Genetic testing for risk of Heart Disease: fact or fiction? (23 Feb 2012)

Uploaded by UCLLHL on 28.02.2012

>> Professor Humphries: Well, thanks very much, Audrey.
And thanks for the opportunity to present some
of our work to you today.
So you've seen up here
on the slide the talk is divided into 4 parts.
I'm going to first of all describe to you about the causes
and the mechanisms of heart disease.
And then I imagine there's plenty of people in the audience
who are not from a genetic background,
so I'll tell you what a gene is or remind you on how it works.
And what is this SNP thing?
You'll find out by the end.
And can we use DNA tests to identify patients.
So if you like the topic of the talk is genetic testing
for risk of heart disease.
Are we talking fact or are we talking fiction?
A number of people who are in my group,
only some of them are mentioned up here, who have contributed
to the work and the work is funded almost entirely
over the last 25 years by the British Heart Foundation.
So I'll be giving them quite a few plugs
as we go through the talk.
So what are the common risk factors for heart disease?
Someone give me some ideas please.
>> Blood pressure.
>> Professor Humphries: Yep.
High blood pressure is very bad for you, yep.
>> Obesity.
>> Professor Humphries: Obesity.
>> Family history.
>> Professor Humphries: Family history.
Yes, that's, of course,
the important one we're going to talk about today.
>> Smoking.
>> Professor Humphries: Smoking.
Very bad. Terrible.
No one here smokes I'm sure.
Yeah. Anything else?
Yes. In the lay press there's bad cholesterol which is --
which from a clinical point of view is called low density
or LDL, LDL cholesterol.
And there's good cholesterol which is called high density
like -- high density like for rating so good
and bad, bad cholesterol.
I mean high levels of lipids is a problem.
Anything else?
Yeah. Lack of exercise.
You've missed the two big ones.
>> [ Inaudible ]
>> Professor Humphries: Yes and no.
That's not one of the -- what are the big ones?
>> [ Inaudible ]
>> Professor Humphries: Yeah.
That's on the list.
Yeah. You're right.
>> [ Inaudible ]
>> Professor Humphries: Diet.
Healthy diet is important.
But what am I?
Major risk factor.
And what am I as opposed to most of the students in the audience?
Age. So these are the list.
These are the main ones.
You got almost all of them.
Well done.
Now, some are modifiable and some are not.
Some have got a very strong genetic contribution
and some are mostly environmental.
Most of these risk factors have got bits of genes and bits
of environment in them.
Okay. So let's talk about what are the processes
of heart disease.
In some of the slides, you'll see it
as CHD an abbreviation coronary heart disease.
So it means that the arteries in the heart are diseased.
You may have seen a couple
of years ago this very challenging advert
by the British Heart Foundation.
For every cigarette we smoke makes fatty deposits stick
in our arteries.
That's a very nice graphic.
And that, of course, is one of the problems
and that's both genes and environment.
So here we've got an artery of maybe quite a young man,
young man or woman, opened up post mortem along its length.
And what you can see here is there are these little
yellow streaks.
They're called fatty streaks.
They've got fat in them.
You can see they're oriented in the direction of the blood flow.
They seem to be very much downstream of the branches.
That's a little blood vessel going off.
These are the very first signs of atherosclerosis.
You can find these in young people
in their teens or their twenties.
And what we hope is and what we know is
if you lower your cholesterol, you can get rid of these.
But they are the precursors of this.
This is an artery from maybe a 50 year old taken post mortem
and now cut across.
You can see interestingly this part of the artery wall is fine.
It looks nice and thin but what you've got here is a big
cholesterol laden plague, an atherosclerotic plague,
that's more -- that's blocking at least half of the lumen here.
That's an advanced lesion.
And this is what something looks really bad.
This is a genetic disease that we do a lot of work
on called familial hypercholesterolemia.
It's a 45 year old smoker.
The whole of the artery
from this individual is just completely stuffed
with this cholesterol.
Obviously bad news.
Here's again, cut-in section, an intact advanced lesion.
Again, part of the blood vessel looks fine;
here is the plaque here.
Here it's no longer cellular.
Cholesterol has killed the cells and they've died.
So there's what's called a necrotic core.
But there's lots of muscle cells, there's lots
of smooth muscle cells,
epithelial cells keeping this happy.
This one's not a problem yet.
