I am a preventive Cardiologist with a strong research interest in lipids, diet and brain health. This data, and yoiur superb article in particular Karin, is a major step forward in the field of Apo E4. I have one apoE4 allele, so this is personal for me. This data around the benefits of very low LDL and apoB are very strong for coronary disease, multi infarct demential, and peripheral artery disease. Ideal apo B levels are about 30 to 40, which is what our levels were when we were born. All wild mammals including hunter gatherer humans had apo B in the range, so it is in the evolutionarily normal range. I virtually never use a statin without ezetimibe, as these drugs are synergistic. I often add PCSK9i (Repatha or Praluent) to the mix. I personally follow a pristine diet and lifestyle but still take ezetimibe 10 mg d, Repatha 140 mg every two weeks and pitavastatin 1 mg Mon, Weds, Fri. Pitavastatin is a VERY cool statin, as it is neutral on glucose metabolism (most statins increase A1c sltightly), and has fewer side effects. This keeps my apoB about 30 to 40. High omega-3 is also key for apo E4 carriers, as the latest data from our group shows these people get outsize benefits from high omega-3 levels. Also agree with you Karin that high fiber, like added psyllium fiber, beta glucan are helpful. Finally, my version of the old adage is that an avocado a day keeps the doctor away. James O'Keefe, MD, FACC.
Thank you so much for the additional information and your perspective! I've noticed an 0.1 increase in my own A1C since starting Pravastatin which could be related...so I'll look into the Pitavastatin now! One thing I didn't include in the article, (but will do an addendum for others who might read it later), is lp(a). Many folks have never even been tested and an elevated lp(a) presence makes lipid control that much more important.
Is NAD also something worthy of adding to a statin and ezetimbe? I have upcoming appoint with a preventative cardiologist and feel a sense of urgency to know all I can! Thoughts on rapamycin vs or with NAD and statin etc…. Sounds like a lot of meds. Thanks in advance I really redirect your research and appreciate the way you explain!
I take NAD regularly but I pulse it.... month on, month off. (see Nick Norwitz' recent post https://staycuriousmetabolism.substack.com/p/never-get-alzheimers-disease-the). I also take once weekly rapamycin. You can take either NMN or NR supplement or opt for NAD injections. I chose injections since I'm older and want to be assured full absorption. If you haven't had lp(a) test, get that done and also ask for ApoB and homocysteine level which are also not routinely checked. Good luck!
@jameshokeefemd. I was just researching the Pitavastatin and see it's lipophilic. Dr. Alan Green, who had over 500 patients with APOE4, always said that APOE4s should only take hydrophilic statins...., any thoughts on this?
Yet again Karin you have blown me away with your extremely helpful and maybe life changing information! Thank you so much! I’m sending this to my Dr. so I can get on low dose statin stat!
I resisted statins for many years. My deeper dive into the lipid altering effects of APOE4 changed my mind as the research is pretty clear and unambiguous.
Thank you so very much for this! Statins are something I am wanting to try, and this piece is very helpful as I prepare for an upcoming doctor's visit.
This is a very thoughtful and balanced synthesis. The relationship between APOE4 and statins is one of those contexts where biology and clinical evidence both matter. APOE4 carriers do have distinct lipoprotein metabolism and a higher baseline cardiovascular and Alzheimer’s risk, but the best trial data we have doesn’t suggest withholding statins in this group; if anything, the absolute risk reduction for atherosclerotic events tends to be greater in those at higher baseline risk.
Where nuance matters is recognizing that risk profile is what drives benefit, not genotype alone. APOE4 may influence LDL response and lipoprotein particle distribution, but statins still reduce cardiovascular events consistently across genotype strata, and they remain a foundational therapy when indicated by risk. I also appreciate the distinction you make between mechanistic plausibility vs clinical outcomes. It’s exactly the type of careful interpretation that helps patients and clinicians move beyond headlines into decisions that actually improve longevity and quality of life.
ApoE4 is not one of the considerations I've honestly ever had around statins. I do think overall benefits of statins, even in the realm of primary prevention of CV events in low risk groups, is surprisingly strong. I'm sure we'll continue to learn about it's influence on the incidence of dementia as the population continues to age. Hopefully that comes with even more data on the ApoE4 populations.
