GLP-1s Don’t Increase Neurologic, Psychiatric Risk

An analysis based on a massive database of US electronic health records (EHRs) adds to evidence favoring the safety of glucagon-like peptide 1 (GLP-1) medications, finding no increased risk for many neurologic and psychiatric ailments when semaglutide was compared with other diabetes drugs.

Riccardo De Giorgi, MD, DPhil, of the University of Oxford, Oxford, England, and colleagues published their research in eClinicalMedicine.

At this time, other research studies are investigating whether GLP-1s have potential for neurologic, psychiatric, and substance use disorders, even amid some concerning reports. The European Medicines Agency and the UK Medicines and Healthcare products Regulatory Agency are reviewing reports of worsening mood and suicidal behavior in people who have used GLP-1 drugs. However, after its preliminary investigation, the US Food and Drug Administration (FDA) found no evidence of harm with GLP-1s.

To address questions about the potential psychiatric complications of semaglutide, De Giorgi and colleagues designed a retrospective cohort study drawn from records from TriNetX US Collaborative Network. This network holds anonymized EHR data from more than 100 million US patients. De Giorgi and colleagues sought to compare the experiences of people with diabetes who had taken semaglutide (Ozempic, Novo Nordisk) sitagliptin (Januvia, Merck), empagliflozin (Jardiance, Boehringer Ingelheim), and glipizide.

“Our findings can inform current regulatory investigations and public health and stimulate clinical trials to test the role of semaglutide for the treatment and prevention of cognitive deficits and substance misuse,” De Giorgi and colleagues wrote.

In an interview with Medscape Medical News, De Giorgi emphasized that findings reported in the paper for cognitive deficits and nicotine use should be interpreted as positive signals.

“I think we need to be extremely cautious about that,” De Giorgi said. “The obvious reason is that this is an observational study. We should refrain from making any causal associative inference.”

The risk for cognitive deficits was significantly lower after semaglutide than after sitagliptin (HR, 0.72; 95% CI, 0.64-0.80; P_adj < .0001) and glipizide (HR, 0.72; 95% CI, 0.63-0.81; P_adj < .0001) but similar between semaglutide and empagliflozin (HR, 0.96; 95% CI, 0.86-1.08; P_adj = .51).

In the same vein, the risk for dementia was lower after semaglutide than after sitagliptin (HR, 0.52; 95% CI, 0.40-0.68; P_adj < .0001) and glipizide (HR, 0.63; 95% CI, 0.46-0.86), although this did not reach significance after correction ( P_adj = .075). The risk was similar between semaglutide and empagliflozin (HR, 0.91; 95% CI, 0.69-1.21; P_adj = .53).

The risk for nicotine dependence appeared to be significantly lower after semaglutide than after glipizide (HR, 0.72; 95% CI, 0.61-0.85; P_adj = .0027) and empagliflozin (HR, 0.77; 95% CI, 0.65-0.90; P_adj =.024) and lower after semaglutide than after sitagliptin, although not after correction for multiple testing (HR, 0.82; 95% CI, 0.70-0.95; P_adj = .23).

There did appear to be an increased risk for migraine after semaglutide compared with after glipizide (HR, 1.20; 95% CI, 1.08-1.33; P_adj = .015) in the study findings. Otherwise there did not appear to be any increased risk for a first diagnosis of a neurologic or psychiatric condition with semaglutide as compared with the other diabetes drugs.

The other neurologic and psychiatric outcomes studied were encephalitis, Parkinson’s disease, epilepsy/seizure, insomnia, nerve disorder, myoneural junction/muscle disease, intracranial hemorrhage, ischemic stroke, alcohol misuse, opioid misuse, cannabis misuse, stimulant misuse, psychosis, bipolar disorder, depression, anxiety, obsessive-compulsive disorder, and suicidality.

De Giorgi cautioned against assuming that the results seen in the study of EHRs of people with diabetes would hold true for those taking semaglutide only for weight loss. Many people with diabetes also are overweight, but the GLP-1 drugs might work differently when people use them only for obesity.

“They might be fundamentally different groups of people, and these are powerful medications,” De Giorgi said. “We need to be extremely, extremely careful.”

The analysis was supported by the National Institute for Health and Care Research Oxford Health Biomedical Research Centre and the Medical Research Council. De Giorgi’s coauthors reported receiving funding for other work from Novo Nordisk, Five Lives SAS, Cognetivity Ltd, Cognes Ltd, P1vital, Lundbeck, Servier, UCB, Zogenix, J&J, and Syndesi.

Kerry Dooley Young is a freelance journalist based in Washington, DC.