But if you were open that up, you'd see this sort of thing,
a bit like a fried egg.
Some -- maybe some blood clotting has gone on here.
You can see it's sort of ulcerated, there's little holes.
It's leading to trouble.
It's not looking clever.
And this is what happens when one of these lesions ruptures.
You get a clot.
This, again, in section, this is maybe a platelet rich clot
covered with protein, the fibrin that's made the fibers
that forms the clot.
And if you would open that up, that's what you'd see.
So here you've got an atherosclerotic lesion,
a bit like this.
It's ruptured and this clot has formed on top
and blocked the blood flow.
Now you can see atherosclerosis in the arteries
in the heart using --
of a living person using a technique called angiography.
An x-ray of opaque dyes injected
into the bloodstream via a catheter.
And what you can see here --
and you can take a series of x-ray pictures.
You can see that some of these blood vessels are really nice.
The blood flow is going through very happily.
And others are looking really bad.
This lesion looks pretty terrible.
There's almost no blood getting through there at all.
And of course the problem is when that happens,
if the blockage is complete,
the clot completely prevents oxygen flowing,
then no oxygen will get through.
That tissue will die.
And that's what causes a heart attack.
So what you've got to think
of that heart disease is an ongoing process.
So we're born with very nice clean arteries here, and the 20s
or 30s, we get the fatty streaks leading to this sort of thing.
And this is clinically silent.
No one knows they've got this.
Maybe by the 50s or so, people might get chest pains
when they run for a bus.
In other words, there's enough blood and oxygen flowing
when they're sitting down,
not doing anything, there's no demand.
But when the heart needs to pump more,
there's not enough oxygen getting through
and you get chest pains.
A big sign.
Go to your doctor.
Try to see what's going on.
Then you get the plaque rupture, the clot forms,
and you get a heart attack.
And then this is the Heart Foundation campaign.
The main problem from people dying from heart disease
that we can fix is not calling 9-9-9 early enough.
Oh, I've got angina.
No, you're not, you're dying.
And so when you get this or when your husband
or someone else starts feeling this crushing chest pain
which is visualized by this very nice poster, call the ambulance,
take an aspirin, get to the hospital as quick as you can.
So understanding this gives us some clues
about the genes we should be thinking about when we're trying
to work out a genetic test.
The thing to remember is
that heart disease risk is multifactorial
as we already showed in an earlier slide.
So what you've got is, some people by luck,
have inherited very few risk raising genes,
risk raising alleles.
And others have inherited lots of cholesterol genes,
blood pressure genes, clotting genes, things like this.
But equally we inherit different environments.
Some of us choose to maintain a healthy environment
and others don't and by all these thesis
of things that we discussed.
But really heart -- the risk of early heart disease only occurs
when you've got a bit of both.
It's this overlap when you've got a number of high risk genes,
a number of high risk environments.
And the thing to remember is that nature loads the gun
and nurture pulls the trigger.
You have to have both to get an explosion
which in this case is the heart attack.
So the research that we're doing is try
to find the genes involved, the way they interact
with each other and with diet,
and we use this for patient benefit.
Okay. So if we're going to try to develop a genetic test,
what are the criteria we need to set out before we do this?
It's got to be predictive over
and above the established risk factors we talked about earlier.
It's clearly got to take environment interactions
into account.
It may be that a particular risk gene is only important
if you smoke.
Doesn't matter if you don't smoke.
It's only important if your -- high blood pressure.
So that's really quite important.
Clearly it's got to be accurate and not to be associated
with negative psychological impact.
And if I've got time at the end, I'll come into that question.
But why are these predictive tests going to be helpful?
Well, in the U.K., if an individual has a more
than 20 percent risk of developing heart disease
over the next ten years, NICHE, the National Institute
of Clinical Health and Excellence,
says you can give him a statin.
Who is currently taking statins in the audience?
At least a half dozen people.
Good. Excellent.
Well done.
And what are the statins doing?
Why are you doing it?
To lower the LDL cholesterol, to lower the bad cholesterol.
And they do it very well.
They lower LDL cholesterol and they lower heart disease.
But they cost money and clearly you want to give it
to the people who are going to benefit most.
So here's the typical U.K. punter.
You can see his age, his LDL cholesterol,
his HDL, blood pressure.
He's a smoker; he has a family history.
If we stick that into a risk algorithm
which the government says we should use,
his 10 year risk comes out at 21%.