What a great breakdown of the data and personal comments which make for a better post. As a PA/prescriber, I am a fan of properly prescribed medications and some advanced testing such as Lp(a). I appreciate the content you created and am sharing this post.
Maybe you can mention this to the very interesting, energetic, popular, but still young 4/4 Dr. Nick Norwitz. He can obviously read, but maybe he has not yet seen some of what you mention. I, too, found the Boston Heart Test to be very comprehensive, and indicating heavy production and retention. I'll take a another look at statins, though ezetimibe, diet, and trail jogging lowered my LDL to 82. Wherever I see your name, I check to see if there is something new and important to read. Thank you!
LDL of 82 sounds good to me. What about your ApoB and hsCRP? Keep an eye on them to decide if statins might be indicated and don't underestimate how fiber can help address LDL.... apart from feeding those butyrate producing gut bacteria we need! As for Nick, he is performing a lot of N=1 experiments on himself and as young as he is, I don't believe he's doing any lasting damage. He doesn't claim his 500+ LDL levels are healthy nor that he is going to keep them there.. it's simply not known where LMHR are positioned on the risk scale.
Sorry, but I think this is a really bad take. The headline is not “APOE4 carriers who initiated a statin had a 40% lower risk of developing Alzheimers”.
The headline is “if we ignore real-life observations and data from randomized controlled trials plus treat the appearance of a statin bottle in someone’s house as the same as them taking this medication long-term and then pretend that healthy user bias is not a thing, proceed to run a number of hidden adjustments to the data and apply inappropriate statistical models to the data (and omit the unadjusted dataset), we manage to calculate a relative hazard ratio of 0.60 in APOE4 carriers when taking a product made corporations we are financially tied to”. That’s not the same.
On the plus side, we can agree on one thing, that ‘this is the strongest genotype-specific human evidence we’ve ever had on this question’. I think what we disagree on is how strong the evidence is. I find it very, very weak.
To zoom out, I often hear claims in support of statins. There are three trends that are repeated here:
1. observational data (subject to ‘healthy user bias’)
2. analyses of this data that uses inappropriate statistical analysis (that just so happens to be easy to introduce bias), uses unreported adjustments and omits the unadjusted data
3. narrow focus on selected groups and selected endpoints (with zero curiosity in running ‘sanity checks’ on the hypothesis generated, either using their own data, that already available in the literature or real-world observations)
4. Claims are made via institutions that maintain ‘partnership’ with the pharmaceutical companies that sell the statins
It is therefore little surprise to see that the ‘pivotal’ 2024 study – one authored by researchers from the industry-partnered Rush Institute of Chicago - shows these exact trends.
To be more specific, this is an observational study that did not track statin use, only whether there were any statin bottles in their house (which they classified as ‘statin INITIATION’). We have no way of knowing what exposure the ‘statin’ group had to statins, which the authors readily admit (albeit in a single sentence in the discussion: “it is unclear if the participants were taking the medication”).
The problem with observational studies is that they cannot account for ‘healthy user bias’, ie. those that prioritize their health (which spills into multiple areas of life, be it going to bed on time, engaging/disengaging in stressful activities, etc etc… but also impacts on how likely to visit a doctor to address health concerns). This is incredibly weak, especially given how much trial data is available on statins in particular. There is no need to use such weird metrics.
In short, tracking statin ‘initiation’ – rather than exposure – will inevitably be confounded by this healthy user bias. The less you visit your doctor, the less likely you are to be recommended a statin. Therefore, it is a perfect example of metric for how often someone visits their doctor (and points heavily to the ‘healthy user bias’ that is rampant in observational studies).
The authors used a co-variate adjusted Kaplan-Meier curve with a Cox regression model. I don’t want to get into the weeds here but Kaplan-Meier estimators treat death as a ‘censored’ event (assumes that a person who died of heart disease at age 75 would have eventually developed Alzheimers at the same rate as those who survived). It’s obviously not possible to develop Alzheimer’s after death. This substantially overestimates the cumulative risk of Alzheimers (artificially increasing any differences in hazard ratios between the two groups, whether such difference be from actual differences in hazards or generated from contrived use of adjustments), and especially so in older populations (who die at an increased risk versus a younger population).
