BACKGROUND: Delirium is a postoperative neurological morbidity in glioblastoma whose risk factors, incidence, and prognostic implications remain undefined. OBJECTIVE: To develop an algorithm using preoperative factors to predict postoperative delirium. METHODS: Retrospective analysis of 554 consecutive patients (mean age = 61.5 yr; 42% female) undergoing first glioblastoma procedure at our institution 2005 to 2011. RESULTS: Postoperative delirium occurred in 7% of patients (n = 38). Patients undergoing biopsy (10%; n = 54) did not experience delirium. In patients undergoing resection (n = 500), multivariate logistic regression identified 5 factors independently predicting postoperative delirium: age, chronic pulmonary disease, psychiatric history, bihemispheric tumors, and tumor size. We developed a score function entitled "GRAD" (Glioblastoma Risk Assessment for Delirium) to stratify patients into risk categories by assigning point(s) to each preoperative factor based on the relative magnitude of its regression coefficient. Point totals were summed for each patient: patients with 0 to 2 (n = 227) and 3 to 7 (n = 221) points were designated as low and high risk with postoperative delirium rates of 2% vs 15%, respectively (chi-square; P < .001), with the model validated using a separate patient cohort. Postoperative delirium lengthened hospital stays (P < .001), decreased likelihood of discharge home (P < .001), and was independently associated with decreased survival (4.5 vs 13.4 mo; hazard ratio = 1.9 [1.2-2.8]) in multivariate analysis. CONCLUSION: We developed a model to predict development of postoperative delirium using 2 tumor-specific (bihemispheric tumors and tumor size) and 3 patient-specific (age, psychiatric history, and chronic pulmonary disease) factors. High-risk patients and their families should be counseled preoperatively, and this risk could be considered in the choice of biopsy vs resection, and resection patients should be monitored closely postoperatively.
BACKGROUND:Delirium is a postoperative neurological morbidity in glioblastoma whose risk factors, incidence, and prognostic implications remain undefined. OBJECTIVE: To develop an algorithm using preoperative factors to predict postoperative delirium. METHODS: Retrospective analysis of 554 consecutive patients (mean age = 61.5 yr; 42% female) undergoing first glioblastoma procedure at our institution 2005 to 2011. RESULTS:Postoperative delirium occurred in 7% of patients (n = 38). Patients undergoing biopsy (10%; n = 54) did not experience delirium. In patients undergoing resection (n = 500), multivariate logistic regression identified 5 factors independently predicting postoperative delirium: age, chronic pulmonary disease, psychiatric history, bihemispheric tumors, and tumor size. We developed a score function entitled "GRAD" (Glioblastoma Risk Assessment for Delirium) to stratify patients into risk categories by assigning point(s) to each preoperative factor based on the relative magnitude of its regression coefficient. Point totals were summed for each patient: patients with 0 to 2 (n = 227) and 3 to 7 (n = 221) points were designated as low and high risk with postoperative delirium rates of 2% vs 15%, respectively (chi-square; P < .001), with the model validated using a separate patient cohort. Postoperative delirium lengthened hospital stays (P < .001), decreased likelihood of discharge home (P < .001), and was independently associated with decreased survival (4.5 vs 13.4 mo; hazard ratio = 1.9 [1.2-2.8]) in multivariate analysis. CONCLUSION: We developed a model to predict development of postoperative delirium using 2 tumor-specific (bihemispheric tumors and tumor size) and 3 patient-specific (age, psychiatric history, and chronic pulmonary disease) factors. High-risk patients and their families should be counseled preoperatively, and this risk could be considered in the choice of biopsy vs resection, and resection patients should be monitored closely postoperatively.
Authors: P R Kappen; E Kakar; C M F Dirven; M van der Jagt; M Klimek; R J Osse; A P J E Vincent Journal: Neurosurg Rev Date: 2021-08-16 Impact factor: 2.800
Authors: Pablo Kappen; Johannes Jeekel; Clemens M F Dirven; M Klimek; Steven A Kushner; Robert-Jan Osse; Michiel Coesmans; Marten J Poley; Arnaud J P E Vincent Journal: BMJ Open Date: 2021-10-01 Impact factor: 2.692