Short answer: Weight change can slow as energy needs and behavior change. Metabolic adaptation is one contributor, not a diagnosis that can be made from a few scale readings. Animal experiments help investigate mechanisms but cannot establish a human treatment result by themselves.

Why plateaus occur

A smaller body generally requires less energy to maintain and move. Biological responses to weight loss can also influence appetite and energy expenditure. A plateau is therefore not proof that a person has stopped trying or that their body is permanently “broken.” Short-term changes in water and digestive contents can also obscure longer trends.

Metabolic adaptation often refers more specifically to an energy-expenditure change beyond what would be predicted from the change in body size and composition. It should not be used as a catch-all term for every slowdown. The size and persistence of adaptation vary with the setting and how it is measured. In a controlled human study, Leibel and colleagues measured energy expenditure before and after experimentally induced weight changes and found compensatory expenditure changes that opposed maintenance of the altered weight. This supports metabolic adaptation as a mechanism; it does not diagnose the cause of an individual's plateau. Leibel and colleagues, original human study; VA and Department of Defense patient guide

What investigational evidence can establish

A well-designed human trial can estimate the benefits and harms of a particular intervention in its enrolled population. Early trials may identify a signal and help determine what larger studies should test. They usually cannot resolve uncommon harms, long-term effects or all relevant clinical subgroups.

As of November 2026, retatrutide has published phase 3 evidence, not only early-phase results. TRIUMPH-1 studied adults with obesity without diabetes, while TRIUMPH-2 studied adults with obesity and type 2 diabetes; both compared retatrutide with placebo over 80 weeks. Their findings strengthen the evidence in those populations without resolving every long-term outcome or establishing superiority over an active comparator. The compound remains investigational, with FDA approval distinct from trial publication. TRIUMPH-1; TRIUMPH-2

Why animal findings are not a shortcut

An animal model can reproduce one feature of a disease while omitting others. Species differ in physiology, receptor biology, lifespan and exposure. Laboratory animals may also have more uniform genetics, diets and environments than people. A controlled injury or induced metabolic disorder does not necessarily match a long-standing human condition.

Translation requires a chain of evidence: a credible mechanism, reproducible experiments, suitable exposure, safety characterization and well-designed human studies. A result in one link cannot stand in for the whole chain. Even human trials require replication and careful interpretation.

The useful question is not whether animal research “counts.” It does. The question is which claim that experiment can support and which claims still need human evidence.