# Semaglutide and Cognition in Psychiatric Care: What a 13,007-Patient Cohort Found > **Author:** Benjamin Caplan, MD (ABFM Board Certified, MA Lic. #240562, NPI: 1962722637) > **Publisher:** CED Clinic (https://cedclinic.com) > **Canonical URL:** https://cedclinic.com/semaglutide-cognition-psychiatric-cohort/ > **Published:** 2026-08-21T08:51:05-04:00 | **Last Modified / Clinically Audited:** 2026-09-27T22:00:12-04:00 > **Clinical Summary:** semaglutide cognition psychiatric cohort: In a retrospective cohort of 13,007 adults with psychiatric diagnoses, semaglutide initiation was associated with f... --- CED Clinical Relevance #97 Clinical Evidence Update A peer-reviewed multicenter US EHR cohort compared 1,261 semaglutide initiators with several antidiabetic treatment strategies among 13,007 adults with psychiatric diagnoses. Clinical Insight | CED Clinic De Giorgi and colleagues studied 13,007 adults with psychiatric diagnoses in a multicenter US electronic health record dataset. Among 1,261 semaglutide initiators, clinician-recorded cognitive signs over 12 months were lower than with no antidiabetic treatment, glipizide, or empagliflozin, but not significantly lower than with sitagliptin after correction. This is a clinically relevant association, not evidence that semaglutide should be prescribed to improve cognition. Residual confounding, treatment selection, missing data, and routine documentation can all shape the estimate. Randomized trials using standardized cognitive and functional outcomes are needed before efficacy or clinical utility can be claimed. GLP-1SemaglutideCognitionMental HealthCohort Study Audience Patients, caregivers, clinicians, and cannabis-science readers interested in cognitive symptoms recorded during psychiatric care Primary Topic semaglutide and clinician-recorded cognitive signs in adults with psychiatric diagnoses Source [Read the full source](https://pubmed.ncbi.nlm.nih.gov/42624612/) In a retrospective cohort of 13,007 adults with psychiatric diagnoses, semaglutide initiation was associated with fewer clinician-recorded cognitive signs over 12 months than no antidiabetic treatment, glipizide, or empagliflozin, but not sitagliptin. The study cannot establish that semaglutide caused cognitive improvement or should be used as a cognitive treatment. What This Study Teaches Us The adjusted 12-month cognitive score was lower among semaglutide initiators than among people receiving no antidiabetic treatment, glipizide, or empagliflozin. The comparison with sitagliptin was not statistically significant after correction. Associations were most consistent in major depression, while the study did not demonstrate a causal treatment effect. Why This Matters Cognitive symptoms are common across major depression, psychosis, and [bipolar disorder](https://cedclinic.com/cannabis-induced-appetite-satiety-bypass-thc-brain/), yet treatment evidence is limited. A large routine-care dataset can identify a clinically relevant association, but prescribing decisions require randomized evidence, validated cognitive testing, and a complete assessment of metabolic and psychiatric risks. Study Snapshot Study Type Retrospective multicenter electronic health record cohort Population 13,007 US adults with a psychiatric diagnosis and routine clinician-rated cognitive assessments before and after antidiabetic treatment initiation Semaglutide Group 1,261 participants Mean Age 62.0 years; 48.4% female Data Source NeuroBlu Data, spanning 1999 to 2024 Comparators Sitagliptin, empagliflozin, glipizide, or no antidiabetic treatment Follow-Up 12 months after treatment initiation Primary Outcome A 0 to 100 composite of clinician-recorded signs across memory, attention, orientation, and other cognitive functions Main Results Semaglutide score 11.14 versus 14.91 with no treatment, 13.46 with glipizide, 13.69 with empagliflozin, and 12.52 with sitagliptin Adjusted Comparisons Mean ratios were 0.75 versus no treatment, 0.83 versus glipizide, 0.81 versus empagliflozin, and 0.89 versus sitagliptin Null Result The sitagliptin comparison was not statistically significant after Bonferroni correction Journal BMJ Mental Health Published August 20, 2026 PMID / DOI 42624612 / 10.1136/bmjment-2026-302828 Major Limitation Nonrandomized EHR data remain vulnerable to residual confounding, treatment-selection bias, missingness, and differences in routine documentation Clinical Bottom Line Semaglutide initiation was associated with fewer recorded cognitive signs than three of four comparison strategies. This is hypothesis-supporting real-world evidence, not proof that semaglutide improves cognition or treats psychiatric illness. Three Comparisons Favored Semaglutide Adjusted cognitive-sign scores