glucose versus cholesterol optimization

[mccoy]

The System allowed me to paste this figure, which is very much IT with the thread. From the Kessler et al. article cited in the sigma nutrition blog.

• Published: 08 March 2017The effect of diurnal distribution of carbohydrates and fat on glycaemic control in humans: a randomized controlled trial

 

From the graph it is evident that a HF (high fat/Low carb) dinner is always very beneficial, in both groups with impaired and non-impaired Glucose tolerance.

Bottom line:

The optimal distribution of carbs and fats across the day may so constitute a viable strategy to decrease BG concentrations.

Now, shall the higher fats concentration at dinner impinge too much on cholesterol levels? Maybe not so much if the fats are taken overall not high and unsaturated.

 

This is something very worth trying, maybe having 3 meals a day where the last one is definitely hypocaloric and low-carb

 

 

 

 

Edited June 21, 2021 by mccoy

[Ron Put]

I have seen some diurnal research but have never been convinced enough to make a change. I time-restrict to 6-8 hours on most days, starting at about noon or early afternoon, since I am always hungry at night and not hungry in the morning.

But perhaps I'll shift my nuts and avocado (and flax?) to the end of my day and see if it makes a difference in three months.

[mccoy]

I'm definitely trying an HC/HF regime and report back. 

 

 

[mccoy]

I just discovered that the Freestyle-Libre CGM from Abbot has a cost of about 110 EURO/US$ per month. Compared to the 500 per month of the DEXCOM 6 systems used in T1D patients. Now I really start being interested, only concern being the point of insertion of the sensor and how to avoid it gets knocked off accidentally.

A bog from Peter Attia discussing the advantages of wearing a CGM

Are continuous glucose monitors a waste of time for people without diabetes?

Edited June 21, 2021 by mccoy

[mccoy]

The Dexcom monitor, used by Peter Attia, has the distinct advantage of providing a real-time glucose curve without any intervention, whereas in the Abbot device values must be read and stored in the cellphone scanning the sensor each time. The Dexcom device is larger and costlier, but much more convenient.

Both devices have a MARD of about 9%, MARD being a measure of accuracy: MARD= MD/Mean where MD= E[|X-Y|] Mean = arithmetic mean.

It means, if I'm not wrong,  since I'm not familiar with this dispersion value, that the average absolute  difference between the measured and the real value is 9% of the mean.

In a few words, if the true value is 100 mg/dL, the GCM can uderestimate or overestimate it on the average by 9 mg/dL ( 91 to 109). Actually, not too precise in terms of thresholds. A 18% error bar.

I wonder if repeated scans a few seconds apart, with averaged results, may yield a more precise measurement.

 

[Ron Put]

1 hour ago, mccoy said:

In a few words, if the true value is 100 mg/dL, the GCM can uderestimate or overestimate it on the average by 9 mg/dL ( 91 to 109)

Hah, I saw 20% error for interstitial fluid sensors. It's quite an error, and I thought I saw that number also for the optical/RF sensors various companies are working on (Apple has a patent on a terahertz sensor measuring RF changes in interstitial fluid).

But I had looked at Dexcom's page and they are talking about 20% either way:

DEXCOM G6 CONTINUOUS GLUCOSE MONITORING SYSTEM (DEXCOM G6) READING AND METER VALUE

[Sibiriak]

15 hours ago, mccoy said:

This is something very worth trying, maybe having 3 meals a day where the last one is definitely hypocaloric and low-carb

Certainly worth trying, I'd say.  

In fact, I'm going to tweak my own regime to shift  a bit more calories/carbs to breakfast time,  even though  I haven't got into regularly tracking glucose levels. Just makes a lot of sense, based on the evidence, and I can't see any downside.

 

 

[mccoy]

With regard to the accuracy of glucose monitors, I found that there is an international ISO standard on them and specific articles. The portable BGM, the traditional ones used by many of us, exhibit different accuracy, the best one (Freestyle Lite from Abbot) has a MARD of about 5%. My Accucheck Aviva has a MARD of about 7%, others have MARD=9% or more.

