Neuropsychiatry

The distinctive psychopathology of NMDAR-antibody encephalitis compared with primary psychoses: an international, multicentre, retrospective phenotypic analysis
Adam Al-Diwani, Jakob Theorell, Tarek Zghoul, Aniruddha Voruganti, Leigh Townsend, Riccardo De Giorgi, Benjamin Griffin, Tomasz Bajorek, David Okai, Belinda Lennox, M Isabel Leite, Carla Y Kim, Arielle Coughlin, Kelsey Martin, Brittany Glassberg, Christian Lachner, Nicola Westerbeek, Veerle Bergink, Kiran T Thakur, Anusha K Yeshokumar, Harald Prüss, Gregory S Day, Carsten Finke, Adam E Handel, Sanjay G Manohar, Dan W Joyce, Sarosh R Iran

Abstract:

Background: N-methyl-D-aspartate receptor (NMDAR)-antibody encephalitis is a life-threatening neuropsychiatric disorder requiring prompt immunotherapy. The earliest features are mental-state changes, often mistaken for primary psychosis. Improved clinical differentiation could assist rational diagnostic investigation and expedite immunotherapy. Inspired by patients' and relatives' lived experience, we aimed to explore the psychiatric phenotype of NMDAR-antibody encephalitis and the features common to and distinct from real-world episodes of psychosis.

Methods: In this international, multicentre, retrospective phenotypic analysis we collected data on episodes of NMDAR-antibody encephalitis from specialised neurology services in Europe (UK, Germany, and Sweden) and the USA. For comparison, we collected similar data from de-identified accepted referrals to a UK early intervention in psychosis service, including consecutively presenting cases (unselected psychosis) and a group defined by having been assessed and admitted to hospital under the Mental Health Act (selected psychosis). Additionally, we included episodes of postpartum psychosis from a mother and baby unit in the Netherlands. In our mental-state inventory we included core features from ICD-11, the Bush-Francis catatonia score, and the Neuropsychiatric Inventory encompassing anxiety, depression, mania, schizophrenia, catatonia, and also more granular transdiagnostic behavioural features including those common to neuropsychiatric and neurobehavioural syndromes, such as delirium and dementia. Ethnicity data were not available. We compared and visualised the neuropsychiatric phenotype of these cohorts.

Findings: We collected data from 100 episodes of NMDAR-antibody encephalitis from 96 patients between 2010 and 2022 (median age 22 years, female:male ratio 3·80), 135 episodes of psychosis from 135 patients between 2018 and 2019 (median age 27 years, female:male ratio 0·75), and ten episodes of postpartum psychosis from ten patients between 2005 and 2012 (median age 30 years, all female sex). Psychopathology in NMDAR-antibody encephalitis was abundant (92 [92%] of 100 episodes) and ultra-rapid in onset (median 1 day [95% CI 1-1; IQR 1-7]) versus unselected primary psychoses (median 180 days [120-210; 91-365]; p<0·0001). 21 (36%) of 58 mental-state features, including catatonic and visual hallucinations, were over-represented in NMDAR-antibody encephalitis and 12 (21%) were under-represented, including features typical of affective (eg, elated mood, flight of ideas, grandiose delusions) and non-affective psychoses (eg, thought broadcasting, thought withdrawal, paranoid delusions; false discovery rate threshold <0·05). Typically, in NMDAR-antibody encephalitis, the complexity sequentially evolved from mood to psychotic to catatonic predominance within 2 weeks.

Interpretation: NMDAR-antibody encephalitis has a rapid-onset, complex, and dynamic neuropsychiatric phenotype, sufficiently distinctive to drive a clinical approach to differentiation.

Funding: UK NIHR, Wellcome, and UK Medical Research Council (MRC)/UK Research and Innovation.

Annotation

The finding: This was a retrospective analysis using multiple patient samples to identify clinical features that could distinguish psychosis due to anti-NMDA receptor encephalitis from that due to a primary psychiatric condition. 100 patients with encephalitis were evaluated across multiple sites in Europe and the US and compared to 135 patients with first-episode psychosis (52 of whom were involuntarily hospitalized) and 10 with postpartum psychosis. Several notable differences were identified in the patients with anti-NMDA receptor encephalitis: they were more likely female (80% vs. ~45%); had rapid-onset symptoms (mostly <7 days vs. ~180 days); and had more frequent catatonia and visual hallucinations, but less delusions and mood symptoms. The authors developed several simple clinical decision rules to identify anti-NMDA receptor encephalitis with sensitivity of >98% and specificity of 90-98%.

