What affects your blood sugar and insulin (other than carbs)?
Does X, Y or Z raise my blood sugar?
This is a common question we get when people start with Data-Driven Fasting.
It’s also an important question for people (like my wife Monica) who are managing diabetes to empower them to make well-informed food choices that will keep their blood sugars stable.
While carbohydrates have the most significant impact on blood sugar and insulin levels, there are several other quantifiable parameters of the food we eat that have a considerable impact.
Five years ago, I shared some insights from my analysis of the Food Insulin Index data (see Making Sense of the Food Insulin Index).
After getting lots of questions from people in the Data-Driven Fasting Facebook Group, I thought it would be useful to revisit the data to understand how the food we eat affects our blood sugar response.
To summarise, the table below summarises the results of the analysis, showing how fibre, protein, fructose and fat influence:
- maximum rise in blood sugars (within two hours),
- glucose rise (area under the curve over two hours), and
- insulin (area under the curve over two hours).
A negative value means that the parameter will decrease our blood glucose/insulin response (and vice versa). A positive value means that it will incrase your blood sugar/insulin response
Before we get too deep into the analysis, let me quickly give you some background on the data.
The original Food Insulin Index testing was done in 1997 at the University of Sydney and documented in An insulin index of foods: the insulin demand generated by 1000-kJ portions of common foods. While only 27 foods were tested initially, we now have data for 120 foods as documented in Clinical Application of the Food Insulin Index to Diabetes Mellitus (Bell, 2014).
The primary goal of the testing was to measure the change in insulin response above baseline in healthy non-diabetic subjects (i.e. uni student volunteers). The chart below shows an example of the insulin response ‘curve’ for a handful of foods.
Before trying to draw any conclusions from this data, it’s important to note a couple of things.
- The testing measured the change against baseline for insulin and glucose. This does not account for basal (or background) insulin which makes up the majority of your insulin demand across the day (i.e. about 50% for people on a high carb diet and up to 80% for someone on a low carb or keto diet).
- Insulin response was only measured over two hours. While simple carbs will be digested quickly, foods and meals that are a combination of carbs and fat will keep glucose and insulin elevated for much longer. So, we need to keep in mind that this testing only represents the initial response to foods.
While this data is useful for people trying to stabilise their blood sugars and insulin levels over the short term, it has limitations when we are considering the long term insulin and blood sugar response across the whole day. The data doesn’t account for the full effect of protein and fat, which take longer than two hours digest and be metabolised.
To illustrate this, the chart below (from this study) shows how blood sugars respond differently to a high carb low fat (green line = Plant Based Low Fat) vs low carb high fat (red line = Animal-Based Low Carb) diet. Blood glucose rises quickly in response to low-fat foods while blood sugars increase rise with low carb high-fat foods. However, for both extremes, after two hours, blood sugars are still elevated above baseline.
We don’t know much about the long term insulin and blood sugar response to the food we eat. However, it’s safe to assume that all calories have some impact on blood sugars and insulin because they will prevent your body from needing to draw down on your internal energy stores.
Just because a food has as minimal blood sugar or insulin response over two hours does not mean that it is a “free food”. Your pancreas will always produce just enough insulin to keep your stored energy locked away while it processes the energy from the food you are eating.
People who are metabolically healthy and lean tend to have more stable blood sugars. However, as shown in the chart below, from our analysis of data from the people who have used Data-Driven Fasting, a lower rise in blood glucose does not necessarily mean that you will be leaner. If you already have healthy blood sugars in the non-diabetic range, manipulating your diet to achieve even more stable blood sugars will not necessarily lead to weight loss.
However, managing your pre-meal blood sugars by delaying your next meal until they have returned to below baseline will help you to reduce your waking blood sugar, weight and improve your metabolic health (see Want to lose fat? DON’T aim for stable blood sugars! (Why your CGM could be making you fat for more detail).
With that disclaimer out of the way, let’s get into the analysis. The thing most people seem to be interested in is the rise in glucose after eating, so we’ll tackle that first.
Blood glucose rise after meals
The GI is useful to help us understand which carbohydrate-containing foods raise blood glucose the most. However, the problem with the GI is that it is based on foods with 50 g of carbs (with no consideration of total calories). Therefore, there is no way to compare the blood sugar impact on a calorie for calorie basis.
The good news is, as part of the Food Insulin Index testing, they also tested the maximum glucose rise after meals (i.e. for both low and high carb meals). They called this the Glucose Index. Compared to the Glycaemic Index, the Glucose Index is much more useful to help us choose foods that will achieve stable blood sugars.
