At the Emory/Georgia Tech Symposium on Predictive Health today Dr. Jeffrey Gulcher (right, from MSNBC) revealed how he did.
But he intentionally buried that lede because it's more important you hear the rest of the story. And so you shall.
Dr. Gulcher is co-founder of deCODE Genetics, which he launched with one of his University of Chicago teachers in the 1990s.
They are based in Iceland, home of co-founder Kári Stefánsson, because it has a homogenous population with genealogys traced back 10 centuries, great health records, a single-payer health system, and patients who do what the doctor tells them.
All this makes it possible to isolate genetic factors from other causes of terrible but common diseases.
Over time deCODE developed tests on genetic risks for six leading causes of death -- heart attack, stroke, diabetes, glaucoma, breast cancer and prostate cancer.
I admit that when Dr. Gulcher mentioned heart attack I perked up. I have a family history of both hypertension and heart problems.
deCODE's work found that, independent of other risk factors, a sequence on the number 9 gene could increase your chance of a heart attack by a factor of 1.3.
This doesn't mean you're going to have a heart attack if you have the defect. This needs to be put into the mix alongside other risk factors -- smoking, obesity, family history, etc.
Many people have a different marker that should lower their risk, he added. Heart attack doesn't pass directly from father-to-son, but skips around in and among generations.
As part of the work commercializing their test deCODE also put together a pamphlet for doctors, highlighting that 1.3 number and what it means.
"This risk factor is similar to that from high LDL cholesterol and is independent of family history," Dr. Gulcher said.
Doctors can plug this into common charts showing your chance of having a heart attack over the next 10 years. Say a 56-year old female smoker comes in. If she has the marker her risk of a heart attack by 66 rises to 13% from 8.8%.
She's no longer average risk, in other words. Now she's intermediate risk. And the doctor can guilt appropriately.
If she doesn't have the marker, you don't say "light up" -- you treat her as you would the average patient. But if you do have the marker, perhaps your target for bad cholesterol should be lowered, from 160 to as low as 100.
"We're just talking about doing a better job of risk prediction, putting people at higher risk under closer study," Gulcher said. "This is easier to measure than a CRP level. The work is widely replicated."
But I know you. You want to hear how this saved his life.
The test showed he faced an increased risk of developing prostate cancer, so he got a PSA test. This is commonly done at 50, not 48.
The PSA test was positive, so he went to have a biopsy. That too was positive so he had a radical prostatectomy, six months ago. "My PSA went down to zero in six months" and he's perfectly healthy -- he talks faster than ever. (He's a very fast talker.)
Note, Dr. Gulcher would not have had the PSA done early without the genetic test, and it was the PSA test which eventually led to an operation. But few men have that test before age 50, and how much worse might the cancer have been then?
deCODE has created other products, including deCODEme, which checks all your genes and can alert you when new risk factors emerge, and deCODEmd, which will be sold to doctors.
The joke is he's not just the chief science officer, he's also a client. But as you can see he didn't go for a hair transplant. Dr. Gulcher went for a longer life, which he's spending trying to make yours longer, too.
There is one more important point to make here which relates to marketing. Obviously deCODE is not the only outfit selling genetic tests. It's not even the most famous.
23andme is far better known, with big-name founders and big-name backing.
But note how deCODE is going to market. Through the profession, with specific tests using genetic data, layering on physician education, and with a ton of scientific journal articles backing their every play.
It's the tortoise method, contrasted with the 23andme hare. Which do you think will work best?