Some of my earliest blogs (Kirsten Lawson – 4/4 – National Elf Service) were about the interface between physical health and mental health and it remains a keen interest of mine. With the well known evidence relating to earlier mortality for people with mental illness and the significant comorbidity of chronic physical health conditions and mental illness (see my earlier blogs for further detail), this remains a vital area for understanding and improvement.
There has been a lot of work done in relation to the link between diabetes and mental illness (Search for “tag diabetes AND mental health” – National Elf Service) and the link between metabolic and psychiatric conditions appears to be bidirectional (Nouwen et al. 2010) meaning that if you have a metabolic disorder you are more likely to have a mental illness and if you have a mental illness you are more likely to have a metabolic disorder. It is really important that by treating one end of this spectrum we don’t impact the other.
Glucagon-Like Peptide 1 drugs aren’t new. These drugs have been used to treat diabetes since early 2000s, but the interest in them has increased since their approved licence for weight loss. There have however been some concerns from international organisations and the World Health Organization regarding increased risk of suicidality, self-injury, and psychiatric adverse events in people prescribed GLP1-RAs.
Formal research in relation to GLP1-RAs and psychiatric effects is limited to date and this blog reports on Pierret et al.’s meta-analysis of randomised placebo-controlled trials to evaluate psychiatric, cognitive, and quality of life outcomes with GLP1-RA treatment, published in JAMA in May 2025.
The relationship between metabolic and psychiatric disorders is bidirectional.
Methods
The researching team completed a search of a range of databases (MEDLINE, Embase, PsycINFO, and Cochrane Central Register of Controlled Trial) in line with PRISMA guidelines, from inception to June 24, 2024.
Specifically, they looked for double-blind placebo-controlled trials comparing GLP1-RA to placebo in adults with overweight/obesity and/or diabetes, reporting on psychiatric, cognition, or quality of life outcomes. If the study included patients with pre-existing psychiatric disorders, it was excluded.
The outcomes looked at were risk of psychiatric adverse events (serious and nonserious) and change in mental health symptom severity, health-related quality of life, and cognition.
Serious psychiatric adverse events commonly reported were major depression, suicidality, and psychosis; nonserious events included anxiety and insomnia.
Results
From the original 19,909 papers identified, 99 trials were included in the systematic review and 80 articles had sufficient data for the meta-analysis. Of the 80:
- 64 reported on psychiatric adverse events,
- 41 reported on QOL, and
- 3 reporting on eating behaviours.
Study population characteristics
- Mean age of participants 60.1 years (SD 7.1).
- Female 40.1% and male 59.9%.
- Patients with type 2 diabetes 62%.
- Patients who were overweight (BMI≥25) or obese (BMI≥30) 29%.
- Patients with type 1 diabetes 9%.
- GLP1-RA Liraglutide used in 30% studies.
- GLP1-RA Semaglutide used in 24.7% studies.
Meta-analysis
Results interpretation notes
- Log risk ratio (RR) compares the risk of an event happening in the two groups. A positive result is increased risk, a negative result lower risk and if the confidence interval crosses 0 then there is no perceived difference.
- I2 is a measure of variation across different studies (aka heterogeneity) that is due to true differences between studies rather than simply due to chance.
- Hedges’ g is a measure of effect size that is used to compare differences in means between two groups in a meta-analysis. A result of 0.2, 0.5, and 0.8 representing small, medium and large effect sizes respectively.
- Z-score indicates how many SDs a data point is away from the mean. Larger z-scores (positive or negative) correspond with smaller p-values.
Psychiatric adverse events – 64 studies
No statistically significant difference was found between groups in serious psychiatric adverse events (log[RR] = −0.02; 95%CI,−0.20 to 0.17; P = 0.90); low heterogeneity (I2=0%, all variance is due to chance).
No statistically significant difference was found between groups in nonserious psychiatric adverse events (log[RR] = −0.03; 95%CI, −0.21 to 0.16, P = 0.76); moderate heterogeneity (I2=40%, a mix of studies with potential cause for concern).
Psychiatric symptoms – 14 studies
Only studies reporting on depression and eating behaviours had sufficient data for meta-analysis.
Depression: Little evidence of an association between change in depressive symptoms and GLP1-RA treatment compared with placebo was found (Hedges’ g = 0.02; 95%CI, −0.51 to 0.55; P = 0.94); low heterogeneity (I2= 0%, all variance is due to chance).
Eating behaviours: Moderate heterogeneity (I2= 61%, a mix of studies and results may be due to this variation as opposed to just chance). GLP1-RA treatment was associated with:
- Improved eating restraint (g = 0.35; 95% CI, 0.13 to 0.57; P = 0.002) and
- Improved emotional eating (g = 0.32; 95% CI, 0.11 to 0.54; P = 0.003)
- But not with external eating/disinhibition (g = 0.23; 95% CI, −0.10 to 0.57; P = 0.17).
