It's the Phenome and Not the Genome: Put Your Money on Mortal Flesh

doctor-David McNew-big.jpgStrong is your hold O mortal flesh . . .

From The Last Invocation, Walt Whitman

Is it just me, or are you also getting a bit tired of all the hype about the genome? Don't get me wrong-- it's pretty incredible that in my lifetime we have mapped out the 25,000 plus genes in our DNA. What's even more amazing is that the price for that chart of the human genome has gone from millions to less than $50,000 and now it takes only a few weeks. I bet by next year it might be a few hundred dollars and take a day! Companies like 23andMe (an innovative venture with a great marketing plan) offer to check you for genetic markers that predict your risk for certain diseases for just a few hundred dollars.

But the fact remains that for most of us, the genotype is much less relevant than the phenotype. What is phenotype? It is the things we can see, the outward or observable physical or biochemical characteristics and they are determined by both your genetic makeup and environmental influences. Your blond hair, your weight, your strange nose, green eyes and that funky shaped little toe of yours --all examples of phenotype.

So what do I mean when I say phenotype is more relevant than genotype? Well, let's say a new patient, a male, walks into my office and he is in his fifties. Let's say he happens to have the outline of a pack of cigarettes showing in his front pocket. As a male he already has one risk factor for coronary artery disease--just being male, alas. The cigarettes tell me that he is four times more likely to have a heart attack than his peers who don't smoke. His risk of sudden death is at least doubled. Let's say I notice he happens to be carrying more than 30 pounds of extra poundage above the belt line: that allows me to predict he has a higher chance of being at risk for diabetes, if he is not already frankly diabetic. Let's say that I notice too the pale outline of a recently-removed wedding ring (I can't help it, my eyes are always looking at the body as text--even when I am out of the hospital), then I know that his risk of death as a recently divorced man can be double that of his married peers.

At this point, before he has even said a word or before I have examined him, I already know so much about his risk of death and disease. Once we talk and I learn more about his job, his stress, his heredity, his habits, his past illnesses, then my predictions get more accurate. Once he disrobes and I examine him, I might find other phenotypic markers that predict risk (such as yellow plaques related to high cholesterol on his eyelids or elbows; high blood pressure; skin tags and velvety darkened areas of skin that predict diabetes; narrowed blood vessels when I look into the back of his eye . . . the list could go on for pages). In short, I'll have an excellent sense of my patient's risk for death or disease. At that point, mapping his whole genome, sexy as it might seem, won't tell me much more than I know and will probably matter much less than getting him to quit smoking, exercise and lose weight.

The famous Whitehall Study of British Civil Servants ranging in ages from 20 to 64 found that the lower grades of civil service had higher mortality rates from heart disease and from all causes than did people in higher grades, even after accounting for risk factors like obesity and smoking. (Yes, it was counterintuitive and that is why we do studies).  Stress was thought to be the factor responsible for this disparity.

The Whitehall studies are ongoing and one of the latest reports from that study made me think of Walt Whitman and reminded me that the phenotype is so relevant. In their report (titled, "Utility of genetic and non-genetic risk factors in prediction of type 2 diabetes: Whitehall II prospective cohort study" and appearing in the British Medical Journal, 2010 Jan 14;340:b4838), the scientists compared a panel of genetic tests for diabetes (common single nucleotide polymorphisms) with non-genetic or phenotypic findings like age, sex, drug treatment, family history of type 2 diabetes, body mass index, smoking status, HDL, triglycerides, fasting glucose.

What they found was that the phenotypic tests did better. Indeed the gene tests added little to the risk already determined by phenotypes. In their own words, "the addition of genotypes to phenotype based risk models produced only minimal improvement in accuracy of risk estimation  . . ."  Translation: use your eyes, take a good history, weigh the patient and get a few simple blood tests, and you can predict risk far better than a panel of genetic tests.

I am not a Luddite (I find I say that a lot) and indeed, I do think the genome studies will help us eventually understand more about causes of disease, and perhaps even point to particular treatments. But until then the message for us in the trenches is: Strong is your hold O mortal flesh and that's where the money (speaking diagnostically) is.

Photo credit: David McNew/Getty Images