Jump to content

Cholesterol paradox

Recommended Posts

Now this is pretty weird. on October 15, 2016, I had a glucose + lipid panel done , when I was still a alcto-ovo vegetarian. By chance, I akready had a 12-days cronometer record, so, even if the timespan is not much, I remember it's pretty representative of my diet at the time. Since I recently had another lipid panel done, and since in the meanwhile I turned vegan, a comparison is in order. I compared 12 days lacto-ovo prior to the test to 3 months vegan, prior to the test. The results are a little surprising:



Glucose: 86

Total cholesterol: 153

HDL : 59

Triglycerides: <70

LDL : 80-90



Glucose: 83

Total cholesterol: 170

HDL : 65

Triglycerides: 69


LDL : 94


My intake of cholesterol when lacto-ovo was 264 mg per day, whereas as a vegan it was 7 mg per day (the occasional tablespoon of butter on the cookies sometimes I indulge in).

So, it appears very weird that, with almost no dietary cholesterol at all, my blood cholesterol increased almost by 20 mg/dl, whereas other values are about the same.

Also, as a lacto-ovo I ate a daily average of 39 g saturated fats per day, whereas as a vegan I ate, prior to tests, 21 grams, about a half.

As a lacto-ovo I ate 133 grams net carbs, whereas as a vegan 176. But fasting glucose was lower as a vegan than as a lacto-ovo. Another strangeness.

As a lacto ovo I ate 250 less kcalories and 32 grams of fiber rather than the vegan 63 daily grams.


I mentioned the cholesterol issue to another friend who eats mainly palnt-based. He had his blood tests doen every year and observed that, as soon as he stopped eating cheese, his cholesterol went up. He also told me about another friend who, as soon as started a stint of total fasts, some of which of significant length (10 days) his cholesterol went up.


I also asked about my first test who was taken at teh local pharmacy. The pharmacist told me that the test is pretty accurate and that, according to him, cholesterol in the summer season may increase.


The paradox remains as why a vegan diet with almost no dietary cholesterol and 63 g daily fibers should result in more blood cholesterol than a lacto-ovo diet with 264 mg cholesterols and 32 g fibers

Edited by mccoy
Link to comment
Share on other sites


Mccoy: He also told me about another friend who, as soon as started a stint of total fasts, some of which of significant length (10 days) his cholesterol went up.


Cholesterol levels can definitely go up (temporarily) following weight loss.



That's sensible, although there remain the cases where the bodyweight was stable (like my own case).

Link to comment
Share on other sites

How long after your last fast did you get your blood tests?    If stored body fat was released into the bloodstream during fasting, it is certainly possible that it could take some time before blood cholesterol levels returned to previous levels (even if body weight returns quickly).

Link to comment
Share on other sites

How long after your last fast did you get your blood tests?    If stored body fat was released into the bloodstream during fasting, it is certainly possible that it could take some time before blood cholesterol levels returned to previous levels (even if body weight returns quickly).


Over 10 days from Longo's FMD. Still, there is my friend's case when he ceased eating cheese. I'll try and collect details about that as well.

Link to comment
Share on other sites

The LDL value of a standard lipid panel isn't measured, it is calculated, and it isn't particularly meaningful at least in isolation.  Here's an interesting Rhonda  Patrick interview/discussion on the topic with Dr. Ronald Krauss.



I heard a lot about the specific lipid particles fractions, but AFAIK there isn't much consensus about its real usefulness among lipidologists. I still cannot grasp well that concept, don't know why. the paleo love that, but they love so many weird things.

Edited by mccoy
Link to comment
Share on other sites

Mccoy: Over 10 days from Longo's FMD.



Not long at all.   That could very well be a factor. 


In any case,  170 TC may (I'm hedging here, I need to review the latest info)  be better than 153 TC when it comes to the association of TC with all-cause mortality (especially as you get older). 



On LDL particle size,  the predictive power of LDL level etc.  see Michael Rae's extensive post here:


Beyond LDL - Is Cholesterol Particle Size Worth Testing?


Edited by Sibiriak
Link to comment
Share on other sites

Todd Allen:  The LDL value of a standard lipid panel isn't measured, it is calculated...


