Novel predictors of postoperative delirium after awake subthalamic nucleus deep brain stimulation for Parkinson’s disease

Background Postoperative delirium (POD) is a common complication of deep brain stimulation (DBS) in patients with Parkinson’s disease, leading to worsened outcomes. However, predictive factors remain unclear. This study aimed to identify predictors of POD following subthalamic nucleus (STN) DBS, with a focus on preoperative radiographic assessment and caudate nucleus (CN) transgression. Methods We retrospectively reviewed 122 patients who underwent awake STN DBS from June 2017 to August 2024. Clinical, laboratory and imaging variables were assessed including postoperative classification of CN transgression. Two Firth logistic regression models were constructed for analysis, with model 1 including and model 2 excluding the Movement Disorder Society Non-Motor Rating Scale. Results POD occurred in 16.4% of the patients. Model 2 (n=62, area under the curve (AUC)=0.904) showed superior performance to model 1 (n=122, AUC=0.884). History of lumbar TPS (trans-pedicle screw fixation) surgery (OR 6.21) and higher preoperative score of Unified Parkinson’s Disease Rating Scale (UPDRS) part II (off-medication) (OR 1.13) were independent predictors. A higher medial temporal lobe atrophy (MTA) score (OR 3.481) showed marginal significance (p=0.054). CN transgression by the electrode or postoperative perilead oedema (PLE) was not associated with an increased risk. Conclusions In this study, a higher preoperative UPDRS II score (off-medication) and history of lumbar TPS surgery are independent predictors of POD following STN DBS. A higher MTA score is marginally significant in predicting POD. In contrast, CN transgression or postoperative PLE appears to be unrelated. These findings may aid in risk stratification and surgical planning.