News & Trends

People on GLP-1s moved less, not more. American football season just opened.

By the US Health Digest editorial team · Published August 30, 2026 · Every claim linked to its primary source

A poster presented at ENDO 2026 — abstract SAT-714, a conference abstract and not a peer-reviewed paper — reported that among 753 adults with obesity in the NIH All of Us program with Fitbit data before and after starting a GLP-1, mean daily steps fell from 5,047 to 4,487 (−560 ± 2,203/day, p<0.001); moderate-to-vigorous activity, reported for 570 people rather than 753, fell from 27.9 to 22.2 minutes a day. We read the abstract. It has no control group, no stated follow-up window, and — despite the headlines about weight loss — no weight measurement at all. Meanwhile, the trials that produced the famous weight numbers prescribed the activity.

The weekend the calendar turned

Per NCAA.com, "The first FBS games are scheduled for Saturday, Aug. 29, while FCS starts a couple of days earlier on Thursday, Aug. 27" — a Week 0 slate that included North Carolina against TCU in Dublin, San Jose State at Southern California, Hawaii at Stanford and New Mexico State at Florida State. ESPN puts its own Week 0 window at "Aug. 28-30." Week 1 runs Thursday, September 3 to Monday, September 7. Then the NFL: per the league's May 14 announcement, "the defending Super Bowl Champion Seattle Seahawks host the New England Patriots at Lumen Field on Wednesday, September 9 (8:20 p.m. ET, NBC)." For scale, the Bureau of Labor Statistics' American Time Use Survey for 2025 (USDL-26-1022) reports that "Watching TV was the leisure and sports activity that occupied the most time (2.6 hours per day), accounting for half of all leisure time, on average (5.2 hours)."

That is the whole of the framing. Nothing below is about football, and nobody's television is anybody's medical business.

What the poster actually reports

The primary document is poster abstract SAT-714, "Losing Pounds, Not Gaining Steps: The Paradox of GLP-1 Receptor Agonist Therapy," by Sajana Maharjan, MD (HSHS St John's Hospital) with Gulshan Dangol, MD and Quang Le, MD, presented on the ENDOExpo poster floor in Chicago on June 13, 2026. In its own words: "We conducted a retrospective pre-post cohort study using the All of Us Research Program, identifying adults with obesity who were initiated on GLP-1 RA therapy (semaglutide, tirzepatide, liraglutide, dulaglutide) and had Fitbit activity data before and after treatment."

The abstract reports that of 1,950 patients with obesity prescribed a GLP-1, 753 (38.6%) had usable Fitbit data — 78.6% female, mean age 52.7 years, with musculoskeletal pain in 81.9% and type 2 diabetes in 48.1%. In this conference abstract, which has not been peer reviewed, daily steps fell from 5,047 ± 3,073 to 4,487 ± 3,133 (−560 ± 2,203; p<0.001). Moderate-to-vigorous activity in the same abstract fell from 27.9 ± 28.2 to 22.2 ± 23.3 minutes/day (−5.7 ± 25.3; p<0.001) — in 570 people, not 753, a detail no coverage we read mentions. Men declined more than women, and people with musculoskeletal pain more than those without; age, morbid obesity, stroke history and heart failure made no significant difference. The authors conclude that "weight loss alone may not promote increased physical activity, highlighting the need for future studies to explore targeted interventions that encourage physical activity alongside pharmacologic therapy."

Three things the poster does not contain

A control group. The design is pre–post within the same people, with no untreated comparison arm, so nothing in it separates the medication from everything else that happened over the same months — including the calendar. A systematic review of 37 studies covering 291,883 participants in eight countries reported that "levels of physical activity vary with seasonality". An uncontrolled design cannot exclude that.

A follow-up window. The abstract says "before and after treatment" and never defines how long either period was, so we could not determine over what span the 560 steps disappeared.

Any weight measurement. This is the one that matters. The Endocrine Society's June 13 release opens: "Adults with obesity losing weight with glucagon-like peptide-1 (GLP-1) receptor agonist medications significantly decreased their physical activity". ScienceDaily ran it on June 14 as "People taking GLP-1 weight loss drugs like Ozempic started moving less". But the abstract reports no body-weight outcome at all. The measured association is with starting the drug, not with losing weight on it — and the release's more careful line, "The study found no evidence that weight loss from these medications led to increased physical activity," still leans on a weight change the poster never reports.

And it is not peer reviewed: as of August 30, 2026 we found no journal publication of this work indexed in PubMed, Europe PMC or Crossref.

