Interference in spectrophotometric analysis of cerebrospinal uid by haemolysis induced by transport through a pneumatic tube system

Transcription

1 Original Article Ann Clin Biochem 2001; 38: 371±375 Interference in spectrophotometric analysis of cerebrospinal uid by haemolysis induced by transport through a pneumatic tube system Philip R Wenham, Tamara Hanson and J Peter Ashby From the Department of Clinical Biochemistry, Western General Hospital, Edinburgh EH4 2XU, UK SUMMARY. The hypothesis that sending blood-stained cerebrospinal uid (CSF) through a pneumatic tube causes in vitro haemolysis has been tested. Spectrophotometric scanning of CSF supernatants demonstrated a signi cantly greater absorbance at 415 nm in those CSF samples that had been sent through the tube system compared to those that had not (P=0 0034). It is concluded that passage of blood-stained CSF down a pneumatic tube system causes in vitro haemolysis, accompanied by the release of oxyhaemoglobin from the lysed cells into the surrounding CSF. In view of this observation, it is recommended that CSF samples requiring spectrophotometric analysis, as part of the investigation of subarachnoid haemorrhage, should not be transported via a pneumatic tube system. INTRODUCTION In patients with suspected subarachnoid haemorrhage (SAH), the identi cation of free oxyhaemoglobin and/or bilirubin in cerebrospinal uid (CSF) is a valuable adjunct to diagnosis, particularly when computed tomography (CT) or magnetic resonance imaging give equivocal results. Although visual examination of CSF appears to be the most common approach used within the UK, 1 there is now general acceptance that spectrophotometric examination of CSF is the method of choice, 2,3 particularly as it is more sensitive than the eye for the detection of oxy- and methaemoglobin. 4±6 To avoid false-positive results for oxyhaemoglobin, it is important to avoid any in vitro haemolysis of red cells that may contaminate the CSF (e.g. following a traumatic tap). The usual way of ensuring this is to centrifuge the CSF on receipt in the laboratory. Recently, it has been shown that red cells, added to CSF in vitro, even up to a concentration of erythrocytes per litre may be removed without lysis by centrifugation within 1 h of sampling. 7 In addition, the method of specimen transport to the laboratory may be important. In an earlier study, it was noted that, in serum samples that Correspondence: Dr Philip R Wenham. philip.wenham@luht.scot.nhs.uk were not visibly haemolysed, the amount of free oxyhaemoglobin detected by spectrophotometry was increased when clotted whole blood was passed through the hospital pneumatic tube system. 8 More recently, spectrophotometric studies in this department on a single patient specimen of CSF showed that the absorbance at 415 nm increased three-fold following passage through the tube system. Therefore, the present study was undertaken to investigate systematically whether transport via the pneumatic tube system can cause haemolysis in CSF contaminated with red blood cells and whether this phenomenon is reproducible. MATERIALS AND METHODS Pneumatic tube system This was supplied by Aircotronic (Nottingham, UK) and operates by both positive and negative pressure being applied to the sample carrier, depending where the carrier is within the system. CSF samples CSF specimens were obtained by lumbar puncture from patients admitted to the Department of Clinical Neurosciences at the Western General Hospital. They were transported by hand to the department then immediately 371

