José M. Ordovás: Personalised nutrition

I think we are not going to be very a different
size of the fence here in terms of what Tim said, probably our differences will be a slightly
in the shade, right? Of a color that we are going to use in this
personalization. Well, for us the future of food is personal
and this is reflected in the article that is in the supplement and the co-authors were
John Mathers, E Shyong Tai, and Lynette Ferguson and I have the pleasure of having two of the
co-authors, John and E Shyong here with us today. In terms of the conflicts and disclosures,
I earn much more modest than Tim, I am just in the scientifical advisory board of habit
and unlike your royalties, I don’t get any economic reward for that. And other than that, my own bias is that I
have been in this business of personalization for the last 30 years or so and I want these
to become a something useful for the society. All right, so one of the things that we have
for a Monday … Uncertainties related to nutrition, one that associates with personalized
nutrition is the definition. There are many terms used for a personalized
nutrition, nutrigenetics, nutrigenomics, precision nutrition and so on, and that produces at
least among the non scientists community, some confusion. We try to define it very simple, we define
it as an approach that uses information of an individual characteristics and as you will
see it, could be the microbiome, it could be genetics, it could be other factors, and
based on those individual characteristics, to develop targeted nutritional advice products
and services. What is the ultimate goal of this? Providing individuals with tools to achieve
a lasting dietary behavior change that is beneficial for health. Now, do we need the personalization? Yesterday we saw many different point of views
about what kind of nutrition is good or bad, and our approach is that probably we need
personalization. We are differently, different externally,
we are different in terms of micro or microbiome as Tim was showing before. And the microbiome is not really internal
to us, this is still external and then we are difference internally, in terms of our
metabolism and this is something that all of you have experienced in terms of the differential
response to the same intervention. These are very old experiments that we did,
we conducted in our MRU many years ago this was published back in the 90s and we had this
variability that you observe here, reproduced for practically any phenotype that we look
at here, the percent change was on LDL cholesterol, the intervention was from going from the traditional
American diet, high in fat and high in cholesterol, to a low fat, low diet … Low cholesterol
diet that was the NCP to recommendations. And well, these people we could say that they
were not cheating because they were in our MRU for several weeks in each case and you
see … Oops, excuse me. And you can see how for some people this healthy
diet didn’t have any effect, or even it went in the opposite direction and other people
were highly responsive. We could do the same thing as I mention for
a response in terms of BMI, a response to drugs, as we have also done in our lab and
we see the same variability. At that point, back in the 80s and early 90s,
we could start finally taking a close look or genome or at least the genetic variance,
and we started looking to see how a genetic variation could be able to identify the people
that were not responsive and the people that were responsive. And at that point we didn’t have the luxury
that we have nowadays of access to the whole genome, but we have access to some interesting
genes. For example APOA, APOA4, both of them involved
in lipoprotein metabolism and what we could identify from our early studies is that for
example, when it comes to people that were highly responsive to these changes in saturated
fat of fat and cholesterol, APOA4 subjects were in this part of the response, so they
were hyper responders. Where people that were low responders, they
could be identified as carriers of one of the common polymorphisms So, but again that was the beginning and the
field has moved, fortunately a quite a bit since those early ages of personalized nutrition. The current strategies a have had limited
to success and we have seen that we have discussed that yesterday. From the implementation of the different Italian
dietary guidelines in the U.S., we have seen that they have not accomplished the objectives
that they were supposed to accomplish in terms of curbing the increase of age-related diseases,
like for example diabetes, as shown here. This is the example in the States and for
example in the UK you have something similar and this is data from the Department of Environmental
Food and Rural Affairs that it’s showed that through the years despite all these recommendations
that we have seen, global recommendations in terms of fat, specifically saturated fat,
there have not been changes and the fiber that Dean mentioned before, also there has
not been changes over time, because you will expect one of them, at least for saturated
fat going down, and for fiber going up, and that’s not the case. So, global recommendations are, although they
are good, they are not making the impact that they should be making in terms of improving
the health or the dietary habits, and the health of the population. So, in terms of we have seen about the genetic
body variability in response, and we have seen the lack of effect of impact of the global
recommendations, and the question is, rather than the dietary guidelines for everybody,
can personalization of the recommendations look better. And then there are, there have been many instances,
and while we have been hearing tell, one size does not fit all, therefore we should be moving
into that personalized nutrition. The personalized nutrition is at the convergence
