People taking GLP-1 receptor agonist drugs like Ozempic logged fewer daily steps after losing weight, according to an analysis of Fitbit wearable data collected through the National Institutes of Health All of Us Research Program. The finding raises a practical concern for the millions of Americans now using these medications: if physical activity drops alongside body weight, the long-term benefits of fat loss could be partly offset by declining muscle mass and reduced cardiovascular fitness. The pattern showed up even as other health markers improved, suggesting the decline in movement was not simply a sign of worsening health.
Why reduced movement on GLP-1 drugs carries real health stakes
When someone loses a significant amount of weight, every step they take burns fewer calories than it did before. The body is lighter, so walking, climbing stairs, and other routine movements require less mechanical effort. That shift in energy cost may be enough to quietly reduce how much a person moves throughout the day, not because they feel less motivated, but because their body no longer has to work as hard. The hypothesis that biomechanical unloading from rapid fat loss lowers the energy cost of movement enough to reduce spontaneous daily steps, independent of changes in motivation or appetite, fits what the wearable data show. Step counts fell as weight dropped, even while users were not reporting any intention to be less active.
This matters because physical activity protects against a range of conditions that weight loss alone does not fully address. Muscle preservation, bone density, insulin sensitivity, and cardiovascular endurance all depend on regular movement. If GLP-1 drugs lead to a slow, unnoticed slide in daily activity, patients and clinicians may need to treat exercise as a separate prescription rather than assuming it will take care of itself. The concern is not that GLP-1 medications are ineffective for weight loss-they clearly are-but that the body’s response to being lighter could unintentionally erode some of the metabolic and functional gains unless activity is maintained.
What Fitbit data from NIH’s All of Us program revealed
The All of Us Research Program, run by the NIH, collects electronic health records, surveys, physical measurements, and wearables data from a large and diverse cohort of participants across the United States. Fitbit devices worn by enrolled participants generate daily summaries of step counts, active minutes, and sleep metrics. These records are stored on the program’s Researcher Workbench, where scientists can link device output directly to clinical information such as diagnoses and prescriptions.
Researchers working with this dataset observed that participants on GLP-1 medications recorded lower step counts after periods of weight loss compared with their own earlier baselines. The decline in steps occurred alongside improvements in other measured health indicators, which makes it unlikely that sickness or disability explained the drop. Instead, the data point toward a behavioral or physiological adjustment: people who weigh less simply move less, possibly without realizing it. Because the analysis compares individuals to their own prior activity, rather than to a separate control group, it strengthens the case that something about the weight-loss process itself is reshaping daily movement.
A separate body of evidence supports the idea that structured exercise must be deliberately added to drug therapy for the best outcomes. A randomized controlled trial published in a major clinical journal tested liraglutide, another GLP-1 receptor agonist, against supervised exercise and a combination of both for weight-loss maintenance. The combination group outperformed either intervention alone, and the trial design gave participants in the exercise arms specific physical activity instructions rather than leaving activity to chance. That result reinforces the idea that GLP-1 drugs do not automatically sustain movement levels and that structured routines produce measurably better results.
Measurement limits and what Fitbit can and cannot confirm
Any conclusion drawn from consumer wearables depends on how accurately those devices capture real activity. Fitbit step counts have been validated against research-grade accelerometers like the ActiGraph and perform well in free-living conditions. But the picture is less clear for moderate-to-vigorous physical activity minutes. A methods paper on Fitbit accuracy within All of Us noted that Fitbit devices may overestimate MVPA minutes, which means researchers need to treat intensity-based metrics with more caution than raw step counts.
That distinction is relevant here. If the observed decline in activity were measured only in MVPA minutes, the finding could partly reflect device error rather than genuine behavior change. Because the core signal comes from step counts, which are more reliable in everyday settings, the pattern carries more weight. Still, wearable data cannot tell researchers why someone took fewer steps. The dataset does not include participant statements explaining whether reduced movement was intentional, nor does it isolate the precise timing of Ozempic prescriptions against the activity timeline with the granularity a controlled trial would offer. Confounding factors such as seasonal changes, work schedules, or unrelated life events could also play a role.
Another limitation is that wearables capture quantity of movement better than quality. A reduction of a few hundred steps might mean one fewer short walk, or it could reflect a shift from walking commutes to more stationary transport. Fitbit records cannot distinguish between someone who trades low-effort strolling for focused strength training and someone who simply sits more. As a result, interpreting changes in step counts as straightforwardly positive or negative requires caution, even when the devices themselves are reasonably accurate.
Open questions about activity decline on semaglutide
Several gaps in the evidence prevent a definitive causal story. The All of Us wearables dataset links Fitbit output with electronic health records, but published analyses have not yet isolated semaglutide users specifically and reported exact pre- and post–step-count differences with confidence intervals. Without that level of detail, the size of the activity drop and how it varies by dose, duration, or patient characteristics remains unclear. It is also unknown whether step counts eventually rebound as people stabilize at a lower weight or whether reduced movement persists long term.
There is also no direct evidence separating the biomechanical explanation-that lighter bodies simply expend less energy per step-from possible drug-related effects on motivation or fatigue. GLP-1 receptor agonists act on brain pathways involved in appetite and reward, and it is plausible that these same circuits could subtly influence the desire to move. Alternatively, some patients report gastrointestinal side effects or transient malaise while starting therapy, which might temporarily dampen activity. Untangling these possibilities would require prospective studies that track symptoms, mood, and motivation alongside detailed activity data.
Future research could use the All of Us infrastructure to design nested studies that more precisely align prescription dates, dose changes, and activity trajectories. Investigators might compare GLP-1 users with matched individuals achieving similar weight loss through diet and exercise alone, to see whether step declines are unique to drug therapy or a general consequence of becoming lighter. They could also look at subgroups, such as older adults or people with existing mobility limitations, where even modest reductions in daily movement might carry outsized risks.
Practical implications for patients and clinicians
For now, the emerging signal from wearables suggests a simple clinical message: do not assume that weight loss from GLP-1 drugs automatically translates into more movement. Patients may need explicit guidance to preserve or increase activity as their bodies change. That could mean setting step-count goals, scheduling resistance training to protect muscle, or using reminders and social support to keep daily movement from drifting downward. Because step counts are easy to track, clinicians can incorporate them into routine follow-up, adjusting goals if a patient’s numbers slide.
At the same time, the data should not deter appropriate use of GLP-1 medications. Instead, they highlight an opportunity to pair powerful pharmacologic tools with equally intentional activity strategies. If clinicians and patients treat exercise as a co-therapy rather than an optional add-on, the combination could lock in more durable metabolic health, maintain functional capacity, and help ensure that the gains recorded on the scale are matched by gains in how people move through their daily lives.
More from Morning Overview
*This article was researched with the help of AI, with human editors creating the final content.
