To identify differences between catatonia due to a medical condition (CMC) and catatonia associated with a psychiatric disorder (CPD) based on clinical, biological, and treatment features.

This was a retrospective cohort study spanning 20 years (2001−2021). Patients diagnosed with catatonia were categorized into CMC or CPD as per DSM-5 criteria. Study outcomes included (1) severity of catatonia at diagnosis, defined using the Bush-Francis Catatonia Rating Scale (BFCRS) and Clinical Global Impression–Severity (CGI–S), (2) catatonia response to lorazepam, (3) time to discharge as a proxy for overall improvement, and (4) identification of clinical features that differentiate between CMC and CPD. To examine the relationship between clinical characteristics and CMC versus CPD, we used Least Absolute Shrinkage and Selection Operator (LASSO) for variable selection, followed by a final model to estimate associations.

Of 249 patients diagnosed with catatonia, 191 had CPD and 58 had CMC (mean age [SD] 48.3 [20.3] years, female n  = 138 [55.4 %]). Patients with CMC exhibited higher severity at diagnosis (CGI-S p   <  0.001, BFCRS p  = 0.005) and more complications (81.0 % vs. 49.7 %, p  <  0.001), including delirium (32.8 % vs. 5.2 %, p  <  0.001) and were less likely to respond to lorazepam (34 % vs. 55.6 %, p  = 0.007). Adjusting for age and sex, Cox proportional hazards model indicated a reduced hazard for discharge in patients with CMC, suggesting a shorter time to discharge for patients with CPD (HR = 0.71, 95 % CI 0.52–0.96, p  = 0.02). Predictors for CMC included negativism, delirium, history of seizures, focal abnormalities on head CT, antipsychotic use history, and electroencephalographic abnormalities, achieving an area under the curve (AUC) of 0.87.

Our study suggests potential differences between individuals with CMC and CPD, pointing to the possibility of more precise, individualized care.