ROLE OF DELAY IN SAMPLE PROCESSING AND OTHER FACTORS IN INTERPRETATION OF CEREBROSPINAL FLUID (CSF) CELL COUNTS IN TRAUMATIC LUMBAR PUNCTURE (LP) FOR SUSPECTED MENINGITIS (M)

Abstract

276 LP attempts were examined prospectively by physician questionnaire and medical record review. Factors occurring at the time of LP as well as time required for CSF receipt at the laboratory (DEL) were compared with clinical and laboratory variables. Occurrence of traumatic LP increased with decreasing experience level of person performing LP (p=0.04) and with younger patient age (p=0.0002). DEL averaged 28 min, with 73% of samples ≤ 30 min DEL but 11.5% ≥ 60 min DEL. A standard correction formula using CSF RBC count and blood RBC and WBC counts for specimens with ≥ 10,000 CSF RBC resulted in 11/15 (73%) of non-M samples with negative corrected CSF HBC. Percent fresh RBC ranged from 0 to 100 and did not correlate with DEL (r=0.02). 2 of 48 M patients had negative or normal corrected CSF WBC with this formula. DEL did not correlate with corrected CSF WBC. Linear regression analysis with multiple independent variables (CSF RBC, blood WBC, blood RBC, CSF WBC, and DEL) failed to produce a meaningful correction formula (r=0.47 with best fit). Over-correction of CSF WBC is likely to occur in traumatic LP and is not explained by DEL. Determination of percent fresh RBC in CSF is not helpful in interpretation of traumatic LP in suspected M. Mathematical models to correct CSF WBC for trauma are not clinically useful and should not be used.