Summary: In a small crossover trial in type 2 diabetes, a dairy-enriched diet increased several circadian gene-expression measures compared with a non-dairy diet with structured meal timing. Reported glycaemic improvements were secondary, and the headline glucose changes were within-phase comparisons with baseline, not treatment differences between diets.
PICO Summary
| Element | Detail |
|---|---|
| Population | 25 adults with type 2 diabetes and HbA1c at least 6.5%, managed with diet or stable oral glucose-lowering therapy, at a single Israeli diabetes centre. Nineteen completed both dietary phases. All 13 starting with dairy completed both phases, compared with 6 of 12 starting without dairy. |
| Intervention | Four weeks of an isoenergetic dairy-enriched diet (YesMilk), incorporating a high-protein breakfast and early daytime-restricted carbohydrate intake. |
| Comparator | Four weeks of a non-dairy isoenergetic diet (NoMilk) with structured meal timing, in the alternate crossover phase; 3-4-week washout between phases. |
| Outcomes | Primary circadian gene-expression findings favoured the dairy phase: BMAL1 +1.8-fold (P=0.0003), REV-ERB alpha +2.2-fold (P<0.001), CRY1 +1.4-fold (P=0.03), and higher PER1 expression (between-diet P=0.01). Secondary changes during the dairy phase versus baseline included fasting glucose approximately -1.7 mmol/L, glucose management indicator -0.7 percentage points and time in range +9 percentage points (all P<0.05). These are not between-diet effect estimates. Hunger and sweet cravings decreased by 15-20%. The abstract does not quantify adverse events or give confidence intervals for these effects. |
| Study Design | Randomised, open-label crossover trial; two 4-week phases separated by a 3-4-week washout. Coin-flip sequence allocation was not concealed. Powered for gene expression, not clinical diabetes outcomes. ClinicalTrials.gov NCT03772067. |
Expert Commentary
I view this as an interesting mechanistic experiment, not a basis for prescribing a particular dairy regimen to improve diabetes outcomes. The study was powered for circadian gene expression, while glucose measures and appetite were secondary. Importantly, the widely quoted glucose changes describe improvement from baseline during the dairy phase; they should not be presented as the causal advantage over the comparator diet. The most consequential limitation is differential completion: every participant starting with dairy completed both phases, whereas only half starting without dairy did so. In a small crossover study, that imbalance can materially change the analysed population and weaken the apparent precision of the findings. Open allocation and several interrelated dietary components further complicate attribution to dairy itself. Can I use this with my patients? I would discuss this as preliminary research with adults interested in meal timing, while individualising dietary advice around preferences, nutritional adequacy and existing treatment. It does not establish that dairy is necessary, that measured gene changes are clinically beneficial, or that medication can be reduced. The absence of detailed adverse-event reporting in the abstract is not proof of safety. I would want a larger, independently replicated trial with concealed allocation, robust retention and prespecified between-diet glycaemic comparisons.
References
Tsameret S, Froy O, Matz Y, et al. Glycaemic, appetite and circadian benefits of a dairy-enriched diet with high-protein breakfast and early daytime-restricted carbohydrate intake in type 2 diabetes: a randomised crossover trial. Diabetologia. 2026;69(4):1021-1034. doi:10.1007/s00125-025-06658-2. PMID: 41578008.
