If There's No Benefit, Why Tolerate Any Risk?

June 1, 2007 — -- We have all grown accustomed to the scare of the week.

Each week we learn about another hazard that is lurking in our environment. We learn of something we are not doing that we must do -- or else.

We must eat fish, but not all fish. Last year if you fed your child butter you were negligent; this year if you feed your child margarine you are negligent. Tomatoes are health foods, or not.

We, the healthy, are taught that life is a minefield. We, the healthy, seem to have an insatiable appetite for scares of this sort. We leap to our own defense regardless of the reliability of the scare, the validity of the remedy or the expense.

With increasing fervor, the ill amongst us are learning a corollary lesson. The vaunted American health care system can be hazardous to your health.

First we learned that hospitals were dangerous places; if evil infections don't get you, errors by the staff might. But you don't need to be hospitalized to be at risk from modern medicine. Every week we have learned that another device or another pill was a Trojan horse, waiting to unleash some horror down the road.

Today I want to consider the question that should precede an assessment of hazard. In order to weigh a hazard, one must have a handle on the benefit.

The more the benefit, the more we might countenance some hazard. Likewise, if there were no important benefits then we would tolerate absolutely no risk.

Several common medical interventions have recently come under the gun and even succumbed to the identification of a measured, important, and putatively likely hazard that has been bellowed by the media -- often perking up the ears of the plaintiff's bar.

There are important lessons regarding the assessment of benefit in these examples:

Drugs to Treat Adult Onset, Type 2 Diabetes

Avandia is one of a newer class of drugs designed to lower the blood sugar of adults whose blood sugar is higher than is said to be good for them.

As we age, our own insulin is less effective in helping our blood sugar enter our cells to provide an energy source. Some of us have this tendency earlier than others, particularly if we have a big gut-to-butt ratio and/or we're poor.

This higher blood sugar and its fellow-travelers (higher blood pressure, higher cholesterol, and lesser wealth) are associated with earlier death, but only if any or all are particularly severe.

For over 50 years medicine has recruited the pharmaceutical industry to smite each of these "risk factors" a mighty blow in order to spare us grief. Avandia is another attempt to tackle persistently elevated blood sugar.

It works. It lowers the blood sugar. Furthermore, the earlier generations of drugs designed to do this also lower the blood sugar. They work too.

However, no one feels better for a lower blood sugar. Some feel worse or get fatter depending on the drug. And no one feels worse for a high blood sugar, except for the rare patient with adult onset type 2 diabetes who can mobilize an extremely high blood sugar.

It's like "high" blood pressure.

So Avandia does nothing for the quality of your life. Does it do something else -- save your life, or postpone the horrid complications some patients can get with adult onset type 2 diabetes and its fellow travelers?

We don't know for Avandia. However the precedents are daunting. Long-term experiments, randomized controlled trials, with earlier generations of drugs that lower blood sugar are not encouraging. One famous trial lasted over a decade.

There is no precedent for any of these drugs saving a life, a limb, an eye, kidney or anything else important. There is no demonstrable benefit except the lowering of blood sugar. Who cares?

I have practiced medicine for 40 years. I have never prescribed a pill to lower blood sugar. I still see no reason to do so. If I am disadvantaging my patients, it's to a trivial degree at most. However, I know I am sparing them known and unknown hazards.

And I won't let you measure my blood sugar or the measure of its persistent elevation, the hemoglobin A1c. I don't care, and I won't care till there is compelling science that something meaningful can be done if it is elevated.

Drugs to Treat Painful Joints

Vioxx and Bextra were pulled off the market because of data suggesting they imparted a tiny hazard for heart disease. It's perhaps a slightly more convincing hazard than that purported for Avandia, but there's little here worthy of alarm.

I am a rheumatologist. I have never prescribed Vioxx, Bextra or their cousins such as Celebrex. But I never prescribed them for reasons other than their possible effects on the heart.

Do you know that no drug of this class has ever been shown to be more effective than aspirin? Do you know that no drug of this class has ever been shown to be safer than aspirin? Only Vioxx convinced some who are swayed by tiny differences. But Vioxx never convinced me.

Why would I ever prescribe this class of drugs? I don't care if there are TV advertisements claiming magical effects like you, too, can skate on ice. I don't care if the occasional patient swears nothing else helped. I am unwilling to prescribe a drug that has no important benefits over the old standbys and has no long-term track record for safety.

I spend much more time explaining my philosophy to my patients than the seconds required to fill out the prescription, or to give a sample. By the way, neither drug company sales people (detail representatives) nor drug samples are allowed in my clinic -- and never have been allowed. I'll take the time to explain my philosophy or stop practicing medicine.

However, I am a "senior" clinician; if my students, and there are many out there, practiced the way I taught them, they'd starve. There is something horribly rotten in the United States, not in Denmark.

Drugs to Lower Cholesterol

Baycol is a statin. That's a drug that lowers cholesterol. It was pulled from the market because of about 50 cases of a complication that almost never occurs unless you are on a statin.

This is a complication that causes the muscles to die, and often the patient follows. It occurs occasionally with all statins (Lipitor, Pravachol, Simvastatin, Crestor, and others). But somehow the number of cases ascribed to Baycol lead to its banishment and not the others. After all, the common wisdom is that lowering cholesterol is too good a thing to pass up.

It is true that cholesterol is a "risk factor" in people who do not have an extraordinary family history of youngsters dying of heart disease. But it is not that much of a risk factor. If you have a particularly high LDL cholesterol (the "bad" cholesterol), and a particularly low HDL cholesterol (the "good" stuff), you are harboring a 2-3 percent mortal hazard.

That means you are at risk for living a year less than others born when you were. Very few of us have even this degree of risk. For most it is a matter of months, if you believe such a tiny risk is measurable.

It is true that these statins can lower your bad cholesterol.

It is true that maybe they can spare you a heart attack, but that's a real MAYBE.

It is also true that no study has shown it can spare me death before others born when I was born. That's true for men of my generation. There are no compelling studies for people who are not men of my generation.

Now you know why I won't let anyone check my blood cholesterol. Why bother? You can lower it, but you can't do anything for me that I would consider meaningful. I'd rather not know if I have any special risk of earlier death, even the tiny risk imparted by high cholesterol.

Merck is faced with a class action law suit because its advertising allegedly suggests benefits from Lipitor for women when there is no supporting science. I don't think there are benefits for anyone in the general population.

I have never prescribed any drug of this class.

Caveat Emptor (Buyer Beware)

I have written commentaries for ABCNEWS.com relating to the lack of benefit from stents and mammography. I will not reiterate these since they are readily available in the archives.

The lesson is obvious. We, all of us, need to demand that no drug be licensed without a demonstration that it is meaningfully effective.

Today, the game is to generate evidence of any effect. We should demand much more than evidence for a surrogate effect such as lowering cholesterol or lowering blood sugar. We should demand more than evidence of an infrequent benefit that is not that crucial. We should demand robust evidence for a meaningful effect.

Then these tiny, remote harms would be countenanced.

Then all the hype, the "me too" drugs, and the profitable silliness would be dark history.

Then we would all be better off.

Dr. Nortin Hadler is professor of medicine and microbiology/immunology at theUniversity of North Carolina at Chapel Hill, and an attending rheumatologist at University of North Carolina Hospitals. He is the author of The Last Well Person.