The chart below shows calories from carbohydrate versus the glucose index (i.e. the rise in post-meal blood glucose). All foods tested were in 1000 kJ (or 239 calorie) servings. The x-axis of the chart is the calories from carbohydrates in each food.
In the top right, we have pure glucose (which has been given a score of 100% to designate the maximum glucose rise). In the bottom left we have high-fat foods such as butter, bacon and cream cheese (which have the lowest increase in glucose after meals).
The number we will be looking more at in the remainder of this article is the correlation coefficient (i.e. R2). The correlation of carbohydrate with maximum rise in blood glucose is 0.628.
While carbohydrates explain the majority of the rise in blood sugars, there are other factors. To better understand the other factors, I have analysed the data to see how other parameters to enable us to better understand the things that raise blood sugars after we eat.
The chart below shows how the correlation changes when we assume that varying proportions of the calories from fat contribute to the rise in post-meal glucose. Somewhat surprisingly, we see that we get a better correlation when we assume 20% of the calories from fat contribute to the maximum rise in glucose.
Mechanistically, this makes sense. If we had glucose alone, it may be absorbed into the liver and muscle quickly and not show up in the bloodstream. However, if we are also filling the fat stores at the same time, glucose seems to rise a little more because both the fat and glucose fuel tanks in our body are being filled.
When we sub this back into the 120 data points (as shown in the chart below), we see that the correlation increases. Considering the effect of fat on blood sugars helps us to better predict the glucose response to foods that are a combination of fat and carbs together in the middle of the chart.
When we look at fibre, we see that we get the best correlation when we assume that 15% of fibre does not raise blood glucose. Not all of the fibre will be digestible to form glucose. Fibre will also slow digestion and hence blunt the maximum rise during the two-hour window.
This next chart shows the glucose rise forecast when we assume that 15% of fibre doesn’t cause a blood sugar rise, with a slight improvement in the correlation coefficient.
Protein is interesting. The highest correlation was achieved when we assume that there is no influence from protein on blood sugars. Both positive and negative factors decreased the correlation. The data indicate that protein doesn’t affect our rise in blood glucose after meals.
When we eat protein, glucagon balances the effects of insulin.
- Most people see a negligible blood sugar response to protein. Many people who are metabolically healthy see their blood sugars drop after a higher protein, low carb meal.
- At the other end of the spectrum, people who have diabetes or significant insulin resistance often see some rise in blood glucose because their glucagon response is not being matched by their insulin.
However, a short term increase in blood sugars should not be seen as a reason to avoid protein. Protein tends to increase satiety and help reduce your overall energy intake across the day. The fact that people who are insulin resistant see more of their protein being converted to glucose means that they actually need to focus on a higher percentage of protein to make up for the protein being lost to glucose. As per the Protein Leverage Hypothesis, their appetite will go in search of more food to make up for the lost protein and to ensure they get the protein they require.
If you’re interested in learning more about this, can check out:
- Why do my blood sugars rise after a high protein meal?
- High protein diets for weight loss: How much you need and why it works
Fructose (i.e. the sugar predominantly found in fruit) is interesting. While fructose is technically a carbohydrate, it can be metabolised into either glucose (if required) or fat (if glucose stores are full).
Anecdotally, some people using Data-Driven Fasting have noted that fruit doesn’t seem to raise their blood sugar as much as they might expect. There are even groups of people with Type 1 Diabetes who seem to do reasonably well on a diet that contains a lot of fruit (although not nearly as well as people using a low carb adequate protein approach).
As shown in the chart below, we get the most significant improvement in correlation with glucose rise when we assume that nearly all (i.e. 90%) of the fructose in food does not contribute to our rise in glucose.
When we apply this to our data, we get a much better prediction of the rise in glucose when we account for fructose.
While fruit (which has more carbohydrate from fructose) may be a better choice than refined grains (which provide carbohydrate in the form of starch), this does not necessarily mean that fruit will not cause a blood sugar rise. You can test and see how it affects you, and avoid it if it causes a rise of more than 1.6 mmol/L or 30 mg/dL after eating.
In addition to fructose, fruit contains a significant amount of glucose and sucrose, which will also raise your blood sugar. Excessive fructose consumption has also been shown to drive Non-Alcoholic Fatty Liver Disease (i.e. a build-up of fat in the liver). So the fact that fructose can be converted to fat rather than glucose is not necessarily beneficial in all cases.