were lower with semaglutide than with no treatment, glipizide, or empagliflozin. The estimates describe associations in routine clinical records and should not be read as randomized treatment effects. The Sitagliptin Comparison Was Null The mean ratio versus sitagliptin was 0.89 with a 95% confidence interval from 0.77 to 1.02. That comparison was not statistically significant after Bonferroni correction, so the paper did not show a consistent advantage over every active comparator. Major Depression Had the Most Consistent Signal Effect sizes were similar across diagnostic groups, but findings were only consistently significant in major depression. Subgroup consistency does not remove confounding or establish that semaglutide treats depression-related cognitive dysfunction. Routine Ratings Are Not a Neuropsychological Battery The outcome came from semistructured clinician-rated assessments documented during usual care. This can capture real-world signs, but it differs from standardized, blinded cognitive testing and may be influenced by visit patterns and documentation practices. Disclosures Matter Two authors reported employment and equity ownership in the company group connected with the NeuroBlu data platform, and one author reported consultancy fees from Holmusk. A disclosure does not invalidate the analysis, but independent replication is important when data access and commercial relationships intersect. How Strong Is This Evidence? The study used a large multicenter cohort, several active comparators, a prespecified cognitive composite, 12-month follow-up, covariate adjustment, and correction for multiple comparisons. Where This Paper Deserves Skepticism Treatment was not randomized, exposure groups may differ in unmeasured ways, routine cognitive documentation may be incomplete, the outcome is not equivalent to a blinded neuropsychological battery, and the abstract does not report a detailed adverse-event analysis. What This Paper Does Not Show The study does not show that semaglutide prevents dementia, reverses [cognitive impairment](https://cedclinic.com/acute-effects-partial-cb1-receptor-agonists-cognition/), treats depression, psychosis, or [bipolar disorder](https://cedclinic.com/cannabis-induced-appetite-satiety-bypass-thc-brain/), improves daily functioning, or is superior to sitagliptin. It also does not establish causality, a cognitive prescribing indication, or a complete safety profile. How This Fits With the Broader Clinical Conversation Observational GLP-1 research has produced signals across neurologic and psychiatric outcomes, but estimates vary with the population, comparator, indication, and outcome definition. For clinical care, cognition should be assessed alongside psychiatric symptoms, metabolic status, medication burden, sleep, substance exposure, nutrition, and functional change rather than inferred from one EHR association. Dr. Caplan’s Take *I find the size and active-comparator design useful, especially because the signal was not identical across all comparisons. The null sitagliptin result is part of the finding, not a detail to hide.* *This paper supports a better randomized question. It does not support prescribing semaglutide as a cognitive enhancer or changing a psychiatric treatment plan on the strength of routine-care associations.* What a Careful Reader Should Take Away A careful reader should retain both halves of the result: fewer recorded cognitive signs appeared with semaglutide versus three comparison strategies, while one active comparison was null and the nonrandomized design cannot establish treatment efficacy. Evidence Interpretation Guide Table of Contents - [How to Read a Cognitive Signal From Routine-Care Data](#How_to_Read_a_Cognitive_Signal_From_Routine-Care_Data)[A Four-Step Reading Frame](#A_Four-Step_Reading_Frame) - [The Same Study Can Mean Different Things Depending on the Question Being Asked](#The_Same_Study_Can_Mean_Different_Things_Depending_on_the_Question_Being_Asked)[Do Not Use This as a Cognitive Prescription](#Do_Not_Use_This_as_a_Cognitive_Prescription) - [The Comparator Pattern Is Informative](#The_Comparator_Pattern_Is_Informative) - [Residual Confounding Remains](#Residual_Confounding_Remains) - [Routine Ratings Have Measurement Limits](#Routine_Ratings_Have_Measurement_Limits) - [The Signal Extends Earlier Neuropsychiatric Work](#The_Signal_Extends_Earlier_Neuropsychiatric_Work) - [Cognition Has Many Competing Explanations](#Cognition_Has_Many_Competing_Explanations) - [Randomized Cognitive Outcomes Are Needed](#Randomized_Cognitive_Outcomes_Are_Needed) - [Avoid Off-Label Cognitive Marketing](#Avoid_Off-Label_Cognitive_Marketing) - [Frequently Asked Questions](#Frequently_Asked_Questions)[Related](#Related) ## How