Diet quality alone did not lower inflammation — the weight loss from fasting did
Published Oct 21, 2026
Methods
Fifty-three adults with metabolic risk factors (77% women, average age 54.7) were randomly assigned to receive either a fully provided Mediterranean diet or their habitual western diet for 8 weeks. Both groups then added two fasting days a week for a further 8 weeks. All food was supplied by the researchers. The pre-specified primary outcome was high-sensitivity C-reactive protein, a standard blood marker of inflammation.
Findings
C-reactive protein did not budge during the diet-only phase, even though blood measurements confirmed the two groups really were eating differently. Once the fasting days began, C-reactive protein fell in the Mediterranean-diet group relative to controls — alongside greater weight loss, better insulin sensitivity, and lower blood pressure. Blood lipids were the exception: those had already improved during the diet-only phase. Multi-omic analysis showed matching reductions in inflammatory signalling, while gut microbiome composition stayed largely stable.\n\nThe authors' conclusion is that diet quality on its own did not lower inflammation over 8 weeks, and that negative energy balance and weight loss were the stronger drivers.
Caveats & Context
This is 53 people over 16 weeks, mostly women, all with existing metabolic risk factors — small, and not a general-population sample. Because every participant did the diet phase first and the fasting phase second, the trial cannot fully separate the effect of fasting from the effect of simply being eight weeks further into the study. All food was provided, so adherence was far better than anyone achieves unaided. And this is not a finding that the Mediterranean diet does nothing: lipids improved in the diet-only phase, and 8 weeks may simply be too short to move C-reactive protein.
Weight loss during intermittent fasting, not Mediterranean diet quality alone, is associated with lower inflammation: A randomized controlled trial
Cell Metabolism · doi.org/10.1016/j.cmet.2026.09.017