, lower consumption of carbohydrate in 277
this study might be contributing to greater mean size of LDL, not higher intake of SFA.
278
However, in our study carbohydrate intake was not significantly associated with the mean size of 279
LDL (data not shown).
280 281
Some probable mechanisms for the relation between intake of SFA and size of LDL particles can 282
be discussed, but further studies are needed to find the mechanism. Apolipoprotein CIII (apo 283
CIII) that exists on the surface of VLDL, LDL and HDL particles has a strong impact on the 284
metabolism of VLDL, LDL and HDL particles. An interventional study showed that diet rich in 285
SFA (15% of energy from SFA, 15% of energy from MUFA) significantly increased the amount 286
of apo CIII in LDL by 33.5% in comparison to diet low in SFA and rich in MUFA (8% of energy 287
from SFA and 21% of energy from MUFA) [27]. It should be mentioned that other 288
macronutrients were kept constant in this study, the intake of carbohydrate was moderate and the 289
changes were independent of the triglycerides level. In addition, another interventional study that 290
used two different diets, including low SFA diet (8% of energy from SFA) and high SFA diet 291
(15% energy from SFA with replacement of SFA by MUFA) showed that lower intake of SFA 292
15 decreased the activity of hepatic lipase [24]. Therefore, SFA might increase the formation of 293
small LDL or decrease the mean diameter of LDL by increasing the concentration of apo CIII 294
and decreasing the hepatic lipase activity.
295 296
Although there is a lot of evidence about the effect of unsaturated fatty acids, especially PUFA 297
on lipoprotein subclasses [12, 28-33], we found only a few relations between unsaturated fatty 298
acids and lipoprotein subclasses and those were not significant after correction for multiple 299
comparisons. In addition, in contrast to the several studies that found the significant relations 300
between intake of fish and lipoprotein subclasses mainly HDL [34-36], we did not find any 301
significant differences in distribution of VLDL and HDL subclasses through categories of fish 302
intake. This contrast might be due to consumption patterns of fish. First, intakes of fish and 303
unsaturated fatty acids among participants were low in the current study as compared to other 304
studies [13, 28-33]. Second, variations in intakes of fish and unsaturated fatty acids may not have 305
been large enough to find an association.
306 307
Using the large population is one strength of this study, but the non-response rate of 36.5%
308
decreased the benefit of that. Since the 750 participants were randomly selected from 3432 309
volunteered women, the sample is representative of cohort population and the result of study is 310
generalizable to the Finnish elderly women. Moreover, availability of diet records and 311
lipoprotein subclasses defined by NMR is strength of this study. Limitations of the study include 312
the cross-sectional study design and therefore we cannot show causal relations. Participant 313
number and the variation in dietary intake might have been too small reducing the power to 314
16 observe significant associations. There can be confounding factors affecting the association of 315
dietary fat and lipoprotein subclasses that we did not measure and were not able to control.
316 317
Our result suggests that higher intake of SFA might be related to the unfavorable size of LDL 318
particles which has been related to increased risk of heart diseases. However, our results do not 319
support significant independent associations between dietary fatty acids and lipoprotein 320
subclasses. This result points to the need for further longitudinal and interventional studies to 321
elucidate the mechanism of the relation between SFA intake and lipoprotein subclasses and to 322
clarify the clinical meaning of this possible association.
323 324
Acknowledgements 325
326
Funding sources for this study included: Finnish Academy (Kröger H, grant number 327
250707), The Ministry of Education and Culture of Finland (grant numbers 116/626/2012, 328
46/627/2011 and 93/627/2010), Kuopio University Hospital Funding (V.T.R. grant number 329
5203024), and Strategic Funding of the University of Eastern Finland (grant number 931053).
330 331 332
Conflict of Interest 333
334
There is no conflict of interest.
335 336
17 5. REFERENCES
337
[1] Kwiterovich P. Dietary and Pharmacologica Treatment of Dyslipidemia and Cardiovascular 338
Diseases. The Johns Hopkins University Textbook of Dyslipidemia. Lippincott Williams and 339
Wilkins, Philadelphia; 2009.
340
[2] Moredich C, Kark D, Keresztes P. Implications of LDL subclass B in patients at 341
cardiovascular risk. Nurse Pract 2005; 30: 16-29.
