Nomogram for Predicting Postoperative Delirium After Deep Brain Stimulation Surgery for Parkinson's Disease

OBJECTIVE: The purpose of the present study was to construct a nomogram for postoperative delirium (POD) after deep brain stimulation (DBS) surgery in patients with Parkinson's disease (PD). METHODS: The present study retrospectively enrolled 165 patients with PD who had undergone DBS surgery. The patients were assessed using the Confusion Assessment Method for the Intensive Care Unit and were divided into 2 groups: the delirium group and nondelirium group. The nomogram for POD after DBS surgery was constructed from the results of univariate analysis and multivariate logistic analysis of the influencing factors for POD after DBS surgery. RESULTS: Univariate analysis revealed that the preoperative length of stay and the presence of preoperative brain atrophy, preoperative pulmonary inflammation, and postoperative cerebral edema were statistically significant. The results from the nonmotor symptoms scale for PD, mini-mental state examination, PD sleep scale, and unified PD rating scale III were also statistically significant (P < 0.05). Multivariate logistic regression analysis showed that the unified PD rating scale III (odds ratio [OR], 2.284; 95% confidence interval [CI], 1.614-3.232), nonmotor symptoms scale for PD (OR, 8.191; 95% CI, 5.629-11.917), PD sleep scale (OR, 0.058; 95% CI, 0.05-0.067), preoperative length of stay (OR, 1.230; 95% CI, 1.053-1.437), and preoperative brain atrophy (OR, 3.912; 95% CI, 3.5-4.255) were independent factors that influenced the occurrence of POD after DBS surgery. A nomogram model was constructed using these indicators. The model predicted an area under the receiver operating characteristic curve after DBS surgery of 0.870 (95% CI, 0.808-0.918), sensitivity of 74.19%, and specificity of 91.54%. CONCLUSIONS: The present study has presented a reliable and useful nomogram that can accurately predict the occurrence of POD after DBS surgery in patients with PD. This tool is easy to use and could assist physicians during the therapeutic decision-making process.