More on Cholesterol, Statins, and JUPITER

Now that we've had something of a vacation from the cholesterol-statin-prevention controversy, we can revisit it thanks to two articles in Archives of Internal Medicine. The more important one is by an international team headed by Dr. Kausik K. Ray of Cambridge. You can view this study as a sort of reply to an earlier meta-analysis of statins for heart disease prevention, which I summarized in this post:
http://brodyhooked.blogspot.com/2009/07/more-on-statins-new-bmj-meta-analysis.html

The previous meta-analysis (by Brugts et al.) concluded that statins, used in high-risk patients without existing cardiovascular disease, do prevent death and vascular disease, but with a high number needed to treat (NNT). This, I noted, was compatible with previous studies. The newer meta-analysis takes issue and notes that Brugts included in their analysis a number of subjects who actually already had existing vascular disease. The analysis focuses on studies that report patients clean of existing disease (primary prevention), and looks at all-cause mortality as the endpoint of concern. (One of the studies they include is JUPITER, of which more below.)

Ray and colleagues then crunch the numbers (from 11 studies looking at 65,229 subjects for 244,000 person-years) and conclude that there's no statistically significant drop in all-cause mortality from prescribing statins for primary prevention of cardiovascular disease.

If you go back and look at the individual studies, most of them were negative, but the major outlier was JUPITER. I previously addressed JUPITER in several posts:
http://brodyhooked.blogspot.com/2008/11/by-jupiter-slick-drug-marketing-great.html
http://brodyhooked.blogspot.com/2008/11/by-jupiter-part-ii-more-skeptical-view.html
http://brodyhooked.blogspot.com/2008/12/more-on-jupiter-trial-and-media.html
http://brodyhooked.blogspot.com/2010/04/forget-science-docs-already-know.html

In the oldest of these posts, I argued that JUPITER may in fact be valid in claiming that if you gave a statin to a group of patients thought to be high-risk for heart disease because of having a positive C-reactive protein test even though they had normal cholesterol levels, you'll see reduced heart attacks and reduced deaths (though again with a high NNT). (Later I had to modify that rosy view as more criticisms of JUPITER came forward.) But the main message I thought had been completely missed by the media is that JUPITER knocked a huge hole in the "lipid hypothesis" of cardiovascular disease-- that statins reduce risks of bad stuff happening by means of reducing "bad" cholesterol. By standard cholesterol guidelines, no one in JUPITER needed to take a statin.

The new article in Archives that further beats up on JUPITER is by another international team, this time led by Dr. Michel de Lorgeril of Grenoble, and including our good friend Dr. John Abramson (of Overdo$ed America fame). A lot of their criticisms are too technical to make very interesting reading, but they note several features about the numbers that raise serious questions of bias. Just to grab one example, they note that in a comparable population, the death rate among people with serious cardiovascular disease is about 40%; but among patients in JUPITER, it can be calculated to be between 5 and 18%. Either these were very unusual patients, or something is fishy. They also note that the JUPITER trial was stopped early, presumably because of a previously selected statistical tripwire. The exact reasons for stopping the trial early were never explained, and close inspection of the data shows that when the trial was stopped, the data in the treatment and placebo groups were converging, suggesting that the statistical difference would have been erased had the trial been continued longer. When trials are stopped early, the effect size reputed to the drug treatment is generally bigger than otherwise. Finally, they claim that while the head of the data safety and monitoring board that decided when to stop the trial was said to have been independent of company sponsorship, the same person has served in that capacity in several commercially-sponsored drug trials.

In short, de Lorgeril at al. claim that evidence of bias due to commercial funding shows through in a number of aspects of the JUPITER trial.

(If you want a brief recap of these facts, plus word of a new report showing that statins might help prevent recurrence in men who have had surgery for prostate cancer, see Melissa Healy's article in the Los Angeles Times: http://www.latimes.com/news/science/la-sci-statins-20100629,0,7388273.story)

Ray KK, Seshasai SRK, Erqou S, et al. Statins and all-cause mortality in high-risk primary prevention: a meta-analysis of 11 randomized controlled trials involving 65,229 participants. Archives of Internal Medicine 170:1024-1031, June 28, 2010.

de Lorgeril M, Salen P, Abramson J, et al. Cholesterol-lowering, vardiovascular diseases, and the Rosuvastatin-JUPITER controversy. Archives of Internal Medicine 170:1032-1036, June 28, 2010.