Statins probably somewhat protective for women

[ApropoE4]

I'm failing to understand experiment design or the difference between causality and correlation

Exactly!

Seriously, selection bias is real. Maybe, for example, statin users are sicker than non-statin users and are therefore more prone to dying of other things. Note that this study cherry-picked an endpoint - LOAD - that confuses matters further. Why not all cause mortality? As a consumer of healthcare science, I care about ACM, not individual disease states.

I propose reading the publication. The authors, too, seem to know a thing or two about selection and survivorship biases and were not studying difference in ACM or mining the data for an effect, as far as I can tell (although of course you could argue that sufficiently many researcher monkeys will eventually find something in the data).

doi:10.1001/jamaneurol.2016.3783

It still remains fascinating that "we can't conclude" based on this publication that statins are safe and probably somewhat helpful but we can conclude based on 10 case studies and a hypothesis paper that statins cause LOAD and that 37 supplements and brisk walks reverse LOAD.

[MAC]

MAC, your initiative, combined with your keenly analytical mind continues to benefit us all. Huge thanks for your efforts and outreach . I see you falling down multiple rabbit holes on various topics as they relate to ApoE4. Would you possibly consider helping us by summarizing some of your research into articles we could plug into our wiki?

Finding out my E4 status and probably more importantly, this forum of well informed, highly motivated and engated kindred spirits has given me a new mountain to climb. I want to beat this thing so bad with relentless ancestral survival tactics willpower! Julie, will help in anyway I can, but like a squirrel, I have an AD folder of many papers, need to remember what paper is associated with what nut? Seriously, anything you would like me to do, I will endeavour. For example, the summary I think I posted on exercise/AD prevention, along those lines? Let me know.

1. Do you agree that the observational study rules out the option that statins cause LOAD in the aging general population?

I don't think we can based upon this study. Cognitive decline is a known side effect of statin use as evidenced by the FDA warning label. Those who experience symptoms, would be very likely to stop using statins and therefore be excluded from this dataset. Variables like this make it difficult to draw conclusions especially from epidemiological studies. FWIW, I suspect that statins may provide both a positive and negative effect for users leading to the mixed results. Frustrating that we don't better understand the beneficial MOA.

Now that is a LOAD-ED question, have not gone down this rabbit hole...send stash of carrots. Based on this ONE study, absolutely not. Long term statin use vs. cognitive function follow-up studies appear non-existent. Since these drugs were created to address potential cardio vascular diseases in hypercholesterolemia persons, there "may" be some narrow vascular dementia benefit?

Would you like a summary on this? Suffice to say that anything effecting cholesterol is likely somehow interconnected with AD. I've briefly come across few papers recently talking about cholesterol not necessarily being causative for AD, but a biomarker? Hand grenade!

Cholesterol as a Causative Factor in Alzheimer Disease: A Debatable Hypothesis (2015 Review Paper)
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3999290/

"Cholesterol levels in serum/plasma and brain of AD patients do not support cholesterol as a causative factor in AD."

(So if statins reduce cholesterol, then why don't they improve classic AD outcome? I don't know the reference ranges of statin cholesterol mitigation, namely, how high a TC to trigger statin vs. what is normal statin modulated lower TC level, and how does this compare to the AD literature on LDL and AD incidence?)

Re FDA label intervention, brief overview, it's not just memory loss:

http://www.fda.gov/Drugs/DrugSafety/ucm293101.htm

> There have been rare reports of serious liver problems in patients taking statins. Patients should notify their healthcare professional right away if they have the following symptoms: unusual fatigue or weakness; loss of appetite; upper belly pain; dark-colored urine; or yellowing of the skin or the whites of the eyes.

> Memory loss and confusion have been reported with statin use. These reported events were generally not serious and went away once the drug was no longer being taken. [MAC: what about long term use of these various statins and cognitive, and filtering for genotype, IR, TC level, etc?"

> The Potential for Muscle Damage

Some drugs interact with statins in a way that increases the risk of muscle injury called myopathy, characterized by unexplained muscle weakness or pain. Egan explains that some new drugs are broken down (metabolized) through the same pathways in the body that statins follow. This increases both the amount of statin in the blood and the risk of muscle injury.

> Increases in blood sugar levels have been reported with statin use (actually induced diabetes, and diabetes and AD are strongly correlated)

FDA’s review of the results from the Justification for the Use of Statins in Primary Prevention: an Intervention Trial Evaluating Rosuvastatin (JUPITER) reported a 27% increase in investigator-reported diabetes mellitus in rosuvastatin-treated patients compared to placebo-treated patients. High-dose atorvastatin had also been associated with worsening glycemic control in the Pravastatin or Atorvastatin Evaluation and Infection Therapy – Thrombolysis In Myocardial Infarction 22 (PROVE-IT TIMI 22) substudy. FDA also reviewed the published medical literature. A meta-analysis by Sattar et al., which included 13 statin trials with 91,140 participants, reported that statin therapy was associated with a 9% increased risk for incident diabetes (odds ratio [OR] 1.09; 95% confidence interval [CI] 1.02-1.17), with little heterogeneity (I2=11%) between trials. A meta-analysis by Rajpathak et al., which included 6 statin trials with 57,593 participants, also reported a small increase in diabetes risk (relative risk [RR] 1.13; 95% CI 1.03-1.23), with no evidence of heterogeneity across trials. A recent study by Culver et al., using data from the Women’s Health Initiative, reported that statin use conveys an increased risk of new-onset diabetes in postmenopausal women, and noted that the effect appears to be a medication class effect, unrelated to potency or to individual statin. FDA’s review of the results from the Justification for the Use of Statins in Primary Prevention: an Intervention Trial Evaluating Rosuvastatin (JUPITER) reported a 27% increase in investigator-reported diabetes mellitus in rosuvastatin-treated patients compared to placebo-treated patients. High-dose atorvastatin had also been associated with worsening glycemic control in the Pravastatin or Atorvastatin Evaluation and Infection Therapy – Thrombolysis In Myocardial Infarction 22 (PROVE-IT TIMI 22) substudy.