In my case, with a 7% MARD, I must always remember that one single measure of fasting glucose has this + -7% average error, so if my reading is 100, the real value might be in between 93 and 107. So, if I want to be reasonably sure I don't go over the 100 mg/dL, pre-diabetic threshold, my reading should be no more than 93 or 94 mg/dL.

Unless I decide to repeat the measurements and average them.

There is a distinct advantage in buying the most accurate one, Freestyle Lite, may be I'll replace mine.

 

Diabetes Technol Ther. 2012 Apr; 14(4): 330–337.

doi: 10.1089/dia.2011.0170

PMCID: PMC3317395

PMID: 22176154

Accuracy Evaluation of Five Blood Glucose Monitoring Systems Obtained from the Pharmacy: A European Multicenter Study with 453 Subjects

[Mechanism]

Sibiriak your quote by Dean & summary sums it up nicely. Other avenue is lowering carbs & substituting with MUFA sources. That’s pretty much it, the rest is following  CGM and lipid panels to ascertain personal data.

[mccoy]

10 hours ago, Mechanism said:

hat’s pretty much it, the rest is following  CGM and lipid panels to ascertain personal data.

Yes, in that context I wonder if it is better to follow some kind of prioritize-and-execute strategy.

Since BG is relatively easy to monitor, maybe it makes sense to prioritize it, trying to keep it as low as possible with less drastic action at the beginning, more drastic if there are no results.

At the same time moderating saturated fats. I wonder if MUFAs are totally neutral as far as cholesterol goes, since some people in this forum seem to have cholesterol-increasing effects from it.

Whereas, PUFAs should be totally neutral with regard to cholesterol. And the effect of fats on blood glucose can be monitored frequently.

Then, once glucose has been optimized, we might monitor & intervene on cholesterol to make its concentration as low as possible, at the same time without disrupting the optimization of the glucose strategy.

In my case, there is the additional challenge/constraint of keeping a lean BMI >24

I like the above, especially if I can get hold of a CGM.

Do they need medical prescriptions in America?

 

Edited June 23, 2021 by mccoy

[mccoy]

The optimization strategy should include some initial targets to build a system of parameters values, which can be later revised, such as the following (my personal case).

LDL-C may be substituted by non-HDL cholesterol or APOB or LDL-C, substituting the threshold value chosen

• 70<BG<150
• 70<Fasting BG<90
• 2hr postprandial BG<140
• 24<BMI<25
• LDL-C<90
• Trygs<100
• S-Blood pressure <= 120  
• D-Blood pressure <= 80

Many of you guys are already within this optimization, I guess, but don't have the rigid BMI constraint I have.

Edited June 23, 2021 by mccoy

[Mechanism]

Sibiriak advice, as usual is very sensible. I'll give you my two cents but keep in mind my goals are not the same.

Yes CGMs in US require prescription.

Yes, the higher BMI you target the less "room" you have without pharmacotherapy, or for a given level of pharmacotherapy

Peter Attia targets <135 mg/dL post-prandially always and ideally <130. I am more stringent: the closer to 100-115 the better, but this is not realistic for most individuals: you have to like what you eat, and your metabolic state, BMI, genetics, etc will all have an impact. 

Impact of PUFA/MUFA depends on total calories, genetics, and what is being substituted out but as a rule-of-thumb, MUFA typically mildly reduces LDL, PUFA to a greater extent.

Which is worse: postprandial peaks vs. lipid profile depends on genotype to a degree, and especially the extent of departures. For all the controversies on cholesterol there is good reason to be conservative and follow guidelines - or in many cases more stringently if possible. For example for BP a BP of 110/70 or lower, to a point, is typically better than 120/80 but there are always exceptions to all the rules, eg, lower BP may be associated with mortality and cognitive decline - part of this association is from confounding but at the same time if you have impaired cerebral perfusion from atherosclerosis later in life, a beneficial rule-of-thumb may be a liability for that situation. Bottom line, check with your doctor or get several second opinions when in doubt since medicine is so personalized. Many doctors are not versed at all in this arena, so your own due diligence and research is very helpful as a quality check and source of dialogue for these conversations with your team.

