Monday, November 10, 2008

Cholesterol, heart attacks and JUPITER

There was a snippet on Radio 4 about the JUPTER study this morning. It was basically an ad for statins for everyone in the UK. Well, 75% of the adult population.

What the JUPITER study found was that in people with low cholesterol levels but elevated hsCRP levels, taking a statin dropped the overall mortality from low to a bit lower.

Bodycounts were 198/8901 in the statin group and 247/8901 in the placebo group. Clearly treating 8901 people for a year and a half will prevent 49 deaths. According to the lipid hypothesis, all of these lives saved should be as a result of less cardiovascular disease, such as fatal heart attacks.



Well Table 3 from the study gives you the total numbers of deaths from heart attacks, but not directly. You have to look at the line "Any myocardial infarction" and subtract the previous line, "non fatal myocardial infarction". Simple arithmetic, which didn't get done in the write up, then gives you the number of fatal heart attacks... Here are the missing numbers:

Heart attack deaths on statin: 9
Heart attack deaths on placebo: 6

Hmmm, 50% increase of being dead from a heart attack on Crestor, with an LDL half the level of the people on the placebo.......... Good job the study was not powered to detect this!

So what did the people die of in the placebo group? Well, it certainly wasn't heart attacks.

hsCRP is a protein produced by your liver as a message to your neutrophils to get out there and do whatever they should be doing, but more so. It probably does lots of other things too. The list of diseases that come with elevated hsCRP is huge. It looks like taking Crestor does blunt this message. Whether that is why the death rate dropped on the statin is anyone's guess. Statins have lots of effects and, believe it or not, not all are acutely bad!

As an aside: Why were there so many cardiac procedures in the placebo group? Perhaps this is what saved their lives? Unlikely. The explanation could be as simple as statins preventing chest pain, which they do by facilitating nitric oxide's vasodilating action (just Pubmed "statin nitric oxide"). Anyone in the USA with chest pain is very likely to get a "soft end point" procedure, as per the JUPITER study. Less chest pain, less Dr visits, less procedures. Dr Davis explains how Dr visits lead to procedures.

So, in summary, rosuvastatin has a minor effect on all cause mortality and no effect on heart attack deaths in people with elevated hsCRP.

What the news clips claimed was:

"The study [JUPITER] of 17,800 men and women with normal cholesterol levels found a new statin drug cut deaths from heart attacks and strokes"

Note the words "deaths", "heart attacks" and "strokes". There were 3 stroke deaths in the statin group, 6 in the placebo group. Exactly enough to reverse the heart attack effect. Phew!

I missed this snippet initially as I was off making tea and came back to find my wife grinding her teeth. Not so much at statinating the nation, we're pretty resigned to this. No, the problem was the deferential supplication of Sarah Montague (normally an infuriatingly aggressive interviewer) to the medic and statination. She would have torn a politician apart for such economy with the truth.

Peter

Thanks to the THINCS members for the full text and kicking these ideas around so thoroughly.

8 comments:

Chris said...

I have put a complaint in to the BBC about crap journalism - lazily rehashing a press release and not examining the data - and giving free advertising to Astra Zenica.

If we can complain about Jonathan Ross we can complain about this.

Peter said...

Good idea. Didn't get time yesterday but it's done now.

Peter

Lee said...

I have just complained too. The Today team have a long history of poor scientific journalsim. Sometimes they seem proud of their lack of scientific understanding. They are so inept that they cannot ask pertinent questions and often give scientists a free ride to hide their ignorance. Perhaps we should nominate Peter to appear as a suitably qualified guest to criticise the information.

Bryan - oz4caster said...

Thanks for the detailed info Peter. I wonder what would have happened if they had a third high CRP group that was treated with a diet that minimized sugar and polyunsaturated fat? Maybe fewer deaths than the statin group? Too bad there's not money to pay for studies on dietary effects. Improving diet instead of treating with drugs would save us all too much money :)

Peter said...

Bryan, the discussion of this paper just goes on and on and I guess you've seen Eades' post? I just find it mind boggling that anyone can describe this study as proving anything.

Peter

Bryan - oz4caster said...

Thanks for the tip Peter. No I hadn't seen what Eades wrote yet. I just went and looked. LOL!

JUPITER = "Justification for the Use of Statins in Prevention: an Intervention Trial Evaluating Rosuvastatin. Which translates to by God we’re going to prove that statins prevent something. We certainly know the mindsets of the people running this study."

Larry the Lipidologist said...

I recently talked at the National Lipid Association meeting in Florida and discussed the number of fatal heart attacks with crestor vs placebo and then was fired by the NLA as host of the successful XMRadio show Lipid Luminations on ReachMD XM 160. Do not try to spread any truth if you want to keep your job. These guys are ruthless

O Primitivo said...

Some JUPITER statistics: http://www.rxfiles.ca/rxfiles/uploads/documents/Lipid-Jupiter-trial-overview.pdf