Content uploaded by Stephan Peters
Author content
All content in this area was uploaded by Stephan Peters on Jan 11, 2019
Content may be subject to copyright.
research into the relationship between
foods and health was mainly been focused
on the eects of nutrients. This is called
the reductionist approach: the study of
individual nutrients in isolation rather
than considering their eects when eaten
in a food component or as a whole food.1
Reductionism
So, how does this reductionist approach
work in general? To begin with, there is
epidemiological science, which tries to
find associations between the intake of
foods and aspects of health among entire
populations or their subpopulations.
Subsequently, if any associations are
found, scientists try to find a mechanism
behind the observed eects, mainly based
on the assumption that nutrients play a
role in this mechanism. The interaction
between the nutrient and the supposed
mechanism can then be tested in vitro or
on animals in the lab. Finally, for ultimate
proof, a placebo-controlled study is
conducted to prove the nutrient’s health
eect on humans. Thanks to this approach
we know a lot about the eects of individual
nutrients in the body. For example, it was
discovered that amino acids, and in
particular the amino acid leucine, stimulate
protein synthesis in skeletal muscles. We
also know a lot about components in food
with antioxidative eects, such as some
vitamins and glutathione.
Reductionist science doesn't
always translate
Unfortunately, science always comes
with uncertainties. This also applies to
nutrition science. Considering food
health eects only by its nutrients may
lead to misinterpretations, with all its
The reductionist approach has delivered essential insights in nutrition
science. However, to fully understand the health effects of whole foods
and food patterns, nutrition science has to move up to a next level.
The concept of the food matrix shows that food is more than the sum
of its nutrients.
The food matrix:
nutrition science shifts from
nutrients to whole foods
TEKST DR. STEPHAN PETERS (NZO, DUTCH
DAIRY ASSOCIATION) AND DR. EMMA FEENEY
(UNIVERSITY COLLEGE DUBLIN, INSTITUTE OF
FOOD AND HEALTH)
Nutrient science
Many of the major discoveries of the
role of nutrients with respect to health
were made decades, and in some cases
centuries, ago. These were related to
severe nutrient deficiencies. In the 17th
century, it was discovered that scurvy
could be prevented by eating citrus fruit.
Nowadays, we know that scurvy was a
severe vitamin C deficiency. Another
famous example is beriberi, vitamin B1
(thiamine) deficiency. The discovery of
the role of nutrients in metabolism is very
important for understanding the health
eects of foods. Until the 1990s, scientific
Food matrix
REPORT
This article was published in Voeding Magazine 3-2018
1
consequences. An infamous example is
the relation between beta-carotene and
cancer. In the 1980s, epidemiologists
found a clear protective association
between fruit intake and lung cancer.
Higher intake of fruit was associated
with higher serum levels of beta-carotene.
Also, beta-carotene is an antioxidant
which scientists expected to have anti-
carcinogenic properties. Based on these
facts, clinical studies were done by giving
smokers supplements with beta-carotene
(or a placebo). Subsequently, the incidence
of lung cancer was investigated. The
scientists expected a protective role of
beta-carotene on the incidence of lung
cancer, but to their great surprise they
found an even higher incidence of lung
cancer with smokers and patients with
asbestosis when compared to a placebo.2
Therefore, the protective eects of fruit
against cancer cannot be explained by
(only) beta-carotene. There must be
something else that comes with fruit
consumption that may explain its health
eect. A second example: There is a clear
scientific consensus that consumption of
fish protects against cardiovascular
disease. For this reason food-based dietary
guidelines advise to consume one portion
of oily fish once a week.3 As this health
eect is mainly found with the consump-
tion of oily fish, its health eects were
explained by its content of omega-3 fatty
acid (fish-oil; EPA/DHA). Thus, many
dietary guidelines still advise taking
fish-oil supplements for those who do not
(want to) eat fish. Very recently, however,
a detailed meta-analysis showed that
taking fish-oil supplements does not
protect against heart disease, stroke
and cardiovascular death.4 Again, as
with the former example, the health
eects of fish consumption are not
explained by the eect of the expected
nutrients individually. Both examples
show that the reductionist approach also
has its flaws. What is going on?
