• 1. Bohn AA, Wells TB, West CL, et al. Cerebrospinal fluid analysis and magnetic resonance imaging in the diagnosis of neurologic disease in dogs: a retrospective study. Vet Clin Pathol 2006;35:315320.

    • Search Google Scholar
    • Export Citation
  • 2. Cook JR, DeNicola DB. Cerebrospinal fluid. Vet Clin North Am Small Anim Pract 1988;18:475499.

  • 3. Di Terlizzi R, Platt SR. The function, composition and analysis of cerebrospinal fluid in companion animals: part I—function and composition. Vet J 2006;172:422431.

    • Search Google Scholar
    • Export Citation
  • 4. Di Terlizzi R, Platt SR. The function, composition and analysis of cerebrospinal fluid in companion animals: part II—analysis. Vet J 2009;180:1532.

    • Search Google Scholar
    • Export Citation
  • 5. Mohammad-Zadeh LF, Sisson AF. Cerebrospinal fluid collection and analysis. NAVC Clin Brief 2005;21–23.

  • 6. Wessmann A, Volk HA, Chandler K, et al. Significance of surface epithelial cells in canine cerebrospinal fluid and relationship to central nervous system disease. Vet Clin Pathol 2010;39:358364.

    • Search Google Scholar
    • Export Citation
  • 7. Greene SA. Anesthesia for patients with neurologic disease. Top Companion Anim Med 2010;25:8386.

Advertisement

Effects of stylet-in versus stylet-out collection of cerebrospinal fluid from the cisterna magna on contamination of samples, sample quality, and collection time

View More View Less
  • 1 1Department of Companion Animals, Atlantic Veterinary College, University of Prince Edward Island, Charlottetown, PE C1A 4P3, Canada.
  • | 2 2Department of Pathology and Microbiology, Atlantic Veterinary College, University of Prince Edward Island, Charlottetown, PE C1A 4P3, Canada.
  • | 3 3Department of Health Management, Atlantic Veterinary College, University of Prince Edward Island, Charlottetown, PE C1A 4P3, Canada.

Abstract

OBJECTIVE

To evaluate safety of stylet-in and stylet-out techniques for collection of CSF from the cisterna magna and to assess whether there were differences between techniques with regard to contamination of samples, sample quality, and efficiency of collection.

ANIMALS

10 adult purpose-bred research Beagles.

PROCEDURES

A prospective crossover study was conducted. Preanesthetic physical and neurologic examinations and hematologic analyses were performed. Dogs were anesthetized, and collection of CSF samples from the cisterna magna by use of a stylet-in or stylet-out technique was performed. Two weeks later, samples were collected with the other sample collection technique. Samples of CSF were processed within 1 hour after collection.

RESULTS

Cellular debris was detected in higher numbers in stylet-in samples, although this did not affect sample quality. The stylet-out technique was performed more rapidly. No adverse effects were detected for either technique.

CONCLUSIONS AND CLINICAL RELEVANCE

Both techniques could be safely performed in healthy anesthetized dogs. The stylet-out technique was performed more rapidly and yielded a sample with less cellular debris. Both techniques can be used in clinical practice to yield CSF samples with good diagnostic quality.

Collection of CSF is a procedure commonly performed at referral practices as part of the assessment of patients with neurologic conditions. Cerebrospinal fluid is usually collected after advanced imaging of the CNS has been performed. Analysis of CSF is particularly useful in the diagnosis of inflammatory diseases of the CNS. In 1 study,1 abnormalities of the CSF were more common than abnormalities detected with MRI in dogs with inflammatory CNS diseases.

