Epidural blood patch (EBP) is propagated for postdural puncture headache with reported success rates between 70 and 98% [1]. EPB is considered a safe procedure, and sideeffects, besides transient local back and leg pain, have been rarely described [2]. We report a case of a patient with symptoms of a communicating hydrocephalus, which occurred after an EBP was performed 2 weeks earlier. This 26-year-old male was first seen in November 2008 because of severe pain at the dorsal aspect of the left leg which had suddenly started 5 days earlier. Previous medical history was unremarkable. At neurological examination a positive straight left leg raising test, decreased reflex of the left Achilles tendon and an impaired sensation in the S1 dermatoma were found. MRI of the lumbar spine showed no abnormalities. Because the symptoms persisted, and radiculitis was considered, a lumbar puncture was performed. CSF analysis showed no pleiocytosis (1 mm), no erythrocytes, and a normal protein level (19 mg/dl). Cytomegalovirus, Epstein Barr, and Borrelia antibodies were negative. Subsequently he developed severe headache which worsened by standing and was relieved in the supine position. Because the headache did not improve 1 week after the lumbar puncture, he underwent an EBP at the L4–5 level using the loss of resistance to saline technique [3]. No cerebrospinal fluid could be aspirated and 20 ml of autologous blood was slowly injected. Immediately after the procedure he complained of paraesthesia in both legs, which resolved after 30 min. Initially the headache resolved, but 2 weeks later worsened, now also present in the supine position. At neurological examination there was no meningismus and no papiloedema. MR imaging of the brain (Fig. 1) showed communicating hydrocephalus and no venous sinus thrombosis or subarachnoidal hemosiderin depositions. Lumbar puncture was repeated, with an opening pressure of 26 cm H2O. After drainage of 30 ml of yellowish cerebrospinal fluid the headache immediately improved. CSF analysis confirmed the presence of bilirubin with no erythrocytes, pleiocyosis (39 mm) and elevated protein (1,000 mg/dl). MRI of the lumbar-thoracic region revealed extensive clumping of nerve roots and linear enhancement of the surface of the nerve roots and spinal cord, consistent with arachnoiditis (Fig. 2). A ventriculoperitoneal drainage was performed which immediately resolved the headache. In our patient arachnoiditis and communicating hydrocephalus were probably caused by intrathecal injection of blood during the EBP, confirmed by the presence of bilirubin in the second CSF analysis. However, we cannot exclude a causative role of the prior lumbar puncture, which previously has been correlated with lumbar arachnoiditis [4]. Another cause of both events and the prior S1 lumbar radiculopathy seems unlikely, because of the initial normal MR and CSF analysis. With epidural anesthesia a 3% incidence of inadvertent injection of drugs into the subarachnoid space has been reported [5]. One case of intrathecal hematoma and arachnoiditis after an attempted EBP has been described [6]. Another report presented a case of neurological deterioration due to further increasing intracranial pressure after EBP in a patient with pre-existent hydrocephalus [7]. For intracranial subarachnoidal hemorrhage (SAH) it has been shown that the amount of N. Al Maach O. J. M. Vogels T. L. Bollen P. H. Wessels (&) Department of Neurology, St Antonius Hospital Nieuwegein, PO Box 2500, 3430 EM Nieuwegein, The Netherlands e-mail: p.wessels@antoniusziekenhuis.nl