Now I appreciate that discussions on methodology are boring. Important, perhaps, when it is via statistical ‘wizardry’ that these figures are generated, but still boring (at least, they bore me). But this is where, for those who aren’t interested in debating the relative merits of Kaplan-Meier vs Fine-Gray Subdistribution Hazard Models, we can simply look at what the researchers are saying and run it through some simple sanity checks.
When it comes to such sanity checks, there is an obvious howler in that they readily admit that they have data for those using statins at baseline but entirely omit this from analysis and discussion. If statins have such a powerful effect on reducing Alzheimers in APOE4 carriers and they are as enthused as they claim about this strategy, then why not at least check what the rate of Alzheimers is among those already taking statins at baseline? If statins have a protective effect against Alzheimers over time in this group then, by definition, we would see more of a protective effect in those who have taken them for a longer time. They choose not to share this data.
But when it comes to these checks, they can be translated to a simple question: if this hypothesis were true, what else would be true?
Well, if statins were as helpful as indicated by this data, then we’d be able to see this in the real world. Policy changes meant that, in the US, statin use rose astronomically from 37 million in 2012–2013 to 92 million users in 2018–2019 (from 12% to 35% of over 40s). 15-25% of the population has at least one APOE4 polymorphism.
Yet the rate of Alzheimers hasn’t dropped – yes, there are caveats and noise when discussing age-adjusted risk and total prevalence, but the Baby Boomers take the most statins and yet have the highest rate of Alzheimers – and overall mortality hasn’t improved (the opposite has happened, something that is historically unprecedented). The hypothesis fails at the first test.
(and PS. yes, there are some RCTs that ‘find’ a benefit, but is there a single RCT that a) reports the endpoints that were pre-registered rather than those that they’ve p-hacked, b) actually compares statins to placebo rather than the scandalous ‘run-in periods’ to distort side-effect profile and c) shows clinically significant benefits?)
You’re right: it’s observational, healthy-user bias is possible, and exposure wasn’t directly measured. I didn’t claim causality, only that this is one of the stronger genotype-specific human signals we currently have. Until we have long-term genotype-stratified RCTs with verified exposure, the evidence will remain imperfect and needs to be interpreted cautiously.
Thanks for this very informative. My question is- having an HDL of 95 and LDL of 150 along with being heterozygote Apoe4- are there still benefits in taking a statin?
The LDL number is less important than looking at the bigger picture which includes: your ApoB number, HDL/Trig. ratio, whether or not you have genetically determined elevated lp(a) and if you are dealing with inflammation (hsCRP). If those aren't pristine, I would look at ways to adjust that number. Ezetimibe (blocking cholesterol absorption) and increased fiber intake would be a good place to start.
A good perspective for the practitioners and I do hope they will be able to translate it for the benefit of their patients. Still, I would request the author to provide a small commoner’s summary here in the comments. Statins have been around for 40-50 years, they are no longer the big pharma’s golden goose. They are bulk manufactured and offered by hundreds of small companies. If there are government schemes for affordable drugs, as in India here (nearly a few hundred drugs and formulations are on offer), they can cost a pittance. For example, under this scheme, a 10 mg Atorvostatin tablet just costs a cent. Like all drugs we have known, statins have their pluses and minuses and the doctors need to be well informed. But their potential in other physically and financially crippling diseases ( cancer, Alzemeirs) should not be missed and must be brought to the table - as prophylactic or therapeutic. Research minded small companies and doctors have a large role here.
The point I’m making is narrower and APOE-specific: APOE4 carriers have distinct lipid transport and neuroinflammatory biology, and responses to statins — particularly regarding cognition — may differ meaningfully from the general population.
I agree that statins have potential pleiotropic effects and may play roles beyond LDL lowering. But in APOE4, the question isn’t access or cost — it’s whether benefits outweigh risks at the brain level, and that’s where more genotype-aware discussion and research are needed.
