Embed Size (px)
Citation preview
The Development of Individual Human Fatty Acid Reference Ranges
by
Cody A. C. Lust
A Thesis presented to
The University of Guelph
In partial fulfilment of requirements
for the degree of Master of Science
in Human Health and Nutritional Sciences
Guelph, Ontario, Canada
© Cody Lust, September, 2020
ABSTRACT
THE DEVELOPMENT OF INDIVIDUAL FATTY ACID REFERENCE RANGES
Cody A.C. Lust
University of Guelph, 2020
Advisor(s):
Dr. David WL Ma
Consumption of dietary fatty acids (FA) are essential for the development and
maintenance of human health. Currently, individual FA reference ranges have yet to be
established in the field of dietary lipid research. Clinical reference ranges for lipids, such as
cholesterol, are routinely utilized in clinical practices for efficient and precise assessments.
However, a number of factors such as dietary consumption, age, sex, and metabolism have been
shown to modify FA status. In this study, the effect of aging on FA status in Australians, and the
correlation between FA concentrations and lipid biomarkers in Singaporean adults was
examined. FA status was modestly correlated to aging across multiple decades and lipid
biomarkers strongly correlated to FA concentrations. Continuing to profile multiple cohorts in
different countries, using consistent methodology, will allow for FA cut-off values to be
developed (i.e. sufficiency vs. deficiency) and provide target values for reducing chronic disease
risk.
iii
ACKNOWLEDGEMENTS
Only two years ago I was working as a bartender in Vancouver B.C., and after receiving an offer
in late-August to pursue a master’s degree, decided to drop everything on ten days’ notice and move to
Guelph. I will forever owe my gratitude to my advisor Dr. David Ma for plucking me from obscurity and
providing me every opportunity to grow as a researcher and open doors for me which I thought had closed
for good. I look forward to the next step in our journey as I begin my PhD and cannot thank you enough
for your continued support and patience as I wind my way through the crazy world of academia.
To the other member of my committee, Dr. Lindsay Robinson, thank you for being so supportive
from the moment I arrived in Guelph and providing valuable insight on not only this project but, my first
publication earlier this year. To all the members of the Ma Lab and friends in HHNS, thank you for
welcoming me with open arms and showing me the ropes. Working alongside so many intelligent, driven
individuals with such a diverse scope of interests pushes me to work harder myself each day. Most
importantly, I would still be in the GC room processing samples for this project if it wasn’t for Lyn
Hilyer. You are the glue which keeps our lab together, and I cannot thank you enough for having the
answer to every question even before I ask.
To my family and friends back home on the west coast, there is only so much more I can say here
which I haven’t told you before. Each one of you has taught me something that I keep with me each day
so, even though we are apart, you are always close by. Finally, Georgia, what is left to say after surviving
undergrad, some awkward gap years, our master’s degree, a soon to be PhD, and a global pandemic
locked-up together for 6 months. You have made me a kinder, smarter, and all-around better person – I
would not have wanted anyone else by my side during this process.
iv
TABLE OF CONTENTS
Abstract ............................................................................................................................. ii
Acknowledgements ............................................................................................................ iii
Table of Contents ............................................................................................................... iv
List of Tables/Figures ........................................................................................................ viii
List of Symbols, Abbreviations or Nomenclature .................................................................. ix
1 INTRODUCTION ....................................................................................................... 1
2 LITERATURE REVIEW | LIPID METHODOLOGY AND REPORTING OF DIETARY FATTY ACID STATUS IN HUMANS ................................................................................. 4
2.1 Assessments of Fatty Acid Status ........................................................................... 4
2.1.1 Food Frequency Questionnaires ...................................................................... 4
2.1.2 24-Hour Recall .............................................................................................. 6
2.1.3 Dietary Records ............................................................................................ 7
2.1.4 Adipose Tissue .............................................................................................. 8
2.1.5 Red Blood Cells .......................................................................................... 10
2.1.6 Blood Plasma and Serum ............................................................................. 12
2.2 Lipid Methodology and Reporting of Fatty Acid Status .......................................... 16
2.2.1 Gas Chromatography ................................................................................... 17
2.2.2 Gas Chromatography Mass-Spectrometry ...................................................... 20
2.2.3 Gas Chromatography with Flame Ionization Detector ..................................... 21
2.2.4 Units of Measure ......................................................................................... 23
2.3 Developing Individual Fatty Acid Reference Ranges ............................................. 25
2.3.1 Disease States ............................................................................................. 28
v
2.3.2 Genetic Differences ..................................................................................... 28
2.3.3 Geographical Differences ............................................................................. 30
2.4 Conclusion ......................................................................................................... 31
3 RATIONALE, OBJECTIVES, HYPOTHESIS, AND EXPERIMENTAL DESIGN ......... 33
3.1 Rationale ........................................................................................................... 33
3.2 Objectives .......................................................................................................... 33
3.3 Experimental Design ........................................................................................... 34
4 AGING IS ASSOCIATED WITH DIFFERENTIAL CHANGES OF CIRCULATING SERUM FATTY ACID LEVELS IN AUSTRALIAN MALES ............................................. 37
4.1 Abstract ............................................................................................................. 37
4.2 Introduction ....................................................................................................... 38
4.3 Materials and methods ........................................................................................ 40
4.3.1 Participants ................................................................................................. 40
4.3.2 Gas Chromatography Protocol ...................................................................... 41
4.3.3 Fatty Acid Analysis ..................................................................................... 42
4.3.4 Statistical Analysis ...................................................................................... 42
4.4 Results ............................................................................................................... 44
4.4.1 General characteristics and lipid biomarkers .................................................. 45
4.4.2 Correlation between fatty acids and age ......................................................... 48
4.4.3 Absolute concentration when stratified by decade ........................................... 50
4.4.4 Percent composition when stratified by decade ............................................... 51
4.4.5 Comparison between absolute concentration and percent composition .............. 52
4.5 Discussion ......................................................................................................... 53
4.5.1 Saturated Fatty Acids ................................................................................... 54
vi
4.5.2 Monounsaturated Fatty Acids ....................................................................... 55
4.5.3 Omega-6 Polyunsaturated Fatty Acids ........................................................... 56
4.5.4 Omega-3 Polyunsaturated Fatty Acids ........................................................... 58
4.5.5 Statistical and Biological Relevance .............................................................. 62
4.5.6 Absolute vs. Relative Measures .................................................................... 63
4.5.7 Limitations ................................................................................................. 64
4.6 Conclusion ......................................................................................................... 65
5 FATTY ACID REFERENCE RANGES AND CORRELATIONS TO LIPID BIOMARKERS OF HEALTHY SINGAPOREAN MEN AND WOMEN .............................. 66
5.1 Abstract ............................................................................................................. 66
5.2 Introduction ....................................................................................................... 67
5.3 Materials and methods ........................................................................................ 68
5.3.1 Participants ................................................................................................. 68
5.3.2 Anthropometric Measures ............................................................................ 69
5.3.3 Blood Samples ............................................................................................ 69
5.3.4 Gas Chromatography Protocol ...................................................................... 70
5.3.5 Fatty Acid Analysis ..................................................................................... 71
5.3.6 Statistical Analysis ...................................................................................... 71
5.4 Results ............................................................................................................... 72
5.5 Discussion ......................................................................................................... 82
5.5.1 Individual Fatty Acid Reference Ranges ........................................................ 83
5.5.2 Anthropometric Measures ............................................................................ 85
5.5.3 Correlations with Lipid Biomarkers .............................................................. 87
5.5.4 Sex Differences ........................................................................................... 91
vii
5.5.5 Limitations ................................................................................................. 93
5.6 Conclusion ......................................................................................................... 93
6 GENERAL DISCUSSION, LIMITATIONS and FUTURE DIRECTIONS ..................... 95
6.1 General Discussion ............................................................................................. 95
6.2 Limitations ......................................................................................................... 96
6.3 Future Directions ................................................................................................ 97
7 REFERENCES .......................................................................................................... 99
8 APPENDIX A ......................................................................................................... 118
8.1 Appendix B ...................................................................................................... 123
viii
LIST OF TABLES/FIGURES
Table 2.1 Studies profiling ranges of fatty acids in different cohorts ........................................... 27
Table 4.1 General characteristics of study population .................................................................. 44
Table 4.2 Correlation between serum fatty acids and age ............................................................ 46
Table 4.3 Correlation and multiple regression of serum fatty acids and age when stratified by decade ........................................................................................................................................... 50
Table 5.1 General characteristics of study population .................................................................. 72
Table 5.2 Range, mean and percentiles of Fatty Acid concentrations (μmol/L) of serum total lipids .............................................................................................................................................. 74
Table 5.3 Concentration (μmol/L) of select fatty acids in males and females .............................. 78
Table 5.4 Reference ranges and clinical cut-offs of lipid biomarkers .......................................... 79
Table 5.5 Linear regression model assessing the correlation between individual fatty acids and lipid biomarkers ............................................................................................................................ 79
Table 5.6 Upper and lower percentile fatty acid concentrations in participants who exceed risk cut-offs for lipid biomarkers ......................................................................................................... 81
Figure 2.1 Screenshot of a chromatogram in OpenLab CDS EZChrom Edition 3.2.1 ................ viii
APPENDIX A
Table 8.1 Absolute measures of major fatty acids in plasma and serum .................................... 118
Table 8.2 Relative percent composition of major fatty acids in plasma and serum ................... 119
Figure 8.1 Bland-Altman plot of the difference of plasma and serum fatty acids ...................... 121
APPENDIX B
Table 8.3 Mean concentration of select fatty acids across multiple decades ............................123
ix
LIST OF SYMBOLS, ABBREVIATIONS OR NOMENCLATURE
24HDR: 24-Hour dietary record
A*STAR: Agency for Science, Technology and Research
ALA: Alpha-linolenic acid
ARA: Arachidonic Acid
AT: Adipose Tissue
BMI: Body Mass Index
CE: Cholesterol esters
CHD: Coronary heart disease
CV: Coefficient variance
CVD: Cardiovascular disease
DHA: Docosahexaenoic acid
DPA: Docosapentaenoic acid
DR: Dietary records
EFA: Essential fatty acids
EPA: Eicosapentaenoic acid
FA: Fatty acids
FAME: Fatty acid methyl esters
FFQ: Food Frequency Questionnaires
GC: Gas chromatography
GC-FID: Gas chromatography with flame ionization detector
GC-MS: Gas chromatography with mass spectrometry
LA: Linoleic acid
LOD: Limit of detection
x
MAILES: The Men Androgen Inflammation Lifestyle Environment and Stress
MUFA: Monounsaturated fatty acids
NA: North America
NEFA: Non-esterified fatty acids
NHANES: National Health and Nutrition Examination Study
OA: Oleic acid
PA: Palmitic acid
PAO: Palmitoleic acid
PC: Phosphatidylcholine
PL: Phospholipids
PUFA: Polyunsaturated fatty acids
RBC: Red blood cells
RT: Retention time
SA: Stearic Acid
SD: Standard deviation
SFA: Saturated fatty acids
t: Trace
TAG: Triacylglycerols
TG: Triglycerides
WC: Waist circumference
1
1 INTRODUCTION The consumption of fats, oils, and lipids is essential to the development and maintenance
of overall human health and well-being (1). Due to their wide variance in structure and
properties, lipids have been classified into eight distinct categories each containing their own
classes and subclasses (2). Of these categories, fatty acids (FA) are a diverse group of molecules
that are most commonly associated with dietary status/intake. Dietary FA are commonly divided
into four classes: saturated fatty acids (SFA), monounsaturated fatty acids (MUFA),
polyunsaturated fatty acids (PUFA), and trans fats (3). Numerous studies have been published
outlining the necessity of obtaining adequate dietary FA to promote optimal health (4). Of these
four classes, MUFA and PUFA are commonly seen as the “healthier” fats, in particular, omega-3
PUFA which has been found to exhibit a number of beneficial health effects (1,4). Omega-6
PUFA are quite prevalent in the North American diet but, are found to negatively impact health
compared to their omega-3 counterparts (1). Increased consumption of SFA and trans fats have
also been found to be harmful to human health and a number of public health initiatives have
attempted to reduce levels in the global food supply (4). Therefore, monitoring FA status can
provide useful insight into changes which may be related to health status.
In epidemiological studies, dietary FA status is most commonly assessed via the use of a
dietary record such as a 24-hour recall, 3 or 7-day food intake record, or a food frequency
questionnaire (5). Although technology has helped improve the capabilities of subjective dietary
measures, inherent bias of self-reporting and the extrapolation of individual FA quantities from
various food sources are significant limitations. FA status can also be assessed via blood lipid
analysis, which can determine individual levels of FA and is reflective of dietary consumption
2
and endogenous lipid synthesis (6). FA status, however, is impacted by a number of biological
and environmental factors. For example, all SFA and MUFA have the capability to be
endogenously produced which is driven via the enzymatic reaction of Δ9-desaturase (3). Only
the PUFA alpha-linolenic acid (ALA) and linoleic acid (LA) are considered to be essential fatty
acids (EFA), which means they must be consumed via the diet and cannot be produced
endogenously (7). The long-chain omega-3 PUFA, eicosapentaenoic acid (EPA) and
docosahexaenoic acid (DHA), can also be produced endogenously, but have very limited
conversion rates of 1% or less (8). Therefore, PUFA more accurately reflect dietary intake based
on our capability to endogenously produce SFA and MUFA (3). However, certain disease-states
are known to alter FA metabolism, impacting circulating FA concentrations independent of
natural endogenous formation (9). An individual’s geographic location also plays a large role in
FA status due to cultural differences in diet and, accessibility to food sources. For example,
countries in South-East Asia and Oceania have far higher reported SFA, and omega-3 PUFA
intakes than North America (NA) whereas, NA has greater omega-6 intake (10). Thus, FA status
can be modified not only by dietary intake but, specific disease-states and individual metabolic
properties.
A current gap in the field of dietary lipid research is that individual FA reference ranges
have yet to be established. Objective measures with established clinical ranges for lipids
categories such as cholesterol, triglycerides, and total free fatty acids are routinely utilized in
clinical practices for efficient and precise assessments (11). Additionally, the reporting and
methodology of such FA data is often inconsistent throughout literature (12). In order to address
these gaps, one must first comprehensively profile baseline FA concentrations in healthy,
3
disease-free populations. The use of consistent methodological approaches will improve
reproducibility and comparisons of results between studies. Furthermore, taking into
consideration differences in the population being assessed such as geographic location, genetics,
age, sex and health status, will help to establish more valid baseline measures. Establishing
reference ranges would allow for meaningful assessments of FA status in individuals (i.e.
sufficiency vs. deficiency) and provide target ranges for health, similar to that of other clinical
measures widely used today.
4
2 LITERATURE REVIEW | LIPID METHODOLOGY AND REPORTING OF DIETARY FATTY ACID STATUS IN HUMANS
2.1 Assessments of Fatty Acid Status
There are a number of subjective and objectives measures currently available to
researchers to assess dietary FA status in humans (13). Food Frequency Questionnaires (FFQ) or,
diet records are subjective measures used to estimate dietary FA status and intake, whereas
measuring blood FA status is an objective measure utilized in certain studies. From a research
design perspective, having access to multiple assessments allows researchers to make appropriate
choices based on their specific study design and desired outcomes. With the goal of developing
dietary FA reference ranges, using accurate and consistent methods which are able to be
conducted effectively in large-scale epidemiological studies is paramount. Three key steps take
place during this process: the collection of samples/data which reflect dietary FA intake, the
analysis of said samples, and the reporting of results. The inconsistency of assessments across
different studies when reporting a singular outcome (values representative of FA status)
however, limits the ability to accurately compare results between studies. Consistent use of
objective measures coupled with validated subjective measures can improve compatibility
between studies and provide a richer understanding of FA status in humans.
2.1.1 Food Frequency Questionnaires
FFQ are a self-reported, indirect measure of assessing dietary FA status in individuals
(14). FFQ consist of three main components: The list of individual foods/group of foods,
frequency of consumption indicated by number of times per day/week/month, and an estimated
portion size of specific items consumed (15). When FFQ include questions regarding the specific
5
quantity of food consumed they are further classified as either semi-quantitative, quantitative, or
non-quantitative (14). The use of FFQ in epidemiological studies is quite high due since it is
relatively simple to fill-out, only taking 20-30 minutes to complete, can be conducted at home or
online, are non-invasive, and cost-effective (14,15). Of 189 studies examined which utilized self-
reported dietary assessment tools in Canada, 64% used some type of FFQ or screener (16).
However, FFQ are not a standardized instrument as they can be designed or adapted for use in-
specific populations or to answer specific dietary questions (15). To determine if a FFQ
accurately reports dietary intake of the population being examined a validity study is almost
always performed (14). A number of studies globally have examined the validity of using FFQ to
estimate the intake of dietary lipids, most commonly omega-3 PUFA, with often moderate-to-
strong correlations (17–25).
As FFQ are an in-direct measure of dietary status, they lack the specificity to accurately
assess individual FA status. In order to validate a FFQ, comparisons must be drawn to other
more specific dietary assessment tools such as 24-hour dietary recalls (24HDR) or blood lipid
analysis (14,15). This is due to FFQ containing a large degree of measurement error and being
limited by the ‘select’ list of foods presented, classifying complex mixtures, bias of participants
to under/over report consumption, limits in memory, ability to estimate portion sizes, and others
(14,15). Studies have attempted to develop mathematical formulas to determine specific FA
status based on FFQ results but, correlating FA with low-levels to FFQ have proven to be
ineffective (26). Additionally, FA metabolism can be modified by disease and age resulting in
the FA quantity consumed not being an accurate reflection of biological status (27). Overall,
although FFQ lack specificity, they provide a simple and informative method for monitoring
6
dietary trends that can inform whether changes in FA status are indicative of metabolic or dietary
changes.
2.1.2 24-Hour Recall
The 24HDR is an indirect, retrospective method of precisely recalling and quantifying all
food and beverages consumed over a 24-hour period, which is commonly completed 2-3 times
(28). Typically, a 24HDR is done in-person or over the phone each time with trained
interviewers and has been systematically utilized by the National Health and Nutrition
Examination Study (NHANES) in the USA (28). The use of a trained interviewer helps improve
precision of the results, especially with low literacy populations (28). Although there are
prevailing 24HDR commonly used in literature, similar to FFQ, they must be validated when one
is created or modified (29). With the rapid growth of technology, online 24HDR have been
developed which participants can fill-out on their own and do not require the supervision of
trained professionals (29,30). The Automated Self-Administered 24-hour Recall has been
modelled based on the same method utilized in the NHANES and has been found to be valid and
perform similarly for assessing true dietary intakes compared to the traditional interviewing
method (29,31). Using an online system can significantly reduce costs of a study and scale-up
research efforts but, was prone to more inconsistencies compared to a standard interview due to
some participants having difficulties with the program (29). When compared to FFQ, 24HDR
have been found to a more accurate assessment of true dietary intake due to the exact
quantification and specificity of the food products consumed (14,32–34).
24HDR have a greater specificity of measuring dietary consumption compared to FFQ
but, are still quite limited at assessing individual FA status. Due to 24HDR also being an in-
7
direct measure of FA status, similar translational limitations to that of FFQ exist. Studies have
found 24HDR can be valid for assessing dietary intake in multiple races but, stronger
correlations were noted in PUFA than the moderate correlations seen in MUFA and SFA (35).
However, other studies have found relatively low correlations with 24HDR and individual FA
levels when compared to blood biomarkers (36). Overall, 24HDR provides an effective tool with
greater specificity than FFQ for reporting dietary FA consumption but, lacks consistency for
reporting individual FA status.
2.1.3 Dietary Records
Dietary records (DR) are an in-direct, prospective method of assessing an individual’s
consumption of food and beverages recorded over a specified period of time, typically 3, 4, or 7
days (37). When assessing dietary consumption, DR are typically the ‘gold standard’ of which
FFQ and 24HDR are often validated against (37). Recording dietary intake as it is happening,
opposed to the retroactive methods used in FFQ and 24HDR, eliminates recall bias allowing for
more accurate and detailed information to be recorded (5,37). As DR are an open-ended
approach, participants still require a training session to ensure they record adequate details of
their diet consistently over the 3-7 day timeframe (5).
Due to the human-bias involved in FFQ, 24HDR, and DR, all three methods of
assessment typically suffer from underreporting of dietary consumption as participants attempt to
exhibit healthier eating habits (38). Furthermore, due to the assessment taking place over
multiple days, participants may begin to change dietary habits or lose motivation to fill out
detailed reports (5,37). Recent advances in technology are attempting to improve upon DR by
using online databases with photos which are easier, less time-consuming, and have been found
8
to reduce underreporting of dietary intake (39). In order to assess the validity of DR, studies
often utilize biomarkers such as blood lipids which provide an integrated perspective of dietary
consumption as it reflects both absorption and metabolism (40). When DR are validated against
biomarkers with respect to FA status they have shown moderate-to-strong correlations
(36,40,41). Comparatively, DR have shown the highest validity of overall and individual FA
intake compared to both FFQ and 24HDR (36,41). Taken together, DR are the most accurate
subjective estimate of dietary status. Although biomarkers are valid objective estimates of
dietary intake, the confounding effect of disease and individual genetics can impact the measured
values (5,27). Thus, the use of DR to assess dietary consumption and provide context to
objective values obtained from the analysis of biomarkers can provide a richer understanding to
changes of individual FA.
2.1.4 Adipose Tissue
Methods of dietary assessment which participants recall/record their food and beverage
consumption over a fixed duration of time are highly common in epidemiological studies but,
lack biological precision. Dietary fat is not only the largest single source of energy; interest
regarding the bioactive capabilities of dietary fat consumption and its effect on chronic disease
has grown in recent years (42). Given this increased interest, the need for more precise
assessments of dietary fat intake has also grown resulting in a number of studies utilizing tissue
and blood samples for strong biomarkers of FA status (3). Adipose tissue (AT) is the largest
storage compartment of FA and is often considered the “gold standard” representation of dietary
FA status (3). AT samples are most commonly obtained from either the gluteal or abdominal
region, which maintain relatively stable FA content over time (42). However, variability between
9
quantities of FA has been seen in adipose tissue found at different sites and should be kept
consistent throughout sampling (43). Due to the half-life of FA in AT lasting up to 600+ days,
AT is often seen as the strongest predictor of long-term FA intake (44,45). Significant positive
correlations, as high as r=0.70, have been seen between dietary intake and AT content of PUFA
(3,35,46) and trans FA (46). Correlations between SFA and MUFA have been more variable
with ranges of r=0.03-65 compared to AT content (3,44) yet strong associations with the dairy
fats SFA pentadecanoic acid (C15:0) and margaric acid (C17:0) have commonly been reported
(47).
The use of AT as a biomarker strongly correlates to dietary intake, and FA content
maintains stable under homeostatic conditions but, it has notable limitations which reduces its
practical use in human research. First, obtaining AT samples is a fairly invasive technique which
proves to be costly and time-consuming when conducted at a large scale (48). Different tissue
sampling sites have also seen notable FA differences, and improper handling of samples or, not
being stored at the correct temperature can oxidize PUFA (42). Drastic changes in weight loss
have also modified AT PUFA content as much as 15% in both abdominal and gluteal samples
(49). Finally, due to the lengthy turn-over of AT FA in weight stable individuals, assessing when
this compartment reaches saturation can take 2-3 years of observation (50). Following
supplementation, one study found that after twenty weeks changes in AT PUFA content was
“almost imperceptible” (51). From a logistical perspective, conducting a longitudinal nutritional
intervention study of >2 years is useful in some contexts but, incredibly difficult and expensive
to be conducted routinely. Furthermore, AT FA concentrations may not be an accurate
“snapshot” of an individual’s current dietary status due to the prolonged time it takes to see
10
changes. Overall, AT may be considered the “gold standard” for representing long-term dietary
FA but, the prolonged half-life and costly invasive procedure render it impractical for use in
large-scale epidemiological studies.
2.1.5 Red Blood Cells
Biomarkers which are best suited for dietary intake are sensitive to small changes, easily
obtained, and are reflective of the period of intake being assessed (short vs. long-term) (44).
Asides from obtaining AT samples to assess individual FA status, multiple blood fractions are
commonly analyzed in epidemiological studies which meet these criteria. The three blood
fractions which are most commonly assessed for individual FA status are erythrocytes/red blood
cells (RBC), whole blood, and plasma/serum (3). RBC, plasma and serum can be further divided
into phospholipids (PL), triacylglycerols (TAG), non-esterified fatty acids (NEFA), cholesterol
esters (CE), phosphatidylcholine (PC), phosphatidylethanolamine (PE), or the total of all
fractions (48). A 2016 paper found that ~90% of dietary FA research assessed at least one of:
total RBC, total plasma/serum, or plasma PL (52). Each of these fractions are each uniquely
modified by dietary and metabolic changes resulting in inherent strengths and limitations of
subsequent assessments (3).
RBC contain a more complete spectrum of membrane PL classes resulting in FA
concentrations which strongly reflects other biological tissues and consumption (3). Consisting
of a half-life of ~120 days, RBC are seen as an effective long-term biomarker of dietary FA
status (53). Compared to the long half-life of AT samples, upwards to two years, the 120 days
seen in RBC is a more accurate reflection of current dietary status/trends (48). The relatively
long half-life results in RBC measures being less sensitive to fasting status, reducing artificial
11
FA spikes via short-term dietary modifications (54). Supplementation studies have seen omega-3
PUFA composition of RBC change in as little as three days but, take beyond twenty-eight days
to become fully saturated (55). As PC and PE classes in the RBC membrane saturate at different
rates, this could account for the relatively quick changes when assessing total RBC composition
(12). Typically, changes to RBC composition show smaller effects with dietary interventions but,
are less variable when compared to other fractions over similar time frames (56). Numerous
studies have validated RBC composition as an accurate indicator of dietary FA intake with the
strongest correlations being found in PUFA, and more moderate correlations to MUFA and
certain trans fats (3,22,44,48,52,53,57). As RBC composition of EPA and DHA have
consistently exhibited strong correlations to dietary consumption, the Omega-3 Index (O3I) was
developed based on the calculation: (%EPA+%DHA)/% Total FA in RBC (58). This index has
been widely cited in literature as a general surrogate to assess overall omega-3 status, with values
of >8% correlating to reductions in cardiovascular disease (CVD) and coronary heart disease
(CHD) (58).
Although assessment of FA status via RBC is common in epidemiological studies, they
do exhibit some well-known limitations. Correlations of RBC values to SFA and specific MUFA
intake have been found to be relatively weak (22,44,53,57). The SFA palmitic acid (PA;C16:0),
stearic acid (SA;C18:0), and the MUFA oleic acid (OA;C18:1 c9) have not only been found to
weakly correlate to dietary intake but, differ between RBC and other fractions (3,12). Similar
discrepancies between fractions have been found with omega-6 PUFA Linoleic acid (LA;C18:2
n6) and Arachidonic acid (ARA;C20:4 n6) (3,12). Our bodies ability to endogenously produce
SFA and MUFA reduces the strength of correlation between dietary consumption and all blood
12
biomarkers, as this trend is seen in RBC and other fractions (3). Furthermore, the preparation and
storage of RBC samples have proven to be troublesome in large-scale studies (56). Standard
collection method of blood samples for scientific research typically utilizes plasma and serum
samples, not RBC (59). The procedures for separating RBC samples from whole-blood and for
lipid extraction differs entirely from plasma or serum and is more time-consuming and labour
intensive to conduct (60). RBC samples display intermediate stability and require very particular
storage conditions immediately following collection to limit the degradation of FA which can
occur in as little as two days (42,56). The additional labour and expertise required to prepare, and
store RBC samples is a significant limitation when conducting large-scale experiments. Overall,
the use of RBC as a biomarker of long-term dietary intake has shown strong correlation to a
number of FA but, weakly correlates to a few key FA of interest and requires very specific and
costly procedures in order to process and store samples effectively.
