Analysis of the USDA Cost of Food at Home database shows that fat is the cheapest macronutrient.
Protein is the most expensive macronutrient, however a reduced carbohydrate diet does not necessarily require an increase in protein.
Reducing the amount of carbohydrate and increasing the amount of fat in your diet is the most effective way to reduce your grocery bill.
One of the common concerns about eating differently from the norm is that it will be more expensive.
Apparently one of the reasons for the relatively low Recommended Daily Intake for protein of 0.8g/kg is that many people can’t afford to eat more protein.  One of the common criticisms of Paleo or the Banting Diet (LCHF) is that it will be too expensive due to the extra protein. 
To see if these concerns were valid I thought it would be interesting to see what the data has can tell us about the relative cost the three macronutrients, protein, carbohydrate and fat.
The chart below shows the cost per calorie versus the percentage of protein in the thousand or so foods in the USDA Cost of Food at Home database. 
Protein is indeed the most expensive of the three macronutrients. As you move to the right in the chart you can see that your weekly grocery bill will increase.
Average intake of both protein and fat in the United States decreased between 1971 and 2004, with an overall increase in carbohydrate. 
While from a nutritional point of view there area lot of good reasons for people to eat higher levels of protein, a low carbohydrate diet is not necessarily high in protein.
People aiming for therapeutic ketosis may aim for lower amounts of protein to minimise insulin.
Tim Noakes’ Banting diet recommends that people get between 20 and 30% of their calories from protein. He says that those with diabetes and / or insulin resistance issues should aim for the lower end of this range, while people who are active and healthy can aim for higher amounts. 
Practically it is difficult to eat much more than 30% to 35% protein from real foods.
The table below shows the relative change in cost if we were to increase our protein from current average levels back to 1970s levels, or to moderate levels such as the Mediterranean diet or even the higher protein Atkins approach.
As shown in the table below, the most expensive high protein foods tend to be seafood. For reference, the average cost of food across the more than one thousand foods in the database is $5.37/kcal.
While protein can be expensive there are some low cost high protein options available.
If you are willing to try organ meats you might get them even cheaper as they are often discarded. The cheaper organ meats also typically have a much higher nutrient density than the more popular muscle meats or even fruits or vegetables.
You often hear the term ‘cheap carbohydrates’, but does this mean that a diet of processed grains and sugars is the most economical way to fill your shopping trolley?
While sugar and corn starch are very cheap food ingredients per calorie, the analysis of the data suggest that a higher carbohydrate diet is actually more expensive overall.
The cheaper high high carbohydrate foods tend to be processed and calorie dense. While the most expensive high carbohydrate foods tend to be natural foods that have a much lower calorie density. The table below shows that someone switching from a typical western diet to a reduced carbohydrate diet could make some significant savings.
So if increasing the proportion of protein and carbohydrate both increase the cost of our food bill then what makes it cheaper? Yes it’s the other macronutrient, fat.
Increasing the proportion of fat in your diet while decreasing the carbohydrates will make your meals tastier, gentler on your blood glucose and cheaper. Not to mention the fact that people typically spontaneously consume less calories when they consume less carbohydrates.
You may pay a premium for coconut oil, butter or olive oil relative to corn oil which is the cheapest food ingredient, however these fats are still much cheaper than the other macronutrients.
It appears that the it’s the very cheapest ingredients that are so prevalent in processed foods – sugar, corn starch, corn oil, high fructose corn syrup. Regardless of cost you’re always going to have to make a value judgement on the nutritional value of your food.
Increasing the protein content of your diet will increase your grocery bill marginally.
While higher levels of protein may be ideal for people who are healthy and active, LCHF is not necessarily high protein, particularly for those who struggle to regulate their blood glucose levels.
The LCHF approach, with its combination of moderate protein, lower carbohydrates and high fat provides an optimal solution with respect to blood glucose management, nutrition and cost.
Dave Asprey’s recipe doesn’t include bacon, but if we add some in we get an increase in the protein and a sight decrease in the nutrient score. The net carbs goes down thought the overall insulin load goes up with the increased protein. Either way a great option for people trying to manage their blood sugars and optimise their nutrition.
A low carb helps reduce blood sugars and insulin levels.
Blood glucose control and improved metabolic health will help to reduce your risk of many of the diseases of modern civilisation such as diabetes, heart disease, stroke, cancer, Parkinson’s and Alzheimer’s.
We become insulin resistant when our body fat can’t store more energy. Once our adipose tissue becomes insulin resistant, excess energy is stored in the liver, pancreas, heart, brain, eyes and other organs that are more insulin sensitive. Once our body fat can’t hold the excess, we see increased levels of energy in our blood in the form of glucose, ketones and triglycerides.
Endogenous ketosisoccurs when we eat less food than we need. Our insulin and blood sugar levels decrease and ketones rise to supply the energy we need.
Exogenousketosisoccurs when we eat lots fat and/or take exogenous ketones. Blood ketones rise, but our insulin levels will also increase because we have an excess of energy coming from our diet.
While a low carb or ketogenic diet helps to stabilise blood sugars, most of the good things associated with ketosis occur due to endogenous ketosis, that is, when we drive lower levels of energy in our system.
If your goal is blood sugar control, longevity or weight loss then endogenousketosiswith lower blood sugars and lower ketones is likely a better place to be than higher blood ketones.
There has been a lot of interest and confusion recently from people following a ketogenic about ideal ketone and blood sugar levels. In an effort to try to clear this up, this article reviews blood ketone (BHB), breath ketone (acetone) and blood sugar data from a large number of people who are following a low carb or ketogenic diet to understand what “normal” and “optimal” look like.
low carb diets reduce blood glucose levels
Many people initiate a low carb diet to manage their blood glucose levels, insulin resistance or diabetes. As shown in the chart below, foods that contain less carbohydrate cause a smaller rise in our blood sugar levels.
If you are insulin resistant, have prediabetes, diabetes (Type 1 or Type 2) it makes sense to reduce the carbohydrates in your diet to the point where you can achieve the blood glucose levels of a metabolically healthy person.
If you cannot produce enough insulin to keep your blood sugars at the levels of a metabolically healthy person, it makes sense to reduce the insulin load of your diet to the point that your pancreas can keep up.
What are optimal blood sugar levels?
