An introduction to carbohydrate counting in type 1 diabetes

Before the discovery of insulin in 1921, the controlled intake of food (carbohydrates) was the only known means of regulating blood glucose levels (Franz, 2004). Indeed, Robert Daniel Lawrence, the British physician who was an early recipient of insulin injections, was put onto the Allen dietary treatment of “absolute starvation” (Hill and Eckman, 1915; Box 1), with the predicted lifespan on this treatment being 3–4 years (Joslin, 1917), before he was treated with insulin.

Around this time, Joslin (1927) was researching the effects of dextrose and starches on blood glucose. As Franz (2004) notes,

“The physician issued the order and the dietitian developed an ‘ideal’ meal plan; the ‘patient’ received instructions on how to comply with the doctor’s orders.”

Restricting carbohydrate consumption continued up until the 1990s. If people with diabetes reported they were hungry, they were told to eat cheese or meat as snacks – foods that were carbohydrate-free but contributed extra fat and calories to their diet. Up until 1994, nutritional recommendations for people with diabetes attempted to define an ideal nutritional prescription that would apply to nearly everyone with the condition (Franz, 2004). Thus, ideal percentages of energy obtained from carbohydrate, protein and fat were recommended. Historically, the use of carbohydrate exchanges (now referred to as counting or estimation) was a restrictive process to eliminate the variance in blood glucose levels created by irregular eating, which could not be managed with biphasic (mixed) insulin (Oliver, 2008). We now know there is no evidence for a recommended ideal amount of carbohydrate for maintaining long-term glycaemic control in people with type 1 diabetes (Dyson et al, 2011).

The 10-g carbohydrate exchange lists used in the 1980s were prescriptive and limited total carbohydrate intake. In the example in Figure 1, a current patient of mine, Vicci, was allowed 125 g of total carbohydrate each day, meaning she could pick 12.5 exchanges from the list provided. This would be distributed throughout the day as three meals and three snacks to match the inflexible twice-daily biphasic insulin injections. Missing these prescribed snacks meant that a hypoglycaemic episode was almost guaranteed.

Indeed, the advice within other diet sheets published at the time, such as the North Warwickshire Health Authority’s Diet Information for Insulin-Dependent Diabetics, was clear:

“There are three rules for diabetics:

1. Eat your meals regularly.
2. Avoid sugar and sugary foods.
3. Eat a controlled amount of carbohydrate to balance your dose of insulin.”

Modern-day carbohydrate counting
Carbohydrate is the main nutritional consideration for glycaemic control in individuals with type 1 diabetes, and both the amount and the type of carbohydrate have an effect on post-prandial blood glucose levels (Dyson et al, 2011).

Carbohydrate counting is a meal-planning approach used to manage diabetes that focuses on the total amount of carbohydrate eaten at meals and in snacks (Gillespie et al, 1998). Taking a greater bolus insulin dose when eating a meal with more carbohydrate makes perfect sense, and most people with type 1 diabetes use this approach every day, based on their experience with the condition. Carbohydrate counting skills enable individuals to assess the impact of different meals on blood glucose levels and their resultant bolus insulin requirements. For example, a perceived “big” meal such as a roast dinner (in which there is lots of food on the plate but in which potatoes and Yorkshire pudding make up the only carbohydrate) may contain half the carbohydrate of a meal such as spaghetti bolognaise.

Randomised controlled trials have shown that carbohydrate counting can improve glycaemic control and quality of life without increasing the risk of severe hypoglycaemia, weight gain or increases in blood lipid levels (Diabetes Control and Complications Trial Research Group, 1993). Carbohydrate counting and insulin adjustment have been shown to be beneficial and cost-effective over the long term (Speight et al, 2007).

To teach the skills needed to carbohydrate count, it is vital to go back to basics and identify all foods that contain carbohydrate; doing this by performing your own research reading the information on that “meal deal” you bought for lunch is helpful. Food labels and packets have lots of information on them, so it is useful to have some handy to show people in clinic. For many of us, breakfast and lunch tend to be quite similar from day to day; indeed, many people are happy to eat the same breakfast every day.

