News · Heart & Metabolic
GLP-1 drugs carried higher hair loss risk than two rivals
Comparing 12,004 people starting a GLP-1 drug against those starting other diabetes medicines, researchers found more diagnosed hair loss. The absolute risk stayed low.
Based on a peer-reviewed target trial emulation in the BMJ
- Researchers used Penn Medicine electronic health records to emulate a trial in adults with type 2 diabetes starting a new medication between 2019 and 2024, publishing in the BMJ.
- Two comparisons were run: 12,004 GLP-1 starters against 15,221 SGLT-2 inhibitor starters, and 11,964 against 11,238 DPP-4 inhibitor starters.
- GLP-1 use came with higher hair loss risk than SGLT-2 inhibitors (HR, 1.37; 95% CI, 1.08-1.73) and than DPP-4 inhibitors (HR, 1.68; 95% CI, 1.28-2.20).
- The signal was specific to non-scarring alopecia, the reversible kind (HR, 1.53 and 1.72 respectively).
- Findings held across sensitivity and subgroup checks, though they weakened after negative control calibration.
- The authors emphasize that the absolute risk is low.
- This is observational. People are prescribed GLP-1 drugs for reasons that differ from why they get the alternatives.
- Hair loss was captured from diagnostic codes, so it only counts people who sought care and got coded.
Hair loss has been circulating as an anecdotal complaint about the GLP-1 weight-loss drugs for a couple of years, mostly in forums and comment sections. A study in the BMJ went looking for it in medical records, and found a signal.
The objective was to examine the association between glucagon-like peptide-1 (GLP-1) receptor agonists and the risk of incident alopecia (hair loss) in adults with type 2 diabetes, compared with sodium-glucose cotransporter-2 (SGLT-2) inhibitors and dipeptidyl peptidase-4 (DPP-4) inhibitors.
Building a fair comparison
The comparator choice is what makes this worth reading. People starting a GLP-1 drug are not comparable to people starting nothing, so the researchers compared them against people starting a different diabetes drug at the same stage of care.
They used a target trial emulation on electronic health records from Penn Medicine, covering new initiation of GLP-1 receptor agonists, SGLT-2 inhibitors, or DPP-4 inhibitors between January 2019 and September 2024.
Two parallel comparisons resulted: 12,004 GLP-1 receptor agonist initiators and 15,221 SGLT-2 inhibitor initiators in one, and 11,964 GLP-1 receptor agonist initiators and 11,238 DPP-4 inhibitor initiators in the other.
Running the question twice against different comparators is a form of internal replication. A result that appears against only one comparator is usually telling you about that comparator.
The result
It appeared against both.
Use of GLP-1 receptor agonists was associated with a higher risk of alopecia than use of SGLT-2 inhibitors, at 1.37 with a range from 1.08 to 1.73, or DPP-4 inhibitors, at 1.68 from 1.28 to 2.20.
Then the researchers split hair loss by type, and the specificity is the most interesting part. The association was specific to non-scarring alopecia, at 1.53 and 1.72 against the two comparators.
Non-scarring means the follicle survives and regrowth is possible. It is the category that includes telogen effluvium, the diffuse shedding that follows a physiological shock - illness, surgery, or rapid weight loss.
The check that tempers it
The authors ran negative control outcomes, which are conditions the drug could not plausibly cause. If those show an association too, the method is generating spurious signal.
The associations were consistent across sensitivity and subgroup work, with attenuation after negative control outcome calibration.
Attenuation means the effect shrank once that correction was applied. It did not vanish, and it did not survive untouched either. That is an honest reporting of a partially confounded result, and it is the sort of detail that usually disappears from coverage.
The confound the study cannot resolve
Rapid weight loss causes telogen effluvium. This is long established, well documented after bariatric surgery and severe calorie restriction, and has nothing to do with any particular drug.
GLP-1 drugs cause substantially more weight loss than SGLT-2 or DPP-4 inhibitors. So a real effect of losing weight quickly would produce precisely this pattern, including the specificity to non-scarring alopecia.
The paper cannot separate the two, and the mechanism matters for what a patient should do. If it is the weight loss, then slowing the rate of loss might help and stopping the drug is not the only option.
Other limits
Hair loss was identified by using diagnostic codes, which means it counts only people who mentioned it to a clinician and received a formal code. Most hair shedding never reaches that threshold, and whether it does may depend on how worried someone is, which in turn may depend on what they have read about their medication.
The data come from one health system. And this is type 2 diabetes, not the much larger population now taking these drugs for weight loss alone, who differ in age, sex and baseline health.
What to take from it
The conclusion is carefully bounded: use of GLP-1 receptor agonists was associated with an increased risk of non-scarring alopecia in adults with type 2 diabetes, and although the absolute risk is low, awareness of this potential effect may help to inform treatment decisions.
Low absolute risk, reversible type, uncertain mechanism. That is a reasonable thing to know about before starting a medication, and a poor reason to stop one.
People also ask
What is a target trial emulation?
A method for squeezing trial-like structure out of routine health records. Researchers specify the randomized trial they would ideally run, including who would be eligible and when follow-up starts, then reconstruct that design from existing data. It does not deliver randomization, but it prevents several classic errors that plague ordinary database studies, particularly around when the clock starts.
Why compare against other diabetes drugs rather than no drug?
Because people who start any medication differ from people who start none, in ways that are hard to adjust for. Comparing GLP-1 starters against SGLT-2 or DPP-4 starters means both groups are people whose diabetes reached the point of needing a new drug. It is an imperfect but much fairer comparison, and it is why two separate comparator drugs were used.
What is non-scarring alopecia?
Hair loss where the follicle survives, so regrowth remains possible. It includes telogen effluvium, the diffuse shedding that follows a physiological stress such as rapid weight loss, illness or major surgery. Scarring alopecia destroys the follicle and is permanent. The signal here was confined to the non-scarring type, which fits a weight-loss-related mechanism rather than direct damage.
Is this the drug, or is it the weight loss?
The study cannot separate them, and that is the most important caveat. Rapid weight loss is a well-established trigger for telogen effluvium regardless of how the weight comes off, and it happens after bariatric surgery and severe dieting too. The comparator drugs cause much less weight loss. So an effect of losing weight quickly would look exactly like an effect of the drug in this design.
Should someone stop their GLP-1 medication over this?
This is general information rather than medical advice, and the answer is no, not on the basis of this study. The authors describe the absolute risk as low, the hair loss type identified is typically reversible, and these drugs are prescribed for serious conditions where the benefits are substantial and well documented. Anyone concerned about hair shedding should raise it with the clinician who prescribed the medication.
References
- Tang H, Zhang B, Lu Y, et al. Risk of hair loss associated with glucagon-like peptide-1 receptor agonists in adults with type 2 diabetes: target trial emulation. BMJ (2026).
- National Institute of Diabetes and Digestive and Kidney Diseases. Insulin, Medicines, and Other Diabetes Treatments.
- MedlinePlus. Hair Loss.