A comparison of prandial insulin dosing algorithms in children and teenagers with type 1 diabetes has revealed the deficiencies in all models – with only just over half of participants typically within the target blood glucose range after a high fat or high protein meal.

The NSW study of 29 participants aged 7-17 years, stable on inulin pump therapy and with HbA1c <8%, compared traditional carbohydrate counting with the Pankowska Equation for insulin dose estimation and the Food Insulin Index (FII).

The randomised crossover designed trial involved continuous glucose monitoring for five hours after two test meals using the three insulin dosing algorithms. Outcome measures included mean glucose excursions from baseline, mean glucose levels, and hypoglycaemic events defined as blood glucose less than 3.9 mmol/l.

Overall, participants counting carbs spent 54-56% of their time in target depending on the meal type, 51-57% for those using FII and 64-75% for those using the Pankowska Equation.

The study found no significant difference in any outcome measures between use of carbohydrate counting and use of the FII.

While the Pankowska Equation reduced mean and peak glucose excursions and increased time within target range, it came at the cost of an increased risk of hypoglycaemia.

“When insulin was dosed according to the Pankowska Equation, the insulin dose was approximately 24% higher for a high protein, and 17% higher for a high fat meal, than for carbohydrate counting.”

“This increase in insulin dose resulted in a significantly lower blood glucose excursion for the high protein meal for 90–240 min as well as a higher rate of hypoglycaemia following the Pankowska Equation for both high fat and high protein meals,” the study authors said.

Dr Carmel Smart, clinical research fellow and senior paediatric dietitian at the John Hunter Children’s Hospital, told the limbic a better insulin dosing was needed.

“And the reasons it matters, is with CGM, people can see it. It’s much more clinical relevant than it was even five years ago because they can actually see the excursions caused by food. We want to minimise excursions from food and for them to spend more time in target.”

“Out of those three algorithms, none of them manage food really well. The equation that is most commonly used for fat and protein in clinical practice, the Pankowska Equation, gives you too much insulin and therefore hypos.”

“What it probably does is add too much insulin for the fat and protein but I don’t think it appreciates that the insulin to carb ratio will also cover some of the fat.”

Dr Smart said FII has been used in studies but not in routine clinical practice.

She said John Hunter Children’s Hospital and the University of Newcastle had developed a new algorithm, currently in clinical trials, which she believed was superior.

“It’s reasonably complicated and adapts to the individual.”

She added some children and adults with diabetes were still not on individualised insulin doses.

“Certainly since 2004 in our centre, we’ve use individualised insulin to carb ratios for all kids so a newly diagnosed child might start on 0.5 units for one exchange or another child might start on three units for one exchange. It just depends on the child, pubertal stage, insulin sensitivity, beta cells remaining, age, sex – it depends on all those things.”

“And I think the issue is that is where some of push for low carbs comes from. People find they are not getting great results but it is partly because their insulin to carb ratios are not appropriate.”

“Diabetes is onerous to start with. I think our kids are good at counting carbs but with protein and fat, I think it’s difficult.”