2.1.6 Blood Plasma and Serum
The two most common biomarkers used to assess dietary FA intake, blood plasma and
serum, are the final biomarkers to be reviewed (52). Both blood plasma and serum also contain
multiple lipid fractions between them such as PL, CE, NEFA, etc. (48). For simplicity, any
discussion of plasma or serum can be assumed to be the total analysis of all individual lipid
fractions, unless otherwise stated.
The process of preparing blood plasma and serum for analysis from whole-blood differs
between each of the fractions. In a whole-blood sample, the component which blood cells are
suspended in is the blood plasma, which comprises ~55% of the total blood volume (61). To
prepare plasma, once a blood sample is collected an anti-coagulant is added which prevents
13
clotting and is then centrifuged to separate the plasma from the remaining blood cells (61).
Serum preparation differs as the collected sample is allowed to clot and is then centrifuged
followed by removal of the separated serum layer (62). Upon collection plasma samples typically
require refrigeration, making serum samples slightly easier to prepare and store (62). One
notable compositional difference between the fractions is the lack of anticoagulants in serum
results in fibrinogen being present in the sample (63). Additionally, the use of anticoagulants in
plasma has sometimes interfered with samples in proteomic research (61). With proper collection
and subsequent storage, plasma and serum samples have the best long-term FA stability of any
biomarker assessed with some studies reporting samples remain stable for up to 10 years (56).
Having samples stored in a freezer at -80°C is the general consensus amongst researchers to be
most effective at reducing FA degradation over long periods of time (12,56). If samples are
unable to be kept at -80°C they must be treated with additives such as butylated hydoxytoluene
to maintain the sample integrity and reduce FA loss but, this process may further modify FA
status during processing (56).
Blood plasma and serum have long been thought to be interchangeable for FA analysis
but, little research has examined this phenomenon. A 2011 study was able to determine
subgroups of PL between plasma and serum samples were strongly correlated to each other but,
they did not examine individual FA (64). In an attempt to directly answer this question, our lab
conducted a study comparing fifty matched plasma and serum samples of middle-aged males
which can be found in Table 8-1 and Table 8-2. No significant differences were found between
plasma and serum samples of the nine select FA analyzed, using both absolute and relative
measures. We concluded plasma and serum samples can be confidently used interchangeably,
14
and results from studies using either fraction to assess individual FA levels are able to be
accurately compared. Thus, the use of plasma and serum as biomarkers of dietary FA intake will
be discussed in tandem.
Plasma and serum are biomarkers which best reflect short-term dietary intake, with
notable FA changes being seen within less than 24 hours (42). The time it takes to fully
incorporate FA via diet/supplementation into plasma or serum varies on the quantity and the
class of FA consumed, which can range from three days to two weeks (3,65–67). Certain lipid
fractions, such as PL, NEFA, and CE, can become fully saturated at a faster rate and achieve
greater relative compositions compared to total plasma and serum analysis (65,67). Additionally,
certain plasma fractions have seen weak correlations to dietary FA intake when compared to AT
(68). A recent study found that total plasma correlates to dietary intake of FA equally, if not
better, than individual fractions (65). Assessment of total plasma and serum reduces the need to
fractionate samples and conduct additional analysis, saving time and resources during large-scale
epidemiological studies (65). The decreased timeframe of which changes to dietary FA levels are
seen in plasma and serum result in a more appropriate biomarker for large-scale cross-sectional
studies as it reflects current dietary FA status more accurately than AT or RBC. Furthermore, a
number of studies have validated plasma and serum as accurate biomarkers for assessing dietary
FA (3,42,44,53,65,66,69–71). Similar to RBC, correlations of omega-3 and omega-6 PUFA to
dietary intake are strong in plasma and serum (3,44,53,65,66,69,72) but, results with ALA have
been inconsistent (6). Correlation of MUFA intakes to plasma and serum tend to be relatively
low, with the exception of OA which has seen moderate correlations in some studies (53,65,73).
SFA correlations seen in plasma and serum are similarly inconsistent to that of RBC but, some
15
studies have found pentadecanoic (C15:0) and margaric acid (C17:0) to have a moderate
correlation to dietary intake (65,66,69,74). Compared to RBC, a majority of the FA classes had a
similar strength of correlation to the dietary intake in plasma and serum samples. When directly
compared, some studies have found plasma to be the more reliable biomarker for assessing
dietary FA status (72). The most notable difference was seen in the moderate correlation of two
odd-chained SFA which are markers of dairy intake and have limited capacity to be
endogenously synthesized in humans, unlike other SFA previously discussed (74).
Large fluctuations which can occur with short-term modifications to dietary intake are a
notable limitation of plasma and serum samples (54). Large spikes in FA status have been seen
in plasma and serum samples mere hours following a meal (42). Due to this rapid change,
participants must fast before a plasma or serum sample is collected or, an artificial increase in
FA may be mistaken as regular intake (54). RBC are less resistant to short-term dietary intakes
due to their ~120 day half-life and certain PL fractions of RBC have actually been found to
reflect intakes of SFA and omega-6 PUFA similar to that of plasma (66). Additionally, when
compared to RBC, plasma exhibits a larger intra-individual variability of PUFA values (53,69).
Although this may reduce the strength of correlations in certain instances, the larger dynamic
range of plasma values could benefit statistical analysis when establishing appropriate reference
ranges. Overall, as plasma and serum samples are easily and commonly collected, exhibit long-
term stability, and strong correlations to dietary FA intake, they are an appropriate biomarker for
use in large-scale epidemiological studies. Establishing consistency amongst the dietary
assessments used in studies for evaluating dietary FA trends will help improve reliability and the
ability to compare results moving forward. This will not only aid in establishing reference ranges
16
of healthy individuals but, be directly applicable for clinical use and monitoring the risk of
chronic disease (75). Thus, the use of plasma and/or serum samples as the primary method of
monitoring dietary FA status in studies which aim to establish FA reference ranges are
encouraged moving forward.
2.2 Lipid Methodology and Reporting of Fatty Acid Status
Once a biological sample has been collected, three steps must occur before individual FA
can be quantified. First, lipids must be extracted from the collected plasma and serum samples in
preparation to be analyzed (76). The gold-standard of lipid extraction protocols are the Folch and
Bligh-Dyer methods which are both routinely performed in scientific research (77–79). Second,
extracted lipids must undergo further derivatization into fatty acid methyl esters (FAME) in
order to analyze individual FA (80). Based on the lipid fraction being analyzed (eg. plasma,
RBC, PL, NEFA) specific techniques to derive FAME must be used which require proper
validation studies (12). For plasma and serum samples, boron trifluoride and hydrochloric acid
are commonly employed for derivatization due their strong validity and time-efficient process
(81). Third, once converted to their FAME derivates, a number of different chromatography
methods can be performed to separate and quantify individual FA from the sample (80). Gas
chromatography (GC) is more commonly utilized than liquid chromatography for FA analysis
(82). Analysis of FA via GC can be further specialized by the use of different detectors, the most
common being: GC with flame ionization detector (GC-FID) and GC with mass spectrometry
(GC-MS) (80,82). These techniques have increased throughput of samples in large
epidemiological studies, with a recent study including 160,000 samples analyzed (12). GC
17
techniques and methods of reporting FA values will be discussed in respect of analysis of total
plasma and serum samples.
2.2.1 Gas Chromatography
GC is a common analysis conducted in analytical chemistry in which compounds are
vaporized, separated and subsequently quantified (83). It is termed ‘gas’ chromatography due to
the mobile phase, containing the compound(s) of interest, consisting of a steady flow of an
inert/unreactive carrier gas (83). In brief, a robotic, automated sampler uses a microsyringe to
introduce the sample automatically to a heat controlled inlet/injector (83). Inside the injector, the
sample is heated and vaporized where it is then injected into the mobile phase and passes into the
GC machine (83). All GC machines contain a long, coiled column that is temperature controlled
by an internal oven which the mobile phase passes through (83). An inner stationary phase that
lines the interior of the column functions to separate individual components contained within the
mobile phase which passes through the column (83). The time it takes the compound, or in this
case individual FA, to travel through the column is known as the retention time (RT) (84). The
size, polarity, and molecular mass of individual FA allows them to separate while passing
through the column, with a longer RT in the mobile phase allowing for a greater separation and
quantification of FA (85). A number of different combinations of columns and stationary phases
can be used which impacts the RT of a sample and the subsequent number of individual FA
included in the profile (42). Increasing the column length (eg. 30m vs 100m) will increase the
RT allowing for a greater separation of individual FA and a larger subsequent pool of FA to be
quantified (84,86). However, this will also increase the run time of each sample, reducing
throughput for the sake of a more comprehensive FA profile (42). Upon leaving the column, the
18
sample reaches a digital detector which is able to quantify the compound and is then exported
digitally as a chromatogram, as seen in Figure 2.1 (83). Internal standards are often added to
samples prior to the lipid extraction phase in order to determine accurate quantification of results
once values are detected (85). Specialized software is used by researchers to check and analyze
the resulting chromatogram and determine the quantity of substance detected by the GC (83). A
calibration factor is able to be derived if internal standards are added quantitatively, allowing
researchers to determine the percent composition and/or absolute concentration of each
individual FA with respect to the sample analyzed (12). The use of odd-chain SFA, which only
trace amounts are produced endogenously and consumed via diet such as: C17:0, C19:0, and
C21:0, are most commonly used (12,85–87).
19
Figure 2.1 Screenshot of a chromatogram in OpenLab CDS EZChrom Edition 3.2.1
20
2.2.2 Gas Chromatography Mass-Spectrometry
Where GC machines primarily differ in their overall function is the detector used for
quantification of the substance being measured. As the name would suggest, GC-MS utilizes
mass spectrometry to measure the mass-to-charge ratio of different ions and produces a signal on
a per mole basis of numerous analytes at once (83,87). This technique coupled with GC allows
for specific structural information, along with quantification, of the FA present to be determined
via the location of double-bonds (87). Mass spectrometry is a powerful tool but, method
optimization such as specific derivation of lipids and optimal instrumental settings are required
to quantify results (82,85). The response of GC-MS to certain FAME can be quite varied if the
machine has not been calibrated to the specified structure of the analytes being analyzed (87).
Parameters specifying a range of mass-to-charge values expected of the analytes must be set and
the total ion count based on integrating peaks in the chromatogram must be completed in order to
accurately quantify values (80). Once this is established the sensitivity and specificity can be
extended to the selective ion monitoring (80). When methods of GC-MS are optimized, it is very
effective at quantifying and identifying geometric isomers of PUFA and conjugate linoleic acids
(CLA) (82). The sensitivity and selectivity of GC-MS allows it to more effectively determine
structural information and using selective ion monitoring, separate peaks from ‘noisy’ samples
(80). Although GC-MS yields similar quantifications compared to GC-FID, the specificity of
methods and the resulting cost of solvents reduces throughput and hinders the use of GC-MS in
large-scale studies (42,82). Current studies are working on procedures to create higher
21
throughput of GC-MS by establishing less specific parameters to accurately quantify individual
FA (87,88).
2.2.3 Gas Chromatography with Flame Ionization Detector
GC-FID is the most well-established, robust, and commonly used detector for quantifying
individual FA status used in scientific literature today (81,82,87,89). Once the mobile phase
containing the analyte passes through the column, it is mixed with hydrogen and undergoes
combustion inside the flame ionization detector where the ions released are detected by an
electrometer (87,90). Each FA is quantified on a chromatogram as a peak on a horizontal
baseline (see Figure 2.1) which represents the amount of ions released during combustion on a
per milligram basis (12,87). Compared to GC-MS, GC-FID requires far less specific parameters
to conduct and produces a nearly identical response factor to all FAME derivatives when
analyzed (87). The response however is far less specific, as no structural information is able to be
determined during GC-FID (80). Individual FA are determined by their RT when passing
through the column which is proportional to the chain length and number of double-bonds
present in the FA (82,87). Flame ionization detectors are able to detect even trace amounts of FA
but, trans-isomers and CLA are quite challenging to precisely determine (42,90). The RT of FA
in both GC-MS and GC-FID have been found to be nearly identical, resulting in unidentified,
mislabeled, or overlapping peaks following GC-FID able to be validated against GC-MS (89).
This process of total combustion results in quantitative measures of GC-FID being nearly
universal amongst organic compounds, exhibit long-term stability, low cost, excellent precision
and strong linearity (80,82,87,90).
22
Other methods of chromatography are also commonly used, such as thin-layer
chromatography which can further fractionate plasma and serum prior to GC analysis (82).
Compared to total plasma or serum analysis alone, this process is time-consuming, and based on
the results of Furtado et al. (65) does not provide stronger correlations to dietary FA intake. The
use of GC-FID is widely applicable for the quantification of individual FA. Although this
detector is limited by a nonspecific response lacking structural information of the analyzed
compound and an inability to detect specific impurities, it is the most stable and accurate method
widely available (80,91). For all GC quantifications, coefficient variance (CV) is used as a
guideline to determine the amount of error present in a sample (92). Using CV as a guideline
when repeating sample measures can be valuable to determine the amount of error which may be
present due to instrumentation, procedure, or inter-tester differences. However, acceptable CV
ranges have yet to be formally established for FA values obtained from GC-FID (65). It has been
previously suggested a CV range of 15-20% is considered to be a reasonable (92), while Furtado
et al. (65) reports less than 20% and upwards to 30% is common. A study examining different
labs from across the world using GC to analyze individual FA levels of plasma and serum
samples found a majority agreed within 20% CV but, there was variability between individual
FA (93). Quantifying FA values of total plasma and serum using GC-FID results in a simplified
procedure, higher throughput, and reduced cost, which is more effective for large-scale studies.
Furthermore, adequately detailing both extraction and derivatization techniques are important for
establishing more consistent practices in lipid research. Recent studies have already begun
utilizing this methodology to comprehensively profile individual FA in global cohorts with the
goal of developing FA reference ranges (94,95).
23
2.2.4 Units of Measure
The final component involved in assessing dietary FA intake is the method in which
results are reported. As mentioned previously, results of GC analysis are exported as a
chromatogram which indicates the quantity of individual FA that was detected as a ‘peak’ on a
horizontal baseline (see Figure 2.1) (83). Internal standards are commonly added to samples
prior to GC analysis and are used to mathematically derive quantitative results (85). Currently,
the method in which dietary FA results are reported is not standardized in scientific literature
(12). The two most common methods of reporting results are as a relative percent composition,
or absolute concentration (12). The manner in which FA status is reported has been found to
significantly impact the correlation and interpretation of dietary status (95). Relative percent
composition and absolute concentration will be reviewed in respect to establishing dietary FA
reference ranges.
2.2.4.1 Percent Composition
Individual FA levels representative of dietary intake are most commonly reported as a
relative percent composition (12,95). Percent composition reflects the total value of an individual
FA divided by the total of the entire FA pool analyzed represented in terms of percentage (12).
This method is very simple to calculate, the units are universal, and is already commonly used to
reflect FA status, such as the Omega-3 Index which correlates to CVD risk (54). Intra-individual
CV when calculated using relative percentages have also been found to be lower compared to
absolute measures (69). This is likely due to values reported as a percentage being limited to a 0-
100 scale, reducing the dynamic range compared to absolute values. Additionally, the total pool
of FA can have large implications on the relative percentage reported. Currently, there is no
24
standardized list for which FA must be included in the pool analyzed, with some studies
reporting <5 and others >60 individual FA (12). The number of individual FA able to be
analyzed at one time is largely dependent on the column used during GC; a longer column allows
for more separation and a larger number of individual FA to be detected (86). As relative
measures are based on the percentage of the total pool of FA included in analysis, the smaller the
pool analyzed equates to a larger apparent percentage of individual FA (12). Although a larger
number of FA assessed has useful implications for establishing reference ranges, CV of difficult
to detect FA, such as a trans-fats and CLA, are more varied (69). Percent composition is also
limited by the direct influence independent FA can have on subsequent calculations of the
remaining pool (69). For example, if large increases in one FA are detected from timepoint A to
timepoint B, the value of the total pool is also increased. When other FA in the same pool are
analyzed, even if no changes are detected from timepoint A to B, their overall percentage would
be decreased as the value of the total pool used in the calculation would have increased. Previous
studies have found the correlation of relative and absolute values of some FA to be poor, which
may be the result of the total FA pool influencing individual FA values (69,95,96). Overall,
although the simplicity of the calculation results in an easily replicable process, the influence of
independent FA on final values limits the accuracy of reporting FA data via this method.
2.2.4.2 Concentration
The other common method of reporting FA values is in terms of an absolute
concentration which reflects the average mass of FA per unit of fluid or tissue (12). Absolute
concentrations are calculated via a calibration factor derived from the use of an internal standard
applied to the detected value of an individual FA (12,85). Concentrations are typically reported
25
as either μg/mL or μmol/L depending on the calibration factor but, are easily converted between
each other if the molecular mass of the FA is known (12). Of note, relative percentages are able
to be derived from absolute concentrations if they are the only method reported but, the same is
not possible for the reverse. Comparatively, absolute values have a larger dynamic range than
relative values as they are not limited to a 0-100 scale, and have values often in the 1000’s. For
statistical analysis of concentrations, data sets may result in non-normal distributions but, for
establishing reference ranges, can be more precise as no variance is being reduced. Additionally,
absolute concentrations are not directly influenced by the status of other FA as the calculation
does not utilize relative values in its computation (12). When both methods of FA reporting have
been used in the same study, trends in the status of certain FA have found to differ depending on
whether concentration or percent composition is reported (69,94,95). Studies reporting both
methods can provide a more in-depth understanding, and greater context to the values being
assessed in cross-sectional studies. For the purposes of developing reference ranges of individual
FA, reporting results as absolute concentrations to eliminate the influence of other FA and
improve generalizability across studies is recommended (94,95).
2.3 Developing Individual Fatty Acid Reference Ranges
Due to their diverse biological function, maintaining adequate dietary FA status is vital for
promoting optimal health and has been implicated in reducing the risk of a number of chronic
diseases (4). However, there are currently no established reference ranges for individual FA
which denotes “healthy” values. This fundamental gap in FA research hinders the translation of
scientific results into clinical practice. Baseline FA values would provide medical practitioners
context regarding the direction and magnitude of change necessary to achieve optimal health in
26
patients with deficient, suboptimal, or even excess amounts of FA. Without context of what
individual FA levels equate to adequate status, understanding the relationship between the values
reported in scientific studies and human health is currently limited. Values which are reported in
scientific literature are often inconsistent for that matter, using different units or methods of
reporting altogether (12). Establishing universal practices in dietary FA assessment and reporting
are necessary steps to create consistent and generalizable results between individual FA status of
different cohorts. A majority of the studies that have profiled FA values have done so in
participants with chronic disease which can result in modified levels of individual FA (9). Other
studies have completed large-scale assessments of FA status in healthy individuals but, results
are reported as a relative percentage which reduces the generalizability of results as individual
concentrations are unable to be derived from percent values alone (97,98). To date, only a
handful of studies have comprehensively profiled plasma/serum FA status in healthy adults with
the primary objective of developing reference ranges, which are highlighted in Table 2.1
(94,95,99,100). Of these studies, only one took place outside of NA, and another contained
participants with diabetes and/or metabolic syndrome (95,100). This section will highlight the
effect FA status can be modified by certain disease-states, along with genetic and regional
differences, and how they are important to consider in the process of developing FA reference
ranges.
27
Table 2.1 Studies profiling ranges of fatty acids in different cohorts
Authors Population # of
Subjects Collection
Year(s) Biomarker Analysis FA
Assessed Units Abdelmagid et al. 2015 Canadian Mixed-Sex 826 2004-2009 Total Plasma GC-FID 61 μmol/L
Sergeant et al. 2016 United States Mixed-Sex 152* 2003-2007 Total Serum GC-FID 29 μmol/L
Bradbury et al. 2011 New Zealand Mixed-Sex 2793 1996-1997 Serum PL, CE,
TAG GC-FID 20 %
Sera et al. 1994 United States Mixed-Sex 130 N/A Total Serum GC-FID 11
mg/L &
μmol/L *59 samples obtained from a diabetes/metabolic syndrome cohort
28
2.3.1 Disease States
All things being equal, changes to dietary consumption plays the largest role in
modifying plasma/serum FA status. For developing reference ranges, there a number of
additional factors which can further impact FA status that need to be considered when
establishing baseline values and assessing health-status. One such factor is the impact chronic
disease may have on FA status. When compared to healthy individuals, participants with a wide
range of health issues such as nonalcoholic fatty liver disease, pancreatitis, type I and II diabetes,
cancer, etc. have displayed modified plasma and serum levels of individual FA (101–109).
Plasma FA values have also been found to strongly correlate with increases in the body mass
index (BMI), as obese individuals, with or without metabolic syndrome, have modified FA status
compared to healthy controls (110–112). These differences in FA status provide strong rationale
for the evaluation of “healthy” individuals to establish reference ranges as an effective biomarker
to assess disease risk. However, further research will be necessary to profile the FA status
reflective of specific disease states in order to establish cut-off FA values which equate to
disease risk. For example, long-chain SFA were found to be increased in colorectal cancer
whereas increased MUFA were seen in patients with pancreatitis (102,106). Whether these
changes in FA status are brought on by the disease itself or genetic alterations which may persist
following remission is important to consider (106). For future research objectives, correlating FA
status to a wide range of biomarkers reflective of chronic disease risk are necessary. (106).
2.3.2 Genetic Differences
Assessing individual genetic differences independent of complications derived from
chronic disease is also important to consider when developing FA reference ranges. Inter-
29
individual differences in metabolism, dietary absorption, or endogenous lipid production can
modify individual FA status (3). Whether healthy ranges require further stratification into groups
such as BMI, age, or sex, have yet to be fully elucidated. As discussed previously, plasma FA
values exhibit a strong correlation to increases in BMI and have found to be increased in most
obese individuals (112,113). In obese adolescents, plasma FA were modified when undergoing a
weight loss program which correlated to improvements in biomarkers related to metabolic
disorders (114). However, one study found obese individuals and those with metabolic syndrome
exhibited very minor differences in plasma FA status compared to healthy controls (111).
Further, a subgroup of obese individuals classified as “metabolically healthy” and disease-free
when compared to “metabolically unhealthy” individuals, exhibited FA levels consistent with
those at decreased risk of CVD (115). Kang et al. (110) found that only overweight individuals
with an increased area of visceral fat had modified plasma FA profiles. Taken together, the use
of obese but, otherwise healthy individuals for developing FA reference ranges may still impact
interpretations of results and stratification may be necessary.
Sex-specific differences in individual FA status have been shown on a number of
occasions (66,111,116,117). Independent of dietary intake, females have a greater genetic
capacity to endogenously produce omega-3 PUFA due to increased Δ6-desaturase activity (118).
The use of oral contraceptives in some studies have found to further impact PUFA metabolism
but, more recent studies found no significant effect (119,120). Additionally, pregnant and
breastfeeding females have obvious metabolic differences and subsequent nutritional needs
compared to males which has also been found to modify FA status (121). Whether these
differences in FA status result in biologically significant relationships to disease biomarkers
30
requires further exploration. Based on these distinct sex-related differences in FA status,
comparing the results of males and females to establish separate baseline values may be
necessary.
Numerous studies have also found significant age-related changes in FA status (122). As
our global population continues to grow older, FA metabolism has been found to be further
modified by adaptations brought on by old age (27,123). Not only is dietary intake of FA
decreased in the elderly, they can also suffer from deficient absorption of essential dietary
nutrients even if intake quantities meet dietary standards (124,125). In particular, MUFA values
have seen significant age-related decreases in dietary consumption via reduced Δ9-desaturase
activity (126). Additionally, previous studies have seen age-related increases in plasma omega-3
PUFA across a number of populations globally (122,127–129). When compared to young
individuals, elderly individuals have a greater capacity of incorporating omega-3 PUFA into
blood plasma (129). As studies in elderly individuals showcase distinct dietary and genetic
modifications which impacts FA status, considering the effect of age as modifying factor for
developing reference ranges is necessary.
2.3.3 Geographical Differences
Dietary intake of FA is the strongest modifier of plasma FA status in humans (130). With
advancements in technology and urbanization, we currently live in a time with the greatest
access to largest variety of food products in human history (131). However, based on national
recommendations, cultural preferences, and accessibility of certain products, the intake of
specific nutrients and macronutrients can differ based on ones geographic location (132). When
examined on a global scale, consumption of FA exhibit dramatic variance between, and even
31
within countries (10). Intakes of SFA in NA are similar to that of Australia, Russia, and many of
the countries in Europe but, are consistently higher than that of China, and all countries in
Central and South America (10). Although global trends dictate trans-fats intake is decreasing,
NA still exhibits higher levels than a majority of the countries in the world (10,133).
Interestingly, some countries may exhibit similar trends in FA intake but, as the result of
consuming different food sources (132). For example, omega-3 PUFA are most commonly
consumed via seafood and to a lesser extent, plant-based foods such as flax (52). In NA, omega-
3 intake via seafood is quite low compared to Southeast Asia but, trends are reversed when
examining plant-based consumption of omega-3 PUFA (10). As our global population has
become more diverse and less homogenous, ethnic/cultural variances in diet are likely to exist
within sampled populations on a national or even regional scale (132). In the context of
developing FA reference ranges, these variances in dietary fat intake can have a significant effect
on the generalizability of results between different cohorts. Whether a global value for individual
FA is possible or, region specific values must be used remains to be seen. It is necessary when
profiling cohorts of different nationalities/ethnicities to assess what regional or cultural
influences on diet may account for the variance in FA levels examined.
2.4 Conclusion
In conclusion, the development of individual FA reference ranges would help to address a
fundamental gap within lipid research. Further unifying best practices for assessing and
quantifying individual FA will help to improve reproducibility and the scale of which studies can
be conducted. The use of blood plasma/serum samples as biomarkers of FA status quantified by
GC-FID have exhibited high throughput with cost-effective and consistent results which have
32
been validated in numerous studies. Reporting FA as absolute concentrations reduces the
influence large fluctuations of individual FA may have on the trends of the pool analyzed and
still allows for relative percentages to be calculated. Differences in both dietary intake and
genetics can modify plasma/serum FA status. In healthy individuals, evaluating the relationship
BMI, sex, age, and variations in dietary status reflective of cultural/ethnic trends should be
considered when conducting studies and comparing results. Overall, developing FA reference
ranges would provide context to values reported throughout scientific literature and give clinical
practitioners a simple and effective tool for evaluating chronic disease risk.
