The role of the electroencephalogram (EEG) in determining the aetiology of catatonia: a systematic review and meta-analysis of diagnostic test accuracy

Background Catatonia is a psychomotor syndrome that has a wide range of aetiologies. Determining whether cata-tonia is due to a medical or psychiatric cause is important for directing treatment but is clinically challenging. We aimed to ascertain the performance of the electroencephalogram (EEG) in determining whether catatonia has a medical or psychiatric cause, conventionally defined.Methods In this systematic review and meta-analysis of diagnostic test accuracy (PROSPERO CRD42021239027), Medline, EMBASE, PsycInfo, and AMED were searched from inception to May 11, 2022 for articles published in peer-reviewed journals that reported EEG findings in catatonia of a medical or psychiatric origin and were reported in English, French, or Italian. Eligible study types were clinical trials, cohort studies, case-control studies, cross-sectional studies, case series, and case reports. The reference standard was the final clinical diagnosis. Data extraction was conducted using individual patient-level data, where available, by two authors. We prespecified two types of studies to overcome the limitations anticipated in the data: larger studies (n >= 5), which were suitable for formal meta-analytic methods but generally lacked detailed information about participants, and smaller studies (n < 5), which were unsuitable for formal meta-analytic methods but had detailed individual patient level data, enabling additional sensitivity analyses. Risk of bias and applicability were assessed with the QUADAS-2 tool for larger studies, and with a published tool designed for case reports and series for smaller studies. The primary outcomes were sensitivity and specificity, which were derived using a bivariate mixed-effects regression model.Findings 355 studies were included, spanning 707 patients. Of the 12 larger studies (5 cohort studies and 7 case series), 308 patients were included with a mean age of 48.2 (SD = 8.9) years. 85 (52.8%) were reported as male and 99 had catatonia due to a general medical condition. In the larger studies, we found that an abnormal EEG predicted a medical cause of catatonia with a sensitivity of 0.82 (95% CI 0.67-0.91) and a specificity of 0.66 (95% CI 0.45-0.82) with an I2 of 74% (95% CI 42-100%). The area under the summary ROC curve offered excellent discrimination (AUC = 0.83). The positive likelihood ratio was 2.4 (95% CI 1.4-4.1) and the negative likelihood ratio was 0.28 (95% CI 0.15-0.51). Only 5 studies had low concerns in terms of risk of bias and applicability, but a sensitivity analysis limited to these studies was similar to the main analysis. Among the 343 smaller studies, 399 patients were included, resulting in a sensitivity of 0.76 (95% CI 0.71-0.81), specificity of 0.67 (0.57-0.76) and AUC = 0.71 (95% CI 0.67-0.76). In multiple sensitivity analyses, the results were robust to the exclusion of reports of studies and in-dividuals considered at high risk of bias. Features of limbic encephalitis, epileptiform discharges, focal abnormality, or status epilepticus were highly specific to medical catatonia, but features of encephalopathy had only moderate specificity and occurred in 23% of the cases of psychiatric catatonia in smaller studies.Interpretation In cases of diagnostic uncertainty, the EEG should be used alongside other investigations to ascertain whether the underlying cause of catatonia is medical. The main limitation of this review is the differing thresholds for considering an EEG abnormal between studies. Funding Wellcome Trust, NIHR Biomedical Research Centre at University College London Hospitals NHS Foun-dation Trust.Copyright (c) 2022 The Author(s). Published by Elsevier Ltd. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).

Automated summary

This study aimed to determine the performance of EEG in determining whether catatonia has a medical or psychiatric cause. The study found that an abnormal EEG predicted a medical cause of catatonia with fair accuracy, but performed poorly in distinguishing psychiatric causes.