The CSF is produced primarily by the choroid plexuses in the brain and, to a smaller extent, the ependymal cells of the ventricular system.2,3 It circulates through the ventricular system before entering the subarachnoid space, where it is accessible for collection through the cisterna magna cranially or the interarcuate space between L5 and L6 or between L4 and L5 caudally.2 The CSF serves many functions, including physically protecting the CNS, assisting with regulation of intracranial pressure, and serving as a medium for transport of metabolites, neurohormones, and neurotransmitters.2 Macroscopically, CSF should be clear and colorless.3,4 Microscopically, the nucleated cells should consist primarily of lymphocytes and monocytoid cells, with a low percentage of mature neutrophils also being acceptable.3,4 It should not contain any (or very few) RBCs, the total nucleated cell count should not exceed 5 cells/μL, and the total protein concentration should not exceed 30 or 45 mg/dL for samples collected from the cerebellomedullary and lumbar regions, respectively.3,4

Proximity of CSF to the brain and spinal cord makes it a useful sample for detecting abnormalities within the CNS. However, not all diseases can be detected through changes in the CSF. In addition, detecting changes in the CSF relies on actual changes of the fluid composition as well as on changes attributable to sample collection because contamination of a sample can make it of nondiagnostic quality or more challenging to interpret cytologically.

Two techniques are commonly used to collect CSF: use of a spinal needle with the stylet in and use of a spinal needle with the stylet out. For the stylet-in technique, the spinal needle is inserted through the skin and directed toward the cisterna magna. Each time the needle is advanced 1 to 2 mm, the stylet is removed and flow of CSF is assessed. When no flow is evident, the stylet is replaced, the needle is then advanced another 1 to 2 mm, and the process is repeated until CSF flow is achieved and a sample is collected. Alternatively, the stylet-in technique can be performed by inserting the needle as previously described and advancing it until a popping sensation is felt, at which time the stylet is removed and flow of CSF is assessed. For the stylet-out technique, the stylet is removed once the spinal needle penetrates the skin and then the needle is slowly advanced until CSF flow is evident and a sample is collected. The stylet is never reinserted into the spinal needle. The preferred technique for collecting CSF has been discussed2–5; however, none of those authors offered reasons for their preferred method.

It could be speculated that lack of a stylet when advancing a spinal needle through the soft tissues may result in occlusion of the needle lumen or more contamination by tissue or blood, compared with results when a stylet is used. However, to the authors’ knowledge, no study has been conducted to compare the 2 techniques regarding quality of the sample and efficiency of sample collection. Because sample quality will influence the ability of clinicians to make a diagnosis, it is vital to understand the manner in which collection can affect the sample. Furthermore, reducing the amount of time an animal must be anesthetized to enable collection of a sample would minimize anesthetic risks for patients. The purpose of the study reported here was to evaluate the effect of the sample collection method on sample quality and collection efficiency. Our null hypotheses were that there would be no difference between the stylet-in and stylet-out sampling methods regarding sample quality or efficiency of collection and that both techniques could be safely performed.

Materials and Methods

Animals

Ten adult purpose-bred research Beagles were used for the study; each dog was anesthetized twice for collection of CSF. Physical and neurologic examinations were performed by one of the authors (PMA) before each anesthetic episode. A CBC and biochemical analysis were performed before the first anesthetic episode, and the PCV and concentrations of total solids, lactate, and glucose were measured before the second anesthetic episode. All procedures were approved by the University of Prince Edward Island Animal Care Committee.

Sample collection

A crossover study was conducted. For the first period of sample collection (which was on August 29), dogs were randomly assigned by use of a random number generator to the stylet-in or stylet-out technique; the order of sample collection among the 10 dogs was similarly randomly assigned. For the second period of sample collection (which was 14 days after the first period), the other sample collection technique was used for each dog, but the order of sample collection again was randomly assigned among the 10 dogs.

