CHAPTER 14 Cerebrospinal Fluid Analysis

M. Desnoyers, C. Bédard, J.H. Meinkoth, M.A. Crystal

Cytology of the cerebrospinal fluid (CSF) may be an important component of the diagnostic evaluation of patients with neurologic diseases. It may also be useful in patients with neck or limb pain and/or fever of unknown origin. As we will see in this chapter, it is important to emphasize that different pathologies may result in similar CSF changes due to the narrow range of inflammatory response in the central nervous system (CNS). Moreover, the absence of changes in the CSF does not eliminate the possibility of a process affecting the central nervous system. Therefore, in most instances a final diagnosis can be reached only with a combination of clinical history, physical findings, medical imaging, CSF findings, and other tests such as a complete blood count (CBC), biochemistry panel as well as serology, polymerase chain reaction (PCR), and histopathology.

COLLECTION TECHNIQUES

CSF collection is performed while the patient is under an-esthesia. Atlantooccipital puncture is performed using a 3.8-cm (1½-inch), 22-gauge spinal needle (occasionally, a 5-cm [2-inch] needle may be needed for dogs over 25 kg or for smaller dogs that are obese). For cats and dogs weighing less than 6 to 8 kg, a 1.2-cm (½-inch), 21-gauge butterfly catheter may be used. Lumbar puncture is performed using a 3.8-cm (1½-in) or 5.4-cm (2⅛-in), 22-gauge spinal needle. Other materials needed for CSF collection include clippers, supplies for sterile preparation of the site, sterile surgical gloves, and ethylene-tetra-acetic acid (EDTA) (purple top) and a sterile evacuated (red top) tubes in which to catch and submit the fluid. The laboratory should be notified before the collection is performed because rapid processing of the sample is important in achieving accurate results.

Before CSF collection, the patient should be assessed for clinical signs that might suggest an increase in intracranial pressure. Signs of increased intracranial pressure are variable, but may include cranial nerve signs (such as anisocoria or mydriatic/nonresponsive pupils), dull mentation/altered state of consciousness, rigid paresis, altered respiratory patterns and heart rhythms (bradycardia), and coma. If these are present, CSF collection should be postponed and appropriate therapy instituted. If CSF collection cannot be postponed, mannitol (2 g/kg, intravenously [IV]) should be given over 30 minutes, 1 hour before induction, and furosemide (2 mg/kg, IV) and either dexamethasone sodium phosphate (1 mg/kg, IV) or prednisolone sodium succinate (20 mg/kg, IV [slowly]) should be given 10 to 15 minutes before induction. Additionally, the patient should be positively hyperventilated for 5 minutes before the start of the procedure.

The location for CSF collection is selected based on the neurologic localization of the lesion. For lesions above the foramen magnum and of the extreme craniocervical spinal cord, CSF is collected from the atlantooccipital space (cerebellomedullary cistern). For lesions below the craniocervical spinal cord, CSF is collected from the lumbar subarachnoid space between lumbar vertebrae 5 and 6 (L5-6). Approximately 0.2 ml/kg of CSF can be safely collected.

For atlantooccipital collection, an area extending 3 to 4 cm from midline in each lateral direction from just cranial to the occipital protuberance to just caudal to the dorsal spinous process of the axis is clipped and sterilely prepared. The patient is then placed in lateral recumbency (right lateral for right-handed individuals), the nose is positioned parallel to the table, and the head is flexed ventrally to maximally open the atlantooccipital space. With sterile gloves, the clinician inserts the needle (bevel facing cranial for lesions cranial to the foramen magnum or caudal for extremely craniocervical lesions) on the midline just in front of the dorsal spinous process of the axis in the caudal area of a triangle formed by the cranial aspects of the wings of the atlas and the occipital protuberance. The needle should be aimed toward the tip of the nose. Usually, a popping sensation can be felt when the needle penetrates the dura mater. The stylet is then removed from the needle and, avoiding the first drop to minimize sample contamination, CSF fluid is allowed to drip into the collection tubes (the EDTA tube is of greater importance and thus should be used first).