Edited June 23, 2021 by Mechanism

[mccoy]

Good points, mechanism. 

In Italy CGMs are apparently without prescription, and Abbot offers a free sample sensor (Freestyle libre)to those who want to test it. It's good for a 2 weeks period. Pretty good encouragement to its use.

I'm in the process of ordering it, I'll report back as soon as I have the first results.

[Ron Put]

I am surprised that CGMs are approved for anything other than broad trend monitoring, given the 20/20 error rate.

I came across a rather instructive (and overall positive) first-hand account:

The disadvantages of continuous glucose monitoring

 

[mccoy]

Ron, I remember an accuracy of 9/9 but I'll go check it more closely.

The lady in the article has T1D, so she's pretty much dependent on these devices for her survival, may have different requirements than people who do that for health reasons. Besides, she wears the Dexcom system which is the best presently available, with really continuos measurement even at night.

I'm pretty sure that the accuracy issue can be worked out by multiple sequential measurements. You average them and, unless there is a systematic bias, the error compensates. Even if there is a systematic bias, it can be figured out and corrected.

Yes, the Freestyle Libre is just a convenient substitute for a finger stick, a very convenient one though. 

Anyway, that's a good question for Peter Attia, with his degree in applied mathematics it's one of the first things he must have thought about.

 

Edited June 25, 2021 by mccoy

[Ron Put]

10 hours ago, mccoy said:

Ron, I remember an accuracy of 9/9 but I'll go check it more closely.

 

Hi, mccoy. I actually had posted a link from Dexcom that describes the "20/20 rule":

Compare the meter and your Dexcom G6 to see how closely the numbers match each other: if your G6 reading and meter value are within what we call the 20 rule (also known as the %20/20 rule), they match closely.

The rule

To use the 20 rule follow these steps using the table below:

1. Look up your meter value in the green middle column.
2. The left G6 – column shows the low range for a G6 reading that’s a close match.
3. The right G6 + column shows the high range for a G6 reading that’s a close match.

For example the orange highlighted row shows that if your meter value is 100 mg/dL, your G6 reading is a close match if it’s between 80 and 120 mg/dL.

It jives with the other estimates I had seen for CGMs and it led to my confusion earlier, since it's the same error I had seen for the optical/RF sensors that Apple, Samsung, and others are working to bring to market. Most of those use interstitial fluids as well, but without the prick.
 

[mccoy]

Yes, by reading those instructions, the so-called 'close match' wouldn't sound so close... But the issue is probably not so simple, the meter is not a precise value itself, having errors ranging from +- 5% to +-9%. So what do they say, they add the 2 error bars?. But it wouldn't make sense since any measures of dispersion are added by quadratic, not simple sums. Assuming a 9/9 error from both parts, SQRT(81+81)= 12.7, which is lower than 18.

More confusion: I listened to a previous AMA from Attia where he said the Dexcom exhibit a +-2% precision. Sounds really too precise to me. Was he seriously wrong, by an order of magnitude? He has a degree in engineering and applied mathematics besides being an MD, so I remain baffled.

One more issue: the blood sugar signal reaches the interstitial fluids after about 20 minutes, so, rigorously, the comparison should be made 20 minutes after reading the meter

But again, I'm trying to go deeper into this important topic.

Edited June 26, 2021 by mccoy

[Todd Allen]

7 hours ago, mccoy said:

More confusion: I listened to a previous AMA from Attia where he said the Dexcom exhibit a +-2% precision.

Precision and accuracy are different.  A meter reading consistently low would have poor accuracy but if it shows small changes it could have good precision.

[mccoy]

Todd, thanks for your precising.

In all the above posts, I assume minimum sistematic error, that is reasonable accuracy, that is, the meter does not usually reads too low or too high.

