Claims of hair loss on Wegovy have circulated for years without much to point to. In July a study in the BMJ addressed the question directly. The short answer: an association was found, but few people actually experience it.
The researchers separated adults with type 2 diabetes into those newly starting a GLP-1 drug and those newly starting a different class of diabetes drug, then compared how often each group was later diagnosed with hair loss.
The GLP-1 class includes Wegovy, Ozempic and Saxenda. It was developed for diabetes and came into use for obesity because of the weight loss it produces.
The data came from clinical records at the University of Pennsylvania health system, covering people who first started a drug between January 2019 and September 2024. The two comparisons together cover more than 50,000 people.
| Groups compared | Numbers |
|---|---|
| GLP-1 vs SGLT-2 inhibitor | 12,004 vs 15,221 |
| GLP-1 vs DPP-4 inhibitor | 11,964 vs 11,238 |
The method is called target trial emulation. A randomised trial is difficult to run on a question like this, so existing clinical records are restructured to resemble one. Factors that could affect the outcome, such as age and comorbidity, were matched statistically before comparison.
Hair-loss diagnoses were more frequent in the GLP-1 group. The figures were as follows.
| Comparison | All hair loss | Non-scarring |
|---|---|---|
| vs SGLT-2 inhibitor | 1.37× (1.08–1.73) | 1.53× (1.18–1.97) |
| vs DPP-4 inhibitor | 1.68× (1.28–2.20) | 1.72× (1.28–2.31) |
The range in brackets is the confidence interval: repeat the same study many times and the result would usually fall inside it. When the lower bound is above 1, the finding is unlikely to be chance. All four rows meet that condition.
Non-scarring hair loss is the form in which the follicle is not destroyed, so regrowth remains possible once the cause is removed. In scarring hair loss the follicle is replaced by scar tissue and recovery is difficult. The association here was with the recoverable form.
The paper’s conclusion has two halves. GLP-1 use was associated with an increased risk of non-scarring hair loss; and the absolute risk is low.
These are not contradictory. When something is rare to begin with, a 1.5-fold increase does not add many people. A hair-loss diagnosis is one of those. The multiplier looks large; the headcount does not.
The authors closed on that point: the absolute risk is low, but being aware of the possibility can help when deciding on treatment.
First, the cohort was adults with type 2 diabetes. Dose, duration and baseline health differ from someone taking the drug for weight loss. The result does not transfer directly.
Second, the drug and the weight loss could not be separated. Rapid weight loss is already known to trigger hair shedding. This design cannot tell the two apart.
Funding came from the US National Institutes of Health, and some authors declared outside consulting fees and research support.
This result is not a reason to stop. Hair-loss diagnoses are uncommon and the association was with the recoverable form. Above all, whether to stop a drug is a decision for you and the doctor who prescribed it.
If shedding is bothering you, it is worth raising at your appointment. Mentioning recent weight change and current medication helps narrow the cause. Hair loss is approached differently depending on the cause, and shedding driven by rapid weight loss often recovers once weight stabilises.
No. The study shows more hair-loss diagnoses in the GLP-1 group. That does not mean an individual will experience it, and the paper states the absolute risk is low.
The association here was with non-scarring hair loss, where the follicle is not destroyed, so regrowth remains possible once the cause is addressed.
This design cannot separate them. Rapid weight loss is independently known to trigger shedding.
Not on the basis of this study alone. It is a decision to make with the prescribing doctor, weighing benefit against burden.
