Adults who started taking GLP-1 receptor agonist drugs such as semaglutide logged fewer daily steps and less time exercising within months of beginning treatment, according to an analysis of Fitbit-linked health records presented at the Endocrine Society’s annual meeting, ENDO 2026, in mid-June. The data, drawn from the National Institutes of Health’s All of Us Research Program, offer the first large-scale, wearable-verified signal that the convenience of weight-loss medication may be quietly displacing physical activity, even as patients shed pounds.
Why a Fitbit-tracked activity drop on GLP-1 drugs matters right now
The finding lands at a moment when tens of millions of prescriptions for semaglutide and similar drugs are active across the United States. Weight loss alone does not guarantee better long-term health. Muscle mass, cardiovascular fitness, and metabolic flexibility all depend on sustained movement, and a drug-driven decline in exercise could erode those benefits over time. The ENDO 2026 presentation reported that both steps and exercise minutes dropped after GLP-1 initiation, a pattern that emerged even as participants continued to lose weight.
One plausible explanation is that the drugs do not create inactivity uniformly. Patients who were already less fit before starting treatment, as reflected by higher resting heart rates and poorer sleep metrics in their Fitbit streams, may have been more susceptible to further reductions in movement. GLP-1 drugs suppress appetite and can cause fatigue and nausea, side effects that would weigh more heavily on someone whose baseline fitness was already low. If the activity decline concentrates among those individuals, the drugs may be amplifying pre-existing behavioral patterns rather than introducing an entirely new problem. The All of Us dataset, which captures daily and intraday metrics for heart rate, sleep, and activity, could eventually allow researchers to test that hypothesis directly, though the ENDO 2026 presentation did not break results down by baseline fitness level.
How NIH wearable data and a 2021 exercise trial frame the evidence
The study’s credibility rests on the quality of its data pipeline. All of Us is a large NIH cohort that links consented participants’ surveys, electronic health records, and wearable device streams. Researchers can access minute-level step counts from Fitbit, giving them a granular, objective record of movement that does not rely on self-reporting. A peer-reviewed characterization of the wearables dataset published in Nature Medicine described tens of thousands of contributors providing daily and intraday metrics for steps, activity, heart rate, and sleep, underscoring that this is not a small or idiosyncratic sample.
A separate methods paper detailed how intraday Fitbit records are made available to registered researchers, laying out the tables and data structures that make this kind of before-and-after analysis possible. Because GLP-1 prescriptions appear in linked electronic health records, analysts can align an individual’s medication start date with their wearable data, comparing activity in the months before and after the first dose. That design strengthens the case that any observed shift in steps is temporally associated with drug initiation, even if it cannot prove causation on its own.
The activity decline also sits in tension with earlier clinical evidence. A randomized controlled trial published in The New England Journal of Medicine compared exercise alone, liraglutide 3.0 mg alone, and the combination for maintaining weight loss after an initial diet phase. That trial found that combining structured exercise with the GLP-1 drug produced better weight-loss maintenance than either intervention by itself. The implication is clear: physical activity adds something the drug cannot replicate on its own. If real-world patients are moving less once they start pharmacotherapy, they may be forgoing the very behavior that maximizes the drug’s long-term value.
That gap between trial design and real-world behavior is where the concern sharpens. In a controlled study, participants receive coaching, schedules, and accountability. They are reminded that medication is meant to complement, not replace, lifestyle changes. In everyday life, a drug that reduces hunger and produces visible weight loss can easily become a reason to skip the gym, cut a walk short, or delay starting an exercise routine altogether. The Fitbit data from All of Us capture that everyday reality in a way that clinical trials, by design, do not.
What the ENDO 2026 Fitbit analysis still cannot answer
Several important questions remain open. The ENDO 2026 presentation has not been accompanied by a publicly available abstract or slide deck with exact step-count changes, sample sizes, or statistical adjustments. Without those details, it is difficult to judge how large the activity decline actually was, whether it was clinically meaningful, or whether confounders such as GLP-1 side effects, reduced caloric intake, or seasonal variation in exercise were adequately controlled.
The individual-level Fitbit tables and linked medication start dates sit behind All of Us registered-researcher access, so independent analysts cannot yet replicate the findings. It is also unclear whether the analysis separated recreational walking from purposeful exercise, or whether it considered changes in non-step activities such as cycling and swimming that wearables may capture imperfectly. No direct statements from the study investigators have addressed whether nausea or fatigue, both common early side effects of semaglutide and liraglutide, could account for much of the drop. If the decline is concentrated in the first weeks of treatment and then recovers, the public-health implications are very different than if it persists for months.
Another unresolved issue is how much of the observed change reflects regression to the mean. People may be more likely to seek GLP-1 prescriptions during periods when they are already discouraged about their weight and moving less, which could exaggerate apparent post-prescription declines. Conversely, some patients might temporarily increase activity just before starting medication in a last attempt to lose weight without drugs, then relax once they have the prescription in hand. Without a carefully matched comparison group, it is hard to separate these behavioral patterns from any direct or indirect effects of the drugs themselves.
Practical takeaways for patients and clinicians
For the millions of Americans already taking these drugs or considering them, the practical takeaway is straightforward: weight loss from medication is not a substitute for movement. The emerging wearable data suggest that it may be especially important to protect or even increase physical activity in the months after starting a GLP-1 drug, precisely when appetite is falling and the scale is finally moving in the right direction. That might mean scheduling walks into the calendar, setting step goals in a Fitbit or similar device, or working with a clinician to build a realistic exercise plan that accounts for possible early side effects.
Clinicians may also need to adjust how they frame these prescriptions. Rather than presenting GLP-1 drugs as a stand-alone solution, they can emphasize that the medications work best when paired with activity, echoing the structure of clinical trials that combined pharmacotherapy with supervised exercise. Simple questions during follow-up visits-about daily walking, strength training, or fatigue-could surface early declines in movement that might otherwise go unnoticed.
For researchers, the ENDO 2026 analysis highlights both the promise and the limits of large-scale wearable data. Programs like All of Us make it possible to observe behavior at a level of detail that was unthinkable a decade ago, but they also demand transparency about methods and careful attention to confounding. As more results emerge from these datasets, the central challenge will be turning descriptive patterns-such as a drop in steps after starting a drug-into actionable guidance that helps people not only lose weight, but also maintain the strength, endurance, and resilience that come from staying active.
More from Morning Overview
*This article was researched with the help of AI, with human editors creating the final content.