to Read a Cognitive Signal From Routine-Care Data Large EHR cohorts can reveal patterns that deserve testing. They cannot reproduce the protection against confounding provided by random assignment and blinded outcome assessment. ### A Four-Step Reading Frame **Design**Identify this as a retrospective cohort, not a randomized trial. **Comparator**Keep all four comparisons visible, including the null sitagliptin result. **Outcome**Distinguish clinician-recorded cognitive signs from standardized neuropsychological performance and daily function. **Causality**Treat the estimates as associations that require randomized confirmation. The Research Question Was semaglutide initiation associated with fewer clinician-recorded cognitive signs over 12 months? The Patient Question Does this mean semaglutide will improve my memory or psychiatric symptoms? The Bottom Line The cohort supports further study but does not establish a cognitive treatment effect. CED Perspective Lens ## The Same Study Can Mean Different Things Depending on the Question Being Asked Scientific papers rarely answer a single question. Patients, clinicians, researchers, policymakers, and critics often read the same data differently. The perspectives below explore how this study looks through several evidence-based lenses. Lens Overview Eight perspectives help separate a useful observational signal from a cognitive treatment claim. #### Do Not Use This as a Cognitive Prescription The study does not show that semaglutide will improve an individual patient’s memory, attention, or daily function. Interpretation should remain matched to this retrospective 12-month cohort. Lens takeaway Medication choices should remain tied to established indications and individualized care. #### The Comparator Pattern Is Informative Three adjusted comparisons favored semaglutide, while sitagliptin did not. Interpretation should remain matched to this retrospective 12-month cohort. Lens takeaway Report the full pattern instead of compressing it into a positive headline. #### Residual Confounding Remains Metabolic severity, prescribing decisions, healthcare contact, and undocumented clinical differences may influence both exposure and recorded cognition. Interpretation should remain matched to this retrospective 12-month cohort. Lens takeaway Adjustment reduces known imbalance but cannot remove every source of bias. #### Routine Ratings Have Measurement Limits Semistructured clinical assessments are pragmatic, but they are not blinded standardized cognitive batteries. Interpretation should remain matched to this retrospective 12-month cohort. Lens takeaway Replication should include objective cognition and functional outcomes. #### The Signal Extends Earlier Neuropsychiatric Work Earlier observational and trial reports have produced mixed domain-specific findings rather than a settled cognitive effect. Interpretation should remain matched to this retrospective 12-month cohort. Lens takeaway The new cohort adds scale, not causal certainty. #### Cognition Has Many Competing Explanations Depression severity, psychosis, sleep, metabolic disease, other medicines, and substance exposure can affect cognitive complaints. Interpretation should remain matched to this retrospective 12-month cohort. Lens takeaway A full clinical assessment matters more than attributing change to one drug. #### Randomized Cognitive Outcomes Are Needed Trials should prespecify validated cognitive tests, functioning, psychiatric symptoms, metabolic mediators, adverse events, and durability. Interpretation should remain matched to this retrospective 12-month cohort. Lens takeaway The next step is causal testing with patient-important outcomes. #### Avoid Off-Label Cognitive Marketing An observational association should not be converted into a promotional claim for cognitive enhancement. Interpretation should remain matched to this retrospective 12-month cohort. Lens takeaway Accurate design labels protect patients and scientific credibility. Join the Conversation Have a question about how this applies to your situation? [Ask Dr. Caplan](https://benjamincaplan.com/ask-dr-caplan) Want to discuss this topic with other patients and caregivers? [Join the forum discussion](https://cedclinic.com/interact/discussion-forums/) [https://cedclinic.com/contact/](https://cedclinic.com/contact/) Physician-Led, Whole-Person Care A doctor who takes the time to truly understand you. Personal care that starts with listening and is guided by experience and ingenuity. Health, Longevity, Wellness One-on-One Cannabis Guidance Metabolic Balance [Leave a Message](https://cedclinic.com/contact/) [Metabolic Care](https://cedclinic.com/metabolic/metabolic-how-it-works/) [Medical