Oscillation analysis as a supporting element toward safe and reliable lumbar catheter intracranial pressure monitoring

Abstract

OBJECTIVE Lumbar drainage of cerebrospinal fluid for treatment of refractory increased intracranial pressure (ICP) is associated with the risk of infratentorial herniation, but real-time biomarkers for signaling herniation at bedside are lacking. Here, the authors tested whether an alteration of pulsatile waveform conduction across the level of the foramen magnum could serve as an indicator of insufficient hydrostatic communication and impending herniation. METHODS This prospective observational cohort study included patients with severe acute brain injury who underwent continuous external ventricular drain monitoring of ICP and lumbar drain pressure monitoring. Continuous recordings of ICP, lumbar pressure (LP), and arterial blood pressure (ABP) were screened throughout a recording period of 4–10 days. Pressure differences between ICP and LP > 5 mm Hg for 5 minutes were defined as a Δ-event, implicating nonsufficient hydrostatic communication. During this period, oscillation analysis of the ICP, LP, and ABP waveforms was performed by determining the eigenfrequencies (EFs) and their amplitudes (AEF) via Fourier transformation scripted in Python. RESULTS Of 142 patients, 14 exhibited a Δ-event, with a median (range) ICP of 12.2 (10.7–18.8) mm Hg and LP of 5.6 (3.3–9.8) mm Hg during 2993 hours of recording time. The AEF ratio between ICP and LP (p < 0.01) and between ABP and LP (p = 0.032) increased significantly during Δ-events compared with the baseline values determined 3 hours prior to the event. The ratio between ICP and ABP remained unaffected. CONCLUSIONS Oscillation behavior analysis of LP and ABP waveforms during controlled lumbar drainage may serve as a personalized, simple, and effective biomarker to signal impending infratentorial herniation in real time without the need for simultaneous ICP monitoring.