People with any type of diabetes would benefit from becoming “carbohydrate aware,” taking note of the total carbohydrate levels displayed on the back of packets. Note that the front of food packaging will only provide information about the “of which sugars” carbohydrate value, whereas the value that is needed – total carbohydrate – will be found on the reverse in the more detailed nutritional information. Examples of common misconceptions are that potatoes are a vegetable, and so do not count, and that fruits are calorie- and carbohydrate-free.

As a dietitian, I sometimes teach carbohydrate counting on a one-to-one basis or conduct a joint appointment with our DSN; however, the teaching appears to be most effective when we hold group education sessions, with time for reflection and personal learning. People report that they react to each blood glucose level in the “here and now,” not having the time to reflect on whatever trends may be occurring in their blood glucose levels. Reviewing these results in a group situation presents a huge learning opportunity, for group members and healthcare professionals alike, as the social learning theory (Bandura, 1977) suggests that people learn by observing, imitating and modelling from each other.

Since 13 December 2014, new food labels suggest an average healthy adult can consume a Reference Intake (RI; replaces Guideline Daily Amount) of 260 g of total carbohydrate per day (based on an average adult consuming 2000 kcal/day). Note that the RI is not a target, just a reference. Our goal is neither to encourage people with type 1 diabetes to eat more carbohydrate nor to suggest unnecessary restriction; rather, it is to suggest that they can eat any type of carbohydrate as long as it can be accurately recorded and the correct amount of bolus insulin administered. Carbohydrate intake varies greatly depending on activity levels, body size and food preferences, to name just a few influences. In a recent type 1 diabetes education group I ran with six people, the variation was huge: from 90 g to 350 g of total carbohydrate in 1 day.

The calculations
When teaching in our clinic, we have chosen the method whereby one unit of bolus insulin is given for a calculated amount of carbohydrate, measured in grams.

Step 1: Total daily dose
The first step is to calculate the total daily dose (TDD); that is, the basal and bolus insulin doses added together. An average figure taken from the past 3 days is most useful.

To calculate the total daily dose (TDD):
Add together the bolus insulin taken in a day.

For example:
Breakfast: 8 units +
Lunch: 10 units +
Dinner: 12 units
Total bolus = 30 units

Add this to the daily basal insulin dose. For example:
Bolus insulin = 30 units +
Basal insulin = 20 units
TDD = 50 units

The worked example above shows the details and, while it is not something that we need to do with all of our type 1 patients, it gives an understanding of the process. However, the 2011 Skills for Life survey revealed that 49.1% of adults in the UK (16.8 million) have numeracy skills equivalent to a 9–11-year-old child (Department for Business Innovation & Skills, 2012), and the potential for panic or anxiety when discussing maths in a group is something we need to be aware of and sensitive to. Giving worked examples and encouraging group members who are more confident to help others is a useful strategy.

Step 2: The 500 and 100 rules
The next step is to use the TDD in the “500 rule” and the “100 rule” (King and Armstrong, 2007; Davidson et al, 2008). In my experience, giving people ratios to use as a starting point and then discussing the maths at a later meeting has been more successful.

The 500 rule to calculate the insulin-to-carbohydrate ratio:
1 unit of bolus insulin covers:
500 ÷ TDD = __ g of carbohydrate

For example, 1 unit of bolus insulin covers
500 ÷ 50 = 10 g of carbohydrate

In this example, this means that 1 unit of bolus insulin will cover 10 g of carbohydrate consumed. So, if eating four crumpets with butter (80  total carbohydrate), use 80 ÷ 10 = 8 units. Whereas, if eating two breakfast wheat biscuits with 200 mL milk (40 g total carbohydrate), use 40 ÷ 10 = 4 units.

Among different individuals, the doses of insulin (and therefore the TDD) will vary greatly depending on their body size, activity levels and duration of type 1 diabetes, so the insulin-to-carbohydrate ratios in a single group could vary greatly – in my experience, from 1 unit per 3 g carbohydrate up to 1 unit per 21 g carbohydrate within a group of eight people.

Using the 100 rule, we can calculate the insulin sensitivity factor (ISF; also known as the correction dose). The example below shows that 1 unit of bolus insulin lowers blood glucose by 2 mmol/L when the TDD is 50 units.