33
3 RATIONALE, OBJECTIVES, HYPOTHESIS, AND EXPERIMENTAL DESIGN
3.1 Rationale
Due to their structural and functional properties, lipids are vital to human health (2). To
date, a number of studies have been published outlining the necessity of obtaining adequate
dietary consumption of specific FA to promote optimal health and reduce chronic disease risk
(4). The use of blood biomarkers to assess dietary FA status is a commonly used, validated
scientific procedure (52). However, interpretation of FA status for assessing chronic disease risk
is currently limited due to the lack of established reference ranges. Established clinical ranges for
other lipids such as cholesterol, triglycerides, and total free fatty acids are routinely utilized in
clinical practices for assessing the risk of chronic disease (11). This gap in literature can be
addressed by utilizing the current best practices in FA assessment and reporting to
comprehensively profile the FA status of large cohorts. To our knowledge, very few studies have
assessed FA with the intent of reporting reference ranges, and no studies have examined
Singaporean or Australian cohorts. Additionally, the reporting and methodology of such FA data
is often inconsistent throughout literature (12). Thus, we will focus on using consistent
methodological approaches that align with current best practices to improve reproducibility and
generalizability of FA values between studies.
3.2 Objectives
The overall objective of this thesis is to assess the relationship between individual FA and
chronic disease biomarkers to add valuable knowledge to the growing work towards establishing
validated FA reference ranges. Secondly, by evaluating different international cohorts it would
34
give us an opportunity to profile FA status in that region and evaluate what impact region
specific diet and cultural factors may affect FA values. Specific objectives and hypotheses
include:
1. Objective 1. To examine the relationship between serum FA values and aging in
Australian men and how reporting results as an absolute concentration or percent
composition may impact the interpretation of results.
i. Hypothesis 1. Overall FA status in Australian males will exhibit negative
correlations as a function of age but, omega-3 PUFA will be positively
correlated with age.
2. Objective 2. To comprehensively profile serum FA concentrations of Singaporean adults
and investigate their association with established lipid biomarkers and health outcomes.
ii. Hypothesis 2. Saturated and unsaturated fatty acids will be positively and
negatively associated with the lipid biomarkers, respectively. It is anticipated
that a comprehensive analysis will reveal that novel, minor fatty acids, may also
differentially influence lipid biomarkers and health outcomes.
3.3 Experimental Design
International Cohorts:
Blood plasma and serum samples were obtained via the Men Androgen Inflammation
Lifestyle Environment and Stress (MAILES) from the University of Adelaide in Australia, and
via the Agency for Science, Technology and Research (A*STAR) in Singapore (134,135). The
35
MAILES cohort was established in 2009 by combining the population of two studies: the Florey
Adelaide Male Ageing Study and North West Adelaide Health Study (134). The entire cohort
consisted of ~2600 males with data collected from 2002-2010 to investigate associations such as
sex steroids, and psychosocial factors with cardio-metabolic disease risk (134). The Singaporean
cohort consisted of ~450 participants with data collected from 2014-2017 to evaluate various
measures of body composition in an Asian population (135).
Gas Chromatography:
A modified Folch method was used for lipid extraction and derivation of plasma/serum
samples which has been described in-detail previously (77). In brief, samples were first thawed
on ice, where 50µl of plasma/serum was pipetted into a 15ml tube. Stock solution of 2:1
chloroform:methanol (v:v) with a 3.33 µg/mL of internal standard (C19:0 free fatty acid) was
prepared and added to the tube. Samples were spun and the lower chloroform layer was
transferred into a clean 15 ml glass tube and dried down under a gentle stream of nitrogen. 300
µL of hexane and 1 ml 14% BF3-MeOH was added and followed by methylation at 100oC for 60
minutes in oven. 1 mL of double-distilled water and 1 mL hexane was added to stop the
methylation process. Sample was once again spun down to separate phases and top hexane layer
was then extracted and then reconstituted in 400µL of hexane. Total FA were quantified on an
Agilent 7890-A GC-FID and separated on a SP-2560 fused silica capillary column (100m x
0.25mm x 0.2um) (Sigma, Cat # 24056). Total run time was ~49.78 minutes per sample. Peaks
were identified and chromatographs were analyzed via OpenLab CDS EZChrom Edition 3.2.1
software. C19:0 internal standard was used to derive absolute and relative FA concentrations
from the results obtained from the exported chromatographs.
36
Statistical Analysis:
Statistical analysis was conducted primarily using SAS University Edition (SAS Institute
Inc.; Cary, North Carolina). In Chapter 4, a Tukey’s HSD test was used to determine significant
differences between biomarkers and FA concentrations of participants stratified by age.
Additional analysis was conducted using R version 3.6.12 (R Foundation for Statistical
Computing, Vienna, Austria). A multiple regression analysis was conducted to determine the
effect of ageing on each FA including multiple covariates for all participants. Correlations
between individual FA concentrations and composition with age were expressed as R. In Chapter
5, a Student t-test was used to determine statistical differences between male and female
participants. A linear regression model was conducted to assess the relationship between FA
concentrations and lipid biomarkers. Correlations were reported as R2 and standardized
parameter estimates/beta-weight (β) of age. Sixty-seven total FA were included in the bulk of the
analysis with ten FA of interest selected for more detailed analysis. A p-value of <0.05 was
considered statistically significant.
37
4 AGING IS ASSOCIATED WITH DIFFERENTIAL CHANGES OF CIRCULATING SERUM FATTY ACID LEVELS IN AUSTRALIAN MALES
4.1 Abstract
Advanced aging is often implicated in a number of detrimental metabolic changes and
modified dietary habits that can lead to fluctuations of individual FA status as one ages. With the
number of individuals >65 years old continuing to increase globally, effective monitoring of
changes in FA status can prove to be a useful tool for assessing adverse health status. Sixty-
seven individual FA were analyzed using GC-FID from serum samples collected from 870
Australian males ranging in age from 40-85 years old. Individual FA status was reported as both
percent composition (%) and absolute concentration (μmol/L). In terms of percent composition,
twenty-two FA exhibited a significant correlation with age (<0.05), with eleven FA exhibiting
negative trends. In terms of absolute concentration, forty-two FA exhibited a significant
correlation with age (p<0.05), with only the omega-3 PUFA EPA and DHA exhibiting
significant positive trends in both measures. Ten biologically relevant FA were further stratified
by decade (40-49 y, 50-59 y, 60-69 y, 70+ y) where a distinct negative correlation with age
beginning at the 50-59 y decade was seen. Overall, metabolic and/or dietary factors associated
with aging negatively impacts individual serum FA status, excluding positive trends seen with
EPA and DHA, and is most accurately assessed using absolute concentration rather than percent
composition.
38
4.2 Introduction
As our global population continues to age, substantial declines in cognitive status coupled
with increased rates of chronic disease such as CVD, rheumatoid arthritis, diabetes and many
other ailments are commonly seen (136–139). In Australia, the 65 year-old+ cohort has increased
to over 15% of the total population and the ≤19 year-old cohort has been in decline over the last
decade (140). CHD was the leading cause of death in Australia in 2012, and there is an estimated
1.2 million Australians currently diagnosed with a CVD-related ailment (136,141). A number of
studies have found significant age-related changes in circulating FA status which from both a
structural and functional standpoint, play a vital role in the maintenance of human health. (27).
Throughout the body, cell membranes are comprised of FA that are an essential component for
proper functioning and aid in processes such as inflammation, cell signaling, growth, and
development (1,142). An immense body of work has been published detailing the importance of
maintaining healthy FA status in chronic disease states such as obesity, cancer, Alzheimer’s
disease, and CVD (143–146). Increasing omega-3 PUFA status has shown to decrease levels of
serum triglycerides and overall risk of CHD mortality and cardiovascular events (147). Increased
omega-3 PUFA consumption has also been found, in certain instances, to improve declines in
cognitive function with old-age and may also have a beneficial effect in the early stages of
Alzheimer’s disease (148,149). As the elderly are quickly becoming the fastest-growing segment
of the population globally, adequate consumption and monitoring of FA status is a valuable
clinical tool for assessing health risks associated with advanced aging (123).
Assessing whether adequate quantities of FA are being consumed poses a unique set of
challenges however, as estimates from dietary assessments such as diet records are imprecise
39
compared to biological measurements (13). Modifying dietary intake is only one of many factors
which impact FA status, as altered metabolism brought on by chronic disease states and genetics
also play a role in altering FA levels (3,94,122). For example, our bodies are able to be
endogenously produce certain SFA and MUFA via the use of Δ9-desaturase, whereas certain
long-chain PUFA are essential and must be consumed via the diet (3). EPA and DHA can also be
produced endogenously but, have very limited conversion rates of 1% or less (8). Long-chain
PUFA therefore exhibit a stronger correlation to dietary intake compared to SFA and MUFA
based on our individual capacity to produce Δ9-desaturase (3). Additionally, genetic carriers of
the apoE e4 allele are not only at an increased risk of Alzheimer’s but, exhibit reduced levels of
EPA and DHA following PUFA supplementation (150). Whole blood and its components:
plasma, serum, erythrocytes, etc. are commonly analyzed biomarkers to determine FA levels
which exhibit strong correlations to dietary intake of FA (42). Blood samples act as an aggregate
reflecting both dietary and genetic/metabolic factors, providing a more specific assessment of FA
levels than diet records alone. Consistent monitoring of FA status could provide valuable
information for assessing trends in elderly adults which may be the result of age-related changes
if dietary status remains unchanged.
Reporting of individual FA status is inconsistent throughout scientific literature. Gas
chromatography is a common method where fasted blood is analyzed to determine individual FA
levels (77). Individual FA levels representative of whole-body status in a fasted state are either
reported as a relative percent composition (%), the proportion of one FA out of the entire pool
analyzed, or as an absolute concentration (eg. μmol/L, μg/mL) (12). Currently, there is no
standardized list for which FA must be included in the pool analyzed, with some studies
40
reporting <5 and others >60 FA (12). The number of individual FA able to be analyzed at one
time is dependent on the column utilized during gas chromatography; a longer column allows for
more separation in the gas phase and a greater amount of FA to be detected compared to a
shorter column (86). As relative measures are based on the percentage of the total pool of FA
included in analysis, the smaller the pool analyzed results in a larger apparent percentage of
individual FA (12). Large fluctuations in one FA could also alter the subsequent percent
composition seen in other FA included in the same pool when no biological changes were
actually seen. Absolute concentrations, however, are not directly influenced by the status of other
FA and thus, may be more indicative of true trends seen in individual FA. This study aims to
examine age-related effects on the status of sixty-seven FA in Australian males. FA status will be
evaluated as both absolute concentration and percent composition to assess the similarities and
differences between each measure during the aging process.
4.3 Materials and methods
4.3.1 Participants
Participants have been previously described by (134). In brief, the MAILES cohort
consisted of Australian males who were 40-85 years of age (n=870) and had completed a 10 to
12 hour fasting procedure prior to blood collection. Participants were later sorted by decade of
life: 40-49 y: n=197, 50-59 y: n=246, 60-69 y: n=237, 70+ y: n=191. Each participant had blood
drawn that was treated by Grant et al. (134) to separate blood serum. Informed consent forms
have been previously completed during the MAILES cohort data collection period (134).
Samples were stored at -80°C as previously discussed by (94). Eight years after initial blood
41
sample collection serum samples were analyzed at the University of Guelph where they
underwent total serum FA analysis via GC-FID.
4.3.2 Gas Chromatography Protocol
A modified Folch method was used for lipid extraction for serum samples which has been
described in-detail previously (77). In brief, samples were first thawed on ice, where 50µl of
plasma or serum was pipetted into a 15ml tube. Stock solution of 2:1 chloroform:methanol (v:v)
with a 3.33 µg/mL of internal standard (C19:0 free fatty acid) was prepared in which 3 mL of
CHCl3: MeOH was added to the tube and capped tightly with a phenolic PTFE liner. After a one
minute of vortex, 550µL of 0.1M KCl was added to each tube and briefly vortexed again for
another 5-10 seconds. Next, samples were spun at 1460 rpm for 10 min (at 21°C) to separate
phases. The lower chloroform layer was transferred into a clean 15 ml glass tube and dried down
under a gentle stream of nitrogen. After, 300 µL of hexane and 1 ml 14% BF3-MeOH was added
and followed with a vortex of 5-10 seconds. Methylation was then completed at 100oC for 60
minutes in oven, where it was then cooled for 10 minutes at room temperature. Next, 1 mL of
double-distilled water and 1 mL hexane was added to stop the methylation process. After a one-
minute vortex, the sample was once again spun down at 1460 rpm for 10 min (at 21°C) to
separate phases. The top hexane layer was then extracted into a clean 2 ml GC vial, dried down
under a gentle stream of N2 and then reconstituted in 400µL of hexane.
Total FA were quantified on an Agilent 7890-A gas chromatograph equipped with flame
ionization detection and separated on a SP-2560 fused silica capillary column (100m x 0.25mm x
0.2um) (Sigma, Cat # 24056). The splitless inlet (pressure of 19.5 psi and a hydrogen flow of
13.7 mL/min) and detector was set at 250°C (hydrogen flow of 30mL/min, air flow of
42
450mL/min and nitrogen flow of 10mL/min), where 1uL of sample was injected onto a splitless
inlet set. The oven was initially set at 60°C and increased by 13°C/min to 170°C, where it was
held for 4 minutes, and then further increased by 6.5°C/min to 175°C (with no hold), by
2.6°C/min to 185°C (with no hold), 1.3°C/min to 190°C (with no hold), and13.0°C/min to 240°C
where it was held for 13 minutes. Total run time was 49.78 min per sample.
4.3.3 Fatty Acid Analysis
Peaks were identified and chromatographs were analyzed via OpenLab CDS EZChrom
Edition 3.2.1 software. An internal standard (C19:0) was used to calculate FA concentrations
(μmol/L) and percent composition (%). In total, sixty-seven total FA were included in the
analysis.
4.3.4 Statistical Analysis
Statistical analysis for Table 4.1 was conducted using SAS University Edition (SAS
Institute Inc.; Cary, North Carolina). A Tukey’s HSD test was used to determine significant
differences between age when participants were stratified by decade of biomarkers and FA
concentrations. Results are expressed as mean ± standard deviation (SD). All remaining analyses
was conducted using R version 3.6.12 (R Foundation for Statistical Computing, Vienna,
Austria). Correlations between individual FA concentrations and composition with age was
expressed as a correlation coefficient (R). Of the sixty-seven analyzed, ten major FA to further
examine correlations between FA status and specific decades of aging. These FA include:
palmitic acid (C16:0), palmitoleic acid (PAO;C16:1c9), stearic acid (C18:0), oleic acid
(C18:1c9), LA (C18:2n6), ALA (C18:3n3), EPA (C20:5n3), docosapentaenoic acid
(DPA;C22:5n3) DHA (C22:6n-3) and ARA (C20:4n6). A multiple regression analysis was
43
conducted with the following covariates included in the model: smoking, blood pressure
medication, depression, anxiety, asthma, BMI and osteoarthritis. A p-value of <0.05 was
considered statistically significant for all analysis.
44
4.4 Results
Table 4.1 General characteristics of study population
Total Population 40-49 50-59 60-69 70+
Population (#) 870 197 246 237 191
Age (yrs) 60 ± 11.16 / / / /
BMI (kg/m2) 28.76 ± 4.52 28.53 ± 4.43 29.07 ± 4.59 29.39 ± 4.93 27.84 ± 3.82c
Glucose (mmol/L) 5.00 ± 1.56 5.14 ± 1.03 5.45 ± 1.73 5.65 ± 1.63a 5.61 ± 1.63a
Insulin (pmol/L) 10.00 ± 8.88 8.86 ± 8.82 9.20 ± 7.42 10.85 ± 10.82 9.45 ± 7.78
HDL Cholesterol (mmol/L) 1.30 ± 0.31 1.28 ± 0.27 1.32 ± 0.32 1.30 ± 0.30 1.37 ± 0.34a
LDL-Cholesterol (mmol/L) 3.10 ± 0.98 3.39 ± 0.90 3.31 ± 0.91 2.97 ± 0.96b 2.72 ± 1.00d
Total Cholesterol (mmol/L) 5.20 ± 1.10 5.45 ± 0.99 5.41 ± 1.06 5.02 ± 1.12b 4.76 ± 1.10b
Triglycerides (mmol/L) 1.72 ± 1.02 1.86 ± 1.12 1.81 ± 1.20 1.69 ± 0.85 1.51 ± 0.78a
45
Total Serum Fatty Acids (μmol/L)
13524.71 ± 4120.15
14409.61 ± 4509.71
13885.52 ± 4058.98
13350.5 ± 4072.61a
12363.46 ± 3539.84b
Data represented as Mean±SD
A p-value of <0.05, determined by Tukey's HSD for differences between groups, was considered statistically significant aSignificant compared to 40-49 bSignificant compared to 40-49 and 50-59 cSignificant compared to 50-59 and 60-69 dSignificant compared to 40-49, 50-59 and 60-69
4.4.1 General characteristics and lipid biomarkers
In Table 4.1, the anthropometric measures, biomarkers, and FA concentrations of 870
Australian males are described. To examine age-related effects, participants were stratified by
age per decade. As participants aged, lipid biomarkers, such as triglycerides (TG), low-density
lipoprotein (LDL), and total cholesterol significantly declined. High-density lipoprotein (HDL)
cholesterol increased significantly in those aged 70+ y compared to those aged 40-49 y. Fasting
glucose increased from 40-49 y to 60-69 y before significant decreases occurred at decade 70+ y.
Mean FA concentrations decreased significantly as participants aged, from 14409.61μmol/L at
40-49y to 12363.46 μmol/L at 70+ y. BMI increased from 40-49 y (28.53kg/m2) to 60-69 y
(29.39 kg/m2) before significantly decreasing at 70+ y (27.84 kg/m2).
46
Table 4.2 Correlation between serum fatty acids and age
Concentration (μmol/L)
Composition (%)
Fatty Acid
Common Name Correlation (R)
p-value Correlation (R)
p-value
12:0 Lauric acid -0.064 0.058 -0.032 0.348
12:1c11
t t t t
14:0 Myristic Acid -0.149 <0.0001* -0.092 0.006*
14:1c9 Myristoleic acid -0.073 0.0301* -0.034 0.311
15:0 Pentadecanoic acid -0.103 0.0023* 0.070 0.0378*
15:1c10
0.047 0.166 0.047 0.166
16:0 Palmitic acid -0.170 <0.0001* -0.010 0.773
16:1t9 Palmitoleic acid -0.077 0.0222* -0.036 0.285
16:1c9 Palmitoleic acid -0.101 0.0028* -0.035 0.307
17:0 Margaric acid -0.103 0.0024* 0.067 0.0486*
17:1c10
t t t t
18:0 Stearic acid -0.205 <0.0001* -0.018 0.599
18:1t4
-0.002 0.958 0.038 0.266
18:1t5 Thalictric acid -0.009 0.788 0.005 0.885
18:1t6-8 Petroselinic acid -0.147 <0.0001* -0.102 0.0025*
18:1t9 Elaidic acid -0.177 <0.0001* -0.119 0.0004*
18:1t10
-0.071 0.0367* -0.048 0.156
18:1t11 trans-Vaccenic acid -0.120 <0.0001* -0.061 0.071
18:1t12
-0.161 <0.0001* -0.100 0.0030*
18:1t13
-0.144 <0.0001* -0.061 0.074
18:1c9 Oleic acid -0.177 <0.0001* -0.066 0.0505*
18:1c11 cis-Vaccenic acid -0.029 0.390 0.078 0.0209*
18:1c12 Octadecenoic acid (cis-12)
-0.092 0.0064* -0.035 0.297
18:1c13
-0.026 0.443 -0.026 0.443
18:1t16
-0.100 0.0030* -0.057 0.090
18:1c14
0.050 0.140 0.057 0.091
47
18:2tt
-0.089 0.0090* -0.043 0.204
18:2t9t12 Linoelaidic acid -0.143 0.0057* -0.035 0.498
18:2c9t13
-0.056 0.100 -0.048 0.161
18:2ct
-0.165 <0.0001* -0.075 0.0273*
18:2c9t12
-0.107 0.00155* -0.012 0.732
18:2t9c12
-0.127 0.0002* -0.066 0.051
19:1c10
-0.027 0.429 -0.027 0.428
18:2n6 Linoleic acid -0.241 <0.0001* -0.079 0.0201*
18:2c9c14
-0.032 0.341 0.012 0.722
18:2c9c15 Mangiferic acid -0.023 0.489 -0.002 0.956
20:0 Arachidic acid -0.126 0.0002* 0.058 0.086
18:3n6 Gamma-linolenic acid -0.229 <0.0001* -0.157 <0.0001*
20:1c5
-0.117 0.0005* -0.059 0.081
20:1c8
-0.039 0.251 0.017 0.620
20:1c11 Gondoic acid -0.097 0.0040* 0.034 0.319
18:3n3 Alpha-linolenic acid -0.059 0.083 0.080 0.0180*
18:2c9t11 CLA Rumenic Acid
-0.126 0.0002* -0.008 0.821
18:2c11t13 Conjugated linoleic
acid -0.016 0.632 0.035 0.305
18:2t10c12 CLA
Conjugated linoleic acid
-0.049 0.150 -0.018 0.605
18:2c/c CLA1
Conjugated linoleic acid
-0.003 0.938 0.030 0.376
18:2c/c CLA2
Conjugated linoleic acid
-0.052 0.123 -0.011 0.755
18:2tt CLA
Conjugated linoleic acid
-0.091 0.0071* -0.005 0.876
18:4n3 Stearidonic acid -0.051 0.129 -0.047 0.161
21:0 Heneicosylic acid -0.146 0.0045* -0.045 0.387
20:2n6 Eicosadienoic acid -0.051 0.135 0.081 0.0171*
22:0 Behenic acid -0.142 <0.0001* -0.010 0.775
20:3n9 Mead acid -0.106 0.0017* -0.028 0.407
20:3n6 Dihomo-gamma-
linolenic acid -0.234 <0.0001* -0.089 0.0086*
48
22:1n9 Erucic acid -0.070 0.0391* -0.045 0.180
20:3n3 Eicosatrienoic acid -0.070 0.0394* -0.041 0.226
23:0 Tricosylic acid 0.043 0.207 0.092 0.0066*
20:4n6 Arachidonic acid -0.185 <0.0001* 0.016 0.640
22:2n6 Docosadienoic acid -0.066 0.051 0.073 0.0308*
24:0 Lignoceric acid -0.129 0.0001* 0.006 0.864
20:5n3 Eicosapentaenoic acid 0.187 <0.0001* 0.248 <0.0001*
24:1n9 Nervonic acid 0.011 0.746 0.141 <0.0001*
22:3n3 Docosatrienoic acid t t t t
22:4n6 Adrenic acid -0.238 <0.0001* -0.117 0.0006*
22:5n6 Docosapentaenoic acid (Osbond acid)
-0.194 <0.0001* -0.069 0.0417*
22:5n3 Docosapentaenoic
acid (Clupanodonic acid)
-0.013 0.701 0.227 <0.0001*
22:6n3 Docosahexaenoic acid 0.190 <0.0001* 0.336 <0.0001*
t= trace amount
*Significant at the level p<0.05
4.4.2 Correlation between fatty acids and age
The relationship between aging and sixty-seven individual FA is described in Table 4.2. In
terms of concentration, forty-two FA reached significance with respect to age, with only EPA
and DHA exhibiting positive correlations. In terms of relative percentage, twenty-two FA
reached significance with respect to age, with eleven FA exhibiting positive correlations. Fifteen
FA reached statistical significance in both absolute and relative measures. Only two SFA,
pentadecanoic acid (C15:0) and margaric acid (C17:0), were statistically significant in both
measures but, exhibited the opposite direction correlations to age when compared between
measures. Sixteen total FA did not reach statistical significance when reported as either measure,
namely: lauric acid (C12:0), thalictric acid (C18:1t5), mangiferic acid (C18:2c9c15), stearidonic
49
acid (C18:4n3), and a majority of the conjugated linoleic acids. Trace (t) values were either zero
or levels below the limit of detection (LOD) (0.05 µg/ml) of the Agilent 7890-A GC-FID as
determined by our internal assessments.
50
Table 4.3 Correlation and multiple regression of serum fatty acids and age when stratified by decade
Absolute Concentration 40-49
(n=197) 50-59
(n=246) 60-69
(n=237) 70+
(n=191) Fatty Acid
Correlation (R)
p-value
Correlation (R)
p-value
Correlation (R)
p-value Correlation (R)
p-value
16:0 0.076 0.542 -0.112 0.109 -0.169 0.017* -0.169 0.045*
18:0 -0.003 0.666 -0.09 0.115 -0.152 0.038* -0.164 0.043*
16:1c9 0.160 0.107 -0.085 0.306 -0.074 0.214 -0.136 0.095
18:1c9 0.016 0.891 -0.175 0.012* -0.191 0.009* -0.145 0.125
18:2n6 -0.001 0.929 -0.028 0.735 -0.172 0.019* -0.094 0.258
20:4n6 -0.001 0.758 -0.105 0.095 -0.148 0.034* -0.030 0.899
18:3n3 0.010 0.970 -0.068 0.439 -0.031 0.418 -0.086 0.333
20:5n3 0.062 0.301 0.069 0.430 0.101 0.077 -0.062 0.113
22:5n3 0.023 0.831 -0.038 0.719 -0.017 0.811 -0.124 0.016*
22:6n3 0.051 0.416 0.053 0.802 0.109 0.084 -0.058 0.112
Percent Composition
16:0 0.143 0.136 -0.097 0.222 -0.04 0.483 -0.093 0.315
18:0 -0.155 0.032* 0.043 0.986 0.078 0.282 0.028 0.907
16:1c9 0.200 0.031* -0.059 0.698 0.006 0.889 -0.116 0.224
18:1c9 -0.020 0.563 -0.157 0.038* -0.109 0.132 -0.073 0.816
18:2n6 -0.072 0.602 0.137 0.056 -0.001 0.884 0.073 0.388
20:4n6 -0.069 0.348 0.019 0.866 0.026 0.865 0.127 0.096
18:3n3 -0.018 0.975 0.014 0.612 0.077 0.545 0.019 0.761
20:5n3 0.018 0.579 0.115 0.181 0.180 0.004* -0.018 0.313
22:5n3 -0.032 0.885 0.100 0.106 0.214 0.004* 0.024 0.389
22:6n3 0.004 0.685 0.149 0.137 0.254 0.000* 0.052 0.801
Multiple regression model included the covariates: BMI, smoking, blood pressure medication, depression, anxiety, asthma, and osteoarthritis. *Significant at the level p<0.05, values indicate significance to multiple regression model
4.4.3 Absolute concentration when stratified by decade
51
Of the 67 FA analyzed, due to their implications to health and prominence in the diet, ten
were chosen to undergo further analysis stratified by decade of age starting at age 40 y, as
described in Table 4.3. Stratification by decade was conducted to assess whether distinct
“breakpoints” in FA exist as one ages.
The SFA, PA and SA, were the only FA to exhibit negative correlations between age and
FA status the latter two decades, 60-69 y and 70+ y, and reach significance. The two MUFA,
oleic acid (18:1c9, OA) and palmitoleic (16:1c9, PAO) were negatively correlated in the decades
starting from age 50 y, but only OA reached statistical significance in the decades 50-59 y and
60-69 y (p=.012; p=.009). Both omega-6 PUFA, LA and ARA, were negatively correlated in
each decade starting at age 40 y and only was found to have a significant p-value in the decade,
60-69 y. The omega-3 PUFA, ALA and DPA, were negatively correlated with age but, only
DPA at decade 70+ exhibited a significant p-value (p=.016). In the decades from 40-49 y to 60-
69 y, EPA and DHA exhibited a position correlation to age (R=0.062-.101; R=0.051-0.109)
before becoming negatively correlated at age 70+ y. In summary, EPA and DHA were the only
two FA to be positively correlated with age while all other FA exhibited negative correlations
between age and FA concentration.
