Article Text


Decreased CSF hypocretin-1 (orexin-A) after acute haemorrhagic brain injury
  1. K Rejdak1,
  2. A Petzold2,
  3. L Lin3,
  4. M Smith4,
  5. N Kitchen5,
  6. E J Thompson6
  1. 1Department of Neurology, Medical University, Lublin, Poland
  2. 2Department of Neuroinflammation, Institute of Neurology, Queen Square, London WC1, UK
  3. 3Center for Narcolepsy, Stanford University, California, USA
  4. 4Neurosurgical Intensive Care Unit, National Hospital for Neurology and Neurosurgery, Queen Square
  5. 5Victor Horsley Department of Neurosurgery, The National Hospital for Neurology and Neurosurgery, Queen Square
  6. 6Department of Neuroinflammation, Institute of Neurology, Queen Square, London, UK
  1. Correspondence to:
 Dr K Rejdak
 Department of Neurology, Medical University of Lublin, 8 Jaczewskiego Street, 20-954 Lublin, Poland;

Statistics from

The novel hypothalamic neuropeptides orexins, or hypocretins, have gained much attention as potent modulators of various different physiological functions.1 Deficient orexin neurotransmission may be responsible for excessive somnolence, as shown in several conditions related to secondary narcolepsy, through direct or indirect damage to the posterior hypothalamus and its connections.2 We have studied for the first time the longitudinal changes of hypocretin-1 concentrations in cerebrospinal fluid (CSF) in patients with acute haemorrhagic brain injury.

Nine patients from a previously reported cohort3 and 21 controls (seven women and five men, median age 38 years, range 17 to 70) with other neurological disease were enrolled in the study (table 1). The patient group included five subjects with intracerebral haemorrhage and five with subarachnoid haemorrhage (table 1). All patients had extraventricular drains inserted within a median of two days of disease onset (range 2 to 36) as a treatment procedure because of increasing signs of hydrocephalus. Morning CSF samples were collected twice: first between day 1 and day 2 after catheter insertion, and second between day 4 and day 10. Patients were assessed using the Glasgow coma scale (GCS) at presentation and the Glasgow outcome score (GOS) after three months.

Table 1

 Patient characteristics

The control group consisted of two patients with primary dementia, two with chronic headache, one with ataxia syndrome, and seven with non-specific neurological symptoms. None reported sleep abnormalities.

All samples were stored at −80°C until analysis. Hypocretin-1 was measured blind to diagnosis by direct radioimmunoassay of 100 μl of CSF (Phoenix Pharmaceuticals, Belmont, California, USA; detection limit 40 pg/ml, intra-assay variation <5%), as described previously.4 Samples with undetectable concentrations (value below 40 pg/ml) were operationally plotted at 0 pg/ml. Statistical tests were carried out with the GraphPad InStat 3.05 software package using the non-parametric Mann–Whitney U test.

There was a significant difference in median CSF hypocretin-1 concentrations between the controls (319.4 (302 to 361) pg/ml) and acute brain injury patients (100.4 (0 to 145.2) pg/ml) (calculated from the means of two measurements; median (range)). No difference was found for sex or age (table 1). The concentration of hypocretin-1 in CSF on the first day of sample collection (24 hours after catheter insertion) was 98.8 (0 to 147) pg/ml and did not change significantly over the observation period (114.0 (0 to 144) pg/ml). All concentrations were either lower than control values (>200 pg/ml), in the intermediate range (110 to 200 pg/ml), or in the very low, narcolepsy range (<110 pg/ml).4 Two patients (Nos 4 and 6) had undetectable levels, while all others had moderately decreased values compared with the cut off level of 196 pg/ml. In patient 6, the hypocretin-1 level increased to 53 pg/ml, but it remained undetectable in patient 4. Both patients suffered from spontaneous intracerebral haemorrhage, which was localised either to the thalamus (patient 4) or to the midbrain (patient 6). Other patients with moderately decreased hypocretin-1 levels were diagnosed as having subarachnoid haemorrhage (six patients) and intracerebral haemorrhage in the frontal lobe (one patient). There was no correlation between hypocretin-1 levels and GCS at disease onset or GOS at three months after disease onset,


This is the first study to show decreased levels of hypocretin-1 in the CSF of patients in the first week after acute brain injury caused by haemorrhagic stroke. Our data are in line with previous observations in patients with traumatic brain injury.4

The findings seem important in the light of a study investigating long term outcome in patients after subarachnoid haemorrhage.2 More than 75% of patients reported excessive fatigue or daytime sleepiness, which persisted for long period (months to years) after the event. However, the exact mechanism responsible for the complaints remained unknown. Although the aetiology may be quite complex, an abnormality in the hypocretin/orexin system could make an important contribution to the phenomenon.

How intracerebral haemorrhage affects hypothalamic function remains obscure. In two of our patients direct damage to the thalamic/brain stem circuits appeared to be responsible. However, the remainder of the patients had a subarachnoid bleed or lesions in remote brain structures unrelated to the hypothalamus. Previously we provided evidence of diffuse axonal injury in these patients5 and speculated that this represented one mechanism of disruption of the hypothalamic circuits. A remote chemical mechanism related to the presence of blood might also contribute to the decrease in hypocretin-1 levels in the CSF. It is well recognised that a large amount of blood entering the CSF compartment or brain parenchyma produces neurotoxic effects through various different mechanisms, including oxidative haem and iron metabolism and secondary oedema with abnormalities of brain perfusion.6

The decreased hypocretin-1 concentrations in our patients might have resulted from the dilution effect of the bleed into the CSF with the development of secondary hydrocephalus in the course of their disease. However, this seems rather unlikely as intraventricular drainage was initiated early and the first samples were collected at least 24 hours after catheter insertion. One might also expect that in hydrocephalus complicating subarachnoid haemorrhage there would be an accumulation of CSF constituents owing to reduced absorption by subarachnoid villi, which would go against our hypothesis; however, an opposite effect was observed and this persisted during the course of the disease.

An important caveat of our study is that control samples were obtained by lumbar puncture, and lumbar CSF is likely to have a different composition from cisternal CSF. However, the concentrations of different neurotransmitters in the ventricular CSF are reported to be higher than in corresponding lumbar puncture specimens.3

Further studies are needed to investigate prospectively the relation between hypocretin-1 production and sleep–wake cycle abnormalities in patients after haemorrhagic stroke.


View Abstract


  • Competing interests: none declared

Request permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.