In an article titled "Statins and the Risk of Dementia" published in the November 11, 2000, issue of Lancet, Jick and colleagues1 from Boston, Mass, reported an epidemiologic, observational, nested case-control study of subjects enrolled in the United Kingdom-based General Practice Research Database. Records of patients aged 50 to 89 years were classified into 1 of 3 groups: group 1 comprised individuals who received at least 1 prescription of lipid-lowering agents (LLAs) within the previous 6 months; group 2, all individuals with a clinical diagnosis of untreated hyperlipidemia; and group 3, a randomly selected group of other individuals.
From this base population, all cases with a computer-recorded diagnosis of dementia or Alzheimer disease (AD) entered between 1992 and 1998 were identified, and each case was matched with up to 4 controls. Excluded from the study were patients with diagnoses of alcoholism or drug abuse, cancer, motor neuron disease, parkinsonism, Down syndrome, chronic liver disease, chronic renal disease, epilepsy, and stroke at any time before diagnosis of dementia. The study comprised 284 patients with dementia and 1080 controls.
The results of this study showed that relative risk for dementia and/or AD (adjusted for age; sex; history of coronary bypass surgery and cerebral ischemia; smoking status; and body mass index) was not statistically significant for individuals with either untreated hyperlipidemia (odds ratio, 0.72; 95% confidence interval, 0.45-1.14) or those treated with nonstatin LLAs (odds ratio, 0.96; 95% confidence interval, 0.47-1.97). Those treated with statins showed a highly significant adjusted risk ratio of 0.29 (95% confidence interval, 0.13-0.63; P = .002), ie, an impressive 71% reduction in the incidence of dementia.
In another study that was published a month earlier in the Archives of Neurology by Wolozin et al,2 the use of both lovastatin and pravastatin sodium was associated with a 60% to 73% (P<.001) reduction of probable AD prevalence. Interestingly enough, simvastatin (the newest and most potent of the 3 studied statins) did not show a similar association. The study was also an epidemiologic, cross-sectional analysis of patient records from 3 hospitals in the United States.
There is accumulating evidence to suggest a relation between lipids and vascular changes involving the brain in dementia. Because both lipid and vascular factors were thought to play a role in the etiology of dementia and AD as well as coronary artery disease and cerebral vascular disease, such studies were carried out. Both of these studies1,2 are based on biologically plausible mechanisms and showed a very interesting association between the use of the statins and a protective effect against dementia and AD. There are 3 possible explanations for these results. First, it could have happened by chance alone; however, given the very low P value for both studies, the likelihood of these results occurring by chance alone is remote (2/1000 and 1/1000, respectively). Second, some unmeasured confounding factors could have played a role. For example, because statins are relatively new and expensive drugs, the users might have been of higher socioeconomic status and more affluent, educated, and compliant with dietary and treatment regimens than nonusers. Unfortunately, in observational epidemiologic studies, controlling for all confounding factors is not feasible. The third explanation is that statins do truly protect against dementia and AD, thus requiring further discussion.
Even though the results of several observational epidemiologic studies have demonstrated a causal relationship when subjected to testing by rigorous clinical trials (ie, aspirin and myocardial infarction; folic acid and neural tube defects; and estrogen use and venous thromboembolism), it is generally accepted that observational studies are good for generating hypotheses only. The gold standard to test any hypothesis is a randomized clinical trial. Given the proven overwhelming benefits of statins in coronary artery disease and cerebral vascular disease prevention, it will be ethically impossible to conduct a placebo-controlled trial to test our hypothesis. One potentially feasible approach is to institute follow-up studies for dementia and AD in subjects who participated in the clinical trials of statins used to prevent myocardial infarction and strokes.3 The lack of protective effect of simvastatin in the second study2 is difficult to explain.
In conclusion, the use of statins was associated with a significant protective effect against dementia in 2 observational studies. Despite the power of both studies, they still contain the usual weaknesses of the epidemiological approach. These results need confirmation in an experimental design before we can advocate the use of statins merely for dementia and/or AD prevention. In the study by Jick and colleagues,1 charts were reviewed from 368 practices that enrolled more than 3 million patients in the United Kingdom-based General Practice Research Database. The number of patients with hyperlipidemia who received at least 1 LLA prescription within the previous 6 months was 24 480, while the number of patients with hyperlipidemia who received no treatment was 11 421. This alarmingly shows that almost one third of the patients with a diagnosis of hyperlipidemia from this cohort received no treatment.
So, should we get our prescription pads ready to give statins for dementia and AD prevention? From the practical point of view and until these findings are confirmed, I think we, as practitioners, have now 1 more reason to aggressively treat low-density lipoprotein hyperlipidemia with statins, especially for women and patients older than 50 years with both diabetes and a family history of dementia.4- 6 It is unacceptable, now more than ever before, to have such a high percentage of untreated hyperlipidemia cases among our patients.
Moawad MA. Possible Relationship Between Statin Use and Decreased Incidence of Dementia: Are We Ready for a New Indication for These Drugs?. Arch Intern Med. 2001;161(15):1909-1910. doi: