Time-restricted eating has been sold as a liver intervention for years, and the strongest test of that claim arrived in JHEP Reports in 2026 (DOI 10.1016/j.jhepr.2026.101956). A 12-week multicenter randomized trial put 197 adults with overweight or obesity into four arms and asked a simple question: does when you eat move liver fat, beyond what the diet education already achieves?
The design was cleaner than most fasting studies. One arm got usual care, a Mediterranean diet-based education program. The other three added an 8-hour eating window on top of it: an early window, a late window, or a self-selected one. Hepatic fat fraction was measured by MRI, and the trial was registered as NCT05310721.
Inside every TRE group, liver fat went down. That part is real and consistent with the smaller trials before it. But the between-group comparisons, the ones that actually isolate the effect of the window, all landed flat: early TRE -0.4 percentage points versus usual care (p=0.95), late TRE -1.5 (p=0.15), self-selected -0.7 (p=0.77). The three fasting arms were indistinguishable from one another too.
The two variables that did separate winners from non-winners were baseline disease and the scale. Participants who already had MASLD lost more hepatic fat than those without it (mean difference -2.7 points, p<0.001), and participants who lost at least 5 percent of body weight lost more than those who did not (-2.6 points, p<0.001).
That weight-loss number deserves the emphasis it gets here, because it is the trial’s most useful finding for a reader with fatty liver. 41 to 44 percent of the TRE groups achieved at least 5 percent weight loss, versus 16 percent of usual care (p=0.001). Whatever the window does to liver fat directly, it is a genuinely effective scaffolding for the calorie deficit that does the work, the same lever the three-diet trial covered earlier this month identified as the real driver.
The microbiome results close the last escape hatch for timing enthusiasts. The trial sequenced fecal microbiota with 16S rRNA and found no between-group differences in composition, so the popular story that an early window feeds a better gut-liver axis did not survive a head-to-head test. The honest summary is that the 8-hour clock is a behavioral container, not a metabolic switch.
For a reader deciding whether to try it, the practical framing matters more than the physiology. If compressing meals into 8 hours is the device that stops late-night snacking, the trial says that device works: nearly half the fasting participants hit a weight-loss threshold that only one in six reached without it. If instead the window makes a person miserable or skipped meals chaotic, the data gives no reason to push through.
There is also a safety note that the internet fasting discourse tends to skip. Compressed eating windows change glucose patterns, and readers taking insulin or sulfonylureas can run low when meals bunch together. That is a conversation for the prescribing clinician, not a forum thread, and it is the same caution we attach to any intervention that changes meal structure this much.
The trial’s limits belong in the open. It ran 12 weeks, it was not powered for the secondary outcomes like elastography and microbiota, and the usual-care arm was not a passive placebo but a real Mediterranean-based education program, which raises the bar any add-on has to clear. A fasting arm beating a weak control proves little; beating a decent one is the test this design set.
How does this fit the rest of the 2026 evidence? The pattern is remarkably consistent. The three-diet trial covered on this site earlier this month found diet labels mattered less than weight loss, and the JHEP Reports pilot that used a very low energy diet showed what an aggressive deficit can do to liver fat. Add the new trial and the lesson sharpens: the liver responds to the energy balance, not to the clock face it happens to run on.
The food side of the equation is where this site keeps its emphasis. An 8-hour window that opens onto ultra-processed snacks and closes on sweetened drinks will not do what the trial’s Mediterranean-educated participants did. The window is a container; the plate inside it is the content, and the two fail or succeed together. A window that opens onto a half-cup scoop of vanilla frozen yogurt and closes on an actual dinner is a different experiment from one that opens on a gas-station pastry.
For readers who like numbers to plan with, the trial used a straightforward definition: 8 hours of eating or less, at least 16 hours of fasting, logged with timestamped photos of the first and last intake of the day. That logging habit is worth copying regardless of the schedule a reader picks, because it converts a vague intention into a countable record.
What would falsify this read? A longer trial in a MASLD-only population, powered for liver outcomes, could still find a timing effect that 12 weeks in a mixed group missed. The CHRONOSTEATOSIS trial now recruiting in France is exactly that study, an 8-hour window without calorie restriction in patients with MRI-PDFF confirmed steatosis. Until it reads out, the honest position is that timing is unproven, not disproven.
For the reader who has already been fasting and wonders whether to switch from a 10 a.m. to a noon window, the trial’s answer is refreshingly boring: pick the one you can keep. Early, late, and self-selected windows produced the same weight loss and the same liver results, and adherence is the only lever the evidence actually supports moving.
The closing line belongs to proportion. Nobody should read this as an attack on time-restricted eating, which helped most of its participants in this trial lose weight, the one change the liver literature rewards most reliably. It is a demotion from magic to tool, and tools are chosen by fit, not by folklore.
Is an 8-hour eating window good for fatty liver?
The 2026 trial found the window itself is not the active ingredient. All three timing groups lost liver fat, but no group beat usual care on head-to-head comparison. What did predict liver fat reduction was losing at least 5 percent of body weight, which the TRE groups achieved far more often (41-44 percent versus 16 percent).
Did the timing of the eating window matter at all?
No. Early TRE differed from usual care by 0.4 percentage points of hepatic fat (p=0.95), late TRE by 1.5 points (p=0.15), and self-selected TRE by 0.7 points (p=0.77). The three TRE arms were also indistinguishable from each other.
Should someone with MASLD try time-restricted eating?
As a weight-loss tool it worked in this trial: TRE participants reached clinically meaningful weight loss at nearly three times the rate of usual care. As a liver-specific trick it did not outperform the diet it was added to. Anyone on diabetes medication should talk to a clinician first, because compressed eating windows change glucose patterns.