342
[3] Ala-Korpela M. Critical evaluation of 1H NMR metabonomics of serum as a methodology 343
for disease risk assessment and diagnostics. Clin Chem Lab Med 2008; 46: 27-42 . 344
[4] Schaefer EJ, Gleason JA, Dansinger ML. Dietary fructose and glucose differentially affect 345
lipid and glucose homeostasis. J Nutr 2009; 139: 1257S-1262S.
346
[5] Gerber PA, Berneis K. Regulation of low-density lipoprotein subfractions by carbohydrates.
347
Curr Opin Clin Nutr Metab Care 2012; 15: 381-5.
348
[6] Sninsky JJ, Rowland CM, Baca AM, Caulfield MP, Superko HR. Classification of LDL 349
phenotypes by 4 methods of determining lipoprotein particle size. J Invest Med 2013;61:942-9.
350
[7] Diffenderfer MR, Schaefer EJ. The composition and metabolism of large and small LDL.
351
Curr Opin Lipidol 2014; 25: 221-6.
352
[8] Kim OY, Chung HK, Shin MJ. Higher levels of serum triglyceride and dietary carbohydrate 353
intake are associated with smaller LDL particle size in healthy Korean women. Nutr Res Pract 354
2012; 6: 120-5.
355
[9] Arsenault BJ, Lemieux I, Despres JP, Gagnon P, Wareham NJ, Stroes ES, Kastelein JJ, Khaw 356
KT, Boekholdt SM. HDL particle size and the risk of coronary heart disease in apparently 357
healthy men and women: the EPIC-Norfolk prospective population study. Atherosclerosis 2009;
358
206: 276-81.
359
18 [10] Pirillo A, Norata GD, Catapano AL. High-density lipoprotein subfractions-what the
360
clinicians need to know. Cardiology 2013; 124: 116-25.
361
[11] Schwab U, Uusitupa M. Diet heart controversies- Quality of fat matters. Nutr Metab 362
Cardiovasc Dis 2015; 25: 617-22.
363
[12] Mukamal KJ, Mackey RH, Kuller LH, Tracy RP, Kronmal RA, Mittleman MA, Siscovick 364
DS. Alcohol consumption and lipoprotein subclasses in older adults. J Clin Endocrinol Metab 365
2007; 92: 2559-66.
366
[13] Annuzzi G, Rivellese AA, Wang H, Patti L, Vaccaro O, Riccardi G, Ebbesson SO, 367
Comuzzie AG, Umans JG, Howard BV. Lipoprotein subfractions and dietary intake of n-3 fatty 368
acid: the Genetics of Coronary Artery Disease in Alaska Natives study. Am J Clin Nutr 2012;
369
95: 1315-22.
370
[14] Kärkkäinen M, Tuppurainen M, Salovaara K, Sandini L, Rikkonen T, Sirola J, Honkanen R, 371
Jurvelin J, Alhava E, Kroger H. Effect of calcium and vitamin D supplementation on bone 372
mineral density in women aged 65-71 years: a 3-year randomized population-based trial 373
(OSTPRE-FPS). Osteoporos Int 2010; 21: 2047-55.
374
[15] Jarvinen R, Tuppurainen M, Erkkila AT, Penttinen P, Karkkainen M, Salovaara K, Jurvelin 375
JS, Kroger H. Associations of dietary polyunsaturated fatty acids with bone mineral density in 376
elderly women. Eur J Clin Nutr 2012; 66: 496-503.
377
[16] Soininen P, Kangas AJ, Würtz P, Suna T, Ala-Korpela M. Quantitative serum nuclear 378
magnetic resonance metabolomics in cardiovascular epidemiology and genetics. Circ Cardiovasc 379
Genet 2015; 8:192-206.
380
[17] Soininen P1, Kangas AJ, Würtz P, Tukiainen T, Tynkkynen T, Laatikainen R, Järvelin 381
MR, Kähönen M, Lehtimäki T, Viikari J, Raitakari OT, Savolainen MJ, Ala-Korpela M. High- 382
throughput serum NMR metabonomics for cost effective holistic studies on systemic 383
metabolism. Analyst 2009; 134:1781-5.
