Plant sterols/stanols and LDL

[Uncayimmy]

Being base two, mortality rates are digital. All other measures are analog, allowing for some grayness.

I'm not sure I understand. Is there something gray about a non-fatal cardiovascular event?

 

[casebro]

I'm not sure I understand. Is there something gray about a non-fatal cardiovascular event?

Well yeah, just exactly how much ischemia can a guy have before it is considered an event? An ER visit? What if you don't do labs? How bad do the labs have to be?

vs: dead.

But OK, go ahead and use the usual "fatal or non fatal cardiac events, or stroke" if you like. But rather than the '50% decrease in the treatment group", tell us how long the end points are delayed between the two groups in any study..

 

[Uncayimmy]

Well yeah, just exactly how much ischemia can a guy have before it is considered an event? An ER visit? What if you don't do labs? How bad do the labs have to be?

vs: dead.

So, you're saying that if we agree upon a definition an event, it's still gray. Is that correct? And because of this, you only want to look at deaths because there is no debate over the definition?

But OK, go ahead and use the usual "fatal or non fatal cardiac events, or stroke" if you like. But rather than the '50% decrease in the treatment group", tell us how long the end points are delayed between the two groups in any study..

Still not following you. Long-term risk assessment is part of life. This study looked at five-year outcomes for statins. Other studies look at long-term outcomes. I recall a four year study with Crestor that was stopped at two years because the benefits were already clear.

So, at this point I have no idea what you are trying to convey to me. Is this a Socratic approach?

 

[casebro]

Methinks UncaYimmy's lack of understanding is an act.

Oh well.

Folks, graphs are cool technology. You can use them to forecast, by extending lines in to the future.

Here are directions to make a pertinent graph:

Bottom line, mark off ten points, years 1-10.
Right hand axis, mark ten points, label by half- 1/2. 1, 1 1/2. etc
Establish end points at the five year line, end points are 5% end point in statin group, and 7 1/2%, end points in the control group. (This data is the ubiquitous "30% improvement on statins" we've all heard so much about. See, after five years, that is how many subjects readed the 'end point'- death, stroke, heart attack) Now, extend the statin line from the 5% percent mark to the right, at the same incline. Note that at about 5 1/2 years, the end point quantity matches the end point quantity of the control group. Note also, that at the 5 year mark, only 2 1/2% of the patients had benefited. The ones under the lines died. The ones above the lines didn't die, treated or not.

So, at this point our graph shows a triangle, 2 1/2% high by six months long. To average a triangle, you cut one dimension in half, so that 2 1/2% of patients average an extra 3 months (90 days) of 'benefit'. 2 1/2%, (.025) times 90 days, give the average benefit for all the patients treated with statins for five years. Final tally, for five years of treatment with statins, 2 1/4 days per patient. What a miracle drug!!!!

Now, how much benefit would you expect if UncaYimmy lowers his LDL from 90 to 70? Answer rounded to nearest whole second please...

(eta: Of course, I did forget to mention that taking statins for 20 years will make the benefit 4 times as much.) (another error- my mistake. 20 years would mean 8 times as much benefit. 20 years, 18 days)

(further eta, I don't remember the name of the high school math class that covered graphs. Algebra? But if you can't manage it, maybe remedial math class would help)

 

[Uncayimmy]

Methinks UncaYimmy's lack of understanding is an act.

When I find that I am unable to express in my own words what someone is saying, I consider that I do not understand it. That could be the quality of the explanation given, my base of knowledge or something else.

Folks, graphs are cool technology. You can use them to forecast, by extending lines in to the future.

Here are directions to make a pertinent graph:

Bottom line, mark off ten points, years 1-10.
Right hand axis, mark ten points, label by half- 1/2. 1, 1 1/2. etc
Establish end points at the five year line, end points are 5% end point in statin group, and 7 1/2%, end points in the control group. (This data is the ubiquitous "30% improvement on statins" we've all heard so much about. See, after five years, that is how many subjects readed the 'end point'- death, stroke, heart attack) Now, extend the statin line from the 5% percent mark to the right, at the same incline. Note that at about 5 1/2 years, the end point quantity matches the end point quantity of the control group. Note also, that at the 5 year mark, only 2 1/2% of the patients had benefited. The ones under the lines died. The ones above the lines didn't die, treated or not.

So, at this point our graph shows a triangle, 2 1/2% high by six months long. To average a triangle, you cut one dimension in half, so that 2 1/2% of patients average an extra 3 months (90 days) of 'benefit'. 2 1/2%, (.025) times 90 days, give the average benefit for all the patients treated with statins for five years. Final tally, for five years of treatment with statins, 2 1/4 days per patient. What a miracle drug!!!!

