Summary: In two acute randomised crossover trials in adults with type 1 diabetes, neither the type of dietary protein (n=16; egg, beef, chicken, salmon, whey, each 30 g) nor the glycaemic index of carbohydrate (n=8; low-GI 52% vs high-GI 76% bread) produced a statistically significant difference in postprandial glucose. Protein iAUC varied almost two-fold but did not differ between sources (p>0.05), and GI curves were virtually identical (iAUC p=0.185 at 1 h, p=0.694 at 5 h).
PICO Summary
| Element | Detail |
|---|---|
| Population | Adults with type 1 diabetes on insulin, dosed by individual insulin:carbohydrate ratio. Two acute randomised crossover trials (Australia): Study 1 n=16, Study 2 n=8. |
| Intervention | Study 1: five protein types (egg, beef, chicken, salmon or whey, each 30 g) with fried rice (45 g carbohydrate). Study 2: high-GI bread (GI 76%) with peanut butter (19 g protein, 30 g fat). |
| Comparison | Study 1: protein sources compared against one another (within-subject crossover). Study 2: low-GI bread (GI 52%), same peanut butter accompaniment. |
| Outcome | Postprandial glucose iAUC, 30 min pre- to 5 h post-meal. Study 1 (null): chicken 203±66, egg 263±100, beef 309±89, salmon 338±83, whey 397±115 mmol·min/L; almost two-fold range but no significant difference between sources (p>0.05). Hypoglycaemia (≤3.5 mmol/L) occurred at least once per protein type with no significant difference in risk (p>0.05). Study 2 (null): glucose curves virtually identical; iAUC not significant at 1 h (p=0.185), 3 h (p=0.538) or 5 h (p=0.694). No CIs, ARR or NNT reported. |
Expert Commentary
The verdict here is a clean null result, and the headline framing should be read as such rather than as evidence that protein or carbohydrate type meaningfully shifts postprandial glucose. Across both small randomised crossover trials, no statistically significant difference was detected: protein iAUC spanned an almost two-fold range yet failed to separate the five sources, and the high- versus low-GI comparison yielded near-superimposable curves with p-values far from significance at every timepoint. The most weighable limitation is power. With only 16 and 8 participants and highly variable individual responses, these trials were realistically positioned to miss modest true effects, so absence of a significant signal should not be over-read as proof of strict equivalence; a wide range that does not reach significance is as much a statement about sample size as about biology. The meals were also high in fat and protein, which blunts and delays glucose excursions and may itself mask GI effects seen with leaner meals. Can I use this with my patients? Yes, in a measured way: for an adult with type 1 diabetes eating mixed, higher-fat meals, it is reasonable to reassure them that switching between common protein sources or fussing over bread GI is unlikely to be the lever that fixes post-meal glucose, and that insulin dosing and timing matter more. Larger confirmatory trials would be welcome before this is written into dosing algorithms.
References
Li X, Wainwright A, Fio CZ, Brodie S, Alexander K, McGill M, et al. Do the types of dietary carbohydrate and protein affect postprandial glycemia in type 1 diabetes? Nutrients. 2025;17(11):1868. doi:10.3390/nu17111868