Cardiology · Medical Literature
Effects of a Fasting-mimicking Diet on MRI-assessed Cardiac Structure and Function in Participants with Type 2 Diabetes.
Abstract
Purpose To evaluate the effects of a monthly 5-day fasting-mimicking diet (FMD) for 1 year on left ventricular (LV) geometry and function in individuals with type 2 diabetes (T2D). Materials and Methods This secondary analysis of the prospective Fasting In diabetes Treatment trial (ClinicalTrials.gov identifier no. NCT03811587) included individuals with T2D randomized to monthly 5-day FMD or usual care for 12 months (November 2018-July 2020). MRI (3.0 T) was performed at baseline, 6 months, and 12 months. LV geometry and systolic function were assessed from two-dimensional cine images, and diastolic function was assessed from four-dimensional flow MRI. Intention-to-treat linear mixed models and linear regression were used to evaluate treatment effects and LV-metabolic associations. Results A total of 82 participants (mean age, 62.9 years ± 8.5 [SD]; 43 female) were included. FMD reduced LV mass at 6 months (mean difference, -5.47 g [95% CI: -9.19, -1.75]; P = .005) but not at 12 months (P = .83). End-diastolic volume remained stable in the FMD group versus controls (mean difference, 7.75 mL [95% CI: 0.19, 15.32]; P = .045). Both the early-to-late peak filling ratio (E/A) (mean difference, 0.27 [95% CI: 0.04, 0.50]; P = .02) and the mitral annular early peak velocity (Ea) (mean difference, 0.84 cm/sec [95% CI: -0.02, 1.67]; P = .04) increased at 12 months. No other treatment effects were observed (all P > .05). ∆E/A was positively associated with change in hemoglobin A1c (HbA1c) level (standardized β = 0.47, P = .01) and negatively associated with changes in high-density lipoprotein cholesterol levels (standardized β = -0.55, P = .01). Conclusion FMD positively influenced LV diastolic function by increasing E/A and Ea in participants with T2D; E/A improvements were associated with reduced HbA1c levels. Keywords: MR Imaging, Cardiac, Left Ventricle, Outcomes Analysis, Metabolic Disorders, Diet, Fasting-mimicking Diet, Left Ventricle Function, Type 2 Diabetes, Cardiac Magnetic Resonance ClinicalTrials.gov identifier number NCT03811587 Supplemental material is available for this article. © RSNA, 2026.
Clinical relevance
Diabetic cardiomyopathy — characterized by LV diastolic dysfunction and adverse remodeling — is a major driver of heart failure risk in the rapidly growing type 2 diabetes (T2D) population, yet dietary interventions targeting cardiac structure and function remain understudied. This RCT secondary analysis addresses whether a structured, cyclical fasting-mimicking diet (FMD) can produce measurable, MRI-quantified improvements in LV geometry and diastolic function over 12 months. The findings are clinically timely given increasing interest in non-pharmacological strategies to mitigate cardiometabolic risk in T2D beyond glycemic and lipid control alone.
Key findings
1. LV mass was significantly reduced in the FMD group at 6 months (mean difference −5.47 g; 95% CI: −9.19 to −1.75; P = .005), though this effect was not sustained at 12 months (P = .83), suggesting early but potentially transient favorable LV remodeling. 2. End-diastolic volume (EDV) remained stable in the FMD group compared to controls at 12 months (mean difference +7.75 mL; 95% CI: 0.19 to 15.32; P = .045), indicating preservation of LV filling capacity. 3. The early-to-late peak filling ratio (E/A), a key 4D flow MRI index of diastolic function, was significantly improved in the FMD group at 12 months (mean difference +0.27; 95% CI: 0.04 to 0.50; P = .02). 4. Mitral annular early peak velocity (Ea) — a marker of myocardial relaxation — was significantly increased in the FMD group at 12 months (mean difference +0.84 cm/sec; 95% CI: −0.02 to 1.67; P = .04). 5. Improvement in E/A (ΔE/A) was positively associated with reduction in HbA1c (standardized β = 0.47; P = .01), linking glycemic improvement to diastolic benefit. 6. ΔE/A was negatively associated with change in HDL cholesterol levels (standardized β = −0.55; P = .01); no other statistically significant treatment effects were observed for remaining structural or systolic parameters (all P > .05).
Practice implications
Clinicians managing T2D patients with early or subclinical diastolic dysfunction should consider that a monthly 5-day FMD protocol, sustained over 12 months, can produce meaningful improvements in LV diastolic indices (E/A and Ea) detectable by cardiac MRI, particularly in patients who achieve concurrent HbA1c reduction — supporting integration of structured dietary cycling into comprehensive cardiometabolic risk management. The attenuation of LV mass reduction between 6 and 12 months warrants further investigation into optimal FMD frequency and duration to sustain structural benefits, and patients should be counseled that long-term adherence may be required for durable cardiac effects. Given the association between diastolic improvement and glycemic control, FMD may be most impactful when deployed in patients with suboptimally controlled T2D (elevated HbA1c) as an adjunct to pharmacotherapy, with cardiac MRI serving as a potential monitoring tool in trial or high-risk clinical settings.