So great. No problems.
The doctor doesn't have a problem giving him a statin.
Here's his colleague at work.
He maybe goes to the gym a couple times a week.
You can see he's not overweight.
His lipid profile's a bit better.
But he's got smoker, blood pressure, family history.
You put that into the algorithm, his 10 year risk is only 18%.
You tell him to stop smoking.
You're not really officially allowed to give him a statin.
Now it maybe that this man here has actually inherited lots
of the bad risk alleles,
his genetic makeup means his risk is much higher
than this individual so what we really want to do is try to work
out tests that would allow us to distinguish people who are
at this intermediate risk into those who are genetically high
or genetically low risk.
So how do we currently estimate an individual's risk?
There are a number of algorithms that are used very widely.
There's one called Pro Cam, one called F'Ham,
there's one called Q Risk, there's one called Score.
They all basically work the same way.
Just focus here on this F'Ham one.
So they stratify by a risk factor and give you points
for high blood pressure, whether you smoke,
whether your HDL is low, whether your LDL is low,
whether you've got diabetes.
And in this case, points don't mean prizes.
Points mean risk of heart disease.
So what you can do if you took a large group of men,
the distribution of their risk would look like this,
and their distribution of score would look like this,
and their risk would go up something like that.
So let's see how it works.
This is a study that we've done a lot of work on.
It's just over 3000 healthy middle-aged men recruited
from 9 general practices throughout the U.K. They were
all free of heart disease on entry.
And then in the first 10 years,
200 of these individuals have an event.
These are the baseline characteristics and if you look
at the event group compared to the event free,
the ones who went on to have a heart attack were a bit older,
a bit fatter.
Their blood pressure was a bit higher,
their cholesterol was a bit higher, more of them smoked.
So very much the sort of risk profile that you would expect.
Okay. So what percentage of the events
that actually occurred do the risk algorithms predict?
I'd like to be able to tell you that they pick
up 60,70,80,90%, they don't.
This, in cartoon form, is the risk score distribution in those
who didn't have an event.
This, in red, is the score of those who did.
If we set a cutoff here so that we pick up a false positive
of 5% of the men who don't have an event,
we correctly identify 14% of those who did
which means we've missed 86% of the events
that actually occurred.
It's a rotten test.
It's the best we've got.
And it's being used by doctors all over the place to decide
if you're going to get a statin or not.
In other words, these folks
up here get a statin and these don't.
The problem is most events occur
in people who've got average levels of risk,
sort of by definition.
So looks like there's plenty of room to improve on this
if we can find some useful individual genetic information.
Okay. That's the heart disease bit.
Let's go on to the genes bit.
Now each chromosome is the storage unit
for thousands of genes.
We now know we've got about 25,000 gene pairs
and they make us who we are.
You, they come in 22 matching pairs,
one from mom and one from dad.
Here they are spread out in what we call a cario type.
And this individual has 2 copies of the X chromosome so this is
from a female individual.
Male has an X and an Y. Genes are made
up of 4 chemical codes known as DNA of course.
Adenine, Cytosine, Guanine,
and Thymine in the classic what someone crypt, the double helix.
A gene can be many thousands of these bases long.
And the information is a code.
It contains what you need to make a whole human being.
The cell reads this and is told what to make,
when to make it, and how much to make.
When a change occurs in the DNA sequence,
the genetic instructions are no longer correct.
It's like having a typo in the manual.
One base is changed to another.
For example, CAT becomes TAT.
It means the cell makes a part that doesn't work,
sometimes bases get missed out, or are copied too many times
and the cell makes too few or too many parts.
Imagine the manual tells how to make an umbrella.
So here's the gene and the gene says, when it rains,
I'm going to make some proteins that will protect the cell.
Let's make some umbrellas.
Okay. So here's a gene that's got a mutation,
a single based change, and it could be that instead
of making lots of umbrellas, it only makes one.
Obviously not protective enough.
Or it could make an umbrella that doesn't work quite right.
It opens and closes again.
Okay. Or it could be that it makes the protein
at the wrong time.
You can see how using this simple analogy,
that mutations either mean
that the gene doesn't make enough protein,
doesn't work right -- well, or is not made
in the right time or the right place.
Genes make proteins that are important for health.
Mutations stop them working.
There are 3 different types of genetic disease.
Chromosomal disease.