384
19 [18] Wang J, Stancakova A, Soininen P, Kangas AJ, Paananen J, Kuusisto J , M. Ala-Korpela M, 385
Laakso M. Lipoprotein subclass profiles in individuals with varying degrees of glucose 386
tolerance: a population-based study of 9399 Finnish men. J Intern Med 2012; 272:562-72.
387
[19] Benjamini Y, Hochberg Y. Controlling the false discovery rate: a practical and powerful 388
approach to multiple testing. J R Statis Soc B 1995; 57:289-300.
389
[21] Adams TH, Walzem RL, Smith DR, Tseng S, Smith SB. Hamburger high in total, saturated 390
and trans-fatty acids decreases HDL cholesterol and LDL particle diameter, and increases TAG, 391
in mildly hypercholesterolaemic men. Br J Nutr 2010; 103: 91-8.
392
[22] Smith DR, Wood R, Tseng S, Smith SB. Increased beef consumption increases 393
apolipoprotein A-I but not serum cholesterol of mildly hypercholesterolemic men with different 394
levels of habitual beef intake. Exp Biol Med 2002; 227: 266-75.
395
[23] Mora S, Otvos JD, Rifai N, Rosenson RS, Buring JE, Ridker PM. Lipoprotein Particle 396
Profiles by Nuclear Magnetic Resonance Compared With Standard Lipids and Apolipoproteins 397
in Predicting Incident Cardiovascular Disease in Women. Circulation 2009; 119: 931-9.
398
[24] Mangravite LM, Chiu S, Wojnoonski K, Rawlings RS, Bergeron N, Krauss RM. Changes in 399
atherogenic dyslipidemia induced by carbohydrate restriction in men are dependent on dietary 400
protein source. J Nutr 2011; 141: 2180-5.
401
[25] Hunter JE, Zhang J, Kris-Etherton PM. Cardiovascular disease risk of dietary stearic acid 402
compared with trans, other saturated, and unsaturated fatty acids: a systematic review. Am J Clin 403
Nutr 2010; 91: 46-63.
404
[26] Dreon DM, Fernstrom HA, Campos H, Blanche P, Williams PT, Krauss RM. Change in 405
dietary saturated fat intake is correlated with change in mass of large low-density-lipoprotein 406
particles in men. Am J Clin Nutr 1998; 67: 828-36.
407
[27] Faghihnia N, Mangravite LM, Chiu S, Bergeron N, Krauss RM. Effects of dietary saturated 408
fat on LDL subclasses and apolipoprotein CIII in men. Eur J Clin Nutr 2012; 66: 1229-33.
409
20 [28] Almario RU, Vonghavaravat V, Wong R, Kasim-Karakas SE. Effects of walnut
410
consumption on plasma fatty acids and lipoproteins in combined hyperlipidemia. Am J Clin Nutr 411
2001; 74: 72-9.
412
[29] Goyens PL, Mensink RP. The dietary alpha-linolenic acid to linoleic acid ratio does not 413
affect the serum lipoprotein profile in humans. J Nutr 2005; 135: 2799-804.
414
[30] Bos G, Poortvliet MC, Scheffer PG, Dekker JM, Ocke MC, Nijpels G, Stehouwer CD, 415
Bouter LM, Teerlink T, Heine RJ. Dietary polyunsaturated fat intake is associated with low- 416
density lipoprotein size, but not with susceptibility to oxidation in subjects with impaired glucose 417
metabolism and type II diabetes: the Hoorn study. Eur J Clin Nutr 2007; 61: 205-11.
418
[31] Neff LM, Culiner J, Cunningham-Rundles S, Seidman C, Meehan D, Maturi J, Wittkowski 419
KM, Levine B, Breslow JL. Algal docosahexaenoic acid affects plasma lipoprotein particle size 420
distribution in overweight and obese adults. J Nutr 2011; 141: 207-13.
421
[32] Perez-Martinez P, Ordovas JM, Garcia-Rios A, Delgado-Lista J, Delgado-Casado N, Cruz- 422
Teno C, Camargo A, Yubero-Serrano EM, Rodriguez F, Perez-Jimenez F, Lopez-Miranda J.
423
Consumption of diets with different type of fat influences triacylglycerols-rich lipoproteins 424
particle number and size during the postprandial state. Nutr Metab Cardiovasc Dis 2011; 21: 39- 425
426 45.
