Content area


Diagnostic and therapeutic challenges in superficial CNS siderosis

Daniel Kondziella1, Mette Lindelof2, Donika Haziri1, 2, Vibeke Andrée Larsen3 & Anders Kruse4 1) Department of Neurology, Rigshospitalet2) Department of Neurology, Herlev Hospital3) Department of Neuroradiology, Rigshospitalet4) Department of Orthopaedic Surgery, Rigshospitalet, Denmark

Article opens in PDF


Introduction: Superficial CNS siderosis was previously almost unknown but is now diagnosed with increasing frequency owing to magnetic resonance imaging. Patients may present with sensory deafness, gait ataxia, various sensorimotor signs and, eventually, cognitive decline. They typically have a history of traumatic brain or spinal cord injury or previous neurosurgery, or may harbour congenital malformations. However, knowledge about treatment outcomes remains scarce.

Methods: We present a series of nine consecutive patients from a large tertiary neuroscience centre in order to highlight the challenges related to the diagnosis and treatment of superficial siderosis.

Results: A potential bleeding aetiology was identified in all patients, but removal of the offending bleeding source was achieved only in three (33%). Symptom progression was halted in just one patient (11%), which suggests that neurodegeneration due to haemosiderin-associated iron toxicity becomes irreversible with time.

Conclusion: Surgical therapy in superficial CNS siderosis is rarely achieved. We suggest that prospective, large-scale multicentre studies are needed to search for non-surgical therapies that reverse (or prevent) ongoing neurotoxicity due to accumulating iron toxicity.

Funding: not relevant.

In the “classic” or “infratentorial” type of superficial CNS siderosis, subpial haemosiderin deposits of the brain and spinal cord accumulate due to chronic subarachnoid haemorrhage, often causing progressive ataxia, deafness, dementia and sometimes a bed-ridden helpless state (Figure 1 and Figure 2) [1-9]. Many different conditions can lead to this form of superficial CNS siderosis, including congenital malformations such as fluid-filled collections in the spinal canal, CNS tumours, post-neurosurgery, head or back trauma, arteriovenous malformations and brachial plexus/root injuries [4, 5]. Surgical removal of the bleeding source remains the only causative treatment option. However, knowledge about treatment outcomes is scarce and typically derived from single case reports [1, 3, 5-9]. We therefore present a consecutive case series from a tertiary neuroscience centre in order to stress the challenges related to diagnosing and treating patients with superficial CNS siderosis.


We report on nine consecutive patients with superficial CNS siderosis (one female, eight male; median age = 49.5 years, range: 29-74 years) who came to our attention in the 2008-2013 period. The diagnosis was based on clinical symptoms suggestive of the disease, characteristic superficial hemosiderin deposits with predilection for the brainstem, cerebellum and spinal cord as revealed by magnetic resonance imaging (MRI), as well as subarachnoid haemorrhage confirmed by lumbar puncture, revealing bloody cerebrospinal fluid (CSF) (without clearance in subsequent CSF tubes) and elevated bleeding metabolites as revealed by spectrophotometry. Spectrophotometric measurement of xanthochromia showed bilirubin (broad peak around 455 nm) with or without oxyhaemoglobin (415 nm), consistent with subarach-noidal bleeding. Patients were assessed for a bleeding source using MRI, including T2-weighted and gradient-echo sequences, at 1.5 or 3 T of the brain and spinal cord. Following surgery, patients were re-assessed by lumbar puncture for clearance of bleeding derivatives.

Trial registration: not applicable. Informed patient consent for publication was obtained.


The complete triad of deafness, ataxia and dementia was seen in seven patients (78%; Table 1). Urinary incontinence occurred in four patients (44%). Four patients (44%) presented with diplopia and/or decreased visual acuity. One patient complained of monosymptomatic tinnitus (suggestive of haemosiderin damage to the vestibulocochlear nerves). The median modified Rankin Scale score was 4 (range: 1-5). MRI showed leptomeningeal haemosiderin deposits with a predilection for the posterior fossa (n = 9). Additional haemosiderin was noticed around the spinal cord, ranging from extensive and involving most of the spinal canal (n = 4) to minor deposits predominantly affecting the cervical medulla (n = 5) as well as on the cerebral cortex (n = 9), mainly localised to the Sylvian fissures (n = 7). CSF analysis revealed elevated erythrocyte counts (mean = 5,340 cells/μl, range: 199-12,700). Bilirubin levels were increased in all patients, but oxyhaemoglobin levels in only five. CSF protein levels (range: 0.63-2.76 g/l; normal: 0.15-0.50) and white blood cell counts (range: 5-866 cells/μl, normal < 5) were increased in all patents. CSF neurofilament light, total tau and phosphorylated tau protein levels were measured, and they were increased in three patients.

A potential bleeding aetiology was identified in all patients. Five patients (56%) had an acquired bleeding; three patients had road traffic accidents leading to brachial plexus (n = 1), respectively, cervicothoracic nerve root avulsions (n = 2), one patient had an odontoid fracture during childhood, and another patient had been operated upon for a medulloblastoma. The mean delay from injury to first symptoms was 17.5 ± 12.5 years. Three patients (33%) had congenital bleeding sources, including spinal arachnoid cysts (n = 2) and brain malformations (n = 1). One patient (11%) had both a possible congenital (thoracic arachnoid cyst) and an acquired bleeding source (previous neurosurgery for a medulloblastoma).

