The Cholesterol Machine

One of the questions I get from almost all of my middle-aged patients is, “How is my cholesterol?”  It seems  an odd question to me, as if I had a personal relationship with “cholesterol” and knew how it was doing.  Can cholesterol “feel” good or “feel” bad?  I’ve never had a patient (OK, one, but he was crazy) come in and say, “My cholesterol feels high.”  If it did “feel high” what would that mean?  High cholesterol does not make one  feel bad.   A person can’t “feel” cholesterol anymore than he can “feel” his blood type or his race.  Not until it gets ridiculously high would one ever see a physical sign of high cholesterol.[i]   “High cholesterol” is a test result that we doctors came up with in our attempts to stomp out disease.  We’ve been trying to figure out what causes heart attacks and strokes for a long time and the truth is we still really don’t know.  You won’t hear that truth from too many doctors – the “we don’t know what causes this” truth.  We’re more likely to put forth some confounding “facts” such as “we know that a high cholesterol level is associated with heart disease.”  We do this because we, and most of America, believe in the deductive reasoning approach to science and health.  If we see that people with high cholesterol levels are having more heart attacks then that must be the cause, right?  It’s almost never that simple.  Often the approach sort of works, or works some of the time, and that is the story with cholesterol.  A famous guy named Gödel, who was a mathematician, showed that just because something was consistent, it wasn’t necessarily true!  We consistently see high cholesterol levels associated with heart disease but to say that high cholesterol CAUSES heart disease is a leap of faith that probably is not true.  It certainly isn’t the whole picture.

We used to have a picture in our “scientific model” of heart disease as cholesterol “plaques” building up on the inside of arteries and finally getting so big that they would block the blood flow to the heart or brain causing a heart attack or stroke.  You can see nice animated cartoons showing this hypothesis on a multitude of silly commercials advertising statins.  It turned out not to be very accurate.  We now “know” (I should probably substitute “think” into every sentence where we doctors use the word “know” if I want to be truthful) that the story is not so simple.  In fact most blockages actually happen when partial blockages that really are not so big (cholesterol plaques blocking 20-50% of the artery) break loose and get caught somewhere downstream.  That’s probably why stents and bypasses to get around the big blockages don’t actually accomplish much.  More on that later.

With the idea in our heads that cholesterol must be the culprit in heart disease, we came up with all kinds of ways to get those evil levels down.  But now we get to one of those critical problems when it comes to primary care/preventative care:  How good is our theory, how good is our treatment, and how much HARM will the treatment cause?  I’m speaking here of people who do not already have a cardiac problem – people who feel fine but went to the doctor for a check-up.  If you can’t feel your cholesterol and it’s not causing you ANY problems, why should we check it?  Well, to prevent something bad from happening, right?  Here is where we doctors have really oversold the whole thing;  that giving someone medicine for their cholesterol is going to prevent a heart attack and save their life.  The idea has proved to be barely, if at all, correct.  There might be a few people who will benefit but MOST will not.  Worse, it may be that we harm as many as we help.  I’ll quote the numbers in a bit, but think about it for a second.  Here some poor innocent law abiding patient is going to his doctor to get “checked out” and he does the normal thing of checking the cholesterol.  It comes back “high” so he goes back for another visit to “discuss the findings.”  He was feeling FINE.  Now he’s a bit scared.  Scared about some nebulous thing called “cholesterol” that he can’t see, feel, hear or touch.  If it weren’t for his doctor telling him so, he might not even know/believe the animal existed.  But now it’s become the  bogeyman.  If he doesn’t  do something about the EVIL CHOLESTEROL he could die!

My biggest issue with our current approach is the psychological damage we inflict on our patients by propagating disease.  High cholesterol is NOT a disease.  It’s an artificially created name made up by doctors to scare patients into doing something about IT.  They mean the best.  Most truly believe they are saving lives by doling all those prescriptions out.  They don’t believe THEY have been influenced by the free pens, golf outings, doughnuts etc.  They are simply following the “guidelines.”