Health-Related Quality of Life (QoL) – 58 studies (41 for meta-analysis)
Compared to placebo, GLP1-RA treatment was associated with:
- Improved mental-health-related QoL (g = 0.15; 95% CI, 0.07 to 0.22; P < 0.001),
- Improved physical-health-related QoL (g = 0.20; 95% CI, 0.14 to 0.26; P < 0.001)
- Improved diabetes-related QoL (g = 0.23; 95% CI, 0.15 to 0.32; P < .001),
- Improved weight-related QoL (g = 0.27; 95%CI, 0.18 to 0.35; P < .001), and
- Improved overall QoL (g = 0.07; 95%CI, 0.02 to 0.12; P = .01),
The authors went on to use meta-analysis to see if specific characteristics had an impact on the results. They found no statistically significant association between change in mental-health-related QoL with any baseline variables or magnitude of weight / HbA1c change. This would suggest that the changes in QoL were not secondary to the changes in weight or blood sugar control; implying there may be a separate underlying mechanism.
For physical-health-related QoL, meta-analysis found that greater improvements were associated with:
- A higher proportion of female participants (z = 3.314, P < 0.001),
- Greater weight loss (z = −2.744, P = 0.006), and
- Lower mean age (z = −2.656, P = 0.008).
Cognitive outcomes – 4 studies
No association was found between GLP1-RA treatment and change in global / composite cognition compared with placebo (g = 0.01; 95%CI, −0.02 to 0.04; P = 0.17).
This study found that GLP1-RA was associated with improved quality of life and did not identify any negative impact on mental health.
Conclusions
The authors said that in populations without pre-existing psychiatric disorders, their study found:
no evidence for increased risk of psychiatric adverse events with GLP1-Ras compared with placebo in people with overweight/obesity or diabetes.
Strengths and limitations
The study was well thought out (except for the one specific flaw that I will come to). The team used a very specific research question, followed PRISMA reporting guidelines, the quality of included studies was considered using the Cochrane risk-of-bias tool, heterogeneity was assessed (low to moderate), and they identified a low risk of publication bias.
The meta-analysis brought together specific quality of life outcomes identified by studies and appeared to deal with these thoughtfully. The number of studies for each subcategory of clinical impact did dwindle in some areas, making the results less robust to apply to wider populations.
Now, what I see as the major flaw: Studies were excluded if they examined patients with pre-existing psychiatric or neurological disorders (e.g. Alzheimer’s disease).
Given where we started, with the highlighting of the bidirectional relationship between metabolic disorders and mental health disorders, this can’t help but feel like an own goal as the results will only be applicable to the overweight/obese and diabetic population without a pre-existing mental health disorder.
It also means that the results cannot be applied to psychiatric patient populations who are overweight/obese or diabetic, and as I highlighted in the introduction, this is a large proportion of people due to this bidirectional relationship between mental health and metabolic disorders.
Patients with pre-existing psychiatric disorders were excluded from this study. Reasonable decision or own goal?
Implications for practice
So, what can we take from this paper?
It will be very reassuring for clinicians managing patients receiving GLP1-RAs to know that a risk of psychiatric adverse effects has not been identified and that quality of life scales across a range of domains seem to improve with GLP1-RA use. I think, however, that is where the story ends.
There are some interesting questions raised by the authors in relation to the aetiology of these improvements and potential future work.
The improvement in relation to mental-health-related quality of life was not associated with the magnitude of weight loss or HbA1c change. So, is there a direct brain effect mediating the psychological benefits of GLP1-RA treatment? There have been areas of the brain identified with GLP1 effect and control of food intake, however, potential pathways for mental health benefits are as yet unknown.
Further work is required to clarify if these mental health associations are secondary to weight loss or reflect a direct neurobiological process and, importantly, to determine the impact of these drugs in psychiatric populations.
GLP1-RA may have their own independent benefit on mental health as well as metabolism
GLP1-RAs have been through an interesting journey to date in relation to their clinical uses and there may be wider applications in the future depending on how we need it…
Statement of interests
I have no conflicting interests to declare.
Links
Primary paper
Pierret ACS, Mizuno Y, Saunders P, et al. (2025) Glucagon-Like Peptide 1 Receptor Agonists and Mental Health: A Systematic Review and Meta-Analysis. JAMA Psychiatry. Published online May 14, 2025. doi:10.1001/jamapsychiatry.2025.0679
Other references
Nouwen A, Winkley K, Twisk J, et al; European Depression in Diabetes (EDID) Research Consortium. Type 2 diabetes mellitus as a risk factor for the onset of depression: a systematic review and meta-analysis. Diabetologia. 2010;53(12): 2480-2486. doi:10.1007/s00125-010-1874-x
