Why science reporters should do their homework

One of the most significant medical advancements of the last few decades has been the use of cholesterol-lowering medications called statins.  These drugs, when used properly, have been shown over and over to lower the risk of heart attacks, strokes, and death.  But like all drugs, they have many effects, both those we like (preventing heart attacks) and those we don’t (in this case, rare liver and muscle problems); the latter we call “side-effects”.  Studies done on drugs before they hit the market can identify common side-effects, but it’s not until many more people are exposed for a long period of time that rare side-effects show up.

A recent Scientific American article wondered if one of these rare side-effects could be memory problems.  At first glance, the idea seems pretty improbable, but the SI article takes some sketchy anecdotes and runs with the idea, managing to cobble together an interesting hypothesis:

It is not crazy to connect cholesterol-modifying drugs with cognition; after all, one quarter of the body’s cholesterol is found in the brain. Cholesterol is a waxy substance that, among other things, provides structure to the body’s cell membranes. High levels of cholesterol in the blood create a risk for heart disease, because the molecules that transport cholesterol can damage arteries and cause blockages. In the brain, however, cholesterol plays a crucial role in the formation of neuronal connections—the vital links that underlie memory and learning. Quick thinking and rapid reaction times depend on cholesterol, too, because the waxy molecules are the building blocks of the sheaths that insulate neurons and speed up electrical transmissions.

It’s not crazy to connect cholesterol-modifying drugs with cognition, but it’s quite a stretch.  We do know that statins affect the central nervous system.  They’ve been proven to reduce the risk of stroke, a devastating central nervous system disease.  If they can prevent brain disease, might they also cause it?  We have some ideas about why statins prevent strokes: they lower cholesterol and stabilize arterial plaques, perhaps by reducing inflammation in these plaques.  They can even cause plaques in some arteries to shrink.  Is there a plausible hypothesis as to why statins might cause memory problems?  What is being posited is that statins actually reduce cholesterol levels so much that cell membranes are damaged and neuronal saltatory conduction* is impaired.  If this were the case, we might also expect to find cognitive differences  when comparing people with high and low cholesterol levels, or to see cognition affected by cholesterol-lowering diets.  This is not the case.

Still, dementia—the most common and severe form of memory loss— is a devastating disease, so if there is even a chance, maybe we should ask the question.   A large  cohort study published in Archives of Neurology in 2005 looked into whether statins might actually help prevent dementia.  They groups of elderly patient who took statins, and those who did not and compared the incidence of dementia in each group. There found neither a protective effect nor a harmful effect.

The idea that lipids (fat molecules) can affect brain function has been supported by certain epidemiologic studies and some animal models.  Omega-3-fatty acids have been touted for possible use in preventing and treating dementia.  Last week, a randomized controlled trial of a particular omega-3-fatty acid was published in JAMA.  The study design was strong, and the study found no evidence that this particular molecule helped dementia patients.

The two most common types of dementia are vascular dementia and Alzheimer’s disease.  The cause of Alzheimer’s disease isn’t known, making prevention difficult.  Vascular dementia, however,  is to a certain extent preventable.  It is caused by a variety of factors that affect blood vessels such as hypertension, and studies have shown that many of the same interventions that prevent stroke can help prevent vascular dementia.  One of the most potent risks for vascular disease is cigarette smoking, so it would make sense that smoking would be a risk factor for vascular dementia.  A surprising result of a study recently published in Archives of Internal Medicine was that smoking is a risk factor not only for vascular dementia but also for Alzheimer’s dementia.

The story of dementia risk is complex, and there is a rich vein of literature to mine.   I was disappointed that the SI article presented anecdotes rather than data, case-reports rather than good studies, and highlighted “experts” who presented fear-mongering testimony rather than the measured caution that we can expect from real experts.


*“Saltatory conduction” describes a way that nerve signals travel quickly.  Nerve cells can function as a sort of wire for electrical signals, and the myelin sheath allows electrical signals to jump from node to node, increasing the speed of conduction when compared to an un-myelinated neuron.  Certain diseases, such as multiple sclerosis, involve destruction of the myelin sheath, decreasing nerve conduction velocity, leading to weakness and other symptoms.    Myelin contains cholesterol, among other things.