4.4.4 Percent composition when stratified by decade
The two SFA, PA and SA, exhibited opposite trends. PA was initially negatively
correlated with age in the 40-49 y decade but, was then seen to be positively correlated in all
subsequent decades. In contrast, SA was negatively correlated to age at decade 40-49 y before
becoming positively correlated with age in all subsequent decades. SA was found to have a
significant p-value in the 40-49 y decade, (R=-0.155; p=.032). The MUFA, PAO exhibited
52
inconsistent direction of correlation, exhibiting a positive correlation at decade 40-49 y (R=0.2;
p=.031) before exhibiting a negative correlation at decade 70+ y (R=-0.116; p=.224). In contrast,
OA, exhibited consistent negative correlations between FA status and age. Only the decade 50-
59 y for OA exhibited a statistically significant p-value (R=-0.157; p=.038). Trends in the
omega-6 PUFA, LA and ARA varied with age. LA alternated between negative and positive
correlations between decades, while ARA was positively correlated with age starting in the
decade starting at age 50 y. The omega-3 PUFA, ALA was positively but, weakly correlated by
decade (R=-0.018-0.077). The omega-3 PUFA, EPA, DPA, and DHA, all exhibited increasingly
positive correlations from decade 50-59 y to 60-69 y and the strength of the correlation
weakened at decade 70+ y. EPA, DPA, and DHA were the only FA to exhibit a statistically
significant p-value between age and FA status at decade 60-69 y or 70+ y.
4.4.5 Comparison between absolute concentration and percent composition
Only the SFA, PA, exhibited the same direction of correlation when expressed as
concentration and percent composition. Whereas OA from the decade 50-59 y onwards exhibited
the conflicting directions of correlation: a negative correlation as an absolute concentration and
positive correlation in terms of percent composition. Both MUFA exhibited nearly identical
directions of correlation between each measure. Statistically significant p-values were seen more
often when evaluated as an absolute concentration. Both omega-6 PUFA, LA and ARA, from
decade 50-59 y onwards exhibited opposite directions of correlations by decade when examined
in both measures. Among the four omega-3 PUFA, ALA and DPA exhibited opposite directions
of correlations between measures at each decade, whereas EPA and DHA, exhibited nearly
identical directions of correlations in both measures. Similar to the other FA at decade 50-59 y, a
53
distinct negative shift was seen for both EPA and DHA in both measures at decade 70+ y.
Overall, directions of correlation between age and FA status were more consistent when
measured as an absolute concentration compared to percent composition.
Additional sensitivity analysis was also conducted but, not reported. Removal of the top
and bottom 5th percentile of participants based on mean FA status yielded statistically similar
results. In a separate analysis when individuals with a BMI of >25kg/mg2 were excluded,
statistically similar results were seen compared to the use of the entire cohort. Thus, individuals
with a BMI of >25kg/m2 were included in our analysis.
4.5 Discussion
Among the sixty-seven FA analyzed in this study, forty-two FA reached statistical
significance when assessed in terms of absolute concentration, and twenty-two FA when
assessed as percent composition. Fifteen FA exhibited statistically significant results in both
absolute concentration and percent composition (Table 4.2). Stratification by decade revealed
that negative correlations to age often started by the 50-59 y decade. This negative correlation
persists once entering the 60-69 y decade, where R becomes larger and the greatest number of
significant values of any decade was seen. This is not surprising as elderly adults typically eat
less and often see a shift in their dietary choices compared to younger individuals (151,152).
Mean FA concentrations across each decade can be seen in Table 8.3 of Appendix B and support
this notion of declining FA status. EPA and DHA were the only FA that exhibited a positive
correlation to aging in both measures unlike all other FA which exhibited negative correlations at
the 50-59 y decade. A number of FA exhibited conflicting correlations between measures and
were consistently shown to be negatively correlated to age when expressed in absolute
54
concentration but, positively as percent composition. When FA status was assessed across
decades, values expressed as percent composition exhibited far more variable direction of
correlation compared to that of absolute concentrations (Table 4.3). Overall, the strength of the
correlations seen across this study were relatively low but, still provide relevant information
regarding FA status with aging.
4.5.1 Saturated Fatty Acids
SFA as a whole exhibited a large number of negative correlations with age in both
absolute and relative measures. Our results indicate a shift in dietary and/or metabolic factors
begins to occur at ~50 y in Australian males which SFA status was impacted by, as seen in Table
8.3. Although we do not have dietary data to report, a study by Waern et al. (151) found 2/3 of
elderly Australian men (75+ y) examined consumed a higher percentage of SFA than
recommended. Compared to the rest of the world, few countries consume more SFA on average
than Australian adults (10). Palm oil, which contains high-levels of PA, is the second most
consumed oil in Australia behind only canola oil, which contains high-levels of OA (153).
Decreases in LDL and total cholesterol were seen in the older decades which is commonly found
with decreased SFA intake (Table 4.1) (154). As the elderly typically consume less calories on
average (124), the actual quantity of SFA consumed, even if the relative percentage is high,
could still likely be lower than that of a middle-aged male. Unfortunately, based on the current
results we are unable to distinguish the specific contributions of either diet or metabolic factors
on SFA levels.
When correlations of SFA to age as an absolute concentration and percent composition
were compared (Table 4.2), a number of conflicting results were seen. When SA was further
55
examined by decade, a negative correlation to age from 40-49 y was seen followed by consistent
positive correlations from 50-59 y to 70+ y when measured as percent composition. However,
when measured as an absolute concentration, PA and SA were the only FA to maintain negative
correlations to age from decade 60-69 y onwards, matching the results seen in Table 4.2. Due to
the prevalence of SFA in the Australian diet, relative percentage of PA and SA compared to the
rest of the diet could see modest increases with a decline in overall calorie/FA consumption as
one ages (124). Thus, the use of absolute concentration or percent composition plays a
significant role in the biological interpretation of dietary FA status, as trends of SFA relative to
the rest of the FA pool are not indicative of actual concentration. Based on our results we can
conclude serum SFA modestly correlates to aging when assessed as an absolute concentration.
4.5.2 Monounsaturated Fatty Acids
Of the large number of MUFA examined in this study, almost all MUFA exhibited
negative correlation with age (Table 4.2). cis-Vaccenic acid (18:1 c11) is a minor FA and was
the only MUFA to exhibit a positive correlation with age (R=0.078; p=.02) in either measure.
trans-Vaccenic acid is commonly found in dairy products which is often consumed by elderly
individuals, however, cis-vaccenic acid is not found in the same sources (155). Mangos are the
most common food source containing cis-vaccenic acid with small amounts also being detected
in vegetable seed oils (156). PAO can be converted to cis-vaccenic acid but, without dietary
records or markers of endogenous lipid production we are unable to say for certain the cause of
the positive correlation seen (157).
Similar to the SFA previously discussed, a distinct negative shift occurs beginning at the
50-59 y decade, where OA exhibited negative correlations with age in both measures. These
56
results once again indicate a shift in dietary and/or metabolic factors may begin to occur at ~50 y
in select SFA and MUFA. The similarity between SFA and MUFA could be attributable in part
through metabolism, as MUFA can be synthesized via substrates derived from the interaction of
the Δ9-desaturase enzyme and SFA (3). Previous experimental models (158) and surrogate
measures of Δ9-desaturase in humans (126) have shown significant age-related decreases in
enzymatic activity, reducing the capability of SFA to be converted to MUFA and subsequent FA
levels. Due to the close biological relationship of the four SFA and MUFA evaluated, seeing
negative correlations in FA status between both FA classes during aging is not surprising. Unlike
SFA, a majority of the MUFA exhibited similar directions of correlation in both measures of
absolute concentration and percent composition. This may be due to the overall smaller
concentrations and percent composition of circulating MUFA, resulting in a smaller dynamic
range of values that are less impacted by the relative abundance of other fatty acids. Based on the
results presented we conclude a consistent and modest negative correlation of MUFA status to
age is seen in Australian males.
4.5.3 Omega-6 Polyunsaturated Fatty Acids
A majority of the omega-6 PUFA assessed in our study exhibited a negative correlation
to aging in at least one measure and displayed the strongest correlations to aging (Table 4.2).
However, only docosadienoic acid (22:2n6) and eicosadienoic acid (20:2n6) were positively
correlated with age. Eicosadienoic acid has been found to increase with consumption of LA but,
LA exhibited a negative correlation to age (159). Due to the relatively small values of these two
omega-6 PUFA found in the diet, the positive correlations could be due to the decrease of other
FA in the pool, resulting in relative increases of the percent composition of other FA.
57
At the 70+ y decade, correlations between LA and ARA status and age in both measures
were slightly higher compared to the preceding decades. When we compared the mean
concentrations of ARA in the 40-49 y decade to the 70+ y decade, the values were found to be
significantly lower in the 70+ y decade (calculation not shown). A decrease in ARA serum status
with aging is similar to results seen in past studies (152,160). Saga et al. (161) also found in
Scandanavian adults those who had higher omega-3 PUFA status exhibited a decrease in ARA
status, which matches the correlations with age and results of Table 8.3 seen between the two
classes in our study (see Table 4.3). A recent shift in the North American-style diet which has
been largely adopted by Australians is associated with a significant influx of omega-6 PUFA
consumption and a decrease in omega-3 PUFA consumption yet, the opposite is seen with aging
in this cohort (162). A survey conducted on older Australians found as participants aged, they
consumed fewer overall calories which typically came from reducing meat and dairy
consumption (124). Common food sources of omega-6 PUFA, primarily ARA, almost
exclusively derive from animal products such as meat, dairy, and eggs (163). These changes in
dietary habits could likely be the cause of the consistent negative correlation with age seen in the
absolute concentration of both LA and ARA.
Omega-6 PUFA are often described as pro-inflammatory and detrimental to human
health but, LA and ARA can both exert their own additional independent properties, suggesting
this could be an overly simplistic view (162). ARA is second only to DHA in terms of
concentration in the brain and increased dietary consumption has found to increase risk of
neurogenerative disease and cognitive dysfunction in elderly participants (164). In older adults,
dietary ARA consumption has actually been found to reduce the risk of hip fractures in men
58
(165). However, the minor benefits of ARA are out-classed by the capabilities of its omega-3
PUFA counterparts. Thus, the negative correlations seen in concentrations of both LA and ARA
with aging are similar to previous studies and is indicative of modifications to the overall lipid
profile which supports the reduction of specific health-risks commonly seen in the aging
population.
When compared to all other FA classes the omega-6 PUFA exhibited the largest number
of discrepancies between direction of correlation when compared between absolute
concentrations and percent composition These contradictions highlight the need to better
understand whether absolute concentration or percent values are more relevant for assessing the
health of older adults. When the mean concentrations of ARA at the 40-49 y and 70+ y decade
were compared, concentrations were lower at 70+ y (Table 8.3). This supports the negative
correlation of omega-6 PUFA to age seen when measured as an absolute concentration. The
positive correlation of ARA when assessed as percent composition could be due the decrease of
the relative mean FA pool as seen in Table 4.1. Although concentrations of ARA were shown to
decrease, the relative decrease of the entirety of the FA pool was so large it produced a false
positive when expressed as percent composition, which other studies have seen previously with
omega-6 PUFA (166). As our results conclude omega-6 PUFA status decline with age, this
provides further reasoning for the use of absolute concentrations over percent composition for
the accurate interpretation of FA status related to health status.
4.5.4 Omega-3 Polyunsaturated Fatty Acids
The long-chain omega-3 PUFA EPA and DHA were the only two FA to exhibit positive
correlations with age within both measures, (Table 4.2). When further examined by decade,
59
correlations of EPA and DHA became increasingly positive from decade 40-49 to 60-69 (Table
4.3). Our study is one of the largest to-date to examining the relationship of FA status and aging,
and one of few to further examine the correlations seen by individual decade in this level of
detail. To our knowledge, no other study has examined the relationship of plasma/serum omega-
3 PUFA concentrations and aging, without a supplement, in in a cohort of this size. Correlations
of omega-3 PUFA status to age seen in our study support the results of previous studies which
have seen a similar age-related increases in blood levels of EPA and DHA across a number of
populations globally (122,127–129,161,167). It is speculated that this data reflects an adaptive
mechanism to utilize blood and/or adipose stores of omega-3 PUFA that could be occurring with
advanced aging. Omega-3 PUFA are mobilized from adipose/blood stores of the mother during
in-utero development of the fetus, and to breast milk postnatally, to support the growth of the
child’s brain (168). In the latter years of life, a similar phenomenon could be occurring where
older adults are mobilizing EPA and DHA from adipose/blood to the brain to support cognition.
However, at some point in the 70+ y decade the mechanism of recruiting omega-3 PUFA from
adipose stores could be starting to fail. We saw a negative correlation of EPA and DHA to age in
the 70+ y decade in Table 4.3 but, mean FA concentrations were greater than the 60-69 y decade
in Table 8.3 of Appendix B. Thus, at some point during the 70+ y decade EPA and DHA
concentration must be beginning to decline to account for the negative correlation to age seen
despite having a higher mean concentration than the previous decade. Based on the age
stratification used in this study we are unfortunately unable to say with confidence what age this
change in omega-3 serum status may be occurring. Alternatively, even if dietary status remained
the same the decline may be due to deficient absorption of essential nutrients (125). Future
60
studies could examine this phenomenon by giving older adults omega-3 PUFA supplements and
monitoring multiple stores of FA to see if any adaptive behaviours occur/change over time.
Consuming fish is the most effective means of improving omega-3 PUFA levels via
regular dietary intake with foods such as eggs, flax, selfish, and some vegetable oils also
providing small quantities (169). Without dietary data we cannot specify whether dietary
changes via an increase in fish occurred with age resulting in the increase seen in this study.
Prior research which describes trends in the dietary habits of older Australians found participants
viewed fish as a healthier alternative to red meat and that it was important to include more fish
into the diet (124). Studies examining the dietary intakes of older Australians however were
mixed, with Flood et al. (170) finding a significant increase in fish consumption from 62-72 y
but, Hill et al. (171) found decreased quantities, and Meyer (172) reported no change in fish
consumption with age. A common dietary supplement that is also used to improve omega-3
PUFA status are fish-oil pills, which are a simple and cost-effective method of obtaining dietary
EPA and DHA compared to consuming fish (173). A majority of older Australians not
consuming any type of fish-oil supplement, even if consuming more than 2 servings of fish per
week, were still found unable to meet the dietary recommendations put forth by the National
Heart Foundation of Australia (174). A similar study found only 20% of all Australians, not only
the elderly, were able to reach the recommended omega-3 PUFA consumption set forth by
Australian Ministry of Health (172). Some studies have reported values as high as 32.6% of
Australians 45+ y consumed an omega-3 supplement in the past month (175). However, the
Adams et al. (175) study did not take into account the frequency of consumption and this value
could be inflated. The trend of increased omega-3 PUFA levels in older vs. younger adults seems
61
paradoxical however, as increased omega-3 PUFA status has found to decrease the rates of CVD
and cognitive impairments of which elderly individuals are at an increased risk for (148,176).
Measures of FA concentrations from blood plasma/serum reflect not only dietary status but, also
serve as an aggregate of genetic and metabolic factors. Independent of fish consumption, these
genetic and metabolic factors can specifically modify omega-3 status. Elderly individuals when
compared to younger individuals, are found to have a greater metabolic capacity for
incorporating dietary EPA and DHA into blood plasma, which aligns with the trends seen in
Table 4.2 (129,177). Although positive correlations with omega-3 PUFA are seen as one ages,
the strength of this association may not be enough to equate to cardioprotective status. The O3I
is a widely utilized tool to assess overall omega-3 PUFA status but, only values of >8% correlate
to reductions CVD and CHD (58). The quantity of omega-3 PUFA needed to be consumed to
reach this threshold can drastically vary between individuals. Genetic carriers of the apoE e4
allele have shown to exhibit a reduced efficacy of fish oil supplementation for increasing plasma
levels of EPA and DHA compared to those who are not a carrier (150). Thus, modified dietary
behaviours or, genetic differences independent of dietary consumption, are both implicated in the
positive correlations between omega-3 PUFA status and age observed but, we are unable to
determine the extent which either factor influenced the trends seen.
When stratified by decade, both ALA and DPA status exhibited opposite directions of
correlation to aging in all four decades when compared between both measures in Table 4.3. For
DPA, the results seen in Table 4.2 are primarily driven by only a single decade in each measure,
as the remaining correlations to age are weaker in comparison. This implies DPA exhibits
relatively consistent correlations to age until decade 70+ y, when correlations of EPA and DHA
62
to age also become negative. As DPA is a direct intermediate in the conversion pathway between
EPA and DHA and found in common food sources, seeing similar results at 70+ y is not
surprising (178). The discrepancies between measures are likely due to the same phenomenon
seen in the omega-6 PUFA which resulted in a false positive when assessed as percent
composition. Based on the number of discrepancies seen between measures in all FA classes, the
use of only one measure can drastically impact the interpretations of results and must be
carefully considered when designing future studies examining the relationships of FA status.
4.5.5 Statistical and Biological Relevance
As a whole, the strength of the correlations between FA status and aging in this study
were relatively low. In terms of absolute concentration, LA had the strongest correlation
coefficient of -0.241 (Table 4.2). When expressed as a coefficient of determination (R2), this
predicts only ~5% of the variance in FA status is attributable to age in our current model.
Whether a 5% variance in FA status is biologically relevant is difficult to determine without
standardized FA reference ranges, which have yet to be established. However, unlike being
prescribed a drug, dietary status is chronically changing and fluctuating. Small dietary
adaptations can accumulate to become greater health benefits over long periods of time. For
example, SFA concentrations decreased across each decade as shown in Table 8.3 of Appendix
B. Decreased SFA levels are consistent with the observed decrease in LDL and total cholesterol
seen in the older decades in this study (Table 4.1) which benefits cardiovascular health (154).
Although modest correlations are seen in Table 4.3, peaking at R= -0.191, small changes over
time could amount to a biologically relevant result. Thus, despite having relatively weak
correlations, we cannot definitively conclude that aging has a negligible biological impact on
63
individual serum FA status. More research is needed in diverse cohorts to determine if a similar
strength of relationship is seen and the role dietary status and endogenous lipid synthesis play.
4.5.6 Absolute vs. Relative Measures
A unique aspect of our study is that we have reported trends of both absolute
concentrations and percent composition for assessing serum FA status. Currently, there is no
established protocol for the consistent reporting of FA data, which makes the interpretations of
the results often difficult to compare between studies. Brenna et al. (12) recently published
twenty-two recommendations for establishing best practices in FA research, specifically stating
sufficient information should be included in reports to convert absolute values to relative values,
and vice-versa. Of the sixty-seven FA assessed in our study, a large majority saw consistent
results between absolute concentrations and percent composition. However, sixteen FA in Table
4.2 exhibited opposing directions of correlations between absolute and relative measures, most
notably: ARA, ALA, and DPA. All sixteen FA exhibited a negative correlation when measured
as an absolute concentration but, a positive correlation was seen when assessed in terms of
percent composition, with respect to age. Our results are similar to that of previous studies which
have also evaluated absolute vs. relative measures (95,166,179,180). When the health outcomes
of Australian women correlated to plasma FA status was examined, absolute and relative
measures were generally similar when predicting all-cause mortality but, did note omega-6
PUFA exhibited the largest disagreement between measures (166). Miura et al. (166)
recommended the use of absolute concentrations to be utilized in future studies examining FA
status and health outcome measures to improve generalizability between studies, as relative
measures are able to be derived from absolute values. A majority of the studies evaluating
64
absolute vs. relative measures echo a similar sentiment advocating the use of concentration
values. This is exemplified by the stronger correlation of absolute concentrations of FA as
compared to percent composition values in a study by Sergeant et al. (95) examining circulating
FA and lipid biomarkers. This difference was also readily apparent in the present study. The
limitation of percent composition is apparent given the limited dynamic range as compared to
concentration values that can differ by several orders of magnitude. Also, percent composition
can be influenced by independent fluctuations of other FA. In our study, as the outcome being
assessed was individual FA status with respect to aging, in terms of the biological question being
asked, we believe FA concentrations are the more accurate measure for basing our interpretations
of healthy individuals compared to percent composition. Percent composition still remains a
useful tool for assessing FA status in certain instances, such as individuals with lipemia who
have high concentrations of all FA where assessing relative trends may be more appropriate (12).
Inclusion of both measures may help to provide a more robust understanding of the relationship
between FA status and different disease biomarkers. We believe all future FA research should
report individual concentrations for the entire pool of FA analyzed to maintain transparency and
improve generalizability which still allows percent composition to be calculated.
4.5.7 Limitations
We performed secondary-analysis on our blood plasma samples and data which was
obtained from the MAILES cohort study (134). We were unable to obtain data regarding dietary
intake which would have provided a more informed view of FA status and dietary trends during
aging. Secondly, as this was a cross-sectional design, we could not assess the effect of aging
within each participant. Finally, our study did not include female participants as the original
65
investigators excluded female participants from their original study. However, studies have
found differences in FA metabolism between men and women, with women typically showing
greater blood FA levels (117). Thus, the age-related effects seen may differ if females are
included in future studies.
4.6 Conclusion
In conclusion, aging has a modest effect on individual serum FA values. Although modest,
the biological implications of aging on individual FA status cannot be overlooked. Of the sixty-
seven FA included in our analysis, only EPA and DHA were found to exhibit positive
correlations with age in this cohort of Australian men. Ten FA were further stratified by decade,
where FA status began to exhibit negative correlations with age in the 50-59 decade that became
stronger at age 60+ y. When FA status was assessed as an absolute concentration, the direction of
the correlations was more consistent and a greater number of significant values were found,
compared to percent composition. Due to the influence of individual FA on the whole pool of FA
assessed when reported as percent composition, it is recommended that absolute concentrations
be the primary reporting method to improve comparability of results.
66
5 FATTY ACID REFERENCE RANGES AND CORRELATIONS TO LIPID BIOMARKERS OF HEALTHY SINGAPOREAN MEN AND WOMEN
5.1 Abstract
Dietary FA are essential for overall human health and well-being yet, individual FA
reference ranges have yet to be established for use in a scientific or clinical setting. Without risk
classifications similar to other lipid biomarkers, reported FA concentrations lack context
regarding the direction and magnitude of change necessary to achieve optimal health in
individuals with deficient, or, excess amounts of FA. Reference ranges of sixty-seven individual
FA (μmol/L) were profiled and analyzed using GC-FID from serum samples collected from 476
middle-aged Singaporean males (BMI: 23.33±2.92) and females (BMI: 21.80±3.56). Measures
of TG, HDL, LDL, and TC (mmol/L) were also collected. The mean FA concentration seen in
this cohort (11458.09±2477.50) was similar to that of overweight North American cohorts
assessed in past studies, and far greater than that of the limited healthy cohorts assessed. A
Student t-test further compared ten biologically relevant FA between sexes, with females
exhibiting higher concentrations in four FA (p<.05; SA, PAO, ALA, DHA). A linear regression
model including age, sex, and BMI as covariates found significant correlations between all lipid
biomarkers and FA concentrations (p<0.05). A majority of the participants who recorded FA
concentrations in the >95th percentile also exhibited TG, HDL, LDL, and TC levels in the “high”
risk classification of developing CVD. Future work examining the relationship between
67
individual FA and lipid biomarkers in multiple global cohorts will help to establish an accurate
measure for cut-off values of individual FA concentrations highly correlated to disease-risk
5.2 Introduction
Adequate consumption of dietary FA is essential for the development and maintenance of
overall human health and well-being (1). Due to their wide array of structural and functional
capabilities, FA status has been implicated in reducing the risk of a number of chronic diseases
(2,4). There is currently a gap in the field of dietary lipid research as individual FA reference
ranges have yet to be established. This fundamental gap hinders the capability of translating
reported values from scientific reports for clinical use. Objective measures with established
clinical ranges for other lipids such as cholesterol, and triglycerides are commonly used in
clinical practices (11). Developing baseline FA values would provide medical practitioners
context regarding the direction and magnitude of change necessary to achieve optimal health in
patients with deficient, suboptimal, or even excess amounts of FA. Additionally, FA values
which are reported in scientific literature are often inconsistent, using either different units or
methods of reporting altogether (12). In order to address this gap, comprehensive profiles of
baseline FA concentrations must be collected and analyzed. Developing validated FA reference
ranges while using consistent methods of assessment and reporting are necessary steps in
creating generalizable results between individual FA status of different cohorts.
To date, only a small number studies have comprehensively profiled plasma or serum FA
levels in healthy adults with the primary objective of developing reference ranges
(94,95,99,100). No studies to our knowledge have comprehensively examined FA status of
individuals in the Singapore-Southeast Asia region of the world. Due to its unique geographical
68
location as an island country located between Malaysia and Indonesia, Singapore has a unique
population consisting of three predominant ethnic groups: Chinese, Malays, and Indian (181).
Independent of ethnicity, rates of obesity in Singapore are steadily increasing as their typical
diets are very energy-dense (182). A rise in western-style fast food and high rates of
Singaporeans typically going out to eat, has been positively correlated with these dietary changes
(182,183). Additionally, the rise in large-scale production of palm oil, high in the SFA PA and
the MUFA OA, from Indonesia and Malaysia have contributed to high levels of consumption in
the region (184). A number of studies have found significant correlations between FA status and
the risk of chronic disease specifically in Singapore Chinese adults (185–194). On a global scale,
countries in Southeast Asia have reported far higher omega-3 PUFA and SFA dietary
consumption than North American countries (10). This study sought to assess and determine the
distribution of sixty-seven individual FA concentrations in healthy, adult Singaporean men and
women. Correlations between FA status and lipid biomarkers associated with overall health
status and differences between males and females were examined. Given that certain disease-
states are known to alter FA status independent of dietary intake (9), using only healthy
participants will reduce potential confounding between FA modifications associated with a
disease-state or natural fluctuation due to daily living.
5.3 Materials and methods
5.3.1 Participants
Participants were originally recruited as described in Bi et al. (135). In brief, 476 healthy
male and female adult Singaporeans were recruited from the general public through
advertisements placed around the National University of Singapore and the Clinical Research
69
Centre website from June 2014 to January 2017. All participants lived in Singapore for at least
the past five years and were excluded if they had been diagnosed with any major disease or were
pregnant. 91% of the participants were ethnically Singaporean Chinese. All participants were
instructed to restrict caffeine and alcohol for 24 hours prior and to fast prior to data collection.
All procedures involving human participants were approved by the National Healthcare Group
Domain Specific Review Board (NHG DSRB, Reference Number: 2013/00783), Singapore.
5.3.2 Anthropometric Measures
Anthropometric measurements were obtained in a fasted state. Weight (kg) in light
clothing without footwear was measured to the nearest 0.1 kg by using an electrical scale and
height (cm) was measured using a stadiometer to the nearest millimeter (Seca 763 digital scale,
Birmingham, UK). BMI (kg/m2) was calculated using weight divided by the height squared.
Waist circumference (WC) was taken at the smallest WC above the umbilicus or navel and
below the xiphoid process. DEXA (QDR 4500A, fan-beam densitometer, Hologic, Waltham,
MA, USA) was used for the measurement of fat mass (%). To ensure accuracy of the
measurement, all metal items were removed from the participants.
