GLP-1 DIABETES DRUGS LINKED TO 37% HIGHER HAIR LOSS RISK
A BMJ study of GLP-1 hair loss in adults with type 2 diabetes found that people who started a GLP-1 drug were 37% more likely to be diagnosed with alopecia than people who started an SGLT-2 inhibitor. The authors said the absolute risk is low, and because the work was observational, it does not prove the drugs caused anyone to lose hair.
The paper was published on July 22, 2026, by researchers at the University of Pennsylvania and drew on electronic health records from Penn Medicine. That system’s database covers more than 6.5 million patients and more than 50 million clinical encounters.
Who was compared with whom
The team looked at adults with type 2 diabetes who newly started a GLP-1 drug, an SGLT-2 inhibitor or a DPP-4 inhibitor between January 2019 and September 2024. The GLP-1 group covered exenatide, liraglutide, dulaglutide, semaglutide, lixisenatide and tirzepatide.
In the main comparison, 12,004 people started a GLP-1 drug and 15,221 started an SGLT-2 inhibitor. The outcome was a new alopecia diagnosis, picked up through diagnostic codes in the records.
The GLP-1 users started out quite different. They were younger (average age 58.1 versus 64.5) and had a higher body mass index (36.2 versus 32.3). The researchers weighted the data to balance those measured differences before comparing the groups.
After that adjustment, the hazard ratio for alopecia was 1.37 for GLP-1 users against SGLT-2 users, and 1.68 against DPP-4 users. Penn Medicine put that as 37% and 68% more likely to have hair loss.
In raw numbers, the adjusted rate was 6.91 cases per 1,000 person-years for GLP-1 users versus 5.04 for SGLT-2 users. Against DPP-4 users it was 6.53 versus 3.89.
The signal sat in non-scarring alopecia, the type that leaves hair follicles intact and may allow regrowth. The hazard ratios there were 1.53 against SGLT-2 inhibitors and 1.72 against DPP-4 inhibitors. The study did not find a clear link to alopecia areata, and diagnoses clustered in the first year after treatment began, peaking just before the 12-month mark.
The detailed results also showed a higher risk of scarring (cicatricial) alopecia against DPP-4 users, with a hazard ratio of 3.83, but not against SGLT-2 users.
The authors pointed to weight loss and reduced calorie intake as a plausible explanation. Rapid weight loss is a known trigger for telogen effluvium, a typically temporary kind of shedding. The study did not measure how much weight loss played a part, so the mechanism is not established.
Earlier postmarketing reports had already raised the question, and the paper notes that the FDA said it was evaluating the possible safety signal.
Alopecia was identified from codes, not scalp exams, so the study did not assess how severe any case was, how long it lasted or whether it reversed. Residual confounding is also possible in an observational design.
All the patients were adults with type 2 diabetes treated in one health system, so the findings do not automatically apply to people taking GLP-1 drugs only for weight loss. The work was funded in part by the US National Institutes of Health, which the authors said had no role in the research or the decision to publish.
The authors wrote that “Although the absolute risk is low, awareness of this potential effect may help to inform treatment decisions.”
The BMJ Group news release the next day described the result as an association and said the absolute risk remained low. Penn Medicine published its own article on August 31, 2026, which opened with the line “Research now shows that this risk is real, albeit muted.”
The researchers called for more work on mechanisms, severity, reversibility and which patients may be at higher risk. In the meantime, they wrote, “Clinicians may consider discussing the possibility of hair loss when initiating GLP-1 receptor agonist therapy and monitoring for symptoms during follow-up, particularly in the first year.”
Penn Medicine’s study page was last updated on October 10, 2026, and still carries the same findings.
