Two Experiments in One
As soon as we started to discuss possibly attending #NLANashville Dave and I began to toy around with the idea of trying out a dual experiment, similar to the identical diet experiment, with a twist. Instead of doing the same experiment the two of us would be doing two different experiments – but both of us would be manipulating our cholesterol numbers intentionally. All while attending a lipid conference and trying to soak up all the knowledge there (because who doesn’t want to add more work on top of an already busy working weekend? Never doubt our masochistic nature).
Originally, two possibilities were floated for my experiment “track”: option 1) fasting for five days or option 2) high calorie ketogenic for five days. I ultimately decided on high calorie, as it would be more interesting to do, and I wanted to be able to concentrate on the presentations while at the conference (fasting longer term often makes me restless). As a result of this choice, I needed something I could easily prepare and eat 1) while traveling to Nashville and 2) while actually at the conference. Luckily, Chris Bair of Ketochow stepped in and offered to sponsor the experiment. This was helpful in a few ways as I would already have most of what I needed before I even left for the conference, it wouldn’t require any cooking, and I could prepare everything for the following day the night before and not have to think about it after that. With that out of the way, the experiment began!
Instead of starting out with fasting, as I did with my Feldman Protocol experiment I instead decided to start with Ketochow intake with baseline calories in the 4 days prior, with some added whey protein powder after day 1 as I had been craving lean meat after a full day of Ketochow only. Calorie intake ended up being around 1500 calories with the added whey. The fat source used was heavy whipping cream – 99 ml per shake, 3 shakes a day. The flavors used were banana and orange cream for baseline, and pumpkin spice caramel and strawberry for intervention (noted here, as some of them have slightly different nutrient info – all Ketochow used was version 2.1).
Since I was aiming to drop my cholesterol over the course of the following five days, the intervention was to increase the amount of fat while keeping all else the same. Unfortunately, I forgot to bring my whey protein powder with me, so I just went without it after day 1 of intervention. I also found that trying to cram 550~ ml of cream into a bottle for the shake ended up making the texture very…. unfortunate. In order to avoid this problem, I kept the amount of powder the same, but split it up into 6 shakes per day for a goal of 6000 calories (with 4000~ calories achieved on days 1-3, and 6000~ calories achieved on days 4 and 5). For those curious this involved 297 ml of cream for baseline, and 1.6 liters of heavy cream for the highest calorie portion of the experiment. The liquid nature of the fat this time definitely made this possible!
All results include data from my 12 1/2 day Feldman Protocol attempt in the shaded area for comparison – new results are highlighted. The correlations listed are for all data points between 8/20 and 9/24.
As expected, Total Cholesterol dropped from the baseline of 313 mg/dL to a new low of 171 mg/dL after 5 days of high fat feeding. This leaves me with a total drop of 142 mg/dL and a correlation of -0.967684 with the 3 day average of dietary fat before the blood draw.
Baseline LDL-C on the 4 days of Ketochow came out to 252 mg/dL and after 5 days of the high calorie protocol it dropped to 105 mg/dL for a total drop of 147 mg/dL. The correlation between 3 day average dietary fat remained high at -0.97497754. To note, this is the lowest LDL-C I have on record, regardless of diet (Standard American Diet, keto, or carnivore). It’s also worth noting that during the highest calorie days I was consuming 554g of total fat, 387g of which were saturated fat.
LDL-P did exactly as expected and dropped during the high fat/high calorie phase – from 3068 nmol/L to 929 nmol/L for a total drop of 2139 nmol/L. This is pretty much what I would expect from the drop in LDL-C but I was still pretty surprised it had dropped so low. I suppose I won’t complain! The correlation for this one was between LDL-P and the 3 day average of dietary fat plus a two day gap and resulted in -0.919329629.
HDL-C surprised me by starting out higher than is typical for me when at baseline calorie/fat intake – especially as fat was a bit lower than my norm during baseline. I have had a few cases of HDL around the upper 40s on record with a baseline diet, but it’s not the usual case for me. Even more surprising was that upon high fat feeding I got the highest level I have on record – suggesting that my baseline HDL-C may have shifted up during the protocol. I suspect this may be due to something that’s in the Ketochow, and I have my suspicions as to what, but I’ll save investigation into that for another time.
