Adherence to the EAT-Lancet reference diet and risk of diabetes mellitus: a systematic review and dose-response meta-analysis of prospective studies

A systematic review and dose-response meta-analysis of 10 prospective studies found that higher adherence to the EAT-Lancet reference diet was associated with a lower risk of diabetes mellitus. The analysis included 722,078 participants and 72,201 diabetes cases. Comparing participants in the highest diet-score category with those in the lowest produced a pooled relative risk of 0.76 (95% confidence interval, 0.68–0.95), corresponding to an estimated 24% lower risk among the highest-adherence group.

The prediction interval was 0.50–1.16, showing that the magnitude of the association could vary substantially across comparable study settings. Because the evidence came from prospective observational studies, the result does not establish that following the reference diet caused the lower risk. The researchers searched PubMed, the Web of Science, Embase and the China National Knowledge Infrastructure for studies published through December 2025.

Given substantial variation among the included studies, they combined estimates with a random-effects model using the DerSimonian and Laird method. For the highest-versus-lowest comparison, that range extended from a relative risk of 0.50 to 1.16. The confidence interval around the pooled estimate supported an inverse association, while the broader prediction interval indicated that the findings were not uniform enough to imply a precise effect in every setting.

Scores were also examined as continuous measures of adherence. Each standard-deviation increase in the EAT-Lancet reference diet score was associated with an 8% lower risk of diabetes mellitus, with a relative risk of 0.92 (95% confidence interval, 0.89–0.95). The prediction interval for this estimate was 0.81–1.04, and statistical heterogeneity was high, with an I² of 93.9%.

The test for nonlinearity was not significant (P = 0.549), while the overall dose-response test did not reach the conventional 0.05 threshold (P = 0.061). The point estimates were consistent with progressively lower diabetes risk as adherence increased, but the analysis did not provide statistically conclusive evidence of a dose-response gradient. Subgroup analyses identified associations in selected parts of the evidence base.

Among participants in studies with a mean age below 50 years, the highest-versus-lowest comparison yielded a pooled relative risk of 0.64 (95% confidence interval, 0.41–1.00; P = 0.05). This estimate was highly heterogeneous (I² = 94.2%; P < 0.001), and its prediction interval ranged widely from 0.08 to 5.44, indicating substantial uncertainty about the effect in a new comparable population. In a subgroup defined specifically by diabetes mellitus as the outcome, the pooled relative risk was 0.56 (95% confidence interval, 0.45–0.69).

These subgroup estimates do not demonstrate that the association was stronger in one group than another, nor do they establish that the diet would have a uniform effect across ages or populations. The review’s central conclusion was that higher EAT-Lancet reference diet adherence is associated with reduced diabetes mellitus risk, with the authors describing adherence to a sustainable plant-based diet as a potential strategy for addressing the global public-health challenge of diabetes. That interpretation must remain within the limits of the observational evidence.

The high heterogeneity across studies limits the precision and generalizability of the pooled estimates, while the prediction intervals show that future findings could differ materially from the average effects reported here. The authors called for larger prospective studies in ethnically diverse populations to confirm the association. Whether the reference diet itself produces the observed reduction remains uncertain because the included studies measured adherence and diabetes outcomes without establishing causation.

Subject of Research: Association between EAT-Lancet reference diet adherence and diabetes mellitus risk in prospective cohorts

Article Title: Adherence to the EAT-Lancet reference diet and risk of diabetes mellitus: a systematic review and dose-response meta-analysis of prospective studies

Article References: Bao, C., Miao, S., Zhang, Z., Shi, G., & Zhu, Z. (2026). Adherence to the EAT-Lancet reference diet and risk of diabetes mellitus: a systematic review and dose-response meta-analysis of prospective studies. BMC Public Health. https://doi.org/10.1186/s12889-026-29725-8

Image Credits: AI Generated

DOI: 10.1186/s12889-026-29725-8

Keywords: EAT-Lancet diet, diabetes mellitus, dietary adherence, dose-response meta-analysis, prospective cohort studies, plant-based nutrition, public health, observational evidence, Adherence, EAT-Lancet, reference, diet