[33] Parlesak A, Eckoldt J, Winkler K, Bode CJ, Schafer C. Intercorrelations of lipoprotein 427
subfractions and their covariation with lifestyle factors in healthy men. J Clin Biochem Nutr 428
2014; 54: 174-80.
429
[34] Li Z, Lamon-Fava S, Otvos J, Lichtenstein AH, Velez-Carrasco W, McNamara JR, Ordovas 430
JM, Schaefer EJ. Fish consumption shifts lipoprotein subfractions to a less atherogenic pattern in 431
humans. J Nutr 2004; 134: 1724-8.
432
[35] Erkkila AT, Schwab US, Lehto S, de Mello VD, Kangas AJ, Soininen P, Ala-Korpela M, 433
Uusitupa MI. Effect of fatty and lean fish intake on lipoprotein subclasses in subjects with 434
coronary heart disease: a controlled trial. J Clin Lipidol 2014; 8: 126-33.
435
21 [36] Lankinen M, Kolehmainen M, Jaaskelainen T, Paananen J, Joukamo L, Kangas AJ,
436
Soininen P, Poutanen K, Mykkanen H, Gylling H, Oresic M, Jauhiainen M, Ala-Korpela M, 437
Uusitupa M, Schwab U. Effects of whole grain, fish and bilberries on serum metabolic profile 438
and lipid transfer protein activities: a randomized trial (Sysdimet). Plos One 2014; 9: 1-12.
439
440
22 Table 1. Clinical and biochemical characteristics of participants.
Women (n=555)
Age (y)b 67.85± 1.87a
Body mass index (kg/m2)b 28.77± 4.73
Current smoker (%)c 5.4
Hypertension (%) 41.1
Coronary heart diseases (%) 16.6
Stroke (%) 1.1
Transient ischemic attack (%) 6.8 Other heart disease (%) 8.8 Diabetes treated with insulin (%) 1.8 Diabetes treated with oral medication
(%)
4.1 Diabetes treated with diet (%) 3.4 Intake of lipid lowering drugs (%) 24.7 Physical activity
2 times or less per weeks (%) More than 3 times per week (%)
40.4 59.6
aMeans ± SD is reported for continuous variable and valid percentage for categorical variables.
bn= 554
cn= 503
23 Table 2. Dietary intake of participants.
Nutrient Means ± SD Minimium Maximum
Energy (kcal/d) 1568± 372 482 3056
Protein (% of energy) 17.6± 3.1 9.3 33.8
Carbohydrates (% of energy) 49.1± 5.8 32.4 66.7
Fat (% of energy) 31.1± 5.6 9.9 50.4
Saturated fatty acids (% of energy) 12.2± 3.1 3.6 24.9
Palmitic acid 16:0 (g/d) 9.91± 3.95 2.3 35.3
Stearic acid 18:0 (g/d) 3.85± 1.63 0.4 14.0
Monounsaturated fatty acids (% of energy) 9.8± 2.4 2.2 21.6 Polyunsaturated fatty acids (% of energy) 5.1± 1.4 1.9 12.7 Linoleic acid 18:2 n-6 (g/d) 5.88± 2.34 1.4 17.9
Linolenic acid 18:3 (g/d) 1.45± 0.78 0.4 5.2
Arachidonic acid 20:4 n-6 (g/d) 0.11± 0.08 0.01 0.73 Eicosapentaenoic acid 20:5 n-3 (g/d) 0.12± 0.14 0.00 0.95 Docosahexaenoic acid 22:6 n-3 (g/d) 0.28± 0.33 0.00 2.04 n=554