Consulting](https://cedclinic.com/book/) [Cannabis Care](https://cedclinic.com/book/) [https://cedclinic.com/contact/](https://cedclinic.com/contact/) Share this Evidence Watch [X](https://twitter.com/intent/tweet?text=Semaglutide+and+Cognition+in+Psychiatric+Care%3A+What+a+13%2C007-Patient+Cohort+Found&url=https%3A%2F%2Fcedclinic.com%2Fsemaglutide-cognition-psychiatric-cohort%2F) [LinkedIn](https://www.linkedin.com/sharing/share-offsite/?url=https%3A%2F%2Fcedclinic.com%2Fsemaglutide-cognition-psychiatric-cohort%2F) [BlueSky](https://bsky.app/intent/compose?text=Semaglutide+and+Cognition+in+Psychiatric+Care%3A+What+a+13%2C007-Patient+Cohort+Found%20https%3A%2F%2Fcedclinic.com%2Fsemaglutide-cognition-psychiatric-cohort%2F) [Instagram](https://www.instagram.com/drcaplan/) [Substack](https://doctorapprovedcannabis.substack.com/) [RSS](https://cedclinic.com/feed/) Source: [Cognitive signs and symptoms in people with a psychiatric diagnosis on semaglutide: a retrospective cohort study of 13,007 patients in the USA.](https://pubmed.ncbi.nlm.nih.gov/42624612/) Related Reading at CED Clinic Continue exploring the evidence [Clinical Evidence on Emotional Effects of GLP-1 Receptor Agonists](https://cedclinic.com/glp-1-cannabis-use-disorder-bmj-cohort-study/) Earlier CED context on mood, emotional experience, and the limits of current GLP-1 evidence. [Compare emotional-effects evidence](https://cedclinic.com/glp-1-cannabis-use-disorder-bmj-cohort-study/) [GLP-1 Receptor Agonist Clinical Evidence: New Findings](https://cedclinic.com/glp-1-receptor-agonist-clinical-trial-evidence/) A broader review of clinical-trial evidence and emerging safety questions across GLP-1 therapy. [Review broader clinical evidence](https://cedclinic.com/glp-1-receptor-agonist-clinical-trial-evidence/) [GLP-1 Receptor Agonist Clinical Evidence: Mental Health Links](https://cedclinic.com/glp-1-receptor-agonist-mental-health-benefits/) Earlier discussion of mental-health associations that helps frame the new cognition-specific cohort. [Read earlier mental-health context](https://cedclinic.com/glp-1-receptor-agonist-mental-health-benefits/) Similar Studies on CEDClinic.com Earlier coverage on closely related evidence When a new paper overlaps with earlier CED Clinic coverage, we preserve the chain instead of hiding the overlap. These links point to older related posts so readers can compare what is new, what is repeated, and how the evidence has moved. Related digest item [Clinical Evidence on Emotional Effects of GLP-1 Receptor Agonists](https://cedclinic.com/glp-1-cannabis-use-disorder-bmj-cohort-study/) GLP-1 receptor agonist emotional effects include reports of mood changes and emotional blunting in patients. Clinical evidence highlights the importance of monitoring these effects during treatment. Understanding these… Covers a closely related topic; this CED post is older by about 101 days and gives earlier context. [Compare earlier coverage](https://cedclinic.com/glp-1-cannabis-use-disorder-bmj-cohort-study/) Related digest item [GLP-1 Receptor Agonist Clinical Evidence: New Findings](https://cedclinic.com/glp-1-receptor-agonist-clinical-trial-evidence/) GLP-1 receptor agonist clinical trial evidence provides critical insights into safety and emerging [side effects](https://cedclinic.com/cannabis-intoxication-memory-disruption/). This evidence supports clinicians in monitoring patient outcomes effectively. Understanding these findings… Covers a closely related topic; this CED post is older by about 122 days and gives earlier context. [Compare earlier coverage](https://cedclinic.com/glp-1-receptor-agonist-clinical-trial-evidence/) Related digest item [GLP-1 Receptor Agonist Evidence: Obesity Drug Breakthroughs](https://cedclinic.com/glp-1-cannabis-use-disorder-bmj-cohort-study/) GLP-1 receptor agonist neuropsychiatric effects are emerging as an important consideration in obesity treatment. These medications influence mood, behavior, and mental health outcomes beyond [weight management](https://cedclinic.com/cannabis-metabolism-weight-management-science/). Clinician… Covers a closely related topic; this CED post is older by about 125 days and gives earlier context. [Compare earlier coverage](https://cedclinic.com/glp-1-cannabis-use-disorder-bmj-cohort-study/) Related digest item [New Study Reveals THC Induced False Memories and Reality Distortion](https://cedclinic.com/thc-induced-false-memories-cannabis-memory-distortion/) A new study highlights THC induced false memories as a significant form of cannabis memory distortion. This [cognitive impairment](https://cedclinic.com/acute-effects-partial-cb1-receptor-agonists-cognition/) affects accurate recall and patient safety. Clinicians should screen for these effects