Dogs were not premedicated. Anesthesia was induced with propofola (4 mg/kg, IV) or alfaxaloneb (2 mg/kg, IV) at the discretion of the attending anesthetist. Dogs were intubated, and anesthesia was maintained with isoflurane. Dogs were positioned in right lateral recumbency on a table. The hair on the back of the head and neck was clipped, and the skin was aseptically prepared with 4% chlorhexidine gluconatec and 70% isopropyl alcohold in a standard manner. An assistant positioned the dog by holding the muzzle parallel to the tabletop and bending the cervical vertebrae so that the head was 90° perpendicular to the vertebral column. The assistant held the skull and muzzle immobile for the entire procedure. The target location for insertion of the spinal needle was selected as the point where a line between the cranial edges of the wings of the atlas intersected the dorsal midline halfway between the spinous process of the axis and the occipital protuberance of the skull. There was an interval of at least 15 minutes between anesthetic induction and collection of a CSF sample. All CSF samples were obtained by one of the authors (PMF), who had used the stylet-in technique for > 15 years. A 22-gauge, 1.5-inch spinal needle with the stylet seated completely in the needle was inserted perpendicularly into the skin at the target location. Once the tip of the needle entered the skin, a stopwatch was started.

After the spinal needle was passed through the skin, the protocol differed depending on the technique. For the stylet-in technique, the stylet was removed and the hub of the needle was observed for several seconds for flow of CSF. When no fluid was observed, the stylet was reinserted into the spinal needle, the needle was advanced approximately 1 mm, the stylet was removed again, and the hub was inspected for flow of CSF. This procedure was repeated multiple times as the spinal needle was slowly advanced toward the subarachnoid space. The procedure was continued until CSF flowed from the hub of the needle or until bone was encountered. When bone was encountered, up to 2 attempts were made to move the needle off the edge of the bone and into the subarachnoid space. When these attempts were unsuccessful, the spinal needle was removed completely and the process was repeated from the beginning with another spinal needle.

For the stylet-out technique, the stylet was removed after the needle was inserted through the skin. The spinal needle was slowly advanced toward the subarachnoid space until CSF flowed from the hub of the needle or until bone was encountered. When bone was encountered, up to 2 attempts were made to move the needle off the edge of the bone and into the subarachnoid space. When these attempts were unsuccessful, the spinal needle was removed completely and the process was repeated from the beginning with another spinal needle.

For both techniques, CSF was allowed to drip from the hub of the spinal needle into a glass vial that contained no anticoagulants. For both techniques, the stopwatch was stopped when CSF was observed flowing into the hub of the spinal needle. The time to collect CSF and number of attempts required were recorded. When multiple attempts were required to collect CSF, the stopwatch was stopped at the end of successful CSF collection; therefore, total sampling time included all attempts.

Once a CSF sample was collected, the spinal needle was removed, and the dog was allowed to recover from anesthesia. Approximately 12 hours after the dogs recovered from anesthesia, physical and neurologic examinations were performed by one of the authors (PMA).

Analysis of CSF samples

Samples of CSF were processed within 1 hour after collection; samples were processed by 2 laboratory technicians who used a specified protocol. A board-certified veterinary pathologist (CVG) evaluated all samples. The pathologist was not aware of the collection technique used to obtain each sample.

Samples initially were subjectively graded for color and turbidity; objective analysis was then performed. Objective measures obtained from the prepared slides included RBC count and protein concentration of the fluid, number of morphologically typical-appearing mature keratinized squamous epithelial cells (total number in 10 randomly selected fields at 200X magnification), amount of myelin-like material (number of fields per slide), degree of cell preservation (number and percentage of poorly preserved cells), and amount of cellular and noncellular debris (number of fields per slide). Except for evaluation of the squamous epithelial cells, the remainder of the slide evaluation was performed with a 50X oil-immersion objective. Overall sample quality was assessed by the pathologist; a sample quality score (0 represented a good-quality sample [ie, diagnostic quality], and 1 indicated a poor-quality sample [ie, likely to not be of diagnostic quality]) was assigned.

