Contraindications to CSF Collection

Cerebrospinal fluid collection is not indicated in cases in which a cause is obvious, such as known trauma or intoxication (Parent and Rand, 1994). Because anesthesia is required for collection of CSF in small animals, CSF collection is contraindicated in cases in which anesthesia is contraindicated (Carmichael, 1998; Cook and DeNicola, 1988).

Cerebrospinal fluid collection is contraindicated in cases with increased intracranial pressure. Increased intracranial pressure should be suspected with acute head trauma, active or decompensated hydrocephalus, anisocoria, papilledema, or cerebral edema. Expansile mass lesions and unstable CNS or systemic conditions may result in increased intracranial pressure or decreased pressure in the spinal compartment relative to the intracranial compartment. In these situations herniation of the brain may result in severe compromise of brain function, tetraplegia, stupor/coma, and/or death (Parent and Rand, 1994). Physical and neurologic examination, history, presentation, and results of imaging studies are of benefit in determining if these conditions are likely before the decision to collect CSF.

Even in cases with potential for brain herniation, the risks associated with CSF collection may be acceptable if the cause of patient deterioration is not apparent. Risk of herniation may be reduced by administration of dexamethasone (0.25 mg/kg IV) just before induction of anesthesia and by hyperventilation of the patient with oxygen during the procedure (Fenner, 2000). Except in cases in which dexamethasone is administered prophylactically because of suspected increased intracranial pressure, CSF collection should predate corticosteroid administration because of potential alteration of CSF composition (Rand, 1995).

Complications of CSF Collection

As with any medical procedure, the risks and benefits of CSF collection should be considered for individual cases. The potential exists for iatrogenic trauma to the spinal cord and/or brainstem from the collection needle, but is minimized by attention to anatomic landmarks and careful collection procedures (Carmichael, 1998; Parent and Rand, 1994). Risk of introduction of infectious agents into the CNS is minimized by adherence to the basic principles of aseptic technique and correct preparation of the site of collection (Cook and DeNicola, 1988).

Slight to moderate blood contamination is a common complication of collection associated with penetration of the dorsal vertebral sinuses or small vessels within the meninges; this may complicate interpretation of the fluid analyses and cytology, but has not been found to be harmful to the patient (Carmichael, 1998; Fenner, 2000).

Ketamine should not be used to anesthetize cats for CSF collection because it increases intracranial pressure and may induce seizures; gas anesthesia should be used (Parent and Rand, 1994).

If three unsuccessful attempts at CSF collection occur, abandonment of the procedure is recommended to decrease the probability of repeated penetration of the spinal cord, which may result in serious complications or death. Practice on cadavers before performing collection in live clinical cases has been recommended.

Equipment for CSF Collection

Clippers, scrub, and alcohol to surgically prepare the site of collection are needed. Sterile gloves should be worn during the procedure. A sterile disposable or resterilizable spinal needle with stylet is used. A 20- to 22-gauge, 1.5-inch needle with a polypropylene hub is recommended for most cases, although smaller needles may be needed in very small dogs and cats and longer needles may be needed in large dogs (Carmichael, 1998; Cook and DeNicola, 1988; Parent and Rand, 1994; Rand, 1995). Several needles should be available since replacement may be needed if the needle is inserted off the midline and enters a venous sinus (Cook and DeNicola, 1988).

Sterile plain tubes for collection of CSF are recommended. Some authors indicate that EDTA is not used because clotting is rare and EDTA may falsely elevate the protein concentration of CSF (Parent and Rand, 1994). However, others recommend addition of samples to EDTA if blood contamination is present or if elevated nucleated cell count, bacteria, elevated protein concentration, and/or the presence of fibrinogen are suspected because these may lead to clotting (Carmichael, 1998). If glucose determination is desired, it is recommended to collect CSF into fluoride/oxalate; this may not be necessary if CSF contains few erythrocytes and is analyzed rapidly.

Plastic containers are recommended because leukocytes may adhere to glass. Use of only a few plain, sterile containers makes collection easier and simpler and increases the probability of maximum yield of CSF.

Collection Volume

Carmichael (1998) indicates that approximately 1 mL of CSF per 5 kg of body weight can be collected safely. It may be dangerous to remove more than 1 mL of CSF per 30 seconds, more than 4 to 5 mL of CSF from the dog, more than 0.5 to 1 mL of CSF from the adult cat or more than 10 to 20 drops of CSF from the kitten. Rand (1990) indicates that 1.0 to 1.5 mL of CSF can usually be collected from the cat. The cat may be susceptible to meningeal hemorrhage if too much fluid is withdrawn.