You can think of Down's Syndrome
where an individual has 3 entire copies of one
of the small chromosomes, chromosome 21.
Single gene defects, like Huntington's
or cystic fibrosis or haemophilia.
But what we're interested in, of course,
is multifactorial diseases where there's lot
of different genes involved.
Hypertension, obesity, diabetes, and heart disease.
Again, I just make this point, the clinical consequence
of all genetic disorders is modified by environment.
You would think, of course, that hair color is genetic,
but many of us in the room are modifying our gene type
by using some sort of chemicals.
And clearly, tall parents have tall children.
But if those children are not fed well, they will be shorter.
So all of these different traits,
there's both genes and environment.
So how can we look at a person's genes?
So what we need -- what we use is something called the
polymerase chain reaction and I'll just spend a couple
of minutes telling you about this because it's so clever.
What you need is the gene sequence that you're interested
in and from that you design what are called primers.
They're homologous to a short region of the DNA, the sequence
of that gene, 20 or so bases.
You'll then need an enzyme that will copy DNA,
and it's useful if it's thermastable.
I'll explain why then.
You need the denaturise buffer energy
and you need a machine called a thermocycler.
Because what you're going to do first
of all is take the DNA here and denaturate it, so you raise it
to 94 degrees, almost boiling.
And the 2 strands fall apart.
You then lower the temperature to about 54.
And what happens then is these primers come along
and find the place in the genome that they are equivalent to.
And they form what's [inaudible] as double strand.
And then this enzyme takes these
and copies them using all the nucleotides.
So what happens is you start off with template DNA with 2 copies
of the gene, you then have 4, 8, 16, 32 copies.
You keep on doing this 35 times and you end
up with 68 billion copies of just the fragment of DNA
that you're interested in.
So you can see why just with a nuclear chain reaction,
this is a polymerase chain reaction.
So then you've got loads of DNA, absolutely bucket loads
and you can sequence it, you can digest it, you can clone it.
You can do all sorts of things.
Now many of you know that the --
there's been a lot of fuss in the press
about the human genome sequence.
This is the press conference -- Clinton --
this is Francis Collins who's the director
of National Institute of Health
in the U.S. This is a venture capitalist called Craig Venter.
And they were announcing the completion
of the first draft of the human genome.
10 years work, cost $2.7 billion,
and you can see the quote there.
And our glorious leader at the time,
never one for missing an opportunity for hyperbole said,
"It's a revolution in medical science.
A breakthrough, it opens the way for massive advances
in the treatment of cancer and hereditary disease
and that's only the beginning."
In this instance, he was right.
You can believe him.
It has revolutionized our research.
Back in the days in the 70s, I won't go into the details,
but we could work out one individual --
we could work out a single based chain
in 20 individuals, it took us a week.
At the Steam Age in the 80s, we could do maybe 200
to 2000 individuals in a week.
In the 1990s, we were up to 3000 to 5000, and for Star Trek fans,
we are now in the warp drive where we can do 100,000 a day.
I mean we've got databases, tons of information.
There is a bioinformatic gold mine out there.
We can now capture DNA from the whole genome in one experiment.
Okay. So what's a SNP?
SNP stands for Single Nucleotide Polymorphism.
It means one base in the DNA is changed.
Polymorphism is from Greek.
It means many shapes.
So it means the difference in DNA sequence found commonly
in the population, but we define it as more than 1% of people.
So here's the sequence, it's there and then in the other --
in another individual, there's an A instead of a T. Clearly
if it occurs in the coding region of the gene,
it's going to affect a number of umbrellas.
Most of them occur outside the gene and are of unknown
or of no effect but we can still use them as genetic markers.
And they're all given a unique ID; it's called a RS number.
So SNP would be called RS #237...
And we're all getting very used to recognizing 7 or 8 numbers
and keeping them in our heads.
How do we choose the genes to study?
Well, clearly, because of what we know about the pathology,
we can think of genes involved in lipid metabolism, clotting
and hypertension, and we can start looking at these.
These are very easy studies to set up.
But most of them were underpowered.
There were too many -- because we were limited
by the technology.
We could look at -- it took us months to look
at 200 patients and 200 controls.
And many of the results didn't stand up to the test of time.
What we need is a hypothesis free approach
and with the technology, we can now do this.
So what we do is we take 2000 cases and --
with heart disease, 2000 controls.
