Cerebrospinal fluid disturbances after transcallosal surgery: incidence and predictive factors

Abstract

OBJECTIVECSF dynamics after transcallosal resection of intraventricular lesions can be altered, and the need for shunt implantation complicates the management of these patients. Because the pathophysiological mechanism and contributing factors are poorly understood and the incidence has largely not been described, the authors conducted a study to elucidate these factors.METHODSThe authors retrospectively reviewed data from patients who had been operated on at their institution via a transcallosal approach between March 2002 and December 2016. They evaluated the need for a shunt implantation up to 3 months after surgery by assessing clinical variables. These variables were age at surgery, the need for perioperative external CSF drainage, histology of the lesion, and the following radiological parameters: pre- and postoperative Evans index, maximal postoperative extension of subdural effusions (SDEs) measured on axial images, and maximal interhemispheric fissure (IHF) width measured on coronal images assessed at 4 different points in time (preoperatively, day 1, days 2–4, and days 4–8 after surgery). To identify potential risk factors, univariate and multivariate regression models were constructed. Receiver operating characteristic (ROC) curves for significant predictors, as well as the area under the curve (AUC), were calculated.RESULTSSeventy-four patients (40 female and 34 male) were identified; their median age at surgery was 17.6 years (range 4 months to 76 years). Shunt implantation was necessary in 13 patients (ventriculoperitoneal [VP] shunt, n = 7; subdural peritoneal [SDP] shunt, n = 6) after a median interval of 24 days (range 10 days to 3 months). Univariate logistic regression models revealed a significant effect of IHF width on days 4–8 (OR 1.31, 95% CI 1.03–1.66; p = 0.027), extension of SDE on days 2–4 (OR 1.33, 95% CI 1.11–1 0.60; p = 0.003), and age (OR 0.932, 95% CI 0.88–0.99; p = 0.02). In the multiple regression model, the effect of the independent variable extension of the SDE remained significant. ROC curves for the predictors IHF width on days 4–8 and extension of SDE on days 2–4 revealed an AUC equal to 0.732 and 0.752, respectively. Before shunt implantation, the ventricles were smaller compared to the preoperative size in 9 of the 13 patients (SDP shunt, n = 5; VP shunt, n = 4).CONCLUSIONSThe rate of shunt-dependent hydrocephalus 3 months after surgery in this heterogeneous group of patients was 17.6% (95% CI 9.7%–28.2%). The authors identified as predictive factors the variables extension of the convexity space, IHF 1 week after surgery, and younger age.