Fatty Liver: What Works, What Fasting Can’t Prove Yet

The surprise is that Cochrane did not pick a winner

If you came here expecting a clean hierarchy for fatty liver treatment, the top-line answer is awkward: the highest-level review could not rank any lifestyle intervention on the outcomes patients care about most. In the 2021 Cochrane network meta-analysis by Buzzetti et al., 59 randomized trials with 3,631 people tested 33 lifestyle approaches over 1 to 24 months. The review set out to compare effects on death, cirrhosis, liver complications, liver cancer, and transplantation. It could not do it, because those events were essentially absent across the entire trial network.

That failure was structural, not a cute statistical quirk. If no one dies, progresses to cirrhosis, or needs a transplant during follow-up, a network meta-analysis has nothing solid to rank. So fasting did not lose to exercise. Mediterranean diet did not beat fasting. Everything hit the same wall: the trials were short, the endpoints were mostly surrogate markers, and the hard outcomes never showed up.

That matters because fatty liver, now often labeled MASLD, is a long game. The Cochrane review notes that detecting mortality differences versus the general population likely requires roughly 8 to 28 years of follow-up. The longest trial in the review lasted 24 months. That is not a small gap. That is a different sport.

What the evidence does show: many lifestyle interventions improve surrogate markers

The evidence is stronger for liver fat on imaging, liver enzymes, and insulin resistance than it is for cirrhosis or survival. Across the literature, several interventions move those markers in the right direction.

Exercise has the cleanest mechanistic support. A 2023 meta-analysis by Stine et al. found that exercise training reduced MRI-measured liver fat in people with NAFLD even without clinically significant weight loss. That is useful because it weakens the lazy assumption that exercise only helps the liver by making people smaller. A separate 2021 meta-analysis by Babu et al. reached a similar conclusion: exercise without weight loss or dietary change still improved NAFLD-related clinical parameters. And a 2022 review by Houttu et al. summarized evidence on aerobic exercise, NASH, and fibrosis while citing current guideline recommendations for 150 to 300 minutes per week of moderate-to-vigorous aerobic exercise.

Diet quality also matters. A 2024 meta-analysis by Xiong et al. found that both Mediterranean and low-fat diets improved liver enzymes and liver fat versus controls, with Mediterranean diet showing particular strength on liver fat content. More recently, a 2025 meta-analysis by Artola Arita et al. found that Mediterranean diet plus exercise outperformed either one alone for body composition and liver outcomes in MASLD. If you insist on a practical frontrunner at the surrogate level, the combined Mediterranean diet and exercise approach has the best case.

That is also why Cochrane suggested aerobic exercise plus dietary advice as a reference comparator for future long-term trials. Not because it won a head-to-head championship belt. It did not. The review could not award one. The point was that exercise and diet have the most coherent mechanistic and surrogate evidence to build on.

Where fasting fits

Fasting is not snake oil here. It has randomized evidence behind it. The issue is scale, duration, and endpoint choice.

A 2025 systematic review by Haddad et al. summarized 12 clinical trials published from 2019 to 2024 on time-restricted feeding, modified alternate-day calorie restriction, and Ramadan-style fasting in NAFLD or MASLD. Trial sizes ranged from 32 participants in a pediatric pilot to 226 in a multicenter Chinese randomized trial. Across those studies, fasting protocols generally improved steatosis, BMI, liver enzymes, and metabolic markers. That is a real signal. It is also a surrogate-level signal.

The strongest direct comparator in this set is the 2023 TREATY-FLD randomized clinical trial by Wei et al., which compared time-restricted eating with daily calorie restriction in people with NAFLD. Its importance is less about proving fasting superior and more about showing that time-restricted eating is at least credible enough to compare against a standard lifestyle strategy in a randomized design. But it still ran only up to 12 weeks and still used surrogate endpoints, not cirrhosis, cancer, transplant, or death.

The fasting literature has also become more specific. Deng et al. in 2024 compared early versus late time-restricted eating for intrahepatic fat reduction, testing the circadian idea that timing may matter, not just the length of the eating window. Tsitsou et al. in 2025 ran the 12-week CHRONO-NAFLD trial, combining a Mediterranean-style diet with time-restricted feeding in people with MASLD. That design makes biological sense: dietary quality plus meal timing is more plausible than pretending a 16:8 schedule can rescue an otherwise lousy diet. Oh et al. in 2025 added newer controlled data on the efficacy and safety of time-restricted eating in MASLD.

