Keto for Fatty Liver: Why Your PNPLA3 Gene May Matter

A ketogenic diet may improve insulin sensitivity and triglycerides quickly, but that does not establish it as a treatment for metabolic dysfunction-associated steatotic liver disease (MASLD), formerly called NAFLD. The best evidence still favors sustained weight loss and Mediterranean-style eating for reducing liver fat. PNPLA3 genetics add a more personal complication: people with the high-risk GG version of rs738409 may process liver fat differently, yet no trial has tested a ketogenic diet specifically in that group.

The sensible conclusion is narrower than the internet version. Keto can be a workable way for some people with MASLD to reduce calories and improve metabolic markers. It has not earned preferred status as a liver-specific treatment, and it is a poor candidate for a universal prescription.

What the ketogenic studies actually show

Two 2024 studies found less liver fat after eight weeks of a very-low-calorie ketogenic diet (VLCKD), but both tested an extreme weight-loss protocol rather than ketosis in isolation.

In a 2024 observational study by Sila et al., 111 adults with overweight or obesity and MASLD completed an eight-week VLCKD providing 650 to 800 kcal per day. Transient elastography measured steatosis. Participants with more severe steatosis at baseline had the largest reductions. That finding may partly reflect more room to improve, but the study had no randomized control group. A diet at 650 to 800 kcal per day produces a major calorie deficit, so the result cannot tell us how much came from carbohydrate restriction, ketosis, weight loss, regression to the mean, or all four.

Rinaldi et al. reported a similar eight-week VLCKD intervention in people with MASLD and found reduced steatosis in both sexes. Men had significantly greater reductions in steatosis and liver stiffness than women. The sex difference is interesting, but it does not prove that men oxidize fat more efficiently or that women respond poorly to keto. The study did not include a control diet or serial serum ketone monitoring, so it cannot separate biological differences from differences in adherence, weight change, body composition, or other factors.

Ketogenic diets can have effects that are not simply weight loss. In a 10-day randomized trial of 29 adults with overweight or obesity and type 2 diabetes, Merovci et al. compared a weight-maintaining standard diet, a weight-maintaining ketogenic diet, and a ketogenic diet plus beta-hydroxybutyrate ketone ester at 8 g every eight hours. Body weight and body fat remained constant, while the ketogenic diet improved glycemic control and insulin sensitivity. That is useful evidence for carbohydrate restriction as a metabolic intervention. It did not measure liver fat, fibrosis, or liver stiffness, so it cannot answer whether keto improves MASLD itself.

Why calorie deficit remains the main explanation

A systematic review and meta-analysis by Houttu et al. concluded that weight loss and caloric restriction drive most dietary reductions in liver fat. Mediterranean, low-fat, and lower-carbohydrate approaches can all help when they produce sustained energy deficit. Across lifestyle research, roughly 5% to 10% weight loss is associated with meaningful hepatic improvement, but this is a population-level gradient, not a cliff where a 4.9% loss fails and a 5.0% loss succeeds.

Duration and adherence also matter. In a six-month randomized trial of 128 people with NAFLD and metabolic syndrome, Abbate et al. compared three lifestyle programs: a conventional diet, a Mediterranean diet with higher meal frequency, and a Mediterranean diet plus physical activity. Each aimed to reduce caloric intake by 25% to 30% and increase energy expenditure by 400 kcal per 70 kg. The interventions reduced liver fat and improved metabolic risk factors; the study also examined kidney-risk markers. It supports a familiar but durable lesson: the diet must work long enough to change energy balance.

Short-term improvement also may not last. A prospective cohort study by Luo et al. followed treatment-naive MASLD patients through weeks 0, 24, and 48 using MRI-PDFF and liver stiffness assessment. It defined a liver-fat response as at least a 30% relative decline in MRI-PDFF without worsening fibrosis, then identified metabolic impairment patterns linked to loss of sustained improvement. A promising eight-week scan is therefore a progress report, not a guarantee.

