Risk of brain herniation after craniotomy with lumbar spinal drainage: a propensity score analysis.

Abstract

Lumbar spinal drainage (LSD) during neurosurgery can have an important effect by facilitating a smooth procedure when needed. However, LSD is quite invasive, and the pathology of brain herniation associated with LSD has become known recently. The objective of this study was to determine the risk of postoperative brain herniation after craniotomy with LSD in neurosurgery overall. Included were 239 patients who underwent craniotomy with LSD for various types of neurological diseases between January 2007 and December 2016. The authors performed propensity score matching to establish a proper control group taken from among 1424 patients who underwent craniotomy and met the inclusion criteria during the same period. The incidences of postoperative brain herniation between the patients who underwent craniotomy with LSD (group A, n = 239) and the matched patients who underwent craniotomy without LSD (group B, n = 239) were compared. Brain herniation was observed in 24 patients in group A and 8 patients in group B (OR 3.21, 95% CI 1.36-8.46, p = 0.005), but the rate of favorable outcomes was higher in group A (OR 1.79, 95% CI 1.18-2.76, p = 0.005). Of the 24 patients, 18 had uncal herniation, 5 had central herniation, and 1 had uncal and subfalcine herniation; 8 patients with other than subarachnoid hemorrhage were included. Significant differences in the rates of deep approach (OR 5.12, 95% CI 1.8-14.5, p = 0.002) and temporal craniotomy (OR 10.2, 95% CI 2.3-44.8, p = 0.002) were found between the 2 subgroups (those with and those without herniation) in group A. In 5 patients, brain herniation proceeded even after external decompression (ED). Cox regression analysis revealed that the risk of brain herniation related to LSD increased with ED (hazard ratio 3.326, 95% CI 1.491-7.422, p < 0.001). Among all 1424 patients, ED resulted in progression or deterioration of brain herniation more frequently in those who underwent LSD than it did in those who did not undergo LSD (OR 9.127, 95% CI 1.82-62.1, p = 0.004). Brain herniation downward to the tentorial hiatus is more likely to occur after craniotomy with LSD than after craniotomy without LSD. Using a deep approach and craniotomy involving the temporal areas are risk factors for brain herniation related to LSD. Additional ED would aggravate brain herniation after LSD. The risk of brain herniation after placement of a lumbar spinal drain during neurosurgery must be considered even when LSD is essential.