Yes, the Friedewald equation is usually used.    LDL=  TC minus HDL-C minus  (TGs/5)


Some interesting info here:


Variations in Lipid Values



Biological variation


Biological variation, i.e., the normal day-to-day variation of total cholesterol, LDL cholesterol, and HDL cholesterol, is on the order of 3-5%. Many studies have also documented a seasonal variation. Although there is discordance between the studies, cholesterol levels tend to be higher in the winter months and lower in the summer months independent of the country of origin, ethnicity, age, sex, and baseline lipids.   The seasonal variation has been reported to be as high as 12%.


Other biological factors that contribute to a patient’s cholesterol level include: Within-day variation – An individual’s serum cholesterol values can vary about 2-3% within the same day. Cholesterol levels are also lower for shorter periods in response to severe pain, surgery, and short-term physical strain.


Posture  – Cholesterol can decrease significantly after a person has been sitting for five minutes.  Changes as large as 10-15% have been observed


Capillary vs. Venous sampling – Most point of care testing devices can utilize capillary blood (obtained with a finger stick) in addition to venous blood. Total cholesterol measured from capillary plasma (obtained via a finger stick) tends to run 2 to 4% higher than venous plasma cholesterol.


In contrast to cholesterol, the normal day-to-day variation of triglycerides is 20 to 30%. The variation in triglycerides has two major clinical implications: 1) when patients are being treated for high triglycerides, it is important to follow the trend in triglycerides, i.e., a single high or low triglyceride value may be due to biological variation and not an improvement or reduction in the efficacy of the triglyceride lowering medication, and 2) if LDL cholesterol is calculated by the Friedewald equation [LDL cholesterol = total cholesterol - HDL cholesterol - triglycerides/5], then the biological variation in triglycerides will add to the variability of the LDL calculation. Since triglycerides are divided by 5 in this equation, this minimizes the effect of triglyceride variation on LDL calculation but will add 4 to 6% variability in the LDL calculation on top of the 3 to 4% variation in LDL per se.


Analytical variation


Analytical variation is the variability in measuring cholesterol and triglycerides that derives from the assay itself due to random error and/or systematic error (described below). LDL cholesterol is the primary target for reducing risk for cardiovascular disease. The reference method or “gold standard” for measuring LDL cholesterol is the beta-quantification procedure. This method subjects serum to ultracentrifugation to separate the VLDL from LDL and HDL. With this method, the LDL cholesterol measured actually includes intermediate density lipoproteins (IDL), LDL, and lipoprotein(a). It is important to recognize that this method for measuring LDL cholesterol includes other lipoproteins. All other analytical tools to determine LDL cholesterol are compared to this method.


The current primary goal for measuring LDL cholesterol is that the total error be within 12% of the true value. For example, LDL cholesterol of 150 mg/dL may translate to a range of 132 to 168 mg/dL. The “total error” combines the imprecision of a given assay (random error) with the inaccuracy or bias (systematic error, i.e., a given assay will always be high or low compared to the true results).


Comparison of methods to measure LDL cholesterol - Although ultracentrifugation is the gold standard, it is very cumbersome to perform in high volume on a routine basis. In 1972, Friedewald developed his formula to estimate LDL cholesterol where LDL cholesterol = total cholesterol – HDL cholesterol – triglycerides/5. It is important to know that this calculated LDL also includes IDL and lipoprotein(a). This method was quickly adopted as the routine means of calculating LDL.  However, the overall variability of calculated LDL between different clinical laboratories was nearly 12%.


The Friedewald equation is most accurate when triglycerides are less than 250 mg/dL and not accurate when triglycerides exceed 400 mg/dL. Hence, postprandial samples can not be used to calculate LDL.


In the past 10 years, direct methods have been developed to determine LDL cholesterol. These “direct” LDL methods initially used chemical precipitation or antibodies to separate LDL from the other lipoproteins. Currently, the third generation of direct LDL methods utilizes various means to block or dissolve lipoproteins to isolate LDL cholesterol. Unlike the calculated LDL, these direct LDL measurements do not include IDL and lipoprotein(a). All things being equal, it is expected that the direct LDL will be 5 to 10 mg/dL lower than the calculated LDL.** In general, these assays are also much less affected by triglycerides and can be used in postprandial samples with little loss of accuracy.



**Contradicting the above claim,  I've found a number of articles that say Direct LDL numbers tend to be HIGHER than calculated LDL numbers (varying with TG levels).   There seems to be conflicting data-- the whole issue  is quite complex


See: http://clinchem.aaccjnls.org/content/clinchem/48/2/236.full.pdf.

