What evidence is there that dietary changes can help inflammation

The honest answer is that dietary changes can lower inflammatory markers in humans, but the evidence that this translates into fewer clinical events is thin, and the best-studied diet is the Mediterranean pattern.

The strongest human evidence comes from a prospective cohort of 1,219 older adults in northern Manhattan followed for a mean of 3.8 years [1]. Those with the highest Mediterranean diet adherence had a hazard ratio for Alzheimer's disease of 0.66 (95% CI 0.41–1.04) compared to the lowest tertile, with a significant trend across tertiles (p=0.04) [1]. The same study measured inflammatory biomarkers: median hsCRP fell from 6.7 mg/L in the low-adherence group to 4.6 mg/L in the high-adherence group [1]. That is a real, dose-responsive drop in a systemic inflammatory marker.

But here is the catch. The protective effect against Alzheimer's was not mediated by those inflammatory biomarkers [1]. Adjusting for hsCRP, insulin, and adiponectin barely changed the hazard ratios [1]. That means the diet's benefit, if real, works through some other pathway, or the biomarkers are not capturing the relevant inflammation.

What about clinical outcomes beyond dementia? The evidence base is dominated by observational studies, not randomized trials. No retrieved record shows a randomized controlled trial where a dietary change reduced hard endpoints like heart attacks or death through an inflammation mechanism. The Mediterranean diet literature is mostly cohorts and surrogate markers. That is not the same as proof.

The mechanism is plausible. Diet changes gut flora, alters fatty acid composition, and shifts cytokine profiles. But plausible mechanism plus observational association is not the same as demonstrated clinical benefit. The burden of proof for "diet cures inflammation" has not been met with randomized data on hard outcomes.

My confidence: moderate that Mediterranean-style diets lower inflammatory markers like hsCRP; low that this translates into fewer clinical events, because the one study that tested mediation found the biomarkers did not explain the benefit [1]. The evidence is real but incomplete. Nobody has run the definitive trial.