A proposed model of cerebrospinal fluid circulation: observations with radionuclide cisternography.

PURPOSE To determine the mechanisms of the tracer distribution at radionuclide cisternography (RC). METHODS Ten patients with venous vasculitis were studied with RC. Flow phantom studies were performed mimicking cerebrospinal fluid (CSF) circulation with and without a main outlet comparable to the pacchionian granulations. RESULTS Nine of the 10 patients had normal findings on RC images, including a maximum uptake over the vertex at 24 hours. In all patients, a second maximum occurred in the lumbosacral area. The flow phantom studies showed no tracer accumulation at an open outlet corresponding to the pacchionian granulations. On the contrary, a maximum arose without such an outlet. A maximum always arose at the closed dead ends of the phantom, including the lumbosacral area. CONCLUSION The commonly accepted flow model for CSF circulation needs to be revised. The pattern of the normal RC cannot be explained by a bulk flow transport of the tracer to an outlet at the pacchionian granulations but rather by a primary mixing caused by pulsatile flow with a secondary dilution by newly formed CSF from the ventricular system. We suggest that the main absorption of the CSF is through the central nervous system to the blood.