Edited by Sibiriak
Link to comment
Share on other sites

Sibirak, according to the material you posted, since cholesterol levels tend to be lower in the summer months, and capillary blood yields higher results, the paradoxic differences in my lacto-ovo and vegan cholesterol values become even more pronounced.

Link to comment
Share on other sites

Any comments? Valter Longo's FMD is getting more kudos from Rhonda.


Keto diet is discussed in the right perspective. Not the panacea some people suggest it is, rather a tool which is  really not suited to everyone (LDL and inflammation skyrockets some times).


Many other things like the IGF-1 soft spot, skepticism toward chronic CR versus cycles of CR and refeeding, I'm still at 2/3 the podcast .

Link to comment
Share on other sites

Here's a very recent podcast between Dr. Peter Attia and Dr. Ron Krauss plus a page of interesting links to studies and related stuff.



I listened to this as well. Interestingly, Ron Krauss does not rule out the importance of total LDL as a CVD risk indicator. His small particles hypothesis is obstensibly sort of a refinement of the  LDL hypothesis, which may work better for some subsets of the population. He also confirms the importance of Trygs and HDL as indicators of CVD.

So, it's not precisely as many paleo guys assert, that cholesterol particle size and number is the only way to go.


Also, it seems (from my limied perspective, I would need to listen again at it 5-6 more times) that Ron Krauss  is reasoning more along the line of what Michael Rae wrote in his post: apoB, Lp(a), inflammation, insulin resistance....



Edited by mccoy
Link to comment
Share on other sites

I also find Ron Krauss's comments on statins interesting.  On one hand he seems to be defending them as highly potent reducers of LDL but he also mentions many disturbing aspects such as negative impacts on mitochondria, muscle and elevating risk of type 2 diabetes and points out they are more effective at lowering the larger LDL particles than the smaller LDL which appear to be more problematic.

Link to comment
Share on other sites

Well, if you put together Krauss's comments on LDL and statins in both the podcasts (Rhonda's and Peter's), he states that without question LDL is causative in CVD; that yes, smaller LDL particles are more deleterious, and yes, statins remove larger particles rather than smaller particles, but according to him, that still is beneficial in the bigger scheme of things. The other aspect, which Peter brought up, and which Krauss agreed with, is that statins may have other beneficial effects in addition to lowering of LDL, most likely mediated through some kind of lowering of inflammation. More targeted drugs, such as PCSK9 inhibitors work directly through retarding cholesterol production in the liver - the benefit is several fold, in that the production of cholesterol in the liver is the big problem and you actually need some cholesterol in other tissues, and statins being less selective can therefore have bad effects on other tissues. But on the other hand, it is possible that if part of the reason for statin effectiveness is inflammation lowering, then perhaps PCSK9 inhibitors don't have that advantage vs statins. In summary - it's complicated (as always).


All that said, I personally found MR's takedown of the whole "small vs large LDL particles" theory very convincing - I therefore don't put as much stock in Krauss's theory. Like I said above "it's complicated" :)

Link to comment
Share on other sites

  • 1 year later...

A new, large retrospective cohort study [1] (popular press article) found statin users were more than twice as likely to develop type-2 diabetes compared with people who had a similar risk profile for heart disease but didn't take statins. The diabetes risk increased to 3x for those who had been taking a statin for more than two years. 

Here is a graph of diabetes risk over time (days) between the two groups:

Screenshot_20190812-094739_Foxit PDF.jpg

One more thing to consider when contemplating statins for elevated serum cholesterol which is unresponsive to lifestyle interventions. At the very least it would seem statin users should monitor their blood glucose and take steps to reverse the trend if it starts to become elevated.



[1] Diabetes Metab Res Rev. 2019 May 24:e3189. doi: 10.1002/dmrr.3189

Statin users have an elevated risk of dysglycemia and new-onset-diabetes.

Zigmont VA(1), Shoben AB(2), Lu B(2), Kaye GL(3), Clinton SK(4), Harris RE(1),
Olivo-Marston SE(1).