I am a preventive Cardiologist with a strong research interest in lipids, diet and brain health. This data, and yoiur superb article in particular Karin, is a major step forward in the field of Apo E4. I have one apoE4 allele, so this is personal for me. This data around the benefits of very low LDL and apoB are very strong for coronary disease, multi infarct demential, and peripheral artery disease. Ideal apo B levels are about 30 to 40, which is what our levels were when we were born. All wild mammals including hunter gatherer humans had apo B in the range, so it is in the evolutionarily normal range. I virtually never use a statin without ezetimibe, as these drugs are synergistic. I often add PCSK9i (Repatha or Praluent) to the mix. I personally follow a pristine diet and lifestyle but still take ezetimibe 10 mg d, Repatha 140 mg every two weeks and pitavastatin 1 mg Mon, Weds, Fri. Pitavastatin is a VERY cool statin, as it is neutral on glucose metabolism (most statins increase A1c sltightly), and has fewer side effects. This keeps my apoB about 30 to 40. High omega-3 is also key for apo E4 carriers, as the latest data from our group shows these people get outsize benefits from high omega-3 levels. Also agree with you Karin that high fiber, like added psyllium fiber, beta glucan are helpful. Finally, my version of the old adage is that an avocado a day keeps the doctor away. James O'Keefe, MD, FACC.
Thank you so much for the additional information and your perspective! I've noticed an 0.1 increase in my own A1C since starting Pravastatin which could be related...so I'll look into the Pitavastatin now! One thing I didn't include in the article, (but will do an addendum for others who might read it later), is lp(a). Many folks have never even been tested and an elevated lp(a) presence makes lipid control that much more important.
Is NAD also something worthy of adding to a statin and ezetimbe? I have upcoming appoint with a preventative cardiologist and feel a sense of urgency to know all I can! Thoughts on rapamycin vs or with NAD and statin etc…. Sounds like a lot of meds. Thanks in advance I really redirect your research and appreciate the way you explain!
I take NAD regularly but I pulse it.... month on, month off. (see Nick Norwitz' recent post https://staycuriousmetabolism.substack.com/p/never-get-alzheimers-disease-the). I also take once weekly rapamycin. You can take either NMN or NR supplement or opt for NAD injections. I chose injections since I'm older and want to be assured full absorption. If you haven't had lp(a) test, get that done and also ask for ApoB and homocysteine level which are also not routinely checked. Good luck!
@jameshokeefemd. I was just researching the Pitavastatin and see it's lipophilic. Dr. Alan Green, who had over 500 patients with APOE4, always said that APOE4s should only take hydrophilic statins...., any thoughts on this?
Yet again Karin you have blown me away with your extremely helpful and maybe life changing information! Thank you so much! I’m sending this to my Dr. so I can get on low dose statin stat!
I resisted statins for many years. My deeper dive into the lipid altering effects of APOE4 changed my mind as the research is pretty clear and unambiguous.
I think this is one of the most important messages an APOE4/4 carrier can receive. Thank you for sharing.
Thank you so very much for this! Statins are something I am wanting to try, and this piece is very helpful as I prepare for an upcoming doctor's visit.
This is a very thoughtful and balanced synthesis. The relationship between APOE4 and statins is one of those contexts where biology and clinical evidence both matter. APOE4 carriers do have distinct lipoprotein metabolism and a higher baseline cardiovascular and Alzheimer’s risk, but the best trial data we have doesn’t suggest withholding statins in this group; if anything, the absolute risk reduction for atherosclerotic events tends to be greater in those at higher baseline risk.
Where nuance matters is recognizing that risk profile is what drives benefit, not genotype alone. APOE4 may influence LDL response and lipoprotein particle distribution, but statins still reduce cardiovascular events consistently across genotype strata, and they remain a foundational therapy when indicated by risk. I also appreciate the distinction you make between mechanistic plausibility vs clinical outcomes. It’s exactly the type of careful interpretation that helps patients and clinicians move beyond headlines into decisions that actually improve longevity and quality of life.
Fantastic work and helpful in the fight against misinfo that runs rampant among APOE4 community. Fear makes an unfortunate friend of misinformation.
ApoE4 is not one of the considerations I've honestly ever had around statins. I do think overall benefits of statins, even in the realm of primary prevention of CV events in low risk groups, is surprisingly strong. I'm sure we'll continue to learn about it's influence on the incidence of dementia as the population continues to age. Hopefully that comes with even more data on the ApoE4 populations.