24 Table 3. Correlation coefficients between intake of fatty acids and lipoprotein subclasses.
MUFA (% of energy) PUFA (% of energy) SFA (% of energy)
Correlation
Coefficienta
p value FDR pb Correlation Coefficient
p value FDR p Correlation Coefficient
p value FDR p Concentration of very large VLDL 0.056 0.192 0.428 0.047 0.272 0.528 0.021 0.622 0.965 Phospholipid in very large VLDL 0.057 0.183 0.428 0.040 0.353 0.605 0.037 0.384 0.791 Total cholesterol in very large VLDL 0.051 0.234 0.456 0.028 0.516 0.694 0.053 0.214 0.567 Triglycerides in very large VLDL 0.056 0.191 0.428 0.051 0.231 0.476 0.011 0.793 0.965 Concentration of large VLDL 0.064 0.136 0.428 0.077 0.072 0.278 -0.004 0.926 0.965 Phospholipid in large VLDL 0.062 0.146 0.428 0.071 0.097 0.310 0.001 0.984 0.984 Total cholesterol in large VLDL 0.054 0.208 0.428 0.061 0.157 0.423 0.006 0.889 0.965 Triglycerides in large VLDL 0.066 0.124 0.428 0.081 0.059 0.269 -0.008 0.852 0.965 Concentration of medium VLDL 0.062 0.149 0.428 0.080 0.061 0.269 -0.007 0.867 0.965 Phospholipid in medium VLDL 0.059 0.166 0.428 0.075 0.079 0.278 -0.003 0.937 0.965 Total cholesterol in medium VLDL 0.047 0.275 0.465 0.059 0.170 0.426 0.011 0.800 0.965 Triglycerides in medium VLDL 0.067 0.120 0.428 0.086 0.043 0.269 -0.017 0.683 0.965 Concentration of small VLDL 0.060 0.162 0.428 0.086 0.045 0.269 -0.012 0.780 0.965 Phospholipid in small VLDL 0.058 0.178 0.428 0.083 0.052 0.269 -0.023 0.594 0.965 Total cholesterol in small VLDL 0.046 0.279 0.465 0.054 0.205 0.476 0.045 0.288 0.673 Triglycerides in small VLDL 0.059 0.165 0.428 0.094 0.027 0.269 -0.042 0.327 0.716 Concentration of very small VLDL 0.034 0.428 0.564 0.033 0.445 0.624 0.035 0.420 0.816
25 Phospholipid in very small VLDL 0.024 0.578 0.630 0.020 0.642 0.816 0.052 0.227 0.567 Total cholesterol in very small VLDL 0.023 0.594 0.630 0.007 0.868 0.893 0.085 0.047 0.165 Triglycerides in very small VLDL 0.033 0.435 0.564 0.067 0.117 0.343 -0.057 0.185 0.539 Concentration of large LDL 0.029 0.497 0.579 -0.013 0.760 0.858 0.115 0.007 0.035 Total cholesterol in large LDL 0.024 0.575 0.630 -0.053 0.219 0.476 0.124 0.004 0.031 Phospholipid in large LDL 0.030 0.481 0.579 -0.014 0.747 0.858 0.124 0.004 0.031 Triglycerides in large LDL 0.016 0.717 0.738 -0.008 0.861 0.893 0.011 0.790 0.965 Concentration of medium LDL 0.045 0.296 0.471 0.002 0.959 0.959 0.111 0.009 0.036 Phospholipid in medium LDL 0.049 0.254 0.465 -0.034 0.432 0.624 0.130 0.002 0.031 Total cholesterol in medium LDL 0.032 0.458 0.573 -0.039 0.363 0.605 0.115 0.007 0.035 Triglycerides in medium LDL 0.012 0.771 0.771 -0.045 0.296 0.546 0.017 0.686 0.965 Concentration of small LDL 0.058 0.174 0.428 0.012 0.788 0.862 0.111 0.009 0.036 Total cholesterol in small LDL 0.037 0.386 0.543 -0.033 0.443 0.624 0.116 0.007 0.035
Phospholipid in small LDL 0.073 0.089 0.428 0.017 0.699 0.844 0.126 0.003 0.031
Triglycerides in small LDL 0.040 0.352 0.536 0.019 0.653 0.816 0.005 0.907 0.965
Mean diameter of VLDL 0.055 0.197 0.428 0.087 0.041 0.269 -0.017 0.684 0.965
Mean diameter of LDL -0.119 0.005 0.182 -0.097 0.024 0.269 -0.080 0.061 0.193
Mean diameter of HDL -0.037 0.388 0.543 -0.034 0.429 0.624 0.004 0.929 0.965
n= 547,
aSpearman and Pearson correlation coefficient were used.
bFDR p value correction by Benjamini & Hochberg for multiple comparisons.
Abbreviations: FDR, false discovery rate; LDL, low-density lipoprotein; MUFA, monounsaturated fatty acids; PUFA, polyunsaturated fatty acids; SFA, saturated fatty acids and VLDL , very low-density lipoprotein.
