G Rennick1, F Shann, J de Campo. 1. Intensive Care Unit, Royal Children's Hospital, Parkville, Victoria, Australia.
Abstract
OBJECTIVE: To see whether the incidence of cerebral herniation is increased immediately after lumbar puncture in children with bacterial meningitis and whether any children with herniation have normal results on cranial computed tomography. DESIGN: Retrospective review of case notes; computed tomograms were read again. SETTING: Large paediatric teaching hospital. SUBJECTS: 445 children over 30 days old admitted to hospital with bacterial meningitis. MAIN OUTCOME MEASURES: Timing of herniation in relation to lumbar puncture; findings on computed tomography in children with herniation. RESULTS: Cerebral herniation was detected in 19 (4.3%) of the 445 children (21 episodes; herniation occurred twice in two children). Herniation occurred in 14 (45%) of the 31 children who died. Nineteen episodes of herniation occurred in the 17 children who had a lumbar puncture; 12 of the episodes occurred in the first 12 hours after the lumbar puncture and seven over six other 12 hour periods (odds ratio 32.6 (95% confidence interval 8.5 to 117.3); p < 0.001). The results of cranial computed tomography were normal in five (36%) of the 14 episodes of herniation in which scanning was performed at about the time of herniation. CONCLUSIONS: The temporal relation between lumbar puncture and herniation strongly suggests that a lumbar puncture may cause herniation in some patients, and normal results on computed tomography do not mean that it is safe to do a lumbar puncture in a child with bacterial meningitis.
OBJECTIVE: To see whether the incidence of cerebral herniation is increased immediately after lumbar puncture in children with bacterial meningitis and whether any children with herniation have normal results on cranial computed tomography. DESIGN: Retrospective review of case notes; computed tomograms were read again. SETTING: Large paediatric teaching hospital. SUBJECTS: 445 children over 30 days old admitted to hospital with bacterial meningitis. MAIN OUTCOME MEASURES: Timing of herniation in relation to lumbar puncture; findings on computed tomography in children with herniation. RESULTS:Cerebral herniation was detected in 19 (4.3%) of the 445 children (21 episodes; herniation occurred twice in two children). Herniation occurred in 14 (45%) of the 31 children who died. Nineteen episodes of herniation occurred in the 17 children who had a lumbar puncture; 12 of the episodes occurred in the first 12 hours after the lumbar puncture and seven over six other 12 hour periods (odds ratio 32.6 (95% confidence interval 8.5 to 117.3); p < 0.001). The results of cranial computed tomography were normal in five (36%) of the 14 episodes of herniation in which scanning was performed at about the time of herniation. CONCLUSIONS: The temporal relation between lumbar puncture and herniation strongly suggests that a lumbar puncture may cause herniation in some patients, and normal results on computed tomography do not mean that it is safe to do a lumbar puncture in a child with bacterial meningitis.