Comparison between 3 infusion methods to measure cerebrospinal fluid outflow conductance

Object There are several infusion methods available to estimate the outflow conductance (Cout) or outflow resistance (Rout = 1/Cout) of the CSF system. It has been stated that for unknown reasons, the bolus infusion method estimates a higher Cout than steady-state infusion methods. The aim of this study was to compare different infusion methods for estimation of Cout. Methods The following 3 different infusion methods were used: the bolus infusion method (Cout bol); the constant flow infusion method, both static (Cout stat) and dynamic (Cout dyn) analyses; and the constant pressure infusion method (Cout cpi). Repeated investigations were performed on an experimental model with well-known characteristics, with and without physiological pressure variations (B-waves, breathing, and so on). All 3 methods were also performed in a randomized order during the same investigation in 20 patients with probable or possible idiopathic normal-pressure hydrocephalus; 6 of these patients had a shunt and 14 did not. Results Without the presence of physiological pressure variations, the concordance in the experimental model was good between all methods. When they were added, the repeatability was better for the steady-state methods and a significantly higher Cout was found with the bolus method in the region of clinically relevant Cout (p < 0.05). The visual fit for the bolus infusion was dependent on subjective assessment by the operator. This experimental finding was confirmed by the clinical results, where significant differences were found in the investigations in patients without shunts between Cout of the visual bolus method and Cout stat, Cout dyn, and Cout cpi (4.58, 4.18, and 6.12 μl/[second × kPa], respectively). Conclusions This study emphasized the necessity for standardization of Cout measurements. An experienced operator could partly compensate for difficulties in correctly estimating the pressure parameters for the bolus infusion method, but for the general user this study suggests a steady-state method for estimating Cout.