If a popping sensation is not felt, and the needle has been advanced enough to potentially be through the dura mater, the stylet should be removed and 2 to 3 seconds allowed for CSF flow. If flow is not detected, the stylet can be replaced and the needle advanced farther. This technique is repeated until CSF flow occurs. It is always best to remove the stylet to check for CSF flow every 2 to 3 mm if the needle-tip position is uncertain because penetration of the medulla oblongata can result in death by cardiorespiratory arrest.

If the needle strikes bone during insertion, slight cranial or caudal redirection can be attempted to enter the atlantooccipital space. In this case, success is usually achieved only by withdrawing the needle, reassessing landmarks, and repeating the procedure. If the CSF drips at an extraordinarily slow rate, bilateral jugular vein compression can be attempted to hasten flow. Once fluid has been collected, the needle is removed and any fluid remaining in the needle is allowed to drip into one of the collection tubes. If a butterfly catheter is used, the needle is inserted as described previously and slowly advanced until fluid is seen entering the clear tubing attached to the needle.

For lumbar puncture, an area extending 3 to 4 cm from midline in each lateral direction from just caudal to the wings of the ileum to the dorsal spinous process of L4 is clipped and sterilely prepared. The patient is then placed in lateral recumbency (right lateral for right-handed individuals) and the lumbosacral area is flexed by bringing the hind limbs forward. With sterile gloves, the clinician inserts the needle on the midline, perpendicular to the spine at the cranial border of the dorsal spinous process of L6. Occasionally, the hind limbs will flinch, indicating penetration of the dura. The stylet is then removed from the needle and, avoiding the first drop to minimize sample contamination, CSF fluid is allowed to drip into the collection tubes. If fluid is not obtained, the stylet is replaced, and the needle is advanced to the floor of the spinal canal. The stylet is then removed and the needle is slowly withdrawn until CSF flow is observed. If the needle strikes bone during insertion, slight cranial or caudal redirection can be attempted, but withdrawing the needle, reassessing landmarks, and repeating the procedure are usually required for success.

Hemorrhage noted during CSF collection is usually a result of needle penetration into a blood vessel or the vertebral sinus and is rarely a result of hemorrhage into the subarachnoid space from a true disease process. If hemorrhage is noted, the needle should be withdrawn and discarded, and the procedure should be repeated with a new needle. An initial hemorrhagic tap does not necessarily lead to blood contamination on subsequent CSF collection attempts.

As mentioned earlier, it is important to emphasize that fluid obtained caudally to the lesion has more chances to be representative of the lesion rather than a fluid taken cranially to that lesion.¹

SAMPLE PROCESSING

Because total protein and nucleated cell count are usually low compared with whole blood, a sample collected in a red-top (serum) tube is adequate for the great majority of CSF samples ² without risk of sample clotting. The cerebrospinal fluid is a fairly labile substance and, unless a stabilizing agent is added, is stable under refrigerated conditions for about 4 to 8 hours but, preferably should be analyzed within 30 minutes to 1 hour.² Samples with higher protein content (> 50 mg/dl) are stable for a longer period than samples with lower protein content (< 50 mg/dl).³ A recent study³ showed a change in the percentage of large mononuclear cells after 2 hours, in small mononuclear cells after 12 hours, and in neutrophils after 24 hours in samples with no added stabilizing agents. Interestingly, the total cell count in those samples did not change significantly over a 24-hour period.³

