Skeptics Stack Exchange is a question and answer site for scientific skepticism. It's 100% free, no registration required.

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

Ben Goldacre, in his sustained critique of the way the pharmaceutical industry and its regulators fall far from the standards of reliable evidence (Bad Pharma), makes an astounding claim (p193 of the paperback, my emphasis):

Trials have been done comparing statins against placebo, and these have shown they save lives rather well. Trials have also compared one statin to another: but these all use cholesterol as a surrogate outcome. Nobody has ever compared the statins against each other to measure which is best at preventing death. This is a truly staggering oversignt when you consider that tens of millions of people have taken these drugs and for many, many years.

There are two pieces of background information that should provide the background required to see why this is an astounding claim.

The first is that statins, designed to reduce cholesterol levels (for a more technical description read the wikipedia entry), are very widely prescribed and make a huge amount of money for the pharmaceutical industry. It has even been suggested that they should be widely used in healthy populations (see BBC news story for background and discussion). They have been the single largest cost category of prescribed drugs in the NHS (2007 BBC story). Total worldwide sales are estimated to exceed $35 billion.

The second is that proxy or surrogate outcomes are not a reliable measure of the effectiveness of a drug. Surrogate outcomes are usually something that is easy or quick to measure (in the case of statins, cholesterol levels). They are used to judge results in trials of drugs because the ultimate goals (the primary outcome) are often harder to measure and take longer to measure (for statins, cardiovascular events and death). The trouble is surrogate outcomes are often misleading. For example, it was long thought that drugs that reduce irregular heart rhythm were good for people who have survived heart attacks as they often have irregular rhythm. But this was shown to be wrong by the famous CAST trial (wikipedia, review article): it turned out that drugs reducing irregular rhythm (the proxy) cause significantly more heart attack deaths (the primary outcome the drugs were intended to prevent).

So, while I would normally just take Ben Goldacre on trust as an expert, this claim is so bold it deserves separate skeptical analysis. He claims there have never been proper trials of one of the biggest classes of drugs that measures what they are intended to achieve.

So has there never been a trial comparing statins using the primary outcome of fewer deaths from cardiovascular events?

share|improve this question
Excellent example of questioning a sceptic. Though I have to mention the caveat that I’m unsure what answer you’d find sufficient. An exhaustive review of Pubmed articles mentioning statins? Oh my … – Konrad Rudolph Nov 18 '12 at 17:43
@KonradRudolph Fair point. But in the case of this claim I might be prepared to accept independent confirmation from other experts. A full meta-analysis of trials would be nice, but I suspect there isn't one. We can dream, though! – matt_black Nov 18 '12 at 19:11
I find it to be a plausible claim. In clinical trials, we attempt to find proximal biomarkers that predict the health outcomes of interest and allow us to conduct a study with reasonable time and expense. Because high cholesterol does not predictably kill people in a short time frame, any study would take a long time, longer than many academic researcher's careers. And the variability in causes of death that you see in a decades long study would mean you'd need a huge sample size. – MattBagg Nov 19 '12 at 4:52
lower cholesterol reducing the risk of a heart attack is well researched and well documented, I don't think maintaining regular heat rhythm ever had a similar body of evidence, which makes their comparison somewhat misleading. – Ryathal Nov 19 '12 at 15:13
@Ryathal The CAST trial isn't the only example where the proxy measure is misleading. But I think it is a fair example given the knowledge at the time. The ALLHAT results for doxazosin (perfectly good at lowering blood pressure, but many more heart attacks) would be another where the proxy is seriously misleading. – matt_black Nov 19 '12 at 17:37

There has not been any statin vs statin comparisons, but there was a comparison of different dosages of the same statin. 10,001 patients were either assigned to treatment with either 10 mg of atorvastatin or 80 mg of atorvastatin. The primary end point was the occurrence of a first major cardiovascular event, defined as death from CHD, nonfatal non-procedure-related myocardial infarction, resuscitation after cardiac arrest, or fatal or nonfatal stroke. The results showed that 80 mg was superior. [1]

A primary event occurred in 434 patients (8.7 percent) receiving 80 mg of atorvastatin, as compared with 548 patients (10.9 percent) receiving 10 mg of atorvastatin, representing an absolute reduction in the rate of major cardiovascular events of 2.2 percent and a 22 percent relative reduction in risk (hazard ratio, 0.78; 95 percent confidence interval, 0.69 to 0.89; P<0.001).

and the conclusion:

Intensive lipid-lowering therapy with 80 mg of atorvastatin per day in patients with stable CHD provides significant clinical benefit beyond that afforded by treatment with 10 mg of atorvastatin per day. This occurred with a greater incidence of elevated aminotransferase levels.

This shows, ceteris paribus, that with respect to LDL-c, lower is better.

Ben Goldacre is concerned that should there be just a 2% difference between the different drugs, there could be vast numbers of lives not saved due to choosing the wrong statin. [2]

A systematic review study protocol 2010 looked at these issues and confirmed [3]

This is because of the fact that most randomized clinical trials have not tested different statins head-­‐to-­‐head. Additionally, almost all of the meta-­‐analyses ‘lumped’ all statins together as one intervention

And they state also

Combination of direct and indirect evidence: Methodological advances in statistical synthesis approaches, called mixed treatment comparisons (also known as network meta-analyses), facilitate the combination of direct and indirect evidence by incorporating both direct (when statins are compared to each other within a trial) and indirect comparisons (when statins are compared between trials with a common comparator treatment, which is often placebo).

So, they claim that statistic methods make head to head comparisons unnecessary.

The answer to the question is, so far there have not been any head to head trials with cardiovascular endpoints.

[1] LaRosa JC, Grundy SM, Waters DD, Shear C, Barter P, Fruchart JC, Gotto AM, Greten H, Kastelein JJ, Shepherd J, Wenger NK, . Intensive lipid lowering with atorvastatin in patients with stable coronary disease. N. Engl. J. Med. 2005 Apr;352(14):1425-35. doi: 10.1056/NEJMoa050461. PubMed PMID: 15755765.



share|improve this answer
Does a comparison of two dosages of the same drug count as a comparison against other drugs? Or is there a typo? – Bobson Jun 6 '14 at 13:59
@Bobson The explanation as to why a same drug comparison is used is given in the answer. No typo. – HappySpoon Jun 6 '14 at 22:51
You start out saying Actually there has been a statin vs statin comparison, but conclude so far there have not been any head to head trials. I suggested an edit which clears it up. – Bobson Jun 8 '14 at 14:52
@Bobson I intended it to be a little confusing so that the reader would think more about whether head to head trials were actually necessary. Oh well ... – HappySpoon Jun 9 '14 at 9:43

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.