According to mainstream medical definitions:
“normal” blood sugar regulation is defined as having an HbA1c of less than 6.0%,
“Prediabetes” is diagnosed when you have an HbA1c between 6.0 and 6.4%, and
Type 2 diabetes is diagnosed when you have an HbA1c of greater than 6.4%.
However, as shown in the charts below, the mainstream definition of “normal” is far from optimal.
By the time you have “pre-diabetes” (with an HbA1c of greater than 6.0%), you are at risk of many of the most common western diseases of ageing and causes of death such as heart disease, stroke, Parkinson’s, Alzheimer’s and cancer.
If you’re interested in optimal rather than what passes for normal, the table below shows some suggested HbA1c and average blood sugar targets.
average blood sugar
In this video, diabetes pioneer Dr Richard Bernstein tells how he refined his understanding of optimal blood sugar levels by asking healthy your sale reps to demonstrate their blood sugar metres on themselves. They always seemed to come out at about 83 mg/dL (or 4.6mmol/L). This correlates well with the level at which we see the lowest level of risk of various common diseases.
ketosis vs. hyperinsulinemia
While high glucose levels are bad news in and of themselves (glucose toxicity and excessive glycation), high blood sugars and high insulin levels typically go hand in hand.
A metabolically healthy person will store excess energy in their fat cells, ready for easy access when required later. But if we continue to fill our fat cells with the excess energy they get to the point where they cannot continue to expand to absorb the excess energy. It is at this point that our adipose tissue becomes insulin resistant.
Ted Naiman does an excellent job explaining insulin resistance in this video.
The “good news” is that becoming insulin resistance will slow the expansion your fat cells. The bad news is that once our fat cells become insulin resistant the excess energy will be re-directed to the parts of our body that are more insulin sensitive such as our liver, pancreas, heart, brain and other vital organs.
Our pancreas works overtime, secreting more insulin to try to keep the energy in the liver and fat cells in storages while we use up the glucose in our bloodstream. High insulin levels mean that we will find it harder to release energy from our fat stores when we go without food. Without easy access to our body fat stores, we will be driven by our appetite to eat again sooner. Without easy access to their fat stores, someone who is insulin resistant may be driven to eat more often.
Fasting insulin levels in healthy populations tend to range between 2 to 6 mIU/L. The average insulin levels in western populations are 8.6 mIU/L. Meanwhile, the official reference range for “normal” fasting insulin is less than 25 mIU/L.Given that the western world is going through a crisis of metabolic health, it is safe to say that this cut off for ‘normal’ is also far from optimal.
how to manage your dietary insulin load
The good news is that you can tailor the insulin load of your diet to suit your current level of insulin resistance. As shown in the chart below, the insulin generated by our pancreas is proportional the net carbs in our diet plus about half the protein it contains.
Reducing the insulin load of your diet can help to reduce your insulin and blood sugar levels to the point where your pancreas (and any insulin resistance you may have) can keep up and maintain normal blood sugars. When our insulin levels are lower, we can access our body fat for fuel, either from our body or our food.
A person with type 1 diabetes is not able to produce enough insulin to keep energy tucked away in their liver, fat stores and even muscle tissue. Without exogenous insulin, people with type 1 diabetes see both their blood glucose and blood ketones rise to very high levels. This is called ketoacidosis which is dangerous and requires exogenousinsulin as soon as possible.
As shown in the image below of JL, one of the first children with type 1 diabetes to receive insulin, people with type 1 diabetes quickly regain weight with exogenous insulin administration.
Rather than minimising insulin, it’s important to find the optimal dose of insulin. Too much insulin can slow the release of fat from your body to be used for fuel. Too little insulin and we fall apart.
People who switch to a low carb diet often find that their blood glucose and insulin levels drop and they are not hungry because energy can more easily flow out of storage.
However, the problem with a verylow insulin load dietary approach is that it may not contain enough of the vitamins and minerals that you need to help you avoid nutrient cravings in the long term.
High-fat foods have a high energy density which can make it hard for some people to control portion sizes. This can be problematic if their goal is weight loss. Very high-fat foods also tend to be less nutrient dense and are less satiating than high fibre or high protein foods.
The most ketogenic foods also contain lower levels of protein which is critical to managing satiety.
We need to find the right balance between a lower insulin load diet that will stabilise our blood sugar while still getting the essential nutrients that our mitochondria need to thrive and produce energy efficiently. If you need help with this process, the Nutrient Optimiser is a useful tool to help you refine your diet to stabilised your blood sugar and insulin levels while also maximising nutrient density as much as possible.
blood glucose and ketones in fasting
When we go without food, our blood glucose levels decrease as the glucose in our bloodstream and liver (glycogen) are used up. With less glucose available our body turns to our body fat stores. Our liver converts our body fat to ketones for use in the brain rather than glucose. This is termed “endogenous ketosis” (endogenous = originating from within an organism).
As explained by Dr Sinclair in this video, there are many of beneficial things that occur during endogenous ketosis such as autophagy, mitogenesis, mitophagy, upregulation of SIRT1 and increase of NAD+). The body effectively goes into repair mode so it can make it through to reproduce in the next time of plenty.
The chart below shows my blood sugar and blood ketone (BHB) during a seven day fast. As glucose levels decrease ketone levels rise to compensate for the lack of energy.
The sum of glucose and ketones continues to increase during fasting. I have termed the sum of glucose and ketones as “total energy” (i.e. glucose + ketones).
Often in fasting, it seems that the body will let the total energy in the blood go quite high. This high energy state enables you to be at a very high state of alertness and be ready to find food to survive. The mental clarity after a few days of fasting great!
Exogenous ketosis vs endogenous ketosis
I think it’s important however to understand the difference between exogenous ketosis and endogenous ketosis.
Endogenous ketosisis when we eat less and our body is forced to use some of our body fat to make up the difference.
Exogenous ketosis occurs when we eat lots of fat (or take exogenous ketones), and we see blood ketones (beta-hydroxybutyrate) build up in the blood.
Higher levels of ketones can build up in the blood when you are eating more fat than you are burning. As a general rule, higher blood ketones are not a cause for concern as long as your blood glucose levels are also low (i.e. low total energy).
You will always be somewhere on the spectrum between exogenous and endogenous ketosis. We need to ingest enough energy to get through the day and not fade away. But keep in mind that most of the good things we attribute to ketosis and the ketogenic diet occurs due to endogenous ketosis when fat is coming from our body (e.g. autophagy, increase in SIRT1, increase in NAD+).
What should your blood sugars and ketone levels be in ketosis?
