A 75-year-old, initially healthy female patient was admitted to the emergency department⁶ after having sustained an Anderson Type II⁷ dens axis fracture (figure 1A) with a cerebral concussion. A CT scan of the scull and brain upon admission showed no signs of hydrocephalus. The concomitant injuries included serial fractures of the ribs and a knee injury. The stabilisation of the dens axis fracture was carried out via a posterior approach using the Harms technique¹ (figure 1B). The wound drainage was prolonged and had a clear yellow colour. A positive 2-transferrin test⁸ showed the fluid to be CSF. To estimate the size of the accumulated CSF, an MRI was performed which showed a subcutaneous CSF accumulation at the operation site. Secondary revision of the wound and application of Duraseal (Confluent Surgical, Waltham, Massachusetts, USA) to seal off the spinal canal from the surrounding muscles 12 days later was not successful. This was probably due to more ventral location of the dural tear. However, the wound remained dry (figure 2) and did not warrant further revisions. After several weeks a therapy-resistant arterial hypertension occurred with systolical values more than 180 mm Hg. The patient progrediently became disoriented and had a dysarthria. A CT scan of the brain performed 4 weeks after the accident showed a ventricular dilatation (figure 3B) with signs of elevated ventricular pressure. A hydrocephalus communicans occurred and a VP shunt (Hakim Codman Ventile; Cilag AG, Switzerland, 70 mm H₂O pressure limitation) was applied. The CSF diagnostics showed no evidence of an infection. After normalisation of the CSF pressure the patient's neurological signs normalised, and the control CT scan performed 5 days later showed almost normally configured ventricles (figure 3C). The orientation and dysarthria significantly improved after application of the VP shunt. The patient was discharged in a neurological rehabilitation centre after 49 days.

Upper spinal fractures in older patients are not uncommon. The trauma mechanism mainly consists of a hyperextension of the neck after a fall. Usually these fractures come along with other injuries such as cerebral concussion or fractures of the extremities. At this cervical level (C2) the patient usually dies when there is a dislocation leading to any neurological manifestation. Dura mater is mostly torn in atlanto-axial fractures as earlier reported in a post mortem study.² The operative approach leaves a cavitation for fluid accumulation. The Harms technique¹ was used to stabilise this fracture without exposing the dural sac. Even if the CSF fistula would have been caused by the initial operative procedure there would not be any traumatic reason to develop a hydrocephalus. The genesis of a hydrocephalus communicans was obviously slow and took a few days. To discuss whether Duraseal or DBX could be the reason for CSF circulation disturbances remains hypothetical, due to a pressure gradient outward the intradural space. Persistent flow out of the dural space inhibits any material translocation towards CSF space. However, the inhibition of re-absorption of CSF after upper traumatic spine injuries cannot be explained as yet. Hydrocephalus as a result of spinal pathology is rare and has been reported with spinal tumours where the pathogenesis may be the disruption of the CSF flow, impaired absorption due to elevated CSF protein content, tumour infiltration into basal cisterns or compression of the spinal venous plexus. The pathogenesis after trauma may be related to the high level of the injury (C2), thus leading to blood accumulation from the site of the injury in the basal cisterns and consequently impairing the absorption of CSF by the arachnoid villi across the hemispheres, which in turn leads to a communicating hydrocephalus, a combined form of the malabsorptive and occlusive hydrocephalus. Several treatment modalities have been recommended for spinal CSF leakage, but only sufficient closure with cadaveric dura or muscle flap can withstand the high hydrostatic pressure.^11–13 Some surgeons favour re-operation with direct repair or application of fibrin sealant on the dural defect.¹⁴ Others suggest continuous closed CSF drainage and/or permanent shunting.¹⁴ The continuous spinal drainage could lower the CSF pressure gradient across the leak and enhance spontaneous closure of the defect. The application of Duraseal was not successful in this case (figure 2). This is indicated by a persistent accumulation of CSF in the occipito-nuchal area and in the second wound drainage suggesting a fistula in the ventral area, a surgically inaccessible location. Whenever a pseudomeningocele⁹ or CSF leak occurs, one should consider the possibility of concurrent hydrocephalus. If this complication is encountered early, an externalised lumbal drain may be sufficient to treat this complication.