2 372 Wenham et al. centrifuged for 10 min at 3000 g ready for analysis. In vitro studies on a CSF pool spiked with blood CSF pool This was prepared by pooling any CSF remaining following the completion of all routine biochemical investigations. All samples that had an oxyhaemoglobin peak at 415 nm following spectrophotometric scanning were excluded from the pool, as were those with a total protein content of greater than 1 g/l, or were accompanied by a serum bilirubin concentration of greater than 50 mmol/l. Preparation of blood-stained CSF From the above pool, 11 further pools were prepared, by the addition of red blood cells, to give nal red cell counts of between 1500 and /L. In experiments 2±11, capillary blood was obtained from a nger prick using a Unistik 2 1 sampling device (Owen Mumford, UK). Ten microlitres of this was immediately mixed with 750 ml of isotonic saline (154 mmol/l NaCl), then varying amounts, between 25 and 200 ml, of the diluted blood were added to 1 2-mL aliquots of the CSF pool to give the desired cell count. In experiment 1, 10 ml of whole blood containing 1 2 mg/ml potassium±edta was used, between 3 and 4 h after venepuncture, and further diluted as outlined above. A range of cell counts was chosen that were representative of the counts obtained both in vivo following a true haemorrhage and in vitro following a traumatic tap. 6 All cell counts were performed on a Fuchs±Rosenthal counting chamber (Weber Scienti c International). From each of the 11 prepared blood-stained pools, six 200-mL aliquots were transferred into each of six plastic tubes. Three of these were left at room temperature in the laboratory and three were carried up to the neuroscience wards then returned to the laboratory via the hospital pneumatic tube system. All six aliquots were then centrifuged simultaneously for 10 min at 3000 g and the supernatants scanned in a Beckman DU 650 spectrophotometer between 350 and 600 nm. The net absorbance due to oxyhaemoglobin was then measured at 415 nm by drawing a tangent from 365 to 500 nm and measuring the perpendicular to this tangent at 415 nm. This task was automatically performed by the spectrophotometer.the absorbances were expressed as the mean (standard deviation) of each of the three sets of readings. Statistical analysis The mean absorbances of each set of three samples before and after passage through the pneumatic tube system were compared by the Wilcoxon signed rank test. The null hypothesis was that there was no difference in the absorbance of the two sets of samples. RESULTS Eleven experiments were performed in total, each carried out in triplicate; red cell counts ranged from 1916 to /L (see Table 1). TABLE 1. Net absorbance at 415 nm of supernatants of 11 CSF pools spiked with blood and either sent or not sent via the pneumatic tube system Transported via pneumatic tube Experiment Red cell Yes No number count610 6 /L Mean (SD) Mean (SD) (0 0179) (0 0020) (0 0122) (0 0036) (0 0149) (0 0028) (0 0033) (0 0047) (0 0304) (0 0203) (0 0175) (0 0044) (0 0199) (0 0002) (0 0131) (0 0094) (0 0053) (0 0008) (0 0013) (0 0020) (0 0957) (0 0124) In experiment 1, blood was anticoagulated with 1 2 mg/ml potassium±edta prior to addition to CSF. In experiments 2±11, no anticoagulant was used. Each absorbance reading is the mean of three estimates. SD=standard deviation; CSF=cerebrospinal uid.

3 Interference in spectrophotometric analysis of CSF 373 FIGURE 1. Absorbance at 415 nm of supernatants obtained from cerebrospinal uid (CSF) pools that were spiked with whole blood. 1=pools not transported through the tube system; 2=pools transported through the tube system. Each point is the mean of three estimates. Different symbols indicate different samples. FIGURE 2. Spectrophotometric scans of supernatants of a cerebrospinal uid (CSF) pool spiked with whole blood. Supernatants A and B were obtained before and after the passage of the pool via the pneumatic tube system. The cell count was /L. All samples showed the presence of an absorbance peak at 415 nm, although those that had been sent via the pneumatic tube system had values that were signi cantly higher than those that had not (P=0 0034) (see Fig. 1). Although the highest absorbance following passage through the tube system was in the sample with the highest red cell count, inspection of the data in Table 1 showed that there was no clear relationship between cell count and net absorbance at 415 nm. The spectrophotometric scans of one of these samples are shown in Fig. 2. DISCUSSION Although it is now generally agreed that spectrophotometric examination of CSF for xanthochromia is superior to visual inspection,