Data analysis

Descriptive statistics were calculated for subject age, body weight, and sex. Dichotomous variables included color (colorless or bloody), turbidity (clear or cloudy), and sample quality (0 or 1). Linear mixed-effects regression models were used to analyze continuous outcome variables. Box-Cox analysis was used to choose the best power transformation for each variable to match the assumptions of a linear model. Dog was treated as a random effect. The method, period, and method-by-period interaction were treated as fixed effects. When the interaction effect was not significant, the interaction term was excluded from the final model. Small sample inference for fixed effects was addressed by use of the Kenward-Roger estimation method. When model predictor variables (ie, the method, period, and method-by-period interaction) were significant and the variable required power transformation, then the pairwise comparisons of the back-transformed estimated medians of the outcome variable were further examined by use of the Wald test. When power transformation was not necessary for an outcome variable, then the pairwise comparisons of estimated means of the predictor variable from the model were reported. All P values were adjusted for small sample inference; values of P < 0.05 were considered significant. The residual analysis method was used to evaluate validity of the mixed-effects regression models. All statistical analyses were performed with commercially available software.e

Results

The dogs comprised 5 sexually intact females and 5 sexually intact males. Median age was 11.5 months (range, 10 to 18 months). Median body weight was 9.6 kg (range, 8.7 to 11.3 kg). Results of preanesthetic physical and neurologic examinations were unremarkable for all dogs. No significant abnormalities were detected in results of the preanesthetic hematologic analyses (CBC and biochemical analysis). Results of physical and neurologic examinations conducted approximately 12 hours after recovery from anesthesia were unremarkable for all dogs.

Both techniques were successfully performed during 2 separate anesthetic episodes on all 10 dogs. Of the 20 CSF samples, 18 were collected on the first attempt, 1 was collected on the second attempt, and 1 was collected on the third attempt. Nineteen of 20 samples were colorless with no turbidity (ie, clear); the remaining sample was slightly red (indicative of blood) and slightly cloudy. Sample quality score was 0 (good) for 17 samples and 1 (poor) for 3 samples (the slightly red, slightly cloudy sample and 2 other samples). Myelin-like material was detected in 2 fields in 1 sample and 1 field in each of 2 other samples. In view of these results, no further statistical analysis was deemed relevant.

Linear mixed-effects regression models were used to analyze each continuous outcome variable (RBC count, protein concentration, degree of cell preservation, and amount of cellular and noncellular debris). Analyses revealed that the predictor variables (collection method and collection period) did not have a significant effect on any of these variables.

For the outcome of variable squamous epithelial cells, mixed-model analysis revealed that the number of squamous epithelial cells within the CSF did not differ significantly (P = 0.57) between the 2 collection methods. Estimated median number of squamous epithelial cells for the stylet-in method was 6.374 (95% CI, 3.630 to 9.886 squamous epithelial cells), whereas the estimated median number for the stylet-out method was 7.433 (95% CI, 4.440 to 11.190 squamous epithelial cells). However, there was a significantly (P = 0.023) higher number of squamous epithelial cells in samples obtained by use of both methods during the second collection period (estimated median, 9.618 squamous epithelial cells; 95% CI, 6.160 to 13.840 squamous epithelial cells) than during the first collection period (estimated median, 4.622 squamous epithelial cells; 95% CI, 2.343 to 7.669 squamous epithelial cells).

The amount of cellular debris was significantly (P = 0.031) less in samples collected by use of the stylet-out technique (estimated mean, 17.1 fields; 95% CI, 11.9 to 22.3 fields) than by use of the stylet-in technique (estimated mean, 26.9 fields; 95% CI, 21.7 to 32.1 fields). The amount of cellular debris also was significantly (P = 0.017) greater in samples obtained during the second CSF collection period (estimated median, 27.6 fields; 95% CI, 22.4 to 32.8 fields) than during the first collection period (estimated median, 16.4 fields; 95% CI, 11.2 to 21.6 fields).

