CSF is thought to flow continuously from the site of production in the ventricles into interconnected spaces; i.e. cisterns and subarachnoid spaces (SASs). Since the SAS of the optic nerve is defined by a cul-de-sac anatomy, it is not evident how local CSF might recycle from that region to the general SAS. The concept of free communication of CSF has recently been challenged by the description of a concentration gradient of beta-trace protein, a lipocalin-like prostaglandin d-synthase (L-PGDS), between the spinal CSF and that in the SAS of the optic nerve, indicating diminished local clearance or local overproduction of L-PGDS here. In fact, computed cisternography with a contrast agent in three patients with idiopathic intracranial hypertension and asymmetric papilloedema demonstrate a lack of contrast-loaded CSF in the SAS of the optic nerve despite it being present in the intracranial SAS, thus suggesting compartmentation of the SAS of the optic nerve. The concept of an optic nerve compartment syndrome is further supported by a concentration gradient of brain-derived L-PGDS between the spinal CSF and the CSF from the optic nerve SAS in the same patients.