Upcoming Statistical Analysis: Macronutrients, Calories, and Weight Loss

I’ve been on-and-off low carbohydrate dieting for a few years now.  My personal problem is lack of discipline.  The trend goes like this:  I realize I’ve put on more weight than I meant to, I get serious about eating better, I cut excess sugar, alcohol, and useless carbohydrates from my diet, I drop some weight, apathy sets in, I get bored following a rigid diet, repeat.

My rationale for following a low carbohydrate diet is quite simple, despite its multi-step formula:

  1.  Eating triggers the pancreas to release insulin in varying amounts depending on the macronutrient (sugar=high, fat=low, protein=sustained medium).
  2. Insulin delivers the broken down components of food (glucose, fatty acids, etc) to our cells to use for energy and storage
  3.  Excess eating triggers excess insulin
  4. Over time, muscle cells become resistant to insulin delivery, which leads to higher levels of insulin in the blood, along with higher levels of glucose in the blood.  This eventually leads to metabolic syndrome, and for many, type 2 diabetes
  5. High levels of insulin causes muscle cells to prioritize glucose uptake, rather than fat metabolism.
    • High insulin encourages free fatty acids to be stored in fat cells.
    • High insulin discourages stored fatty acids (triglycerides) from being broken down and released from fat cells
    • This creates an environment where stored fat tends not to be used for energy, which leads to fat accumulation
  6. Limiting the initial secretion of insulin prevents this fat accumulation in fat cells
    • This can be done by caloric restriction
    • Theoretically, it could also be done by eating foods that are less insulinogenic (insulin secreting)

Over the past few months, I have become receptive to the idea of time-restricted feeding (also sometimes referred to as intermittent fasting).  One problem overweight people can face is that their base insulin levels tend to be higher than non-overweight people.  It can literally be the case that an overweight person’s fasting insulin level is higher than a thin person’s post-eating insulin level!

*Source:  Twenty-Four-Hour Profiles and Pulsatile Patterns of Insulin Secretion in Normal and Obese Subjects; Polonsky, Given, and Cauter

One strategy I recently tried was to use time restricted feeding (a daily routine of going 16-20 hours without eating – eating within a 4-8 hour window) without limiting carbohydrates.  This did not work very well, and I found myself stagnating in weight loss while I was doing it.  When I combined low carbohydrate eating with intermittent fasting, weight loss resumed.  One thing I do not know is whether my caloric intake was high when I did this (I suspect it was).

As I mentioned above, I tend to lose momentum in my weight loss efforts, and then gain the weight back.  So, I thought I would add some extra incentive for myself by adding an extra layer of intrigue to the process.

One of the things I have wondered for a long time is how caloric intake relates to this weight loss process, versus macronutrient composition.  For example, if I eat 2500 calories comprised mostly of fat (say, at least 65% fat), will that affect weight loss (or gain) differently than if those 2500 calories came mostly from carbohydrates?

Different camps of researchers have different opinions about this, but the majority of researchers (at least from what I can tell) advocate a calories-in-calories-out model, rather than assuming the above insulin model implies one can increase their fat proportion of their diet to lose wight.

My gut instinct, based on research from Kevin Hall (source 1), is that the differences in weight change will be negligible, regardless of macronutrient composition (given consistent caloric intake/deficit) – one article Hall authored even suggested fat restriction is more effective than carbohydrate restriction; I suspect the reason low carb diets tend to be more successful than other diets has to do with satiety – high fat diets lead people to be less hungry.  It also seems to be the case that it is much easier to overeat high carbohydrate food than it is to eat high fat food.  However, Hall has also stated that protein metabolism is different enough from carbohydrate and fat that people can eat more protein calories than the other macronutrients.

I have to admit, I am drawn to the idea that we can alter macronutrient composition and be able to eat more than we otherwise would in a high carbohydrate diet.

My plan is to build statistical models to answer the following questions:

  1.  Is caloric intake predictive of daily weight change?
  2. Does macronutrient composition data improve the models I create?
  3. Is macronutrient composition a better predictor of weight change than caloric intake? (especially when caloric intake is high)
  4. Is there anything surprising from the (to-be-created) models?

This process will require tedious calorie counting, food weighing, and abstinence from eating foods that I don’t know enough about.  I’ve been collecting data for 5 days, and the data I have is shown below.  Once I have more data, I can begin building statistical models.  The trouble is my current data tells me very little right now, as my low carb+time restricted feeding has led me to lose weight every day so far.  Clearly my basal metabolic rate is more than 1800-2000 calories per day…at least when time restricting and limiting carbohydrates.


My goal is to collect many different days with varying caloric intake, time restricted feeding, and macronutrient composition.  As the dataset grows, the model will be more reliable…at least for my own body chemistry.

Author: Tim...Stepping Out

Tim Stepping Out

13 thoughts on “Upcoming Statistical Analysis: Macronutrients, Calories, and Weight Loss”

  1. I’d say the best way to improve glucose utilization, sensitivity and insulin spikes is by exercise (both aerobic and anaerobic). I’ve been doing IF recently and I was surprised at the result although at the exchange of loss of muscle mass which I don’t like. I agree that reducing the intake of high glycemic index food stuffs reduces the insulin spikes since high insulin levels has a nasty effect on the brain by increasing serotonin release which makes a person feel sleepy.