One of the trials referenced by FDA in their review of labelling:

Randomized trial of the effects of simvastatin on cognitive functioning in hypercholesterolemic adults.
https://www.ncbi.nlm.nih.gov/pubmed/15589485

n= 308, 35-75 yrs age, 6 month only trial

This study provides partial support for minor decrements in cognitive functioning with statins. Whether such effects have any long-term sequelae or occur with other cholesterol-lowering interventions is not known.

[MarcR]

ApropoE4, why would we readers of the paper reach stronger conclusions than the authors did? They are scrupulous throughout about referring to associations rather than causes and effects.

I have not disparaged the authors' statistical work. I am simply pointing out that ascribing causal relationships to correlations is illogical and unscientific. I completely understand why some of us allow studies like this to inform our personal choices as none of us have the luxury of acting on perfect information. But particularly in the Science and Research area, I think we should strive to be careful not to misrepresent the science. Compare your conclusion - "Statins probably somewhat protective for women" - with the authors' more guarded and nuanced one:

This suggests that certain patients, facing multiple, otherwise equal statin alternatives for hyperlipidemia treatment, may reduce AD risk by using a particular statin.

It still remains fascinating that "we can't conclude" based on this publication that statins are safe and probably somewhat helpful but we can conclude based on 10 case studies and a hypothesis paper that statins cause LOAD and that 37 supplements and brisk walks reverse LOAD.

The last time you and I had this conversation, you had no response. I stand by what I said then and challenge you to identify where anyone said "that statins cause LOAD and that 37 supplements and brisk walks reverse LOAD".

[ApropoE4]

I had no response because there is no response that would satisfy an ideologue.

In scientific research (especially such that takes place in industry), we often, very very often, are happy to use known associations even if we don't understand the underlying mechanisms of action - of course we are happy when we do understand them but still aware that usually we are wrong even when we have a mechanism of action that seems to predict almost all experimental results (see Newtonian Mechanics, Electromagnetics during the late ether era, etc. - all excellent theories).

Researchers in medicine, where mechanisms of actions are usually a prerequisite for clinical trial approval, are often in a bind and may end up making mechanisms of action up where in fact all they have are observations (see Bexarotene), with the hope that they can get funded. It is better, imo, to say "this seems to work, even after we've adjusted for all known biases, and we don't understand the disease process very well anyway, and the sample size is huge, so let's say it's probably helpful and certainly not the cause of AD", which going back to the beginning of this thread - "Statins probably somewhat protective for women" states.

p.s. If you had to make a bet - helpful or harmful or neither, which would you choose?

[MAC]

MAC, you're correct, I should have used age-specific rates. Do you think the decline of age-specify coronary disease might correlate with the mixed AD/vascular disease?

Actually, the incidence rate of death rate by heart disease has dropped dramatically, whereas AD incidence has apparently simply plateaued, no evidence yet of a drop.

Deaths per 100,000 persons:

YEAR CVD AD

1950 590 NA
2000 258 18
2014 167 27

Screen Shot 2017-01-19 at 2.39.06 PM.png

If the age-adjusted vascular (no AD) dementia does decline, would not the percentage of only amyloid & tau (more pure AD) increase, hence a higher Apoe4 percentage of those with AD?

I didn't find any breakdown of AD type incidence to comment on your assertion. Vascular is typically about 10% of all dementia diagnosis, although it's very complicated, as most all cases are considered some type of mixed classifications. I am not sure about causation of increased % amyloid/tau due to any possible reduction in vascular dementia. But yes, if the incidence rate is set, and vascular is decreasing, then simple math would equate to higher amyloid/tau type, as a percentage of the total. As for E4, I haven't dug into the data on the correlation between E4 and type of alzheimer classification.

Quick google:

http://www.karger.com/Article/Abstract/106730

"The frequency of the Ε4 is increased in familial and sporadic late-onset Alzheimer's disease, but its prevalence in non-Alzheimer dementias in Caucasian populations is unknown. We found that the frequency of the E4 was 0.45 in 93 Alzheimer's disease patients, 0.46 in 23 vascular dementia patients."

So alzheimer vs. vascular similar E4 frequency one study reference.

[KatieS]

MAC, this is what I was referring to as declining age-specific rates of AD.

[MAC]

Duly noted. Do any of these studies breakout by genotype?

So how do you see the answer to your question?