Our understanding of the ketogenic diet is still evolving. Keto Clarity co-author Eric Westman recently admitted that there is still a lack of clarity ideal ketone levels.
The problem, however, is that it’s hard for most people to achieve “optimal ketone levels” (i.e. 1.5 to 3.0mmol/L) without fasting for a number of days or eating a lot of additional dietary fat (which may be counterproductive if you are trying to lose weight).
Recently, I had the privilege of having Steve Phinney (pictured below in our kitchen making his famous blue cheese dressing) stay at our house when he spoke at a Low Carb Down Under event in Brisbane.
Since the publication of this chart in the Art and Science books, Phinney has noted that well-trained athletes who are long-term fat adapted (e.g. the athletes in the FASTER study) actually show lower levels of ketones than might be expected.
Over time, many people, particularly metabolically healthy athletes, move beyond the ‘keto adoption’ phase and are able to utilise fat as fuel even more efficiently and their ketone levels reduce further.
Urine ketone strips are often considered to be of limited usefulness because the body stops excreting ketones as it learns to use them. It seems that a similar thing happens with blood ketones (BHB) over a longer period of time.
As our blood glucose levels decrease, our ketones increase to make up for the lack of fuel. When we have less glucose available our insulin levels go down, more fat is burned, and blood ketone levels rise.
Crowdsourced ketone and glucose values
The chart below shows the sum of the blood glucose and ketones (i.e. total energy) from nearly three thousand data points from a broad range of people following a low carb or ketogenic dietary approach.
[In late 2015 I pooled a range of data from myself and a number of people on the Optimal Ketogenic Living (OKL) Facebook group. After sharing this data initially, a number of other people sent me their data. Later, Michel Lundell from Ketonix agreed to share an extensive set of anonymised data for me to analyse. I hope that this crowd-sourced data will help to provide more clarity about optimal ketone levels in a similar way to Dr Bernstein surveying his glucose metre sales reps helped to provide a better understanding of what normal ketone values are.]
On the right-hand side of the chart, we have a high energy situation from both glucose and ketones. While not as extreme, high energy situation is similar to someone with Type 1 diabetes with high glucose and high ketone levels due to inadequate insulin. High levels of energy in the blood causes the pancreas to secrete insulin to hold the glycogen back in the liver and stop lipolysis (i.e. the release of fat from storage) until the energy in the bloodstream is used up.
On the left-hand side of the chart, we have a low energy situation. These people do not have too much energy floating around in their bloodstream. They are also likely insulin sensitive and can easily access their body fat stores for fuel. In this low energy state, we see autophagy, NAD+ and SIRT1 increase.
As shown in my fasting ketone data above, we may see high blood ketone levels when we go without food for an extended period. However, trying to replicate high ketone levels with high levels of exogenous ketones or an oversupply of dietary fat will not necessarily provide the same positive benefits as endogenous ketosis.
Based on this crowd-sourced data it seems the body tries to maintain a blood glucose level of around 4.9mmol/L and a blood ketone level (BHB) of around 1.5mmol/L. The table below shows this data in terms of average as well as the 25th percentile and 75th percentile points.
total energy (mmol/L)
blood glucose and ketone levels during exogenous ketosis
People following a therapeutic ketogenic diet may be fuelling up with MCT oils to achieve high ketone levels and low glucose:ketone index values (GKI) value for therapeutic purposes. Others will target high levels of ketones for brain performance. Others will load up on exogenous ketones and glucose to ‘dual fuel’ the system for athletic performance.
One of the benefits of a low carb or keto diet is that it tends to eliminate a lot of hyper-palatable processed foods. Stabilising blood sugar and insulin levels can also help normalise appetite and cause people to eat less.
The danger, however, with trying to drive high levels of ketones primarily through exogenous ketosis is that it will lead to an energy excess which will drive insulin to slow the use of fat and glucose stores until the energy in the bloodstream is used up.
Even though we are getting most of our energy from fat rather than carbs, the body will still increase work to bring maintain lower levels of energy in the bloodstream using insulin, and stop the release of stored glycogen and body fat until you have used up the energy in your blood from your food. I think this is why Thomas Seyfried and Dominic D’Agostino talk about the benefits of a calorie restricted ketogenic diet. 
blood glucose and ketone levels for weight loss and endogenous ketosis
Many people embark on a ketogenic dietary approach to manage their diabetes and achieve long-term weight loss. They want lower insulin levels to enable them to burn more body fat for long-term insulin sensitivity and health.
On the left-hand side of the total energy chart below, we have a situation where we are generating endogenous ketosis, meaning that our own body fat is being burned for fuel. With lower levels of energy in our blood, our body needs to pull fat from our body fat stores as well as use excess stored fat and old proteins in our liver, pancreas, brain and other organs (i.e. autophagy).
To understand what all this means I have divided the three thousand data points into five quintiles, from lowest to highest total energy. For each quintile, I have calculated the average, 25th percentile and 75th percentile blood ketone (BHB) value. Half of the ketone values will fit between the 25th and 75th percentile values.
The table below shows the quintiles numerically.
average BG (mmol/L)
total energy (mmol/L)
In the discussion above we see that the lowest risk of the modern diseases of ageing and metabolic health occurs when we have an HbA1c of less than 5%. It seems that as a general rule (other than when we are fasting or aiming for therapeutic ketosis) being somewhere to the left of this chart is (in the first or second quintile) with blood glucose levels less than 5.0mmol/L (or 90mg/dL) is optimal.
In summary, it appears that the lowest blood sugar levels are associated with lower ketone levels and a lower total energy overall. Or to put it another way, higher ketones are not an indicator of better metabolic health.
Virta study ketone values
This crowd-sourced data above seems to also align well with ketone data from the results from the recently published Virta study (Phinney et al., 2017). In this study, they aimed to get their patients into nutritional ketosis (which was defined as blood ketone levels between 0.5 and 3.0mmol/L as per the optimal ketone zone chart above) using a ketogenic diet.
The distribution of BHB levels is shown in the chart below. The blood ketone level over the first 10 weeks of the study was 0.6mmol/L with a standard deviation of 0.6mmol/L. In spite of consuming a ketogenic diet under the supervision of Steve Phinney and the Virta team, many of the study participants did not achieve ketone levels that qualified as nutritional ketosis. The chart below shows the distribution of blood ketone levels.
In spite of what many might consider to be lower blood ketones, the reduction in HbA1c was still significant.