And companies now make gene chip devices which have 300,000
or a million of these SNPs spread throughout the
entire genome.
We can cover the entire genome with one little chip like that.
So what we do is for each of these SNPs we look to see
if the frequency is different in the cases and the controls.
If the frequency of this SNP, this allele,
is base change is higher in the cases than the controls,
significantly, then we say Ah, this gene is marking --
this SNP is marking a gene that increases your risk
of developing heart disease.
Now, of course, if you've got a million SNPs,
you have done a lot of contrasts and many
of these will be statistically significantly different
by chance alone.
So what you have to do is set a very low P value
as your threshold.
And what you have to do, more importantly,
is to replicate it in the second study.
That's what people have been beavering away doing
for many years now.
And if you look on this website, in the middle of last year,
there were more than 237 traits
that had been mapped to the human genome.
And they're for all sorts of things.
Some of them are very interesting and some
of them are very weird and wonderful.
It's amazing what you can do with genetics.
You can find the gene causing restless leg syndrome.
But some of these, of course, are very important genes
and we want to know about them.
The way that the data is presented is
in what's called a Manhattan plot.
You present it along the X chromosome here
as the chromosomes and up here, in fact, it's the log --
minus the log with the P value so what you want is to find lots
of different P values that are statistically significant being
associated with your trait.
In 2007, there were 3 GWA studies,
genome wide association studies
for heart disease published at once.
All 3 of them looked like this.
There was one hit and one hit only on chromosome 9.
It's not really a Manhattan skyline,
it's more of a Oxford skyline.
Now, replicating many independent data sets,
a major breakthrough.
What the heck is this gene?
It wasn't anything that we've ever thought of.
It occurs in a gene desert.
The nearest gene or codes or proteins that we know
about are 58,000 bases away, miles away.
The common SNPs are, however, strongly associated with risk
and compared to the AA, to people who've got 2 alleles,
if you've got one gene allele, your risk is 30% higher.
Two G alleles, 60% higher.
But interestingly, these GG people weren't --
the fact that they didn't have high cholesterol,
they didn't have high blood pressure,
we still don't really know what the mechanism of this is.
But it's certainly -- I'm going to skip this
because we're running out of time.
But it certainly looks like it's going to be of clinical utility
and you can actually buy it now.
You can actually go online at 23andMe.
If you send them $399 they will test you for this single SNP
and that's its RS number.
If you go to deCODE, they will also do it
and they will just send a SNP.
Now is it worth $399?
Might be. Let's see.
So we looked -- went back
to Northwood Park [assumed spelling]
and found exactly the same results compared
to the AA group.
Let's see if you had 1 G allele, your risk was 38% higher;
2 G alleles, it was about 60% higher.
This is adjusted for age, cholesterol, triglycerides, BMI.
It certainly looks like it's working well.
So the question is, is it going to add over
and above the Framingham score?
The one I showed you earlier.
So we went back to [inaudible] again and we looked at it
and we got a 3% improvement.
That curve was shifted slightly to the right
but they still overlapped.
It wasn't significantly better.
The two -- when we did tests,
that wasn't better than the other.
We're getting there.
But one single SNP is not enough.
Just as you wouldn't predict an individual's risk
by simply measuring cholesterol.
You'd put them all together.
Simply looking at one gene won't do it.
It's because heart disease is multifactorial.
You've got to have lots of different genes.
So here's the chromosome 9 SNP and we're very quickly able
to find by looking at data, a whole bunch of others.
And then earlier, about a year ago now,
a whole bunch of others were found.
We are now doing GWAs by combining data
in 100,000 individuals and it gives you the power
to find very small effects.
But we now have over 50 heart disease risk genes.
The alleles are common
but they're all having this modest effect.
They're increasing your risk by 10%, 20%,
the chromosome 9 one is the biggest one.
That's why we're able to find it first.
So what are we going to do?
Well, what we want to do is put them all together.
We would need to combine them in what we would call a gene score.
So we put together a number of genes involved
in lipid metabolism, clotting, and endothelial function.
So 13 SNPs in these genes.
7 of these, GWAs SNPs, the earlier ones,
and we look at the frequency distribution.
Now, we've simply used an additive model.
We say at each SNP if you've got no risk alleles,
we'll score you 0.
We'll score you 1 if you're a carrier.
And 2 if you've got 2.
It's very simple.