Cerebellomedullary Cistern Collection

Collection at this site is indicated to classify lesions affecting the meninges of the head and neck when the clinical signs involve seizures, generalized incoordination, head tilt, or circling.

Preparation of the site should include clipping of the hair from the head and neck, from the anterior margin of the pinna to the level of the third cervical vertebra and laterally to the level of the lateral margins of the pinnae. This area should be scrubbed for a sterile procedure (Cellio, 2001).

The animal is positioned in lateral recumbency with the head and vertebral column positioned at an angle of approximately 90 degrees. Excessive flexion of the neck may result in elevation of intracranial pressure and increase the potential for brain herniation (Fenner, 2000) or may result in occlusion of the endotracheal tube (Carmichael, 1998). The nose should be held or propped so that its long axis is parallel to the table and it should not be allowed to rotate in either direction. The point of insertion is located on the midline approximately half way between the external occipital protuberance and the craniodorsal tip of the dorsal spine of C2 (axis) and just rostral to the anterior margins of the wings of C1 (atlas). The needle is inserted at the intersection of a line connecting the anterior borders of the wings of the atlas and a line drawn from the occipital crest to the dorsal border of the axis along the midline. Puncture of the skin first with an 18-gauge needle or a scalpel blade is helpful in overcoming skin resistance in thick-skinned animals. Alternatively, the skin can be pinched and lifted so that the needle can be safely pushed through the skin with a twisting motion.

The needle should be inserted with the bevel oriented cranially. It should be held perpendicular to the skin surface and gradually advanced with the stylet in place. Periodically the needle should be stabilized and the stylet withdrawn to determine if CSF is present. Occasionally, a sudden loss of resistance may be felt as the subarachnoid space is entered, but this may not be recognized in all cases. If the collector suspects that the needle has been inserted too deeply, the stylet may be removed and the needle slowly withdrawn a few millimeters at a time, watching for the appearance of fluid in the hub. If the needle hits bone during insertion, slight redirection of the needle cranially or caudally should be attempted to enter the atlanto-occipital space.

If opening pressure readings are taken, CSF fluid sample is taken by directing the flow of CSF through the manometer by way of a three-way stopcock. If pressure readings are not obtained, CSF may be collected directly from the spinal needle hub by dripping into a test tube or gentle aspiration of drops as they collect at the hub using a syringe. Attachment of a syringe to the needle with aspiration of CSF is not recommended because suction may result in contamination with blood or meningeal cells or obstruction of CSF flow by aspirated meningeal trabeculae. Careful aspiration is acknowledged to be necessary for collections in some cases. Passage of the needle through the spinal cord to underlying bone should be avoided at the cerebellomedullary cistern because it may cause damage to the cord and/or cause blood contamination of the CSF sample. On completion of collection of CSF, the needle is smoothly withdrawn. Replacement of the stylet is not necessary.

If the fluid appears bloody at the onset of collection, replacement of the stylet for 30 to 60 seconds may result in clearing of the blood. If the first few drops of CSF are still slightly bloody, they can be collected separately from the following drops that are often clear. If rate of flow of CSF is slow, the needle should be rotated slightly to make sure that it is clear at the luminal tip. If this not effective, rate of flow may be increased by compression of the jugular veins, resulting in expansion of the venous sinuses and increased CSF pressure.

Appearance of abundant fresh blood from the collection needle indicates that the point of the needle is most likely off the midline and in a lateral venous sinus. A new approach with a fresh, clean needle is recommended if the first attempt results in frank blood consistent with puncture of the venous sinus.

Lumbar Cistern Collection

Both cerebellomedullary and lumbar cistern specimens may be collected. Collection of a cerebellomedullary specimen is recommended prior to thoracolumbar myelography to ensure that a diagnostic CSF sample will be obtained because lumbar puncture alone may not be sufficient. The collection of CSF from the lumbar cistern is more difficult and more likely to be contaminated with blood than that from the cerebellomedullary cistern (Chrisman, 1992). Sometimes no fluid or only a very small amount of fluid can be obtained owing to the small size of the lumbar subarachnoid space. Lumbar puncture may be preferred in cases with localized spinal disease because it may be more likely to confirm abnormality than cerebellomedullary cistern collections (Thomson et al., 1990).

The dorsal midline is clipped and prepared between the midsacrum and L3, extending laterally to the wings of the ilium. The animal is placed in lateral recumbency and the back is flexed slightly to open the spaces between the dorsal laminae of the vertebrae. The L5-6 or L6-7 spaces are most commonly used in dogs because the subarachnoid space rarely extends to the lumbosacral junction. In cats collection can frequently be made from the lumbosacral space.