Heart disease is a disease.  Strokes are diseases.  High cholesterol is the result of a blood test that a doctor did trying to prevent a real disease.  We even give it exotic scary sounding names like hyperlipidemia or hypercholesterolemia or dyslipidemia or the mouthful hypertriglyeridemia!  Just saying it three times fast might make you winded.  Now a healthy person who should be focusing on caring for his family is all worked up about a blood test.

So, before I carry this too far and give you the impression that it’s totally not worth ever testing cholesterol, let’s look at the actual chances it will help you.  Remember again, I’m talking about the average 40-50 something year old with no history of heart attacks or strokes.  There have been many, many studies with lots and lots of people/subjects testing the idea that taking statin drugs for high cholesterol will decrease your risk of heart attack, stroke or death.  The results are, how shall we say, not terribly impressive.  The study that got this whole statin story started was called the “West of Scotland” study.  The drug companies were looking to expand the market for their product.  Treating people who had already had a heart attack is a big market but if they could treat everyone with high cholesterol, that’s an enormous market!  The “West of Scotland” study was set up to try to prove that giving healthy men ages 45-64 a drug called pravastatin (name brand Pravachol®) would prevent heart attacks and death.  They got almost 6,600 patients to take either the real drug or a placebo for an average of five years.  Like most studies that look to test a drug’s benefit, this one was paid for by the drug company that makes name brand pravastatin.  When they finished collecting the data it wasn’t what was called “statistically significant.”  Statistically significant usually means, in scientific terms, that you have a “P” value of less than .05.  It basically means that if you did the whole thing over again, at least 95 out of 100 times, you would get the same results.  If you had a “P Value” of .01 then you should get the same results 99 out of 100 times.  The cut off is arbitrary but 95 out of 100 is “good enough” according to most “experts.”  The problem with the West of Scotland Study was it didn’t quite get to that magic P value.  There had been 52 heart attacks in the group of 3,293 patients taking the placebo and 38 in the 3,302 taking the real drug.  They were SO close.  So they went back to the drawing board and, as my colleague Nortin Hadler[ii] likes to say, “tortured the data.”  They looked at all the results and said to themselves (well, I wasn’t there so I don’t know exactly what they said, but…) “Hey, if we…..”  And they found something!  They saw 10 deaths that had occurred in the placebo group that had been labeled “Noncardiovascular death” and moved them over to the “Deaths by heart attack” category.”  That successfully got the magic P value down to .042.  Now the company had a blockbuster;  they could say that their drug decreased heart attacks by 0 .6 percent!  Of course they didn’t quote that number – the ABSOLUTE risk reduction.  They quoted the RELATIVE risk reduction – because it sounds so much better.  Remember that number needed to treat?  How many patients have to take the drug for ONE to benefit?  It’s so easy to calculate but they never ever tell you it in the 10 page FDA insert you get with your prescription.  You have to look at the ABSOLUTE risk reduction, never the relative risk reduction.  In the West of Scotland Study there was a 98.3 percent chance you were NOT going to die of a heart attack if you took a placebo.  If you took the “real thing” you had a 98.8 percent chance.  So somewhere hidden in all that mumbo jumbo statistical manipulation is the real answer – over 200 people have to take the drug for 5 years to prevent ONE heart attack death.[iii]

After this study came out in 1995 huge numbers of patients started leaving their “annual physicals” with a prescription for a statin.  I was a young doctor at the time and fell prey to the fallacy myself.  I even attended a few “events” to get “educated” about the results.  These events usually occurred at a high end restaurant and were sponsored by either the drug company themselves or some umbrella corporation set up to “educate physicians and patients” as to the danger of high cholesterol.  Of course they were funded by the drug companies!  The food was alwa ys good and they often had some big name from an academic institution lined up to give the lecture.  These doctors were on “speaker’s bureaus” for the pharmaceutical corporations.  They would be paid anywhere from $500-$5,000 to give a little talk to us unsuspecting doctors.  The next day we would go forth with a bellyful of great information and a prescription pad loaded for action.  In fact, many of them would even provide cute little prescription pads with my name and office already printed.  The trick was that every 5th page would have an ad for the drug and every 10th page would have a pre-printed prescription for their drug!