5.3.3 Blood Samples
Blood samples were collected as previously described in Bi et al. (195). In brief, 10 mL
of venous blood was collected into Vacutainers (Becton Dickinson Diagnostics, Franklin Lakes,
NJ USA). Blood samples were separated by centrifugation at 1500 xg for 10 min at 4°C within 2
h of being drawn and aliquots were stored at - 80°C until analysis. Fasting lipid parameters
including total cholesterol, HDL, and LDL were measured using the COBAS c311
(Roche)chemistry analyzer. Measurements were done in duplicate and readings were averaged.
70
5.3.4 Gas Chromatography Protocol
A modified Folch method was used for lipid extraction for serum samples which has been
described in-detail previously (77). In brief, samples were first thawed on ice, where 50 µl of
plasma or serum was pipetted into a 15ml tube. Stock solution of 2:1 chloroform:methanol (v:v)
with a 3.33 µg/mL of internal standard (C19:0) was prepared in which 3 mL of CHCl3: MeOH
was added to the tube and capped tightly with a phenolic PTFE liner. After a one minute of
vortex, 550 µL of 0.1 M KCl was added to each tube and briefly vortexed again for another 5-10
seconds. Next, samples were spun at 1460 rpm for 10 min (at 21°C) to separate phases. The
lower chloroform layer was transferred into a clean 15 ml glass tube and dried down under a
gentle stream of nitrogen. After, 300 µL of hexane and 1 ml 14% BF3-MeOH was added and
followed with a vortex of 5-10 seconds. Methylation was then completed at 100oC for 60
minutes in oven, where it was then cooled for 10 min at room temperature. Next, 1 mL of
double-distilled water and 1 mL hexane was added to stop the methylation process. After a one-
minute vortex, the sample was once again spun down at 1460 rpm for 10 min (at 21°C) to
separate phases. The top hexane layer was then extracted into a clean 2 ml GC vial, dried down
under a gentle stream of N2 and then reconstituted in 400 µL of hexane.
Total FA were quantified on an Agilent 7890-A GC-FID and separated on a SP-2560
fused silica capillary column (100m x 0.25mm x 0.2um) (Sigma, Cat # 24056). The splitless inlet
(pressure of 19.5 psi and a hydrogen flow of 13.7 mL/min) and detector was set at 250°C
(hydrogen flow of 30 mL/min, air flow of 450mL/min and nitrogen flow of 10mL/min), where
1uL of sample was injected onto a splitless inlet set. The oven was initially set at 60°C, where it
was increased to 13°C/min to 170°C and was held for 4 min, then followed 6.5°C/min to 175°C
71
with no hold to 2.6°C/min to 185°C with no hold, to 1.3°C/min to 190°C with no hold, to
13.0°C/min to 240°C and was held for 13 min. Total run time was 49.78 min per sample.
5.3.5 Fatty Acid Analysis
Peaks were identified and chromatographs were analyzed via OpenLab CDS EZChrom
Edition 3.2.1 software. Sixty-seven total FA were included in the pool analyzed and a C19:0
internal standard was used to calculate FA concentrations (μmol/L) and percent composition (%).
Trace (t) values were either zero or levels below the detectable limit (0.05 µg/ml) of the Agilent
7890-A GC-FID.
5.3.6 Statistical Analysis
Statistical analysis was conducted using SAS University Edition (SAS Institute Inc.;
Cary, North Carolina). The range, mean, and percentiles of the concentration of all sixty-seven
FA and TG, LDL, HDL, and total cholesterol were profiled. Of the sixty-seven FA analyzed, ten
were chosen for further analysis due to their implications to health and prominence in the diet. A
Student t-test was used to determine statistical differences between male and female participants
with results expressed as mean ± SD. A linear regression model adjusted by age, gender, and
BMI was conducted to assess the relationship between FA concentrations and TG, LDL, HDL,
and total cholesterol. Results were reported as coefficient of determination (R2) of the entire and
standardized parameter estimates/beta-weight (β) of age. A p-value of <0.05 was considered
statistically significant.
72
5.4 Results
Table 5.1 General characteristics of study population
Total Population Males Females p-value Population (#) 476 186 290 /
Age (yrs) 38.9±14.62 38.41±14.57 39.21±14.68 0.5601
BMI (kg/m2) 22.4±3.41 23.33±2.92 21.80±3.56 <0.0001* Waist Circumference
(cm) 73.93±9.22 79.28±8.09 70.49±8.21 <0.0001* DEXA Fat (%) 30.97±7.68 24.66±5.66 35.05±5.84 <0.0001*
Glucose (mmol/L) 4.55±0.49 4.63±0.51 4.50±0.46 0.0032* HDL Cholesterol
(mmol/L) 1.68±0.43 1.51±0.35 1.79±0.44 <0.0001* LDL Cholesterol
(mmol/L) 3.35±0.92 3.41±0.93 3.32±0.91 0.2771 Total Cholesterol
(mmol/L) 5.27±1.03 5.21±0.99 5.31±1.05 0.2868 Triglycerides
(mmol/L) 0.97±0.45 1.01±0.44 0.95±0.46 0.1328 Total Serum Fatty
Acids (μmol/L) 11458.09±2477.50 11379.0±2421.9 11508.8±2515.4 0.5775
Data represented as Mean±SD * Considered statistically significant at p<0.05 following Student t-test DEXA; Dual-energy X-ray absorptiometry scan, HDL; High-density lipoprotein, LDL; Low-density lipoprotein
The general characteristics of the entire study population, including differences between
males and females, are presented in Table 5.1. Participants ethnicity was not included in our
calculations as a large majority reported to be Singaporean Chinese (>90%), as previously
described (135). All participants self-reported to be healthy, disease-free and on average
recorded normal BMI values (22.4 kg/m2). Significant differences were noted between sexes in
measures that are commonly seen in other populations such as: BMI, WC, and fat %. Females
had significantly greater levels of HDL cholesterol (1.79 to 1.51 mmol/L) but, both groups
averaged relatively high concentrations compared to normal values. Both sexes exhibited
73
moderately high concentrations of LDL-cholesterol, total cholesterol and fat % but, TG were
well below reported normal levels of 1.7 mmol/L.
Table 5. 1 Concentration (μmol/L) of select fatty acids in males and females
74
Table 5.2 Range, mean and percentiles of Fatty Acid concentrations (μmol/L) of serum total lipids
Range Percentile Fatty Acid
Common Name Min. Mean SD Max. 5 10 25 50 75 90 95
12:0 Lauric acid t 10.86 12.38 159.9 t 3.45 5.22 7.93 13.19 20.94 28.97 14:0 Myristic Acid 25.39 82.75 46.99 342.7 33.60 39.52 48.89 69.79 100.00 146.08 180.10 15:0 Pentadyclic acid t 39.17 28.18 207.7 9.59 11.85 16.72 26.38 61.92 77.39 87.82 16:0 Palmitic acid 1400.4 2667.5 625.1 6161 1809.78 1991.27 2256.22 2562.95 2967.94 3529.16 3908.25 17:0 Margaric acid t 37.79 15.42 161.96 17.59 22.22 28.13 35.62 45.04 56.15 66.34 18:0 Stearic acid 415.16 745.21 155.5 1550.6 522.53 568.83 641.51 720.94 828.47 946.03 1021.83 19:0* Nonadecylic acid
20:0 Arachidic acid 1.77 20.06 5.27 50.27 12.82 13.98 16.60 19.49 22.88 26.38 29.83 21:0 Heneicosylic acid 0.72 8.65 7.22 120.05 3.51 4.30 5.70 7.38 9.82 12.98 16.90 22:0 Behenic acid 11.04 45.17 11.63 91.93 28.58 31.11 37.47 44.48 52.07 59.05 65.52 23:0 Tricosylic acid t 6.96 9.2 36.27 t t t t 15.46 20.69 23.02 24:0 Lignoceric acid 4.99 42.66 11.21 94.13 26.16 29.36 34.87 41.73 50.23 55.98 61.68
12:1c11 t t / t t t t t t t t 14:1c9 Myristoleic acid t 10.38 9.15 46.63 t t 2.76 9.09 17.13 22.33 25.94 15:1c10 t t / t t t t t t t t 16:1c9 Palmitoleic acid 6.52 195.97 84.2 526.43 88.79 105.94 135.08 180.22 238.06 309.04 359.35 17:1c10 t t / t t t t t t t 18:1c9 Oleic acid 263.11 2095.7 633.05 5070.1 1311.13 1431.03 1683.16 1975.62 2389.16 2969.18 3381.20 18:1c11 cis-Vaccenic acid 13.1 163.91 43.21 355.91 106.54 115.73 134.73 159.52 186.73 218.78 240.30 18:1c12 t 4.2 3.34 17.34 t t 2.13 3.86 5.85 8.06 10.77 18:1c13 t t / t t t t t t t t
18:1c14 t t / t t t t t t t t
75
19:1c10 t t / t t t t t t t t
20:1c5 t 6.31 2.94 18.81 2.95 3.34 4.33 5.60 7.44 10.73 12.17 20:1c8 t 3.01 1.61 18.01 1.18 1.49 2.04 2.73 3.73 4.81 5.86 20:1c11 Gondoic acid 2.61 16.23 5.48 52.52 9.77 10.81 12.68 15.44 18.52 23.02 26.45 22:1n9 Erucic acid t 4.23 4.77 74.29 t 1.87 2.40 3.29 5.23 7.86 8.92 24:1n9 Nervonic acid 2.59 63.85 17.67 187.13 41.14 44.12 52.68 61.63 72.69 84.22 92.30 16:1t9 Palmitoleic acid t 7.92 5.67 27.13 2.01 2.22 2.89 6.84 10.92 16.29 19.93 18:1t4 t 0.18 1.35 26.78 t t t t t t 1.05
18:1t5 Thalictric acid t 6.57 8.78 62.18 t t t 3.62 9.07 17.87 25.01 18:1t6-8 Petroselaidic acid t 3.33 2.45 16.54 0.73 0.93 1.54 2.66 4.57 6.45 8.42 18:1t9 Elaidic acid t 6.68 3.33 46.96 2.74 3.52 4.71 6.25 8.13 10.01 11.91 18:1t10 t 4.37 3.13 29.71 1.13 1.75 2.56 3.64 5.36 7.50 9.36
18:1t11 trans-Vaccenic acid t 7.06 5.04 59.38 1.67 2.41 3.88 6.04 8.87 12.06 15.91
18:1t12 t 4.43 3.76 59.85 1.47 1.94 2.66 3.61 5.14 7.84 9.78 18:1t13 t 11 7.04 59.34 4.25 5.13 6.86 9.39 13.16 17.98 21.92 18:1t16 t 3.62 2.4 26 t t 2.43 3.44 4.59 6.36 7.45 18:2tt t 4.06 4.39 25.93 t t 0.91 2.71 5.53 10.40 13.64
18:2t9t12 Linoelaidic acid t 1.17 3.03 25.63 t t t t t 4.56 6.68 18:2c9t13 t 1.05 2.64 17.83 t t t t t 4.91 7.01
18:2ct t 5.17 3.2 24.07 1.97 2.47 3.27 4.44 5.97 8.83 11.72 18:2c9t12 0.98 13.68 4.58 38.48 8.12 9.38 10.82 12.63 15.59 20.02 23.23 18:2t9c12 t 7.85 3.16 21 3.24 4.19 5.77 7.67 9.46 11.93 13.41 18:2c9c14 t 1.28 2.39 18.22 t t t t 2.14 3.90 6.30 18:2c9c15 Mangiferic acid t 2.49 3.58 21.31 t t t t 4.54 7.20 9.22 18:2c9t11
CLA Conjugated linoleic acid t 8.68 2.59 23.37 4.80 6.20 7.44 8.54 9.68 11.48 13.11
76
18:2c11t13 Conjugated linoleic acid t 0.96 2.04 10.73 t t t t t 4.33 5.89
18:2t10c12 CLA
Conjugated linoleic acid t 0.58 1.75 13 t t t t t 2.55 5.09
18:2c/c CLA1
Conjugated linoleic acid t 0.26 1.2 12.7 t t t t t t 1.66
18:2c/c CLA2
Conjugated linoleic acid t 2.13 2.4 16.2 t t t 1.71 2.97 4.68 6.56
18:2tt CLA
Conjugated linoleic acid t 8.88 4.42 38.36 0.55 4.71 6.77 8.59 10.74 13.07 15.23
18:2n6 Linoleic acid 1652.5 3700.2 777.5 7244.2 2559.24 2766.63 3181.02 3640.64 4090.33 4771.69 5081.19
18:3n6 Gamma-linolenic acid 2.94 27.04 19.71 124.53 7.50 8.67 12.68 22.47 35.00 51.62 63.23
20:2n6 Eicosadienoic acid 2.52 23.95 7.82 52.69 12.63 15.80 18.84 22.84 28.41 33.59 38.79
20:3n6 Dihomo-gamma-linolenic acid 2.03 115.34 46.26 273.92 56.59 63.95 80.09 106.35 141.41 181.09 206.44
20:4n6 Arachidonic acid 104.28 685.78 177.42 1538.7 433.66 480.04 572.97 673.92 784.63 908.26 973.53
22:2n6 Docosadienoic acid t 7.88 4.77 33.49 2.49 3.22 4.56 6.48 10.33 14.34 16.74
22:4n6 Ardenic acid 3.6 17.72 5.88 50.39 9.25 11.06 14.14 17.21 20.92 25.16 27.35
22:5n6 Docosapentaenoic
acid (Osbond acid)
3.72 19.67 6.24 44.73 10.99 12.59 15.39 18.99 23.40 27.45 31.08
18:3n3 Alpha-linolenic acid 7.28 57.63 24.65 175.42 28.66 32.95 41.08 52.37 69.74 91.59 103.29
18:4n3 Stearidonic acid t t / t t t t t t t t
20:3n3 Eicosatrienoic acid t 16.58 11.32 47.14 t 2.28 4.54 19.22 26.17 30.87 33.76
20:5n3 Eicosapentaenoic acid 4.22 92.97 80.94 596.25 20.06 29.04 39.96 68.19 115.16 193.42 285.06
22:3n3 Docosatrienoic acid t t / t t t t t t t t
22:5n3 Docosapentaenoic acid 14.7 43.5 15.53 118.69
23.68 26.98 32.49 40.66 50.54 65.28 71.54
77
(Clupanodonic acid)
22:6n3 Docosahexaenoic acid 44.88 255.52 94.44 663.35 121.88 150.71 193.56 240.45 303.95 391.38 430.12
20:3n9 Mead acid 3.03 8.89 4.06 29.36 4.34 4.90 6.21 8.05 10.24 13.92 17.05 *Internal Standard t; Trace amount
78
Table 5.3 Concentration (μmol/L) of select fatty acids in males and females
Table 5.2 details the mean, range, and percentiles of sixty-seven individual FA
concentrations, and the C19:0 internal standard, expressed as μmol/L for the entire study
population of 476 participants. Of this large pool, ten common FA which consisted of more than
90% of the total FA concentration, were selected to be analyzed further due to their health
implications and prominence in the diet. Significant differences in the concentrations of SA,
PAO, ALA, and DHA (p<0.05) were found between males and females. Females demonstrated
higher mean FA concentrations in all four FA which reached significance, with males
demonstrating slightly higher means in only PA, OA, and ARA (Table 5.3).
Fatty Acid
Common Name Males
(n=186)
Females (n=290)
p-value
16:0 Palmitic acid 2683 ± 630.6 2658 ± 622.4 0.66 18:0 Stearic acid 716 ± 145.2 764 ± 159.2 0.001*
16:1 c9 Palmitoleic acid 185 ± 82.5 203 ± 84.7 0.027* 18:1 c9 Oleic acid 2117 ± 645.0 2082 ± 625.9 0.55 18:2 n6 Linoleic acid 3652 ± 737.4 3731 ± 801.9 0.28 20:4 n6 Arachidonic acid 687 ± 176.2 685 ± 178.5 0.86 18:3 n3 Alpha-linolenic acid 53.9 ± 23.0 60.0 ± 25.4 0.0084* 20:5 n3 Eicosapentaenoic acid 84.4 ± 74.2 98.5 ± 84.7 0.06 22:5 n3 Docosapentaenoic acid 42.6 ± 14.8 44.1 ± 0.9 0.3 22:6 n3 Docosahexaenoic acid 240 ± 86.8 266 ± 97.8 0.0033*
Represented as Mean ± SD. *Statistically significant at the level p<0.05
79
Table 5.4 Reference ranges and clinical cut-offs of lipid biomarkers
Range Percentile
Class Min Mean ± SD Max 5 10 25 50 75 90 95
TG 0.36 0.97 ± 0.45 3.06 0.5 0.55 0.66 0.87 1.125 1.64 1.91 HDL 0.8 1.68 ± 0.43 3.44 1.07 1.16 1.38 1.65 1.91 2.29 2.48 LDL 0.99 3.35 ± 0.92 7.64 2.06 2.29 2.74 3.23 3.87 4.48 4.99 TC 2.7 5.27 ± 1.03 9.44 3.74 4.08 4.58 5.17 5.85 6.67 7.05
Class Desirablea Borderline/High Very High
TG 1.7 - 2.2 mmol/L 2.3 - 4.4 mmol/L ≥ 4.5 mmol/L HDL 1.0 - 1.5 mmol/L >1.6 mmol/L / LDL 2.6 - 3.3 mmol/L 3.4 - 4.8 mmol/L ≥ 4.9 mmol/L TC < 5.2 mmol/L 5.2 - 6.1 mmol/L ≥ 6.2 mmol/L
HDL; High-density lipoprotein cholesterol, LDL; Low-density lipoprotein cholesterol, TC; Total cholesterol, TG; Triglycerides aTai et al. (196)
Table 5.5 Linear regression model assessing the correlation between individual fatty acids and lipid biomarkers
Fatty Acid
Common Name Triglycerides HDL LDL Total
Cholesterol
R2 β R2 β R2 β R2 β
16:0 Palmitic acid 0.70 0.78** 0.23 0.02 0.45 0.55** 0.49 0.64** 18:0 Stearic acid 0.50 0.62** 0.25 0.18** 0.49 0.60** 0.57 0.73**
16:1 C9 Palmitoleic acid 0.59 0.70** 0.23 -0.05 0.28 0.31** 0.26 0.38**
18:1 C9 Oleic acid 0.75 0.81** 0.24 -0.1* 0.37 0.46** 0.36 0.50** 18:2N6 Linoleic acid 0.36 0.45** 0.28 0.24** 0.54 0.62** 0.60 0.72**
20:4N6 Arachidonic acid 0.20 0.17** 0.29 0.25** 0.38 0.43** 0.40 0.52**
18:3N3 Alpha-linolenic acid 0.50 0.61** 0.23 -0.06 0.27 0.29** 0.24 0.34**
20:5N3 Eicosapentaenoic acid 0.18 0.00 0.26 0.20** 0.24 0.22** 0.22 0.29**
80
22:5N3 Docosapentaenoic acid 0.27 0.33** 0.26 0.19** 0.33 0.40** 0.35 0.49**
22:6N3 Docosahexaenoic acid 0.24 0.27** 0.27 0.21** 0.36 0.44** 0.37 0.52**
TOTAL 0.56 0.67** 0.24 0.12* 0.49 0.59** 0.55 0.69** *p<0.05 ** p<0.0001 R2 represents the entire model including covariates: BMI, age, and sex
The mean, range, and percentiles of TG, HDL, LDL, and total cholesterol for the entire
study population of 476 participants is profiled in Table 5.4. The cut-off values for lipid
biomarkers used by the Singapore Ministry of Health which denotes a health risk are included in
the bottom of the table (196). A linear regression model was conducted to determine the
coefficient of determination (R2) and standardized parameter estimates (β) which correlated FA
values to cholesterol and TG, as seen in Table 5.5. Correlations between FA concentrations and
lipid biomarkers was significant for almost all cholesterol and TG values. Only EPA did not
exhibit a significant correlation with TG (β =.00). All ten FA correlated significantly with LDL
(β =.22-.62) and total cholesterol (β =.29-.73). A majority of the FA exhibited significant xs=-
0.06-0.02). When the total mean FA concentrations were evaluated, significant correlations were
found within all four measures but, HDL was only significant to the degree of p<.05 compared to
the other measures which were p<.001.
81
Table 5.6 Upper and lower percentile fatty acid concentrations in participants who exceed risk cut-offs for lipid biomarkers
16:0 18:0 16:1c9 18:1c9 18:2n6 20:4n6 18:3n3 20:5n3 22:5n3 22:6n3
Lipid Class
Exceeded Cut-offa
(#) >95% <5% >95% <5% >95% <5% >95% <5% >95% <5% >95% <5% >95% <5% >95% <5% >95% <5% >95% <5%
TG 11 10 0 8 0 7 0 11 0 3 0 0 0 6 0 0 0 1 0 1 0 HDL - Low 11 1 1 0 3 0 0 1 0 0 2 0 2 0 3 1 2 1 2 0 2
HDL - High 262 5 11 9 5 5 13 3 15 14 9 14 11 10 6 16 11 16 8 15 8 LDL 92 13 0 16 0 8 0 11 0 15 0 15 1 8 0 9 0 14 0 15 1
TC 82 13 0 15 0 11 0 13 1 17 0 17 1 11 1 10 0 15 0 17 0
TOTALb 137 20 12 21 8 16 13 19 15 21 11 20 13 19 4 21 13 22 10 22 11 Values represented as number of participants. Each FA has 24 participants in the upper and lower 5th percentile. aTG ≥ 2.3 mmol/L; HDL < 1.0 mmol/L; HDL ≥ 1.6 mmol/L; LDL ≥ 4.1 mmol/L; TC ≥ 6.2 mmol/L bTotal number of individual participants who exceeded cut-off
82
Based on the cut-off values described in Table 5.4, the number of participants who
exhibited levels of TG, HDL, LDL, or TC which exceeded recommended values are tallied in
Table 5.6. Using the reference ranges of this cohort described in Table 5.2, FA concentrations in
the ≥ 95th percentile and ≤ 5th percentile, outside of the normal range, of each at risk participant
was evaluated. Each FA in Table 5.6 had a possible 24 participants in the top and bottom 5% of
their recorded reference range. In participants with high TG values, SFA and MUFA
concentrations in the ≥ 95th percentile was found in a majority of participants, compared to only
a small number of omega-3 and omega-6 PUFA. Participants with low HDL cholesterol values
exhibited the fewest scores in the ≥ 95th percentile and only a modest number of concentrations
in the ≤ 5th percentile of omega-3 and omega-6 PUFA. Comparatively, participants with high
HDL cholesterol saw a large number of omega-3 and omega-6 PUFA concentrations in the ≥ 95th
percentile and simultaneously, the most participants with FA concentrations in the ≤ 5th
percentile. LDL and TC exhibited nearly identical trends, with a large number of participants
with FA concentrations in the ≥ 95th percentile across all ten FA. Omega-6 PUFA and SFA
concentrations in the ≥ 95th percentile was the most commonly seen in LDL and TC, with the
omega-3 PUFA, ALA and DPA, exhibiting the lowest frequency. Overall, FA concentrations in
the ≥95th percentile were far more prevalent in participants exceeding lipid cut-off values than
concentrations in the ≤ 5th percentile.
5.5 Discussion
In this study we determined the percentiles, mean, and ranges of sixty-seven individual FA
collected from 476 healthy, adult Singaporean males and females with ten FA highlighted for
further correlation analysis with TG and cholesterol. Significant correlations between FA
83
concentrations and TG, LDL, HDL, and total cholesterol were seen in almost all of the ten FA
highlighted. Mean total FA concentrations significantly correlated with all four lipid biomarkers
but, exhibited the weakest relationship with HDL cholesterol. Individual FA concentrations
significantly correlate with all four lipid biomarkers and high concentrations of FA were more
often seen in individuals exceeding established cut-off values. Female participants exhibited
significantly higher concentrations SA, PAO, ALA and DHA when compared to males (Table
5.3), but total FA concentrations did not statistically differ between sexes (Table 5.1). To our
knowledge, the inclusion of sixty-seven individual FA concentrations is the largest pool of FA to
be examined to date for the purposes of profiling FA reference ranges in any cohort.
5.5.1 Individual Fatty Acid Reference Ranges
A very small number of studies have, in detail, profiled FA concentrations in healthy
adults for the purposes of developing reference ranges (94,95,99,100). Glew et al. (197) have
also reported mean percent composition of twenty-six FA from serum samples collected in
healthy adult males and females from Northern Nigeria. Of these studies, Bradbury et al. (100)
and Glew et al. (197) reported their values in percent composition, whereas absolute
concentrations were reported in our study. Sergeant et al. (95) (n=93; obese American women)
and Abdelmagid (94) (n=826; healthy, multiethnic Canadians) also reported their results as
absolute concentrations (μmol/L) which allows for the most direct and reliable comparisons to
our results. Percent composition is determined by the percentage of one FA from the total pool of
FA included in the analysis; a smaller pool analyzed results in a larger apparent percentage of
individual FA (12). Large fluctuations in a single FA could significantly modify the total value
of the pool and subsequent percent composition seen in other FA when no changes in
84
concentration actually occurred. Absolute concentrations of an individual FA are independent of
changes in the pool analyzed and have a larger dynamic range of values, as percent is limited to a
0-100 scale. Reporting results as absolute concentrations has been found to reduce the loss of
statistical significance between FA compared to when values were assessed as weight
percentages (198). Furthermore, relative percent can be derived from absolute concentrations
but, the same is not possible in reverse as the values of the internal standard used in the study are
needed to calculate absolute concentration (12). As our study included over sixty individual FA,
whereas Bradbury et al. (100) and Glew et al. (197) only included up to twenty-six, it is difficult
to accurately compare our reported values even in terms of percent composition. Thus, the use of
absolute concentrations is recommended for future studies profiling FA reference ranges to
improve comparability and generalizability of results.
Three of the five studies mentioned previously have examined adult, North American
cohorts and the remaining studies examined cohorts from New Zealand and Nigeria
(94,95,99,100,197). To our understanding, ours is the first study to examine individual FA
concentrations of a large cohort of Southeast Asian, specifically Singaporean, adults in this level
of detail. Previous work done in a multiethnic Singaporean cohort found plasma FA composition
is a strong, reliable biomarker reflecting dietary consumption in epidemiological studies (199).
The mean FA concentration in our study (11,458 μmol/L) more closely resembled that of
Sergeant et al. (95), as Abdelmagid (94) reported a mean concentration which was ~40% lower
(6,948 μmol/L) . Of the ten FA highlighted in Table 5.3; SA, PAO, AA, and ALA concentrations
were noticeably lower in our study whereas EPA and DHA were higher compared to Sergeant et
al. (95). Nearly all ten FA concentrations were greater in our study than Abdelmagid (94), with
85
notable higher concentrations in PA, SA, OA, LA, ARA, EPA, DPA, and DHA. A study
conducted in NA by Asselin et al. (200) found a cohort of overweight, older participants
(BMI=25.9; age=59) reported mean plasma FA concentrations (15,169 μmol/L) similar to that of
Sergeant et al. (95) and over double that of the cohort in Abdelmagid et al. (94). The primary
difference that may account for the high concentrations reported in Sergeant et al. (95) is their
use of obese middle-aged women (BMI=33.3; age=56.9) compared to healthy, young university
students (BMI=22.8; age=22.6) in Abdelmagid et al. (94). FA concentrations have been found to
strongly correlate with increases in BMI, as obese individuals exhibited modified FA status
compared to healthy controls (110–112). Furthermore, as one ages, genetic adaptations occur
which can modify FA status, primarily omega-3 PUFA (177). Taken together, the FA
concentrations reported in Sergeant et al. (95) may be more reflective of the average FA status of
North Americans due to the high prevalence of obesity in recent years. The use of healthy
university-aged students in Abdelmagid et al. (94) could reflect FA concentrations lower than
what is expected of the average NA and is representative of a cohort that has reduced risk factors
of chronic disease.
