Wednesday, November 10, 2010

Consumer Genetic Testing for heart attack risk? Worthless!

Here are the top ten reasons why in its current state, direct to consumer or otherwise, genomic testing for cardiovascular disease risk is dead in the water

1. Family History Risk paints a far better picture and IT IS FREE

2. Reynolds and Framingham risk paint a more accurate picture

3. An independent panel has reviewed 58 variants, 29 genes, and gave the thumbs down.

4. The highest increased risk from any of these tests is 30%, Fam Hx can be as high as 500%

5. Kif6 was just shot down as a useful marker.

6. Clinical Utility has not been evaluated in ANY of these tests.

7. Spit Parties don't lower cholesterol

8. The FDA is hunting down these type of crazy claims!

9 . Topol's heart attack gene didn't pan out, why would these?

10. A recent 23 gene panel failed to make the grade as well.

Let me be crystal clear.

I am glad that the number one reason for ordering a DTCG test was curiosity and not true medical concern in the "early adopters"

But I am concerned that may not be the case for the next wave. I am concerned they will take these genetic tea leaves and use them.

The problem, most of these tests are disproven or will be in the next couple of years.

Loose associations with small increased risks sounds a lot like fortune telling or phrenology. Or hell, even birth order....

Someday we will have good predictive models, 10-15 years from now. But NOT Now! Do you hear that VC country, SV, NYC, Hedgies?

10 year exit strategy. Not 2 not 8. So stop hyping this bull$h!t and go invest in Gold or Commodities or something for the love of god!

The Sherpa Says: Did you hear the one about the research geneticist? He keeps telling his wife how great their sex life WILL BE! Someday we will have this tool, let's try not to burn out and cynicize the public yet.....HT Francis Collins


Keith Grimaldi said...

Of course they ask the wrong questions again:

"However, the estimated additional benefit from adding genomic markers to traditional risk factors was found to be negligible"

" Traditional risk factors such as those used in the Framingham Risk Scores have an advantage in clinical screening and risk assessment strategies because they measure the actual targets for therapy (e.g., lipid levels and blood pressure). To add value, genomic testing should lead to better outcomes than those achievable by assessment and treatment of traditional risk factors alone."

Why is this needed to add value? If I have high blood pressure I expect my doctor to sort it out, not to ignore it and do a genetic test instead. Where is it claimed that genetics is supposed to replace classical risk factors?

Framingham is more accurate? It damn well should be, it is measuring things that have already gone wrong. As in your previous post, the value in genetics will be in keeping the Framingham score low.

Family history is better? Maybe, maybe not, I don’t think that has been formally assessed (the subject of your previous post is not that), also there are no RCTs that actually prove clinical benefit for family history (yet this is apparently required for genetics). This point though is irrelevant. Genetics should not, yet, replace family history, it should be complementary. As you know very well a proper, accurate family history takes time and effort – it’s not always done properly and anyway is not possible for all.

You might be right though to be worried about the next wave. If it really happens that companies will start selling panels of variants for use by doctors in the clinical setting to assess risk then there could be problems with overblown claims and proper misuse of the genetics. Actually it is happening – perhaps there should be more focus on that area.

Red Herring said...


Steven Murphy MD said...

Here's the skinny

1. Genetic tests add little to traditional risk models when USED IN CONJUNCTION. That has been shown for 9p21.3, TCF7L2, and a NEW 23 gene multiplex test.

2. Family history with these tests has not been fully evaluated.

But the comparison is clear. Regular medicine for predicting risks vs DTCG spit kits = DTCG as Loser

I think you are on to something though. Why would any doctor use these tools when they add nothing?

Why would a person use these tools if they can go to their doctor and get the other tools?

Oh, wait, how about if you are trying to completely replace the doctor?

Well, then why not just use the regular data points?

Oh, you mean I have to get blood drawn and my bp checked???.....God Forbid!

Get it? The only people that have an interest in using these tools are looking to disintermediate healthcare from people who have the training to provide healthcare.


Do they care about the quality of their tools? I don't see 23andMe going around and evaluating if their tools compare to standard care for risk prediction.....

Get it yet Keith?

PhDs are not clinicians

Businessmen and women with a B.S. in biology from Yale are not clinicians

Clinicians are clinicians

So why use a rusty blade for surgery?

These tests add one thing and one thing alone


Keith Grimaldi said...