Therefore, if a CSF sample cannot be analyzed within 1 hour of collection, the following procedure should be followed if there is sufficient sample (> 1 ml): the CSF should be separated into two aliquots: (1) one aliquot (unaltered) should be used for total protein and nucleated/red blood cell (RBC) count determination and (2) an aliquot containing 20% (volume:volume) fetal calf serum (FCS) or 10% (volume:volume) autologous serum should be used to perform the differential cell count. If the CSF sample is small (< 0.5 ml), adding hetastarch at a ratio of 1:1 should be done and all tests can be performed on the sample because adding hetastarch does not influence total protein. Keep in mind to correct the total cell count due to the dilution effect of the hetastarch.^3,⁴

Cell Count

Usually, the CSF cell count is too low to be analyzed by automated machines such as the Cell-Dyn or Advia. Therefore, a cell count using a hemacytometer is still the gold standard for nucleated and red cell count.² Some clinical pathologists prefer to stain the cells with new methylene blue (NMB) prior to counting, whereas others count unstained cells.

One technique to stain cells with NMB is to draw NMB into an unheparinized capillary tube until the capillary tube is about one-third full. Then, the stain is emptied from the tube and excess stain is blotted from the tip of the tube. The tube is then filled with CSF. The NMB adhered to the side of the tube stains the cells sufficiently to reveal cellular morphology. Due to the small amount of stain that is retained within the tube, this method does not influence cell count.³

Another technique is to draw a small amount of NMB into a capillary tube. The tube is tilted to allow the small amount of NMB to migrate to the middle of the tube, leaving an air pocket on either side of the NMB. A small amount of CSF is then drawn into the tube. The two liquids do not touch each other due to the presence of the air pocket. The capillary tube is then rocked back and forth several times so that the two columns of liquid move side to side, without touching. The tube is then allowed to sit for 10 minutes. The presence of the air pocket prevents the CSF and NMB from mixing together. Therefore, this technique does not influence the cell count.⁵

Cell counts (both nucleated and red cells) are performed by counting all cells (each cell type separately) in the nine large chambers of the hemacytometer. Both sides of the hemacytometer are counted and an average count is obtained by dividing the total of both sides by 2. Because the total area of one side of the hemacytometer is 0.9 mm³, multiplying the cell count by 1.1 gives the number of cells/μl². Experience may be needed to differentiate small nucleated cells from RBCs: usually nucleated cells have a granular appearance, whereas RBCs have a smooth appearance or may be crenated. If NMB is used, leukocytes will have a blue appearance, whereas RBCs will be unstained (Figure 14-1).

Figure 14-1 Numerous erythrocytes and two leukocytes present on a hemacytometer. The two nuclei of the two leukocytes in the center of the field stain dark purple (arrows). (NBM, original magnification 50×.)

Total Protein Determination

CSF fluid protein concentration is too low to be measured accurately by conventional biochemistry total protein techniques such as biuret or refractometer method. Instead, CSF total protein concentration is measured by using a Coomasie blue or pyrogallol red assay (Fisher Scientific, St-Laurent, QC). A crude method to determine the presence of excess protein in the clinical setting is to use urine chemstrip dipsticks (e.g., Multistix, Bayer Corporation, Elkhart, Ind.; Chemstrip 9, Roche Diagnostics, Laval, QC): the presence of a positive reaction indicates the presence of albumin because globulins are not detected by the dipsticks. Trace to 1+ protein is considered normal.⁶

The Pandy test is a screening test for the presence of immunoglobulin. It is performed by adding 2 or 3 drops of CSF to 1 ml of a 10% carbolic acid solution (10 mg of carbolic acid in 100 ml of distilled water²). Turbidity indicates the presence of globulins. The test is fairly sensitive with a sensitivity of approximately 50 mg/dl.^5,⁷ Agarose gel protein electrophoresis can also be performed on CSF.⁸

Cytologic Slide preparation

Because CSF is almost always poorly cellular, concentration techniques are required. In a university or private laboratory setting, the use of a cytocentrifuge is the preferred method of preparation. At the University of Montreal, we typically use 200 μl of CSF for slide preparation and prepare between two and four slides (if enough sample is available) in order to get an adequate number of cells. This results in the presence of an adequate number of evenly distributed cells that can be readily recognized. Slides are routinely stained with a modified Wright’s stain with some slides retained unstained if special stains are required.