Strength and weaknesses: The international multicenter design and the detailed 58-item symptom assessment represent notable strengths, alongside creation of a practical diagnostic tool. Limitations include its retrospective design, reliance on chart review for psychosis comparators, modest absolute sample sizes, lack of ethnicity data, and limited generalizability to non-specialist settings. To minimize bias, the optimal way to perform this analysis would be to routinely test for anti-NMDA receptor antibodies in thousands of patients with first-episode psychosis, then compare the features of those with positive tests to those with negative tests; otherwise, we risk simply reconfirming those features that led us to test for autoimmune encephalitis in the first place. But considering that the above approach is quite financially and clinically difficult, this study offers a solid alternative.

Relevance: For C-L psychiatrists, this study offers concrete clinical guidance for differentiating NMDAR-antibody encephalitis from primary psychosis at bedside. The emphasis on quick onset, atypical psychotic features, and early catatonia or delirium-like presentations supports a clinically driven approach to raising diagnostic suspicion and advocating for CSF testing and neurologic evaluation. As psychiatric symptoms often precede overt neurologic findings and patients are frequently first admitted to psychiatric units, this work directly informs consultation practice, clinically grounded diagnostic advocacy, and timely escalation to immunotherapy. Of note, the tools in question are likely to be fairly accurate in ruling out anti-NMDA receptor encephalitis (NPV>99%) but less accurate in ruling it in (PPV 10-50%).

Management of Anxiety in Parkinson's Disease
Alex J Berry, Harry Costello, Silvia Jesús, Gary Price, Ashwani Jha

Abstract:

Background: Anxiety is a common, distressing, hard-to-diagnose and hard-to-treat symptom in Parkinson's disease. No formal guidelines exist to assist management.

Objective: We provide a pragmatic guide to detecting and managing anxiety in Parkinson's disease.

Methods: In this educational review, we describe the phenomenology, diagnostic challenges, hypothesized neurobiology, and the rationale for treatments for anxiety in Parkinson's disease. We review the major drug-classes and non-pharmacological treatment approaches in current use. We also present a meta-analysis of pharmacological treatments of anxiety derived from previously published systematic reviews of RCTs for depression in Parkinson's disease in which anxiety was a secondary outcome.

Results: In our meta-analysis, anxiolytic medications showed a moderate overall anxiolytic effect compared to placebo (standardized mean difference: -0.45 [95% CI: -0.78, -0.12], p = 0.02). There were significant limitations with the studies included in this meta-analysis, with studies generally having small cohort sizes, and each specific pharmacotherapy was not studied in more than one randomized control trial. We also describe a pragmatic algorithm for individualized pharmacological management, based on our own experience of selecting treatments to optimize side-effect profile and treatment of comorbid symptoms.

Conclusions: Detecting and treating anxiety likely benefits people with Parkinson's disease, though the current evidence-base for specific treatments or specific pharmacotherapy remains limited. Further work is needed to investigate the different evidence-based approaches for managing anxiety in Parkinson's disease.

Keywords: Parkinson's disease; anxiety; meta‐analysis; neuropsychiatry; non‐motor.

Annotation

The finding: This article includes a review of the major treatments for anxiety in Parkinson’s disease (PD) and a meta-analysis of the anxiolytic pharmacotherapies. Using published systematic reviews to identify studies for the meta-analysis, the authors incorporated five placebo controlled RCTs in which PD anxiety was a secondary outcome measure. Seven total medications (citalopram, desipramine, rasagiline, paroxetine, venlafaxine, atomoxetine, buspirone) across five pharmacological classes were included. Compared to placebo, pharmacological treatments had an overall moderate effect in PD anxiety (standardized mean difference (SMD) -0.45 [95% CI -0.78, -0.21], p=0.02). Treatment with citalopram and atomoxetine significantly favoured the treatment arm (SMD -1.06 [-1.87, -0.25] and -0.59 [-1.13, -0.05], respectively).  In the review, the authors covered additional pharmacotherapies (mirtazapine, trazodone, bupropion, agomelatine, benzodiazepines, quetiapine, pregabalin, and cannabinoids) as well as non-pharmacological approaches. Based on their findings, they constructed a proposed treatment algorithm for anxiety in PD. Different approaches were recommended for anxiety occurring in combination with other clinical features.

