Metabolic Syndrome: What Beats Drugs, and What Adds to Them

The ranking looks dramatic. The interpretation needs brakes.

A 2026 meta-analysis by Werndle et al. tried something useful and unusual: it ranked diet-only, exercise-only, combined lifestyle, and pharmacotherapy on one continuous measure of metabolic syndrome severity, the MetS z-score. Across 46 studies, 85 intervention arms, and 12,128 participants with overweight or obesity, diet-only produced a pooled MetS-Z change of -0.72, exercise-only about -0.63, combined diet plus exercise -0.68, and pharmacotherapy just -0.07 versus controls, with the drug result reported as non-significant (p=0.134).

If you stop there, the story writes itself: food and exercise crush drugs, and drugs barely move the needle. That is the clean headline. It is also too clean to trust.

The better reading is narrower. Lifestyle treatment has strong evidence as first-line care for metabolic syndrome. The pooled pharmacotherapy result in this analysis does not prove that modern obesity drugs barely help metabolic syndrome. It mostly shows what happens when you average unlike things, use different populations, and ask a small subgroup of drug trials to answer a question they were not designed to answer.

What the meta-analysis actually found

MetS-Z is a composite score. Instead of asking whether someone still meets the yes-or-no definition of metabolic syndrome, it tracks movement across the syndrome's component risks as a continuous number. That matters because small improvements across waist circumference, lipids, blood pressure, and glucose can disappear inside a binary diagnosis.

The appeal of Werndle et al. is that it puts different intervention categories onto the same scoreboard. Diet-only had the largest pooled pre-post reduction at -0.72. Exercise-only was close behind at about -0.63. Combined diet plus exercise came in at -0.68. Pharmacotherapy, pooled across just five drug arms, landed at -0.07 versus control and missed conventional statistical significance.

Those MetS-Z shifts are not trivial bookkeeping. In a 2023 randomized trial by Sandsdal et al., the authors cite prior risk data showing that people in the highest MetS-Z quartile had a hazard ratio of 5.1 for coronary heart disease over more than 11 years and 17.4 for future type 2 diabetes over eight years, compared with the lowest quartile. A z-score is still a surrogate outcome, not a heart attack prevented. But it is a serious surrogate.

The catch is that Werndle et al. pooled within-group pre-post changes across different studies and different populations. That design can show that each category tends to help. It cannot cleanly prove that diet beats exercise, or that combined lifestyle is truly worse than diet-only because -0.68 is numerically smaller than -0.72. Different baseline severity, age, intervention intensity, and co-interventions can change the size of the pre-post shift. Regression to the mean can help too. A ranking built this way is informative, but it is not a podium ceremony.

Why the drug result probably underestimates modern drugs

The pharmacotherapy bucket in Werndle et al. pooled mechanistically mixed agents: metformin, DPP-4 inhibitors, and GLP-1 receptor agonists. That is a statistical convenience, not a biological class. Metformin and DPP-4 inhibitors are not expected to remodel body weight and cardiometabolic risk the way GLP-1 drugs do. Average them together, then keep the subgroup small, and the modern signal gets flattened.

The analysis also ran into a design problem common to obesity-drug trials: major studies usually require lifestyle advice or lifestyle modification in both arms. That is not sloppiness. It is standard practice. But it means the published evidence often does not isolate a pure drug-only MetS-Z effect. The pooled estimate of -0.07 is better understood as the effect of mixed pharmacotherapy evidence in the context of concurrent lifestyle care, not a definitive reading on what contemporary GLP-1 or dual agonist therapy can do by itself.

Other drug evidence makes that point harder to ignore. In a 2024 post hoc SURPASS analysis by Nicholls et al., tirzepatide reduced the proportion of patients with type 2 diabetes who met metabolic syndrome criteria by 16% to 51%, depending on dose. That does not directly overturn the Werndle result. SURPASS used a binary metabolic syndrome definition, not continuous MetS-Z, and it studied people with type 2 diabetes, a population Werndle excluded. Still, it does refute the lazy version of the story that pharmacotherapy is broadly metabolically weak.

A 2025 meta-analysis in Nature Medicine reached a similar bottom line from a different angle: obesity drugs produce robust weight loss and improve obesity-related complications, but the translation from strong single-endpoint effects to a pooled composite metric can vary. Composite outcomes are useful. They are not magic. Sometimes they blur the very effect people care about.