The point is not that fasting lacks evidence. The point is that its evidence lives in the same neighborhood as other lifestyle interventions: short trials, imaging endpoints, small or moderate samples, and no hard clinical outcomes.

Why fasting still does not outrank exercise and Mediterranean diet

First, the comparator problem. There is no fasting-versus-exercise trial and no fasting-versus-Mediterranean-diet trial in the evidence set summarized here. The question many people actually care about, "Should I fast instead of doing the boring stuff?", is not answered directly.

Second, the calorie-confound problem. In several fasting studies, reduced eating time may reduce calories whether or not the protocol tells participants to count them. That makes it hard to separate meal timing from plain old eating less. The 2023 crossover trial by Feehan et al. tried to isolate this by testing 16-hour daily time-restricted fasting without prescribed calorie restriction against standard care with diet and lifestyle advice in 32 people with NAFLD. Over 12 weeks, fasting reduced hepatic steatosis compared with standard care, with reported p values of 0.038 for steatosis, 0.005 for weight, 0.001 for waist circumference, and 0.005 for BMI. Useful result, but the crossover design carries carryover risk, and without rigorous dietary logging, incidental calorie reduction remains hard to rule out.

Third, durability. Most post-2021 fasting trials lasted 12 weeks or less. That can tell you whether liver fat budges. It cannot tell you whether the benefit persists. A 2025 cohort study by Luo et al. examined which metabolic impairments predict loss of sustained liver improvement in MASLD treatment, reinforcing a broader point: short-term imaging gains do not guarantee durable benefit.

Fourth, bias and generalizability. In the Cochrane review, 57 of 59 included trials were judged high risk of bias. That warning applies across the lifestyle field, not just to fasting. In fasting specifically, some studies involve restricted populations, including exclusively overweight or obese Chinese adults or small pediatric samples, and one trial in the review base is only a protocol paper. Haddad et al. also note heterogeneity by sex in one study, with women showing larger improvements than men. Pooling those studies can hide real differences.

What about supplements and drugs?

They matter as context because lifestyle does not own the surrogate-outcome space. A 2025 meta-analysis by Zhang et al. compared lifestyle modifications, silymarin combinations, Mediterranean hypocaloric diets, and medications for NAFLD surrogate outcomes. Silymarin combinations showed benefit, but they did not leap past well-structured lifestyle approaches on liver fat endpoints. Meanwhile, Lavynenko et al. in 2022 reported 6-year data in people with newly diagnosed type 2 diabetes showing less hepatic steatosis and fibrosis prevalence with initial triple therapy than with stepwise conventional therapy. That does not make drugs the universal answer for fatty liver, but it does kill the fantasy that fasting sits unchallenged at the top of every intervention list.

The practical verdict

If you have fatty liver and want the most evidence-backed lifestyle base, start with aerobic exercise plus a Mediterranean-style diet. That combination has the strongest surrogate-level support, and exercise has repeated evidence for reducing liver fat independent of major weight loss. Guidelines already recognize aerobic exercise; fasting has not received equivalent guideline placement in the evidence summarized here.

Time-restricted eating is still a defensible add-on. The randomized evidence suggests it can reduce liver fat and improve metabolic markers over short periods, especially when paired with better diet quality. It may also be easier for some people to stick with than deliberate daily calorie restriction. Adherence matters because the best protocol on paper is useless if it lasts three Wednesdays.

But the honest conclusion is plain: fasting has not beaten exercise or Mediterranean diet on clinical outcomes, and neither has anything else. The Cochrane review did not find a winner because the field has mostly studied short-term biomarkers, not the long arc of liver disease. Until longer registry-based or cohort-embedded trials run for years rather than weeks, the hierarchy remains provisional. Fasting can help. It just has not earned a crown.

Best-supported stack for now: 150 to 300 minutes per week of aerobic exercise, Mediterranean-style eating, and, if it improves adherence, a fasting schedule layered on top rather than used as a substitute.