Proposal of a predictive score for the occurrence of postoperative cerebral vasospasm: analysis of a large single institution retrospective series and literature review.

Abstract

Despite being uncommon, postoperative vasospasm (PoVS) present notably high morbidity and mortality rates. Our aim was to identify prognostic factors associated with this condition and introduce a scoring system to improve subsequent clinical and radiological surveillance strategies. We conducted a retrospective analysis of our institutional database covering patients aged over 18 who underwent craniotomic or transsphenoidal surgery for elective tumor removal at the Neurosurgical Unit of our institution between January 2016 and August 2023. A comprehensive search was conducted using the Cochrane Database of Systematic Reviews and PubMed database to identify the most correlated risk factors. Literature review included a final group of 32 studies (52 patients) and identified SAH, vessel encasement or vessel manipulation, hypothalamic disfunction, meningitis, younger age, tumor size > 3 cm, and long operative time as predictive factors for PoVS. Our cohort included 2132 patients, with only 13 individuals (0.61%) presenting PoVS. To predict the occurrence of PoVS, we developed a logistic multivariate regression model that identified thick (defined as Fisher grade ≥ 3) subarachnoid hemorrhage (coeff. 6.7, p < 0.001), intraparenchymal hemorrhage (coeff. 3.44, p < 0.001), lesion located in the parasellar region (coeff. 2.1, p = 0.064), and lesion size ≥ 4 cm (coeff. 2.0, p = 0.069) as potential independent predictors of PoVS. Based on statistical model for these variables was assigned a score: thick SAH 7 points, intraparenchymal hemorrhage 3 points, parasellar lesion site 2 points, and lesion size ≥ 4 cm 2 points. The cumulative scores ranged from 0 to 14. PoVS is a rare complication but its association with significant morbidity and mortality underscores the importance of early identification and treatment. In our study we proposed a stratified risk score to identify high risk patients. However, due to rarity of this condition, our score proposal should be considered as a training set a to be validated in future studies with a multicenter setting.