In a private clinic setting, preparation of adequately cellular slides is more problematic. Sometimes, laboratories at local human hospitals can be used to prepare concentrated slides using a cytocentrifuge. These slides can then be examined in-house or mailed off for evaluation. If this is not available, a sedimentation technique is probably the most practical method. One such technique can be found in reference 2 and utilizes the barrel of a tuberculin syringe (the end where the needle is normally attached has been cut off). A hole that is slightly larger than the inside diameter of the syringe barrel, but less than the outside diameter is punched on a piece of filter paper, which is used to absorb excess fluid. The filter paper is placed on top of a clean glass slide, with the hole centered on the slide. The syringe barrel is placed, cut side up, on top of the filter paper so that the hole of the syringe barrel aligns with the hole of the filter paper. The flanged portion of the syringe barrel must then be clamped to the slide to remain steady during the sedimentation process. A sample of CSF is placed within the barrel of the syringe and allowed to sit for 30 to 60 minutes. At the end of the sedimentation period, excess fluid is aspirated or wicked off, and the slide is dried and then stained. Cell recovery is acceptable and representative of the CSF, but this technique is more time consuming.

NORMAL CSF

Normal CSF is a clear, colorless, water-like fluid. The presence of any color or turbidity must be considered abnormal (Table 14-1).

Table 14-1 Reference Values for Canine and Feline Cerebrospinal Fluid

Type of cells

The great majority of cells in the CSF are mononuclear in nature with a mixture of small (Figures 14-2 and 14-3) and large (Figures 14-4 and 14-5) mononuclear cells. The presence of less than 10% nondegenerate neutrophils is considered normal by most authors^12,¹³. Some studies have found up to 25% nondegenerate neutrophils in normal CSF.² The term macrophage should be restricted to mononuclear cells showing evidence of material ingestion in their cytoplasm¹³ (Figure 14-6). If contamination from skin occurs, a few anucleate superficial keratinized cells may also be present (Figure 14-7).

Figure 14-2 Small mononuclear cells in a cerebrospinal fluid sample. (Wright-Giemsa stain.)

Figure 14-3 Small mononuclear cell in a cerebrospinal fluid sample. The two cells to the left have slightly increased amounts of cytoplasm. (Wright-Giemsa stain.)

Figure 14-4 Large mononuclear cell with cytoplasmic vacuolation in a cerebrospinal fluid sample. (Wright-Giemsa stain.)

Figure 14-5 Numerous large foamy mononuclear cells and two neutrophils (arrows) in a sample of cerebrospinal fluid. Large mononuclear cells may have nuclei that are similar in shape to band cells or neutrophils, only larger (arrowheads). (Wright-Giemsa stain.)

Figure 14-6 Macrophages in cerebrospinal fluid showing evidence of (A) erythrophagia and (B) cytophagia. (Wright-Giemsa stain.)

Figure 14-7 CSF sample from a dog. A single large keratinized epithelial cell is present (arrowhead). Squamous epithelial cells represent cutaneous contamination. (Wright-Giemsa stain.)

Other Parameters

1) Glucose: the CSF glucose concentration is usually 60% to 80% that of the blood glucose concentration in both dogs and cats.²

2) Albumin quota: this value is easily obtained by dividing CSF albumin by serum albumin (CSF albumin/serum albumin). In dogs, the normal value varies between 0.17 and 0.3. An increase of the albumin quota indicates a blood-brain barrier disturbance.⁸

3) Protein electrophoresis: if total protein is increased in the CSF, agarose electrophoresis is possible and as little as 10 μl is required for this test. Increased values of immunoglobulins in the CSF associated with a normal albumin quota indicate intrathecal production of immunoglobulins.⁸

INTERPRETATION OF ABNORMAL CSF

CSF cytology can be abnormal even though the cell count is within reference range.¹⁴ Therefore, even when CSF cell count and total protein values are within reference limits, a cytologic examination can be of value.