It assumes equal and additive effect.
We can do much better than that, but this is a simple first-go.
And we went back to the Northwood Park map.
This is the distribution in Northwood Park.
This is the distribution in U.K. Caucasian individuals.
Median number is 15.
What you can see there's a group of people
that have got 12 or fewer.
We imagine they should be protected.
A group who've got 18 or more,
our guess is they're going to have higher risk.
How does it actually work out?
This is how it looks.
It looks just what we wanted.
So compared to this group in the middle here,
we've got average risk, we can identify that these people
who are in the lower, the lower three tenths
of the risk, they're protected.
The average risk is about half of this group.
And this group here, in the top three tenths,
their risk is almost 2, 2 1/2.
Just to put that in context, a risk of 2 is about what smoking,
lifetime risk of smoking is.
If you smoke, it roughly doubles your risk.
So what we can do with this score is identify men,
maybe 20% of men who have a genetic risk as high
as if they were smoking.
We all know doctors try to stop you from smoking
so this really just puts this into context.
So can we go home?
Are we all done?
No. The heritability of heart disease are 45 to 50%
so roughly half the causes of heart disease are genetic
and roughly half of them are environmental
which is pretty much what we'd expect.
So here's a pie chart, half environment.
We have maybe found 10% of the genetic cause of heart disease
with the SNPs we're looking at.
The problem we've got is we've only found this bit.
We've still got to find all that lot.
There's a number of possible reasons why,
I won't go through them in great detail,
but we've still got a lot to learn.
Can we find more by sequencing an individual's genome?
Now you may remember the first genome cost how much was it?
$2.7 billion -- oh, it's got a lot cheaper.
You can now sequence a whole genome
for an individual for $5000.
You can sequence just the part that codes
for proteins for about $1000.
You can just take these 40 genes that I've been talking about
and you could probably get them sequenced for $500.
This sounds like good value.
This sounds like it would really be useful.
Now the problem is the bioinformatics.
We are drowning in data and we don't know what to do with it.
One of the projects I'm working with is funded
by the Wellcome Trust and it's called The U.K. 10,000 Genomes
Project and they're doing just that.
They're finding the genome,
they're sequencing the entire genome
of about 6000 individuals.
And just the exomes, the part that codes
for proteins, in about 4000.
And we're providing samples from people with this disease, FH,
this inherited high cholesterol disease.
Here's the results for the first 22 of these samples.
Each part -- each coding part is covered more than 74 times,
we get 30 million bits of DNA information.
This is what's startling.
So for every patient, and for everybody in this room as well,
we got about 42,000 SNPs.
And about 1100 of them have never been seen before.
They aren't in any of the databases.
They're completely novel.
And we use could bioinformatics and predict that about 350
of them are likely to affect the protein and 10
of them are likely to make the protein shorter
than it should be, to truncate it.
So we could guess that each individual,
everyone in the room, has got about 350 variants
that are going to affect the protein.
This is why you shouldn't marry your cousin
because it's very likely that there will be --
any children will have 2 copies of this, one of these genes,
and that could really affect disease.
So we're in the needle in the haystack business
and it's a big challenge.
So yes, we can do it technically;
we can't actually analyze the data yet.
So these are the criteria that I set out in the beginning.
Criteria for useful heart disease tests.
I'm quite certain with several genes, we're going to get there.
We certainly are not taking interactions into account.
I told you the model is additive.
Work to do.
I hope I've persuaded you that we do have accurate
and reproducible risk estimates.
But all the studies almost so far have been in white,
Caucasian, middle class aged men.
We don't know anything about women.
And we certainly don't know anything
about different ethnic groups.
And then what we don't have -- maybe I'll have to skip this --
I've got 5 minutes; okay.
Let me just explain to you about the psychological impact.
What's the concern?
The concern is that is if about -- what we call genetic fatalism
or forced reassurance.
There are, of course, issues of confidentiality
but let's not worry about those right now.
This is a cartoon illustrating this
and here's Freda and his Fanny.
They've just been given their genetic test results.
They've gone home.
They've lit up cigarettes; they're drinking martinis.
And Freda says, "I haven't got the genes so it's fine for me
to smoke" and Fanny says, "I've got the genes.
So I guess I'm doomed anyway."
So all out technology is to no avail.
And there was a lot of concern about both of these aspects.
And it really seems now we've got more data,
that these are unjustified.
