Early, small, conflicting, or animal-only. Not actionable alone.
Emulsifiers such as CMC and carrageenan inflamed the guts of mice. In short human trials they changed gut bacteria and, in two trials, how leaky the gut was, but none of the four trials in healthy people found more inflammation. In Crohn's disease, two trials split.
Cocoa took about 1.8 mmHg off blood pressure in trials of 2 to 18 weeks, mostly of specially made drinks, capsules and bars. The largest heart trial, 21,442 older adults on cocoa capsules for 3.6 years, did not clearly cut heart attacks, strokes and other heart events, and the US regulator says its cocoa claim does not apply to chocolate. The EU allows a narrower claim, about blood vessels, only on dark chocolate that delivers 200 mg of cocoa flavanols a day.
A neck-worn stimulator sold on a health channel quotes a 56% drop in stress symptoms. In the seller-funded study behind it, 56% was the fall on a depression questionnaire, and nobody got a sham. The sham-controlled trials of neck stimulation are small and mixed. Cold water on the face and slow breathing do shift the heart's rhythm, while you do them.
In trials of a few weeks to a few months, turmeric extract capsules eased knee arthritis pain a little more than a dummy did, and about as much as a modest dose of ibuprofen. The big pooled numbers rest on small trials, most of them paid for by the extract's maker, and the claims for blood sugar, fatty liver and PCOS rest on small, split evidence.
The blood-sugar study behind a clove in your coffee gave 13 people a capsule of a clove extract its maker calls de-flavoured, with no comparison group, and the maker paid for it. Nobody has tested a clove in coffee, and the stomach claims rest on dishes and animals.
In small trials lasting weeks, hibiscus, as tea or capsules, took about 6 points off the top blood-pressure number compared with a dummy or no treatment, more in people whose pressure started higher. Against prescription pills the trials split: in nine head-to-head trials of hibiscus products, the drug did more in three, hibiscus did more in one, four against captopril found no clear difference, and the ninth favoured hibiscus only after treatment stopped. The strictest review found too few good trials to say whether it works in high blood pressure.
Six small one-year trials found aged garlic extract slowed a number on a heart scan; five came from one Los Angeles lab, and at least five had money, capsules or staff from the extract's maker. Total plaque did not shrink in either trial that measured it, and the one long trial that counted heart deaths found no fewer.
Eleven randomised trials of Akkermansia supplements for weight or blood sugar are small and split: the largest, 142 adults, missed its main goal, and the clearest result was 4.4 lb less weight regained after a diet. Four reported GLP-1, and they disagree. The cancer link is a stool finding, and the one such trial to report results had nine patients and no comparison group.
In trials of adults with constipation, two or three kiwifruit a day added roughly one to three complete bowel movements a week to where people started, about what a fiber supplement did, over four weeks. The one randomised trial against a dummy found no clear difference across everyone in it. No trial of the fruit we found reported the first 24 hours, and a kiwifruit marketer paid toward every gut study of the whole fruit on this page.
Cooling cooked rice, pasta or potatoes turns a little of the starch resistant to digestion. Of the 28 single-meal trials we found and could read in people without diabetes, 15 found a clearly smaller blood-sugar rise, 11 found no clear difference and 2 reported no test result for it. The peak often barely moved, and no trial of chilled food alone that measured blood sugar lasted longer than a day. Resistant-starch supplements are a separate, weaker story.