Simvastatin

Simvastatin is the compound that makes the argument of this batch on its own.

The Interventions Testing Program fed it to genetically heterogeneous mice at two concentrations, 12 and 120 ppm, starting at 9 months of age, in the same cohort that produced one of the programme's landmark positives. Rapamycin extended median survival by an average of 10 percent in males and 18 percent in females at each of three sites. Resveratrol did nothing. Simvastatin did nothing, at either dose, in either sex.

The human record is the opposite of thin. In the Scandinavian Simvastatin Survival Study, 4,444 people with angina or prior myocardial infarction were randomised to simvastatin or placebo and followed a median of 5.4 years. Twelve percent of the placebo group died against 8 percent on simvastatin, a relative risk of death of 0.70. In the Heart Protection Study, 20,536 UK adults at high vascular risk were randomised to 40 mg daily or placebo, and all-cause mortality fell from 14.7 percent to 12.9 percent, with major vascular events down about a quarter.

These findings do not contradict each other. Mouse lifespan in a well-housed laboratory colony is not limited by atherosclerotic vascular disease, and a drug that prevents heart attacks in people with coronary disease has nothing to prevent in a mouse that was never going to have one. The lifespan null is informative about the geroscience claim, which is that statins slow ageing itself. It says nothing about the cardiovascular claim, which is settled.

A lifespan null from this programme is a real, expensive, well-powered finding, and it is narrower than it sounds. What was tested was one concentration in the diet, started at one age, in one mouse strain, with death as the endpoint. A null tells you that this dose, in these animals, did not move that endpoint. It does not tell you the compound is inert, that a different dose would also fail, that the mechanism is wrong, or that the reason people actually take it has been refuted. Those are separate questions and most of them were never asked. The failure mode this site is trying to avoid runs in both directions: treating a mouse lifespan positive as proof that a supplement works, and treating a mouse lifespan null as proof that it does nothing.

Simvastatin is the reference case for that paragraph. Somebody reading only the mouse result would conclude statins do nothing. Somebody reading only the human trials would conclude statins are a longevity drug. Both readings are wrong, and the reason is that the two literatures measured different things in populations with different causes of death.

Well-housed laboratory mice do not die of atherosclerotic coronary disease. A drug whose entire demonstrated benefit is preventing that disease therefore has nothing to prevent, and the null was arguably predictable. It is still worth having, because the hypothesis on test was not the cholesterol one. It was that statins have pleiotropic anti-inflammatory effects that slow ageing generally. At the doses used, that hypothesis failed.

The Interventions Testing Program is funded by the National Institute on Aging and run in parallel at three independent laboratories, in Bar Harbor, Ann Arbor and San Antonio. It uses UM-HET3 mice, a four-way genetic cross, so no result can be an artefact of a single inbred background. Both sexes are studied, cohorts are sized to detect roughly a 10 percent shift in lifespan, compounds are fed in the diet from a stated starting age, and the programme commits in advance to publishing negative results alongside positive ones. That last commitment is why this batch can exist at all: almost no other part of ageing research reliably tells you what did not work.