    1. One of my struggles is that I have NEVER felt great on a low carb diet…regardless of actual food intake. Conversely, I feel great after a big bowl of rice and chicken. Exercise when fasted is also great for insulin sensitivity…especially into a deeper fast (>16 hours). The rationale is that insulin falls while growth hormone spikes, which is helpful for muscle maintenance during caloric deprivation…something that simple caloric reduction does not do well.


      1. As an Asian man whose staple food has always been rice, I can definitely relate to the struggle. But it did paid off this month after restricting my carb intake into lower quantities and choosing ones with low glycemic index like brown/black rice, quinoa and oat bran. I don’t think I can go full vegan but I’m currently just eating vegetables with milk and eggs. I’m sure vegans consider this as a blasphemy.


  2. The classic experiment was done with a couple of Scandinavian explorers. Under the auspices of a major medical hospital, they ate (actually one dropped out) noting but fat and protein for a year (I think it was 80% fat). At the end of the year the blood workup of the guy who made it that long was identical to what it was when they started. It is mentioned in Taube’s “Good Calories. Bad Calories.”

    Liked by 1 person

  3. Unlike you, I’m a T1 diabetic and am extremely motivated to stay on a VLC diet (almost Carnivore). Over the past year, I have reduced my insulin by 50%, gotten my HbA1c’s into the non-diabetic range and lost around 30 lbs., no longer take a myriad of pills I was once on and feel so much better.

    I need variety in my diet and cook something new almost every day so nothing is boring (I’m also the cook in the family, so I need to make a tasty meal for non-Keto family members as well). It took me about 3 months to adjust to this WOE (Way Of Eating) but it is sustainable. BTW, I gave up most grains, most fruits (except berries) and all ‘seed’ oils which are inflammatory.


    1. One of the things I find difficult to separate is the weight loss difference and the health benefits. In my mind, Keto is the best first option for t2d, despite what seems to be some pretty odd recommendations from nutritionists. I really have no idea about t1d, but apparently it’s good for at least some.

      All I’m really trying to do here is determine if, for me, keto is more effective, calorie-for-calorie, than other macronutrient composition


      1. In my mind, Keto is the best first option for t2d, despite what seems to be some pretty odd recommendations from nutritionists.

        No it’s not for a couple of reasons. Insulin levels is not the problem with DMT2 but the lack of receptors in cells and since low glucose means high HGH, GH might actually make DM worse even though it’s promoting fat burning. Still the best way to increase insulin sensitivity is through exercise and medication. Keto might trigger ketoacidosis which could be life threatening.


  4. I have another theory that has so far been very effective in its application.

    I have impaired glucose tolerance, I have fasting glucose levels near pre-diabetics level.

    I decided to try a low carb diet by using nuts as my fill gap, and replaced potatoes and other starches like pasta and rice with falafel (chickpea based and homemade) while maintaining my normal intake of free range lamb and full cream dairy. The results were quite remarkable, for about 4 weeks I barely needed to eat as I lost about 5 kilograms around my tummy. I got less migraines and found I had more energy to want to exercise more.

    After the first 4 weeks my appetite returned but i maintained my low glycemic/low carb diet and have maintained the same weight for the last two years.

    The reason it works I believe is not what you think, in fact it is the opposite. Insulin is often portrayed as the bad guy, but it is an essential hormone for managing the level of glucose in the blood. Human populations who are adapted to high starch diets have high insulin sensitivity.


    More copies of the amylase gene make you more insulin sensitive, and this helps you manage your blood sugars.

    After a few months on my new regime I thought I would have reversed my pre-diabetes, however my fasting blood glucose was the same, my lipids dropped, and most of the inflammatory markers had returned to normal. I was puzzled, because I had assumed weight loss would equal decreased insulin resistance. I then realised that my high fasting glucose was not caused then by insuiin resistance, but low insulin production. My low carb diet may have resulted in weight loss, but this was a side effect of keeping my blood glucose in check.

    I figured out that the whole issue is not about calories in and calories out, but managing blood glucose. The whole reason we secrete insulin is to keep our blood glucose in the right zone, and if it gets too high or too low our body is not happy. In the case of high blood glucose, this leads to small arterial damage. If cells are stressed they emit signals, ie inflammation responses. Weight gain is then the result of the body reacting to the inflammation, and the reason for this is probably protective. The inflammation chemicals stop our abdominal fat cells from releasing their energy between meals, as the liver is still struggling to process the energy. This inflammatory state means we metabolise high after meals, but drop to very low metabolism between meals.

