Ben-Yacov · Diabetes care 2021 · randomized controlled trial · n=225

Personalized Postprandial Glucose Response-Targeting Diet Versus Mediterranean Diet for Glycemic Control in Prediabetes.

Cited 150 times in the scientific literature.

Level 2 - randomized trial

Individual randomized controlled trial

PubMed 34301736 · doi:10.2337/dc21-0162 · record verified 2026-08-30

What was done

Researchers randomized 225 adults with prediabetes (mean age 50 years, BMI 31.3 kg/m², baseline HbA1c 5.9%) to either a Mediterranean (MED) diet or a personalized postprandial-targeting (PPT) diet for a 6-month active intervention followed by a 6-month observation period. The PPT diet was generated by a machine learning algorithm integrating clinical and microbiome data to predict postprandial glucose responses. Participants tracked food intake via a smartphone application and wore continuous glucose monitors (CGM).

What was found

A total of 200 participants completed the 6-month intervention, and 177 completed the 12-month follow-up. At 6 months, reduction in daily time with glucose >140 mg/dL was significantly greater in the PPT group compared to the MED group (-1.3 ± 1.5 h/day vs -0.3 ± 0.8 h/day; between-group difference 95% CI -1.29 to -0.66, P < 0.001). HbA1c reduction was also greater in the PPT group (-0.16 ± 0.24% vs -0.08 ± 0.19%; between-group difference 95% CI -0.14 to -0.02, P = 0.007). These differences persisted at 12 months. No significant between-group difference was observed on a CGM-measured oral glucose tolerance test.

Why it matters

This study shows that algorithm-guided personalized dietary advice can produce modestly greater improvements in glycemic control than an established healthy dietary pattern like the Mediterranean diet in prediabetes.

Limits

The absolute difference in HbA1c reduction between groups was small (-0.08% absolute difference). The intervention did not yield significant differences on continuous glucose-measured oral glucose tolerance tests, diet tracking relied on self-reports, and the trial did not assess progression to clinical type 2 diabetes or cardiovascular outcomes.

Cited by