Regardless of the slightly higher baseline at 48 mg/dL HDL-C still did as expected and went higher during the high calorie ketogenic phase all the way up to 57 mg/dL. This is a total increase of 9 mg/dL which is about what I would expect. The correlation with the 3 day average of dietary fat was positive, at 0.815920699.
HDL-P, like HDL-C, went up during the protocol from 21.6 umol/L to 35.7 umol/L for a total increase of 14.1 umol/L and a positive correlation between the 3 day average of dietary fat (plus a two day gap) of 0.922353315.
Triglycerides were around normal baseline levels at 63 mg/dL which dropped further upon the high fat/high calorie feeding all the way down to 45 mg/dL – the lowest I have on record. Apparently, despite the high fat intake, I was able to use or store the fat-based energy quite efficiently! This was a total drop of 18 mg/dL and a correlation of -0.857910104.
Out of all the markers I had been looking at, I was most curious about what lipoprotein(a) would do. During my Feldman Protocol attempt in August, I was surprised to find that it seemed to be following the inversion pattern. As dietary fat intake went higher, lipoprotein(a) dropped lower. But, there were some questions as to why this could be – was it perhaps that having the high calorie phase immediately after a 7 day fast influenced it? Was it something else? Obviously, testing with Ketochow helped isolate out most other possible influences and once again lipoprotein(a) dropped like a rock upon the high calorie ketogenic phase. It also started off a bit higher than my typical 130-140 nmol/L range, but this could be due to the macro composition I was consuming due to using Ketochow + whey protein isolate (resulting in slightly higher protein via whey, and slightly lower fat than usual).
Regardless – despite starting out at 171 nmol/L lipoprotein(a) quickly dropped to an astounding 58 nmol/L after the high fat/high calorie phase of the experiment. Not only was this 7 nmol/L lower than the previous low from August, but it resulted in a 113 nmol/L drop in 5 days of high fat/high calorie feeding (a 66% decrease). Wow! Lipoprotein(a) can move!
Not only that, but the correlation between dietary fat stayed strong here, at -0.93368339.
It will definitely be worth exploring how energy status impacts lipoprotein(a) in the future, along with other possible influences on its levels day-to-day. This experiment (especially back to back with the other protocol) confirms that just like other lipid markers lipoprotein(a) appears to be slightly more dynamic than anticipated.
On the Locals
Although I didn’t get to do much (any) sightseeing while in Nashville, except to walk to the conference area from where I was staying, I did get to interact with some of the locals. I had several long conversations with my AirBnB hostess, Meredith, for example, who was surprisingly delighted to hear about what I did for work and what I was in town for. Within about five minutes of me arriving at the AirBnB we were discussing lipidology, and she even pulled up a recent lipid panel she’d had so I could discuss what the markers were referring to. Funnily enough, she’d also had a CAC (a test I’m fairly interested in) recently, citing that she wanted to see any signs of actual disease (I brought up Ivor Cummins after this, of course!). She’d often ask how the conference was going, and said she could discuss lipidology with me for hours as it was all so fascinating (something I would have, of course, been happy to accommodate if I’d had more free time!) and over all was one of the highlights of my time in Nashville.
On the Conference
There were many interesting presentations throughout the weekend, ranging from lipoprotein(a), to lipid disorders (I find that understanding how things look when they’re going wrong can help me figure out how they should work when they’re going right), and intermittent fasting was even mentioned during a presentation regarding dietary habits! There was also a section of vendors at the conference, including ones that offered genetic testing, and more in-depth lipid testing which I found intriguing. I’ll definitely have to do a little more research into those sometime in the future.