PNPLA3 changes the question

The PNPLA3 I148M variant, rs738409, is a major common genetic contributor to MASLD susceptibility and severity. Its GG genotype does not dictate anyone's outcome, but it may magnify the consequences of diet quality.

In a 2024 population-based cohort study by Chen et al., polygenic hepatic risk that included PNPLA3 amplified the association between diet quality and steatosis, inflammation, and fibrosis. Higher-quality diets tracked with greater hepatic protection among people at higher genetic risk, while lower-quality diets tracked with worse outcomes. This was an observational gene-diet interaction, not a ketogenic-diet trial. It supports taking diet quality seriously in high-risk people; it does not identify a specific carbohydrate target for them.

A 2025 cross-sectional metabolomics study offers a reason for caution about aggressive ketogenic diets in PNPLA3 GG carriers. Jegodzinski et al. profiled 353 people with MASLD; the fasting GG subgroup included about 49 people. Compared with other genotypes, GG homozygotes had worse baseline liver-injury markers and higher fasting ketone bodies and tricarboxylic-acid-cycle metabolites, along with lower lysine and higher intermediate-density lipoprotein particle concentrations. The pattern is consistent with altered fasting hepatic fat oxidation.

Consistent is doing a lot of work there. The study did not assign anyone to keto, measure liver outcomes after carbohydrate restriction, or prove that ketogenesis causes hepatocellular stress in GG carriers. A fasting metabolite pattern is a clue, not a dietary contraindication. Still, adding a very-low-carbohydrate diet to a genotype already associated with elevated fasting ketones deserves more caution than a blanket "keto is good for fatty liver" claim.

Mediterranean eating has the stronger liver evidence

Among named dietary patterns, Mediterranean-style eating has the more mature evidence base for MASLD and NAFLD. A 2024 meta-analysis of randomized trials found that both Mediterranean and low-fat diets reduced ALT, AST, and liver fat in NAFLD, with greater liver-fat reduction in the Mediterranean-diet group.

The MEDINA randomized trial followed people with NAFLD for 12 weeks and reported hepatic and metabolic improvements, including MRI-measured liver-fat reduction, with a Mediterranean diet. A separate 12-week randomized trial, the CHRONO-NAFLD study, reported hepatic and metabolic improvements with Mediterranean-type time-restricted feeding in MASLD. These trials do not prove that one eating pattern fits every patient. They do give Mediterranean approaches a better controlled evidence base for liver endpoints than keto currently has.

Broader diet quality also tracks with fibrosis risk. In cross-sectional analyses of 2,532 Framingham Heart Study participants and 3,295 NHANES participants, Gao et al. found that higher DASH, Alternative Healthy Eating Index, and Mediterranean-style diet scores were associated with lower liver stiffness after demographic and lifestyle adjustment. Further adjustment for liver fat or BMI weakened those associations, which fits the possibility that body weight and steatosis explain part of the relationship.

What to do with this evidence

If you have MASLD, aim for a dietary pattern you can sustain that creates an appropriate calorie deficit when weight loss is indicated. Mediterranean-style eating has the strongest controlled evidence for liver-fat improvement. Time-restricted eating may be useful when it helps adherence, not because a clock can outvote total intake.

Keto is reasonable as one option when it improves adherence, glycemic control, or triglycerides and is medically supervised when needed. Do not assume its early metabolic effects equal fibrosis regression or durable liver recovery. Exogenous ketone salts and esters have no clinical-trial evidence for MASLD liver endpoints, so they do not belong in a liver-treatment plan on the basis of ketogenic-diet studies.

If you know you carry PNPLA3 rs738409 GG, favor high overall diet quality over an aggressive attempt to drive ketosis until genotype-stratified trials answer the question. Discuss genetic results and liver monitoring with a clinician, especially if you have elevated liver enzymes, diabetes, advanced fibrosis risk, or are considering a very-low-calorie diet. The available biology justifies caution; it does not justify certainty.

For fatty liver, the most defensible target is sustained improvement in weight, diet quality, metabolic health, and measured liver risk—not a particular ketone number.