Many of the people on insulin were able to reduce or eliminate their insulin requirements.
Also, most of the participants lost a significant amount of weight.
Do ketone values change after long-term adaption?
Anecdotally it seems that people who are very physically fit may find they have lower blood ketone levels as well as lower blood sugar levels.
In fact, you could even save your money on testing ketones and just focus on making sure your blood sugars are closer to optimal. One of the simplest and most accurate blood glucose machines is the Abbott FreeStyle Lite.
If you want to test your ketone levels and blood glucose levels, then the Abbot Optium Neo is the one to go for.
If you don’t have health cover for diabetes, the test strips can get expensive if you want to check regularly. However, you can quickly get a feel for whether you are insulin resistant and which nutritional approach you need to follow.
Is it the ketones or NAD+ that is really doing the good stuff?
A smart friend of mine, Robert Miller, has been challenging my thinking recently around ketones. Robert’s theory is that the benefits that people attribute to ketones are actually largely due to an increase in nicotinamide adenine dinucleotide (NAD+).
Adequate levels of NAD are critical to moving energy around our body, from our food to our mitochondria. NAD+ declines with age, increases in fasting, during a ketogenic diet or in response to exercise. When NAD+ rises, SIRT1 helps our body to repair and improve our insulin resistance.
Our NAD+/NADH balance is critical to controlling our appetite and telling the body whether we need more fuel, or if it’s time to tap into our body fat stores.
We can increase our circulating levels of NAD+ by eating a nutrient dense diet, particularly with adequate niacin (vitamin B3). There is a range of NAD+ supplements that seem to have positive benefits.
Nicotinamide Riboside is a close precursor to NAD+ which can be helpful for people who struggle to make NAD+ directly from food.
NAD+ metabolism is a fascinating rabbit hole if you want to do some research, but for now, let’s say that it’s important to have adequate NAD+ to enable your mitochondria to convert the food you eat to energy rather than having it back up in your body.
What about breath ketones (acetone)?
You will likely be aware that three separate types of ketone bodies:
acetone (breath ketones),
acetoacetic acid, and
beta-hydroxybutyric acid (BHB in the blood).
Fatty acids come to the liver either from our food or lipolysis (mobilisation of fat from our body) and converted to acetoacetate. Unfortunately, we have no way to directly measure acetoacetate, only BHB in the blood and acetone in the breath.
“Endogenous acetone production is a by-product of the fat metabolism process… Breath acetone concentration (BrAc) has been shown to correlate with the rate of fat loss in healthy individuals… A strong correlation exists between increased BrAce and the rate of fat loss.”
The data below provided by Michele Lundell from Ketonix data shows a similar relationship between breath ketones and blood ketones but with a lot of scatter!
Most people at this point throw in the towel and go back to blood ketones (BHB) which appear to be more reliable. However, this may not be advisable given that breath acetone is likely a stronger marker of burning ketones for fuel, not just buffering energy in the blood. A higher NAD+:NADH ratio means that less AcAc is stored as BHB and is available to be oxidised, at which point we see more breath acetone.
What I’m trying to illustrate with the dodgy clip art in the chart above is that having high blood ketones and low levels of breath acetone appear to be a sign that you are not burning the fat you’re eating.
By contrast, a high level of breath acetone with a low level of blood ketones (BHB) and low blood glucose may be a sign that you are very efficiently burning the fat you eat.
If you had a choice, I think it’s better to have high breath acetone (BrAce) and lower blood ketones (BHB) (which would indicate that you were efficient at burning fat) than high BHB and low BrAc (which would indicate that you were good at eating fat but not necessarily burning fat).
The figure below shows BHB vs BrAce for the different purposes that people nominate for using the Ketonix. What we see is that people with diabetes (red trend line) have both higher blood glucose and higher blood ketone (BHB) levels (i.e. higher total energy). Fuelling the mitochondria with the right nutrients that increase NAD+ will both help to burn off the excess energy and manage appetite through the NAD+: NADH ratio.
The second law of thermodynamics states that the total entropy (chaos) of an isolated system can only increase over time. It seems that an organism that is ageing will decay and not be able it’s energy packed in storage, but instead we will see increasing levels of energy floating around in the bloodstream (e.g. diabetes).
What can we use blood ketones for?
So while my take-home message is that blood ketones are likely more useful than worrying about blood ketones, I think tracking BHB can still be useful in diabetes management.
You will remember that people with uncontrolled type 1 diabetes will have blood glucose and blood ketones that are sky high. This is called ketoacidosis.
The mainstream approach to treating diabetes is by ramping up the insulin to suppress both blood glucose and blood ketones. Unfortunately, without modification of diet to reduce the insulin load, excessive amounts of exogenous insulin will just slow fat burning as well as suppressing blood glucose levels.
Having low blood glucose and very low ketone levels (say less than 0.4 mmol/L) can just be a sign that you are taking too much exogenous insulin and need to back off the exogenous insulin dosing. If you are taking exogenous insulin, this would be worth checking periodically to make sure you’re not overdoing the insulin
Metabolic related diseases such as heart disease, cancer, stroke, diabetes, Alzheimer’s disease, kidney disease are the leading causes of death.
People with the best metabolic health (i.e. low HbA1c, insulin and blood sugar levels) have the lowest risk of dying from these common western diseases of ageing.
Keeping your processed and starchy carbohydrates low will help to keep your blood glucose and insulin levels and reduce your risk of obesity.
When we fast, our glucose levels decrease, and ketones increase to maintain our energy levels.
People who are metabolically healthy and insulin sensitive typically have a lower level of total energy (i.e. glucose + ketones) in their bloodstream.
We can mimic the rise in ketones with added fats or exogenous ketones that we see in fasting. However, the real benefits occur when the body is forced to draw on its own stored energy, and we experience autophagy, upregulation of SIRT1, and a rise in NAD+.
The benefits that we often attribute to ketones may also be due to increased NAD+ levels which occur in fasting and/or with a nutrient-dense diet.
We can only measure beta-hydroxybutyrate (in the blood) and acetone (in the breath). We can think the BHB as a buffer ready for use, and acetone as the exhaust showing that the ketones have been burned for fuel. Higher levels of fat burning with a lower need for buffering is a better place to be than a large buffer in the blood and minimal fat burning.