It was honestly a bit surreal to walk past people having a conversation and hear them discussing LDL, particle counts, and other lipid-y things in passing – it’s not often I get to experience that in person! I also got the opportunity to meet some interesting people, from clinicians, to nurses, and dietitians. I was also delighted to see that one of the complimentary goodies the conference was offering its attendees was the most recent volume of The Journal of Clinical Lipidology. I got to have a bit of a study session with Dave before the conference day started in earnest as we both read through the studies it contained. It ended up being pretty productive and (dare I say it) fun.
Over all, it was a great learning experience, and I’m glad I decided to go.
On the Experiment
I’d say, in this case, the inconvenience of having to drink 6 shakes a day was far outweighed by the data I got in the process. Lipoprotein(a) coming from a baseline diet to a high calorie/high fat phase provided some useful information that – of course – leaves me with even more questions, and possible future experiments in mind, for sure. Plus, it wasn’t too bad, as the very kind staff at the hotel where the conference was taking place offered to store my shakes in the front office fridge so they could stay refrigerated until I needed them. I’m extremely grateful, as this made the whole process much less of a hassle, and allowed it to go as smoothly as it did. The food served at the conference actually did look quite appetizing, but in this case the sacrifices of citizen science won out over the freshly carved meat they were serving.
Even with that said, I must say that my diet over the conference weekend was definitely more appetizing than what Dave was eating – as he expressed multiple times! I’m sure he’ll be mentioning that himself, however, in part 2….
WOW! that’s all I got! Totally amazing. Congrats to both of you my lipid spys…
I would be interested if you keep the high calorie going for say another 5 days and see what happens to TC and other markers. Do they stay “switched”? Do they return to baseline? Do they start to climb higher?
Anyway, great stuff!
That is a LOT of fat to consume in one day- you are brave!!- in regards to how that must have been on your stomach…did you experience any digestive discomfort? Did you monitor your weight? (I would be very interested to see if this type of protocol changed your body composition more than your weight, for the purpose of knowing if that extra energy is stored…) fascinating!
Hi Erin! 🙂
It actually wasn’t too uncomfortable, I think because of the liquid nature of the fat – plus there is some fiber in ketochow which may have slowed absorption somewhat. I didn’t experience any digestive discomfort, or feelings of being unwell. It felt pretty much like nothing.
I did monitor weight – I couldn’t monitor it as closely as I’d have wanted as there wasn’t a scale I could find at the airbnb I was staying at. But before the high calorie phase I weighed 145, and after I got back (immediately after the high calorie phase) I weighed 147. So not really significant. Over the weeks following I gained another 3 pounds – I think because my appetite was messed up from the protocol and I was consistently eating much more than usual.
I’ve gone back down to 143 since after I cut dairy to get control of my appetite again.
Fascinating Siobhan! It would be really interesting, although harder to control for, to compare consuming the same amount of energy from day, fatty meat, as the liquid protocol you folllowed, as far as how all the numbers changed. I’m curious if the food type, and not just the energy density and macro distribution, affects numbers differently…
It would indeed be very difficult! I tried to achieve 6000 calories per day for the other protocol I did, but it was pretty much impossible for me to get past 3000 calories per day when utilizing solid food, or at least it felt impossible! Perhaps 3000 cal of whole food vs 3000 cal of ketochow would work better just to make it more feasible.
Does Keto Chow regimen have enough fiber to keep bowels moving ok?
I don’t usually eat *any* plant fiber (usually eat a carnivorous diet) and don’t have any issues on that front. With Ketochow I was going a bit more often, but otherwise all was normal.
Is it the increased fat that triggers the drops or is it the increased calories? I typically consumer 100g of fat per day, if I doubled my fat intake, would I see a drop in LDL, LDL-P, etc.?
Oh! I missed this the first time around. Yes, I imagine if you doubled your fat intake for 3-5 days LDL-C and LDL-P would drop. So far there’s about an 80%ish success rate for the protocol, so chances are good.
Also yes, I believe it’s increased dietary fat resulting in less need to pull fat from storage. You can technically get a similar effect via high carb low fat feeding (by switching to glucose based metabolism) but I really wouldn’t recommend it.