OBJECTIVE: Statins are one of the most widely prescribed medications in the

United States; however, there is a concern that they are associated with
new-onset-diabetes (NOD) development. We sought to understand the risk of
dysglycemia and NOD for a cohort of individuals that reflect real-world physician
prescribing patterns.
METHODS: A retrospective cohort study was conducted among individuals with
indications for statin use (n = 7064). To examine elevated glycosylated
hemoglobin (>6.0%), logistic regression with inverse probability weighting was
used to create balance between incident statin users and nonusers. To evaluate
the risk of NOD development, Cox PH models with time varying statin use compared 
NOD diagnoses among statin users and nonusers.
RESULTS: A higher prevalence of elevated HbA1c (PD = 0.065; 95% CI: 0.002, 0.129,
P = 0.045) occurred among nondiabetic incident users of statins. Additionally,
statin users had a higher risk of developing NOD (AHR = 2.20; 95% CI: 1.35, 3.58,
P = 0.002). Those taking statins for 2 years or longer (AHR = 3.33; 95% CI: 1.84,
6.01, P < 0.001) were at the greatest risk of developing NOD
; no differences were
observed by statin class or intensity of dose.
CONCLUSION: As lifestyle programs like the Diabetes Prevention Program are
promoted in primary care settings, we hope physicians will integrate and insurers
support healthy lifestyle strategies as part of the optimal management of
individuals at risk for both NOD and cardiovascular disease. The relationships
between statin use and glycemic control should be evaluated in large cohort
studies, medical record databases, and mechanistic investigations to inform
clinical judgment and treatment.

© 2019 John Wiley & Sons, Ltd.

DOI: 10.1002/dmrr.3189 
PMID: 31125480 

Link to comment
Share on other sites

Back to cholesterol with more data in my head. All right, I understood at last that cholesterol is synthetized by cells and that dislypidemia is a problem of faulty homeostasis even though often wrong dietary habits may be a cofactor.

When the paleo say that we need to ingest cholesterol which is fundamental to life, they ignore that the body is very well capable to provide for himself. 

 In Peter attia's series on lipidology with Tom Daysprings, statins are discussed with unusual detail, a suggested listening to those who are taking them.

Edited by mccoy
Link to comment
Share on other sites

My own experience which I’ve reported here several times is this.

high 35-40% fat raises my Hdl by about a 1/3, cuts triglycerides in half and has zero effect on my ldl when compared to a low fat diet ~ 12-15%

The fat thing is still an issue for some and of course high fat diets tend to injure the endothelial tissue by causing chronic issues with flow  mediated dilation. Of course this will effect some more than others such as smokers, sedentary folks, junk foodies, diabetics etc. wine is an excellent accompaniment with high fat meals and may partly explain the French paradox.

Link to comment
Share on other sites

In the context of statins and diabetes, I wonder what impact combination therapy of metformin and a statin would be. I understand that metformin also mildly lowers LDL, so that should not be counter to the action of a statin, meanwhile, if it lowers the odds of diabetes, that could be a plus. From my understanding, metformin is prescribed for pre-diabetes in many parts of the world, including Europe, but has not been cleared for that indication by the FDA statewide.

Here's an article making this exact claim:


"Statin monotherapy in some patients will not be sufficient to achieve an LDL-C target and increase the incidence of new-onset diabetes

in a dose dependent manner. Metformin, a safe and efficient medication for improving glucose metabolism, has been found to reduce

LDL-C levels and has cardiac benefits for patients with T2D. Therefore, statins plus metformin may further reduce LDL-C levels and partly

counteract statin-induced diabetes, which may be a promising therapeutic strategy."

There's a whole raft of articles about statins and 2TDM and adding metformin, but I'm talking about using metformin preventatively, before 2TDM appears. 

Again, as so often you need multiple agents, because you need to counteract deletrious effects of a given intervention, which is what makes all of this so complicated.

Certainly, one could explore any number of other 2TDM preventative agents in combination with statins (f.ex. acarbose), but I don't think there's been much research devoted to such combinations. Yet again, it seems, we know very little about the drugs we take and their interactions. 

Link to comment
Share on other sites

  • 11 months later...

Had a standard blood panel again after 2 years.

This time the results are a little worse:


Total cholesterol=196

LD colesterol=111

HDL colesterol=67


Still reasonable numbers, but higher up than the previous analyses. I'm eating more saturated fats from dairy products and that may of course be an explanation. 

My other results in the recent thread on basic blood tests

Edited by mccoy
Link to comment
Share on other sites

  • 2 weeks later...
3 hours ago, Sibiriak said:

Keep in mind that cholesterol levels can be transiently increased for over a month after a fast.  

Yes, I remember you pointed that out in a previous post, blood was drawn 16 days after a FMD, but my previous analyses were also done about 2 weeks after an FMD and the comparison shows increased unfavorable values..... 

Link to comment
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.

Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

  • Create New...