Studies actually suggest
that DNA risk information doesn't increase fatalism
and it actually may motivate people to change behavior,
to actually take their medication and things like this.
So I think this is an issue
that we probably don't need to worry about yet.
So heart disease risk test, I think it is possible.
We need to use several genes.
It's got to be based on good data and currently,
there are some gaps in that.
Don't bother buying over the internet because without use --
without having the information
about your other risk factors, it's not valuable.
It will help us to risk stratify those 2 men that I talked about
and the idea would be that before you come for a clinic,
we would ask you to send us a saliva through the post
and we could then test it.
We could test 20 SNPs, 40 SNPs, a thousand SNPs.
And we would then present you your risk information,
your 10 year risk information,
based both on your classical risk factors
and your genetic risk factors, and then we would work with you
to try to reduce your overall risk.
We still need to compile how to best present the data
for maximum understanding.
So yes, heart disease DNA testing is ready now.
And finally, the take home message is
that small differences in your genes make big differences
to how you look, but also to your health.
And, Audrey, I'll stop there.
Thank you.
[ Applause ]
>> Audrey: Thank you very much.
That was a tremendous talk.
We have 5 minutes for questions.
Would anybody like to raise a -- yeah.
>> Professor Humphries: Microphone's coming.
Just wait.
>> Hi. Thanks for a really interesting talk.
I was just wondering do you think doctors are well equipped
at the moment to give that information?
>> Professor Humphries: Not in the slightest I'm afraid.
I know of a number of colleagues in the medical school,
people are -- having the tests.
They're getting 23andMe reams of information.
They're going to their cardiologists and saying,
look I've got this and that and the other.
I've got this chromosome 9 SNP.
And it's a bit -- one of my jobs is to educate doctors.
We need to be doing continued professional development
so they know about this sort of thing.
And as always, it's about the balance
between hype, fact and fiction.
In other words, to what extent is this information
clinically useful?
I mean I think it is, but 23andMe
and deCODE don't present it to you in a way
that you can understand or in a way that the doctor can.
I can but that's a bit different, isn't it?
So these are the things that we need to be doing.
>> Audrey: We have 2 final questions.
>> Professor Humphries: One right here.
>> Excuse my voice.
It's gone a bit.
I'm afraid I missed the first few minutes of your lecture
so you may have covered this.
This is anecdotal but all my family, for as far back
as medical records go, have suffered from only one thing
which is high cholesterol.
No high blood pressure, no cancer, anything.
And it's always scared me out of my wits that that might transfer
over to cancer, but nobody's ever had cancer.
We don't smoke; we don't eat animals,
and we don't drink, so --
>> Professor Humphries: Are you taking a statin or --
>> I don't want to take a statin because I'm -- somebody I knew,
one of teachers from I think it was University of --
College School, a math teacher took statins for a while
and then he threw himself under a train.
It struck me that I stood more chance away
from the railway lines.
>> Professor Humphries: One of the things that I just touched
on very briefly is
that inherited disease called familial hypercholesterolemia.
High cholesterol runs through families.
I don't know.
It sounds like that's what's happening in your family.
It's a single -- it's a mutation in a single gene.
It knocks out this gene
and it prevents you removing cholesterol from the blood.
So you have high cholesterol and early atherosclerosis.
People who've got this disease have maybe ten-fold higher risk
of early heart disease.
And if we find them and give them a statin, that risk reduces
to the same as the general population.
We have shown that these individuals, people with FH,
don't have a high risk of cancer and when they're treated
with statin, they actually are --
I guess you could say paradoxically,
they're actually carrying less cancer
than the general population because we tell them not
to smoke and they don't smoke.
So they're not dying of lung cancer or other cancers.
Statins are in general remarkably safe.
We're very lucky to have such a powerful and such a safe drug.
There are some side effects.
There are some muscle pains.
But my recommendation to you is that you do get a --
you ask a GP for a referral to a lipid clinic.
We've got a very good lipid clinic at UCL
and get a proper diagnosis.
And then you really should consider taking a statin.
>> I feel confident --
>> Professor Humphries: You know what?
I think we better stop.
>> It's just that --
>> Professor Humphries: I'm happy to talk
to you afterwards but --
>> Audrey: I think that's actually all we have time
for today.
Quite a day.
Join me in thanking Professor Humphries for a great lecture.
[ Clapping ]