5.5.2 Anthropometric Measures
The high total FA concentrations seen in the present study are in contrast to the lower
total FA reported in Abdelmagid et al. (94) but, the BMI are nearly identical to that of the
Singaporean cohort. Our cohort, although reporting a standard healthy BMI, WC, and fasting
blood glucose, had mean FA concentrations similar to that of the obese group profiled in
Sergeant et al. (95). The Singaporean participants were nearly twice as old as the participants
reported in Abdelmagid et al. (94), which may account for some of the differences. In addition,
86
cultural and dietary differences are important modifying factors affecting FA concentrations.
First, SFA intake is greater in Singapore than that of NA (10). The countries surrounding
Singapore are the largest producers of palm oil on the planet, which is high in PA, OA, and LA
(184). When examining the mean difference between PA, OA, and LA concentrations between
our Singaporean cohort and the Canadian cohort of Abdelmagid et al. (94), those three FA alone
account for ~75% of the difference in mean total FA concentrations (calculation not shown). Due
to the abundance of palm oil produced in Southeast Asia, dietary intakes have increased which
have also been associated with a rise in CVD-related deaths (201). Palm oil has also been found
to strongly correlate with increased LDL cholesterol levels (201), which were reported to be
borderline high in our study (94). Similarly, frequency of omega-3 PUFA consumption via fish is
low in NA but, consumption in Southeast Asia are some of the highest in the world (10,52). The
mean difference between EPA, DPA, and DHA between studies accounts for ~5% of the total
FA concentrations (calculation not shown). Nearly 34.3% of adults in Singapore are overweight
or obese and 33.1% are not meeting daily physical activity requirements as of 2014 (202,203).
Measures of body fat % in both males and females in the present study were also quite high
compared to other Asian cohorts (204). In Caucasian men and women, an obese BMI
classification of 30kg/m2 corresponds to approximately 25% and 35% body fat (205), equating to
roughly the same body fat % reported of both males and females in this study, seen in Table 5.1.
This phenomenon of a low BMI but, high body fat % in Singaporean adults has also been
described in previous studies (206). Widely reported differences between Caucasian populations
and Asian populations have resulted in the WHO to make amendments to the current BMI cut-
offs for Asian cohorts which was adopted by Singapore in 2005 (207,208). Current WHO BMI
87
cut points for being considered overweight is 23-27.5 kg/m2 and obese is ≥ 27.5 kg/m2 for Asian
populations (208). Although the participants included in our study exhibited healthy levels of TG
and HDL cholesterol, and were screened for any chronic diseases, their BMI and body fat %
should not be considered to be optimal. Due to this unique paradox found in the Singaporean
population, interpretations between FA concentrations and health outcomes related to BMI in
other reports should be compared cautiously. Thus, the high FA concentrations seen in our study
can be interpreted as a risk factor to chronic disease, similar to that seen in other populations
with higher reported body fat % and BMI.
5.5.3 Correlations with Lipid Biomarkers
Unlike FA concentrations, established clinical reference ranges for TG and cholesterol
have been established and are routinely utilized in clinical practices to assess health outcomes
(11). By determining the correlation between FA concentrations and lipid biomarkers, individual
FA can be targeted to determine levels which may be deemed adequate or in excess. Mean LDL
and total cholesterol levels in this study were found to be borderline high, which is associated
with CVD disease risk (3.35 ± 0.92; 5.27 ± 1.03) (196). HDL levels were classified as being high
(1.68 ± 0.43), which has shown to promote cardioprotective effects (196). Due to age, increased
BMI, and gender exhibiting abilities to modify FA and lipid concentrations independently, they
were included as covariates in our linear regression model, seen in Table 5.5. Unfortunately,
dietary data was not available for this study. Nearly all FA exhibited a significant positive
correlation to HDL, LDL, and total cholesterol levels in our study. Only PA, PAO, and ALA did
not reach statistical significance when correlated to HDL cholesterol, which has also been in
previous studies (209–211). EPA and DHA were found to be significantly positively correlated
88
with HDL cholesterol, which is similar to many previous studies citing strong positive
correlations between omega-3 PUFA consumption and the antiatherogenic properties of HDL
cholesterol levels (176). Interestingly, the omega-6 PUFA LA and ARA, exhibited significant
positive correlations to HDL cholesterol and the strongest positive correlations to LDL
cholesterol in our study. Strong positive correlations between PA and SA to LDL and total
cholesterol levels have been reported a number of times previously; SFA intakes positively
correlate to total cholesterol levels and the subsequent risk of developing CHD (212). Results of
our study support these findings with both PA, SA, and LA exhibiting the strongest positive
correlations to LDL and TC in Table 5.5. Compared to other common vegetable cooking oils low
in SFA, palm oil, which is widely consumed in Singapore, has been found to increase LDL
cholesterol levels significantly and subsequently increasing the risk of developing type II
diabetes (201). Thus, the correlations of SFA reported could be disproportionately high based on
the dietary habits of Singaporean adults. When the R2 values in Table 5.5 are compared to those
of Abdelmagid et al. (94), our study exhibited stronger positive correlations overall. As the
cohort described in Abdelmagid et al. (94) was healthy Canadians and the Singaporeans in this
study exhibit characteristics of the low BMI and high body fat % paradox, it is not surprising that
stronger correlations with lipid biomarkers were observed in the current cohort.
The cohort of Singaporeans in the current study had generally very healthy TG levels,
with only participants in the 95th percentile and up reaching borderline high values (based on
Asian cutoffs) and none of our participants were classified as having very high TG levels (Table
5.4) (196,213). All SFA and MUFA exhibited significant positive correlations with TG (Table
5.5). These results are similar to correlations between PA, SA, PAO, OA and TG concentrations
89
as reported in previous studies (94,214,215). High levels of TG have been found to increase the
risk of developing CVD, atherosclerosis, and are often seen in obese individuals and those with
type-2 diabetes (216,217). Numerous studies, including work published by the European Food
Safety Authority, have reported omega-3 PUFA consumption reduces disease risk and benefits
heart health via reductions in TG levels (218). In our study ARA, EPA, and DHA, exhibited the
weakest correlations to TG concentrations, with EPA being the only FA not to reach statistical
significance. Southeast-Asians on average consume some of the highest levels of omega-3 PUFA
via fish in the world, which is reflected by the concentrations of EPA and DHA reported in Table
5.2 (10). Greater consumption of omega-3 PUFA is positively correlated with HDL cholesterol,
and negatively correlated with TG, resulting in a reduction of developing CVD and CHD
(176,218,219). Results in Table 5.5 however, indicate omega-3 PUFA exhibited significant
positive correlations with TG, LDL, and total cholesterol levels. High levels of these three lipid
biomarkers are considered to put an individual at an increased risk of developing CVD
(176,219). Our findings of long-chain omega-3 PUFA exhibiting positive correlations to TG
levels differs to the negative correlations commonly seen in other studies (214). As we reported
larger mean concentrations for nearly the entire FA pool profiled compared to that of a very
healthy cohort seen in Abdelmagid et al. (94), correlations between high concentrations of health
promoting FA and lipid biomarkers may be influenced by increased concentrations of all other
FA. We posit the increased omega-3 PUFA concentrations seen are still resulting in beneficial
health effects for disease reduction despite the positive correlations to TG and cholesterol
reported. This would provide reasoning for the relatively healthy values of TG and cholesterol
seen in Table 5.1, although mean total FA concentrations were similar to that of an obese cohort
90
(95). By establishing correlations between validated biomarkers and individual FA
concentrations more precise clinical recommendations can be made to reduce chronic disease
risk.
Further work is necessary to establish cut-off values of individual FA concentrations that
are deemed to be deficient or in excess. A large body of evidence from multiple cohorts across
multiple countries is necessary to establish baseline or “normal” values, and whether region, sex,
and/or, age specific ranges are required (220). Cohorts from the same country have exhibited
significantly different mean FA concentrations, showcasing that large variances are possible, and
values from all studies should not be compared equally until a better understanding of what
ranges are considered normal. Correlations between FA status and validated biomarkers in
cohorts consisting of low or high FA concentrations, contrasted with healthy populations, will
provide a more accurate measure as to where cut-off values highly correlated to disease-risk
could be established. For example, all participants between the 5th and 95th percentile in our
study exhibited optimal TG levels but, this does not equate to all participants exhibiting optimal
OA concentrations in that same range although strong correlations were seen between the two. In
Table 5.6, we observed that participants outside of the normal FA range also commonly
exceeded current lipid biomarker cut-offs, increasing chronic disease risk. Specifically,
individual FA concentrations in the ≥ 95th percentile were more commonly seen in participants
who exceeded TG or cholesterol cut-off values compared to concentrations in the ≤ 5th percentile.
In certain instances, 100% of participants who were above a lipid biomarker cut-off value was
also found to have a FA concentration in the ≥ 95th percentile. A number of our high
concentration FA exhibited normally distributed curves, similar to those displayed Abdelmagid
91
et al. (94), but, some smaller FA exhibited a skewed distribution (data now shown). Due to the
sample size of our study, the ≥ 95th percentile and ≤ 5th percentile contained only 24 participants
each, not enough to account for the total participants who exceeded LDL, high HDL, or total
cholesterol cut-offs. If we were to broaden our criteria to include the ≥ 75th percentile for both
PA and SA, 119 participants exhibit concentrations in that range of both FA. Of the 92
participants in Table 5.6 who exceeded the cut-off for LDL cholesterol, 67% and 70% exhibit
FA concentrations in the ≥ 75th percentile of PA and SA, respectively (calculation not shown).
Thus, specific FA may pose a greater health risk at values outside of the normal range and others
could be detrimental at lower relative concentrations. FA reference ranges will need to be further
profiled in multiple cohorts to assess whether our results are generalizable to the population at
large or, specifically Singaporean adults. Overall, we saw a higher proportion of participants who
exceeded a lipid biomarker cut-off also having a FA concentration in the ≥ 95th percentile
compared to those in the ≤ 5th percentile. However, at this point in time we cannot confidently
make any conclusions for cut-off values designating adequate/inadequate status of individual FA.
The findings in Table 5.6 provides evidence for individual FA and percentiles to target in future
projects for assessing FA status and chronic disease risk. Regression analysis to specifically
determine cut-off values will be required in the future once a larger database of cohorts has been
compiled and stronger correlations between biomarkers and specific percentiles of FA
concentrations have been established.
5.5.4 Sex Differences
Along with anthropometric measures, we examined concentrations of circulating FA
between sexes, (Table 5.1 and 5.3). Although Singapore consists of a unique ethnic make-up,
92
where cultural differences can influence dietary habits (181), our population was >90%
Singaporean Chinese and thus, ethnic differences were not evaluated. Commonly reported
differences in anthropometric measures between males and females were seen, such as higher
BMI and WC in males and body fat % in females. Based on the World Health Organization
Global Status Report (202) the BMI and fasting blood glucose of both males and females in this
cohort were below national averages but, males were considered to be overweight (23.3 kg/m2)
based on the BMI cut-offs for Asian populations (208).
We observed mean values for BMI and WC to be below the risk cut-off of an Asian
population for developing CVD, diabetes, and dyslipidemia in both sexes but, a number of
participants in the higher percentiles fell into a risk category (221). Of the ten FA concentrations
which were compared between males and females, significant differences were found between
SA, PAO, ALA, and DHA with females reporting the greater concentration in all four. Sex-
specific differences in individual FA status have been shown on a number of occasions
(66,111,116,117). In Abdelmagid et al. (94) when males and females were compared PAO,
ALA, and DHA were also all significantly greater except, males had greater concentrations of
ALA. In terms of percent composition, Glew et al. (197) found no significant difference between
males and females in any of the ten FA highlighted in our study. The larger dynamic range when
reporting FA concentrations compared to percent composition can result in greater statistical
significance but, may not always equate to biological significance. Independent of dietary intake,
females have exhibited a greater genetic capacity to endogenously produce omega-3 PUFA
(118). The use of oral contraceptives in some studies have found to further impact PUFA
metabolism but, more recent studies found no significant effect (119,120). Additionally,
93
pregnant and breastfeeding females have obvious metabolic differences and subsequent
nutritional needs compared to males which has also been found to modify FA status (121). Our
study excluded female participants who were taking oral contraceptives and who were pregnant
or breastfeeding during data collection. Although omega-3 PUFA were significantly greater in
females in this study, the mean FA concentration between sexes was not significantly different
(see Table 5.1). As participants were nearly all the same ethnicity and from the same
geographical area, dietary habits were likely similar. Genetic differences between sexes may
account for some of the variability shown but, dietary consumption still remains the strongest
individual modifier of FA concentrations.
5.5.5 Limitations
This study is limited by a lack of dietary data and measures of physical activity recorded.
Both physical activity and dietary consumption, particularly of fat, are modifiers of serum FA
concentrations. Inclusion of these two variables in the linear regression model would have
provided a more sensitive analysis for assessing the correlation of FA concentrations to lipid
biomarkers. Including all ten FA in the linear regression model violated the collinearity tolerance
and VIF which, dietary data may have provided a better understanding as to why specific FA
were related due to which foods were consumed.
5.6 Conclusion
In conclusion, our study provides valuable knowledge for the growing desire to develop
individual FA reference ranges for assessing chronic disease risk. We are one of the few studies
to examine a cohort outside of NA which showcases the effect of geographical and cultural
differences on individual FA concentrations. Inclusion of more cohorts with very high or low
94
concentrations of FA will help establish a greater understanding of FA levels that are reflective
of chronic disease states. Future research should be conducted in multiple global cohorts while
maintaining generalizability of the results to allow further correlations used to established FA
cut-offs to be easily compared across studies.
95
6 GENERAL DISCUSSION, LIMITATIONS and FUTURE DIRECTIONS
6.1 General Discussion
The current lack of guidelines surrounding plasma/serum FA concentrations is a major
gap for translating scientific results into clinical practice for promotion of optimal health and
reducing chronic disease risk. Validated FA reference ranges would allow for meaningful
assessments of FA status in individuals (i.e. sufficiency vs. deficiency), and provide target ranges
for health, similar to that of other lipid biomarkers commonly used in clinical practice. This
study is one of the few that have aimed to specifically address this gap in scientific literature.
Strong correlations between FA concentrations and health-related biomarkers such as age, sex,
TG, and cholesterol were found in Australian and Singaporean adults. In Chapter 4, age
significantly correlated to the status of a number of individual FA, notably EPA and DHA, in
Australian males. In Chapter 5, while Singaporean females and males exhibited a small number
of significant differences in FA status, mean total FA concentrations were similar between both
groups. When compared to similar studies conducted in NA, vastly different profiles of
individual FA were seen in this thesis. Cultural and geographical differences that impact dietary
FA consumption vary from region-to-region and are major modifying factors of FA status that
likely played a large role in the differences seen in this thesis. Whether separate reference ranges
for sex, the elderly, or geographic location are necessary will require additional studies building
upon the foundational knowledge described in this thesis. Further assessments of cohorts with
broad ranges of FA concentrations (low vs high) will provide specific targets for establishing
cut-offs that correlate to disease risk. Overall, the findings from this thesis support correlations
between FA status and health-related biomarkers in previous studies (94) and the use of large-
96
powered, healthy, international cohorts contributes foundational knowledge for developing
baseline FA reference ranges to assess chronic disease risk across various groups.
6.2 Limitations
The entirety of the data sets and biological samples analyzed in this study was secondary
data received from our international collaborators. This influenced the decision-making process
of developing our research questions based on what variables were available to be analyzed.
Primarily, the lack of dietary data available in both cohorts limits the strength of our correlation
analysis. Highlighting trends in specific food sources would have strengthened inferences made
based on blood concentrations and provided more in-depth comparison to region-specific food
consumption. Similar limitations due to the lack of physical activity data also persists. Further, as
the MAILES cohort was developed to assess prostate cancer risk, the reasoning to exclude
female participants is justified based on the original research question developed. However, the
inclusion of female participants in our study would have provided results which are generalizable
to the entire population and further expanded on sex-related differences in FA status.
Serum FA concentration is an aggregate of both dietary intake and endogenous lipid
synthesis. As SFA and MUFA have the capacity to be produced endogenously, metabolic
biomarkers of desaturase activity would help to distinguish the relative effect which dietary or
genetic factors modify FA concentrations. Endogenous lipid production could limit the
correlations seen for disease risk and certain SFA and MUFA as abnormal metabolic activity
may be the largest contributor to FA concentration, not dietary intake.
97
Finally, the processing of blood serum and subsequent analysis of chromatograms
conducted using OpenLab CDS EZChrom Edition 3.2.1 is limited by potential human error. For
the Australian cohort, two different investigators performed lipid extraction and chromatogram
analysis. A comparison analysis was conducted between the resulting FA concentrations of each
investigator which yielded a CV <10% (results not shown). The lipid extraction technique used
in this study remained consistent for all serum samples but, day-to-day intra and inter-individual
variation was still possible. The internal standard (C19:0) was added to the sample manually and
may exhibit small variabilities in quantity between samples. Similarly, chromatogram analysis
requires training to manually identify individual peaks based on the RT of the sample using a
standard reference. Peaks could also be mislabeled which is the responsibility of the investigator
to manually check intra and inter-individual variation is possible to influence the values of
smaller peaks between samples. All samples from the Singapore cohort were extracted and
analyzed by the same investigator.
6.3 Future Directions
Future studies should continue to profile large-scale, global cohorts and follow the current
best standards of practice in FA research to maintain comparability and generalizability of
results. Reporting full profiles of individual FA as absolute concentrations is recommended over
percent composition. When possible, the collection of dietary data coupled with blood
plasma/serum samples should be conducted to provide a more robust portrait of what factors may
be modifying individual FA status, and differences in global populations. Correlations between
FA concentrations and established biomarkers related to chronic disease should continue to be
expanded past TG, cholesterol, age, BMI, and sex. Continuing to assess FA reference ranges in
98
various populations of healthy participants is necessary to establish what values are considered
“normal”, and whether region/sex/age specific recommendations should be developed. Profiling
“unhealthy” cohorts which exhibit low or high FA concentrations will begin to provide targets as
to where cut-offs could be developed for use as a clinical tool for assessing chronic disease risk.
The methodology described in this study was chosen for its efficiency, generalizability, and
high-reproducibly which may be used as a guide for future work.
99
7 REFERENCES 1. Ratnayake WMN, Galli C. Fat and fatty acid terminology, methods of analysis and fat
digestion and metabolism: A background review paper. Ann Nutr Metab. 2009;55(1–3):8–43.
2. Fahy E, Subramaniam S, Brown HA, Glass CK, Merrill AH, Murphy RC, et al. A comprehensive classification system for lipids. J Lipid Res. 2005;46(5):839–61.
3. Hodson L, Skeaff CM, Fielding BA. Fatty acid composition of adipose tissue and blood in humans and its use as a biomarker of dietary intake. Prog Lipid Res. 2008;47(5):348–80.
4. Billingsley HE, Carbone S, Lavie CJ. Dietary fats and chronic noncommunicable diseases. Nutrients. 2018;10(10):1–16.
5. Shim J-S, Oh K, Kim HC. Dietary assessment methods in epidemiologic studies. Epidemiol Health [Internet]. 2014 Jul 22;7(2):e2014009. Available from: https://linkinghub.elsevier.com/retrieve/pii/0002914961902892
6. Astorg P, Bertrais S, Laporte F, Arnault N, Estaquio C, Galan P, et al. Plasma n-6 and n-3 polyunsaturated fatty acids as biomarkers of their dietary intakes: A cross-sectional study within a cohort of middle-aged French men and women. Eur J Clin Nutr. 2008;62(10):1155–61.
7. Uauy R. Dietary fat quality for optimal health and well-being: Overview of recommendations. Ann Nutr Metab. 2009;54(SUPPL. 1):2–7.
8. O’Neill CM, Minihane AM. Irish Section Postgraduate Meeting the impact of fatty acid desaturase genotype on fatty acid status and cardiovascular health in adults. Proc Nutr Soc. 2017;76(1):64–75.
9. Li SW, Wang J, Yang Y, Liu ZJ, Cheng L, Liu HY, et al. Polymorphisms in FADS1 and FADS2 alter plasma fatty acids and desaturase levels in type 2 diabetic patients with coronary artery disease. J Transl Med. 2016;14(1):1–9.
10. Micha R, Khatibzadeh S, Shi P, Fahimi S, Lim S, Andrews KG, et al. Global, regional, and national consumption levels of dietary fats and oils in 1990 and 2010: A systematic analysis including 266 country-specific nutrition surveys. BMJ [Internet]. 2014;348(April):1–20. Available from: http://dx.doi.org/doi:10.1136/bmj.g2272
11. Anderson TJ, Grégoire J, Pearson GJ, Barry AR, Couture P, Dawes M, et al. 2016 Canadian Cardiovascular Society Guidelines for the Management of Dyslipidemia for the Prevention of Cardiovascular Disease in the Adult. Can J Cardiol. 2016;32(11):1263–82.
12. Brenna JT, Plourde M, Stark KD, Jones PJ, Lin YH. Best practices for the design, laboratory analysis, and reporting of trials involving fatty acids. Am J Clin Nutr.
100
2018;108(2):211–27.
13. Brady LM, Lesauvage SVM, Saini N, Williams CM, Lovegrove JA. Comparison of dietary fat and fatty acid intake estimated by the duplicate diet collection technique and estimated dietary records. J Hum Nutr Diet. 2003;16(6):393–401.
14. Naska A, Lagiou A, Lagiou P. Dietary assessment methods in epidemiological research: Current state of the art and future prospects. F1000Research. 2017;6(0):1–8.
15. Rodrigo CP, Aranceta J, Salvador G, Varela-Moreiras G. Métodos de frecuencia de consumo alimentario. Nutr Hosp. 2015;31:49–56.
16. Kirkpatrick SI, Vanderlee L, Raffoul A, Stapleton J, Csizmadi I, Boucher BA, et al. Self-Report Dietary Assessment Tools Used in Canadian Research: A Scoping Review. Adv Nutr An Int Rev J. 2017;8(2):276–89.
17. Sublette ME, Segal-Isaacson CJ, Cooper TB, Fekri S, Vanegas N, Galfalvy HC, et al. Validation of a Food Frequency Questionnaire to Assess Intake of n-3 Polyunsaturated Fatty Acids in Subjects with and without Major Depressive Disorder. J Am Diet Assoc [Internet]. 2011 Jan;111(1):117-123.e2. Available from: https://linkinghub.elsevier.com/retrieve/pii/S0002822310016433
18. Herter-Aeberli I, Graf C, Vollenweider A, Häberling I, Srikanthan P, Hersberger M, et al. Validation of a food frequency questionnaire to assess intake of n-3 polyunsaturated fatty acids in switzerland. Nutrients. 2019;11(8):1–12.
19. da Silva DCG, Segheto W, de Lima MFC, Pessoa MC, Pelúzio MCG, Marchioni DML, et al. Using the method of triads in the validation of a food frequency questionnaire to assess the consumption of fatty acids in adults. J Hum Nutr Diet. 2018;31(1):85–95.
20. Goto C, Tokudome Y, Imaeda N, Takekuma K, Kuriki K, Igarashi F, et al. Validation study of fatty acid consumption assessed with a short food frequency questionnaire against plasma concentration in middle-aged Japanese people. Scand J Food Nutr. 2006;50(2):77–82.
21. Warensjö Lemming E, Nälsén C, Becker W, Ridefelt P, Mattisson I, Lindroos AK. Relative validation of the dietary intake of fatty acids among adults in the Swedish National Dietary Survey using plasma phospholipid fatty acid composition. J Nutr Sci. 2015;4:1–9.
22. Poppitt SD, Kilmartin P, Butler P, Keogh GF. Assessment of erythrocyte phospholipid fatty acid composition as a biomarker for dietary MUFA, PUFA or saturated fatty acid intake in a controlled cross-over intervention trial. Lipids Health Dis. 2005;4:1–10.
23. Zhang B, Wang P, Chen CG, He QQ, Zhuo SY, Chen YM, et al. Validation of an FFQ to estimate the intake of fatty acids using erythrocyte membrane fatty acids and multiple 3d
101
dietary records. Public Health Nutr. 2010;13(10):1546–52.
24. Garneau V, Rudkowska I, Paradis AM, Godin G, Julien P, Pérusse L, et al. Omega-3 fatty acids status in human subjects estimated using a food frequency questionnaire and plasma phospholipids levels. Nutr J [Internet]. 2012;11(1):1. Available from: ???
25. Patterson AC, Hogg RC, Kishi DM, Stark KD. Biomarker and Dietary Validation of a Canadian Food Frequency Questionnaire to Measure Eicosapentaenoic and Docosahexaenoic Acid Intakes from Whole Food, Functional Food, and Nutraceutical Sources. J Acad Nutr Diet [Internet]. 2012;112(7):1005–14. Available from: http://dx.doi.org/10.1016/j.jand.2012.03.030
26. Shen W, Weaver AM, Salazar C, Samet JM, Diaz-Sanchez D, Tong H. Validation of a dietary questionnaire to screen omega-3 fatty acids levels in healthy adults. Nutrients. 2019;11(7):1–16.
27. Gormaz JG, Rodrigo R, Videla LA, Beems M. Biosynthesis and bioavailability of long-chain polyunsaturated fatty acids in non-alcoholic fatty liver disease. Prog Lipid Res [Internet]. 2010;49(4):407–19. Available from: http://dx.doi.org/10.1016/j.plipres.2010.05.003
28. Castell GS, Serra-Majem L, Ribas-Barba L. ¿Qué y cuánto comemos? Método de recuerdo 24 horas. Nutr Hosp. 2015;31:46–8.
29. Kirkpatrick SI, Subar AF, Douglass D, Zimmerman TP, Thompson FE, Kahle LL, et al. Performance of the Automated Self-Administered 24-hour Recall relative to a measure of true intakes and to an interviewer-administered 24-h recall. Am J Clin Nutr. 2014;100(1):233–40.
30. Rowland MK, Adamson AJ, Poliakov I, Bradley J, Simpson E, Olivier P, et al. Field testing of the use of intake24—An online 24-hour dietary recall system. Nutrients. 2018;10(11):1–12.
31. Pannucci TE, Thompson FE, Bailey RL, Dodd KW, Potischman N, Kirkpatrick SI, et al. Comparing Reported Dietary Supplement Intakes between Two 24-Hour Recall Methods: The Automated Self-Administered 24-Hour Dietary Assessment Tool and the Interview-Administered Automated Multiple Pass Method. J Acad Nutr Diet [Internet]. 2018 Jun;118(6):1080–6. Available from: https://linkinghub.elsevier.com/retrieve/pii/S2212267218302545
32. Subar AF, Kirkpatrick SI, Mittl B, Zimmerman TP, Thompson FE, Bingley C, et al. The Automated Self-Administered 24-Hour Dietary Recall (ASA24): A Resource for Researchers, Clinicians, and Educators from the National Cancer Institute. J Acad Nutr Diet [Internet]. 2012;112(8):1134–7. Available from: http://dx.doi.org/10.1016/j.jand.2012.04.016
102
33. Procter-Gray E, Olendzki B, Kane K, Churchill L, B. Hayes R, Aguirre A, et al. Comparison of Dietary Quality Assessment Using Food Frequency Questionnaire and 24-hour-recalls in Older Men and Women. AIMS Public Heal. 2017;4(4):326–46.