Combining carbohydrate, fructose, fibre and fat
Finally, we can combine all these insights to give us a better prediction of the glucose rise to foods.
As shown in the chart below, when we account for the calories from carbohydrates, minus 90% of fructose, minus 20% of fibre, plus 15% of fat we get a better overall correlation (i.e. the R2 increase from 0.628 to 0.7586).
What does this all mean?
The first point to make here is that, if your blood sugars are in the healthy, non-diabetic range, then trying to stabilise them further by avoiding carbohydrates will not necessarily lead to better metabolic health.
People with good metabolic health tend to have more stable blood sugars, but manipulating your diet to have flatline blood sugars will not necessarily lead to weight loss or better health outcomes. If that is your goal, you should optimise your diet for greater nutrient density and satiety, and more stable blood sugars will follow.
However, this analysis indicates that:
- Foods with more fibre will help to stabilise your blood sugars,
- Naturally occurring fructose in fruit and veggies are unlikely to raise your blood sugars as much as starch in grain-based foods. However, foods that are sweetened with high fructose corn syrup are still nutrient-poor empty calories and should ideally be avoided, particularly as the fructose may be metabolised directly to fat and stored on your body.
- While high-fat foods don’t appear to raise blood sugar much in the short term, foods that are a combination of fat and carbs together will likely increase your blood glucose more than you would expect from the carbs alone.
Which foods will raise my blood glucose the least?
The table below shows the calculated glucose rise with nutrient density score for popular foods. To minimise your rise in blood glucose after you eat, you should avoid foods towards the top of this list with a high glucose rise. But if you also want to lose fat, improve satiety and get enough nutrients without excessive calories, you should avoid foods with a low nutrient density score and prioritise those with a higher nutrient density.
|name||Glucose Rise||Nutrient Density|
|milk (low fat)||50%||60%|
|yogurt (whole milk)||37%||50%|
|half and half||26%||28%|
|cottage cheese (low fat)||19%||23%|
|Mozzarella cheese (whole milk)||14%||45%|
|ground beef (80% lean)||11%||46%|
|sirloin steak (fat eaten)||8%||51%|
|ground beef (85% lean)||8%||49%|
|sirloin steak (fat not eaten)||4%||59%|
|chicken breast (no skin)||4%||52%|
The chart below shows nutrient density vs glucose rise. If you want to minimise your rise in blood glucose, you should choose foods lower on this chart while also trying to stay towards the right to ensure you get the nutrients you require. For more detail, click here to view an interactive version of this chart.
If you would rather think in terms of recipes rather, rather than foods, you can check out our series of 22 nutrient-dense recipe books. If your blood sugars are elevated after meals (i.e. a rise of greater than 1.6 mmol/L or 30 mg/dL) then you may want to try the Low Carb and Blood Sugar or Nutritional Keto recipe books. If your blood sugars are already stable but you have fat to lose you could use the Maximum Nutrient Density or Fat Loss books. If you have elevated blood sugars and fat to lose (as many people do) then the Blood Sugar and Fat Loss book will help.
Glucose score (area under the curve)
As part of the Food Insulin Index testing, they also measured the Glucose Score, which is the area under the curve glucose response to foods over two hours.
Foods with a higher Glucose Score will keep your blood sugar elevated higher for longer. This is potentially more important than the maximum rise in blood sugar.
The chart below shows the calories from carbohydrate versus the Glucose Score. Low carb high-fat foods in the bottom left have a low glucose score, while cornflakes and pure glucose have a glucose score of 100% in the top right. We have an R2 correlation of 0.6468 when we only consider the effect of carbohydrate.
I’ll spare you the detail of the correlation charts this time around, but the analysis found that the best correlation was achieved with:
- minus 70% fructose,
- minus 20% fibre, and
- plus 10% fat.
Considering the negative impact of fructose and fibre and the positive impact of fat give us a better prediction of the blood sugar area under the curve over two hours (i.e. the R2 correlation coefficient increases from 0.6468 to 0.7122).
In practical terms, this means that:
- foods with naturally occurring fructose (i.e. fruit and vegetables) won’t keep your blood sugars elevated as much as other carbohydrates (e.g. grains), and
- foods with more fibre (which are partially indigestible) will also help to blunt the blood sugar response, and
- fat has some influence on keeping blood sugars elevated over the short term and likely has a more significant effect over the longer term too.