26 Table 4. Concentration of lipoprotein particles in tertiles of saturated fatty acid intake.
Lowest tertile (3.6-
10.8 % of energy)
Middle tertile (10.8- 13.3 % of energy)
Highest tertile (13.3- 24.9 % of energy)
P trenda
FDR p- valuesb
N 163 157 156
Concentration of very large VLDL (nmol//l) 0.688± 0.727c 0.565± 0.498 0.702± 0.688 0.564 0.596 Phospholipid in very large VLDL (mmol/L) 0.011± 0.012 0.009± 0.008 0.011± 0.011 0.585 0.596 Total cholesterol in very large VLDL
(mmol/L)
0.015± 0.015 0.013± 0.010 0.015± 0.014 0.528 0.596 Triglyceride in very large VLDL (mmol/L) 0.042± 0.045 0.034± 0.031 0.043± 0.043 0.596 0.596 Concentration of large VLDL (nmol/L) 4.527± 3.875 3.736± 2.649 4.512± 3.671 0.467 0.596 Phospholipid in large VLDL (mmol/L) 0.047± 0.41 0.39± 0.028 0.047± 0.038 0.420 0.596 Total cholesterol in large VLDL (mmol/L) 0.061± 0.053 0.051± 0.037 0.061± 0.052 0.555 0.596 Triglyceride in large VLDL (mmol/L) 0.154± 0.133 0.126± 0.091 0.153± 0.125 0.467 0.596 Phospholipid in medium VLDL (mmol/L) 0.114± 0.061 0.100± 0.043 0.114± 0.062 0.165 0.432 Total cholesterol in medium VLDL
(mmol/L)
0.160± 0.085 0.145± 0.062 0.162± 0.085 0.224 0.473 Phospholipid in small VLDL (mmol/L) 0.155± 0.042 0.144± 0.034 0.153± 0.049 0.099 0.432 Triglyceride in small VLDL (mmol/L) 0.250± 0.097 0.220± 0.071 0.246± 0.113 0.046 0.418 Concentration of very small VLDL (nmol/L) 47.635± 10.287 46.317± 8.953 48.054± 11.394 0.282 0.536 Free cholesterol in very small VLDL
(mmol/L)
0.097± 0.024 0.095± 0.021 0.099± 0.026 0.554 0.596
27 Triglyceride in large LDL (mmol/L) 0.121± 0.033 0.115± 0.027 0.122± 0.036 0.146 0.432 Triglyceride in medium LDL (mmol/L) 0.056± 0.017 0.053± 0.014 0.057± 0.019 0.141 0.432 Concentration of small LDL (nmol/L) 170.823± 46.934 170.440± 43.127 181.791± 52.546 0.182 0.432 Phospholipid in small LDL (mmol/L) 0.143± 0.028 0.142± 0.025 0.150± 0.030 0.066 0.418 Mean diameter of LDL (nm) 23.664± 0.1350 23.680± 0.142 23.639± 0.152 0.040 0.418 Log transformed variables were used for analysis.
aThe Multivariate ANOVA model was adjusted for BMI, age, physical activity, lipid lowering drugs and smoking.
bFDR p value correction by Benjamini & Hochberg for multiple comparisons.
cValues are means ± SD.
Abbreviations: FDR, false discovery rate; LDL, low-density lipoprotein and VLDL, very low-density lipoprotein.
28 Figure 1. Flow chart of the subsample of the OSTPRE-FPS trial
Eligible subjects of OSTPRE- cohort (n=5407)
3432 Randomized
1718 to calcium + vitamin D group 1714 to control group
750 randomly selected subsample 375 to both groups
237 Withdrew after randomization 132 From intervention group
132 Withdrew consent 105 From control group
83 Withdrew consent 15 Died before start 7 No contact information 1975 Excluded
701 did not return enquiry 962 Not willing to participate 312 Did not meet inclusion criteria
290 Allocated to calcium + vitamin D group and received allocated
intervention
313 Allocated to control group
554 provided food record data
547 had both diet and serum lipoprotein data
29 Figure 2. Error bar graph showing the diameter of LDL particles by tertiles of saturated fat intake. n=476. Means, 95% confidence intervals and p-value were derived from ANCOVA.
Abbreviations: CI, confidence interval; LDL; low-density lipoprotein and SFA, saturated fat.