A therapeutic ketogenic diet has a very low insulin load from non-fibre carbohydrates and a higher amount of dietary fat to achieve higher ketone to manage chronic conditions such as cancer, epilepsy, alzheimer’s, dementia etc.
The chart below shows our insulin response versus insulin load which considered fibre and protein as well as carbohydrates. People wanting to following a ketogenic diet should eat foods towards to the bottom left of this chart.
We can quantify the insulin load using the following formula:
insulin load = total carbohydrates – fibre + 0.56 x protein
The chart below shows the nutrition provided by this high fat approach. The therapeutic ketogenic dietary approach does not provide the DRI levels of:
alpha linolenic acid,
Hence this style of therapeutic approach is idea for a shorter term intervention with a higher nutrient density approach being adopted when possible.
other dietary approaches
The table below contains links to separate blog posts and printable .pdfs detailing optimal foods for a range of dietary approaches (sorted from most to least nutrient dense) that may be of interest depending on your situation and goals. You can print them out to stick to your fridge or take on your next shopping expedition for some inspiration.
A therapeutic ketogenic diet can be helpful for a range of chronic health conditions such as cancer, epilepsy, Alzheimer’s disease or dementia.
Domonic D’Agostino is doing interesting research into the possible uses for ketosis, both through diet and supplementation. His initial funding was from the US Military to research the applications of ketosis for navy seal divers in order to avoid oxygen toxicity seizures.
He has continued this research into how ketosis can starve cancer and be used in conjunction with normal treatments to aid recovery from chemotherapy and slow tumour growth.  His more recent research demonstrates that body builders can maximise their power to weight ratio and recovery using a ketogenic approach.
Dr Mary Newport has received a lot of coverage after treating her husband’s advanced Alzheimer’s with coconut oil. 
Terry Whals is undertaking clinical trials of her high nutrient density ketogenic diet that has worked to reverse her own multiple sclerosis.
The ketogenic diet for epilepsy has made a resurgence since director Jim Abrahams  found success with the ketogenic diet for his son Charlie and then made a movie of his experience. 
Jimmy Moore’s Keto Clarity spends three chapters profiling the various conditions that the ketogenic diet has been claimed to be beneficial for.
Solid science (chapter 16)
Polycystic ovary syndrome (PCOS)
Irritable bowel syndrome (IBS)
GERD and heartburn
Non-alcoholic fatty liver disease (NAFLD)
Good evidence (chapter 17)
Schizophrenia, bipolar and other mental illnesses
Narcolepsy and other sleep disorders
Emerging areas (chapter 18)
Traumatic brain injury
Gum disease and tooth decay
Amyotrophic lateral sclerosis (ALS)
Multiple sclerosis (MS) and Huntington’s disease
Restless leg syndrome (RLS)
Alopecia and hair loss
GLUT1 deficiency syndrome
The therapeutic ketogenic diet is similar to the LCHF approach but takes it one step further, with net carbs typically restricted to 25g per day (or sometimes less) and protein restricted to the minimum necessary for muscle repair.
People trying to slow or reverse cancer growth or stop seizures will often also resort to more aggressive measures including supplementing with larger amounts of butter, coconut oil, MCT oils and ketone salts to drive their ketones to higher levels (i.e. 1.5 to 3.0mmol/L).
I figured we could use the food ranking system to prioritise foods with a low insulinogenic load over and above nutrition or the other parameters.
The table below shows the weighting I have used for this ranking. Still considering nutrient density, cost and calorie density will help to optimise these other elements of nutrition even though we are primarily targeting a low insulin load.
ND / cal
fibre / cal
ND / $
ND / weight
cal / 100g
$ / cal
I have also used a filter using Wilders’ formula to show only foods that have a ratio of ketogenic to anti ketogenic calories greater than 1.5. This is the commonly accepted parameter in therapeutic ketosis circles to determine whether a food or a meal is sufficiently ketogenic.
You could also use this calculator to check the percentage of insulinogenic calories of your food. If you’re aiming for therapeutic ketosis you’ll probably need to have an overall average of less than 15% and any individual food should ideally have a percentage of insulinogenic calories less than 25%.
The resultant foods are listed below. This approach will obviously prioritise the liberal use of fats and oils along with higher fat dairy products and meats.
Not all of the vegetables have a Wilder’s ketogenic ratio greater than 1.5 but it would still be desirable to include adequate vegetables for nutrition as long as they give they fit your tolerances whether they be net carbs, ketones or something else.
Someone using this approach may choose to supplement vitamins and minerals or use organ meats to achieve their nutrition in order to minimise carbohydrates from the vegetables.
People battling chronic illnesses also often have allergies that will mean that they further need to refine this list.
vitamins & minerals
I hope that these lists will be useful for people who need to maximise ketosis for therapeutic purposes, as well as possibly others with diabetes, insulin resistance or people looking to lose weight who want to use a more aggressive approach for a period.
My nine year old son was asking what he could have for breakfast that would be healthy.
This is pretty much the simplest and healthiest recipe that I could design that a nine year old can put together unsupervised.
I’m teaching him to just put some frozen spinach, eggs, cream and cheese into a microwave bowl.
Eggs are a wonderfully complete protein. Cheese and cream add to the taste as well as adding in other proteins and fat not in the eggs.
Spinach is such a great superfood with so many micronutrients while being low in net carbs due to the high fibre content.
You can add coconut oil or butter to get extra good fats into your meal to create something wonderfully indulgent, nutritious and that will be gentle on your blood sugars.
If you’re short on time frozen spinach is quicker than using fresh spinach. I figure if you’re going to skip the spinach because it takes too long to fry up, then it’s better to go with the frozen option. I will use frozen kale sometimes too which gives a slightly different taste and texture.
If I’m cooking for the family and have bit more time I’ll do the fresh spinach in the fry pan. If I’m putting this together for the family before we run out the door in the morning I will use some frozen spinach or kale.
With only 7g net carbs and heaps of well rounded nutrition (in terms of both vitamins and minerals and amino acids) it’s pretty hard to go wrong with this for any meal.
The stats for a 500 calorie serving are shown below. The net carbs are low, the fibre is fairly high with the spinach and most of the insulin requirement is for the slower digesting protein rather than the carbohydrates.
The more spinach the better if you want to maximise the nutrition. The highest overall score occurs when we we use 700g of spinach. Though this may be pushing the limits of optimising nutrition, not necessarily palatability or eating pleasure.
