Carb counting means adding up the grams of carbohydrate you eat, because carbohydrate raises blood sugar more than protein or fat does. This guide covers Canadian and US labels, fibre, 15 g carb choices, portions, insulin-to-carb ratios and what the research shows.
To count carbs, you add up the grams of carbohydrate in everything you eat and drink at a meal or snack, then keep the total within a target range you have set with your diabetes care team 12. Carbohydrate gets this attention because it has the greatest effect on blood sugar 1. If you take insulin, the count can also help decide how much to take 12.
You can count in grams, or in “carb choices” of about 15 g each 6. Either way, the skills are the same: knowing which foods contain carbohydrate, reading a label and judging how much is on your plate. A dietitian can teach you to measure portions and read food labels 1.
This guide covers each of those skills, the two main ways a carb count is used, what studies show and what guidelines say about how much to eat. For how photo, label and barcode estimates compare, see carb counting with photos, labels and barcodes; for why the type of carbohydrate matters, see glycemic index and glycemic load.
Carbohydrate in food comes in three main kinds: sugars, starches and fibre 2, 13.
Diabetes Canada notes that meat and alternatives, most vegetables and fats contain little carbohydrate 3.
During digestion, sugars and starches break down into glucose, which is why carbohydrates have the greatest effect on blood sugar 1. Johns Hopkins Medicine puts it the same way: carb-rich foods affect blood sugar more than other foods do 12. Fibre is the exception, because it does not raise blood sugar 3.
Protein and fat still matter, mostly through timing. The ADA’s 2026 Standards of Care say fat and protein can affect both early and delayed glucose after a meal, with an effect that appears to depend on the amount eaten 8. A meal high in protein and fat can also change how quickly carbohydrate is absorbed 6. See why the same meal spikes differently for more.
Check the serving size first, then read the carbohydrate line in grams; on both Canadian and US labels, that number already includes fibre and sugars 2, 4.
On a Canadian nutrition facts table, the serving size is at the top, in a household measure followed by grams or millilitres. Health Canada says it is the amount typically eaten in one sitting, not a recommendation, so your portion may not match it 14. US labels base every number on the serving size, with servings per container shown just above it 5. Eat two servings and you double the carbohydrate 6.
In Canada, fibre and sugars must be listed under carbohydrate, and starch may be listed if the manufacturer chooses 2. According to the US Food and Drug Administration (FDA), total carbohydrate includes fibre, total sugars and sugar alcohols, and the label lists grams of fibre, total sugars and added sugars 4. Total sugars include both natural and added sugars 4.
Diabetes Canada teaches people to subtract fibre, because it does not raise blood sugar. What is left is called available carbohydrate: a food with 26 g of carbohydrate and 6 g of fibre has 26 g − 6 g = 20 g 3. Its clinical guideline says that when people match insulin to carbohydrate, fibre and sugar alcohols should be subtracted from the total 15.
The ADA takes a different line. It notes that terms such as “net carbs” haven’t been defined by the FDA, doesn’t recommend their use, and advises checking total carbohydrate first 5. Use the method your dietitian or diabetes educator teaches, the same way each time.
| Label item | Canada | United States |
|---|---|---|
| Serving size | Household measure, then grams or mL; amount typically eaten in one sitting 14 | All numbers are based on it; servings per container shown above it 5 |
| Carbohydrate total | Includes fibre, starch and sugars 2 | Includes fibre, total sugars and sugar alcohols 4; starch too 5 |
| Listed underneath | Fibre and sugars; starch optional 2 | Fibre, total sugars and added sugars 4 |
| Fibre in the count | Diabetes Canada subtracts it 3 | ADA starts from the total 5 |
In plain terms: Both countries put fibre inside the carbohydrate total. Diabetes Canada’s method takes it back out, so 26 g of carbohydrate with 6 g of fibre counts as 20 g. The ADA starts from the total. Either way, multiply by the servings you actually eat.
You can count either, because a carb choice is simply a portion with about 15 g of carbohydrate, so 45 g is three choices 6.
Choices come from exchange lists. Mayo Clinic explains that one serving in a food category is called a choice, and that on lists covering starch, fruit and milk, each choice is 15 g of carbohydrate 1. The ADA describes choices as a more basic version of carb counting, and counting every gram as the advanced form used to dose mealtime insulin 6.
Grams look more precise, but that doesn’t always make the count more accurate. In one study, 102 young people aged 8.3 to 18.1 on intensive insulin therapy and 110 caregivers estimated the carbohydrate in 17 standard meals containing 8 to 90 g. Seventy-three percent of estimates were within 10 to 15 g of the true amount, and teaching in grams did not improve accuracy compared with portions or exchanges 7.