Blood Contamination and Hemorrhage

Due to the sometimes difficult nature of CSF collection, blood contamination is possible. For instance, in human medicine, one in five lumbar punctures are contaminated with blood.¹⁵ This blood contamination may result in a pink to frankly reddish appearance. When blood is present, the first thing to determine is if the blood present is contamination (iatrogenic) or secondary to hemorrhage in the CNS (Figure 14-8). If hemorrhage is older than 12 hours, evidence of erythrophagia by macrophages will be seen on cytology. Theoretically, the presence of fresh blood contamination should result in a combination of RBCs without erythrophagia and presence of platelets but, from our own experience, platelets may be difficult to visualize. Also, if hemorrhage of several days’ duration is present, the CSF supernatant should be xanthochromic (have a yellow to yellow-orange discoloration). However, some recent human studies suggest that the human eye cannot reliably detect xanthochromia in CSF. This means that hemoglobin breakdown products may be present in the CSF, but not detectable by the naked eye.¹⁶^–¹⁷ Therefore, subarachnoid hemorrhage may be present in some instances, but a CSF discoloration will not be recognized. It can also be difficult to differentiate xanthochromia from icterus: a recent human study showed that about 80% of CSF samples with a significant amount of bilirubin appeared red instead of yellow macroscopically.¹⁷ Only spectrophotometry can differentiate xanthochromia from icterus adequately.

Figure 14-8 Hemorrhage versus blood contamination.

Effect of Blood Contamination on Total and Differential Cell Counts and Total Protein Concentration

The influence of blood contamination on total and differential cell counts and total protein concentration depends on the degree of blood contamination (Table 14-2). Some studies have shown that blood contamination yielding RBC counts of up to 15,000 RBC/μl does not influence either nucleated cell count or total protein concentration.^18,¹⁹ Another study, however, found that an increase of 1 white blood cell (WBC)/100 RBCs could be found in some cases of blood contamination.¹³ Therefore, if the observed WBC count is higher than the predicted increase of 1 WBC/100 RBCs, the increased WBC count is likely real (i.e., it cannot be attributed to blood contamination alone). However, if the CSF WBC count is less than or equal to the predicted blood contamination value, the increase in WBCs might be real or might be secondary to blood contamination. If blood contamination of CSF appears likely from its gross appearance, the best approach is to try to repeat the CSF tap. As a general rule, RBC count > 3000/μl warrants collection of another sample (J. Parent, neurologist, Université de Montréal, Canada, personal communication).

Table 14-2 Influence of Blood Contamination on Nucleated Cell Count and Total Protein

Red Blood Cell Count	Influence on Total Protein	Influence on Nucleated Cell Count
Less than 15,000/μl	No^18,¹⁹	No^18,¹⁹
Greater than 15,000/μl	Yes: overestimate^18,¹⁹	Yes: overestimate^18,¹⁹

Use of Formulas to Establish Reliable Leukocyte Differential Counts in Cases of Blood Contamination

Several studies in human and veterinary medicine ^18,^20,²¹ have shown that correction formulas do not reliably estimate “uncontaminated” WBC values in CSF. This is probably due to the fact that contamination of CSF samples by blood usually results from small meningial or spinal cord vessels that do not have the same proportion of WBCs/RBCs as large peripheral veins.¹⁸

CHANGES IN THE PERCENTAGE OF LEUKOCYTES WITHOUT PLEOCYTOSIS

Neutrophils

The presence of an increased percentage of neutrophils (10% to 20%) without a pleocytosis can be seen in several conditions (Figure 14-9). Acute trauma to the spinal cord, intravertebral disease,⁵ low-grade inflammation secondary to neoplasia, and thiamine deficiency (in cats)²² can result in an increased percentage of neutrophils without a pleocytosis.