If you found 7g of net carbs raised your blood sugar you could certainly reduce some of the spinach, although most people find that counting net carbs in real food (rather than processed / manufactured foods) are OK (see the article fibre… net carbs or total carbs for more info)
Since working through the nutrition analysis comparing different meals, this meal has basically become our preferred go to meal, which you can make as simple or as complex as you want. If you’re an insulin dependent diabetic you could use this as a regular meal and refine your insulin dosing based what you see on your BG metre after the meals.
I’ve run the analysis below with the kale rather than the spinach. As much as people rave about kale, it actually does not do as well on the nutrient balance score as spinach! Spinach scores better in both the nutrient completeness and the amino acids. Kale also has more total carbohydrates and less fibre which makes less diabetic friendly.
It’s generally difficult for healthy people to eat too much protein. However the fact that protein requires some insulin to metabolise is an important consideration for people who need to inject extra to keep their blood glucose levels stable.
A better understanding of the insulin response to various foods would be useful for diabetics calculating their insulin dose or even to help refine food choices to manage insulin load.
Since launching the optimising nutrition blog I have had many interesting discussions and learned a lot about protein and how it affects insulin and blood glucose.
The Most Ketogenic Diet Foods article which reviews the food insulin index data and what we can learn about our food choices has received almost 200,000 view. Given the level of interest, I thought it would be useful to review this topic in more detail.
the food insulin index… a quick refresher
If you’ve been reading Optimising Nutrition blog you would have come across discussion of the recent food insulin index testing undertaken at the University of Sydney as detailed in Kirstine Bell’s PhD thesis Clinical Application of the Food Insulin Indexto Diabetes Mellitus  (Sept 2014).
The primary learning from the recently expanded food insulin index data is that the carbohydrate content of a food only partially explains the insulin response. The protein, fibre and fructose also affect our insulin response to our food.
The cluster of data points on the left-hand side of the figure below shows that:
low carbohydrate, high fat foods trigger a negligible insulin response, while
low carbohydrate high protein foods cause a significant insulin response.
When we assume that fibre is indigestible and protein has about half the insulinogenic effect of carbohydrates we get a much better prediction of insulin response.
The insulin requirement of a particular food is described better by the following formula:
insulin load = total carbohydrates – fibre + 0.56 * protein
digestion time for protein versus carbohydrates
One of the limitations of the food insulin index data is that the insulin area under the curve was measured over only three hours. This is not a big deal for foods that are high in carbohydrates as they are generally fully digested within three hours.
If we were to repeat the food insulin index testing over a longer period it is likely that the measured insulin response would be significantly greater and even more-so in people with diabetes. That is, the insulin response to protein may be greater than the 56% of the insulin response to carbohydrate indicated by the analysis of the food insulin index data if we were to measure the insulin response over a longer period.
Wilder’s ketogenic formula
Dr Russell Wilder of the Mayo Clinic was the first to coin the term ‘ketogenic diet’.  Wilder developed the diet as an alternative to fasting in the treatment of epilepsy in the 1920s.
Wilder also developed the formula shown below to determine whether a diet would be ketogenic. If the number from this calculation was greater than 1.5 (ideally greater than 2.0) then the diet would be considered to be ketogenic and appropriate for the treatment of epileptics. 
This formula is based on the understanding that:
100% of carbohydrate is glucogenic (i.e. converts to glucose),
54% of protein is glucogenic,
46% of protein is ketogenic, and
10% of fat is glucogenic.
I had previously searched for detail of how Wilder had arrived at the 56% / 46% split for protein and only found references suggesting that the 56% glucogenic potential of protein comes from the analysis of nitrogen in the urine of dogs.  However I recently came across this paper which details Wilder’s thinking in more detail.
Wilder’s conclusion that a diet needs to have more than two times the ketogenic precursors compared to glucogenic precursors is still the basis of the formulation of diets used to treat epilepsy.
According to George Cahill, Krebs also found that 57g of glucose may be derived from 100g of protein.  Again, this is similar to the insulin demand for protein observed in the food insulin index tests.
The most straight forward approach is to assume that protein has no impact on insulin or blood sugars.
Dr Richard Berstein and Dr Robert Atkins pioneered the concept of carbohydrate counting for weight loss and diabetes management in the 70s and 80s. There have been various waves of popularity of low carbohydrate diets with many people finding success.
Carbohydrate counting alone is a reasonable approach that is likely to work for most people, particularly if they are not highly insulin resistant.
However, there are some people that reducing carbohydrates alone doesn’t work for. The fact that protein also generates insulin suggests that managing protein as well as carbohydrates may be necessary to manage insulin levels.
thermic effect of food
You may have heard of the concept of the thermic effect of food where different foods require different amounts of energy for the digestion process. For example, a mushroom, which has a very low calorie density and a lot of fibre and protein, may require more energy to digest than is obtained from the digestion of the mushroom.
The maximum and minimum thermic effect (also known as the specific dynamic action) for each macronutrient is shown below. 
Compared to carbohydrates and fat, protein only yields between 76% and 84% of the energy per calorie ingested because of losses in digestion. This is useful to know if you’re trying to minimise calorie intake.
As discussed in the Why We Get Fat V2 article, part of this thermic effect of food is also likely to be due to the fact that there is a significant loss of energy when we convert protein to glucose to be used as energy. The body doesn’t like to do this other than in an emergency.
Steve Phinney’s “well formulated ketogenic diet”
One of the key observations from Steve Phinney’s well formulated ketogenic diet (WKFD) chart is that we need to strike a balance between carbohydrates and protein in order to maximise the ketogenic potential of our diet.
You can have 30% protein and 5% carbs or 20% carbs and 10% protein and still be within the bounds of a ketogenic diet. However if you have 30% protein and 20% carbs you will be outside the realms of a ketogenic diet because you will be producing too much glucose.
According to Nuttall and Gannon  the body requires between 32 and 46g per day of high quality dietary protein to maintain protein balance. This equates to around 6 to 7% of calories in a 2000 to 2500 calorie diet being taken ‘off the top’ for growth and maintenance, with everything else potentially available as ‘excess’ protein for gluconeogenesis. This should not be considered optimal, but simply a minimum reference point for the absolute minimum amount of protein.
