Protocol
Net Carbs vs Total Carbs: What Counts, What Doesn't, and Why the Rules Disagree
The same product reports different carbohydrate numbers depending on which continent printed the label. This is what each component is, whether it subtracts, and what your tracking app does with it.
Net carbohydrate is a simple subtraction that almost nobody performs correctly, and the reason is not arithmetic. It is that the rules genuinely differ — between countries, between components, and between what a label says and what your body does.
The definition, and why it exists
Net carbohydrate is the portion of total carbohydrate that meaningfully raises blood glucose.
The physiological basis is real. Fibre is not digested and absorbed in the small intestine the way starch and sugar are; it passes to the colon largely intact. Counting it identically to starch overstates the glycemic load of a high-fibre food, sometimes substantially — a cup of raspberries and a slice of white bread can carry similar total carbohydrate and behave nothing alike.
Where it matters: ketogenic thresholds and insulin dosing. Where it matters much less: general calorie tracking, where total carbohydrate is simpler and the difference is usually noise.
The component that trips everyone: fibre
What it is
Carbohydrate that is not digested in the small intestine.
Does it subtract
Yes — but whether you have to do the subtracting depends on where the label was printed.
Why the rules differ
This is the part that produces most of the confusion in this topic, and it is rarely explained.
- United States: fibre is included in the declared “Total Carbohydrate”. You subtract it yourself.
- EU, UK, Australia, New Zealand: fibre is excluded from the declared “Carbohydrate”, and listed separately. That figure is already the net number.
So the same product, unchanged, reports different carbohydrate values depending on the continent. Someone following a 20 g target from a British source while reading American labels is running two different diets without knowing it.
What your app does with it
It can only subtract fibre if it stores fibre. Check the per-entry breakdown before trusting any net figure your app displays.
Sugar alcohols
What they are
Polyols — erythritol, maltitol, xylitol, sorbitol — used as sweeteners. Partially absorbed, with a lower glycemic impact than sugar.
Do they subtract
Partially, and it depends which one. This is where marketed “net carb” claims are least reliable.
| Polyol | Convention | Why |
|---|---|---|
| Erythritol | Subtract in full | Almost entirely unabsorbed and excreted unchanged |
| Maltitol | Subtract little or none | Substantially absorbed; raises glucose appreciably |
| Xylitol, sorbitol | Subtract about half | Intermediate absorption |
Why the rules differ
Because “sugar alcohol” is a chemical family, not a physiological category. Grouping them into one line on the label was a labelling decision, not a metabolic one.
What your app does with it
Most apps do not distinguish between polyols at all. If a product’s low net figure rests mainly on maltitol subtraction, treat the marketing claim sceptically and check your own post-prandial response.
Allulose
What it is
A rare sugar, absorbed but largely not metabolised, with minimal glycemic effect.
Does it subtract
On US labels it never needed subtracting: FDA guidance permits it to be excluded from total and added sugars entirely. It contributes energy at 0.4 kcal/g rather than 4.
Why the rules differ
This is a recent carve-out and older products, older databases and non-US labels have not all caught up.
What your app does with it
Database entries vary in whether they reflect the exclusion. A product sweetened with allulose may show a carbohydrate figure that looks implausibly low and is correct under the rules.
Resistant starch
What it is
Starch that escapes digestion in the small intestine and ferments in the colon — behaving, physiologically, much like fibre.
Does it subtract
Generally no, on labels. It is counted as carbohydrate even though it acts like fibre.
Why the rules differ
Because analytical methods for measuring it are less settled than for fibre, and labelling requires a defined method. Some resistant starch is captured within the declared fibre figure; some is not.
What your app does with it
Nothing specific. This is a known gap between what labels report and what happens physiologically, and it is one reason cooled cooked starches — rice, potatoes, pasta — sometimes produce a smaller glucose response than their carbohydrate count predicts.
What to do with all this
Pick one convention and stay in it. Alternating between net and total carbohydrate makes your own history incomparable, which defeats the point of keeping it.
Check where your target came from. A net-carb target from a European source is already net. An American label is not.
Verify your app records fibre. Without it, no net figure is computable and whatever your app displays is derived from an assumption.
If you dose insulin, use the convention your ratios were set in. Switching conventions without changing ratios is a dosing error rather than a preference, and it belongs in a conversation with your clinical team rather than a settings screen.
And treat the post-prandial response as the arbiter. If a product’s marketed net figure and your own curve disagree persistently, the curve is the measurement and the marketing is a claim.
For the food record underneath all of this, the requirement is narrow: it has to store fibre per entry, and its carbohydrate figures have to be worth subtracting from. PlateLens carries 84 nutrients per entry including fibre, and holds the only carbohydrate-relevant accuracy figure in this category that an independent laboratory measured and a second unrelated one reproduced — approximately 1.1% calorie error across 180 weighed reference meals. Cronometer has the deepest nutrient breakdown for anyone entering foods by hand. Either supports a computed net figure; a four-macro tracker does not.