Which foods will keep your blood glucose elevated the most vs least?
The table below shows the calculated glucose score with the nutrient density score. To keep your blood sugars stable over two hours, you should avoid foods towards the top of this list. But if you also want to lose fat, improve satiety and get enough nutrients without excessive calories, you should prioritise foods with a higher nutrient density score.
|name||Glucose Score||Nutrient Density|
|milk (low fat)||47%||60%|
|yogurt (whole milk)||30%||50%|
|cottage cheese (low fat)||17%||23%|
|half and half||15%||28%|
|Mozzarella cheese (whole milk)||3%||45%|
|sirloin steak (fat not eaten)||0%||59%|
|chicken breast (no skin)||0%||52%|
|sirloin steak (fat eaten)||0%||51%|
|ground beef (85% lean)||0%||49%|
|ground beef (80% lean)||0%||46%|
This chart shows nutrient density vs glucose score. If you want to stabilise your blood glucose, you should choose foods from the bottom of this chart while staying towards the right to ensure you get the nutrients you need. For more detail, click here to view an interactive version of this chart.
Insulin response to food
Finally, I wanted to revisit my previous analysis to understand how different foods elicit a different insulin response.
This first chart shows carbohydrate calories versus the food insulin index (i.e. area under the curve insulin response over two hours). We see an overall correlation of R2 = 0.369.
The analysis found that the best correlation was achieved with:
- plus 55% protein,
- minus 70% fructose,
- minus 45% fibre, and
- plus 10% fat.
By considering these other factors, we get a significant improvement in our ability to predict the insulin response to food.
Which foods will keep my insulin stable?
The table below shows the calculated food insulin index score with the nutrient density score. To stabilise your insulin levels over the first two hours after you eat, you should avoid foods towards the top of this list. But if you also want to lose fat, improve satiety and get enough nutrients without excessive calories, you should prioritise those with a higher nutrient density score. While very high-fat foods will keep your insulin levels stable, they will not provide a significant amount of nutrients and tend to have a low satiety value, meaning that you will tend to eat more and your insulin levels will be higher across the whole day.
|name||Food Insulin Index||Nutrient Density|
|milk (low fat)||67%||60%|
|cottage cheese (low fat)||56%||23%|
|yogurt (whole milk)||47%||50%|
|chicken breast (no skin)||44%||52%|
|sirloin steak (fat not eaten)||44%||59%|
|ground beef (85% lean)||30%||49%|
|sirloin steak (fat eaten)||30%||51%|
|half and half||28%||28%|
|Mozzarella cheese (whole milk)||25%||45%|
|ground beef (80% lean)||22%||46%|
This chart shows nutrient density versus food insulin index response. If you want to stabilise your insulin response to the food you eat, you should avoid foods towards the top of this chart while also staying towards the right to ensure you get the nutrients you need. For more detail, click here to view an interactive version of this chart.
Practical implications of this analysis
This more in-depth understanding of the insulin response to food is particularly useful for people who need to calculate how much insulin they need to inject. Being able to calculate short term insulin demand and matching the food they are eating with their medication requirements is critical.
For the rest of us, this information is interesting, but we need to keep in mind that this data only represents the insulin and glucose response over two hours. Most foods will require insulin and keep blood sugars significantly elevated over much longer than two hours.
If you are testing your blood sugars and find that they rise significantly after meals (i.e. by more than 1.6 mmol/L or 30 mg/dL) then choosing food that causes a lower blood sugar and insulin response will stabilise your blood sugars. However, if your goal is fat loss, flatline blood sugars should not be your primary goal, especially if it requires you to consume more low satiety, higher fat, nutrient-poor foods.
While large swings in blood sugars are not ideal and can increase hunger and appetite, blood sugar swings in the healthy non-diabetic range are not cause for concern.
If your blood sugars are elevated after meals (i.e. a rise of greater than 1.6 mmol/L or 30 mg/dL) then you may want to try the Low Carb and Blood Sugar or Nutritional Keto recipe books which will stablise your blood sugar and insulin levels.
If your blood sugars are already stable but you have fat to lose you could use the Maximum Nutrient Density or Fat Loss books. If you have elevated blood sugars and fat to lose (as many people do) then the Blood Sugar and Fat Loss book will help.
Then, managing your pre-meal blood sugars by waiting a little longer until they return to below baseline (e.g. using Data-Driven Fasting) will also be useful if your goal is fat loss or reversal of Type 2 Diabetes.