Traumatic lumbar punctures in children and contributory factors: A 10 year retrospective study

Abstract

Background: Traumatic lumbar puncture (TLP) occurs due to puncturing venous plexus or vessels along the corda equina, causing bleeding in subarachnoid space1. TLP may cause diagnostic uncertainty, unnecessary antibiotic use, prolonged hospital stay and harm to the patients2-4. There are varying definitions for TLP in the literature ranging from >400 to >1000 red blood cells (RBC) in the cerebrospinal fluid (CSF)2-4. Aim: This study reviewed the lumbar punctures performed at the Queensland Children's Hospital and the Royal Children's Hospital over a 10-year period. The primary aim for this study was to evaluate the identifiable risk factors related to traumatic lumbar punctures in a large cohort of paediatric patients at tertiary care facility. Methods: A retrospective study of lumbar punctures was conducted on children seen at two tertiary paediatric centres in Brisbane. Data were extracted from the hospital database and medical records of children who underwent a lumbar puncture during the study period from 1/1/07 to 30/6/16. Non-identifiable data including sex and age as well as variables related to the lumbar puncture including CSF cytology, location and time of the lumbar puncture was collected. A TLP was defined as a lumbar puncture (LP) with >400 red blood cells (RBC) present. An alternative definition of >1000 RBC was also considered. Statistical analyses were performed using STATA 15.0 software. Categorical variables were described in proportions with 95% confidence intervals (CI). Continuous variables were presented in means and standard deviations if normally distributed or medians with interquartile ranges if non-parametric. Results: There were 16,196 LPs performed during the study period. 3083 LPs [19.04% (CI 18.43%, 19.65%)] had >400 RBC, 2301 LPs [14.21% (CI 13.67%, 14.75%)] had >1000 RBC and 2219 LP's [13.70% (CI 13.17%, 14.24%)] were grossly haemorrhagic. In univariate analysis, younger children particularly infants less than one month had a statistically significant increase in risk of TLP with Odds Ratio (OR) 4.3 (CI 3.6, 5.1). There was significant increase in the TLP when performed afterhours with OR 3.6 (CI 3.3, 3.9). There was an apparent increased risk of TLP (>400 RBC) in ED only in the univariate analysis. In the multivariate analysis, adjusted odds ratios indicated significantly increased risk of TLP in infants less than one year and especially in the newborns and for LP's performed afterhours. Conclusions: TLP can occur in approximately 20% of LPs. Children under 12 months especially newborns are at higher risk of TLP as well as for LP's performed afterhours, with a peak after 8 pm. The study results suggest that providing supervision for LPs for infants, particularly for newborns and for LP's performed afterhours is likely to decrease the TLP rates with a potential decrease in unnecessary treatment, antibiotic use and length of stay in hospital.L6279138192019-06-06 <br />