Fluidic Considerations of Measuring Intracranial Pressure Using an Open External Ventricular Drain

Abstract

Measurement of intracranial pressure (ICP) during cerebrospinal fluid (CSF) drainage with an external ventricular drain (EVD) typically requires stopping the flow during measurement. However, there may be benefits to simultaneous ICP measurement and CSF drainage. Several studies have evaluated whether accurate ICP measurements can be obtained while the EVD is open. They report differing outcomes when it comes to error, and hypothesize several sources of error. This study presents an investigation into the fluidic sources of error for ICP measurement with concurrent drainage in an EVD. Our experiments and analytical model both show that the error in pressure measurement increases linearly with flow rate and is not clinically significant, regardless of drip chamber height. At physiologically relevant flow rates (40 mL/hr) and ICP set points (13.6 - 31.3 cmH₂O or 10 - 23 mmHg), our model predicts an underestimation of 0.767 cmH₂O (0.56 mmHg) with no observed data point showing error greater than 1.09 cmH₂O (0.8 mmHg) in our experiment. We extrapolate our model to predict a realistic worst-case clinical scenario where we expect to see a mean maximum error of 1.06 cmH₂O (0.78 mmHg) arising from fluidic effects within the drainage system for the most resistive catheter. Compared to other sources of error in current ICP monitoring, error in pressure measurement due to drainage flow is small and does not prohibit clinical use. However, other effects such as ventricular collapse or catheter obstruction could affect ICP measurement under continuous drainage and are not investigated in this study.

Keywords: bioengineering; catheter; cerebrospinal fluid; continuous measurement; external ventricular drain; intracranial hypertension; neuro-monitoring; neuro-surgery; neurology and critical care; traumatic brain injury.

Figure 1. Comparison of Experimental and Clinical Setup

A: Experimental setup for EVD characterization. The head model was fed by a syringe pump with programmable flow rate and fluid drained from the head model directly into a catheter which was placed into the column as shown. This junction was set at the same height as the pressure transducer. P_trans corresponds to the pressure measured with a digital transducer at the same height as the zero point. R₁ and R₂ are the fluidic resistances of tubing segments before and after the transducer. R_c is the resistance of the Medtronic 9025 catheter (Medtronic, Minneapolis, USA). Since the drip chamber is vented, pressure in it is equal to atmospheric pressure (zero gauge pressure). H is the distance above the zero point that the tubing terminates. B: Diagram of standard EVD setup. The zero point on the scale is set to the same height as the patient’s external auditory meatus by medical staff and the patient head is immobilized. EVD: external ventricular drain; ICP: intracranial pressure

Figure 2. Error in Pressure Measurement Is Linear and Small

Linear regressions of pressure measurement error compared to the prediction in a physiologic flow rate range. The analytical prediction follows experimental data, confirming that error in pressure measurements scales linearly with flow rate and is not affected by drip chamber height. Most importantly, the observed and predicted errors are very small, mostly under 1 cmH₂O. Data is shown as mean ± standard deviation (n = 3). Slope values, statistical analysis and plots of results from higher flow rates and individual data points are included in Table 5.

Figure 3. Error in Pressure Measurement Over a Wider Range of Flow Rates

This linear regression of all aggregate data points across all drip chamber heights and flow rates has a slope of 0.0192 cmH₂O*hr/mL and an R² of 0.987. The regression has been constrained to have no offset. The Pearson correlation coefficient is r(81) = 0.9934, p = 0.00001. 95% confidence intervals are plotted alongside the analytical prediction. The prediction line is based on Equation 2 and has a slope of 0.0223 cmH₂O*hr/mL.