Objective: To study the safety and accuracy of ventriculostomy by neurosurgical trainees.
Methods: Initial computed tomographic studies of 346 consecutive patients who underwent bedside ventriculostomy were reviewed retrospectively. Diagnosis, catheter tip location, midline shift, and procedural complications were tabulated. To analyze catheter placement, we used a new grading system: Grade 1, optimal placement in the ipsilateral frontal horn or third ventricle; Grade 2, functional placement in the contralateral lateral ventricle or noneloquent cortex; and Grade 3, suboptimal placement in the eloquent cortex or nontarget cerebrospinal fluid space, with or without functional drainage. Statistical analysis was performed using Fisher's exact test and a weighted kappa coefficient.
Results: Diagnoses included the following: subarachnoid hemorrhage, n = 153 (44%); trauma, n = 64 (18%); intracerebral hemorrhage/intraventricular hemorrhage, n = 63 (18%); and other, n = 66 (20%). There were 266 (77%) Grade 1, 34 (10%) Grade 2, and 46 (13%) Grade 3 catheter placements. Hemorrhagic complications occurred in 17 (5%). Four patients (1.2%) were symptomatic, with two (0.6%) requiring surgery. Inter- and intraobserver agreement was almost perfect (kappa = 0.846 and 0.922, respectively) as applied to our grading system. Rates of suboptimal placement were highest in patients with midline shift (P = 0.059) and trauma (P = 0.0001). Rates of optimal placement were highest in patients with subarachnoid hemorrhage (P = 0.003) and when the catheter was placed ipsilateral to the side of midline shift (P = 0.063). Neither the resident's training experience nor the side of placement seemed to affect accuracy.
Conclusion: Bedside ventriculostomy is a safe and accurate procedure for intracranial pressure monitoring and cerebrospinal fluid drainage.