Sampling efficiency was evaluated by counting the number of collection attempts and determining the amount of time required to collect a CSF sample. Eighteen of 20 CSF samples were collected in 1 attempt; both of the collections that required multiple attempts (2 and 3 attempts) were performed during the second collection period.

Collection times during the first collection period did not differ significantly (P = 0.570) between the stylet-in and stylet-out techniques. However, during the second collection period, the stylet-out technique was performed significantly (P < 0.001) more rapidly. Because sample collection may fail for a number of reasons, further analysis was performed to investigate whether there was a significant difference in collection time between methods when only 1 attempt was needed for successful collection of CSF. Data for 2 dogs were excluded from this analysis because 2 or 3 attempts were needed for collection of CSF from these dogs during the second collection period. For this analysis of data for the remaining 8 dogs, period of collection did not have a significant (P = 0.927) effect on collection time for the stylet-in (estimated median, 48.11 seconds; 95% CI, 40.45 to 55.77 seconds) and stylet-out (estimated median, 47.57 seconds; 95% CI, 39.91 to 55.23 seconds) techniques. However, there was a significant (P = 0.009) effect for method of collection on collection time. Collection time for the stylet-in technique (estimated median, 58.39 seconds; 95% CI, 50.73 to 66.05 seconds) was approximately 21.1 seconds longer than the collection time for the stylet-out technique (estimated median, 37.29 seconds; 95% CI, 29.63 to 44.95 seconds).

Discussion

Results of the study reported here indicated that both techniques could be safely used to obtain a CSF sample from healthy dogs. There was a significant difference in samples between the 2 techniques because cellular debris was found in a higher number of fields in samples obtained by use of the stylet-in technique. Cellular debris consisted of remnants of ruptured cells (cell type was not recognizable). Cells can also be damaged during sample handling and processing, but because all CSF samples were processed in a standardized manner, damage during handling and processing should not have been a factor that impacted the amount of cellular debris in the samples. However, the amount of cellular debris would not likely be of clinical importance because it did not affect results of sample analysis or the sample quality score. Furthermore, sample quality score did not differ significantly between collection techniques. Therefore, the null hypothesis that there would be no difference in CSF sample quality between the 2 collection techniques was accepted.

Higher numbers of squamous epithelial cells were detected in samples obtained during the second collection period. Individual or randomly scattered morphologically typical-appearing keratinized squamous epithelial cells are surface contaminants commonly seen in cytologic evaluations that are attributable to the patient or as a result of inadvertent contamination during handling of the samples and glass slides. The presence of these cells in CSF is likely of no importance and not related to neurologic disease.6 Therefore, the clinical relevance of this finding is likely to be negligible.

Regarding efficiency of sample collection, it was found that CSF samples could be obtained significantly more rapidly by use of the stylet-out technique than with the stylet-in technique. Interestingly, sample collection time during the first collection period did not differ significantly between the collection techniques. During the second collection period, multiple sampling attempts were required to obtain CSF from 2 dogs (one dog required 2 attempts, and the other required 3 attempts). Both of these samples were for the stylet-in technique. Therefore, analysis was conducted without the data for these 2 dogs. This analysis revealed that a CSF sample was obtained significantly more rapidly by use of the stylet-out technique (approx 21.1 seconds more rapidly) than with the stylet-in technique. Therefore, the null hypothesis that there would be no difference in efficiency between the 2 collection techniques was rejected. This finding may not be of clinical importance because CSF samples were obtained in < 60 seconds by use of both techniques. However, it should be mentioned that the author who performed all CSF collections (PMF) was experienced only with the stylet-in technique. The difference in efficiency (amount of time required to collect a sample) between the 2 techniques may be more important when less-experienced veterinarians collect samples. In addition, minimizing the amount of time for anesthesia should minimize risks associated with anesthetizing patients with neurologic abnormities.7