The dorsal spinous process of L7 lies between the wings of the ilia and is usually smaller than that of L6. To collect from the L5-6 intervertebral space, the needle is inserted just off the midline at the caudal aspect of the L6 dorsal spinous process and advanced at an angle cranioventrally and slightly medially to enter the spinal canal between the dorsal laminae of L5 and L6. Misdirection laterally into the paralumbar muscles or underestimation of the length of needle required might result in advancement of the needle to the hub without encountering bone.

Cerebrospinal fluid may be collected from the dorsal subarachnoid space, or the needle may be passed through the nervous structures to the floor of the spinal canal and CSF collected from the ventral subarachnoid space. The stylet is removed and the needle may be carefully withdrawn a few millimeters to allow for fluid flow. The rate of flow is usually slower than from the cerebellomedullary cistern. Rate of flow may be increased by jugular compression.

Cerebrospinal Fluid Opening Pressure

CSF pressure is measured with a standard spinal fluid manometer as the fluid is collected (Simpson and Reed, 1987). CSF opening pressure should be measured to confirm a supposed increase in intracranial pressure due to a space-occupying mass or cerebral edema. Its normal range is less than 170 mmH₂O (Lipsitz et al., 1999) and 100 mmH₂O, for the dog and cat, respectively (Chrisman, 1991).

Effect of Blood Contamination

Various formulas have been used to predict the effect of blood contamination on protein concentration and nucleated cell count in CSF (Parent and Rand, 1994; Rand et al, 1990). Rand (1995) indicates that red blood cell (RBC) counts greater than 30 cells/μl in CSF will have a profound effect on the total and differential cell counts. However, in a study (Hurtt and Smith, 1997) of iatrogenic blood contamination effects of total protein and nucleated cell counts in CSF, the RBC count was not significantly correlated with nucleated cell count or protein concentration in CSF from clinically normal dogs or those with neurologic disease. The study concluded that high CSF nucleated cell counts and protein concentrations are indicative of neurologic disease, even if samples contain up to 13,200 RBC/μl. Although blood contamination may make interpretation of CSF more difficult, red or pink CSF or CSF with a high RBC count should not be discarded as a useless specimen because cytologic evaluation may detect abnormalities (Chrisman, 1992).

Macroscopic Evaluation

Normal CSF is clear, colorless, and transparent and does not coagulate Deviations from normal should be recorded as part of the macroscopic evaluation and are often graded 1 to 4 + or as slight, moderate, or marked. Turbidity (Fig. 14-1A) is reported to be detectable if greater than 500 cells/μl are present (Fenner, 2000) or if at least 200 leukocytes/μl or 700 erythrocytes/μl are present (Parent and Rand, 1994).

FIGURE 14-1 A, Turbid CSF. Dog. There is marked turbidity or cloudiness as apparent against the background of newsprint. This animal had a steroid-responsive meningitis with a count of 760 nucleated cells/μl. Turbidity is reported to be detectable if greater than 200 leukocytes/μl are present. B, Xanthochromic CSF. Dog. The fluid has a yellow-orange discoloration and moderate turbidity from a dog with subarachnoidal hemorrhage of inflammatory origin.

Red to pink discoloration may be associated with iatrogenic contamination with blood or pathologic hemorrhage. Erythrophages or siderophages in a rapidly processed CSF specimen with fixative added immediately following collection support pathologic hemorrhage as an underlying cause. Xanthochromia is the yellow to yellow-orange discoloration (Fig. 14-1B) associated with pathologic hemorrhage due to trauma, vasculitis, severe inflammation, disc extrusion, or necrotic or erosive neoplasia. Occasionally xanthochromia will be seen with leptospirosis, cryptococcosis, toxoplasmosis, ischemic myelopathy, coagulopathy, or hyperbilirubinemia.

Quantitative Analysis

Other Tests

Other tests that have been recommended by various authors or used in specific situations include electrophoretic determination of albumin and determination of total immunoglobulin levels. In combination with the serum albumin level and serum immunoglobulin, these can be used to calculate the albumin quotient (AQ) and immunoglobulin G (IgG) index. The AQ is equal to the CSF albumin divided by serum albumin times 100. AQ greater than 2.35 suggests an altered blood-brain barrier with increased protein in CSF associated with leakage from plasma. The IgG index is equal to the (CSF IgG/serum IgG) divided by (CSF albumin/serum albumin). An IgG index greater than 0.272 with a normal AQ suggests intrathecal production of IgG. An increased IgG index and increased AQ are suggestive of an altered blood-brain barrier as the source of IgG (Chrisman, 1992).