Measure when you can: Diabetes Canada notes that you may need measuring cups and spoons to be accurate. When you can’t measure, your hands give a rough guide 3, 16.
Compare your measured portion with the label’s serving size to work out the carbohydrate 3. Because Canadian serving sizes also give grams or millilitres 14, a kitchen scale lets you match grams directly. For foods without a label, such as whole fruit, the ADA points to apps and the US Department of Agriculture’s food composition database 6.
Diabetes Canada’s handy portion guide suggests 16:
Estimates drift with meal size. In the study above, large meals tended to be underestimated and snacks overestimated 7. Restaurants are harder still, with no scales or labels to check against 3.
People on fixed insulin doses are advised to eat consistent amounts of carbohydrate at consistent times, while people who take insulin with meals match each dose to the carbohydrate they eat 6, 8. Mayo Clinic notes that a diabetes meal plan will likely include carb counting, especially if you have type 1 diabetes or use insulin 17.
The ADA’s 2026 Standards recommend counselling people on fixed insulin doses about consistent carbohydrate intake in both timing and amount, taking into account how long their insulin acts, because this can improve glucose and reduce the risk of lows 8. Diabetes Canada suggests trying to meet your target within 5 g per meal or snack 3.
The ADA describes this as the advanced form of carb counting, recommended for people on intensive insulin therapy by injection or pump, such as people with type 1 diabetes and some with type 2 6. It uses an insulin-to-carbohydrate ratio to work out a mealtime insulin dose from the grams of carbohydrate in a meal 6. Your ratio is set and adjusted with your care team; this article gives no numbers to use.
Diabetes Canada’s guideline says matching injections of rapid-acting insulin to carbohydrate allows flexibility in meal size and frequency 15. The ADA recommends education on how carbohydrate, fat and protein affect glucose, tailored to each person’s insulin plan 8. Insulin from an earlier dose may still be working; see insulin on board and types of insulin.
The ADA calls carb counting “a more advanced skill” and describes the diabetes plate method as a simpler option: a 9-inch plate with half nonstarchy vegetables, a quarter protein and a quarter carbohydrate. It cites a randomized trial in which both methods helped improve A1c 8. Automated insulin delivery (AID) systems, which adjust pump insulin with software, still need basic carb counting, but their algorithms work with less accurate carbohydrate entries 8.
Training people with type 1 diabetes to match insulin to carbohydrate has lowered A1c in trials, though the pooled benefit is clearest in adults 9, 18.
The best-known trial is DAFNE (Dose Adjustment For Normal Eating), published in the BMJ in 2002 9. It randomly assigned 169 adults with type 1 diabetes and moderate or poor glucose control to a five-day outpatient course, taught in groups of six to eight, or to a waiting list. The course taught people to match insulin to the carbohydrate they wanted to eat, meal by meal. After six months, A1c was 8.4% (68 mmol/mol) with the course against 9.4% (79 mmol/mol) on the waiting list. Quality of life and dietary freedom improved, while severe lows, weight and blood fats did not change. At one year, A1c was still 0.5 percentage points below where it started 9.
A 2014 meta-analysis pooled seven trials with 599 adults and 104 children 18. Overall, carb counting did not significantly improve A1c compared with other advice or usual care (−0.35 percentage points; 95% confidence interval −0.75 to 0.06). In five parallel-design trials in adults, it lowered A1c by 0.64 points (7.0 mmol/mol). The authors concluded there is some evidence for carb counting in adults with type 1 diabetes 18.
A 24-week trial assigned 273 people with type 2 diabetes using background insulin (glargine) and mealtime insulin (glulisine) to set mealtime doses or to an insulin-to-carbohydrate ratio, with doses adjusted weekly from meter readings in both groups 19. A1c fell by 1.46 points with set doses and 1.59 points with carb counting, a difference that was not significant 19.
A small study gave 31 children and teenagers aged 9.5 to 16.8 an insulin dose calculated for 60 g of carbohydrate, then lunches with 50, 60 or 70 g on different days 20. A 10 g difference either way made no difference to glucose for two and a half hours. After that, the 70 g lunch ran higher, by up to 1.9 mmol/L (34 mg/dL) at three hours, though glucose stayed within international targets 20. The ADA notes that smartphone apps, CGM and AID systems may reduce the need for precise counting 8.