Epidemiological flaws
One explanation of the surprising
results mentioned above, can be found
in the flaws of epidemiological studies.
Epidemiological studies can find asso-
ciations between the intake of foods or
nutrients and risks of, for example,
non-communicable diseases. Associations,
however, do not automatically imply causal
relationships. In order to confirm an actual
causal relationship, placebo-controlled
studies are needed and, preferably,
more than one. But, as shown above,
the hypothecation based on results in
epidemiological studies can go wrong.
When few studies are available, there is
a chance that an association found is not
a causal one but that this is confounded
with another factor. These so-called
confounders are factors that disturb or
even reverse a suspected causal relation-
ship. Confounders can be age or a healthy
lifestyle. Imagine that an association is
found between the intake of fruit and
vegetables and a decreased risk on
cardiovascular disease. Is this a causal
relationship or are the people who eat
a lot of fruit and vegetables the same
people with a healthy and active lifestyle?
Because of such confounders, association
can be mistakenly ascribed to a causal
relationship. Luckily, there are techniques
in epidemiological studies to rule out
confounders. In addition, when several
epidemiological studies lead to the same
conclusions with respect to a relationship,
this may decrease the risk of confounders.
These studies can be pooled in meta-
analyses. The more studies that a
meta-analysis includes, the greater the
probability that the association found is
a causal one.
Foods versus nutrients
Often high-quality epidemiological
evidence of health eects of foods cannot
be explained by single nutrients alone.
For example, think of the protective eects
of fruits and vegetables on cardiovascular
disease, the protective eects of dairy on
type 2 diabetes, cardiovascular disease
and colon cancer and the protective eects
of whole-grain products on cardiovascular
disease.3 The mechanisms behind these
associations are still unknown. In order to
find explanations for these health eects
we have to leave the reductionist approach
of food science and examine the health
eects of whole foods following a more
holistic or multi-mechanistic approach.
The dairy food group is very interesting
from a health perspective, because it
generally contains a relatively high level
of saturated fatty acids and salt (in many
cheeses for instance), which would predict
an increased risk of cardiovascular
disease. However, studies have shown no
or even a protective eect of dairy intake
with respect to coronary heart disease.5-7
In addition, although there is a convincing
Evidence of health
effects of foods can
often not be
explained by single
nutrients alone
2
amount of evidence that yogurt protects
against type 2 diabetes, this eect is
dicult to explain by one of the individual
nutrients alone.8-11
The food matrix
Food is more than just the sum of its
nutrients. A diet does not consist of single
nutrients but of whole foods, either alone
or combined with other foods, consumed
as part of a meal. Additionally, foods have
both physically and nutritionally complex
structures, which interact with each other.
This can have an influence on digestion
and absorption of the nutrients, and could
even alter the bioactive properties of the
nutrients themselves. These matrix eects
or interactions are not predictable from
the nutrient content of the food alone.
The concept of the food matrix may help
us explain the dierent health aspects
found in foods or its nutrients. Recently, a
report of an expert meeting was published
to help define the food or dairy matrix.12
This report reviews what is known about
matrix eects and what research is needed
to further explain or explore food matrix
eects.
The cheese matrix effects
Although cheese intake is a contributor
to sodium and saturated fatty acid intake
in the Western diet (Dutch National
Food Consumption Survey, RIVM 2018),
it is not associated with an increased risk
of cardiovascular disease (CVD),
hypertension or stroke.13,14 A recent
meta-analysis of 15 cohort studies confirms
that, overall, eating cheese is linked to a
significant 10% reduction in the risk of
cardiovascular diseases (14% for coronary
diseases, 10% for strokes). A dose-response
analysis suggests that the "optimum"
consumption is about 40 g/day.6 Given the
sodium and SFA content, this suggests an
eect of the overall nutrients in cheese on
heart health, which is in agreement with
many larger observational studies and
with an RCT into cheese consumption,
showing that SFA when eaten in the form
of cheese is significantly less LDL-raising
in comparison with other dairy products.15,16
Various components (calcium, phosphorus,
the milk fat globular membrane (MFGM)
or sphingolipids and starter cultures) in
cheese are all suggested to have an impact
on this.17 In addition, calcium in cheese is
suggested to increase faecal fat excretion
through the formation of insoluble soaps
with the fat and, consequently, reducing
fat absorption in the digestive tract.