All the contradictory data seem to lie in the precision, or presumed precision of these devices. This does not rule out the potential presence of inaccuracy of course.

But, if the precision is too low, everything sound unreliable, which sounds contrary to the use of the CGMs by T1 diabetics, whose survival is really dependent on the readings.

I have a relative with T1D, who is a medical doctor, who wears a dexcom device and it seems to be reasonably accurate for the purpose. So, maybe our requirement of extreme precision is essentially desiring an overkill.

But it is good to have clear ideas.

Edited June 28, 2021 by mccoy

[mccoy]

I just installed a Freestyle libre sensor and I'm reporting the details in a specific topic: 

So far so good. This definitely will impact my glucose control beneficially, since there is a behavioural feedback: I know when my meals cause a glycemic peak and can accordingly adjust my strategy real-time. Peaks also constitute an undesired outcome, so meals and food which will spike glucose will be avoided usually.

Keeping one parameter monitored all the time, it's going to be easier to monitor the other parameter, cholesterol. While optimizing carbs intake, I'm going to limit the intake of saturated fats and have cholesterol measured by simple finger pricks in the pharmacy. This will allow me to monitor cholesterol on a monthly or bimonthly basis.

After one week of glucose monitoring, I'm also going to take berberine and see the results.

[mccoy]

My new toy (the Freestyle Libre CGM) has inspired me to start a moderately low-carbs diet. It's not hard after all, but I had to redesign my meals and start back using cronometer.

I'm eating plenty of vegetables, little or no fruit (only in the morning), no cereals at all, little legumes, lots of nuts and seeds, about 50 grams of EVOO per day. Nonfat yogurt and greek yogurt, tofu, sometimes a sugarless quest bar or other lowcarb sweets at the end of a meal.

My first concern is that, notwithstanding the amount of vegetable fats, my energy input decreased (sometimes I feel bloated or nauseos) and am losing weight. I'm not being able to keep a level of 2500 kCAls at least. I'm eating lots of protein, but that's not enough to avoid loss of bodyweight with the decreased energy input.

Cholesterol is optimized by eating soy and only nonfat dairy. Not a big deal so far.

Pretty easy after all, but potentially problematic for muscle mass.

Edited July 9, 2021 by mccoy

[BrianA]

Nuts and protein both may not give you the expected listed calories on the product label:

 

The truth about counting calories

https://www.popsci.com/health/calorie-counting-apps-accuracy/

[Dean Pomerleau]

29 minutes ago, BrianA said:

Nuts and protein both may not give you the expected listed calories on the product label:

 

The truth about counting calories

https://www.popsci.com/health/calorie-counting-apps-accuracy/

There is an entire thread on the topic of nut and calories:

--Dean 

[mccoy]

By the way, I just checked, Cronometer still indicates 654 kCal for 100 grams of walnuts. Anyway, I'm not bothered that much by it. If I take cronometer as a benchmark, for example, I know that, for a particular average dietary and lifestyle regime, what cronometer says amounts to an average of about 2700  will tend to keep me at a constant bodyweight, increasing or decreasing accordingly.

If I change my regime, as I'm doing now, of course I'll have to use a different calories benchmark.

Peanut butter sounds like a good calorie boosting, healthy item. It fills up very rapidly though, so I'll have to experiment.

Those who practice ketogenic diets without saturated fats tend to eat huge amounts of EVOO and avocados, but I am not able to drink EVOO or gorge on avocados, unfortunately.

Another caloric item that seems not to spike my blood sugar is, surprisingly, cheesecake, or rather the Italian millefoglie cake, a very fatty concoction. It is rich in cholesterol though, and definitely belongs to the group of junk foods.

On the other side, very healthy, vitamin-C concentrates like kiwifruit spike my BG horrendously if eaten alone. I'm going to report the details in the GCM thread I opened.

[mccoy]

Today I managed to hit 2500 kCals according to cronometer, by eating 60% fats: 80 grams EVOO, 5 tablespoons peanut butter, various seeds and nuts. Hope I'm going to do better tomorrow.