The Seesaw of Risk
If it’s not clear by now, I am an advocate of early intervention. It’s such a simple premise towards disease treatment: the earlier you intervene, the better your chances of reducing - and even preventing - disease. That’s true on an individual/clinical level, as well - and more importantly - on a population level (a little understood paradox of the health insurance industry, for instance, is that they perceive disease prevention to be cost prohibitive, since the average enrollee only stays in plan for a few years, too short a time to wield any profit on preventive medicine).
The basic principle of early intervention is that you identify risks of disease, rather than wait for causes or symptoms. And one fascinating result of that principle is that we now have a range of diseases which are fundementally risks. Metabolic syndrome, which I’ve written about, is one example; high cholesterol is probably the paragon of risk-based disease.
By and large, the emergence of high cholesterol treatments - including statins - has been a boon to the fight against heart disease. In the US, about 10,000 lives a year are saved by statins, according to a recent study (aspirin, surprisingly, has helped save more than double that amount). This has led to demand from some quarters, particularly in the UK, that statins be sold over the counter - making them more available and more likely to save more lives.
Now it’s important to remember that high cholesterol (or, more precisely, high LDL cholesterol and low HDL cholesterol) is just a risk of heart disease; it doesn’t constitute any real disease - understood as progressive injury to the body - in itself. But that hasn’t stopped people from making the misstatement that cholesterol itself is a “killer”, or from outlining risks for developing high cholesterol (in other words, risks of risks of disease). Indeed, try to find out exactly what the risk of heart disease is from high cholesterol, and you’ll likely be stymied as I have been for the past half hour. While the websites for the American Heart Association and the Mayo Clinic and WebMD all talk about cholesterol as a risk factor, they don’t quantify it at all (is the risk 20% higher for heart disease? 40%? 80%?). UPDATE: Finally found it - the ATP III [PDF link], a 2004 NIH report on treatment for high blood cholesterol defines it thus, for LDL cholesterol levels over 190 in the “highest risk” category:
The category of highest risk consists of CHD and CHD risk equivalents.
The latter carry a risk for major coronary events equal to that of established
CHD, i.e., >20% per 10 years (i.e., more than 20 of 100 such individuals
will develop CHD or have a recurrent CHD event within 10 years).
By comparison, the ATP III says people with no risk factors have a 10 year risk of less than 10%.
Statins reduce that risk significantly: a 2005 metastudy found that a 40-point reduction in LDL levels reduced a patient’s risk of heart disease by 20%. This is where the math gets hard for an individual: If you had an LDL of 190 - and thus had a 20% risk of developing heart disease, and statins reduce your level to 150, does that mean you subtract 20% of the 20% - and thus you now have a 16% risk of developing heart disease? Or do you recalibrate your risk level from the ATP III? It’s not clear to me.
My math troubles aside, these results have been hailed as dramatic, and have turned statins into the biggest blockbuster drug of the past decade (sales of statins run somewhere around $13 to $15 billion a year).
But: now there’s a new wrinkle. A new study out today found that statin use is associated with slightly higher cancer rates. (The Journal of the American College of Cardiology, which published the study, is subscription only, so a thorough WebMD story is here).This study found association, not causation, so it’s not clear that the increased cancer is due to the statins themselves. And again, I can’t track down a specific figure for what “increased risk” means - it says 1 case of cancer per 1,000 patients, but given that the meta-study had 90,000 patients, that means only 90 cases were found, not enough, I’d gather, to create a bonafide increased relative risk figure.
So my point is this: It’s great that we’re moving into an era of risk-based medicine, where we start acting earlier and, overall, will no doubt save more lives and extend the quality of life for others. But this is fundementally a numbers game based on odds and percentages, and these aren’t so easy to measure, grasp, or to compute. Indeed, we may be able to crunch them out for populations, but that doesn’t necessarily mean anything to one patient facing personal decisions.
Play this out: Patient A has some condition B that elevates his risk of disease C by D percent. He takes drug E, which reduces his risk of disease C by F percent. But drug E increases his risk of disease G by H percent. What we want to know, really, is I - What is the net result? Have we made the patient healthier, or sicker? And has this battery of drugs and diagnoses improved his quality of life overall or reduced it?
In an age of algorithms and biostatistics and metaanalyses and databanks, there is the impression (the illusion?) that we are well on our way to figuring all this out - that these formulae are being cranked out and that in a few years we will have decisive numbers on all this. Me, I hope that’s true. But I wouldn’t bet my house on it.
Leave a Comment