Adding protein to the bedtime snack steadied the early night

Design

randomised crossover, exploratory pilot, blinded CGM

Population

12 children and adolescents with type 1 diabetes on glargine/lispro multiple daily injections

Primary outcome

overnight time in range 70–180 mg/dL, 21:00–09:00

Effect

first 6 hours, carbohydrate+protein vs carbohydrate+fat: time in range 91.2% vs 71.4% (adjusted p = 0.011); peak-to-nadir 88.1 vs 123.9 mg/dL (adjusted p = 0.041)

Twelve children and adolescents with type 1 diabetes on glargine and lispro by multiple daily injections completed four bedtime snack protocols in random order: carbohydrate alone, carbohydrate plus protein, carbohydrate plus fat, and carbohydrate plus fibre. Each snack was eaten at 21:00 for three consecutive nights with no additional bolus, and blinded CGM was analysed from 21:00 to 09:00. There was no sample size calculation; the authors call it an exploratory pilot.

Overnight time in range differed across protocols (p = 0.014), highest after carbohydrate plus protein at 85.8% and lowest after carbohydrate plus fat at 70.9%, though adjusted whole-night pairwise comparisons were not significant. The clearer signal was in the first six hours, where carbohydrate plus protein gave higher time in range than carbohydrate plus fat (91.2% vs 71.4%, adjusted p = 0.011) and a smaller peak-to-nadir excursion (88.1 vs 123.9 mg/dL, adjusted p = 0.041). Hypoglycaemia and hyperglycaemia event rates did not differ.

The practical reading is narrow but real: if a family is already giving a bedtime snack, a fat-heavy one is the least helpful choice for early-night stability. It says nothing about whether a snack is needed at all — there was no no-snack arm — and the snacks were not matched for energy.

• Ask what the bedtime snack actually is, not just whether one is given
• Where a snack is used, a carbohydrate-plus-protein option is the more sensible default
• Do not add a snack on this evidence alone in a child who does not currently have one
• Review overnight CGM traces before and after any change rather than assuming the effect

Why it matters

The composition of a snack families already give may matter more than whether they give it.

Don't overread it

The authors are explicit that this does not show any snack composition protects against nocturnal hypoglycaemia.

The statistics, in plain English

Twelve participants is very small, and with four comparisons the chance of a spurious result is high — which is why only two pairwise differences survived correction. The whole-night difference did not survive at all. Unmatched energy content means the snacks differed in more than the nutrient being studied, so the protein itself may not be what mattered.

Read the rest in the app

You have read your two free briefings this month. The app carries all 27 specialties, every morning, free — and this finding is waiting in it.

Get Daily Doctor — free

Scan to keep reading on your phone. No account needed to start.