Rusanen, M., Kivipelto, M., Quesenberry, C., Zhou, J., & Whitmer, R. (2010). Heavy Smoking in Midlife and Long-term Risk of Alzheimer Disease and Vascular Dementia Archives of Internal Medicine DOI: 10.1001/archinternmed.2010.393

Quinn, J., Raman, R., Thomas, R., Yurko-Mauro, K., Nelson, E., Van Dyck, C., Galvin, J., Emond, J., Jack, C., Weiner, M., Shinto, L., & Aisen, P. (2010). Docosahexaenoic Acid Supplementation and Cognitive Decline in Alzheimer Disease: A Randomized Trial JAMA: The Journal of the American Medical Association, 304 (17), 1903-1911 DOI: 10.1001/jama.2010.1510

Rea, T. (2005). Statin Use and the Risk of Incident Dementia: The Cardiovascular Health Study Archives of Neurology, 62 (7), 1047-1051 DOI: 10.1001/archneur.62.7.1047

Forette F, Seux ML, Staessen JA, Thijs L, Birkenhäger WH, Babarskiene MR, Babeanu S, Bossini A, Gil-Extremera B, Girerd X, Laks T, Lilov E, Moisseyev V, Tuomilehto J, Vanhanen H, Webster J, Yodfat Y, & Fagard R (1998). Prevention of dementia in randomised double-blind placebo-controlled Systolic Hypertension in Europe (Syst-Eur) trial. Lancet, 352 (9137), 1347-51 PMID: 9802273

Posted in: Science and Medicine

Leave a Comment (17) ↓

17 thoughts on “Why science reporters should do their homework

  1. kirkmc says:

    Hold on a second…

    “We do know that statins affect the central nervous system.””

    Saying that they affect the central nervous system because the prevent strokes is like saying that eating too much affects your pancreas because obesity can cause diabetes…

    Statins only act on the circulatory system; by doing so, they can prevent strokes, as you go on to explain. But they don’t in any way affect the central nervous system.

    And this in a post about how science reporters should do their homework…

  2. Peter Lipson says:

    One, I’m giving them the benefit of the doubt. Two, strokes can be hemorrhagic or thromboembolic. Among thromboembolic strokes, many are “in situ” strokes, not caused by thromboemboli of distant origin. Vascular dementia is caused mainly by, well, vascular disease of the brain.

    Statins have a significant impact on CNS disease–they affect the CNS.

    Whether they also affect the CNS in the other ways posited by the article, I doubt.

  3. daedalus2u says:

    Kirkmc, you have no basis for saying statins don’t affect the CNS until there is a pretty conclusive research that looks at that particular subject in detail and finds there actually is no effect. In the absence of data our default has to be “we don’t know”.
    There are many effects of statins that are not mediated through cholesterol synthesis.

    It isn’t even clear that the major therapeutic effects on reducing cholesterol is mediated through inhibition of cholesterol synthesis. One of the major effects of statins is an upregulation of eNOS and increased NO availability. More NO would certainly affect the CNS, and there would be no threshold for increased NO having an effect because NO is already being actively used as a control signal by the CNS. More NO adds to that signal and changes the outcome. Knowing that statins affect NO synthesis, our default can’t be that statins have no effect on the CNS.

    However, I think the idea that statins would have neurological effects mediated through cholesterol reductions is (I think) extremely unlikely. It isn’t like statins reduce cholesterol to zero, and it isn’t like there isn’t any cholesterol synthesis still going on, and it isn’t like there are no dietary sources of cholesterol. The anecdotes that were used relate to prompt effects of statins. These are unlikely to be due to changes in cholesterol levels because those can’t change promptly.

    In any case more NO would likely improve memory not worsen it.

    However, statins do interfere with synthesis of coenzyme Q-10. Reductions in that would cause mitochondrial problems which always show up as more superoxide. More superoxide would cause adverse cognitive effects (usually), and enough certainly would cause dementia. Prompt dementia due to superoxide would likely be completely reversible if the superoxide source were removed. If it isn’t removed promptly, then there will be neurodegeneration and the defects will become permanent. That likely takes months or longer (depending on the severity).