of four major trends that we have now in the society. One is the customized experience and we have
seen and we can see that for example in the case of, an example from the Chipotle or The
Counter, in which you select your own burritos and how specifically do you want to, what
do you want to have on them or what kind of burgers you, what do you want in your burgers
and so on. So, it’s personalization, it’s customization
and the same applies to, for example the music that we listen to Spotify is not like turning
the radio and just basically waiting or listening to what is there. All the movies that we want to see. S So this is part of what the society is experiencing,
customization. It’s true that the society and it’s good,
is looking for … Is recognizing continuously, the need and the improvements that you can
get with healthier aging. So there is a growing a trend for eat healthy,
organic foods and so on. Also, the personalized medicine. We can consider the personalized nutrition
as part of personalized medicine, that has been or is being facilitated by the advances
in genomic science. And also the quantified self, now people,
anything way that they want customized experience, they want to know more and they can do it
through the tools that we have available to us, about what is happening in our inner environment,
in our bodies all the time. So, in addition to those trends or to facilitate
those trends, we have for scientific advancements. One is the ability of diagnostics, the monitoring
tools, the digital capabilities that we have nowadays, the nutrigenomics or precision nutrition,
and the ability to modify the foods. One of the steps or one of the first step
in terms of achieving that personalization, is that we have to assess ourselves. We need that individual or assessment, or
diagnosis, and for that we need the compilation of all the things that you can see here. We need to collect traditional information,
that is like a gender, medical history and health, and so on. And then we have to take advantage of the
developments that we’re experiencing in genetics, the microbiome that it has been so nicely
illustrated by Tim, but also things that we know are important, but they are not part
of the mainstream yet, like epigenetics. Then, we need to have these continuous follow-up
through biomarkers, and in this case we can use the traditional biomarkers, like cholesterol
or glucose, presenting some of the previous studies, but now we have to go a little longer,
we can do deeper than that, by using metabolomics, which is part of those -omics that we are
including here. And then, to look at our environment our exposure,
we have to be … We can measure with great precision the genetic aspect with the epigenetics
or so, the microbiome but one of the problems that we have, and it has been illustrated
during the meeting, is that we cannot quantify it as precisely as we want, things are important
obviously for what we are discussing here. Our lifestyle, what we eat, our physical activity
or behavior and so on. So, in terms of the next few slides, I am
going to a present three different aspects of the genetics, as they relate to personalization. And first, I am going to do it with an example
that was published a couple of years ago in the New England Journal of Medicine. You think that the strategy that John was
mentioning before, that is used by the geneticists, which is just great, create large consortium
and get together in order to extract the information from individual studies. In this case, this study involve these populations
that you see here, they are to study the women’s genome health study and the Malmo diet and
cancer study. The outcome was to look at the 10 year insulin
coronary events, and the research question is the one that you see here, to what extent
can genetic risk of coronary artery disease be offset by healthy lifestyle. So here in terms of a lifestyle, it was something
very simple. A smoking non-smoking and to have a healthy
BMI, at least not obese. Physical activity at least once a week and
a healthy dietary pattern that was very simple. It was just increased fruits, nuts, vegetable,
dairy, fish, daily fish and whole grains and avoid refined grains, red and processed meats,
laboratories and so on. And then, in terms of the genetics, what it
was done in this case, is to have a genetic risk score using 50 SNPs that in previous
genome-wide association studies, they were significant in terms of the phenotypes of
interest. And they, as you know, it’s very simple to
calculate the individual participant’s course. The results are shown here, and what we have
is first, the importance of genetics, because according to that a genetic risk score, you
can classify people in low, intermediate and high genetic risk and this is for heart, this
is for the women genome, how to study, and this is for the Malmo study and you see quite
consistent findings, in terms of … Oops. People with with higher genetic risk score,
they have more events for coronary disease than those that were in the low category,
but what is nice is that when you sub-classify each one of these genetic groups based on
their lifestyle, good, intermediate and bad you can see a ladder indicating that both
things are important, the genetics and the lifestyle. And the
the conclusion of this study is that among participants with high genetic risk, a favorable
lifestyle was associated with nearly 50% lower relative risk of coronary disease, artery
disease that was those that didn’t have the proper lifestyle. So that was a convincing evidence of the interaction
between genetics and lifestyle. Now another study in this case restricted
to one single polymorphism in the tcf7l2 gene, which was one of the first genes identified
with diabetes and later on it has been associated with other risk phenotypes, including for