Interestingly, the slope of the line along the face of Phinney’s WFKD triangle corresponds with the assumption that 7% of protein goes to muscle growth and repair (protein synthesis) with 75% of the remaining ‘excess’ protein being glucogenic. This 75% value is in the “ball park” (although a little higher) of our previous estimate of the glucogenic potential of protein based on the analysis of the food insulin index data.
amino acid potential
We also have an understanding of which amino acids are glucogenic, which are ketogenic and which are a bit of both.  The table below shows the various amino acids divided up on the basis of their ketogenic versus glucogenic potential and also which are essential versus non-essential. 
Only two amino acids are exclusively ketogenic. There is a handful that are both glucogenic and ketogenic. However most of the amino acids are glycogenic, meaning that they will most likely turn into glucose if not required for protein synthesis.
According to David Bender“In fasting and on a low carbohydrate diet as much of the amino acid carbon as possible will be used for gluconeogenesis, an ATP-expensive, and hence thermogenic process.”
Hence it appears likely that in a low carbohydrate diet situation excess amino acids that fit under the “both” classification will be turned to glucose rather than ketones because the body needs the extra glucose which it is not getting from ingested carbohydrates.
Conversely, if someone is consuming a high carbohydrate diet the excess amino acids that fit into the “both” category will be converted to ketones rather than glucose because the body is getting more than enough glucose from the diet.
So, to some extent, protein is versatile depending on the body’s need. But at the same time, it is only a small portion of the amino acids that are able to do this. The fate of the majority of the amino acids is pre-destined.
the krebs cycle
The figure below shows the process of catabolism of amino acids. 
I am not an organic chemist, but from what I understand this means that:
The amino acids Leucine and Lysine cannot be converted back to glucose as they are ketogenic;
Isoleucine, Tyrosine, Phenylalanine, Tryptophan, Threonine all enter into the amino acid catabolism cycle and can be used for various functions, such as muscle repair and growth, but can also be converted back into glucose if required (glucogenic) or turned into fatty acids (ketogenic); and
The remaining amino acids enter the cycle and can be used for a variety of functions in the body, but cannot be converted into fatty acids. If they are not required they can be turned into glucose and potentially stored as body fat.
The majority of the amino acids obtained from the digestion of protein have the potential to be turned into glucose through gluconeogenesis.
The reason that we don’t see a sharp rise in blood glucose is partly because amino acids from digestion circulate in the blood until they are required. Gluconeogenesis is a demand driven process. Glucose is pulled from amino acids when there is no other source rather than pushed into the bloodstream due to ingestion of excess protein.
By contrast, glucose from carbohydrates will be used to refill glycogen stores (liver and muscle) and then find their way quickly into the bloodstream. In most people, the amino acid stores in the blood are not saturated and hence there is plenty of capacity to store amino acids until they are required, at least if you have good insulin sensitivity and are not diabetic.
The body does need glucose, and it is fine to get it from carbohydrates or protein via gluconeogenesis. However many people struggle to produce enough insulin and / or are insulin resistant and hence struggle to keep their blood sugars in normal range. For these people it makes sense to reduce the insulin load their diet (the portion that requires insulin) to a point that they can maintain normal blood glucose levels.
tallying up the amino acids
I figured I could use this knowledge of the categorisations of the various amino acids to better understand how much of the proteins in the 8000 foods listed in the USDA food database are glucogenic versus ketogenic.
For each food in the USDA database I tallied up the weight of the glucogenic and ketogenic amino acids and the amino acids that fell onto the ‘both’ category and found that:
ketogenic amino acids make up only 12% by weight of the total protein across the 8000 foods in the database,
glucogenic amino acids comprise 74% of the foods, and
amino acids that fit in the “both” comprise 14% of the total weight of amino acids.
This means that somewhere between 78% and 89.5% of protein has the potential to turn into glucose, depending on whether you considered the amino acids in the ‘both’ column to be glucogenic or ketogenic, or somewhere in between.
For someone eating a low carbohydrate diet nearly 90% of ‘excess’ protein could be turned to glucose in the blood stream.
Why is this different to the observation from the food insulin index testing that approximately 56% of protein raises insulin? Perhaps the following factors come into play:
When we consider the glucogenic potential of the individual amino acids we are considering the maximum potential of protein if it is not first used for protein synthesis. The amount of protein synthesis will be greater for say an athlete or a body builder, with less protein remaining for gluconeogenesis.
Converting protein to glucose requires energy and hence some of the energy from ingested protein is lost in the process and hence is not converted to glucose.
The insulin index testing is undertaken over only three hours. Protein takes much longer to digest and be metabolised into glucose hence the insulin index testing may underestimate the full glucogenic potential of protein.
which foods have the most ketogenic protein?
So I bet you are wondering which forms of protein have the highest amount of ketogenic protein. Maybe not? Well, I was, and I am going to share it with you.
The table below shows the foods from the USDA database that have the most ketogenic protein (assuming the ‘both’ amino acids are split 50/50 glucogenic / ketogenic) in terms of grams of ketogenic amino acids per 100 grams of the food.
ketogenic aminos ( per 100g)
% ketogenic protein
Seal, Bearded Alaskan
Chicken, breast with skin
It is hard to know what to make of this list other than noting that the seal, whale and cod have the highest amounts of ketogenic protein. Perhaps there is something about cold water animals that cause them to store more ketogenic amino acids? This seems to align with what we see in the traditional diets of humans who may eat more fat if they are living further away from the equator but eat more carbohydrates from fruits if they live closer to the equator.
Although seal, whale and cod have high amounts of ketogenic amino acids, overall they are still quite insulinogenic. In view of the high proportion of insulinogenic properties of some meats it is not surprising that people can thrive on a 100% meat zero carb diet because the body can get as much glucose they need from the meat. At the same time though, I’m not sure that an all meat diet can provide an optimal array of vitamins and minerals unless you are emphasising organ meats.
In view of the fact that a large amount of protein can be converted to glucose through gluconeogenesis, it seems better to focus on foods that have a lower percentage of insulinogenic calories if you are insulin resistant or do not have a fully functioning pancreas.
Rather than worrying about whether you’re eating too much protein, most people will do fine if they limit their processed grains and sugars and eating as much protein as their appetite directs them to. If you are aiming for a therapeutic ketogenic diet to manage chronic conditions such as cancer, epilepsy or dementia, then you may want to consider moderating your protein intake to drive ketosis.
The table below shows a comparison of a range of glucogenic factors for protein relative to carbohydrate, summarising the discussion above. Most of the approaches to understanding the insulinogenic portion of protein give an even higher value than suggested by the analysis of the food insulin index data.