The 100 rule to calculate the insulin sensitivity factor/correction dose:
1 unit of bolus insulin will lower blood glucose by 100 ÷ TDD = __ mmol/L

For example, 1 unit of bolus insulin will lower blood glucose by 100 ÷ 50 = 2 mmol/L

Blood glucose meters offering bolus advice are available in the form of the Accu-Chek^® Aviva Expert (Roche Diagnostics, Burgess Hill) or the FreeStyle^® InsuLinx meter (Abbott, Maidenhead). These can be a real help in giving people more confidence, as they do the maths while allowing individuals the autonomy to review the dose suggested. It is always necessary to count the total carbohydrate in a meal, after which blood glucose is measured and the amount of carbohydrate to be eaten is entered, and the meter advises how much bolus insulin is required. This suggests an insulin dose both for the carbohydrate and to determine how much insulin could also be added for the ISF to bring blood glucose into the range set on the meter. The meters are set up individually so that people can agree their own goals. The experience of one of my patients with one of these meters is described in Box 2.

Practical tips on educating people with diabetes
The Dose Adjustment for Normal Eating (DAFNE) programme originated in Germany and has been running there since the 1980s, first coming to our shores in the early 2000s (DAFNE Study Group, 2002). This structure was then adapted by the Bournemouth Diabetes and Endocrine Centre team, who introduced the Bournemouth Type 1 Intensive Education (BERTIE) model (Everett, 2003). After training in BERTIE, our team now delivers its own version of the course to groups of eight participants with type 1 diabetes, a DSN and a diabetes specialist dietitian. Across the UK, 84 centres in the Diabetes Education Network run their own type 1 diabetes education programmes. Our course runs over four successive weeks, allowing a week between each session for practical learning from the 21 meals (plus snacks) that will be eaten throughout the 7 days, and possibly some exercise.

Eating lunch together in a group over these successive weeks is a useful and safe way to learn, allowing group members to help each other out with the sums. This has also highlighted the fact that some supermarkets do not label their pre-packed sandwiches with the total carbohydrate values per pack, they are only listed per 100g. One solution to this is to weigh the sandwiches and make a calculation. However, while scales are freely available in this learning environment, they are not so common in the “real world.” An alternative approach is to estimate the carbohydrate values of the bread and then identifying other carbohydrate sources, such as pickle, in that sandwich. By weighing a chocolate crispy cake or tiny muffin, we can get around the fact that fresh bakery items are currently exempt from nutrition labelling. One such experience that occurred in one of our education sessions is described in Box 3.

Taking photos of actual dinners is helpful in teaching this concept – some people are happy to share their own photos for this exercise; otherwise, they have the pleasure of looking at what is eaten in my household.

Conclusion
Carbohydrate counting is a proven and robust method to help manage type 1 diabetes and I hope that this article has given a clearer understanding of what it is and how it can work for the people with diabetes that we work with. Using the resources that are now available will help the process. As with all new skills, it will take some time to implement, but seeing the results in improved confidence among people who are more in control of their eating and diabetes is worth the extra work.

Helpful resources

• The Bournemouth Diabetes and Endocrine Centre has an excellent website that takes you through the carbohydrate counting process, and is useful for both people and healthcare providers (available at: www.bdec-e-learning.com).
• The Diabetes UK Carbs Count resource is available to download for free from the Diabetes UK shop (available at: http://bit.ly/1DFRXe8), and takes you through the process with worked examples.
• The Diabetes Education Network (available at: www.diabetes-education.net) offers support and resources allowing diabetes teams to provide high-quality self-management education for their patients. The Type 1 Diabetes Workbook (available at: http://bit.ly/1xYmxRC) is an invaluable part of all our locally run sessions.
• Other resources to help with carbohydrate counting can be found at the Carbs & Cals website (available at: www.carbsandcals.com), which allows you to “eyeball” a portion of carbohydrate on a plate, from over 3500 photos. This resource is available as a book or as an app for Android and iOS devices.
• For home-cooked recipes, the Cook and Count app (available at: www.healthapps.uk.com) allows you to build your own recipes from a list of 2000 ingredients, divide into portions and see the carbohydrate content of each portion. You can create your own digital recipe file, upload photos and share via social media.

Acknowledgements
Thank you to Vicci, Louise and Sarah for giving permission for me to share some of their stories.