Strength and weaknesses: This article’s strengths include the thorough review of anxiolytic medication classes used in PD anxiety, incorporation of a meta-analysis with quantification of a therapeutic effect size, and creation of a practical algorithm to guide a pharmacological approach. However, since only prior systematic reviews were used, potentially relevant RCTs may not have been captured. Additionally, the meta-analysis was based solely on secondary outcome measures in the studies, the RCTs had small sample sizes, and each medication included was studied only in a single RCT, limiting generalizability and reproducibility.

Relevance: Anxiety is common in PD and causes considerable morbidity. Yet, trials focused on its treatment in this patient population are lacking, and there is insufficient literature to guide clinicians. The consultation-liaison psychiatrist is often called upon to aid in the diagnosis and management of anxiety in PD, which can be particularly challenging due to overlapping features in PD and high rates of neuropsychiatric comorbidities. Therefore, an informed approach is essential. This review calls for more rigorous studies examining the efficacy of anxiety treatments in PD to inform evidence-based management guidelines for this difficult to treat condition.

A randomized controlled trial of the safety and efficacy of dronabinol for agitation in Alzheimer’s disease
Paul B Rosenberg, Halima Amjad, Haroon Burhanullah, Milap Nowrangi, Ryan Vandrey, Mersania Jn Pierre, John D Outen, Meghan Schultz, Christopher Marano, Marc Agronin, James M Wilkins, David Harper, Todd Laffaye, Eilis Reardon, Kathryn Turner, Rosain Ozonsi, Mia Drury, Andre Nguyen, Tuna Hasoğlu, Julia Cromwell, Jeannie-Marie Leoutsakos, Brent P Forester

Abstract:

Importance: Agitation in Alzheimer's disease (AD) is a great source of distress for patients and caregivers and a major public health burden. Current treatments are only modestly effective and many have safety issues including mortality risk. Novel therapeutic options are needed. There is preliminary evidence for the safety and efficacy of dronabinol (tetrahydrocannabinol, THC) for agitation in AD.

Objective: Assess the safety and efficacy of dronabinol (THC) to decrease agitation in AD.

Design: THC-AD was a 3-week randomized parallel double-blind placebo-controlled clinical trial, conducted between 2017 and 2024.

Setting: 5 inpatient and outpatient academic clinical research centers in the Eastern U.S.

Participants: Volunteer sample of 75 participants meeting inclusion criteria for agitation of AD (International Psychogeriatric Association Provision Criteria) with Neuropsychiatric Inventory Clinician Version Agitation or Aggression (NPI-C A/A) domains total score of 4 or greater. Major exclusion criteria included seizure disorder, delirium, and non-AD dementia.

Interventions: 3 weeks dronabinol vs. placebo titrated up to target dose of 10 mg daily in divided twice-daily.

Main outcomes and measures: Prespecified co-primary agitation outcomes were the Pittsburgh Agitation Scale (PAS) and NPI-C A/A total score.

Results: The majority of participants were female and were taking concomitant psychotropic medications (antidepressants and antipsychotics) at baseline. Study participants were moderately agitated at baseline, were diverse in ethnic background (9% Black, 11% Hispanic/Latina/Latino), and had severe cognitive impairment evidenced by MMSE or SIB-8. 84% completed the 3-week trial. Dronabinol decreased agitation on both primary outcomes greater than placebo to a clinically relevant extent. The fitted between-arm difference in PAS decline/week was -0.74 (SE 0.3, p = 0.015, effect size = 0.53) and for NPI-C A/A the decline was not significant at -1.26 (SE 0.67, p = 0.094, effect size = 0.36). No secondary outcomes differed between treatment arms including sleep, activities of daily living, Cohen-Mansfield Agitation Inventory (CMAI), cognition, intoxication, or use of 'as-needed' lorazepam or trazodone. Dronabinol treatment was not associated with greater intoxication nor with other adverse events (AEs) except for somnolence.