There is no single right amount: the ADA’s 2026 Standards state that there is no ideal percentage of calories from carbohydrate, protein or fat for people with diabetes 8.
The ADA’s 2019 nutrition consensus report said the mix should be based on each person’s eating patterns, preferences and metabolic goals 10. Diabetes Canada’s 2018 nutrition guideline says the balance is flexible within recommended ranges, which include 45% to 65% of energy from carbohydrate, depending on individual goals 15. Its carb-counting page says most people aim for 45 to 60 g at meals and 15 to 30 g at snacks 3. Your own target is set with your care team 12.
The ADA consensus found that reducing overall carbohydrate intake has the most evidence for improving glucose. It defined low-carbohydrate eating as 26% to 45% of calories from carbohydrate and very-low-carbohydrate eating as under 26% 10. In one meta-analysis it cites, the A1c benefit was more pronounced with very-low-carbohydrate eating at 3 and 6 months, but not at 12 and 24 months 10. The 2026 Standards suggest considering lower carbohydrate intake for some adults, for example by eating fewer processed foods. Very-low-carbohydrate plans are not currently recommended during pregnancy or breastfeeding, for children, for people with kidney disease, or for people with or at risk for disordered eating, and should be avoided with SGLT2 inhibitors, a group of glucose-lowering medicines, because of the possible risk of ketoacidosis 8.
Diabetes Canada’s 2020 position statement says healthy low-carbohydrate (50 to 130 g a day) or very-low-carbohydrate (under 50 g a day) diets can be considered as one healthy eating pattern for people with type 1 or type 2 diabetes. For type 2 diabetes, it found that very-low-carbohydrate diets may improve glucose control more than higher-carbohydrate diets and reduce the need for medicines for up to 12 months, but that evidence on longer-term benefit is limited 11.
Cutting carbohydrate can lower glucose quickly. The ADA consensus says starting a very-low-carbohydrate plan can cause diuresis (passing more urine) and swiftly reduce glucose, so a knowledgeable practitioner should be involved from the start to prevent dehydration and to reduce insulin and other glucose-lowering medicines to prevent lows 10. Diabetes Canada adds that insulin or sulphonylurea (a type of diabetes pill) doses may need to be reduced or stopped, and that SGLT2 inhibitors may raise the risk of ketoacidosis on a low-carbohydrate diet 11. Don’t change a medicine or dose on your own.
There is no single right number; the ADA says there is no ideal percentage of calories from carbohydrate for people with diabetes 8. Diabetes Canada’s carb-counting page says most people aim for 45 to 60 g at meals and 15 to 30 g at snacks, and your own target is set with your care team 3, 12.
Diabetes Canada says yes: fibre does not raise blood sugar, so it subtracts fibre from total carbohydrate 3. The ADA advises checking total carbohydrate first and doesn’t recommend “net carb” terms 5.
About 15 g 6. On exchange lists for starch, fruit and milk, each choice is 15 g of carbohydrate 1.
Not always. The ADA cites a trial in which the plate method and carb counting both helped improve A1c, and advises people on fixed insulin doses to keep carbohydrate consistent 8. In 273 people with type 2 diabetes on mealtime insulin, set doses adjusted weekly lowered A1c about as much as an insulin-to-carbohydrate ratio 19.
In 31 children and teenagers, a dose calculated for 60 g covered lunches of 50 or 70 g with no difference in glucose for two and a half hours, though the 70 g lunch ran higher afterwards 20. The ADA notes that AID systems work with less accurate carbohydrate entries 8.
A carb count is easier to judge next to the readings that follow it. In Glyve you can log a meal by typing, by voice or with a photo of the meal, its nutrition label or its barcode, and every estimate is a draft you check before saving. Food, insulin, exercise, sleep and stress sit on one timeline with your glucose, and the daily glucose graph shows your own target range, so you can see how a logged meal lined up with what came next. The week view shows whether a pattern repeats. Glucose can come from Apple Health, from Dexcom or FreeStyle Libre through your own Dexcom Share or LibreLinkUp follower account (readings can arrive later than in the CGM’s own app), or from meter readings you type in; a CGM is not required. A PDF summary, CSV or AGP report lets you share your logs with your care team when you choose. Glyve is not a medical device, gives no medical advice and recommends no insulin doses; your carbohydrate targets and any insulin-to-carbohydrate ratio come from your care team.
How we write about diabetes. Spotted a mistake? Email info@glyve.ca and we’ll check it against the source.
Glyve keeps glucose, food, insulin, exercise, sleep and stress on one timeline.
For iPhone, iPad and Apple Watch.