Keto vs. Mediterranean vs. Low-Fat: What Happens When You Control For Weight Loss?
Updated: Oct 7
# Understanding Diets and Metabolic Health: Insights from Recent Research
By JoBeth Augustyniak, DO
Board-Certified Family Medicine (AOBFP)
Published September 25, 2026 · Medically reviewed by JoBeth Augustyniak, DO on September 21, 2026
About the author →
Every diet debate eventually encounters a common issue: if one group loses more weight than another, it’s hard to determine whether the diet itself played a role or if the weight loss was the main factor. Most nutrition research carries this flaw.
A new trial from Washington University, published in Cell Metabolism in August 2026, cleverly addressed this issue. The findings provide some of the clearest evidence yet on a question I often hear in my practice: does the composition of your diet matter, or is weight loss just weight loss?
Here’s what the researchers did, what they found, and — importantly — where we should exercise caution before making broad conclusions.
The Setup: Three Diets, One Identical Amount of Weight Loss
Researchers studied 55 adults with obesity, prediabetes, and fatty liver disease (confirmed by MRI). They randomized participants into one of three fully food-provided diets:
A very-low-carbohydrate ketogenic diet (about 4% of calories from carbs, 73% from fat)
A Mediterranean diet (50% carbs, 35% fat)
A very-low-fat, plant-forward diet (70% carbs, 15% fat)
Every meal was prepared and delivered by the study team. Participants received the food itself, tracked their intake via an app, and met weekly with a dietitian. Adherence was over 95% in every group.
The critical design choice was that researchers adjusted each participant's calorie intake so that every group lost nearly the same amount of weight — about 10% of body weight — over four to five months, followed by a few weeks of weight stabilization before final testing. Forty-two individuals completed the full protocol.
In other words, the same weight loss occurred through three completely different eating patterns. Any remaining differences can't be attributed to one group simply losing more weight than another.
What They Found: Your Muscles and Your Liver Respond Differently
Here’s the most intriguing result. Researchers measured insulin sensitivity in two areas: skeletal muscle and the liver, using gold-standard clamp testing.
In muscle, all three diets performed identically. Insulin sensitivity improved by about 50% across the board, with no significant difference between the ketogenic, Mediterranean, or low-fat groups. Weight loss itself seems to be the driving force here — the composition of the diet didn’t matter for how your muscles respond to insulin.
In the liver, the story was entirely different. Hepatic insulin sensitivity improved two to three times more on the ketogenic diet than on either of the higher-carbohydrate diets. Liver fat showed a similar trend: it dropped from 19.4% to 6.2% on keto (a 67% reduction) — right at the edge of the threshold used to define fatty liver disease — compared to roughly 45% reductions on the other two diets.
The glucose and insulin numbers reflected this trend. HbA1c fell 0.6 points on the ketogenic diet versus 0.1 and 0.4 points on the other two. Fasting insulin dropped nearly 15 points on keto, while it decreased by about 10 points on the others. Twenty-four-hour insulin exposure fell 74% on keto, compared to 44% and 27% on the comparison diets.
On lipids, the picture was reassuring — with one important caveat. Fasting triglycerides fell most significantly on the ketogenic diet, and markers of small, dense, artery-unfriendly VLDL particles improved the most as well. Importantly, LDL cholesterol and apoB — the particle count we monitor closely for cardiovascular risk — did not show meaningful differences between any of the three diets. This is a notable finding for a diet that derives nearly three-quarters of its calories from fat.
However, here's the caveat worth considering: while fasting triglycerides looked best on keto, triglyceride exposure measured over a full 24-hour period was essentially the same across all three diets. Why? Because the high-fat ketogenic meals temporarily raised triglycerides after eating. A single fasting lab number, in other words, doesn’t provide the complete picture of what’s happening in your bloodstream throughout the day.
The Honest Limitations
It’s crucial to discuss what this study cannot prove.
It's a small trial. Only 14 people completed each diet arm — enough to detect the two effects the researchers specifically aimed to measure (the muscle and liver insulin sensitivity difference), but not enough to be fully confident about many secondary findings.
It's short. Four to five months of active dieting, followed by a few weeks of retesting. We don’t know what happens to these numbers after a year or after weight regain, which research consistently shows is common.
It's not just a carbohydrate study — it's three complete dietary patterns. Protein, saturated fat, and food sources all varied between groups. This tells us how three specific diets compare, not that carbohydrate percentage alone caused the differences.
The population was narrow. Participants were aged 18 to 55, with obesity, prediabetes, and fatty liver. Those taking diabetes medications, and notably for many of my patients, postmenopausal women, were excluded from this trial entirely. We can’t assume these results apply directly to a 58-year-old managing perimenopausal metabolic changes.
This was a highly supported, food-provided study. Every meal was prepared for participants, and a dietitian checked in weekly. This reality is very different from planning and cooking a ketogenic diet on your own in daily life.
No long-term outcomes were measured. There were no heart attacks, liver biopsies, or five-year follow-ups — just physiology and lab markers over a few months.
What This Means for Your Health Plan
I don’t interpret this study as "keto wins, case closed." Instead, I see it as valuable evidence that helps us match the right dietary strategy to the right metabolic phenotype.
If your labs resemble those of the participants in this study — insulin resistance, fatty liver, high triglycerides, and that stubborn hyperinsulinemia that doesn’t budge — this trial supports trying carbohydrate restriction as a targeted tool. Importantly, it didn’t come at a significant cost to LDL or apoB.
If your metabolic profile differs from that, this study reminds us that a well-constructed Mediterranean or plant-forward approach produced nearly identical weight loss, similar improvements in muscle insulin sensitivity, and substantial liver-fat reduction too. The "best" diet remains the one that aligns with your specific biology, preferences, and lifestyle — not a one-size-fits-all solution.
Practical Takeaways for Your Health Journey
Here are a couple of practical takeaways I’m incorporating into visits:
We’ll continue to track apoB and LDL particle data individually rather than assuming any diet is automatically "safe" for your lipid panel. This trial’s average results were reassuring, but averages don’t reflect individual experiences.
If liver health is a concern, ALT and AST on a basic panel may not capture what’s truly happening. This study’s liver-fat changes were measured with MRI, which goes beyond what routine bloodwork reveals.
All three diets in this study caused some loss of lean muscle mass alongside fat loss. This serves as a reminder that adequate protein intake and resistance training deserve attention, regardless of the eating pattern you choose.
If you’re navigating decisions about carbohydrate restriction, liver health, or lipid management, this is the kind of nuanced conversation worth having at your next visit. Bring your labs, and we’ll discuss what the evidence supports for you specifically.
This article is for educational purposes and reflects a review of peer-reviewed literature current as of 2026. It is not a substitute for individualized medical advice. Talk with your physician before making significant changes to your diet, especially if you have diabetes, kidney disease, liver disease, or are taking medications that affect blood sugar or blood pressure.
Comments