example stroke, as illustrated in this case, that uses the Predimed that yesterday didn’t
receive very favorable critiques by John. Well, in this case randomization does not
apply, because as an important tool, because here the randomization comes from the genetics. And what we are seeing here in the control
group, that was the group that did, they had a low-fat diet, you see the expression of
the risk associated with this, the tcf7l2 a SNP polymorphism, TT higher risk, CT intermediate
risk, and CC higher risk, in this case as I mentioned for a stroke. Now, for those that were in the intervention
a group with Mediterranean diet, what you can see very nicely is that the Mediterranean
diet benefits in general to everybody, but more specifically as we saw in the previous
study, in the New England Journal of Medicine, the people with a higher genetic risk score
were benefiting more of the a proper lifestyle. Here we are seeing the same situation, it’s
the people that are at higher risk, genetic risk are the ones that benefit the most out
of the, in this case the personal proper diet for this group, which is the Mediterranean
diet. So, during the life of the study that was
about six years, what you have is that you could prevent or you prevented in fact, about
52 events of a stroke, because these people were getting the right diet for their tcf7l2
polymorphism. When you extrapolate these findings to the
population at large, in this case Spain because that’s where the Predimed was conducted, there
are about 90,000 new strokes per year, and according to these findings, you could prevent
about nine thousand of them, yes by providing the right dietary recommendation for the right
genotype. In yes, to conclude, let’s move into something
more practical, which was the food for me a study. And this study was a proof of principle of
how personalized nutrition could be helpful in order to change the dietary behaviors of
the individual. Those were their research questions, is personalized
nutrition advice more effective than general healthy eating guidelines, is phenotype or
genotype information more effective than dietary-based advice alone, and is internet a successful
delivery a method. These were the different countries involved,
and young mothers were leading this study, and the main question was, they were randomized
to four arms and the question was, those were just getting the global advice or doing the
same of work in terms of changing their dietary habits, than those that were receiving some,
any kind of personal recommendation, diet, based on diet, based on diet and phenotype,
or based on diet phenotype and genomics. And in terms of the research question that
they presented before, we can say that personalized nutrition advice was in fact more effective
than conventional one size fits all approach. Now, what this study did not confirm is that
going further and looking into the phenotype and genotype of these individuals or doing
better than just any type of personalization. And you can see it is here, in terms of, for
example how people, who were changing their behavior in terms of eating red meat or adhering
to a healthy eating index, and in this case, is people that were getting just the global
advice, they change for the best their habits in terms of red meat, but the people that
were receiving any type of personalization, they do had a statistically significant improvement,
compared to the controls. And the same is true for the healthy eating
index. However, as you can see here, whether they
did it based on diet, or based on diet and phenotype and genotype, didn’t make any difference. And for people that receive a specific advice,
for example in terms of saturated fat or folate intake, you can see that the improvement was
even better. So, the take-home messages is, is that personalized
nutrition works, at least in these studies there was no added benefit to a phenotypic
or genetic information, and internet-based delivery is effective. So let me finish with, what is the situation
now for example, and this is something that we’ll probably discuss later about the offering
that we have nowadays, by different companies, right? This is the situation that we have nowadays,
based on the statistical analysis of the studies that we do, whether observational or clinical
trials, what we can observe is that there are statistically significant difference for
a specific genotypes in terms of diet A and diet B, however when we look under these means,
we see that for some people this recommendation is less than a half really a benefit, because
these people are not benefiting of say, of the the diet … These people here, that are
supposed to benefit from the diet B, and in reality don’t do it. So, what we want to do is yes, to go to true
personalization, which means that we can classify people in one group or another, based on their
specific genome, on the epigenome, their microbiota, their metabolome, and then yes, at this point
the conclusion is that diet B is better than diet A for the individuals that we classify
in that group. All right, so in terms of the implementation
for more effective personalized nutrition, this is the model that we propose. We need those participant characteristics
that we have been discussing in the previous slide, we need to identify the barriers and
the facilitators of the change, we need to know the aspiration of individuals, in one
case could be a weight loss, in other cases more fitness or decrease risk of diabetes,
and then implement that personalized nutrition, based on that self monitoring that we are
able to do nowadays, then we can feed back into this information here and keep tuning
up the personalized nutrition until we reach the goal that we want, which is the healthy
well-being and with this, I think we can move to the discussion. Thank you very much.