Evaluation of a method for noninvasive intracranial pressure assessment during infusion studies in patients with hydrocephalus

Object. A mathematical model previously introduced by the authors allowed noninvasive intracranial pressure (nICP) assessment. In the present study the authors investigated this model as an aid in predicting the time course of raised ICP during infusion tests in patients with hydrocephalus and its suitability for estimating the resistance to outflow of cerebrospinal fluid (Rcsf). Methods. Twenty-one patients with hydrocephalus were studied. The nICP was calculated from the arterial blood pressure (ABP) waveform by using a linear signal transformation, which was dynamically modified by the relationship between ABP and cerebral blood flow velocity. This model was verified by comparison of nICP with “real” ICP measured during lumbar infusion tests. In all simulations, parallel increases in real ICP and nICP were evident. The simulated Rcsf was computed using nICP and then compared with Rcsf computed from real ICP. The mean absolute error between real and simulated Rcsf was 4.1 ± 2.2 mm Hg minute/ml. By the construction of simulations specific to different subtypes of hydrocephalus arising from various causes, the mean error decreased to 2.7 ± 1.7 mm Hg minute/ml, whereas the correlation between real and simulated Rcsf increased from R = 0.73 to R = 0.89 (p < 0.001). Conclusions. The validity of the mathematical model was confirmed in this study. The creation of type-specific simulations resulted in substantial improvements in the accuracy of ICP assessment. Improvement strategies could be important because of a potential clinical benefit from this method.