Neuroimaging revealed potentially excisable lesions in five patients (56%). One patient declined surgery and another was surgically explored but no bleeding source was detected. Definite surgical excision with post-operative clearance of CSF bleeding derivatives was thus achieved in three out of five patients. However, symptom progression was halted in only one patient (11%). Medical therapy was attempted in two patients (iron chelation, oral prednisolone, and/or hyperbaric oxygen) but was unsuccessful, as expected.


Establishing a potential bleeding aetiology seems possible in most patients with superficial CNS siderosis. All our patients had acquired or congenital bleeding aetiologies, the significance of which is well-established in previous reports [1, 3-5, 7, 9]. It is believed that slow and chronic venous leakage of blood products into the subarachnoid space is the cause of superficial haemosiderin deposits [2, 4, 5]. Given that all patients were referred to our tertiary neuroscience centre, there is a high likelihood of referral bias, which suggests that the present patients were affected by a rather high degree of morbidity. Indeed, the incidence of asymptomatic superficial CNS siderosis is unknown since intra vitam diagnosis of haemosiderin deposits requires MRI which is typically performed because of neurological complaints. Of note, one patient in the present case series (patient 1) was monosymptomatic, and superficial CNS siderosis was a rather unexpected finding during MRI. Although sensorineuronal deafness and tinnitus due to haemosiderin damage to the vestibulocochlear nerves are well-recognised features of the disease [2, 4], to the authors’ knowledge, monosymptomatic tinnitus due to superficial CNS siderosis has not been reported previously.

As our data show, identifying a potential bleeding aetiology (e.g. a previous traffic accident) does not necessarily mean that an active bleeding can be found, and definitive surgical removal of the culprit bleeding source is achieved only rarely. Nevertheless, patients may deteriorate, probably because neurodegeneration induced by haemosiderin-associated iron toxicity becomes irreversible [2, 3]. Important mechanisms, akin to those seen in other neurodegenerative disorders such as Alzheimer’s disease, include hyperphosphorylation of tau protein and oxidative stress [3]. Despite post-surgical decrease of CSF bleeding derivatives in all three patients in whom a bleeding source was removed, two patients (66%) continued to deteriorate. This suggests a point-of-no-return when critical amounts of haemosiderin have accumulated [2, 3]. It remains unknown why some patients apparently tolerate larger amounts of haemosiderin deposits than others, but genetic factors may play a role [6]. Furthermore, some patients with superficial CNS siderosis clearly suffer from concurrent brain pathologies, which contribute to their neurological decline, e.g. patients 6 and 9 with a history of radiation and chemotherapy for a medulloblastoma. However, the point-of-no-return hypothesis of haemosiderin accumulation is supported by e.g. patient 4 who progressed clinically despite serial MRI showing unchanged amounts of haemosiderin during a period of 11 years.


Neurosurgeons and neurologists must become familiar with the diagnosis of superficial CNS siderosis in order to quickly establish the bleeding aetiology, identify the source of haemorrhage using high-resolution MRI of the brain and spinal cord, and, if possible, schedule patients for timely surgery. However, definite removal of a culprit bleeding source is achieved only occasionally and even then further clinical progression is possible. With the increasing use of MRI and better knowledge of the disease among physicians, the reported incidence is expected to rise in the future. This may allow for prospective large-scale multicentre studies searching for non-surgical therapies that reverse (or prevent) ongoing neurotoxicity due to accumulating iron toxicity.

Correspondence: Daniel Kondziella, Neurologisk Afdeling, Rigshospitalet, Blegdamsvej 9, 2100 Copenhagen, Denmark.

Accepted: 17 March 2015

Conflicts of interest:Disclosure forms provided by the authors are available with the full text of this article at

Patient consent: Informed consent to publication was obtained from the patients (or their next of kin, if necessary).


Aquilina K, Kumar R, Lu J et al. Superficial siderosis of the central nervous system following cervical nerve root avulsion: the importance of early diagnosis and surgery. Acta Neurochir 2005;147:291-7.

Koeppen AH, Michael SC, Li D et al. The pathology of superficial siderosis of the central nervous system. Acta Neuropathol 2008;116:371-82.

Kondziella D, Zetterberg H. Hyperphosphorylation of tau protein in superficial CNS siderosis. J Neurol Sci 2008;273:130-2.

Kumar N, Cohen-Gadol AA, Wright RA et al. Superficial siderosis. Neurology 2006;66:1144-52.

Lewy M, Turtzo C, Llinas RH. Superficial CNS siderosis: a case report and review of the literature. Nature Clin Pract Neurol 2007;3:54-8.

Matsushima A, Yoshida T, Yoshida K et al. Superficial siderosis associated with aceruloplasminemia. Case report. J Neurol Sci 2014;339:231-4.

McCarron MO, Flynn PA, Owens C et al. Superficial siderosis of the central nervous system many years after neurosurgical procedures. J Neurol Neurosurg Psychiatry 2003;74:1326-8.

Payer M, Sottas C, Bonvin C. Superficial siderosis of the central nervous system: secondary progression despite successful surgical treatment, mimickin amyotrophic lateral sclerosis. Acta Neurochir 2010;152:1411-6.

Posti JP, Juvela S, Parkkola R et al. Three cases of superficial siderosis of the central nervous system and review of the litterature. Acta Neurochir 2011;153:2067-73.

Kumar S, Goddeau RP Jr, Selim MH et al. Atraumatic convexal subarachnoid hemorrhage: clinical presentation, imaging patterns, and etiologies. Neurology 2010;74:893-9.

Lummel N, Bernau C, Thon N et al. Prevalance of superficial siderosis following singular, acute aneurysmal subarachnoid hemorrhage. Neuroradiology 2015;57:349-56.

Bib ref: 
Dan Med J 2015;62(5):A5079


Right side