In the present study, we attempted to standardize the CSF collection process to minimize bias by having the same author perform all CSF collections, the same 2 technicians perform laboratory processing of the samples within 1 hour after collection, and the same veterinary pathologist analyze all slides. The preinduction process was also standardized, as was the time of CSF collection after induction. However, we did not standardize the induction agent. Two agents (propofol and alfaxalone) were used because the dogs were concurrently included in another unpublished study involving laryngeal examination. There was an interval of at least 15 minutes between anesthetic induction and collection of CSF, which should have allowed for clearance of the induction agent. Effects of the induction agent on sample collection and quality were not investigated in this study. In addition, a further limitation of the study was the small number of dogs. It would be ideal to perform a larger study with standardization of the induction agent.

Both techniques were safely performed in healthy anesthetized dogs in the present study. The stylet-out technique was performed slightly more rapidly and yielded CSF samples with less cellular debris, whereas the stylet-in technique yielded samples that contained more cellular debris. The clinical impact of these findings is not likely to be of importance; therefore, either technique can be used in clinical practice to obtain CSF samples of good diagnostic quality.

Acknowledgments

Supported by the Companion Animal Trust Fund, Atlantic Veterinary College, University of Prince Edward Island. The funding source did not have any involvement in the study design, data analysis and interpretation, or writing and publication of the manuscript.

The authors thank Dr. Henrik Stryhn and Gabrielle Monteith for assistance with statistical analyses.

ABBREVIATIONS

CI

Confidence interval

Footnotes

a.

Baxter International Inc, Deerfield, Ill.

b.

Alfaxan, Jurox Pty Ltd, Kansas City, Mo.

c.

Germi-Stat gel chlorhexidine gluconate 4%, Germiphene Corp, Brantford, ON, Canada.

d.

Healthcare plus isopropyl 70% rubbing compound, Canadian Custom Packaging, Toronto, ON, Canada.

e.

Stata statistical software, version 15, StataCorp LP, College Station, Tex.

References

  • 1. Bohn AA, Wells TB, West CL, et al. Cerebrospinal fluid analysis and magnetic resonance imaging in the diagnosis of neurologic disease in dogs: a retrospective study. Vet Clin Pathol 2006;35:315320.

    • Search Google Scholar
    • Export Citation
  • 2. Cook JR, DeNicola DB. Cerebrospinal fluid. Vet Clin North Am Small Anim Pract 1988;18:475499.

  • 3. Di Terlizzi R, Platt SR. The function, composition and analysis of cerebrospinal fluid in companion animals: part I—function and composition. Vet J 2006;172:422431.

    • Search Google Scholar
    • Export Citation
  • 4. Di Terlizzi R, Platt SR. The function, composition and analysis of cerebrospinal fluid in companion animals: part II—analysis. Vet J 2009;180:1532.

    • Search Google Scholar
    • Export Citation
  • 5. Mohammad-Zadeh LF, Sisson AF. Cerebrospinal fluid collection and analysis. NAVC Clin Brief 2005;21–23.

  • 6. Wessmann A, Volk HA, Chandler K, et al. Significance of surface epithelial cells in canine cerebrospinal fluid and relationship to central nervous system disease. Vet Clin Pathol 2010;39:358364.

    • Search Google Scholar
    • Export Citation
  • 7. Greene SA. Anesthesia for patients with neurologic disease. Top Companion Anim Med 2010;25:8386.

Contributor Notes

Dr. Shamir's present address is Department of Clinical Sciences, College of Veterinary Medicine and Biomedical Sciences, Colorado State University, Fort Collins, CO 80523.

Dr. Hagen's present address is Department of Clinical Studies, Ontario Veterinary College, University of Guelph, Guelph, ON N1G 2W1, Canada.

Dr. Amsellem's present address is Department of Veterinary Clinical Sciences, College of Veterinary Medicine, University of Minnesota, Saint Paul, MN 55108.

Address correspondence to Dr. Amsellem (amsellem@umn.edu).