Alterations in electrophoretic protein fractions have been reported to be useful in identifying inflammatory, degenerative, and neoplastic disease in combination with clinical signs. In general, dogs with canine distemper often have elevated CSF gamma globulins, most likely related to intrathecal production, while dogs with granulomatous meningoencephalitis (GME) may have elevated CSF beta and gamma globulins (Chrisman, 1992).

In a more recent work, Behr and co-workers (2006) found, using high-resolution protein electrophoresis, a strong linear correlation between CSF total protein concentration and AQ suggesting that an increased CSF total protein concentration is an indicator of blood-brain barrier dysfunction; moreover, although unexpected, the same authors found that electrophoretic profiles in a series of 94 dogs with different neurologic diseases were not characteristic of any particular disease concluding that high-resolution electrophoresis of paired CSF and serum samples cannot be considered a valuable ancillary diagnostic tool for canine neurologic diseases.

Detection of specific antibodies within the CSF and comparison with serum levels may be useful in diagnosis of infectious meningoencephalitides, including infectious canine hepatitis, canine herpesvirus, canine parvovirus, canine parainfluenza virus, canine distemper virus, ehrlichiosis, Rocky Mountain spotted fever, Lyme disease (borreliosis), Toxoplasma gondii, Neospora caninum, Encephalitozoon cuniculi infection, Babesia spp. infection, cryptococcosis (Berthelin et al., 1994), and blastomycosis. On the contrary, measurement of anticoronavirus IgG in CSF of cats with confirmed feline infectious peritonitis involving the central nervous system is considered of equivocal clinical use (Boettcher et al., 2007). Serial titers for serum IgG to show rising titer are helpful to demonstrate active disease. The presence of IgM in serum or CSF is considered more specific than IgG or total immunoglobulin levels for detection of active disease (Chrisman, 1992).

Glucose measurement in CSF and comparison with serum or plasma glucose levels are frequently cited. Normal CSF glucose is approximately 60% to 80% of the serum or plasma concentration (Fenner, 2000). However, changes in CSF glucose concentration in serum or plasma are not immediately reflected in CSF and may take 1 to 3 hours before they are apparent in CSF (Cook and DeNicola, 1988). The ratio between blood glucose and CSF glucose is frequently reduced in bacterial infections of the CNS in humans and has been reported to occur in some cases of pyogenic infections of the CNS, CSF hemorrhage, or blood contamination that may result in increased utilization of glucose by cells. However, the relationship between bacterial encephalitis and decreased CSF glucose compared with serum or plasma glucose may depend on multiple factors, including the blood glucose level, degree of permeability of the blood-brain barrier, and presence or absence of glycolytic cells or microorganisms. Fenner (2000) states that the reduction in glucose does not occur in dogs. Significant reductions in CSF glucose concentrations have been reported in human malignant disorders involving the leptomeninges and it is considered a relatively specific finding for this condition (Chamberlain, 1995).

Measurement of various electrolytes and enzymes in CSF has been reported. Their interpretation may be limited because of increases associated with altered blood-brain barrier permeability, concurrent evaluation of serum values, low benefit for cost, and poor correlation or specificity for particular pathologic processes or conditions (Chrisman, 1992; Cook and DeNicola, 1988; Parent and Rand, 1994).

Aerobic and anaerobic bacterial cultures are recommended for all CSF samples with degenerated neutrophils or when bacteria are identified cytologically.

Nevertheless, the CSF culture rarely assist in the identification of microorganisms: in a series of 8 confirmed cases of canine bacterial meningoencephalomyelitis, the CSF culture was positive in only 1 sample and the cultured bacteria was Corynebacterium spp. (Radaelli and Platt, 2002). Several factors likely contribute to poor culture performance in veterinary medicine —e.g., small volumes of CSF are collected for culture, the organisms are mostly confined to the brain parenchyma, some organisms are slow growing or require nonstandard culture techniques, and animal has received antibiotic therapy before sampling (Fenner, 1998).

Staphylococcus, Streptococcus, Klebsiella, Escherichia coli and Pasteurella are aerobic bacteria that may cause CNS infection; Fusobacterium, Bacteroides, Peptostreptococcus, Clostridium, and Eubacterium are anaerobic species that have been reported.

The immunophenotype of cerebrospinal fluid mononuclear cells in the dog has been examined (Duque et al., 2002). According to authors, the main advantage of standardized immunophenotyping is having a more objective classification of mononuclear cells compared with the relative inconsistency of this classification when based solely on microscopic appearance.

