Can't seem to find this article to link it, but I found it in my library. It's actually just a comment, by J Abramson and JM Wright from Harvard Med School and the Univ of British Columbia Dept of Anesthesiology.

They reviewed the 8 big randomized trials of 47,925 patients comparing statin with placebo in "primary prevention populations at increased risk". But they admit their analysis is "imperfect because these trials are not solely primary prevention. 8.5% of patients had occlusive vascular disease at baseline". 6570 patients had had a stroke, or peripheral vascular occlusion, but they hadn't had a cardiac event, so they were considered primary prevention. But they were at much higher 5-year risk for a cardiac event (25-30%) than the true primary prevention group (9%), and that lumping them together made statins seem more beneficial.

They basically said these studies "cheated"; they weren't really primary prevention, because these people in question had the "equivalent" of heart disease. Despite this cheating, there still wasn't statistically significant proof that statins helped any woman, or anyone over age 69.

The main complaint was that the "Cholesterol Treatment Trialists collaboration (CTT) have access to all the data, and they could re-analyze the existing studies to find out just how inneffective statins were for the truly primary prevention patients. But they haven't done it. The CTT is a group that does periodic meta-analysis of all the available trials, and they are aparently shirking their jobs.

The article estimates that if it's shown that statins have no benefit (just risk!) for primary prevention patients, about 3/4 of all statin recipients could stop their drugs.

There is an interesting disclaimer though. A footnote states that Dr. Abramson is "an expert consultant to plaintiff's attorneys on litigation involving the drug industry, including Pfizer for it's marketnig of atorvastatin". I guess this could tarnish the credibility of this piece. In the eyes of some