Patients with type 2 diabetes who received glucagon-like peptide-1 receptor agonists had more documented new smell and taste disturbances than matched patients receiving other diabetes medications, according to findings published in JAMA Otolaryngology–Head & Neck Surgery.
Researchers analyzed electronic health records from the TriNetX Global Collaborative Network to evaluate whether glucagon-like peptide-1 receptor agonist (GLP-1 RA) use was associated with incident smell and taste disturbances. The retrospective cohort study included adults with type 2 diabetes and no prior record of smell or taste disturbances. After propensity score matching, the analysis included 438,474 patients prescribed GLP-1 RAs and 438,474 matched controls prescribed other antidiabetic medications without GLP-1 RA exposure.
Outcomes were assessed from 3 months to 2 years following the index prescription, with a 90-day lag intended to reduce bias related to early treatment instability and discontinuation. The primary outcome was new-onset smell or taste disturbance identified through diagnostic codes. Researchers also assessed smell and taste disturbances separately, including anosmia, parosmia, and parageusia.
Overall smell and taste disturbances were documented in 1,615 patients receiving GLP-1 RAs, or 0.37%, compared with 960 matched controls, or 0.22%. This corresponded to an absolute risk increase of 0.15%, or about 1.5 additional documented cases per 1,000 treated patients, and a number needed to harm of 670.
In time-to-event analyses, GLP-1 RA use was associated with a 48% higher hazard of documented smell or taste disturbance compared with other antidiabetic medications. Smell disturbances were documented in 0.15% of patients receiving GLP-1 RAs and 0.07% of controls, while taste disturbances were documented in 0.18% and 0.10%, respectively.
Among specific smell disorders, anosmia was diagnosed in 0.13% of patients receiving GLP-1 RAs compared with 0.07% of controls. Parosmia was diagnosed in 0.05% and 0.02%, respectively, and showed the largest relative association.
“Our study found that the use of GLP-1 RAs is associated with an increased incidence of smell and taste disturbances among patients with T2D,” wrote Jonathan Zontag, BSc, and Nir Zontag, BSc, of Hadassah Medical Center, Faculty of Medicine, Hebrew University, Jerusalem, Israel. The researchers added that the findings suggest “a potential multifactorial association, possibly involving both peripheral sensory receptors and central neural pathways.”
The researchers noted that prior evidence has been mixed. Some experimental studies have suggested potential protective effects of GLP-1 RAs on olfactory function, while prior pharmacovigilance analyses have reported increased rates of olfactory adverse events among users of the medications.
The absolute incidence of documented sensory disturbances remained low. Fewer than 0.5% of patients receiving GLP-1 RAs had a documented smell or taste disturbance during the study period. The researchers emphasized that discontinuation of therapy should not be routine and should instead be guided by shared decision-making, overall clinical context, and the established cardiometabolic benefits of the drug class. In patients who develop symptoms, clinicians should follow standard diagnostic evaluation and avoid prematurely attributing smell or taste changes to GLP-1 RA use.
The study has several limitations. The observational design does not establish causality. Outcome identification relied on diagnostic coding rather than standardized smell or taste testing, and the authors noted that these diagnoses are often assigned based on patient-reported symptoms rather than objective testing—measures they described as inherently subjective and subject to potential misclassification.
The researchers also reported that residual confounding could not be excluded. Because outcome identification relied on diagnostic codes for symptoms that are often patient-reported, the researchers noted that differences in health care utilization could influence symptom reporting and diagnostic ascertainment. They also noted that variables not fully captured in propensity score matching, such as psychiatric comorbidities, may still influence the findings.
The researchers further noted that data on adherence, treatment duration, medication dose, and indication for therapy were unavailable. The control group comprised a range of other antidiabetic medications, and all GLP-1 RAs were grouped together despite molecular differences that the researchers acknowledged could have distinct effects. The cohort was limited to patients with type 2 diabetes, and the researchers noted that treatment indication was unavailable; as a result, the findings should be interpreted cautiously when considering patients using these agents for other indications, such as obesity.
In an invited commentary, Charles A. Riley, MD, of Veteran Affairs Boston and Boston University Chobanian & Avedisian School of Medicine, and Edward D. McCoul, MD, MPH, of Ochsner Clinic Foundation, said the findings highlight a potential adverse effect that may warrant greater attention as use of the drug class expands. They noted that smell and taste function play important roles in nutrition, safety, and quality of life and suggested that physicians consider directed questions about baseline sensory function before initiating therapy.
“For patients with uncontrolled diabetes, cardiovascular disease, or severe obesity, the risk of adverse effects, including sensory disturbance, may be acceptable. However, when used for marginal weight loss or cosmetic purposes, an equivalent probability of harm may be weighted differently,” Riley and McCoul wrote.
Disclosures: Zontag and Zontag reported no conflicts of interest. McCoul reported personal fees from 3D Matrix, Advanced Rx, and Sanofi/Regeneron and stock options from Zsquare outside the submitted work. No other disclosures were reported.