Hi Siobhan …. have you seen this study?
Fasting increases serum total cholesterol, LDL cholesterol and apolipoprotein B in healthy, nonobese humans.
Sävendahl L1, Underwood LE.
Voluntary fasting is practiced by many humans in an attempt to lose body weight. Conflicting results have been published on the effects of food deprivation on serum lipids. To study the effect of acute starvation on serum lipids, 10 nonobese (93-124% of ideal body weight), healthy adults (6 men, 4 women, 21-38 y old) fasted (no energy) for 7 d. Fasting increased total serum cholesterol from 4.90 +/- 0.23 to 6.73 +/- 0.41 mmol/L (37.3 +/- 5.0%; P < 0.0001) and LDL cholesterol from 2.95 +/- 0.21 to 4.90 +/- 0.36 mmol/L (66.1 +/- 6. 6%; P < 0.0001). Serum apolipoprotein B (apo B) increased from 0.84 +/- 0.06 to 1.37 +/- 0.11 g/L (65.0 +/- 9.2%; P < 0.0001). The increases in serum cholesterol, LDL and apo B were associated with weight loss. Fasting did not affect serum concentrations of triacylglycerol and HDL cholesterol. Serum concentrations of insulin-like growth factor-I (IGF-I) decreased from 246 +/- 29 (prefast) to 87 +/- 10 microg/L after 1 wk of fasting (P < 0.0001). We conclude that, in nonobese subjects, fasting is accompanied by increases in serum cholesterol, LDL and apo B concentrations, whereas IGF-I levels are decreased.
PMID:539776 DOI: 10.1093/jn/129.11.2005
Yes, I believe I’ve seen that one a couple times. In fact I demonstrated it in my previous protocol attempt where I also fasted for 7 days 🙂
In a way fasting is the reverse of the Feldman protocol. The funny thing is that when Peter Attia did a 7 day fast his total cholesterol went from 121 to 90 mg/dL and his LDL cholesterol went from 64 to 37 mg/dL. How could this be?
Correct… Attia has mentioned elsewhere he’s on several medications (including lipid lowering therapy). There’s nothing wrong with this, but the truth is that once you throw medication into the mix there’s no knowing how the lipid metabolism will respond. It just adds too many confounders. I believe he also exercised during the fasting phase, which could also interfere with the results.
Yes … he exercised every day at almost the same level as when he is normally eating. I think he also takes metformin and I suspect he takes rapamycin.
Right… all of which introduces variables that likely haven’t been studied (or at the very least certainly not with the same combination of medications, etc). So I’m not surprised he didn’t get the expected results, I’m not even sure what results I’d expect with medications thrown into the mix. There’s just no way to know how it affected the results.
I also wonder if I can get these results but eating more fat without the extra calories. Could be overworried but don’t want to gain weight or lose control of hunger. Also do you think exercise would hurt the hoped for decrease in ldl?
Anecdotally I’ve seen a couple cases where exercise during the protocol did seem to interfere with results – a re-test with no exercise got the desired result.
As far as the extra fat with no extra calories (so all calories coming from fat?) I’m not sure – but for what it’s worth both times I’ve done the protocol the weight gain was insubstantial and came off pretty quickly. The appetite weirdness went away after a week or so and was largely helped by returning to basics. You could try it and see what happens though – but the point is to make it so you’re relying less on energy from storage so I’m not entirely sure how much of an effect it would have.
Great stuff Siobhan! I just posted my results after doing the protocol for 3 days. Seems like I am a LMHR although currently and as i did the protocol was not exercising regularly. Was wondering what your thoughts were on that and m thinking of doing something similar (5 days) to hopefully further drop my LDLs. I will do a mixed diet of pseudo keto chow (whey, cream, butter) blended up and whole food. Was wondering if the psyllium husk is a “must” cuz like u, m mostly carnivore. I felt really full with the psyllium husk added, so on the 3rd day i took it out. Also i know we are too stay away from mct oil and coconut oil, does this include coconut flour and aminos as well? Thank u for ur time and looking forward to your feedback in this post and the other one.