34. Trijsburg L, De Vries JHM, Hollman PCH, Hulshof PJM, Van’t Veer P, Boshuizen HC, et al. Validating fatty acid intake as estimated by an FFQ: How does the 24 h recall perform as reference method compared with the duplicate portion? Public Health Nutr. 2018;21(14):2568–74.
35. Knutsen SF, Fraser GE, Beeson WL, Lindsted KD, Shavlik DJ. Comparison of adipose tissue fatty acids with dietary fatty acids as measured by 24-hour recall and food frequency questionnaire in black and white adventists: The adventist health study. Ann Epidemiol. 2003;13(2):119–27.
36. Yuan C, Spiegelman D, Rimm EB, Rosner BA, Stampfer MJ, Barnett JB, et al. Relative Validity of Nutrient Intakes Assessed by Questionnaire, 24-Hour Recalls, and Diet Records as Compared with Urinary Recovery and Plasma Concentration Biomarkers: Findings for Women. Am J Epidemiol. 2018;187(5):1051–63.
37. Ortega RM, Perez-Rodrigo C, Lopez-Sobaler AM. Métodos de evaluación de la ingesta actual: Registro o diario dietético. Nutr Hosp. 2015;31:38–45.
38. Subar AF, Freedman LS, Tooze JA, Kirkpatrick SI, Boushey C, Neuhouser ML, et al. Addressing Current Criticism Regarding the Value of Self-Report Dietary Data. J Nutr. 2015;145(12):2639–45.
39. Boushey CJ, Spoden M, Zhu FM, Delp EJ, Kerr DA. New mobile methods for dietary assessment: Review of image-assisted and image-based dietary assessment methods. Proc Nutr Soc. 2017;76(3):283–94.
40. Brunner E, stallone D, Juneja M, bingham S, Marmot M. Dietary assessment in Whitehall II: comparison of 7 d diet diary and food-frequency questionnaire and validity against biomarkers. Br J Nutr. 2001;86(3):405–14.
41. Persson PG, Carlsson S, Grill V, Östenson CG, Wallén A, Hagman U, et al. Food frequency questionnaire versus 7-day weighed dietary record information on dietary fibre and fat intake in middle-aged Swedish men. Scand J Public Health. 1998;26(1):75–80.
42. Arab L. Biomarkers of Fat and Fatty Acid Intake. J Nutr. 2003;133(3):925S-932S.
43. Pittet PG, Halliday D, Bateman PE. Site differences in the fatty acid composition of subcutaneous adipose tissue of obese women. Br J Nutr. 1979;42(1):57–61.
44. Baylin A, Campos H. The use of fatty acid biomarkers to reflect dietary intake. Curr Opin Lipidol. 2006;17(1):22–7.
103
45. Field CJ, Clandinin MT. Modulation of adipose tissue fat composition by diet: A review. Nutr Res. 1984;4(4):743–55.
46. Baylin A, Kabagambe EK, Siles X, Campos H. Adipose tissue biomarkers of fatty acid intake. Am J Clin Nutr. 2002;76(4):750–7.
47. Brevik A, Veierød MB, Drevon CA, Andersen LF. Evaluation of the odd fatty acids 15:0 and 17:0 in serum and adipose tissue as markers of intake of milk and dairy fat. Eur J Clin Nutr. 2005;59(12):1417–22.
48. Silva V, Barazzoni R, Singer P. Biomarkers of fish oil omega-3 polyunsaturated fatty acids intake in humans. Nutr Clin Pract. 2014;29(1):63–72.
49. Hudgins LC, Hirsch J. Changes in abdominal and gluteal adipose-tissue fatty acid compositions in obese subjects after weight gain and weight loss. Am J Clin Nutr [Internet]. 1991 Jun 1;53(6):1372–7. Available from: https://academic.oup.com/ajcn/article/53/6/1372/4732431
50. Beynen AC, Hermus RJJ, Hautvast JGAJ. A mathematical relationship between the fatty acid composition of the diet and that of the adipose tissue in man. Am J Clin Nutr. 1980;33(1):81–5.
51. HIRSCH J, FARQUHAR JW, AHRENS EH, PETERSON ML, STOFFEL W. Studies of Adipose Tissue in Man. Am J Clin Nutr [Internet]. 1960 Jul 1;8(4):499–511. Available from: https://linkinghub.elsevier.com/retrieve/pii/B9781483196985500306
52. Stark KD, Van Elswyk ME, Higgins MR, Weatherford CA, Salem N. Global survey of the omega-3 fatty acids, docosahexaenoic acid and eicosapentaenoic acid in the blood stream of healthy adults. Prog Lipid Res [Internet]. 2016;63:132–52. Available from: http://dx.doi.org/10.1016/j.plipres.2016.05.001
53. Sun Q, Ma J, Campos H, Hankinson SE, Hu FB. Comparison between plasma and erythrocyte fatty acid content as biomarkers of fatty acid intake in US women. Am J Clin Nutr. 2007;86(1):74–81.
54. Harris WS, Varvel SA, Pottala J V., Warnick GR, McConnell JP. Comparative effects of an acute dose of fish oil on omega-3 fatty acid levels in red blood cells versus plasma: Implications for clinical utility. J Clin Lipidol [Internet]. 2013;7(5):433–40. Available from: http://dx.doi.org/10.1016/j.jacl.2013.05.001
55. M.B. K, J.P. D, A.P.J.M. VB, M. P, M. Z. Kinetics of the incorporation of dietary fatty acids into serum cholesteryl esters, erythrocyte membranes, and adipose tissue: An 18-month controlled study. J Lipid Res [Internet]. 1997;38(10):2012–22. Available from: http://ovidsp.ovid.com/ovidweb.cgi?T=JS&PAGE=reference&D=emed4&NEWS=N&AN=1997337154
104
56. Metherel AH, Stark KD. The stability of blood fatty acids during storage and potential mechanisms of degradation: A review. Prostaglandins, Leukot Essent Fat Acids [Internet]. 2016 Jan 1;104(2):33–43. Available from: ttp://dx.doi.org/10.1016/j.plefa.2015.12.003
57. Fuhrman BJ, Barba M, Krogh V, Micheli A, Pala V, Lauria R, et al. Erythrocyte membrane phospholipid composition as a biomarker of dietary fat. Ann Nutr Metab. 2006;50(2):95–102.
58. Harris WS, Von Schacky C. The Omega-3 Index: A new risk factor for death from coronary heart disease? Prev Med (Baltim). 2004;39(1):212–20.
59. Tuck MK, Chan DW, Chia D, Godwin AK, Grizzle WE, Krueger KE, et al. Standard operating procedures for serum and plasma collection: Early detection research network consensus statement standard operating procedure integration working group. J Proteome Res. 2009;8(1):113–7.
60. ROSE HG, OKLANDER M. Improved Procedure for the Extraction of Lipids From Human Erythrocytes. J Lipid Res. 1965;6:428–31.
61. Luque-Garcia JL, Neubert TA. Sample preparation for serum/plasma profiling and biomarker identification by mass spectrometry. J Chromatogr A. 2007;1153(1–2):259–76.
62. Dunn WB, Broadhurst D, Begley P, Zelena E, Francis-Mcintyre S, Anderson N, et al. Procedures for large-scale metabolic profiling of serum and plasma using gas chromatography and liquid chromatography coupled to mass spectrometry. Nat Protoc [Internet]. 2011;6(7):1060–83. Available from: http://dx.doi.org/10.1038/nprot.2011.335
63. Köhler S, Schmid F, Settanni G. The Internal Dynamics of Fibrinogen and Its Implications for Coagulation and Adsorption. PLoS Comput Biol. 2015;11(9):1–20.
64. Yu Z, Kastenmüller G, He Y, Belcredi P, Möller G, Prehn C, et al. Differences between human plasma and serum metabolite profiles. PLoS One. 2011;6(7):1–6.
65. Furtado JD, Beqari J, Campos H. Comparison of the utility of total plasma fatty acids versus those in cholesteryl ester, phospholipid, and triglyceride as biomarkers of fatty acid intake. Nutrients. 2019;11(9).
66. Hodson L, Eyles HC, McLachlan KJ, Bell ML, Green TJ, Skeaff CM. Plasma and Erythrocyte Fatty Acids Reflect Intakes of Saturated and n–6 PUFA within a Similar Time Frame. J Nutr. 2014;144(1):33–41.
67. Arterburn LM, Hall EB, Oken H. Distribution, interconversion, and dose response of n-3 fatty acids in humans. Am J Clin Nutr. 2006;83(6).
68. Walker CG, Browning LM, Stecher L, West AL, Madden J, Jebb SA, et al. Fatty acid profile of plasma NEFA does not reflect adipose tissue fatty acid profile. Br J Nutr.
105
2015;114(5):756–62.
69. Yuzyuk T, Lozier B, Schwarz EL, Viau K, Kish-Trier E, De Biase I. Intra-individual variability of long-chain fatty acids (C12–C24) in plasma and red blood cells. Prostaglandins Leukot Essent Fat Acids [Internet]. 2018;135(March):30–8. Available from: https://doi.org/10.1016/j.plefa.2018.06.006
70. Hodge AM, Simpson JA, Gibson RA, Sinclair AJ, Makrides M, O’Dea K, et al. Plasma phospholipid fatty acid composition as a biomarker of habitual dietary fat intake in an ethnically diverse cohort. Nutr Metab Cardiovasc Dis. 2007;17(6):415–26.
71. Kobayashi M, Sasaki S, Kawabata T, Hasegawa K, Akabane M, Tsugane S. Single measurement of serum phospholipid fatty acid as a biomarker of specific fatty acid intake in middle-aged Japanese men. Eur J Clin Nutr. 2001;55(8):643–50.
72. Baylin A, Mi KK, Donovan-Palmer A, Siles X, Dougherty L, Tocco P, et al. Fasting whole blood as a biomarker of essential fatty acid intake in epidemiologic studies: Comparison with adipose tissue and plasma. Am J Epidemiol. 2005;162(4):373–81.
73. Kuriki K, Nagaya T, Tokudome Y, Imaeda N, Fujiwara N, Sato J, et al. Plasma Concentrations of (n-3) Highly Unsaturated Fatty Acids Are Good Biomarkers of Relative Dietary Fatty Acid Intakes: A Cross-Sectional Study. J Nutr. 2003;133(11):3643–50.
74. Wolk A, Furuheim M, Vessby B. Fatty Acid Composition of Adipose Tissue and Serum Lipids Are Valid Biological Markers Of Dairy Fat Intake in Men. J Nutr. 2001;131(3):828–33.
75. Kim OY, Lee SM, An WS. Impact of blood or erythrocyte membrane fatty acids for disease risk prediction: Focusing on cardiovascular disease and chronic kidney disease. Nutrients. 2018;10(10).
76. Ulmer CZ, Jones CM, Yost RA, Garrett TJ, Bowden JA. Optimization of Folch, Bligh-Dyer, and Matyash sample-to-extraction solvent ratios for human plasma-based lipidomics studies. Anal Chim Acta [Internet]. 2018;1037:351–7. Available from: https://doi.org/10.1016/j.aca.2018.08.004
77. Folch J, Lees M, Sloane Stanley G. a Simple. 1987;55(5):999–1033.
78. Bligh, E.G. and Dyer WJ. Canadian Journal of Biochemistry and Physiology. Can J Biochem Physiol. 1959;37(8).
79. Breil C, Abert Vian M, Zemb T, Kunz W, Chemat F. “Bligh and Dyer” and Folch methods for solid–liquid–liquid extraction of lipids from microorganisms. Comprehension of solvatation mechanisms and towards substitution with alternative solvents. Int J Mol Sci. 2017;18(4):1–21.
106
80. Dodds ED, McCoy MR, Rea LD, Kennish JM. Gas chromatographic quantification of fatty acid methyl esters: Flame ionization detection vs. electron impact mass spectrometry. Lipids. 2005;40(4):419–28.
81. Ostermann AI, Müller M, Willenberg I, Schebb NH. Determining the fatty acid composition in plasma and tissues as fatty acid methyl esters using gas chromatography - a comparison of different derivatization and extraction procedures. Prostaglandins Leukot Essent Fat Acids [Internet]. 2014;91(6):235–41. Available from: http://dx.doi.org/10.1016/j.plefa.2014.10.002
82. Dołowy M, Pyka A. Chromatographic methods in the separation of long-chain mono- and polyunsaturated fatty acids. J Chem. 2015;2015.
83. Bjerrum JT veite. Metabonomics: methods and protocols: preface. Vol. 1277, Methods in molecular biology (Clifton, N.J.). 2015. v.
84. Etxebarria N, Zuloaga O, Olivares M, Bartolomé LJ, Navarro P. Retention-time locked methods in gas chromatography. J Chromatogr A. 2009;1216(10):1624–9.
85. Chiu HH, Kuo CH. Gas chromatography-mass spectrometry-based analytical strategies for fatty acid analysis in biological samples. J Food Drug Anal [Internet]. 2020;28(1):60–73. Available from: https://doi.org/10.1016/j.jfda.2019.10.003
86. Shantha NC, Napolitano GE. Gas chromatography of fatty acids. J Chromatogr A [Internet]. 1992 Oct;624(1–2):37–51. Available from: https://linkinghub.elsevier.com/retrieve/pii/002196739285673H
87. Brenna JT. Fatty acid analysis by high resolution gas chromatography and mass spectrometry for clinical and experimental applications. Curr Opin Clin Nutr Metab Care. 2013;16(5):548–54.
88. Gallego SF, Hermansson M, Liebisch G, Hodson L, Ejsing CS. Total fatty acid analysis of human blood samples in one minute by high-resolution mass spectrometry. Biomolecules. 2019;9(1):1–16.
89. Pacchiarotta T, Nevedomskaya E, Carrasco-Pancorbo A, Deelder AM, Mayboroda OA. Evaluation of GC-APCI/MS and GC-FID as a complementary platform. J Biomol Tech. 2010;21(4):205–13.
90. Poole CF. Ionization-based detectors for gas chromatography. J Chromatogr A [Internet]. 2015;1421:137–53. Available from: http://dx.doi.org/10.1016/j.chroma.2015.02.061
91. King B, Westwood S. GC-FID as a primary method for establishing the purity of organic CRMs used for drugs in sport analysis. Anal Bioanal Chem. 2001;370(2–3):194–9.
92. Shah VP, Midha KK, Findlay JWA, Hill HM, Hulse JD, Mcgilveray IJ, et al.
107
art%3A10.1023%2FA%3A1007669411738. 2000;17(12).
93. Schantz MM, Powers CD, Schleicher RL, Betz JM, Wise SA. Interlaboratory analytical comparison of fatty acid concentrations in serum or plasma. Clin Chim Acta. 2016;462:148–52.
94. Abdelmagid SA, Clarke SE, Nielsen DE, Badawi A, El-Sohemy A, Mutch DM, et al. Comprehensive profiling of plasma fatty acid concentrations in young healthy canadian adults. PLoS One. 2015;10(2):1–16.
95. Sergeant S, Ruczinski I, Ivester P, Lee TC, Morgan TM, Nicklas BJ, et al. Impact of methods used to express levels of circulating fatty acids on the degree and direction of associations with blood lipids in humans. Br J Nutr. 2016;115(2):251–61.
96. Kish-Trier E, Schwarz EL, Pasquali M, Yuzyuk T. Quantitation of total fatty acids in plasma and serum by GC-NCI-MS. Clin Mass Spectrom [Internet]. 2016;2:11–7. Available from: http://dx.doi.org/10.1016/j.clinms.2016.12.001
97. Saadatian-Elahi M, Slimani N, Chajès V, Jenab M, Goudable J, Biessy C, et al. Plasma phospholipid fatty acid profiles and their association with food intakes: Results from a cross-sectional study within the European Prospective Investigation into Cancer and Nutrition. Am J Clin Nutr. 2009;89(1):331–46.
98. Khan SA. Effects of Short Term Supplementation of Fish Oil Capsules on the Blood Fatty Acid Profile of Vegetarians : A Pilot Study. 2017;6(5):19–23.
99. Sera RK, McBride JH, Higgins SA, Rodgerson DO. Evaluation of reference ranges for fatty acids in serum. J Clin Lab Anal. 1994;8(2):81–5.
100. Bradbury KE, Skeaff CM, Crowe FL, Green TJ, Hodson L. Serum fatty acid reference ranges: Percentiles from a New Zealand national nutrition survey. Nutrients. 2011;3(1):152–63.
101. Ma DWL, Arendt BM, Hillyer LM, Fung SK, McGilvray I, Guindi M, et al. Plasma phospholipids and fatty acid composition differ between liver biopsy-proven nonalcoholic fatty liver disease and healthy subjects. Nutr Diabetes. 2016;6(7):e220.
102. Khan J, Solakivi T, Seppänen H, Lappalainen-Lehto R, Järvinen S, Ronkainen J, et al. Serum lipid and fatty acid profiles are highly changed in patients with alcohol induced acute pancreatitis. Pancreatology [Internet]. 2012;12(1):44–8. Available from: http://dx.doi.org/10.1016/j.pan.2011.12.006
103. Gambino R, Bugianesi E, Rosso C, Mezzabotta L, Pinach S, Alemanno N, et al. Different serum free fatty acid profiles in NAFLD subjects and healthy controls after oral fat load. Int J Mol Sci. 2016;17(4).
108
104. Seigneur M, Freyburger G, Gin H, Claverie M, Lardeau D, Lacape G, et al. Serum fatty acid profiles in type I and type II diabetes: Metabolic alterations of fatty acids of the main serum lipids. Diabetes Res Clin Pract. 1994;23(3):169–77.
105. Szczuko M, Kaczkan M, Drozd A, Maciejewska D, Palma J, Owczarzak A, et al. Comparison of fatty acid profiles in a group of female patients with chronic kidney diseases (CKD) and metabolic syndrome (MetS)–similar trends of changes, different pathophysiology. Int J Mol Sci. 2019;20(7):1–13.
106. Kondo Y, Nishiumi S, Shinohara M, Hatano N, Ikeda A, Yoshie T, et al. Serum fatty acid profiling of colorectal cancer by gas chromatography/mass spectrometry. Biomark Med. 2011;5(4):451–60.
107. Jung Y, Cho Y, Kim N, Oh IY, Kang SW, Choi EK, et al. Lipidomic profiling reveals free fatty acid alterations in plasma from patients with atrial fibrillation. PLoS One. 2018;13(5):1–14.
108. Silva V, Singer P. Membrane fatty acid composition of different target populations: Importance of baseline on supplementation. Clin Nutr Exp [Internet]. 2015;1:1–9. Available from: http://dx.doi.org/10.1016/j.yclnex.2015.07.001
109. Qiu JF, Zhang KL, Zhang XJ, Hu YJ, Li P, Shang CZ, et al. Abnormalities in Plasma Phospholipid Fatty Acid Profiles of Patients with Hepatocellular Carcinoma. Lipids. 2015;50(10):977–85.
110. Kang M, Lee A, Yoo HJ, Kim M, Kim M, Shin DY, et al. Association between increased visceral fat area and alterations in plasma fatty acid profile in overweight subjects: A cross-sectional study. Lipids Health Dis. 2017;16(1):1–10.
111. Tremblay-Franco M, Zerbinati C, Pacelli A, Palmaccio G, Lubrano C, Ducheix S, et al. Effect of obesity and metabolic syndrome on plasma oxysterols and fatty acids in human. Steroids [Internet]. 2015;99(PB):287–92. Available from: http://dx.doi.org/10.1016/j.steroids.2015.03.019
112. Pickens CA, Sordillo LM, Comstock SS, Harris WS, Hortos K, Kovan B, et al. Plasma phospholipids, non-esterified plasma polyunsaturated fatty acids and oxylipids are associated with BMI. Prostaglandins, Leukot Essent Fat Acids [Internet]. 2015 Apr;95(1):31–40. Available from: https://linkinghub.elsevier.com/retrieve/pii/S0952327814002038
113. Boden G. Obesity and FFAs. Endocrinol Metab Clin North Am. 2008;37(3):635.
114. Guerendiain M, Montes R, López-Belmonte G, Martín-Matillas M, Castellote AI, Martín-Bautista E, et al. Changes in plasma fatty acid composition are associated with improvements in obesity and related metabolic disorders: A therapeutic approach to overweight adolescents. Clin Nutr. 2018;37(1):149–56.
109
115. Telle-Hansen VH, Christensen JJ, Formo GA, Holven KB, Ulven SM. A comprehensive metabolic profiling of the metabolically healthy obesity phenotype. Lipids Health Dis. 2020;19(1):90.
116. Tsou PL, Wu CJ. Sex-dimorphic association of plasma fatty acids with cardiovascular fitness in young and middle-aged general adults: Subsamples from NHANES 2003–2004. Nutrients. 2018;10(10).
117. Lohner S, Fekete K, Marosvölgyi T, Decsi T. Gender differences in the long-chain polyunsaturated fatty acid status: Systematic review of 51 publications. Ann Nutr Metab. 2013;62(2):98–112.
118. Burdge GC, Wootton SA. Conversion of α-linolenic acid to eicosapentaenoic, docosapentaenoic and docosahexaenoic acids in young women. Br J Nutr. 2002;88(4):411–20.
119. Fehily AMA, Dickerson JWT, Meade BW, Ellis FR. Plasma and erythrocyte membrane fatty acids in oral contraceptive users. Clin Chim Acta. 1982;120(1):41–7.
120. Jensen MD, Levine J. Effects of oral contraceptives on free fatty acid metabolism in women. Metabolism. 1998;47(3):280–4.
121. Marangoni F, Cetin I, Verduci E, Canzone G, Giovannini M, Scollo P, et al. Maternal diet and nutrient requirements in pregnancy and breastfeeding. An Italian consensus document. Nutrients. 2016;8(10):1–17.
122. Fortier M, Tremblay-Mercier J, Plourde M, Chouinard-Watkins R, Vandal M, Pifferi F, et al. Higher plasma n-3 fatty acid status in the moderately healthy elderly in southern Québec: Higher fish intake or aging-related change in n-3 fatty acid metabolism? Prostaglandins Leukot Essent Fat Acids [Internet]. 2010;82(4–6):277–80. Available from: http://dx.doi.org/10.1016/j.plefa.2010.02.010
123. Jura M, Kozak LP. Obesity and related consequences to ageing. Age (Omaha). 2016;38(1).
124. Brownie S, Coutts R. Older Australians’ perceptions and practices in relation to a healthy diet for old age: A qualitative study. J Nutr Heal Aging. 2013;17(2):125–9.
125. Freitas HR, Ferreira GDC, Trevenzoli IH, Oliveira KDJ, Reis RADM. Fatty acids, antioxidants and physical activity in brain aging. Nutrients. 2017;9(11):1–23.
126. Warensjö E, Öhrvall M, Vessby B. Fatty acid composition and estimated desaturase activities are associated with obesity and lifestyle variables in men and women. Nutr Metab Cardiovasc Dis. 2006;16(2):128–36.
127. Crowe FL, Skeaff CM, Green TJ, Gray AR. Serum n-3 long-chain PUFA differ by sex
110
and age in a population-based survey of New Zealand adolescents and adults. Br J Nutr. 2008;99(1):168–74.
128. De Groot RHM, Van Boxtel MPJ, Schiepers OJG, Hornstra G, Jolles J. Age dependence of plasma phospholipid fatty acid levels: Potential role of linoleic acid in the age-associated increase in docosahexaenoic acid and eicosapentaenoic acid concentrations. Br J Nutr. 2009;102(7):1058–64.
129. Plourde M, Chouinard-Watkins R, Vandal M, Zhang Y, Lawrence P, Brenna JT, et al. Plasma incorporation, apparent retroconversion and β-oxidation of 13C-docosahexaenoic acid in the elderly. Nutr Metab [Internet]. 2011;8(1):5. Available from: http://www.nutritionandmetabolism.com/content/8/1/5
130. Li K, Brennan L, McNulty BA, Bloomfield JF, Duff DJ, Devlin NFC, et al. Plasma fatty acid patterns reflect dietary habits and metabolic health: A cross-sectional study. Mol Nutr Food Res. 2016;60(9):2043–52.
131. Zobel EH, Hansen TW, Rossing P, von Scholten BJ. Global Changes in Food Supply and the Obesity Epidemic. Curr Obes Rep [Internet]. 2016;5(4):449–55. Available from: http://dx.doi.org/10.1007/s13679-016-0233-8
132. Herforth A, Arimond M, Álvarez-Sánchez C, Coates J, Christianson K, Muehlhoff E. A Global Review of Food-Based Dietary Guidelines. Adv Nutr. 2019;10(4):590–605.
133. Wanders AJ, Alssema M, Van Greevenbroek MJ, Elshorbagy A, Zock PL, Dekker JM, et al. Comment on Sergeant et al.: Impact of methods used to express levels of circulating fatty acids on the degree and direction of associations with blood lipids in humans. Br J Nutr. 2016;115(11):2077–8.
134. Grant JF, Martin SA, Taylor AW, Wilson DH, Araujo A, Adams RJT, et al. Cohort profile: The men androgen inflammation lifestyle environment and stress (MAILES) study. Int J Epidemiol. 2014;43(4):1040–53.
135. Bi X, Loo YT, Henry CJ. Body fat measurements in Singaporean adults using four methods. Nutrients. 2018;10(3):1–11.
136. Cardiovascular disease, What is cardiovascular disease? - Australian Institute of Health and Welfare [Internet]. [cited 2020 May 24]. Available from: https://www.aihw.gov.au/reports/heart-stroke-vascular-diseases/cardiovascular-health-compendium/contents/what-is-cardiovascular-disease
137. Fortique F. Nutrición Hospitalaria. 2019;
138. Boots AMH, Maier AB, Stinissen P, Masson P, Lories RJ, De Keyser F. The influence of ageing on the development and management of rheumatoid arthritis. Nat Rev Rheumatol [Internet]. 2013 Oct 18;9(10):604–13. Available from:
111
http://www.nature.com/articles/nrrheum.2013.92
139. de Souza Moreira B, Ferreira Sampaio R, Rossana Cavalcanti Furtado S, Corrêa Dias R, Noce Kirkwood R. The Relationship Between Diabetes Mellitus, Geriatric Syndromes, Physical Function, and Gait: A Review of the Literature. Curr Diabetes Rev. 2016;12(3):240–51.
140. 3101.0 - Australian Demographic Statistics, Jun 2019 [Internet]. [cited 2020 May 24]. Available from: https://www.abs.gov.au/ausstats/[email protected]/0/1CD2B1952AFC5E7ACA257298000F2E76?OpenDocument
141. Nichols M, Townsend N, Scarborough P, Rayner M. Cardiovascular disease in Europe 2014: Epidemiological update. Eur Heart J. 2014;35(42):2950–9.