4 374 Wenham et al. the de nition of xanthochromia remains a point of contention. 3 One de nition was an absorbance exceeding at a wavelength of 415 nm and/or a peak between 450 and 460nm. 2 This was later challenged since any CSF with a total protein content greater than 1 g/l would have an absorbance at 415 nm that exceeded More recently, the de nition was modi ed to mean either one or the other of the original two criteria above, but not both together. 10 This implies that the presence of oxyhaemoglobin alone in a CSF specimen supports the diagnosis of SAH, although it has been argued that oxyhaemoglobin alone is not a speci c indicator for SAH because in vitro haemolysis cannot be excluded. 3,7 The results of these in vitro studies support the hypothesis that passage through a pneumatic tube system can cause lysis of any red cells that might be present in CSF. In all 11 experiments there were increases in the absorbance of the CSF supernatants at 415 nm, whether or not the blood had been anticoagulated prior to addition to the CSF pool (see Fig. 1). Furthermore, using the above absorbance cut-off of absorbance units at 415 nm, several of the scans obtained in this study would have been interpreted differently, depending on whether or not the sample had traversed the tube system. This point is illustrated by the two scans shown in Fig. 2, one of which would have been interpreted as negative for xanthochromia and the other positive and, therefore, on the basis of the above de nition, consistent with SAH. In this study, even samples that had not been sent via the tube system demonstrated the presence of an oxyhaemoglobin peak, indicating that minor haemolysis occurred during the experiment. Inspection of Table 1 shows that there was a range in absorbance values obtained, not only between experiments but also within each experiment, as illustrated by the variation in the observed standard deviation of each triplet set of absorbance readings. This suggests that not only are some samples more prone to haemolysis than others, but also that the same sample may behave differently on different occasions. Taken together, these ndings are in agreement with previous studies that have shown that the haemolysis of red cells added to CSF is not predictable. 7 The mechanism by which haemolysis arises is not immediately obvious. In part it may be due to the effect of the forces of acceleration and deceleration that occur as the samples travel down the pneumatic tube and their impact on the red cell membranes. It is possible that the susceptibility of red cells to these forces varies from one sample to another. Furthermore, the forces themselves might vary between each passage through the tube system. This might explain the absence of any clear relationship between cell count and haemolysis. Although the speci city of oxyhaemoglobin for SAH may be questionable, the presence of bilirubin or methaemoglobin is strongly supportive of it. 3 Since these pigments are more dif cult to detect in the presence of oxyhaemoglobin, for this reason alone our results indicate that CSF samples requiring spectrophotometric analysis should not be transported through a pneumatic tube system. Acknowledgement We wish to thank Dr R Beetham for helpful discussion. REFERENCES 1 Mendelow AD, Cartlidge N. Xanthochromia in subarachnoid haemorrhage. J Neurol Neurosurg Psychiatry 1990; 53: 270±1 2 Vermeulen M, Hasan D, Blijenberg BG, Hijdra A, van Gijn X. Xanthochromia after subarachnoid haemorrhage needs no revisitation. J Neurol Neurosurg Psychiatry 1989; 52: 826±8 3 Beetham R, Fahie-Wilson MN, Park D. What is the role of CSF spectroscopy in the diagnosis of subarachnoid haemorrhage? Ann Clin Biochem 1998; 35: 1±4 4 Kjellin KG, SoÈ derstroè m CE. Diagnostic signi cance of CSF spectrophotometry in cerebrovascular diseases. J Neurol Sci 1974; 23: 358±69 5 SoÈ derstroè m CE. Diagnostic signi cance of CSF spectrophotometry and computer tomography in cerebrovascular disease. Stroke 1977; 8: 606±12 6 Kronholm V, Lintrup J. Spectrophotometric investigations of the cerebrospinal uid in the near-ultraviolet region. Acta Psychiatr Scand 1960; 53: 314±29 7 Fahie-Wilson M, Park D. Spectrophotometric examination of CSF in suspected subarachnoid haemorrhage ± what should we look for? In: Martin SM, Halloran SP, eds. Proceedings of the XVI International Congress of Clinical Chemistry, 1996 July 8±12, London. London: Association of Clinical Biochemists, 1996: 85 8 Wilmot R, Clark NL, Davidson W, Ashby JP. Blood sample haemolysis following transport through a pneumatic tube system. In: Martin SM, Halloran SP, Green AJ, eds. Proceedings of the ACB National Meeting, 1995 May 15±19, Glasgow. Cambridge: Association of Clinical Biochemists, 1995: 143

5 Interference in spectrophotometric analysis of CSF Beetham R. Spectrophotometric examination of CSF for xanthochromia. Lancet 1992; 339: Van der Wee N, Rinkel GJE, Hasan D, van Gijn J. Detection of subarachnoid haemorrhage on early CT: is lumbar puncture still needed after a negative scan? J Neurol Neurosurg Psychiatry 1995; 58: 357±9 Accepted for publication 22 February 2001