dur… Covers a closely related topic; this CED post is older by about 129 days and gives earlier context. [Compare earlier coverage](https://cedclinic.com/thc-induced-false-memories-cannabis-memory-distortion/) Related digest item [Editorial: New York’s Cannabis-Related Toxicity Challenges](https://cedclinic.com/managing-cannabis-related-toxicity-new-york/) Cannabis-related toxicity is an emerging concern following New York’s legalization of recreational cannabis. Rising emergency department visits and poison control calls highlight the need for updated clinical protocols…. Covers a closely related topic; this CED post is older by about 130 days and gives earlier context. [Compare earlier coverage](https://cedclinic.com/managing-cannabis-related-toxicity-new-york/) Related digest item [GLP-1 Receptor Agonist Clinical Evidence: Mental Health Links](https://cedclinic.com/glp-1-receptor-agonist-mental-health-benefits/) Emerging research highlights GLP-1 receptor agonist mental health benefits beyond metabolic effects. Semaglutide shows promise in improving mood and psychiatric outcomes. Clinicians should consider these findings in pat… Covers a closely related topic; this CED post is older by about 131 days and gives earlier context. [Compare earlier coverage](https://cedclinic.com/glp-1-receptor-agonist-mental-health-benefits/) Related digest item [Daily Digest: Cognitive Safety, End-of-Life Cannabis Access, and Regulatory Challenges — March 06, 2026](https://cedclinic.com/end-of-life-cannabis-access-regulatory-gaps-2026/) End-of-life [cannabis access](https://cedclinic.com/medical-cannabis-delivery-services-growth-dmv-jamaica/) remains a critical issue for patients and healthcare providers. This digest highlights regulatory gaps and hospital policies impacting terminally ill patients. Cognitive safety in older adult… Covers a closely related topic; this CED post is older by about 167 days and gives earlier context. [Compare earlier coverage](https://cedclinic.com/end-of-life-cannabis-access-regulatory-gaps-2026/) ## Frequently Asked Questions Did semaglutide improve cognition in this study? The study found lower clinician-recorded cognitive-sign scores in several adjusted comparisons. Because treatment was not randomized, it cannot establish that semaglutide caused improvement. How many people were included? The cohort included 13,007 adults with psychiatric diagnoses, including 1,261 who initiated semaglutide. Which comparisons favored semaglutide? Scores were lower versus no antidiabetic treatment, glipizide, and empagliflozin in the adjusted analyses. Was semaglutide better than sitagliptin? No statistically significant difference was detected after correction for multiple comparisons. What did the cognitive score measure? It combined clinician-recorded signs across memory, attention, orientation, and other cognitive functions on a 0 to 100 scale. Does the result apply to dementia prevention? No. The study did not establish dementia prevention or reversal of [cognitive impairment](https://cedclinic.com/acute-effects-partial-cb1-receptor-agonists-cognition/). Did the study prove benefit for depression? No. Associations were most consistent in major depression, but subgroup findings from a retrospective cohort do not prove treatment efficacy. Were adverse events fully assessed? The PubMed abstract does not provide a detailed adverse-event analysis, so this paper should not be used as a complete safety assessment. Why can residual confounding matter? People receiving different diabetes treatments may differ in health status, access, prescribing context, and other factors that are incompletely captured in EHR data. What is the practical takeaway? The finding justifies randomized study, not semaglutide use as a cognitive enhancer or unsupervised medication changes. ### Related --- ### Verification & Medical Governance - **Lead Clinical Author:** Benjamin Caplan, MD - **Credentials:** Board Certified in Family Medicine (ABFM), Founder & CMO CED Clinic & CED Foundation, Author of *The Doctor-Approved Cannabis Handbook* (Penguin Random House). - **Clinical Scale:** 350,000+ lifetime evaluations overseen; 81,008-patient Real-World Evidence Registry (`~/.ced_rwe_vault.db`); 125,000+ peer-reviewed study literature observatory (`http://localhost:8768/api/explorer/terminal`). - **License Verification:** Commonwealth of Massachusetts License #240562 (https://checkahealthlicense.mass.gov/). - **NPI Registry:** 1962722637 (https://npiregistry.cms.hhs.gov/provider-view/1962722637). - **Ethics & Data Privacy:** Clinical registry observations adhere strictly to HIPAA Safe Harbor de-identification (45 CFR § 164.514(b)) and criteria for secondary research exemption under 45 CFR § 46.104(d)(4).