Semaglutide, suicidal ideation and behaviour: A resting state functional magnetic resonance imaging perspective

In recent months, there has been a lot of talk about the use of glucagon-like peptide 1 (GLP-1) receptor agonists, not only in type 2 diabetes but also for its effects on weight loss. Because of its wide use for this purpose and recent reports of suicidal thoughts and self-injury in patients taking liraglutide and semaglutide, GLP-1 receptor agonists have also come under the scrutiny of the European Medicines Agency (EMA), which has launched an investigation into these new adverse events. There is a wealth of research on brain changes in major depressive disorder (MDD) and suicidal ideation and behaviour. Previous research has highlighted changes in resting state functional connectivity (RSFC) in brain regions thought to be critical for emotion regulation and impulsivity, such as the amygdala, orbitofrontal cortex and anterior cingulate cortex.1, 2 A recent review of neuroimaging evidence for suicidal thoughts and behaviours in MDD highlighted the importance of prefrontal cortices and their interplay with the limbic system.3 Research on resting state networks has also focused on the aforementioned regions, with particular attention to changes in and between default mode, salience and frontoparietal networks in patients with suicidal ideation and attempts.4 Patients with MDD with suicidal ideation and/or behaviour show declines in RSCF in mentioned areas compared with patients with MMD without suicidal ideation and/or behaviours.2-4 Given recent reports of adverse effects and interest in GLP-1 receptor agonists, we compared RSFC in brain regions identified in previous studies as important for suicidal ideation and behaviour in a 16-week, placebo-controlled study comparing the semaglutide intervention group with the placebo group. The goal of this study was to evaluate if there is any indication that semaglutide intervention could have an effect on RSFC that would reflect the change observed in suicidal ideation/behaviour. The data for this study were a part of the clinical trial registered at ClinicalTrials.gov NCT04263415, which has been described elsewere.5 Briefly, 30 women with obesity and polycystic ovary syndrome were enrolled in a randomized, single-blind, placebo-controlled 16-week study where they were administered either 1.0 mg of semaglutide or placebo. In addition, patients underwent magnetic resonance imaging (MRI) at the Center for Clinical Physiology, University of Ljubljana, Slovenia. Before enrolment in the study, participants gave informed consent. In addition, the study was approved by the Slovene National Medical Ethics Committee and was conducted in accordance with the Declaration of Helsinki and Good Clinical Practice guidelines. Regions of interest (ROI) analysis was done on ROIs defined by the Brainnetome6 atlas (listed in Table S1 in Appendix S1) based on regions linked to suicidal ideation and behaviour in MDD from previous research.1-4 After quality control and preprocessing, and addressing the missing data by excluding participants that had large amount of movement artefacts or missing sessions, RSFC data from 18 participants were included in the analysis (10 in the semaglutide and eight in the placebo group). Of the 30 subjects, nine had >20% of the volumes with excessive motion in the MRI data and three had data for only one of the MRI sessions and were therefore not included in the analysis. First level ROI correlation analysis and subsequent multivariate group analysis were performed using AFNI programs. Comparison between the groups was done using a paired t-test, using a significance level of .05. Further details on the doses of semaglutide administered and on MRI acquisition, preprocessing and analysis using AFNI programs are described in the Supplementary Appendix S1. Demographic and clinical data for the 18 participants included in the analysis are displayed in the Table S2 in Appendix S1. Our analysis showed that the change in RSFC between visits did not differ significantly between the groups [χ2 (df = 1, N = 18) = 0.104, p = .747]. Looking at visit differences in RSFC, the results showed significant differences between visits [χ2 (df = 1, N = 18) = 27.609, p < .001]. More specifically, the semaglutide group mostly showed higher RSFC after the intervention (Table S3 in Appendix S1), whereas the placebo group mostly showed lower RSFC in a selection of investigated ROIs (Table S4 in Appendix S1). These results are included in the Supplementary Appendix (Tables S2-S4 in Appendix S1). Our results show that there is no significant change in RSFC after our semaglutide intervention in participants who received semaglutide compared with the placebo group in the investigated ROIs. This result is encouraging given the reports of adverse effects that have been reported to the EMA. However, upon closer examination of the results, we note that there are some significant changes in RSFC in the ROIs associated with suicidal ideation and behaviour in our sample. After the intervention, the semaglutide group showed higher connectivity compared with the placebo group, which had lower RSFC in some of the regions studied (Tables S2 and S3 in Appendix S1). The direction of this change in connectivity was in the opposite direction (i.e. connectivity between regions was increased), compared with what previous studies on suicidal ideation and behaviour have found in patients with MDD.2-4 This decrease in RSFC is often explained by impaired decision making and emotional processing. However, we believe that the mechanism behind the semaglutide intervention in our study may be different, mainly because of the characteristics of our sample; sample included patients with polycystic ovary syndrome and obesity who did not have a history of MDD and MDD was not assessed. In obesity research, the changes in functional connectivity between salience regions (cingulate cortex, insula, amygdala, precuneus) have been associated with reward processing,7 which could be altered during the semaglutide intervention. The increase in RSFC could also be because of the improved clinical outcomes (reduced weight and waist circumference) observed in the semaglutide group in the clinical trial, and this positive change could therefore have an impact on greater interoceptive self-awareness.8 However, we must be cautious in interpreting our results because this study has several limitations. Although the clinical trial included a substantial number of participants, only 18 were included in this study because of excessive movement during MRI and some missing data, resulting in eight and 10 participants per group. The limited number of participants and their characteristics might be the reason why we found significant differences in RSFC at visit 1. In addition, participants were not assessed for MDDs or other emotional disorders, and their mental status, or possible suicidal thoughts or behaviours were not assessed during or after the clinical trial and therefore cannot be correlated with RSFC. For this reason, we can only speculate about the reasons for the observed change. Further studies should look into this relationship more closely, assessing MDD and connected behaviours to be able to draw conclusions about the relationship between MDD, RSFC and semaglutide treatment. We also believe that the lower dose may also play a role in the results of this study. In contrast to the typical dose of 2.4 mg per week9 for the treatment of weight loss, the clinical trial used 1 mg doses. The reason behind the lower dosage is because at the time of the clinical trial, the semaglutide dosage for weight control had not yet been approved and the usual dosage for glycaemic control was used. Therefore, based on the results of this study, we believe that the lower dose of semaglutide for the weight loss application has no observed adverse effects that are currently under investigation by the EMA. The results also show an increase in RSFC in the intervention group after the 16-week trial, which could be interpreted as a secondary effect of semaglutide (possibly because of improved reward processing and self-perception). It is important to emphasize that further research is needed, particularly the inclusion of psychological assessments to correlate RSFC with depressive symptoms and behaviours would further illuminate the relationships between RSFC changes, semaglutide and suicidal thoughts/behaviours. Furthermore, the underlying mechanism of these changes is unknown and needs further investigation. We would like to thank our colleagues at the Department of Endocrinology, Diabetes and Metabolic Diseases, Division of Internal Medicine, University Medical Centre Ljubljana for recruiting the participants and organizing the data collection. Our special thanks go to Professor Dr Andrej Janež and Dr Mojca Jensterle for providing resources and their collaborative spirit. The authors have no conflicts of interest to declare. The peer review history for this article is available at https://www.webofscience.com/api/gateway/wos/peer-review/10.1111/dom.15363. The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions. Appendix S1. Supporting Information. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.