Calcium and phosphorus may also
precipitate together and bind both bile
acids and fat, thus impacting hepatic
eux and therefore blood lipid content,
although these mechanisms remain to be
confirmed.12 The sphingolipids content in
cheese may also impact the cholesterol
metabolism through genetic regulatory
pathways associated with MFGM.18
This appears to be confirmed in a recent
intervention study showing that the
addition of MFGM to a high-fat palm
oil-based meal reduces post-prandial total
Food matrix
REPORT
Figure 1. Matrixeffects of cheese and
cheese-equivalents on changes in
plasma cholesterol concentrations.20
Changes in total cholesterol (∆Total-C),
HDL-cholesterol (∆HDL-C) and LDL-chole-
sterol (∆LDL-C) after six weeks consump-
tion of full-fat cheese (FFC), reduced-fat
cheese + equivalent butter (RFC) and
only butter with additional calcium
caseinate and a calcium supplement
(CC+B). Statistically tested with AN-
COVA. More information in paragraph
‘Study with cheddar cheese’.
* p<0,05 compared to FFC.
** p<0,05 compared to RFC.
+0,4
+0,2
0
-0,2
-0,4
-0,6
∆ Total-C ∆ HDL-C ∆ LDL-C
*
**
*
**
FFC
RFC
CC+B
3
and LDL-cholesterol.19 Finally, cheese
starter cultures may impact the resultant
gut microbiota, leading to a higher
number of short-chain fatty acid (SCFA)
producing colonies and so increasing
SCFA absorption in the gut, which in
turn impacts the LDL-cholesterol
metabolism.17
Study with cheddar cheese
A recent study varied the degree of the
content of main components within a
cheese matrix.20 (Figure 1) Participants
were divided into three groups and took
approximately 42g dairy fat per day for six
weeks in the form of (A) full-fat cheddar
cheese, (B) reduced-fat cheese + equivalent
butter or (C) all in the form of butter with
additional calcium caseinate and a calcium
supplement, such that the degree of the
matrix was the main variable manipulated.
After six weeks, a significant reduction in
LDL-c was observed in the full-fat cheese
vs. the butter group with the reduced-fat
cheese falling between the two. This study
demonstrated that the eect of the nutrients
eaten in the whole food significantly
diered from that of the nutrients or
components eaten separately and this
appears to verify that the "whole" is indeed
more than the sum of its parts when it
comes to cheese.
Conclusion
In this article, and the above-mentioned
review of Thorning and some others, it is
suggested that the purported detrimental
eects of saturated fatty acids on cardio-
metabolic health may in fact be nullified
when they are consumed as part of food
matrices, such as those in cheese, yogurt
and other dairy foods.5,7,17,21 Cheese in
particular has been a focus, as most
casein-based cheeses are relatively high
in fat, (23-35%), approximately 60% of
which is saturated fat, and yet they do not
appear to raise LDL-cholesterol in human
feeding studies nor are they associated
with negative cardiometabolic eects.
Thus, the focus on low-fat dairy products
in current guidelines is not entirely
supported by the existing literature and
may need to be revised on the basis of
this evidence.
References
1 Hoffmann I. Transcending reductionism in
nutrition research. Am J Clin Nutr.
2003;78(3 Suppl):514S-516S.
2 Druesne-Pecollo N, Latino-Martel P, Norat T,
et al. Beta-carotene supplementation and cancer
risk: a systematic review and metaanalysis of
randomized controlled trials. Int J Cancer.
2010;127(1):172-184.
3 Gezondheidsraad. Richtlijnen goede voeding. 2015.
4 Abdelhamid A, Brown T, Brainard J, et al. Omega-3
fatty acids for the primary and secondary preven-
tion of cardiovascular disease. Cochrane Database
of Systematic Reviews 2018. 2018(7).
5 Alexander DD, Bylsma, L.C., Vargas, A.J., Cohen,
S.S., Doucette, A., Mohamed, M., Irvin, S.R., Miller,
P.E., Watson, H., Fryzek, J.P. Dairy consumption
and CVD: a systematic review and meta-analysis.