Now, how much benefit would you expect if UncaYimmy lowers his LDL from 90 to 70? Answer rounded to nearest whole second please...

(eta: Of course, I did forget to mention that taking statins for 20 years will make the benefit 4 times as much.) (another error- my mistake. 20 years would mean 8 times as much benefit. 20 years, 18 days)

(further eta, I don't remember the name of the high school math class that covered graphs. Algebra? But if you can't manage it, maybe remedial math class would help)

All you gave me was an endpoint value. How can I construct a trend line with just one value? I guess technically you gave me two values with zero being the implied value as the start point. Are you saying that statins have an immediate effect that is linear? Since age is a factor in risk, I wouldn't even expect the controls to be linear. You can just give me a table with data in it, and I'll chart it myself.

Beyond that, what justification do you have to calculate this down to time by person? In what way is that relevant to the choices I make about taking statins? Basically, I see it as a choice (using your numbers) between having a 5% chance of dying over the next five years or a 7.5% chance. I would weigh that against the costs and risks of complications with the drug.

 

[Perpetual Student]

The data is quite lacking about the quantity of plant sterols that might reduce LDL. Nevertheless, since some studies have "hinted" that there could be risks associated with excessive intake, my decision is to include some wheat germ and canola oil in my daily diet (perhaps about a tablespoon of each). Any thoughts?

 

[casebro]

Basically, I see it as a choice (using your numbers) between having a 5% chance of dying over the next five years or a 7.5% chance. I would weigh that against the costs and risks of complications with the drug.

Bingo! I've made a convert. Most people taking a statin seem to expect a miracle.

So far as linearity goes, the only study results that I know of that gave any intermediate data was a ten year study. Actually, it was two consecutive five year studies, with the control group given the drug starting at the five year conclusion of the first study. Same results for the second five year span, so linearity is proven.

And obviously, the chart lines start at zero time, zero end points. Where else would you start a study?

AAand, "Basically, I see it as a choice (using your numbers) between having a 5% chance of dying over the next five years or a 7.5% chance" brings up another point. The typical cholesterol study is done at the 50-60 age groups. Yet most of us die in out 70s. Why not study efficacy in that group?

 

[Uncayimmy]

Bingo! I've made a convert. Most people taking a statin seem to expect a miracle.

Since I never claimed it was a miracle drug, you converted no one. You did confuse me.

So far as linearity goes, the only study results that I know of that gave any intermediate data was a ten year study. Actually, it was two consecutive five year studies, with the control group given the drug starting at the five year conclusion of the first study. Same results for the second five year span, so linearity is proven.

Actually, no, that doesn't prove linearity. How about citing the one study you know of?

And obviously, the chart lines start at zero time, zero end points. Where else would you start a study?

If I am looking at the number of deaths over a period of time, then having N deaths at one point in time does not a line make. If I use 0/0, then I get a linear trend line when in reality I have no idea when the deaths occurred during my time line.

AAand, "Basically, I see it as a choice (using your numbers) between having a 5% chance of dying over the next five years or a 7.5% chance" brings up another point. The typical cholesterol study is done at the 50-60 age groups. Yet most of us die in out 70s. Why not study efficacy in that group?

Since you are only aware of one study, who says they don't? It's actually quite typical to adjust for age and other factors. Standard practice, really.

 

[xjx388]

The data is quite lacking about the quantity of plant sterols that might reduce LDL. Nevertheless, since some studies have "hinted" that there could be risks associated with excessive intake, my decision is to include some wheat germ and canola oil in my daily diet (perhaps about a tablespoon of each). Any thoughts?

I think that's perfectly sensible. I envy your numbers! Statins may not give you much in a "cost/benefit analysis" while there's no harm in increasing rich natural sources of sterols/stanols. Update us if you get a chance!

 

[casebro]

Actually, no, that doesn't prove linearity. How about citing the one study you know of?

I can't find it. Very few studies go beyond five years, here is a different one using Niacin: "Coronary Drug Project — 1,119 men who had survived a heart attack were given nicotinic acid (niacin)* and compared with a larger control group. After six years, total cholesterol was 10 percent lower. Nine years after the study, total mortality in the group that had received the drug was still 11 percent lower than that of the control group."

(eta: but wait, wasn't the Crestor study cut short at two years? End point data? Above or below my straight line? Does a 2-5 year schedule find most of the patients that benefit? I dunno)

If I am looking at the number of deaths over a period of time, then having N deaths at one point in time does not a line make. If I use 0/0, then I get a linear trend line when in reality I have no idea when the deaths occurred during my time line.