The Famous “Combined Endpoint”

A more recent study striving to show that statins were great for you was called the “JUPITER” study.  It was published in the New England Journal of Medicine in late 2008[iv]  When statins first came out they were used for people who had had heart attacks.  Now we were supposed to use them for people who didn’t actually have any heart disease but had high cholesterol.  What about those with normal cholesterol?  Surely they could benefit from a statin!  Enter the “C Reactive Protein.”  C Reactive protein, or “CRP,” is a marker of inflammation that lives in your blood.  We can measure it.  The people who were running this study realized if they could find another group of otherwise normal people who would “benefit” from statins they could make a LOT of money.  This business plan works particularly well if you can catch them when they are fairly young and fairly healthy.  Then they will live a long, long time all the while taking your $4.70/day drug.[v]  Pretty good income stream if you can make it work.

To run these huge trials costs a ton of money.  The people paying for it want to make sure that the numbers are going to work.  In a study you have what are called “primary” and “secondary endpoints.  These are basically stating “I’m doing this to check that and I’m also going to look at this other thing too.  The “that” is the primary endpoint and if you don’t get a good number with that one then usually you don’t get published.  In order to stack the deck they often combine a whole bunch of things together and call them “combined endpoints.” In the JUPITER study the intervention was giving people with high CRP but with normal cholesterol 20 mg of Crestor®.  They had 17,802 patients.  Yes, you read that right:  A whole stadium full of patients.  Half got the Crestor and half got the placebo.  The primary “combined” endpoint was myocardial infarction, stroke, arterial revascularization,

hospitalization for unstable angina, or death from cardiovascular causes.  Did you catch that?  1,2,3,4,5

things.  Note also that two of the five are very subjective.  “arterial revascularization” and

“hospitalization for unstable angina.”  The “need” for arterial vascularization and hospitalization

is often quite arbitrary.  In fact many studies in the past few years have questioned why we do all of the bypass surgery and stents we do.  In any case my point about the JUPITER study is that to prove the drug is working, they look at a whole bunch of things together in the so called “combined” endpoint.  In this case the five things they were asking the study to answer was: Did Crestor reduce or prevent:

a) Death

b) Heart Attacks

c) Strokes

d) Need to be admitted to the hospital with chest pain (angina)

d) Need to get “revascularized” (an angioplasty, stent or bypass)

In the end they were ecstatic.  The study was designed to run for up to five years but by 1.9 years it had achieved “statistical significance.”  It was so good they stopped the study right then and there and declared victory.  If you can’t sense the caustic tone of my voice in the last sentence I’ll elaborate;  the results were O.K.., and to my mind just barely O.K.  The drug achieved it’s goals but at a rate that was  just barely meaningful. Truly meaningful to me would mean that it helped MOST of the people who took it.  There are very few drugs that achieve that level of meaningful.  Insulin for type I diabetics, food and water.  Not too many others.  Most drugs are lucky they actully help one in 10 people.  Let’s see what the Crestor numbers were.

The results showed that people who took the drug had one of the 5 things listed above happen to them at an extrapolated rate of .77 times per 100 “person years.”  The people who took the placebo had it happen at 1.36 times per 100 person years. The actual numbers were 251 “events” in 8,901 people in the placebo group and 142 “events” in the Rosuvastatin (Crestor©) group.  The study group then extrapolates,  in the “results”  section of the paper, that one would have to treat 25 patients with 20mg/day of Crestor© for 5 years to prevent one “primary end point.”  Without taking any issues with their math, that calculates out to $214, 437.50 to prevent one of the above.  Now if you are the guy who wins the drug lottery here you might say, “O.K. It was worth it,” but if you are one of the 24 who wasted $8,577.50 for nothing you might be upset.  The problem is we never know who will be the winner.  I don’t even have a problem paying for it.  I mean, they did prove the drug works.  I just have a problem with the marketing.  If you look at what makes it to the newspapers and certainly if you watch the ads on TV or in the print media (or even online these days) you would believe that this drug is going to help most of the people most of the time.   How likely would you be to buy it if they said, “Well, there’s a 24 out of 25 chance this isn’t going to do jack for you in the next 5 years?”

If we look at the really important number in this study, how many deaths were prevented, it’s very underwhelming.  Out of 8,901 patients in each group 198 died in the Crestor© group and 247 in the Placebo.  It’s statistically significant, but what if you look at it a different way?

