Depending on how you slice the epidemiological pie, the use of cholesterol-lowering statins could appear to decrease, increase, or do nothing to the risk of developing Alzheimer disease, says a study published this week in the Archives of Neurology.

Using data from 2,798 older adults in the Cardiovascular Health Cognition study, Thomas Rea and colleagues at the University of Washington in Seattle and collaborators in several US cities found no change in the risk of AD or vascular dementia between people who had ever used statins compared to those who had never used the drugs. But a different analysis, comparing current users and never-users, gave a different result: The risk of dementia was cut in half in the current users. Yet another comparison, between former users compared to never-users, revealed a higher risk of dementia with statin use.

The results are consistent with previous studies showing that statin treatment is associated with a lower risk of AD when disease incidence is measured at one time in users vs. non-users (the case-control or cross-sectional study design), but that the benefit evaporates under the kind of prospective, follow-up design favored by Rea et al. (see ARF related news story). One proposed explanation for the inconsistent findings is that current statin use could, in fact, be a marker for good health among elders, while discontinuing statin use indicates declining health. The results provide a cautionary tale about the limitations of epidemiological studies and lead the authors to conclude that “additional investigation is needed to determine whether and for whom statin may affect dementia risk.

The gold standard of clinical investigation is, of course, the prospective, controlled trial, of which two small versions have yielded promising results for statins (see ARF related news story and Simons et al., 2002). The results of two larger studies, now underway, are eagerly awaited.—Pat McCaffrey


  1. The new paper raises legitimate questions regarding the potential for artifactual associations emerging from epidemiological studies. My position remains cautiously optimistic because of the faint but positive signal emerging from the Sparks et al. trial (see ARF related news story). Randomized, double-blind placebo-controlled clinical trial data trump epidemiological data every time. The size of the Sparks et al. study (

  2. The cholesterol and statin story in AD has been a never-ending battle since its inception in the late 1980s, and the current paper sends a
    mixed message. It seems that if the authors exclude the final year of
    medications from consideration, there is no reduced hazard risk (HR), but if the final year of current statin use is included in the
    analysis, there is a near significant or significant (for AD with or without vascular factors) reduction in the hazard ratio. One must also consider that an
    individual who may have taken a statin for, say, 1 month would be
    included in the "ever statin use." I would suggest the take-home message
    may be that longer exposure to statins produces a reduced risk of AD later
    in life.

    I am sure that the statin story with regard to treatment of AD will be
    sorted out by the results of LEADe and CLASP: the two large multicenter
    trials testing atorvastatin and simvastatin, respectively. The way to
    determine the effect of statins on prevention of AD (reduced risk) is to
    directly test for benefit in a double-blind, placebo-controlled prevention
    trial of 6-10 years' duration. In advance of expending the time and monies
    to perform such a prevention trial, I would suggest identifying benefit of
    statin therapy in delaying the transition from MCI to AD in a double-blind,
    randomized placebo-controlled MCI treatment trial would strongly support
    possible success in a prevention trial.

  3. The recent epidemiological study by Rea and colleagues adds yet more complexity (and confusion) to the issue of statin use and AD risk. It’s difficult to draw any firm conclusions from the study, since the reported outcomes vary so distinctly as a function of analysis parameters. The gold standard will always be double-blind, case-controlled studies, and for good reason. The results from the statin clinical trial(s) in the pipeline will hopefully shed more light on this important issue.

    The Rea study does, however, bring to light a couple of general issues (some of which have been discussed previously on Alzforum) that may or may not be resolved in the upcoming prospective clinical trials (e.g., CLASP). If statin use indeed influences AD risk, what duration of use is needed to achieve the effect? I don’t think the “ever use” versus “never use” in the Rea paper is useful in sorting this out. And perhaps more importantly, when do statins need to be taken in order to achieve proposed protection? AD pathology is known to begin years, perhaps decades, prior to cognitive symptoms. Is it mid-life use that protects against future late-life dementia, or will late-life statin use also be of some benefit? By enrolling only demented patients (mild and moderate), the CLASP study will be able to address the issue of efficacy of statin use in preventing clinical progression, but what about preventing or slowing pathologic progression prior to symptoms? This is a much more difficult question to answer and even test. I think it’s still unclear from the human literature whether statins are able to influence the disease process itself or if they merely influence the symptoms. While we’ll take any protective effect we can get, finding an AD treatment that would stop or even slow disease progression at its earliest stages (especially preclinically) is clearly the ultimate goal.