No study ever gives a graph in it's results. They only give the five year figure. I seem to be the only person in the world advocating a graph at all.

But, combining the "11% after ten years" from the study I quoted just above, with "30% after five years" would make a non-linear line, a downward curve, meaning my point is even stronger.

But believe my original point was that comparing "relative risk" causes an exaggeration of the expected benefit. We, as patients, don't care about death rate except as it implies longer, better life. My point of graphing the results of the studies is to forecast just how much benefit we can expect.

My own summation of the benefits of cholesterol: If the whole population takes Statins for our whole lives, our average life span will increase by a few days." Yet some people advocate putting it into our water supply.

Now it's time for me to take a break. I think I'll go have a triple bacon cheeseburger deluxe, with fries. All four fat groups in one bag.

 

[casebro]

Perhaps a comparison of all-cause mortality of two groups, high and low cholesterol, would simplify things? The end point of all those studies seems to be "heart attack, stroke, or a 'cardiac intervention' meaning fixing clogged arteries. Do you care what you die of, or how to extend your life?

 

[Uncayimmy]

I can't find it.

Bummer.

No study ever gives a graph in it's results. They only give the five year figure. I seem to be the only person in the world advocating a graph at all.

I regularly see graphs from studies in PubMed, but only when I have access to the full text. If only the abstract is available for free, you're not going to see the graphs. You can filter the results in PubMed to show only free full text articles (upper right side).

If you want to see charts for the Crestor study, you can look at the full study right here. They have a number of pretty little charts tracking things over time and showing hazard ratios.

But, combining the "11% after ten years" from the study I quoted just above, with "30% after five years" would make a non-linear line, a downward curve, meaning my point is even stronger.

If you read this, I think you will change your conclusion. The use of the word "still" in your quote was to point out that niacin "still" had an effect after discontinuation of the treatment, not that it was "still" 11%. During the study niacin did not decrease mortality. Nine years after discontinuing the treatment, they found a benefit.

But believe my original point was that comparing "relative risk" causes an exaggeration of the expected benefit. We, as patients, don't care about death rate except as it implies longer, better life. My point of graphing the results of the studies is to forecast just how much benefit we can expect.

I disagree as to the necessity of a chart for me to make my decision. What I want to know are the end point numbers of my risk group (age, sex, history of CV disease). I can look at the placebo group versus the treatment group and determine the potential benefit over that period of time.

My own summation of the benefits of cholesterol: If the whole population takes Statins for our whole lives, our average life span will increase by a few days." Yet some people advocate putting it into our water supply.

I had never heard of people advocating putting statins in the water supply. That's funny. I had to Google it. The following puts it into better context, which is to say that it looked like a publicity stunt:

http://news.bbc.co.uk/2/hi/health/3931157.stm
Dr John Reckless, chairman of Heart UK and a consultant endocrinologist at Bath University, put forward the case.

"The whole point of the debate is to bring out the fact that we are under-treating and the fact that a lot more people could benefit.

"The whole population should be following diet, lifestyle and weight loss measures. We shouldn't have our high- fat meals and we shouldn't lounge around, we should all be taking exercise and so on.

"Of course we all need that. But on the other hand, rather more people do need statins than are currently getting them.

"So maybe people should be able to have their statin, perhaps if not in their drinking water, with their drinking water.

"The issue is how far we should be encouraging wider use," he said.

 

[xjx388]

If you are interested in statins, LDL and mortality, I would suggest you look up the JUPITER study. It's starting to change the focus from "LDL" to "C-Reactive Protein." Interesting stuff.

 

[Uncayimmy]

If you are interested in statins, LDL and mortality, I would suggest you look up the JUPITER study. It's starting to change the focus from "LDL" to "C-Reactive Protein." Interesting stuff.

That's the study I linked to regarding Crestor.

 

[casebro]

If you want to see charts for the Crestor study, you can look at the full study right here. They have a number of pretty little charts tracking things over time and showing hazard ratios.

They all look linear enough to use for predictive purposes to me.

Check out their fig 1 (D), "Death from any cause". The difference is less than .01 after five years. And the lines are very, very linear.

Use your own math, whatever way you chose, and give my the answer to "how much longer can we expect to live, on average, if we all take statins for the last 40 years of life?" because, isn't that the point?

 

[Uncayimmy]

They all look linear enough to use for predictive purposes to me.

It was only for four years. I presented that study because you claimed (incorrectly) that studies don't show graphs, and that you were the only one talking about it. You can't even find the study you cited, so don't know if it had charts or not.