A lower end sensitivity assuming that no protein is converted to glucose (i.e. as per standard carbohydrate counting).
Food insulin index
Based on testing of > 100 foods in healthy individuals
Thermic effect of food
Average of additional in digestion losses minus 7%.
Used in initial ketogenic formula
Krebs / Janney
Based on nitrogen excretion in dogs
Glucogenic potential (min)
Based on summing amino acids in USDA foods database, excluding “both” aminos.
Glucogenic potential (max)
Based on summing amino acids in USDA foods database, including “both” aminos.
Steve Phinney WFKD
Assuming that the first 7% of calories goes to growth and repair with 75% of the remaining amino acids being glucogenic.
the most ketogenic foods… updated
I have calculated the insulinogenic potential of the foods shown in this previous article (The Most Ketogenic Diet Foods) using the following approaches:
food insulin index data (i.e. protein is 56% insulinogenic);
thermic effect (i.e. protein is 77% insulinogenic); and
maximum glucogenic potential of the amino acids for each food (varies for each food based on data in USDA foods database).
This updated data illustrates the difference in standard carbohydrate counting and the full insulinogenic potential of the food. While there is a range of values due to the varying amounts and types of protein overall, there is a reasonable alignment between the food insulin index (56%), thermic effect of food (77%) and maximum glucogenic potential values, particularly when we compare it to the carbohydrate only approach for the lowest carbohydrate foods.
When my wife Monica was diagnosed with type 1 diabetes at ten she was advised to eat at least 130g of carbohydrates with every meal.
The insulin dose was kept fixed to cover this fixed amount of carbohydrates. If she went low she had to eat more carbs to bring her blood glucose back up.
Welcome to the everyday blood sugar roller coaster that takes over your life when you have diabetes!
It wasn’t till after we were married in 2002 and started thinking about having kids that she found a doctor with an interest in diabetes who told her that she could tailor her insulin dose to what she wanted to eat.
Up until this time even the visits to the endocrinologist were to get more scripts for insulin and thyroid medication. No useful advice was provided about how to manage diabetes.
It’s amazing that the concept of carbohydrate counting was new and shiny in 2003 when Richard Bernstein developed the concepts back in 1970s!
Today, the standard of care for diabetes seems to have incorporated Bernstein’s carbohydrate counting, however the nutritionists and diabetes associations still advises that diabetics should not have to deprive themselves of any food in the pursuit of health. And like everyone else, they should eat a diet full of “healthy whole grains”.
It wasn’t until we discovered Paleo and then low carb through family members and social media that she found that she could improve blood sugar control through diet.
More recently by refining our diet to prioritise low insulin load, high fibre and high nutrient density foods I’m pleased to say that she has been able to find another level of improved blood sugar control, increased energy and reduced depression and anxiety that so often comes with blood sugar dis-regulation.
It is still not easy and we are still learning, however she is now able to enjoy working as a supply teacher rather than just getting through the morning and needing to sleep during the afternoon before picking up the kids from school.
Her big regret is that she did not discover this earlier, which would have saved her from spending decades living in a fog with limited energy.
The chart below shows the difference diet can make in the management of blood glucose, particularly for a type 1 diabetic (notice that these plots are only two months apart!). People who find success with this dietary approach find a substantial improvement in quality of life and their state of well being that makes it well worth the effort.
Below is a recent post on the TYPEONEGRIT Facebook group from a mother of a type 1 diabetic child describing their interaction with her health care team. It’s still not simple to go against the main stream dietary advice.
We had our team meeting today to discuss LCHF… they are so terrified of this, even though we have great BG readings, behaviour improvements and learning improvements (noted by us, family, friends and his school) which they didn’t even acknowledge.
The nutritionist is concerned that he won’t be getting the micronutrients that only come from grains and the higher carb vegetables (grains are fortified), then her concern was the B vitamins 1, 3 and 6.
Then the concern about Iron (what?! have you seen the meat and spinach listed?). Then it was calcium and magnesium (clearly they don’t have a clue about LCHF).
They said they are afraid this diet may cause future developmental harm. We said your diet WILL cause future harm and way more than developmental. Back and forth and on it went. We addressed their concerns with peer reviewed research, and respect to their limited knowledge.
We will be an open book and comply because I want them to learn that T1D care can be so much better than it has been up to now, and pave the way for the next families that wishes to do LCHF.
They will check for vitamins and minerals at his 3 month blood work (again special for our case, which we have to pay for).
The good news here is that after running an intense battery of tests they decided to use this child as Canada’s first case study in LCHF paediatrics for the management of type 1 diabetes.
This post inspired me to run some numbers on a range of diets to see whether there was any issue with the nutritional content of higher fat diets. It turns out that diets with higher levels of fat can be very nutritious while the grain based diet that everyone is recommended does very poorly, particularly when you take the insulin of these higher carbohydrate diets into account (see the Diet Wars… Which One is Optimal article for more details).
It breaks my heart to see diabetics living with a highly diminished quality when there is the potential to greatly reduce the impact of diabetes by more informed food choices.
For people with diabetes and their carers diet is important and maybe a matter of life and death, or at least a decision that will greatly affect their quality and length of life.
Organ meats top the list when it comes to nutrient density, however I haven’t found a lot of recipes using organ meats that rank really well. Perhaps it’s because people feel they need to put a lot of not so healthy things with their organ meats to drown out the taste?
This recipe uses heart, ground beef along with a range of spices to build a solid nutritional profile.
If you look closely at the chart below you’ll see that I’ve used pork heart as there is no data for beef heart in the NutritionSELF database.
Robb has had a massive influence on my thinking, along with many others. I like the way he brings everything back to evolutionary principles that need to make sense in the broader context rather than just looking at isolated studies. I also like the way he promotes quantifying nutritional density as a way to beat the nutritionists at their own game.
I’m hoping that nutritional density combined with the insulin index component will take things one step further to provide a quantitative basis to demonstrate that one meal is better than another.
The details for a 500 calorie serving (the recipe is for two servings) are shown below. This recipe does really well on the weight loss ranking because it has a low calorie density and a solid, great nutrient levels and a solid amount of fibre.
If you are insulin resistant or are sensitive to tomatoes you may want to reduce the tomatoes. I’ve re-run the numbers with 5 ounces of tomatoes rather than the 26 ounces in the original recipe. Still not perfectly diabetic friendly, but an improvement and slightly closer to the Bernstein target of 12g carbohydrates per meal.