  4. Recently there has been much debate as to whether statin therapy offers a benefit for Alzheimer disease (AD), and whether statins reduce AD incidence and/or progression remains an open question (Jick et al., 2000; Wolozin et al., 2000; Shepherd et al., 2002; Zandi et al., 2005; Sparks et al., 2005). The prospective cohort study by Rea and colleagues is certainly interesting, and several important factors are brought into consideration, including analysis of the effects of statin use duration, the type of statin used (lipophilicity profile) and patient characteristics. Most importantly, however, this study demonstrates how analysis of the same data set in two different ways can lead to diverging conclusions. Their analysis indicates that antecedent statin use in the population of elderly patients examined was not associated with a lower risk of dementia when primary analysis incorporated a 1-year lag. However, if the data is analyzed in a way similar to that of case-controlled studies, whereby analysis was based on current statin use compared to non-use, without a lag period, statin use appears somewhat beneficial in protecting against dementia. Thus, this study not only indicates that different results are obtained when a lag period is incorporated into the study design, but also ultimately questions whether statins affect the development of dementia at all.

    However, it should be considered that AD is a slow and insidious disease, with abnormal physiological alterations likely preceding the clinical manifestations of the disease by several decades. Indeed, familial AD (FAD) mutations are known to elevate the levels of Aβ42 and thus, while patients with FAD likely overproduce Aβ42 from an early age, clinical symptoms are not typically observed until the third to fifth decades. In the current study, participants had a mean age of 75 years and, as discussed, it may well be the case that statin exposure much earlier in life may be required in order for these drugs to beneficially affect dementia risk. In addition to examining a younger population, it may also be of benefit to extend the lag period; given the insidious nature of AD, a year may not be sufficient for statins to beneficially affect the underlying pathophysiology.

    In summary, while there is no doubt that this study provides important insights for the analysis of cohort studies, future prospective studies involving the analysis of a younger population and incorporating longer lag times in the study design would be informative. Ultimately, to provide the definitive answer to the question, are statins protective in AD?, it may be necessary to perform primary prevention studies to determine whether continuous long-term statin use in mid-life (long before the onset of dementia) could prevent or significantly delay AD.

  5. I disagree with an earlier comment that this study found that a greater duration of statin use reduces the risk of Alzheimer's. The authors stated, for all-cause dementia, that the hazard ratio (HR) for less than 1 year was .98, 1.41 for 1 to 3 years, and 0.74 for greater than 3 years.

    It seems very clear to me that what happened is all those who were experiencing the worst side effects from statins dropped out of statin use in the 1-to-3-year range. Statin use may have triggered dementia in these people. Since those experiencing adverse effects were weeded out during this period, naturally the >3 years will reflect the most positive results, simply because only those who can tolerate statin use and not develop dementia and other adverse side effects will stay with the treatment!

    As I comment in the most recent issue of my newsletter, the fascinating take-home message of this study is not the results comparing ever-use to current use, but the massive increase in dementia risk in the former use category, where all-cause dementia was nearly doubled at an 88 percent increase and Alzheimer's was much more than doubled, being multiplied by a factor of 2.54.

    The authors explain this away by suggesting that doctors and patients would not be interested in using statins if they see health deteriorating, and thus that those with the worst health were the most likely to quit statin use.

    It seems much more likely that patients would opt to drop statin use because they were experiencing adverse effects. Most people whose health is declining would prefer to receive treatment for it. On the other hand, people who are perceiving declining health to be attributable to their treatment are likely to discontinue that treatment.

    I find it fascinating that the authors did not even entertain the possibility that the statin use in and of itself could have triggered the dementia in the portion of patients who chose to discontinue its use.

    View all comments by Chris Masterjohn

Make a Comment

To make a comment you must login or register.


News Citations

  1. Philadelphia: All Is Not Well with the Statin Story
  2. Atorvastatin, Vaccine Trial Data Published

Paper Citations

  1. . Treatment with simvastatin in normocholesterolemic patients with Alzheimer's disease: A 26-week randomized, placebo-controlled, double-blind trial. Ann Neurol. 2002 Sep;52(3):346-50. PubMed.

Further Reading


  1. . Statins--a cure-all for the brain?. Nat Rev Neurosci. 2005 Apr;6(4):325-31. PubMed.
  2. . Statins cause intracellular accumulation of amyloid precursor protein, beta-secretase-cleaved fragments, and amyloid beta-peptide via an isoprenoid-dependent mechanism. J Biol Chem. 2005 May 13;280(19):18755-70. PubMed.
  3. . Modulation of statin-activated shedding of Alzheimer APP ectodomain by ROCK. PLoS Med. 2005 Jan;2(1):e18. PubMed.

Primary Papers

  1. . Statin use and the risk of incident dementia: the Cardiovascular Health Study. Arch Neurol. 2005 Jul;62(7):1047-51. PubMed.