Check out their fig 1 (D), "Death from any cause". The difference is less than .01 after five years. And the lines are very, very linear.

Did you expect something different in a chart of death from any cause?

Use your own math, whatever way you chose, and give my the answer to "how much longer can we expect to live, on average, if we all take statins for the last 40 years of life?" because, isn't that the point?

Not to me it isn't, no. Why do I care about average life spans, specifically, how long other people live? I care about how long I personally live. Heart disease is but one factor in life span, and statins are but one factor for heart disease.

I guess we shouldn't make recommendations about how babies sleep because even if we wiped out sudden infant death syndrome completely, it would only add a matter of minutes to the average lifespan.

 

[casebro]

If you are interested in statins, LDL and mortality, I would suggest you look up the JUPITER study. It's starting to change the focus from "LDL" to "C-Reactive Protein." Interesting stuff.

Yup, I do agree with the fact that cholesterol is not the culprit. Artery damage is. And the inflammation as a result of the damage. Or maybe it is damage caused by the inflammation.

I think up-thread I mentioned the fact that aspirin alone does 80% of the same benefit as statins, that the two together are only an additional 20% benefit, therefor, statins do the same thing as aspirin, to the tune of 80%. ( and aspirin is lots cheaper.) And aspirin has no effect on cholesterols. Therefor, the statin/cholesterol link is pretty bogus, there is some other boogie man involved.

Try Googling "allergic angina" or "allergic myocardial infarction" for an eye opener. Treat heart attacks with anti-histamines. Think what "mild anaphylactic shock" would be.

My own history is of diabetes and myopathy. The myopathy precludes use of statins. My heart attack at 49 was treated with an angioplasty. Needed another angioplasty two years later. Third angioplasty 18 months after that. Six months later, angina getting severe again, looking another procedure in the face. (anybody care to look for a non-linear graph here?) At that point I coincidentally went gluten free in an attempt to aid a digestive problem. Within a week, people around me noticed the difference in my activity level. At the two week mark I threw my nitro pills away. That was about five years ago. Five years of home made sausage, I might add. Well, I lied about throwing away the nitro. I've needed them several times, from eating hidden gluten as "modified food starch", soy sauce, or wheat beer. So I take a nitro, and a claritin. Without the claritin, the angina lasts a couple days. With the antihistamine, gone in a day. And never another agioplasty involved.

 

[casebro]

Hey guys, those of you who looked at the Jupiter graphs, explain the numbers across the bottom of the graphs?

Called "number at risk". It gets smaller and smaller as time goes by. Starts at a total of 17,802, and by the end of the study they were down to fewer than 500? I guess that accounts for the "mean follow up of 1.9 years"? So this study had drop out rates of 95%?

Whoa, so a five year study of 18,000 people is actually a 23 month study of 500? And this study is the basis of billions of dollars of drug sales?

UY, did I get that right?

 

[Uncayimmy]

Hey guys, those of you who looked at the Jupiter graphs, explain the numbers across the bottom of the graphs?

Called "number at risk". It gets smaller and smaller as time goes by. Starts at a total of 17,802, and by the end of the study they were down to fewer than 500? I guess that accounts for the "mean follow up of 1.9 years"? So this study had drop out rates of 95%?

Whoa, so a five year study of 18,000 people is actually a 23 month study of 500? And this study is the basis of billions of dollars of drug sales?

UY, did I get that right?

As I understand it that study was ended early (19 months), and we're looking at data that went past that. I know that they held a two-week washout period where everyone got the placebo. They lost about 7% for non-compliance before starting the real trials. Amazing. 7% of people can't even take 80% of the pills for 14 days.

I don't want to get all sidetracked, but you should look at all the issues with adherence to treatment. I recently cited a 5-year follow-up study to look at compliance rates for prescriptions for women who had breast cancer, a highly motivated group. It was less than 50%. I cited another study where after oral surgery only 14% actually followed the recommended treatment (oral rinsing and antibiotics) for a few weeks.

It really makes you wonder about the NNT (number needed to treat) since in practice you'll need to treat a lot more than the theoretical number. It also makes me wonder how well they track compliance in studies like this. Are we underestimating the actual effect since it would seem that many of those taking the treatment are probably not compliant? It's not always clear how well they track compliance.

 

[Perpetual Student]

I recently cited a 5-year follow-up study to look at compliance rates for prescriptions for women who had breast cancer, a highly motivated group. It was less than 50%.

I don't know what that study was, but if it involved tamoxifen, the adverse effects can be severe for some, perhaps forcing many women to stop the treatment.