Abstract
National guidelines for analysis of cerebrospinal fluid (CSF) in suspected subarachnoid haemorrhage (SAH) were published by a UK National External Quality Assessment Scheme (NEQAS) Immunochemistry Working Group in 2003. 1 These guidelines provided recommendations on sampling requirements, spectrophotometric analysis and interpretation of results. The same group, having evolved into the UK NEQAS Specialist Advisory Group for External Quality Assessment (EQA) of CSF Proteins and Biochemistry, published in 2007 the results of the audit it conducted into current UK laboratory practice for the biochemical investigation of CSF. 2 As a result of the findings of this audit and in the light of experience in the use of the 2003 guidelines, amendments have been made to the original guidelines and the revised guidelines are published in this journal. 3
The main aims of the revision are threefold: to ensure that CSF is always analysed spectrophotometrically if sufficient sample is received; to downgrade the clinical significance of detecting small amounts of haemoglobin with no increase in bilirubin; and to emphasize the clinical significance of increased bilirubin with a visible oxyhaemoglobin absorbance peak irrespective of CSF protein or serum bilirubin concentrations.
The National Audit of CSF Testing identified 78 laboratories which offered spectrophotometric analysis of CSF either on site or elsewhere. 2 Twenty-one of these 78 laboratories rejected samples delivered by pneumatic tube, 17 rejected heavily blood-stained samples and 14 rejected samples taken outwith sample timing guidelines. These sample rejection policies are indefensible because useful information may be gained from spectrophotometry in all of these circumstances. CSF usually will be blood-stained following SAH – blood staining does not necessarily reflect a traumatic tap. The detection of increased bilirubin is significant irrespective of the mode of sample delivery or sample timing. Use of pneumatic tube would not be expected to cause false-negative bilirubin results. CSF is precious. Repeat sampling is invasive and interpretation of any positive results highly problematic. Laboratories are duty bound to analyse any CSF sample if sufficient is received and to interpret the results in the light of all the available information.
It is the combined experience of the group over several years that the probability of SAH in a patient whose correctly timed CSF sample contains only small amounts of oxyhaemoglobin is vanishingly small. We know of one patient with a posterior communicating artery aneurysm whose CSF sample had a borderline net bilirubin absorbance (0.0067 AU) with a net oxyhaemoglobin absorbance of 0.067 AU. We are aware of no other cases of SAH where the CSF sample had a net oxyhaemoglobin absorbance <0.1 AU and a net bilirubin absorbance ≤0.007 AU. The reporting comment recommended in the original guidelines for such samples ‘oxyhaemoglobin on its own has a low predictive value for SAH but does not exclude’ confuses and frustrates clinicians. We recommend that it be replaced by the less ambiguous comment ‘no evidence to support SAH’.
In the original guidelines, the recommended interpretation effectively downgrades the likelihood of SAH when net bilirubin absorbance >0.007 AU but CSF protein concentration is >1 g/L. The 1 g/L cut-off is an arbitrary figure selected in recognition of the fact that high CSF protein concentrations (and therefore bilirubin absorbances) may reflect alternative pathology such as meningitis. However, it is important that biochemists realize that by its nature SAH usually results in a raised CSF protein concentration, sometimes >1 g/L. A blood-stained CSF sample with increased oxyhaemoglobin and bilirubin absorbances and CSF protein concentration >1 g/L is entirely consistent with SAH. For example, of 18 cases of SAH with net bilirubin absorbance >0.007 AU identified in one laboratory last year, five had CSF protein concentrations >1 g/L. The relevant comment in the 2003 guidelines for these cases is ‘This finding may be consistent with: SAH; an increased bilirubin accompanying the increased CSF protein; or other sources of CSF blood.’ Clinicians do not understand this comment. Our priority should be to alert them to the fact that there is a real possibility that a patient has had a SAH, not confuse matters by citing other potential causes of increased CSF bilirubin. The key finding is a net bilirubin absorbance >0.007 AU along with a visible oxyhaemoglobin peak. We recommend that such findings be reported as ‘consistent with SAH’, irrespective of CSF protein concentration. Conversely, the absence of a visible oxyhaemoglobin peak suggests that an alternative diagnosis such as meningitis is more likely, and therefore a CSF protein concentration >1 g/L should be factored into interpretation along with an increased serum bilirubin concentration.