"The first large study" — what we found when we checked

The Society's release states: "This is the first large study analyzing data from wearable fitness trackers among adults taking GLP-1 receptor agonists." We looked for counter-examples before repeating it, and found two.

Ten weeks before ENDO, Butensky and colleagues published in Obesity Surgery (online March 31, 2026) a peer-reviewed Fitbit analysis of the same database — 669 patients, 581 on semaglutide, modeled by mixed-effects regression adjusted for demographics, BMI and comorbidities. It points the same way: "Step counts did not show sustained improvement, with semaglutide and RYGB patients remaining below baseline". Pointing the other way, Grosicki and colleagues tracked 66 people on a wrist wearable for 12 weeks after starting a GLP-1 against a propensity-matched control group, and found that "Trends (Ps < 0.10) suggested that increases in weekly physical activity were associated with GLP-1 RA medication (31.5 ± 13.2 min)" — not a significant result, in a far smaller sample, but the only one of the three with a comparison group.

We are not calling the "first large study" line wrong; "large" is undefined, and the ENDO cohort is the biggest of the three. But a reader weighing a poster deserves to know that a peer-reviewed analysis of the same program got there first, and that the one study with a control arm points the other way.

The comparison nobody is making: the trials prescribed the activity

The weight-loss figures everyone quotes were not produced by a drug alone, but by a drug plus a written lifestyle protocol. Section 14.2 of the Wegovy (semaglutide) prescribing information on DailyMed states: "In Studies 2, 3, 5, 7, 8 and 9, all patients received instruction for a reduced-calorie diet (approximately 500 kcal/day deficit) and increased physical activity counseling (recommended to a minimum of 150 min/week) that began with the first dose of study medication or placebo and continued throughout the trial." The label identifies its Study 2 as NCT#03548935 — the 68-week, 1,961-patient trial published as STEP 1. Section 14.1 of the Zepbound (tirzepatide) prescribing information says the same of its Study 1, NCT04184622 — SURMOUNT-1: "all patients received a standard lifestyle intervention which included instruction on a reduced-calorie diet (approximately 500 kcal/day deficit) and increased physical activity counseling (recommended minimum of 150 min/week) that began with the first dose of study medication or placebo and continued throughout the trial."

That prescribed minimum is the floor of the federal recommendation: per the Office of Disease Prevention and Health Promotion, "adults need at least 150 to 300 minutes of moderate-intensity aerobic activity, like brisk walking or fast dancing, each week."

Read the last row against the first three. The trial numbers were achieved with a calorie target and an activity target written into the protocol and repeated at every visit. The All of Us cohort is not a trial — it is people getting prescriptions in ordinary American care, and the abstract documents no lifestyle program at all. That does not make the decline a drug effect; there is no control arm to support that reading either. It does mean the drug-versus-real-world gap and the protocol-versus-no-protocol gap are confounded in every discussion of this poster, including the coverage that framed it as a paradox.

The half of the lean-mass story that is modifiable

The Society's release gives its reason for caring: "GLP-1 receptor agonists like semaglutide, liraglutide, dulaglutide and tirzepatide reduce not only fat but also lean muscle mass." That needs a qualifier the release does not give: lean mass is not the same thing as muscle, the fall is roughly proportional to the weight lost, and almost no trial has measured whether strength holds up. The full accounting is in our page on GLP-1s and lean mass. Activity is the part of that picture pharmacology does not settle — which is why even a weak activity finding is worth reading properly.

Adjacent evidence: protein-first eating on a GLP-1, eating on a GLP-1, stopping a GLP-1 and weight regain, and the prescriber questions in our GLP-1 FAQ for women. Nothing here is a reason to start, stop or change any medication, and none of it is instruction about how anyone should spend their week.

Bottom line — and what we could not verify

We could not establish the observation windows in SAT-714, because the abstract does not state them; nor whether the cohort lost weight, because no weight outcome is reported; nor find any peer-reviewed publication of the work. We also found no US dataset tying step counts to the football calendar.

What is left still matters. The finding is directionally supported by a peer-reviewed analysis of the same database — and it is 560 steps a day reported in a conference abstract that has not been peer reviewed, in an uncontrolled cohort with no stated window and no weight data. The sharper fact is in the labeling rather than the conference hall: the 14.9% in STEP 1 and the 20.9% at 15 mg in SURMOUNT-1, different estimands from different trials and not comparable to each other, were each produced alongside a 500 kcal/day deficit and a minimum of 150 minutes a week of counseled activity, from the first injection onward. That protocol is written into the label. Whether it is written into anyone's care is a separate question the poster does not answer.