Steve - I suppose I should say read my first comment again. The potential value of genetics is not to add them to traditional risk factors, or to replace them, but to use then BEFORE trad factors come into play. That is where the studies should be.

Most SNPs will actually be in or link to genes that code proteins that affect the TRF so it's most likely that the majority of SNPs will not add to TRF because they are part of the same mechanism.

“But the comparison is clear. Regular medicine for predicting risks vs DTCG spit kits = DTCG as Loser”

Really? Maybe, but only when the risk factors are present. Here is a 25 year old healthy person, no family history to be worried about anything in particular. Normal BMI, fat mass, lipids, blood pressure, glucose, insulin HbA1C, and so on… Will regular medicine and TRF testing still win? No, it can’t. I’m not saying here that genetics will definitely win, but it’s certainly favourite, at least it has the possibility of scoring where regular medicine does not.

But it’s not genetics vs. regular healthcare, where you see a battle there is not really one - it’s when and where to use genetics in healthcare.

By the way, I’m not talking only about DTC genetics, but all personal genetics

Steven Murphy MD said...

I do understand. But I am having a hard time swallowing the concept that our genomes will give us that much of a risk right now.

They Won't Keith.

Listen, I hate to tell you the story about the intellectually disabled child who, when CGH was done, we found a CNV region. Which we were sure caused the problem......

Until we tested both parents, high functioning lawyers and we found THE SAME CNV issue in Mom!

Keith, right now, most "risk factors" to be used are likely to be
1. Overturned
2. Watered down in risk

Thus we might as well call out Santa Claus and the Easter Bunny!

They work just as well to adjust behavior and are far cheaper than one of these scans!

So that is the test I am looking for
A true scan versus a "sham" scan with results and see what we get for behavioral modification.


Keith Grimaldi said...

Some thoughts, only a few though, Sunday morning...

1. Overturned? Probably not most of them, they seem quite robust as far as they go, it's not like the old days of association studies on single SNPs in 50 people.
2. Watered down? Maybe, but I hope that they will actually get more robust as environment interactions are factored in. Genetic data on it’s own is not so valuable
3. Risk – most collections of genes give small risk changes, plus most people will be at raised risk for some common disease and the advice in healthy people will almost always be good diet & lifestyle which holds for cancer, CVD, diabetes, etc. So until more environmental interactions are brought in the advice will be pretty non-specific so I would not call the panels of scans “actionable”. MTHFR 677TT needs more folate to keep homocysteine normal. That is actionable -. In a healthy person the claim is not that this will definitely reduce your risk of a stroke but it will reduce your risk of a stroke caused by high homocysteine (let’s not get into a Hcy discussion here though, this is just for example!)
4. Information – here there is personal utility. Individual learns about genes, proteins, metabolism and disease and is more likely to improve lifestyle. This is a reasonable assumption. Positive evidence is here – but… will it have the same effect in people who are not the curious early adopters? Studies required…

THIS is where the research should be – the point I am making is that the current research efforts, like those of Eng and EGAPP are wide of the mark, are not helpful, are not going to reach any other conclusion apart from the obvious. Genetics will not replace trad risk factors so enough with these comparisons. Genetics hopefully will complement FH, but if you want to find out, don’t do a study that looks at something that is not tested by the gene scan used!

Research should be:

1. Studying the effects (benefits/harm) of this information given to healthy individuals
2. Exploring far more deeply the effects of gene-gene and gene-environment interactions, because only by studying the modifiable components will we ever be able to deliver something where precise action can be prescribed

On point 2 I am not very optimistic yet, things are improving but unfortunately the research cash is going to get hoovered up by the much more seductive exon and whole genome deep sequencing.

Steven Murphy MD said...

So what you are saying is by presenting data that is "you" you will learn more about biology and genetics and epigenetics etc. etc.

In the decades which we have shown patients their cholesterols, CRPs, BPs, etc that IS most definitely them it has not, I repeat it has not greatly influenced health or motivation by patients to any significant degree.

Lipitor on the other hand has resulted in health benefits.

As for Hcy.....dude, not here.....I don't want to even start this debate until I have a good week off.....

Should we study what genetic data makes you "want" to do? Sure. But we should also study sham genetic data too.....

I think a few visits with the MD would work just as well as the DTCG data.....

Would love to see it put to the test.