  4. tuck says:

    “Sketchy anecdotes”?

    Dr. Graveline is also a medical doctor and a NASA flight surgeon, which is not mentioned in the article, but 2 seconds on Google reveals it:

    He also doesn’t claim that statins cause dementia: in his case it was amnesia, which passed, but occurred again after he went back on statins some time later.

  5. kirkmc says:

    “Kirkmc, you have no basis for saying statins don’t affect the CNS until there is a pretty conclusive research that looks at that particular subject in detail and finds there actually is no effect. In the absence of data our default has to be “we don’t know”.”

    So should the author of the article have said, “We do know that statins affect the central nervous system.”?

  6. Calli Arcale says:

    It was unclear, but I believe he meant that they affect it indirectly by affecting plaques in the vascular system, as he explained in his reply above.

    Tuck was right to point out that the anecdote has nothing to do with dementia; it’s about temporary amnesia. The article itself did not tie dementia to statins; the very title says “It’s Not Dementia” (emphasis mine). While I agree that science reporters should do their homework, I’m not sure this was the best article to hold up as an example of sloppy reporting.

  7. Peter Lipson says:

    Reasonable people can disagree, but given the title of the post, and this:

    they are calling for increased public awareness of the possible cognitive side effects of statins—symptoms that may be misdiagnosed as dementia in the aging patients who take them

    There was an impression left that statins can cause dementia-like memory deficits (the temporary amnesia story notwithstanding). There is also considerable buzz in the less-than-reputable altmedosphere about statins “causing” dementia, which i didn’t link to.

  8. transposition says:

    It should be clear to all that statins have far ranging effects beyond simply lowering serum cholesterol. In part, statins disrupt intracellular signaling mediated through Ras and related proteins through inhibition of covalent binding to cholesterol, farnesyl, and related moieties.

    A fair amount of evidence suggests that they may have a role in treating the CNS effects of type I neurofibromatosis. In fact, a clinical trial (phase I) for their use is underway.

    Please, let’s think before we post—there’s a lot out there that is known that we may not be aware of personally.

  9. Peter Lipson says:

    I should add ( or should have in the post) that while this article was not good (including lots of anecdotes to support a hyperbolic hypothesis and a failure to examine the existing evidence), it was meant to be seen in a broader context of how to look at similar questions, and to present some recent interesting scientific articles.

    As someone who writes daily, I’m not always going to hit the mark with all my readers, and I’m quite open to legitimate criticism.

  10. daedalus2u says:

    Kirkmc, the article should not have used as its title the statement that “It’s Not Dementia, It’s Your Heart Medication: Cholesterol Drugs and Memory”.

    Whether a particular drug can have a side effect, or does have a side effect is an idiosyncratic interaction between a particular patient and a particular drug. I don’t find it helpful to make inaccurate and imprecise statements. Blanket statements (which happen to also be wrong) such as “Statins only act on the circulatory system; by doing so, they can prevent strokes, as you go on to explain. But they don’t in any way affect the central nervous system.” are not helpful.

    A statement that statins do affect the CNS is more accurate that a statement that statins do not affect the CNS.

    The topic of the article was dementia and statins. Affecting the CNS and causing dementia are two completely different things. Not everything that affects the CNS causes dementia but everything that does cause dementia has to do so by affecting the CNS. Some things can affect the CNS, cause dementia and prevent dementia but at different doses. Statins may prevent vascular dementia at one dose in some individuals and may cause mitochondrial mediated dementia via reduced coenzyme Q-10 synthesis (which by the way is not an “antioxidant”, it is a charge carrier in the respiration chain) in other individuals at another dose.

  11. Woody says:

    It is not unreasonable to postulate a link between statins and cognitive symptoms, based on the wealth of published data regarding the impact of Apolipoprotein E (a key cholesterol trafficking protein in the central nervous system) genotype on risk of developing dementia. The E4 genotype in particular confers a greatly increased risk of developing Alzheimer’s dementia, with additive risk in E4 homozygotes (those carrying 2 E4 alleles).

