Significant association of slow vasogenic ICP waves with normal pressure hydrocephalus diagnosis

Abstract

Objective: We aimed to test whether there is an association of slow vasogenic wave (SVW) occurrence with positive response to external lumbar drainage (ELD) and ventriculoperitoneal shunting and to design a method for the recognition and quantification of SVWs in the intracranial pressure (ICP) signal. Materials and methods: We constructed SVW templates using normalized sine waves. We calculated the cross-correlation between the respective SVW template and the ICP signal. This was followed by shifting the templates forward and performing the cross-correlation analysis again until the end of the recording. Cross-correlation values above a threshold were considered to be indicative of SVWs. This threshold was previously determined and validated on a sample of ICP records of six patients. We calculated the root mean square of the recognized SVW periods as a measure of signal strength. Time-averaged signal strength was calculated over the full recording time (ICPSmean) and over the wave periods (ICPS). Results: We determined ICPS and ICPSmean in recordings of 2 groups of patients presenting with Hakim’s triad: 26 normal pressure hydrocephalus (NPH) patients and 20 non-NPH patients. We then tested whether there was an association between ICPS or ICPSmean and the respective diagnosis using a Mann–Whitney test. We found significant association between ICPS (p = 0.014) and ICPSmean (p = 0.022) and the diagnoses. Conclusions: The described method based on pattern recognition in the time domain is suitable for the detection and quantification of SVWs in ICP signals. We found a significant association between the occurrence of SVWs and independent NPH diagnosis.