Conclusions and relevance: Adjunctive dronabinol treatment was safe and effective for treating agitation in AD.

Annotation

The finding: This was a randomized controlled trial comparing the synthetic cannabinoid dronabinol (2.5 mg at 8 am and 2 pm for one week, then 5 mg at 8 am and 2 pm) vs. placebo for treatment of agitation in 75 adults with dementia due to Alzheimer’s disease. The patients were mostly inpatients on psychiatric units, but some outpatients were also enrolled; had severe dementia (mean MMSE 8.6); and had an average age around 80. Of the co-primary endpoints, one showed a significant beneficial effect (reduction in the Pittsburgh Agitation Scale, with an effect size of 0.53) and one showed a nonsignificant beneficial effect (reduction in NPI agitation subscale, with an effect size of 0.36). No dramatic differences in adverse effects were noted, but dronabinol appeared to cause increased rates of somnolence, with possible (though far from definitive) increases in falls and delirium.

Strength and weaknesses: This study took nearly a decade to complete, which is a testament to the challenges of conducting trials in this population, along with the added disruption of the COVID pandemic – so merely having completed an RCT of a novel medication for agitation is a major strength and laudable accomplishment. Weaknesses include the relatively small sample size and the mixture of inpatient and outpatient settings.

Relevance: Agitation in patients with Alzheimer’s disease is a common reason for psychiatric consultation, and is often challenging to manage pharmacologically, so adding one more medication to your armamentarium is no doubt a welcome development, even if it’s not dramatically effective. For well-selected patients – avoiding those already dealing with excess daytime sedation and prioritizing those with other indications for dronabinol such as loss of appetite – it’s a reasonable option to trial that should be fairly well-tolerated.

Prevalence of clinically significant radiological abnormalities in people with first episode psychosis
Thomas J Reilly, Matt Butler, Ahmed Abdelsamie, Ishaac Awatli, Sneha Barai, Luke Baxter, Theo Boardman-Pretty, Rebecca Phelps, Anam Waqar, Amelia Jewell, Graham Blackman, Dominic Oliver, Ashwin V Venkataraman, Matthew J Kempton, Philip McGuire

Annotation

The finding: This was a retrospective analysis of MRI findings in a sample of ~25,000 first-episode psychosis patients in the UK, ~1500 of whom had a brain MRI performed. Via detailed chart review, the authors were able to determine both the indication for the MRI as well as its consequences in most cases. Among patients with no specified indication for MRI, or where the indication was “routine screening”, 16% of MRIs identified some abnormality and 4% of them led to some change in management – most commonly this was referral to another specialty, and none involved a change in diagnosis or identifying a secondary cause of psychosis. By contrast, when there was any specific indication for the MRI (including cognitive impairment, head injury, “atypical psychosis presentation”, suspected encephalitis, and several others), 41% of MRIs were abnormal, 17% led to a change in management, and 6% led to a change in diagnosis. Beyond indication, the one other factor that predicted management-altering MRIs was age – in patients over 65, three-quarters of MRIs were abnormal and a third led to changes in management.

Strength and weaknesses: The devil’s in the details with this sort of study, as not all imaging abnormalities or changes in management are created equal. But this study does about as good a job as possible of illuminating those details and helping the reader distinguish between more vs. less meaningful MRI findings.

Relevance: For patients with new psychosis and no obvious neurologic abnormalities, there’s immense debate and clinical uncertainty on if/when to order a brain MRI. This study suggests that the yield of such imaging is not insignificant. Even if you just order MRIs blindly in all patients with new-onset psychosis, you’ll identify an alternative cause in 2% of cases – and if you restrict yourself to instances where there’s any additional indication for MRI (even vague indications such as an atypical presentation of psychosis), you’ll change your patient’s diagnosis in 6% of cases. While I wouldn’t advocate for uniformly ordering MRIs in every newly psychotic patient, this study should provide some reassurance that you won’t be wasting resources if you check an MRI anytime you have an inkling that there’s something unusual about a patient’s presentation of psychosis.