    Since this cholesterol trafficking protein clearly influences risk of developing dementia, and cardiovascular risk factors increase risk of developing dementia, the hope among some researchers was that drugs impacting cholesterol biosythesis might modulate the APOE effect or even other as yet unknown CNS effects of elevated cholesterol, leading to clinical trials to test the hypothesis:

    In short, the results thus far are underwhelming for a positive or negative impact of statins on dementia. There are too many unanswered questions. For example, does the APOE effect on dementia risk have anything to do with cholesterol trafficking or is it a separate mechanism? Does high cholesterol impact risk of developing dementia via atherosclerosis or some other mechanism? Perhaps there is a benefit of statins for dementia, but maybe they need to be administered earlier in life to reap dividends. A primary prevention study is impractical given the numbers of individuals and the length of follow-up required. Not to mention it would be unethical to randomize individuals with high cholesterol to a placebo arm since we know definitively that statins reduce risk of cardiovascular events in such individuals.

  12. Brett says:

    I agree with Woody that one of the best papers on this issue is the Cochrane review on statins and dementia – – which found no effect in either direction on dementia.

    I don’t really agree that “it would be unethical to randomize individuals with high cholesterol to a placebo arm since we know definitively that statins reduce risk of cardiovascular events in such individuals.” The effectiveness of statins is not very well researched in significantly older adults. As far as I am aware, no randomised controlled trials of statins that looked at clinically meaningful endpoints in primary prevention have enrolled subjects over age 82 (the upper age limit for the PROSPER trial – ). That study was controversial as cardiovascular benefit seemed offset by an increase in cancer incidence. As Mangin et al point out, this effectively meant that pravastatin in the elderly changed the mode of death for some people from cardiovascular disease to cancer ( ). Thus, the question of whether statins have overall benefits in primary prevention in the elderly is still open. I think equipoise exists on this issue, and a trial of a statin other than pravastatin could ethically be performed in an elderly population. Cognitive outcomes could be prespecified as secondary endpoints.

  13. tmac57 says:

    “As Mangin et al point out, this effectively meant that pravastatin in the elderly changed the mode of death for some people from cardiovascular disease to cancer ”
    Similarly,an elimination of ALL disease and accidental deaths would only lead to an increase in people dying from old age.Good catch!

  14. weing says:

    I recall Aubrey de Grey saying, on a recent SGU interview, that if we got rid of all cardiovascular disease, our lifespan would increase by 5 years. I don’t recall him giving a reference for it. But it sure sounds depressing, however, it may shed light on why if you prevent cardiac deaths, you end up increasing cancer deaths.

  15. daedalus2u says:

    Weing, an important reason is because the “chance” of death is 100%. Reducing the cause of death by one mechanism necessarily increases it by another.

  16. Woody says:

    @Brett – I agree that a statin primary prevention study for dementia with randomization in the very old might be reasonable. The biggest concern would be whether that would be too late to impact the earliest pathologic changes of dementia which likely occur years before the onset of any recognizable cognitive symptoms. If there is a statin effect on dementia risk, perhaps it has to take place earlier in life, years (or even decades) before the age at which dementia incidence starts to pick up. If you started the primary prevention study in individuals in their 60s, you might have to follow the randomized groups for two decades before you could detect enough incident cases to run your statistics, which is why primary prevention studies in dementia are impractical. What we really need is a more reliable biomarker that identifies those destined to develop dementia long before they do, at which point a prevention trial would be more feasible.

  17. daedalus2u says:

    In the PROSPER trial, they lumped together trauma and suicide (which was increased in the placebo leg). Depression is highly correlated with neurodegenerative diseases which are also highly correlated with vascular disease.

    I consider it quite likely that the statins do reduce dementia via their increase in NO levels. I think the increased incidence of cancer is not due to increased cancer causation, but rather to increased growth of tumors present at the start of the trial (which is why there was a big increase in the first year and not in following years).

    A different dosage scheme might prevent the growth of tumors. Instead of the same dose every day, suspending treatment for a period of time sufficient to get “wash out” plus a few weeks every few months might prevent the continuous growth effects of NO and allow for immune system disposition of those tumor cells.

Comments are closed.