ABSTRACT
Background: Growing interest in very low carbohydrate diets, and in particular the ketogenic diet, has been met with some resistance. Important gaps exist regarding what diet to compare to the ketogenic diet. The objective of this study was to compare a Well Formulated Ketogenic Diet (WFKD) with a Mediterranean-Plus diet (Med-Plus;Mediterranean with emphasis on eliminating added sugars and refined grains), in a crossover study, stratified by diabetes status (T2D vs Prediabetes). Methods: The intervention involved having participants follow the WFKD and Med-Plus, for 12 weeks each, in random order. All meals were provided for the first 4 weeks of each diet phase (food delivery);then participants were responsible for purchasing and preparing their own foods (self-provided). The primary outcome was glycosylated hemoglobin (HbA1c). Secondary outcomes included weight, glucose as measured by continuous glucose monitor (CGM), and cardiometabolic risk factors, such as fasting insulin, glucose, and lipids. Results: Among participants randomized (n = 42), 33 had complete data at both diet phases (some missing data attributable to COVID disruptions). Participants were 60 ± 9 (mean ± sd) years of age, 61% men, with BMI 31 ± 5 kg/m2. Adherence for both diets was higher during the food delivery than the self-provided phase, but similar between diets for both phases. HbA1c concentrations were not significantly different between diets, but average CGM glucose levels were significantly lower during the WFKD compared to Med-Plus (p = 0.03). Additionally, WFKD induced a significantly greater decrease in triglycerides (-16% vs -5%, p = 0.02) and greater increase in LDL-C levels (10% vs -5%, p = 0.01), compared to Med-Plus. Weight change on WFKD vs Med-Plus was -8% vs -7% (p = 0.05). Sensitivity analyses largely confirmed the main findings. Conclusions: Participants improved in glucose control and weight management on both diets relative to baseline;however, glucose control was superior on the WFKD. Some caution is warranted when interpreting these results due to pandemic disruptions and a small sample size. A fair comparison of the two diets should also take into consideration non-glycemic effects.