Methods of Cytologic Preparation

Standardization of the volume used for cytologic evaluation may be of benefit in minimizing analytic variation and aid in interpretation, although evaluation of multiple preparations or preparations from larger volumes of CSF may increase the likelihood of detection of minor abnormalities. Because CSF is normally of low cellularity and increases in cellularity may not result in large numbers of cells, concentration of the cells is required. Cytocentrifugation, sedimentation, or membrane filtration techniques may be used for concentration of cells.

Cytocentrifugation is most commonly available in reference or commercial laboratories in which the volume of submissions justifies purchase of specialized equipment. The membrane filtration technique requires special staining that is not commonly available in practice, but which may be available at some reference or commercial laboratories. Several sedimentation techniques have been described and are suitable for use in practice or commercial or reference laboratories to which rapid submission of CSF specimens is possible (Cook and DeNicola, 1988; Parent and Rand, 1994). Readers are referred to these sources for more detail on construction of a sedimentation apparatus and preparation of sedimentation specimens. A sample device is demonstrated in Fig. 14-2A&B.

FIGURE 14-2 A-D, In-house CSF sedimentation device. A, Unassembled sedimentation device with materials needed including 1 mL modified insulin syringe barrel, filter paper with hole punched, glass slide, two binder clips, and Eppendorf tube for CSF collection. B, Partially assembled sedimentation device. C, Assembled sedimentation device demonstrating the attachment of the binder clips to the barrel flanged portions. D, The tube made from the syringe barrel is filled with as little as 100 μl CSF by transfer pipette or as the figure demonstrates using a Butterfly needle. The added fluid is allowed to sit undisturbed for 1 hour. Cells concentrate and settle on to the exposed area of the glass slide.

Sedimentation preparations should be made if a specimen cannot be delivered immediately to the laboratory for cytologic processing. Prepared slides can then be sent to a commercial or reference laboratory for interpretation or stained and evaluated by clinicians at the practice.

Cytocentrifuge or sedimentation preparations are most commonly air-dried and stained with Romanowsky stains that are commonly available in commercial or reference laboratories and clinical practice laboratories. Membrane-filtration specimens require wet-fixation and stains appropriate for this method, commonly Papanicolaou, Trichrome, or H & E. Wet-fixation and these staining methods may also be used on cytocentrifuge or sedimentation preparations and are appropriate for formalin- or alcohol-fixed specimens. Cytocentrifuge or membrane-filtration preparatory and staining techniques may vary with laboratory, technical training, and pathologist preference. Summaries of cytopreparatory and staining techniques for cytospin and membrane filtration specimens and specimens fixed in formalin or alcohol are available from a variety of sources. Interested readers are referred to Keebler and Facik (2008) as a recent comprehensive review.

Special stains may be indicated in some cases. Gram stain may be useful for confirmation and identification of categories of bacteria. India ink or new methylene blue preparations have been reported to be helpful in identification of fungal infections, especially cryptococcosis. Periodic acid-Schiff stain may be used to demonstrate positive intracellular material in dogs with globoid cell leukodystrophy. Luxol fast blue can be used to demonstrate myelin in CSF specimens (Mesher et al., 1996).

Cytologic Features of Normal CSF

Normal CSF from healthy dogs and cats contains primarily mononuclear cells (Fig. 14-3) and is indicated to be a mixture of lymphocytes and large mononuclear (monocytoid) cells. The percentages of lymphocytes and monocytoid cells may vary with the method used for cytologic preparations, but lymphocytes are reported to be the predominant nucleated cell type in normal canine and feline CSF. However, Parent and Rand (1994) report monocytoid cells as the predominant type in normal CSF from healthy cats. They indicate monocytoid cells compose 69% to 100% of the nucleated cells, lymphocytes 0% to 27%, neutrophils 0% to 9% macrophages 0% to 3%, and eosinophils 0% to less than 1% of nucleated cells. Occasional neutrophils or eosinophils as within normal limits may be present as long as these cell types do not represent more than 10% or 1% of the nucleated cells, respectively. Low numbers of mature, nondegenerate neutrophils are occasionally seen in normal CSF from healthy dogs and cats and that rare eosinophils may be present. Occasional choroid plexus cells, ependymal cells, meningeal lining cells, or mitotic figures may be seen in normal CSF from dogs or cats (Chrisman, 1992; Rand, 1995).

FIGURE 14-3 Cell types found in CSF. Dog. Two small mononuclear cells (lymphocytes), one large mononuclear (monocytoid) at (arrow), one nondegenerate neutrophil, and one erythrocyte are present. (Wright-Giemsa; HP oil.)