Comparison of Fluid Analysis and Cytologic Findings of Cerebrospinal Fluid Between Three Collection Sites in Adult Equids With Neurological Disease.
Authors: Young Kimberly A S, Hepworth-Warren Kate L, Dembek Katarzyna A
Journal: Frontiers in veterinary science
Summary
# Editorial Summary: CSF Collection Site Affects Cell Counts in Equine Neurological Cases Young and colleagues analysed 113 horses with suspected neurological disease to establish whether CSF samples collected from the three standard sites—atlanto-occipital, atlantoaxial, and lumbosacral—yielded comparable results for diagnostic purposes. Using retrospective data from 2008–2020, the team compared total nucleated cell count (TNCC), red blood cell concentration, total protein, and cytological findings across collection sites. Lumbosacral centesis consistently produced higher RBC contamination than cranial sites, whilst protein concentrations and cell counts varied between locations depending on whether cytological abnormalities were present; importantly, abnormal cytology itself strongly predicted non-survival (p = 0.0002), and elevated TNCC >24 cells/µL showed modest predictive value for mortality (sensitivity 54%, specificity 72%, negative predictive value 78%). For practitioners, these findings underscore the critical importance of knowing the collection site when interpreting results—lumbosacral samples require particular caution regarding RBC contamination artefacts—and emphasise that whilst most parameters show minimal clinically relevant differences between sites, cytological abnormalities carry significant prognostic weight regardless of collection location.
Read the full abstract on PubMed
Practical Takeaways
- •When comparing CSF results between horses, account for higher RBC counts from lumbosacral puncture—this is normal variation from site choice, not necessarily pathological
- •Collection site has minimal practical impact on clinical interpretation when cytology is abnormal, so choose your preferred site based on safety and accessibility rather than diagnostic accuracy concerns
- •Total nucleated cell count >24 cells/μL carries poor prognosis (only 45% positive predictive value), but normal counts are reassuring with 78% negative predictive value—interpret alongside clinical signs and imaging
Key Findings
- •Lumbosacral CSF collection yielded significantly higher RBC counts compared to atlanto-occipital and atlantoaxial sites (p < 0.05), with highest concentrations in cytologically unremarkable samples (p < 0.01)
- •Total protein was lower in atlantoaxial samples than lumbosacral samples (p < 0.05), but differences were minimal and clinically insignificant in most cases
- •Total nucleated cell count ≥24 cells/μL predicted non-survival with 72% specificity and 54% sensitivity (AUC 0.67), and abnormal cytology correlated with non-survival (p = 0.0002)
- •Within cytologically abnormal samples, no significant differences existed between collection sites for any fluid analysis parameter, limiting the clinical impact of site selection in diseased animals