If you took the placebo there was a 97.22% chance you were going to be alive at the end of the study.  If you took Crestor® there was a 97.78% chance you would still be alive.  In other words about a half of one percent chance better.  In drug company terms that’s a blockbuster.  In real terms that means that 179 out of 180 people who took the drug in the study didn’t cash in on the death prevention benefit.

What does it all mean to you?  Well, if you have 3rd party insurance, have a low deductible and the drug doesn’t make you feel bad, maybe it’s a reasonable insurance purchase.  It’s unlikely to help you but you could be the lucky one.  If, on the other hand, $1,715.50 is a lot of money to you for one years’ worth of drug, then you might think twice.  Lots of other studies have shown that your socioeconomic status is the most important predictor of longevity and health.  If buying expensive medications is making you poor and causing stress then maybe this is a false bargain.  Part of the problem with drug studies is that the results are almost always better in the study than in the real world.   First of all, the drugs are almost always provided free during the study.  Second, they have nurses and consultants always contacting patients to make sure they are taking the drug correctly.  Often the patients are actually getting paid for their time and to be the guinea pigs.

The last thing that really bothers me philosophically about this particular study is that it went looking for healthy people to treat with a drug.  They went and measured a substance in the blood that you can’t feel, taste or see.  If it was higher that “x,” an arbitrary number, they declared you “sick.”  To be fair, they were testing the hypothesis.  Now that they’ve “proven” that it works they can declare every healthy person out there with a high CRP level lurking in their blood a “sick” person who “needs” to take this wonderful medicine to “cure” them of all their woes.  “It could save your life” might read one of the potential ads, hiding the relative ineffectiveness (179/180) under the glossy print.  We are creating 179 “sick” people who have to take medicine for every one we save.  To my mind that’s not a straight up bargain.  The dollar value is $583,441 to save that life, but the cost in time, psychological despair and to the approximately 1/3rd who will have physical symptoms (muscle aches, etc.) makes the cost even higher.

In the end, my point about cholesterol is that:

a) We know a lot less about cholesterol’s relationship to diet, heart disease, strokes, drugs and death than we let on.

b) We way, way over sell the benefits of the drugs used to lower cholesterol.

c) You shouldn’t spend too much time worrying about it.  There are a few things you might consider doing, but they probably won’t affect you anyway.

d) If worrying or paying for cholesterol medicine//testing etc. is causing you a lot of stress it’s certainly  not worth it.


Addendum 12/2013

The powers that be came out with new recommendations recently and decided that we shouldn’t be treating simple numbers.  I certainly agree with there initial intentions, but they set a threshold value of 7.5% “risk” for the future 10 years as to when to recommend statins.  Unfortunately at least 2 major experts take issue with the online calculator used to calculate that risk, stating it vastly overestimates the risk!  I’m not sure if I had a 92.5% chance of being ok that I would agree to take a statin but if the calculator is overestimating the risk in the first place the whole thing gets really fishy!  Stand by for details.

[i]           xanthomas are benign cholesterol tumors that grow in the skin of people with certain -usually inherited.

[ii]   Professor of Rheumatology at UNC Chapel Hill and author of “Worried Sick” and “The Last Well Person”

[iii]  Absolute Risk Reduction equals 98.8%-98.3% equals 0.005.  Half of one percent.  1/0.005=200 equals the NNT

[iv]            Rosuvastatin to Prevent Vascular Events in Men and Women with Elevated C-Reactive Protein

Paul M Ridker, M.D., Eleanor Danielson, M.I.A., Francisco A.H. Fonseca, M.D., Jacques Genest, M.D.,

Antonio M. Gotto, Jr., M.D., John J.P. Kastelein, M.D., Wolfgang Koenig, M.D., Peter Libby, M.D.,

Alberto J. Lorenzatti, M.D., Jean G. MacFadyen, B.A., Børge G. Nordestgaard, M.D., James Shepherd                          M.D.,James T. Willerson, M.D., and Robert J. Glynn, Sc.D., for the JUPITER Study Group*

New England Journal of Medicine, November 20, 2008 vol. 359

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s