142. Huffman SL, Harika RK, Eilander A, Osendarp SJM. Essential fats: How do they affect growth and development of infants and young children in developing countries? A literature review. Matern Child Nutr. 2011;7(SUPPL. 3):44–65.
143. Borges MC, Schmidt AF, Jefferis B, Wannamethee SG, Lawlor DA, Kivimaki M, et al. Circulating Fatty Acids and Risk of Coronary Heart Disease and Stroke: Individual Participant Data Meta-Analysis in Up to 16 126 Participants. J Am Heart Assoc. 2020;9(5):e013131.
144. Simopoulos AP. The importance of the omega-6/omega-3 fatty acid ratio in cardiovascular disease and other chronic diseases. Exp Biol Med. 2008;233(6):674–88.
145. de Roos B, Romagnolo DF. Proteomic Approaches to Predict Bioavailability of Fatty Acids and Their Influence on Cancer and Chronic Disease Prevention. J Nutr. 2012;142(7):1370S-1376S.
146. Bogie JFJ, Haidar M, Kooij G, Hendriks JJA. Fatty acid metabolism in the progression and resolution of CNS disorders. Adv Drug Deliv Rev [Internet]. 2020; Available from: https://doi.org/10.1016/j.addr.2020.01.004
147. As A, Tj B, Js B, Biswas P, Gc T, Hj M, et al. Abdelhamid AS, Brown TJ, Brainard JS, Biswas P, Thorpe GC, Moore HJ, Deane KHO, Summerbell CD, Worthington HV, Song F, Hooper L. 2020;
148. Devassy JG, Leng S, Gabbs M, Monirujjaman M, Aukema HM. Omega-3 Polyunsaturated Fatty Acids and Oxylipins in Neuroinflammation and Management of Alzheimer Disease. Adv Nutr. 2016;7(5):905–16.
149. Canhada S, Castro K, Perry IS, Luft VC. Omega-3 fatty acids’ supplementation in Alzheimer’s disease: A systematic review. Nutr Neurosci. 2018;21(8):529–38.
112
150. Plourde M, Vohl MC, Vandal M, Couture P, Lemieux S, Cunnane SC. Plasma n-3 fatty acid response to an n-3 fatty acid supplement is modulated by apoE 4 but not by the common PPAR- L162V polymorphism in men. Br J Nutr. 2009;102(8):1121–4.
151. Waern RVR, Cumming RG, Blyth F, Naganathan V, Allman-Farinelli M, Le Couteur D, et al. Adequacy of nutritional intake among older men living in Sydney, Australia: Findings from the Concord Health and Ageing in Men Project (CHAMP). Br J Nutr. 2015;114(5):812–21.
152. Kuriki K, Nagaya T, Imaeda N, Tokudome Y, Fujiwara N, Sato J, et al. Discrepancies in dietary intakes and plasma concentrations of fatty acids according to age among Japanese female dietitians. Eur J Clin Nutr. 2002;56(6):524–31.
153. Oilseed A, Review I, Plan AOF. Oilseeds Federation. 2015;
154. Schwingshackl L, Bogensberger B, Benčič A, Knüppel S, Boeing H, Hoffmann G. Effects of oils and solid fats on blood lipids: A systematic review and network meta-analysis. J Lipid Res. 2018;59(9):1771–82.
155. Iuliano S, Poon S, Wang X, Bui M, Seeman E. Dairy food supplementation may reduce malnutrition risk in institutionalised elderly. Br J Nutr. 2017;117(1):142–7.
156. Shibahara A, Yamamoto K, Nakayama T, Kajimoto G. cis-Vaccenic acid in mango pulp lipids. Lipids. 1986;21(6):388–94.
157. Djoussé L, Matsumoto C, Hanson NQ, Weir NL, Tsai MY, Gaziano JM. Plasma cis-vaccenic acid and risk of heart failure with antecedent coronary heart disease in male physicians. Clin Nutr. 2014;33(3):478–82.
158. Kumar VB, Vyas K, Buddhiraju M, Alshaher M, Flood JF, Morley JE. Changes in membrane fatty acids and delta-9 desaturase in senescence accelerated (SAMP8) mouse hippocampus with aging. Life Sci. 1999;65(16):1657–62.
159. Angela Liou Y, Innis SM. Dietary linoleic acid has no effect on arachidonic acid, but increases n-6 eicosadienoic acid, and lowers dihomo-γ-linolenic and eicosapentaenoic acid in plasma of adult men. Prostaglandins Leukot Essent Fat Acids. 2009;80(4):201–6.
160. Rizzo AM, Montorfano G, Negroni M, Adorni L, Berselli P, Corsetto P, et al. A rapid method for determining arachidonic: Eicosapentaenoic acid ratios in whole blood lipids - Correlation with erythrocyte membrane ratios and validation in a large Italian population of various ages and pathologies. Lipids Health Dis. 2010;9:1–9.
161. Saga LC, Liland KH, Leistad RB, Reimers A, Rukke EO. Relating fatty acid composition in human fingertip blood to age, gender, nationality and n-3 supplementation in the Scandinavian population. Int J Food Sci Nutr. 2012;63(7):790–5.
113
162. Burns JL, Nakamura MT, Ma DWL. Differentiating the biological effects of linoleic acid from arachidonic acid in health and disease. Prostaglandins Leukot Essent Fat Acids. 2018;135(May):1–4.
163. Kawashima H. Intake of arachidonic acid-containing lipids in adult humans: Dietary surveys and clinical trials. Lipids Health Dis. 2019;18(1):1–9.
164. Thomas MH, Pelleieux S, Vitale N, Olivier JL. Dietary arachidonic acid as a risk factor for age-associated neurodegenerative diseases: Potential mechanisms. Biochimie. 2016;130:168–77.
165. Farina EK, Kiel DP, Roubenoff R, Schaefer EJ, Cupples LA, Tucker KL. Dietary Intakes of Arachidonic Acid and α-Linolenic Acid Are Associated with Reduced Risk of Hip Fracture in Older Adults. J Nutr. 2011;141(6):1146–53.
166. Miura K, Hughes MCB, Ungerer JPJ, Smith DD, Green AC. Absolute versus relative measures of plasma fatty acids and health outcomes: example of phospholipid omega-3 and omega-6 fatty acids and all-cause mortality in women. Eur J Nutr. 2018;57(2):713–22.
167. Walker CG, Browning LM, Mander AP, Madden J, West AL, Calder PC, et al. Age and sex differences in the incorporation of EPA and DHA into plasma fractions, cells and adipose tissue in humans. Br J Nutr. 2014;111(4):679–89.
168. Mozurkewich E, Berman DR, Chilimigras J. Role of omega-3 fatty acids in maternal, fetal, infant and child wellbeing. Expert Rev Obstet Gynecol [Internet]. 2010 Jan 10;5(1):125–38. Available from: http://www.tandfonline.com/doi/full/10.1586/eog.09.66
169. Cutuli D. Functional and Structural Benefits Induced by Omega-3 Polyunsaturated Fatty Acids During Aging. Curr Neuropharmacol. 2016;15(4):534–42.
170. Flood VM, Burlutsky G, Webb KL, Wang JJ, Smith WT, Mitchell P. Food and nutrient consumption trends in older Australians: A 10-year cohort study. Eur J Clin Nutr. 2010;64(6):603–13.
171. Hill E, Hodge A, Clifton P, Shivappa N, Hebert JR, Dennerstein L, et al. Longitudinal nutritional changes in aging Australian women. Asia Pac J Clin Nutr. 2019;28(1):139–49.
172. Meyer BJ. Australians are not meeting the recommended intakes for omega-3 long chain polyunsaturated fatty acids: Results of an analysis from the 2011–2012 national nutrition and physical activity survey. Nutrients. 2016;8(3).
173. Kris-Etherton PM, Hill AM. n-3 Fatty Acids: Food or Supplements? J Am Diet Assoc. 2008;108(7):1125–30.
174. Pittaway JK, Chuang LT, Ahuja KDK, Beckett JM, Glew RH, Ball MJ. Omega-3 dietary
114
fatty acid status of healthy older adults in Tasmania, Australia: An observational study. J Nutr Heal Aging. 2015;19(5):505–10.
175. Adams J, Sibbritt D, Lui CW, Broom A, Wardle J. Q-3 fatty acid supplement use in the 45 and Up Study Cohort. BMJ Open. 2013;3(4):1–7.
176. Leslie MA, Cohen DJA, Liddle DM, Robinson LE, Ma DWL. A review of the effect of omega-3 polyunsaturated fatty acids on blood triacylglycerol levels in normolipidemic and borderline hyperlipidemic individuals. Lipids Health Dis [Internet]. 2015;14(1). Available from: http://dx.doi.org/10.1186/s12944-015-0049-7
177. Leveille P, Chouinard-Watkins R, Windust A, Lawrence P, Cunnane SC, Brenna JT, et al. Metabolism of uniformly labeled 13C-eicosapentaenoic acid and 13C-arachidonic acid in young and old men. Am J Clin Nutr. 2017;106(2):467–74.
178. Drouin G, Rioux V, Legrand P. The n-3 docosapentaenoic acid (DPA): A new player in the n-3 long chain polyunsaturated fatty acid family. Vol. 159, Biochimie. 2019. 36–48 p.
179. Bradbury KE, Murray Skeaff C, Green TJ, Gray AR, Crowe FL. The serum fatty acids myristic acid and linoleic acid are better predictors of serum cholesterol concentrations when measured as molecular percentages rather than as absolute concentrations. Am J Clin Nutr. 2010;91(2):398–405.
180. Mocking RJT, Assies J, Lok A, Ruhé HG, Koeter MWJ, Visser I, et al. Statistical methodological issues in handling of fatty acid data: Percentage or concentration, imputation and indices. Lipids. 2012;47(5):541–7.
181. Tan KHX, Tan LWL, Sim X, Shyong Tai E, Lee JJM, Chia KS, et al. Cohort profile: The Singapore Multi-Ethnic Cohort (MEC) study. Int J Epidemiol. 2018;47(3):699-699J.
182. Whitton C, Ma Y, Bastian AC, Fen Chan M, Chew L. Fast-food consumers in singapore: Demographic profile, diet quality and weight status. Public Health Nutr. 2014;17(8):1805–13.
183. Henry CJ, Kaur B, Quek RYC. Are Asian foods as “fattening” as western-styled fast foods? Eur J Clin Nutr [Internet]. 2020;74(2):348–50. Available from: http://dx.doi.org/10.1038/s41430-019-0537-3
184. Koushki M, Nahidi M, Cheraghali F. Physico-chemical properties, fatty acid profile and nutrition in palm oil. J Paramed Sci. 2015;6(3):117–34.
185. Jiang YW, Sheng LT, Pan XF, Feng L, Yuan JM, Pan A, et al. Midlife Dietary Intakes of Monounsaturated Acids, n-6 Polyunsaturated Acids, and Plant-Based Fat Are Inversely Associated with Risk of Cognitive Impairment in Older Singapore Chinese Adults. J Nutr. 2020;150(4):901–9.
115
186. Seah JYH, Ong CN, Koh WP, Yuan JM, Van Dam RM. A Dietary Pattern Derived from Reduced Rank Regression and Fatty Acid Biomarkers Is Associated with Lower Risk of Type 2 Diabetes and Coronary Artery Disease in Chinese Adults. J Nutr. 2019;149(11):2001–10.
187. Brostow DP, Odegaard AO, Koh WP, Duval S, Gross MD, Yuan JM, et al. Omega-3 fatty acids and incident type 2 diabetes: The Singapore Chinese Health Study. Am J Clin Nutr. 2011;94(2):520–6.
188. Butler LM, Yuan J-M, Huang JY, Su J, Wang R, Koh W-P, et al. Plasma fatty acids and risk of colon and rectal cancers in the Singapore Chinese Health Study. npj Precis Oncol [Internet]. 2017;1(1):1–9. Available from: http://dx.doi.org/10.1038/s41698-017-0040-z
189. Gago-Dominguez M, Yuan JM, Sun CL, Lee HP, Yu MC. Opposing effects of dietary n-3 and n-6 fatty acids on mammary carcinogenesis: The Singapore Chinese Health Study. Br J Cancer. 2003;89(9):1686–92.
190. Koh WP, Dan YY, Goh GBB, Jin A, Wang R, Yuan JM. Dietary fatty acids and risk of hepatocellular carcinoma in the Singapore Chinese health study. Liver Int. 2016;36(6):893–901.
191. Sun Y, Koh HWL, Choi H, Koh WP, Yuan JM, Newman JW, et al. Plasma fatty acids, oxylipins, and risk of myocardial infarction: The Singapore Chinese Health Study. J Lipid Res. 2016;57(7):1300–7.
192. Soh AZ, Chee CB, Wang Y-T, Yuan J-M, Koh W-P. Dietary Cholesterol Increases the Risk whereas PUFAs Reduce the Risk of Active Tuberculosis in Singapore Chinese. J Nutr. 2016;146(5):1093–100.
193. Stern MC, Butler LM, Corral R, Joshi AD, Yuan JM, Koh WP, et al. Polyunsaturated fatty acids, DNA repair single nucleotide polymorphisms and colorectal cancer in the Singapore Chinese Health Study. J Nutrigenet Nutrigenomics. 2010;2(6):273–9.
194. Sun Y, Koh W-P, Yuan J-M, Choi H, Su J, Ong CN, et al. Plasma α-Linolenic and Long-Chain ω-3 Fatty Acids Are Associated with a Lower Risk of Acute Myocardial Infarction in Singapore Chinese Adults. J Nutr. 2016;146(2):275–82.
195. Bi X, Loo YT, Henry CJ. Ultrasound measurement of intraabdominal fat thickness as a predictor of insulin resistance and low HDL cholesterol in Asians. Nutrition [Internet]. 2018;55–56:99–103. Available from: https://doi.org/10.1016/j.nut.2018.04.003
196. Tai ES, Chia BL, Bastian AC, Siang Jin Chua T, Chih Wei Ho S, Koh TS, et al. Ministry of health clinical practice guidelines: Lipids. Singapore Med J. 2017;58(3):155–66.
197. Glew RH, Chuang L Te, Berry T, Okolie H, Crossey MJ, VanderJagt DJ. Lipid profiles and trans fatty acids in serum phospholipids of semi-nomadic fulani in Northern Nigeria. J
116
Heal Popul Nutr. 2010;28(2):159–66.
198. Schwertner HA, Mosser EL. Comparison of lipid fatty acids on a concentration basis vs weight percentage basis in patients with and without coronary artery disease or diabetes. Clin Chem. 1993;39(4):659–63.
199. Whitton C, Neelakantan N, Ong CN, van Dam RM. Reproducibility of Dietary Biomarkers in a Multiethnic Asian Population. Mol Nutr Food Res. 2019;63(8):1–7.
200. Asselin C, Ducharme A, Ntimbane T, Ruiz M, Fortier A, Guertin MC, et al. Circulating levels of linoleic acid and HDL-cholesterol are major determinants of 4-hydroxynonenal protein adducts in patients with heart failure. Redox Biol [Internet]. 2014;2(1):148–55. Available from: http://dx.doi.org/10.1016/j.redox.2013.12.009
201. Epidemiology N, Sun Y, Neelakantan N, Wu Y, Lote-oke R, Pan A, et al. Palm Oil Consumption Increases LDL Cholesterol Compared with Vegetable Oils Low in Saturated Fat in a Meta-Analysis of Clinical. J Nutr. 2015;145:1549–58.
202. World Health Organization. Global Status Report On Noncommunicable Diseases 2014. 2014;
203. Lee YS, Biddle S, Chan MF, Cheng A, Cheong M, Chong YS, et al. Health promotion board–ministry of health clinical practice guidelines: Obesity. Singapore Med J. 2016;57(6):292–300.
204. Kim CH, Park HS, Park M, Kim H, Kim C. Optimal cutoffs of percentage body fat for predicting obesity-related cardiovascular disease risk factors in Korean adults. Am J Clin Nutr. 2011;94(1):34–9.
205. Deurenberg-Yap M, Chew SK, Deurenberg P. Elevated body fat percentage and cardiovascular risks at low body mass index levels among Singaporean Chinese, Malays and Indians. Obes Rev. 2002;3(3):209–15.
206. Deurenberg-Yap M, Schmidt G, Van Staveren WA, Deurenberg P. The paradox of low body mass index and high body fat percentage among Chinese, Malays and Indians in Singapore. Int J Obes. 2000;24(8):1011–7.
207. Nishida C, Barba C, Cavalli-Sforza T, Cutter J, Deurenberg P, Darnton-Hill I, et al. Appropriate body-mass index for Asian populations and its implications for policy and intervention strategies. Lancet. 2004;363(9403):157–63.
208. Jih J, Mukherjea A, Vittinghoff E, Nguyen TT, Tsoh JY, Fukuoka Y, et al. Using appropriate body mass index cut points for overweight and obesity among Asian Americans. Prev Med (Baltim) [Internet]. 2014 Aug;65(1):1–6. Available from: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3624763/pdf/nihms412728.pdf
117
209. Clandinin MT, Cook SL, Konard SD, French MA. The effect of palmitic acid on lipoprotein cholesterol levels. Int J Food Sci Nutr. 2000;51(SUPPL.):4–10.
210. Wendland E, Farmer A, Glasziou P, Neil A. Effect of α linolenic acid on cardiovascular risk markers: A systematic review. Heart. 2006;92(2):166–9.
211. Hiraoka-Yamamoto J, Ikeda K, Negishi H, Mori M, Hirose A, Sawada S, et al. Serum lipid effects of a monounsaturated (palmitoleic) fatty acid-rich diet based on macadamia nuts in healthy, young japanese women. Clin Exp Pharmacol Physiol. 2004;31(SUPPL.2):37–8.
212. Williams CM, Salter A. Saturated fatty acids and coronary heart disease risk: The debate goes on. Curr Opin Clin Nutr Metab Care. 2016;19(2):97–102.
213. Triglycerides: Why do they matter? - Mayo Clinic [Internet]. [cited 2020 Jun 7]. Available from: https://www.mayoclinic.org/diseases-conditions/high-blood-cholesterol/in-depth/triglycerides/art-20048186
214. Grimsgaard S, Bønaa KH, Bjerve KS. Fatty acid chain length and degree of unsaturation are inversely associated with serum triglycerides. Lipids. 2000;35(11):1185–93.
215. DiNicolantonio JJ, O’Keefe JH. Effects of dietary fats on blood lipids: A review of direct comparison trials. Open Hear. 2018;5(2):1–5.
216. Talayero BG, Sacks FM. The role of triglycerides in atherosclerosis. Curr Cardiol Rep. 2011;13(6):544–52.
217. Alexopoulos A-S, Qamar A, Hutchins K, Crowley MJ, Batch BC, Guyton JR. Triglycerides: Emerging Targets in Diabetes Care? Review of Moderate Hypertriglyceridemia in Diabetes. Curr Diab Rep [Internet]. 2019 Apr 26;19(4):13. Available from: http://link.springer.com/10.1007/s11892-019-1136-3
218. Panel E, Nda A. Scientific Opinion on the substantiation of health claims related to eicosapentaenoic acid (EPA), docosahexaenoic acid (DHA), docosapentaenoic acid (DPA) and maintenance of normal cardiac function (ID 504, 506, 516, 527, 538, 703, 1128, 1317, 1324, 1325),. EFSA J. 2010;8(10).
219. Arca M, Borghi C, Pontremoli R, De Ferrari GM, Colivicchi F, Desideri G, et al. Hypertriglyceridemia and omega-3 fatty acids: Their often overlooked role in cardiovascular disease prevention. Nutr Metab Cardiovasc Dis [Internet]. 2018;28(3):197–205. Available from: https://doi.org/10.1016/j.numecd.2017.11.001
220. Häggström M. Establishment and clinical use of reference ranges. WikiJournal Med. 2014;1(1):1–7.
221. Lam BCC, Koh GCH, Chen C, Wong MTK, Fallows SJ. Comparison of Body Mass Index
118
(BMI), Body Adiposity Index (BAI), Waist Circumference (WC), Waist-To-Hip Ratio (WHR) and Waist-To-Height Ratio (WHtR) as predictors of cardiovascular disease risk factors in an adult population in Singapore. PLoS One. 2015;10(4):1–15.
8 APPENDIX A
Table 8.1 Absolute measures of major fatty acids in plasma and serum
Fatty Acid Plasma (μg/mL) Serum (μg/mL) Difference (μg/mL) p value *
16:0
(PA)
805.75 ± 25.90 853.61 ± 29.07 -47.86 ± 13.00 0.22
16:1c9
(POA)
101.04 ± 7.45 111.12 ± 10.66 -10.08 ± 4.89 0.44
18:0
(SA)
231.67 ± 5.66 245.74 ± 6.13 -14.06 ± 3.49 0.09
18:1c9
(OA)
805.54 ± 29.94 855.85 ± 37.60 -48.31 ± 18.45 0.32
18:2n6
(LNA)
822.23 ± 21.37 863.39 ± 23.62 -41.16 ± 11.39 0.20
18:3n3
(ALA)
22.23 ± 1.12 23.07 ± 1.21 -0.83. ± 0.34 0.47
119
20:5n3
(EPA)
48.26 ± 3.83 51.45 ± 4.41 -3.19 ± 0.90 0.59
20:4n6
(ARA)
223.63 ± 6.18 235.11 ± 6.45 -11.47 ± 3.25 0.20
22:6n3
(DHA)
71.88 ± 2.80 73.70 ± 3.24 -1.83 ± 1.49 0.67
Data are mean ± Standard error of means. Absolute (μg/mL) concentration measures of individual fatty acid plasma and serum mean based on 98 samples from the MAILES cohort. Difference mean was composed of the difference between each participants plasma and serum (plasma minus serum) levels. *Significant difference compared participant levels of plasma and serum, p<0.05.
Table 8.2 Relative percent composition of major fatty acids in plasma and serum
Fatty Acid Plasma (%) Serum (%) Difference (%) p value*
16:0
(PA)
22.53 ± 0.18 22.75 ± 0.19 -0.22 ± 0.11 0.41
16:1c9
(POA)
2.71 ± 0.13 2.74 ± 0.14 -0.04 ± 0.03 0.85
18:0
(SA)
6.60 ± 0.06 6.69 ± 0.09 -0.09 ± 0.07 0.40
120
18:1c9
(OA)
22.48 ± 0.34 22.40 ± 0.38 0.08 ± 0.17 0.88
18:2n6
(LNA)
22.58 ± 0.45 23.66 ± 0.49 -0.08 ± 0.18 0.90
18:3n3
(ALA)
0.61 ± 0.02 0.60 ± 0.02 0.01 ± 0.00 0.99
20:5n3
(EPA)
1.41 ± 0.12 1.42 ± 0.12 0.00 ± 0.01 0.99
20:4n6
(ARA)
6.48 ± 0.16 6.51 ± 0.16 -0.03 ± 0.05 0.88
22:6n3
(DHA)
2.09 ± 0.08 2.04 ± 0.09 0.05 ± 0.04 0.66
Data are mean ± standard error of means. Relative percentage (%) composition measures of individual fatty acids in plasma and serum mean based on 98 samples from the MAILES cohort. Difference mean was composed of the difference between each participants plasma and serum (plasma minus serum) levels. *Significant difference compared participant levels of plasma and serum, p<0.05.
121
Figure 8.1 Bland-Altman plot of the difference of plasma and serum fatty acids
100 200 300 400 500
-300
-200
-100
0
100
200
diffe
renc
e pl
asm
a - s
erum
(ug/
ml)
average plasma + serum (ug/ml)
A
500 1000 1500 2000 2500
-1500
-1000
-500
0
500
diffe
renc
e pl
asm
a - s
erum
(ug/
ml)
average plasma + serum (ug/ml)
B
500 1000 1500
-400
-200
0
200
400
diffe
renc
e pl
asm
a - s
erum
(ug/
ml)
average plasma + serum (ug/ml)
C
5 10 15
-6
-4
-2
0
2
diffe
renc
e pl
asm
a - s
erum
(%)
average plasma + serum (%)
D
10 20 30 40 50
-5
0
5
10
15
diffe
renc
e pl
asm
a - s
erum
(%)
average plasma + serum (%)
E
10 20 30 40
-10
-5
0
5
10
15
diffe
renc
e pl
asm
a - s
erum
(%)
average plasma + serum (%)
F
122
Differences in major fatty acids between 98 matched plasma and serum samples in absolute (A-C) and relative (D-F) terms. Major fatty acid Bland-Altman Plot of: 18:0 (A, D), 18:1 c9 (B, E), and 18:2 n6 (C, F) is displayed. For each participant, the difference was calculated subtracting the amount of plasma to serum and was compared to the average between the amount of plasma and serum. The dashed line The dashed line ( - - - ) represents the mean difference between plasma and serum, with each dotted line ( . . . . . ) reflecting the lower limit (LL) and upper limit (UL) with 95% confidence below and above the mean. In 2A the LL is – 81.71 µg/mL, mean is -14.06 µg/mL, and UL is 53.59 µg/mL. In 2B, the LL is - 406.4 µg/mL, mean is - 48.31 µg/mL, UL is 309.70 µg/mL. For 2C, LL is – 105.00 µg/mL, mean is -10.08 µg/mL, UL is 84.84 µg/mL. For 2D, LL is -1.40 %, mean -0.09 %, and UL is 1.22 %. For 2E, LL is -3.31 %, mean 0.08 %, and UL is 3.50 %. For 2F, LL is – 3.53 %, mean -0.08 %, and UL is 3.36 %.
123
8.1 Appendix B
Table 8.3 Mean concentration of select fatty acids across multiple decades
Total Decade Fatty Acid Common Name Mean SD 40-49 50-59 60-69 70+
16:0 Palmitic acid 3377.02 1229.73 3584.67 ± 1302.78
3490.30 ± 1336.35
3338.58 ± 1174.77
3064.66 ± 997.68
18:0 Stearic acid 902.32 275.66 947.86 ± 284.72
940.09 ± 294.50
905.53 ± 274.19
802.70 ± 212.39
16:1c9 Palmitoleic acid 410.45 276.39 441.18 ± 272.27
419.65 ± 256.32
411.05 ± 262.23
366.17 ± 316.43
18:1c9 Oleic acid 3044.74 1221.38 3326.91 ± 1396.73
3090.00 ± 1164.05
2965.07 ± 1146.44
2794.29 ± 1132.38
18:2n6 Linoleic acid 3096.36 909.12 3372.08 ±
924.59 3198.70 ±
872.85 3013.36 ±
912.55 2783.13 ±
828.40
20:4n6 Arachidonic acid 764.14 222.03 797.51 ± 212.24
799.23 ± 215.93
755.50 ± 243.19
695.26 ± 191.72
18:3n3 Alpha-linolenic acid 88.28 50.15 90.60 ± 51.54
88.80 ± 45.97
90.55 ± 60.36
82.38 ± 38.47
20:5n3 Eicosapentaenoic acid 158.58 121.37 121.95 ±
67.27 151.46 ±
99.22 173.24 ± 109.71
187.35 ± 180.49
22:5n3 Docosapentaenoic acid 70.83 23.47 69.12 ± 22.07
71.70 ± 24.40
72.97 ± 24.35
68.83 ± 22.39
22:6n3 Docosahexaenoic acid 217.71 94.8 190.29 ±
81.60 208.59 ±
94.04 233.47 ±
95.56 238.19 ±
99.60
Values expressed as μmol/L ± SD
SD, standard deviation