I don’t think the psyllium husk is necessarily a must – I believe Dave has done fat shakes which were just cocoa powder, stevia, and heavy cream (no psyllium husk).
Coconut flour, and aminos would be fine, I’d imagine. Or at least I’ve not heard of them causing issues for people.
I’ll look for your other comment 🙂
[…] If you haven’t already, be sure to first read Part 1 of this experiment by Siobhan Huggins. […]
I think there is something important here – but the experimental method & data presentation are impossible to interpret. What is the justification for averaging 3 days of fat intake? Is that axis in grams per day? Why is the horizontal/time scale non-linear with 2 day increments, then 20, then 5 day increments? I still can’t parse the methodology except in a sea of verbiage and personal anecdote.
The linear connection between the shaded and non-shaded regions appears spurious.
You don’t need to create a formal scientific paper but the basic layout would improve readability.
1) As mentioned in the Feldman Protocol post linked, LDL-C tends to follow average daily fat intake over three days. LDL-P follows 3 day average, with a two day gap. Thus for LDL-C it’s averaged to show that correlation.
2) Yes it’s grams, I’ll be sure to clarify that in future graphs.
3) It’s from two different sets of data, as stated in the intro to that section. The shaded area is from my previous attempt at the protocol, then there was a gap of 20 days, and then the non-shaded area is from this experiment. It is in different increments because I was getting more frequent tests in the first experiment, compared to this one. They’re on the same graph so people who’ve been following along can compare the two sets easily.
I’ll be sure to read the linked article – there’s no point in posting data (even if just on a blog post) if it isn’t easy to understand, after all. Thanks for the feedback! 🙂
[…] known cardiovascular risk marker. Levels are generally considered to be mostly genetic, however recent experiments have demonstrated this may not necessarily be true – and high values could be the result of […]
[…] Keto Chow has supplied both Siobhan and I some of their product for experiments (like the Tandem Drop), but this company (nor any other) is an affiliate or in any way compensates CholesterolCode or has […]
I was thinking of taking a lipid and Hba1c test again before and after.
I know that my cholesterol numbers are bad as per the doctor.
I would like to replicate the diet you did from 9-20-18 thru 9-26-18 to drop your numbers from 296 to 83.
Your video talk said you had white bread, processed meat and some supplements. I was wondering if you could help me by letting me know exactly how much and what you ate and at what times to have your numbers change in just a matter of days. I am just concerned that my Hba1c would go up.
Thanks for your time.
Hi Siobhan, thank you for sharing this data.
I am a little bit confused about exactly what you ate in each of the 5 days leading up to the blood test.
First, you say:
“Calorie intake ended up being around 1500 calories with the added whey”
But, then you also say:
“but split it up into 6 shakes per day for a goal of 6000 calories (with 4000~ calories achieved on days 1-3, and 6000~ calories achieved on days 4 and 5).”
And in the first chart displayed, the running average dietary intake of Fat is shown as ZERO on 8/20, 8/22, and 8/24.
Perhaps if you could list out the calories (and the macros) you ate in each of the few days leading up to the day of the blood test, that would be very helpful.
There are a couple things noted in the post which may help clarify.
1) On the graphs I note that the shaded part of the graph (the three 0 fat days you mentioned and 2 high fat days after that; 8/20 to 8/29) are from a prior experiment for comparison to the current one. The write up on that one is here.
Thus only the last two datapoints on each graph are from this experiment, with the prior data points in the shaded areas showing replication of the effect seen in a prior experiment.
2) The 1500 calorie portion was a four day baseline. Each day had a controlled calorie intake during the baseline, at 1500 calories after day 1.
3) The intervention was high calorie and ranged from 4000-6000 calories per day (hence the difference between this part and the baseline 1500 cal/d).
I hope that helps clarify! But, better yet, I do eventually intend to go back and release all data from all posted experiments as an excel file which would have the individual day breakdowns. When I do this I’ll be sure to announce when posts are updated so people can go look at the full data to clarify any questions they had.