What a great breakdown of the data and personal comments which make for a better post. As a PA/prescriber, I am a fan of properly prescribed medications and some advanced testing such as Lp(a). I appreciate the content you created and am sharing this post.
Thanks, Karin, as always for sharing helpful and accurate information!
Great article! Thanks for sharing the data!
Maybe you can mention this to the very interesting, energetic, popular, but still young 4/4 Dr. Nick Norwitz. He can obviously read, but maybe he has not yet seen some of what you mention. I, too, found the Boston Heart Test to be very comprehensive, and indicating heavy production and retention. I'll take a another look at statins, though ezetimibe, diet, and trail jogging lowered my LDL to 82. Wherever I see your name, I check to see if there is something new and important to read. Thank you!
LDL of 82 sounds good to me. What about your ApoB and hsCRP? Keep an eye on them to decide if statins might be indicated and don't underestimate how fiber can help address LDL.... apart from feeding those butyrate producing gut bacteria we need! As for Nick, he is performing a lot of N=1 experiments on himself and as young as he is, I don't believe he's doing any lasting damage. He doesn't claim his 500+ LDL levels are healthy nor that he is going to keep them there.. it's simply not known where LMHR are positioned on the risk scale.
Sorry, but I think this is a really bad take. The headline is not “APOE4 carriers who initiated a statin had a 40% lower risk of developing Alzheimers”.
The headline is “if we ignore real-life observations and data from randomized controlled trials plus treat the appearance of a statin bottle in someone’s house as the same as them taking this medication long-term and then pretend that healthy user bias is not a thing, proceed to run a number of hidden adjustments to the data and apply inappropriate statistical models to the data (and omit the unadjusted dataset), we manage to calculate a relative hazard ratio of 0.60 in APOE4 carriers when taking a product made corporations we are financially tied to”. That’s not the same.
On the plus side, we can agree on one thing, that ‘this is the strongest genotype-specific human evidence we’ve ever had on this question’. I think what we disagree on is how strong the evidence is. I find it very, very weak.
To zoom out, I often hear claims in support of statins. There are three trends that are repeated here:
1. observational data (subject to ‘healthy user bias’)
2. analyses of this data that uses inappropriate statistical analysis (that just so happens to be easy to introduce bias), uses unreported adjustments and omits the unadjusted data
3. narrow focus on selected groups and selected endpoints (with zero curiosity in running ‘sanity checks’ on the hypothesis generated, either using their own data, that already available in the literature or real-world observations)
4. Claims are made via institutions that maintain ‘partnership’ with the pharmaceutical companies that sell the statins
It is therefore little surprise to see that the ‘pivotal’ 2024 study – one authored by researchers from the industry-partnered Rush Institute of Chicago - shows these exact trends.
To be more specific, this is an observational study that did not track statin use, only whether there were any statin bottles in their house (which they classified as ‘statin INITIATION’). We have no way of knowing what exposure the ‘statin’ group had to statins, which the authors readily admit (albeit in a single sentence in the discussion: “it is unclear if the participants were taking the medication”).
The problem with observational studies is that they cannot account for ‘healthy user bias’, ie. those that prioritize their health (which spills into multiple areas of life, be it going to bed on time, engaging/disengaging in stressful activities, etc etc… but also impacts on how likely to visit a doctor to address health concerns). This is incredibly weak, especially given how much trial data is available on statins in particular. There is no need to use such weird metrics.
In short, tracking statin ‘initiation’ – rather than exposure – will inevitably be confounded by this healthy user bias. The less you visit your doctor, the less likely you are to be recommended a statin. Therefore, it is a perfect example of metric for how often someone visits their doctor (and points heavily to the ‘healthy user bias’ that is rampant in observational studies).
The authors used a co-variate adjusted Kaplan-Meier curve with a Cox regression model. I don’t want to get into the weeds here but Kaplan-Meier estimators treat death as a ‘censored’ event (assumes that a person who died of heart disease at age 75 would have eventually developed Alzheimers at the same rate as those who survived). It’s obviously not possible to develop Alzheimer’s after death. This substantially overestimates the cumulative risk of Alzheimers (artificially increasing any differences in hazard ratios between the two groups, whether such difference be from actual differences in hazards or generated from contrived use of adjustments), and especially so in older populations (who die at an increased risk versus a younger population).