When drugs are tested the right way, adding one helps

The cleanest piece of evidence here is not the pooled ranking. It is the direct randomized comparison. In the 2023 trial by Sandsdal et al., adults with overweight or obesity and metabolic syndrome were randomized to liraglutide plus exercise or exercise alone. The combination reduced metabolic syndrome severity, abdominal obesity, and inflammatory markers more than exercise alone.

That does not mean liraglutide replaces exercise. It means the add-on question was tested directly, and the answer was yes. For patients who can access and tolerate a GLP-1 receptor agonist, the best current evidence supports combining it with structured exercise rather than treating medication as a separate lane.

The practical conclusion is simple: the pooled drug null is not a verdict on GLP-1 therapy. The head-to-head randomized trial says a GLP-1 drug can add to exercise.

Exercise is not one thing

The lifestyle category in the ranking looks tidy. Real exercise prescriptions are less tidy, and the details matter.

In a 2021 randomized trial by Reljic et al., obese patients with metabolic syndrome undergoing caloric restriction improved more with resistance training than with whole-body electromyostimulation. Passive electrical stimulation did not reproduce the cardiometabolic benefit of lifting against resistance. Muscles appear to prefer actual work. Unfair, but true.

Intensity matters too. Another 2021 trial from Reljic et al. found that very low-volume high-intensity interval training, about 10 minutes per session three times per week, produced clinically meaningful improvements in obese patients with metabolic syndrome and performed at least as well as moderate-intensity continuous training despite a far smaller time commitment. A 2024 meta-analysis by Poon et al. backed that up, finding HIIT improved several metabolic syndrome components, including waist circumference, triglycerides, and HDL cholesterol, and was superior or equivalent to moderate continuous training on those endpoints.

That time-efficiency argument got another boost from a 2022 trial by Reljic et al., where very low-volume interval training improved broader cardiometabolic markers and the NAFLD fibrosis score in adults with obesity and metabolic syndrome. Liver disease is common in this population and sits outside the five components included in MetS-Z. So the exercise benefit may be broader than the composite score captures.

The microbiome angle is interesting, not settled

A 2025 pooled analysis by Reljic et al. found that exercise increased fecal short-chain fatty acids in an intensity-dependent manner. Higher-intensity protocols produced larger SCFA increases. That is plausible as part of the exercise story, because SCFAs are often discussed as mediators of insulin sensitivity, adipose metabolism, and inflammation.

Still, the cited analysis shows an exercise-SCFA association. It does not prove that SCFA changes caused the improvement in metabolic syndrome severity, and it does not compare exercise with diet for SCFA production. The mechanism may matter. The clinical result matters more, and the mechanism here remains partly speculative.

Why diet still ranks first, with an asterisk

The diet-only estimate of -0.72 is the largest pooled effect in the 2026 ranking, and it fits clinical common sense. Caloric restriction can rapidly shift waist circumference, triglycerides, glucose, and blood pressure. But the number should not be read as one universal diet score.

A 2022 meta-analysis by Jahrami et al. found that Ramadan fasting improved metabolic syndrome severity, but the size of the effect varied by ethnicity and sex. That suggests the pooled diet category probably hides substantial variation across dietary patterns and populations. Some diets likely work better than others. The ranking tells you diet matters. It does not tell you that every diet deserves the same credit.

Population also changes the hierarchy. A 2024 network meta-analysis in reproductive-aged women with overweight or obesity and a 2025 review in PCOS both suggest that the relative payoff from diet, exercise, and pharmacologic strategies can shift when hormonal comorbidity changes the metabolic terrain. A general-population ranking is useful. It is not universal law.

What this means if you actually have metabolic syndrome

The evidence supports a hierarchy, but not the cartoon version of one.

• Start with structured dietary restriction. It had the largest pooled MetS-Z reduction in the 2026 meta-analysis.
• Add structured exercise, with a bias toward resistance training or HIIT rather than assuming any movement dose is interchangeable.
• If you are eligible for a GLP-1 receptor agonist or a dual agonist, the best current evidence supports using it as an adjunct to lifestyle, not as a substitute for it.
• Ignore the idea that the pooled -0.07 pharmacotherapy result settles the value of modern obesity drugs for metabolic syndrome. It does not.

The strongest conservative conclusion is this: lifestyle treatment has the best direct evidence base for lowering metabolic syndrome severity, and GLP-1-based pharmacotherapy appears to add benefit when tested alongside structured exercise. The weak-looking pooled drug number is real as a published estimate. It is not reliable as a final answer.

That is less catchy than "diet beats drugs by tenfold." It is also much closer to the science.