A 2026 systematic review and meta-analysis published in Nutrients set out to answer a question that pure weight-loss numbers often obscure: when people use GLP-1 receptor agonist and incretin-based therapies alongside structured diet and exercise programs, what actually changes in the body, and by how much? The researchers, led by Bruna-Mejias and colleagues, pulled together data from 35 independent trials to produce one of the more comprehensive looks at this combination to date.
The distinction the paper draws early on matters: total body weight is only one metric. Changes in fat mass, lean muscle mass, physical function, and cardiometabolic markers can each tell a different story, and each may shift differently depending on what kind of lifestyle program accompanies the peptide therapy. That framing shapes the entire analysis.
The review followed a pre-registered protocol, searched six major academic databases, and applied standard risk-of-bias tools to each included study. The result is a layered picture of both the promise and the remaining uncertainty in this area of research.
How incretin-based peptides work
Glucagon-like peptide-1, or GLP-1, is a hormone the gut releases after eating. It signals the pancreas to release insulin, tells the brain that food has arrived, and slows the pace at which the stomach empties. Peptides that mimic or amplify this signal are called GLP-1 receptor agonists, and they belong to a broader family called incretins, which includes related hormones with overlapping roles in appetite and blood sugar regulation.
Because these peptides act on appetite signaling rather than simply blocking nutrient absorption, researchers have been interested in whether pairing them with behavioral interventions, meaning changes to eating patterns and physical activity, produces effects that neither approach alone would generate. The systematic review was designed specifically to test that idea in a rigorous, pooled way.
How the review was conducted
The research team searched PubMed, Web of Science, Scopus, CINAHL, SPORTDiscus, and the Cochrane Central Register from database inception through the final search dates. After removing duplicates and ineligible record types, 1,326 records were screened. Seventy-two reports were read in full, and 48 reports representing 35 independent trials or trial clusters were ultimately included in the qualitative summary.
Risk of bias was assessed using the Cochrane RoB 2 tool, a widely used framework that grades the methodological quality of randomized trials across several domains. The statistical approach used random-effects meta-analysis with a restricted maximum likelihood estimator and a Hartung-Knapp adjustment, which tends to produce more conservative confidence intervals than simpler pooling methods. That conservative choice is worth noting because it means the reported intervals are, if anything, wider than some older analyses might have shown.
The team also made a deliberate methodological decision to analyze absolute weight change in kilograms and percentage body-weight change as separate outcomes rather than combining them. The reason is straightforward: a 10-kilogram loss means something very different for a person starting at 90 kilograms versus one starting at 160 kilograms, and mixing the two scales in one pool can distort results.
Primary weight-change findings
The primary kilogram-scale meta-analysis drew on eight independent comparisons. It found that GLP-1 receptor agonist or incretin-based therapy delivered alongside a lifestyle background was associated with a mean difference of roughly 10 kilograms more weight reduction compared with placebo or control conditions. The 95 percent confidence interval ran from about 7.4 to 12.8 kilograms, suggesting the pooled estimate is reasonably precise around that central figure.
The percentage body-weight analysis, which pulled in 11 independent comparisons, pointed in the same direction. The mean difference in percentage body-weight change was approximately 9.5 percentage points in favor of the peptide-plus-lifestyle group, with a confidence interval spanning roughly 7.1 to 11.9 percentage points.
Both analyses also reported prediction intervals, a statistic that tries to capture the range in which a future single trial's result would likely fall given the observed heterogeneity. The prediction intervals were wide, from about minus 18 to minus 2 percentage points for the percentage analysis, which tells a cautionary tale: while the average effect across trials was substantial, individual future trials could land anywhere across a broad range. That width reflects real variability across study populations, peptide types, dosing schedules, and the nature of the accompanying lifestyle programs.
What the heterogeneity means
Both meta-analyses showed very high statistical heterogeneity, with I-squared values above 95 percent. In plain terms, this means the results varied considerably from trial to trial, far more than would be expected from chance alone. High heterogeneity is common in nutrition and behavioral research because it is difficult to standardize what people eat, how much they exercise, or how closely they follow any given protocol across many different study sites and populations.
The review authors point to inconsistent reporting of lifestyle co-interventions as a core problem. Some trials described dietary guidance in detail; others mentioned it only briefly. Exercise protocols ranged from supervised sessions to general advice to move more. Without knowing the precise dose of the lifestyle component, it becomes hard to say how much of the measured effect came from the peptide, how much came from the behavioral change, and how the two might interact.
Risk of bias was also a concern across the included trials, which the authors flag as limiting the certainty of the overall conclusions. These caveats do not invalidate the findings, but they do mean the estimates should be treated as informative rather than definitive.
Body composition and physical function gaps
One of the review's recurring themes is that body weight alone is an incomplete outcome. The authors emphasize that changes in fat mass, lean mass, and physical function matter independently, particularly for older adults or people with conditions that affect muscle quality. A therapy that reduces total weight substantially but also erodes lean muscle would have a different risk-benefit profile than one that preferentially reduces fat mass.
The included trials did not report body composition and physical function outcomes consistently enough to support separate pooled analyses for those endpoints. That gap becomes a call to action in the paper's recommendations: future trials should standardize how they measure and report fat mass, lean mass, grip strength, walking speed, or whatever functional markers are most relevant to their population.
Adherence reporting was similarly inconsistent. Understanding how faithfully participants followed both the peptide regimen and the lifestyle program is essential for interpreting any measured effect, yet many trials did not provide enough detail to assess this properly.
What the review calls for next
The authors conclude that GLP-1 receptor agonist and incretin-based therapies delivered within lifestyle interventions are associated with meaningful reductions in body weight in adults with overweight or obesity. At the same time, they are careful to note that the magnitude of benefit varies substantially across trial contexts and that current evidence carries real uncertainty.
Their recommendations center on standardization. Future trials should describe dietary and exercise co-interventions in enough detail that readers and future meta-analysts can understand what the lifestyle component actually involved. Trials should also report body composition separately from total weight, track physical function outcomes, monitor and report adverse events systematically, and measure adherence to both the pharmacological and behavioral components.
From a research perspective, the review reinforces that studying a peptide therapy in isolation, without characterizing the behavioral context around it, may produce results that are difficult to generalize. The interaction between a metabolic peptide and the person's actual eating and movement patterns is likely part of the story, and future study designs will need to take that seriously to produce results the field can build on.