    Understanding this process led me to refine how I needed to eat, and the simple rule was: do not eat more than a glycemic load of 30 grams (glucose equivalent) in any three hours. To this end, I have added some extra rules to the usual GI table, to include alcohol, fructose and polyunsaturated fats as being equivalent in an inflammatory sense.
    This means I get my energy from resistant starches, proteins, saturated and mono-unsaturated fats. This diet of predominantly chickpeas, coconut cream, nuts, lamb, low-starch vegetables, egg and dairy provides sufficient energy without causing spikes of blood glucose. I also avoid foods that are excessively sweet, which means no low/no-calorie sweeteners either (no stevia!). The science behind the dangers of excessive sweetness is not mature, but there is evidence that non-calorie sweeteners impair insulin sensitivity through as yet undetermined means (probably physiological).

    This theory then explains why some people have no problems with a high carbohydrate diet while others do. People with poor insulin sensitivity (for example people with less amylase gene copies) are more prone to higher and longer spikes of blood glucose that causes an inflammation response. This inflammation stops the visceral fat cells from releasing energy between meals, and causes a retarded metabolism between meals and weight gain above the normal weight set-point.
    As the condition progresses, insulin resistance develops when fat cells get full and fatty liver develops. This (insulin resistance) being a secondary effect that reinforces the condition. It is at this stage, doctors generally diagnose the problem, hence why most doctors would assume insulin resistance is the initial cause.

    So, to sum up:

    Insulin is good, but if you don’t produce enough you need to manage your blood glucose with a glycemic aware diet.

    A low carb diet is such a diet. I would not recommend going keto because it is very restrictive and difficult to maintain.
    Legumes are a good source of energy if you can tolerate them, as resistant starches are certainly safe for people with poor insulin sensitivity.


    1. I don’t mean to oversimplify the role of insulin. Insulin can be very useful in the right context. I was simply talking about insulin in the context of obesity, not necessarily concerned about its role in glucose management (which admittedly is its primary function).

      I agree the more copies of amy1b are probably very useful in carbohydrate metabolism, especially if that translates to salivary amylase secretion.

      Have you had your insulin levels tested? The bulk of the research I’ve seen with insulin as it relates to obesity is that fasting and post-prandial levels are too high, but I suppose that wouldn’t necessarily mean that the pancreas is always over secreting…


      1. Yes, I mean to request a fasting insulin test next time. I don’t expect it to be low, just normal, as it is not so much that I cannot produce sufficient insulin, but that it responds to glucose intake too slowly.

        High fasting insulin would indicate cellular resistance, and this occurs as diabetes melitus progresses (even if the underlying issue is a sluggish response normally). Ie, when insulin resistance occurs, the pancreas might still be responding slowly to glucose increases, however it will keep plugging on and increasing insulin levels until glucose levels drop. Insulin resistance means more insulin is required to cause the lowering effect, so insulin levels will tend to stay high and go even higher even if the response is slow.
        The third stage of diabetes melitus would be the pancreas failing, and insulin levels plateauing or dropping off and glucose levels going to extreme dangerous levels.

        Note that my fasting glucose is still high, and that is because my body can make it from protein (as well as my resistant starch sources) and in the morning, the ‘dawn effect’ sees the liver release it. This morning peak is still below the threshold of inflammation and is harmless, however I am more wary of carbohydrates in the morning for this reason.

        I do not believe excess insulin plays any significant role in obesity for a few reasons.

        1./ Insulin is also an important player in metabolism, it acts on thyroid hormones to activate them
        2./ In general, the pancreas secretes the right amount of insulin if it can, sometimes too slowly for some of us, but rarely ever too much. (except in the rare condition of an insulin secreting tumor or from some drug’s side effects)
        3./ The body has many mechanisms to manage a constant weight regardless of type of carbohydrate and fat intake. A good artlcle about weight set-point theory is that of Gina Kolata’s : https://www.nytimes.com/2007/05/08/health/08fat.html. It demonstrates the scientific history of this theory.

        Diabetes is the exception to the weight set-point theory above, where the normal feedback effects are being dulled by another factor, this factor I believe must be inflammation.

        The factors that lead to diabetes are:

        1./ Refined carbohydrates that release glucose very quickly (ie high GI)

        2./ Excessive sweetness in foods that for some reason cause temporary insulin resistance (and maybe long term decreased insulin sensitivity). My hypothesis is that sweetness warns the body of fructose (regardless of what the sweetness is), which non liver-cells should not take up – perhaps cells are being hyper-vigilant to fructose which is like putting extra security at the gate?

        3./ Increased intakes of polyunsaturated fats – these are prone to oxidation and therefore add to the oxidative stress of a high glucose environment and therefore more inflammation.

        4./ deep fried potato based foods – these are so damned tasty that you are very likely to take your blood glucose levels into the stratosphere even if you generally have good insulin control. While the oil may slow down the absorption, this is more than countered for by the amount you eat. The fats (even if they are not polyunsaturated) then become part of the problem, taking up the livers resources and therefore causing the dangerous glucose peak to last longer.

        5./ genetic factors then pre-determine our susceptibility. For those of us with impaired glucose tolerance, two pieces of toast is likely to get us to the threshold of inflammation.

        I am sorry for going on too long, however my theory is actually quite basic – as you said, insulin’s primary role is managing blood glucose, we have become confused with all these other issues about calories and obesity, but blood glucose is everything. Manage your blood glucose and you are putting the horse in front of the cart!


Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s