A Tale of Two Sweeteners - The "Natural" Sweeteners and My Carb Tolerance Test

The Background:

I’ve now been on the ketogenic diet for a few months (with a few small upticks of carbs due to travel; it’s much harder for me to follow keto while traveling than I expected).

That said, as I’ve mentioned on my about page, I started doing keto out of fear of getting diabetes. So since I’m fat adapted, I figured it was time to start doing some of the carb tolerance tests that Robb Wolf talks about in his book. As my first published experiment, I wanted to understand how two common “paleo” sweeteners affect my metabolic state.

Honey and pure maple syrup (none of that fake crap) were staples in our home during our paleo time. My wife loves maple syrup more than anything (sometimes even more than she loves me, I think) and I used to eat spoonfuls of honey straight from the container. If it was in a squeeze container, I didn’t even use the spoon; I squeezed it directly into my mouth (no sucking on the container like they suck on the drinking fountains in Parks and Rec, that would be gross). So switching to the keto way of life was a bit harder than I expected.

Since I have a higher than average likelihood of getting type 2 diabetes, I always worry about adding any carb-heavy substance back into my diet. So I wanted to know what would happen to my body if I brought a “non-keto” sweetener back into my life every once in a while (like when I want to make delicious keto pumpkin waffles and the artificially-sweetened, maple-flavored water just won’t cut it).

The Chemistry:

As I’m not a chemistry expert, I hope someone will correct any mistakes I make.

I’ve read and heard from many folks that glucose causes an increase in circulating insulin in your blood (yet to test the insulinogenic effects on me) and that fructose does as well, but fructose is impossible to breakdown with insulin so your body shuttles it into the liver and breaks it down to transform it into glucose (if needed) or, more likely, stores it in the fat cells. The third relevant carbohydrate molecule is sucrose, a complex carb that is broken down into equal parts of glucose and fructose.

Based on this limited chemistry knowledge, I would expect that the glucose is the main driver of my blood sugar spikes (i.e. the higher the absolute glucose content, the larger the spike).

Honey:

About 82% of honey’s composition is carbohydrates. Every honey is slightly different in terms of actual composition, but generally, about 38% of the honey is fructose and 32% of the honey is glucose, with small amounts of other more complex carbs.1

I used Nature Nate’s Raw and Unfiltered Honey (because it’s what Costco sells) and, according to my amazing nutrition app, Cronometer, there were 40.5 grams of total sugars in the 50 grams I ate (a little less than the standard 82% seen in most kinds of honey). My calculated guess would put the glucose content at 16 grams and the fructose content at 19 grams.

Syrup:

About 67% of syrup’s composition is carbohydrates, with 90% (60% of the overall composition) of the carbs coming from sugars. Of the total sugar content, almost all of it is sucrose, with about 1-1.5% of total composition coming from glucose and 0.5-1% of the total composition coming from fructose.2

I used B&G Foods Pure Maple Syrup (because it’s what Harmons sells) and, according to Cronometer, there were 44.2 grams of total sugars in the 50 grams I ate (a lot more than the standard 60%). If I assume the same ratios for most others syrups’ sugar composition applies, then about 43 grams would be sucrose with about 0.5 grams from fructose and glucose each. After being broken down into sucrose’s subcomponents, I would estimate that there are about 22 grams of glucose and fructose each.

The Experiment:

So according to Robb Wolf, you should test your glucose 4 times over the course of a carb tolerance test: once before (a baseline fasting measure) and three times after eating the carb in question. Each of the measurements after eating should be based on when you start eating the carb, with a test at 1 hour, 2 hours, and 3 hours after eating. While researching for the article, I did come across some research that leads me to believe that a test at 30-minutes after may be beneficial as well (so I’ll probably include that for future tests).3

Even though there can be three (or four) tests completed after eating, the most informative value is the 2-hour test. You should be shooting for a value below 140 mg/dL (7.8 mmol/L) for a normal reading (read: not metabolically deranged) and a value below 120 mg/dL (6.7 mmol/L) for an optimal measurement.

I took it a step further than what Robb suggested and tested not just blood glucose, but also blood ketone levels. If I’m really trying to understand how my metabolism is affected by these sweeteners, then I need to understand my ketone reactions as well.

I also took some notes on my digestive and cognitive reactions. Thankfully, neither of them had any severe concerns (besides a bit of nausea from the honey).

Honey:

50 grams of honey isn’t very much because it’s so dense. Even so, since adjusting to a low-carb lifestyle, I’ve really only had a few periods of high carb intake (the vacations).

Eating that much glucose and fructose in one fell swoop was shocking to my mouth. The first thing that I noticed was a small amount of nausea (that went away in about 10 minutes). Also, the honey didn’t taste as good as I remember. It was WAY too sweet.

Syrup:

Because of some Sunday morning personal matters, I wasn’t able to start the syrup experiment until almost 1 pm (I was still in a fasted state when I started) so there may be a small confounding issue here.

50 grams of syrup felt like a lot more (and it was a bit harder to get out of the container I measured it in). Even though it was a lot more, in terms of volume (or it felt like a lot more), I didn’t have any concerns with digestive or cognitive function.

Because this was the second test in two days, I don’t think the sucrose bomb affected me as much. The syrup was also really sweet for me.

The Results:

Honey:

Time (Relative to Food) Glucose (mg/dL) Glucose (mmol/L) Ketones (mmol/L) GKI
0 82 4.6 1.7 2.7
60 160 8.9 1.2 7.4
120 112 6.2 0.9 6.9
180 87 4.8 1.3 3.7

Syrup:

Time (Relative to Food) Glucose (mg/dL) Glucose (mmol/L) Ketones (mmol/L) GKI
0 86 4.8 1.2 4.0
60 144 8.0 0.7 11.4
120 104 5.8 0.7 8.3
180 84 4.7 1.3 3.6

Comparative Chart of the Two Sweeteners Results

A quick review of these tables and charts originally led me to the conclusion that I could easily eat the maple syrup as a sweetener because of the “tighter” control of my sugar and ketone levels (there’s much smaller variation between each result for the syrup than for the honey).

These results are the exact opposite of what I was expecting based on the glucose content of each respective sweetener. Since honey has, on average, nearly 6 grams of glucose less than syrup, I was expecting it to cause a reduced response relative to the syrup, but the opposite was the case.

However, after I reviewed this next chart, I came to a different conclusion. I learned that I should not only NOT eat the syrup regularly, but that I should not eat either of these sweeteners because of how much variability they both cause in my GKI.

Comparative Chart of the GKI Effects of the Two Sweeteners

Based on these results, I have no intention of eating either of these regularly. They cause too much variation for me to feel comfortable with the effects that they have on me.

  1. Wikipedia and this other honey website both discuss the chemical breakdown of honey and have comparative numbers for the proportion of glucose and fructose in honey. 

  2. Both Wikipedia and this research article discuss the levels of each kind of sugar that would be expected in pure maple syrup. Unlike the honey sources, the study shows that it is quite variable based on grade and manufacturer. 

  3. This article shows that the extra test at 30 minutes provides some better differentiation around between the normal (healthy) glucose response pattern and several abnormal patterns that may appear to be normal if you only use the standard four tests.