A peer-reviewed retrospective cohort study published online September 22, 2026 in Obesity (DOI: 10.1002/oby.70301) found that 1,133,953 US adults were prescribed a GLP-1 receptor agonist without an apparent FDA-approved indication between January 2021 and December 2025 (PubMed-indexed abstract). Those prescriptions came from a denominator of 92,415,648 adults in Epic Cosmos electronic health records who lacked a documented indication — no recorded diagnosis of type 2 diabetes and no qualifying obesity diagnosis. That works out to 1.2% of the group.
The prescribing rate in this population rose from 0.1% in 2021 to 1.5% in 2025, a 15-fold increase (study abstract; Medscape). The study, led by Babak J. Orandi of New York University, tracked three active ingredients — liraglutide, GLP-1S, and GLP-2T — not a single brand. Its conclusion: GLP-1 prescribing without an apparent approved indication has increased rapidly and disproportionately involves socioeconomically advantaged populations, including individuals without obesity, and the risk-benefit profile in these populations remains uncertain (study abstract).
This article is for general informational purposes only and is not medical advice. Talk to a licensed clinician about personal medical decisions.
What the study measured
The researchers defined "no apparent FDA-approved indication" as the absence of a documented diagnosis of type 2 diabetes or obesity with qualifying comorbidities — a BMI of 30 or higher, or a BMI of 27 to below 30 with a related comorbidity (Medscape). Two design choices matter for interpreting the numbers:
- Drugs were tracked by active ingredient, not brand name. As Medscape's report explains, a patient without a qualifying diagnosis might still receive a prescription for the diabetes-labeled version of a drug because it is easier to obtain or costs less, even if the practical goal is weight loss. Sorting by compound let the researchers catch those cases.
- Patients without a recorded weight from at least six months before their first prescription were excluded (study abstract). This safeguard against misclassification meant the analysis could assess measured BMI against obesity thresholds rather than relying solely on diagnosis codes.
The study was retrospective and used the Epic Cosmos database, a large multi-institutional electronic health record system. Prescriptions filled outside participating health systems, and compounded versions of these drugs, were not captured (Medscape).
The prescribing trend, 2021 to 2025
| Year | Prescribing rate among adults without an apparent indication |
|---|---|
| 2021 | 0.1% |
| 2025 | 1.5% |
| Change | 15-fold increase |
The absolute count matters as much as the rate: more than 1.1 million adults over five years (Medscape; study abstract). For context, Medscape's reporting on the study notes that public fascination with GLP-1 drugs has grown well beyond their original use for diabetes, with roughly 1 in 8 US adults now saying they have tried one — often for weight loss rather than a diagnosed medical need.
Who received these prescriptions
Compared with non-recipients in the same denominator, GLP-1 recipients without an apparent indication were (study abstract; Medscape):
- More often female: 85.5% vs. 54.2%
- More often White: 60.0% vs. 51.0%
- Higher median BMI: 25.9 vs. 24.5 kg/m² — but 35.1% had a normal BMI
- Far more likely to have private insurance: 78.0% vs. 47.8%
- More likely to live in the least socially vulnerable neighborhoods, as measured by the Social Vulnerability Index
The clinical profile was mixed: recipients had higher rates of hypertension (32.5% vs. 23.5%) and dyslipidemia (6.8% vs. 2.9%), but lower rates of heart failure (2.6% vs. 4.5%) and coronary artery disease (study abstract).
One finding stands out: recipients were six times as likely to have a history of eating disorders (1.8% vs. 0.3%). Most of these diagnoses were logged without a specific subtype, so the researchers could not determine which conditions were most associated with the pattern (Medscape).
What "no apparent indication" does and does not mean
The study's central measure is documentation, not appropriateness. As Medscape's summary of the limitations puts it, missing documentation does not always mean a missing diagnosis. Several scenarios could explain the gap:
- Uncoded diagnosis. A patient may have had type 2 diabetes or qualifying obesity that was not coded in the record at the time of prescribing.
- Earlier qualification. Some patients may have qualified before joining Cosmos, or before losing weight (Medscape).
- Clinical judgment not captured by codes. Electronic health records capture what clinicians document, not everything they know or intend.
The authors do not characterize these prescriptions as improper. Their own framing, quoted by Medscape: "[T]hese patterns suggest that GLP-1 receptor agonists may be reaching populations at higher risk of disordered eating or being used for weight control in the absence of obesity," and "The risk-benefit profile for those using GLP-1 receptor agonists in the absence of an FDA-approved indication is unknown and merits additional study."
That said, the 35.1% normal-BMI subgroup is clinically notable. Under the BMI thresholds used in the study (BMI ≥30, or 27 to <30 with a comorbidity), these individuals fall below the qualifying obesity criteria. Whether GLP-1 receptor agonists confer net benefit in this group is, per the authors, an open question.
Why the pattern matters
Two implications deserve attention:
Access tracks advantage. Prescribing without a documented indication concentrated among privately insured, less socially vulnerable, predominantly White women, and the authors' conclusion explicitly states that the pattern "disproportionately involves socioeconomically advantaged populations" (study abstract). This is the study's most policy-relevant finding.
EHR-based surveillance has a documentation gap. Electronic health records capture what clinicians code, not necessarily what they know or intend. As GLP-1 use expands — the investigational pipeline PeptidePrices tracks includes GZC8072 and survodutide — large-scale prescribing surveillance will increasingly depend on documentation quality. The Orandi study describes where prescribing currently sits relative to documented indications; it does not predict where future agents will land.
Limitations
- Missing documentation is not necessarily a missing diagnosis. Some patients may have qualified before joining Cosmos or before losing weight (Medscape).
- Compounded and outside-system prescriptions were not captured.
- Eating disorder subtypes were mostly unspecified, so the types most tied to the pattern could not be determined.
- Retrospective design. The study describes associations over time, not causation or clinical outcomes.
- No outcome data. The study did not track adverse events, weight change, or other health outcomes in this population.
The authors reported no specific funding for the study, though some authors disclosed grants, honoraria, advisory board participation, consulting relationships, and other ties with multiple pharmaceutical companies (Medscape).
Continue your research
Sources
- Trends in GLP-1 Receptor Agonist Prescribing Without an Apparent FDA-Approved Indication, Obesity (Silver Spring), September 22, 2026, DOI: 10.1002/oby.70301, PMID 42768905
- Off-Label GLP-1 Surges Since 2021, Uneven Across Groups, Medscape, September 23, 2026
- Life Science Network publication record, September 22, 2026