Now I appreciate that discussions on methodology are boring. Important, perhaps, when it is via statistical ‘wizardry’ that these figures are generated, but still boring (at least, they bore me). But this is where, for those who aren’t interested in debating the relative merits of Kaplan-Meier vs Fine-Gray Subdistribution Hazard Models, we can simply look at what the researchers are saying and run it through some simple sanity checks.
When it comes to such sanity checks, there is an obvious howler in that they readily admit that they have data for those using statins at baseline but entirely omit this from analysis and discussion. If statins have such a powerful effect on reducing Alzheimers in APOE4 carriers and they are as enthused as they claim about this strategy, then why not at least check what the rate of Alzheimers is among those already taking statins at baseline? If statins have a protective effect against Alzheimers over time in this group then, by definition, we would see more of a protective effect in those who have taken them for a longer time. They choose not to share this data.
But when it comes to these checks, they can be translated to a simple question: if this hypothesis were true, what else would be true?
Well, if statins were as helpful as indicated by this data, then we’d be able to see this in the real world. Policy changes meant that, in the US, statin use rose astronomically from 37 million in 2012–2013 to 92 million users in 2018–2019 (from 12% to 35% of over 40s). 15-25% of the population has at least one APOE4 polymorphism.
Yet the rate of Alzheimers hasn’t dropped – yes, there are caveats and noise when discussing age-adjusted risk and total prevalence, but the Baby Boomers take the most statins and yet have the highest rate of Alzheimers – and overall mortality hasn’t improved (the opposite has happened, something that is historically unprecedented). The hypothesis fails at the first test.
(and PS. yes, there are some RCTs that ‘find’ a benefit, but is there a single RCT that a) reports the endpoints that were pre-registered rather than those that they’ve p-hacked, b) actually compares statins to placebo rather than the scandalous ‘run-in periods’ to distort side-effect profile and c) shows clinically significant benefits?)
You’re right: it’s observational, healthy-user bias is possible, and exposure wasn’t directly measured. I didn’t claim causality, only that this is one of the stronger genotype-specific human signals we currently have. Until we have long-term genotype-stratified RCTs with verified exposure, the evidence will remain imperfect and needs to be interpreted cautiously.
Thanks for this very informative. My question is- having an HDL of 95 and LDL of 150 along with being heterozygote Apoe4- are there still benefits in taking a statin?
The LDL number is less important than looking at the bigger picture which includes: your ApoB number, HDL/Trig. ratio, whether or not you have genetically determined elevated lp(a) and if you are dealing with inflammation (hsCRP). If those aren't pristine, I would look at ways to adjust that number. Ezetimibe (blocking cholesterol absorption) and increased fiber intake would be a good place to start.
Good news. I have FH so I have been on statins a long time.
A good perspective for the practitioners and I do hope they will be able to translate it for the benefit of their patients. Still, I would request the author to provide a small commoner’s summary here in the comments. Statins have been around for 40-50 years, they are no longer the big pharma’s golden goose. They are bulk manufactured and offered by hundreds of small companies. If there are government schemes for affordable drugs, as in India here (nearly a few hundred drugs and formulations are on offer), they can cost a pittance. For example, under this scheme, a 10 mg Atorvostatin tablet just costs a cent. Like all drugs we have known, statins have their pluses and minuses and the doctors need to be well informed. But their potential in other physically and financially crippling diseases ( cancer, Alzemeirs) should not be missed and must be brought to the table - as prophylactic or therapeutic. Research minded small companies and doctors have a large role here.
The point I’m making is narrower and APOE-specific: APOE4 carriers have distinct lipid transport and neuroinflammatory biology, and responses to statins — particularly regarding cognition — may differ meaningfully from the general population.
I agree that statins have potential pleiotropic effects and may play roles beyond LDL lowering. But in APOE4, the question isn’t access or cost — it’s whether benefits outweigh risks at the brain level, and that’s where more genotype-aware discussion and research are needed.
This is an intriguing and great post. It continues to show the pleomorphic and beneficial effects of statins.