Br J Nutr. 2016;115:737-750.
6 Chen GC, Wang, Y., Tong X., Szeto, I.M.Y., Smit, G.,
Li, Z-N., Qin, L-Q. Cheese consumption and risk of
cardiovascular disease: a meta -analysis of pros-
pective studies. Eur J Nutr. 2017;56:2565-2575.
7 de Goede J, Soedamah-Muthu, S., Gijsbers, L.,
Geleijnse, J. Dairy Consumption and Risk of Stroke:
A Systematic Review and Updated Dose–Response
Meta-Analysis of Prospective Cohort Studies.
J Am Heart Assoc 2016;DOI: 10.1161/
JAHA.115.002787.
8 Brouwer-Brolsma EM SD, Singh-Povel CM, Feskens
EJM. Dairy product consumption is associated
with pre-diabetes and newly diagnosed type 2
diabetes in the Lifelines Cohort Study. Br J Nutr.
2018;119(4)
9 Gille D, Schmid A, Walther B, Vergeres G. Fermen-
ted Food and Non-Communicable Chronic Disea-
ses: A Review. Nutrients. 2018;10(4).
10 Guo J, Givens I, Astrup A, et al. Verzion 180829:
The impact of dairy products in development of
type 2 diabetes: where does the evidence stand
in 2018?
11 Panahi S, Tremblay, A. The potential role of yogurt
in weight management and prevention of type 2
diabetes. J Am Coll Nutr. 2016;35(8):717-731.
12 Thorning TK, HC B, JP B, et al. Whole dairy matrix
or single nutrients in assessment of health effects:
current evidence and knowldge gaps. Am J Clin
Nutr. 2017;105(5):1033-1045.
13 Feeney EL, Nugent AP, Mc Nulty B, Walton J, Flynn
A, Gibney ER. An overview of the contribution of
dairy and cheese intakes to nutrient intakes in the
Irish diet: results from the National Adult Nutrition
Survey. Br J Nutr. 2016;115(4):709-717.
14 Drouin-Chartier JP, Brassard D, Tessier-Grenier M,
et al. Systematic Review of the Association between
Dairy Product Consumption and Risk of Cardiovas-
cular-Related Clinical Outcomes. Adv Nutr.
2016;7(6):1026-1040.
15 Hostmark AT, Tomten SE. The Oslo health study:
cheese intake was negatively associated with the
metabolic syndrome. J Am Coll Nutr.
2011;30(3):182-190.
16 de Goede J, Geleijnse JM, Ding EL, Soeda-
mah-Muthu SS. Effect of cheese consumption on
blood lipids: a systematic review and meta-analysis
of randomized controlled trials. Nutr Rev.
2015;73(5):259-275.
17 Thorning TK, Raben, A., Tholstrup, T., Soedamah-
Muthu, S., Givens, I., Astrup, A. Milk and dairy
products: good or bad for human health? An
assessment of the totality of scientific evidence.
Food and Nutrition Research.60(32527).
18 Rosqvist F, Smedman A, Lindmark-Mansson H, et
al. Potential role of milk fat globule membrane in
modulating plasma lipoproteins, gene expression,
and cholesterol metabolism in humans: a rando-
mized study. Am J Clin Nutr. 2015;102(1):20-30.
19 Demmer E, Van Loan MD, Rivera N, et al. Addition
of a dairy fraction rich in milk fat globule mem-
brane to a high-saturated fat meal reduces the
postprandial insulinaemic and inflammatory
response in overweight and obese adults.
J Nutr Sci. 2016;5:e14.
20 Feeney EL, Barron R, Dible V, et al. Dairy matrix
effects: response to consumption of dairy fat
differs when eaten within the cheese matrix-a
randomized controlled trial. Am J Clin Nutr. 2018.
21 Drouin-Chartier JP, Cote JA, Labonte ME, et al.
Comprehensive Review of the Impact of Dairy
Foods and Dairy Fat on Cardiometabolic Risk.
Adv Nutr. 2016;7(6):1041-1051.
For more information contact Stephan Peters of the Dutch Dairy Association (peters@nzo.nl)
4