When Is Lumbar Puncture Contraindicated in Children and Young People with Suspected Bacterial

When is lumbar puncture contraindicated in children and young people with suspected bacterial meningitis?

When is lumbar puncture contraindicated in children and young people with suspected meningococcal septicaemia?

Should lumbar puncture be performed prior to stopping antibiotic treatment in children less than 3 months of age with bacterial meningitis?

Bibliographic information / Study type & evidence / Study details / Number of patients / Patient characteristics / Intervention & comparison / Follow up & outcome / Results & effect Size / Reviewer comment
Selby 1994 86 / Prospective survey
Evidence level: 3 / Aim: to identify risks of performing lumbar puncture (LP) and poor outcomes associated with not performing lumbar puncture.
Study site: Children’s hospital 1st July 1991 to 30th April 1992. / N=218 / Children admitted to hospital with suspected meningitis.
Age range : 6 weeks to 13 years.
Included: children with febrile convulsions, children transferred from another hospital with no LP performed.
Excluded: neonates, children transferred from another hospital who had already undergone an LP. / Lumbar puncture vs. no lumbar puncture – survey of usual practice. / Outcomes: Timing of lumbar puncture in relation to admission; diagnosis; mortality; neurological morbidity; adverse effects of LP inc. cerebral herniation.
Follow up: In patient data only (timescale not defined). / Timing of LP:
Immediate LP n=195
Delayed LP n=17
No LP n=6
No reported adverse effects of LP
Diagnosis:
Initial LP: Bacterial meningitis n=18
Viral meningitis n=31
Not meningitis n=146
Delayed LP:
Bacterial meningitis n=3
Meningitis - unconfirmed organism n=1
No LP:
Bacterial meningitis n=5
No diagnosis n=1
No reported neurological morbidity on discharge.
1 child died (unrelated to LP)
Eleven of the lumbar punctures were defined as traumatic and two children required repeated attempts
In nine of the 18 children with bacterial meningitis the lumbar puncture provided information that was defined by the authors as useful in deciding the appropriate management of the children. Twenty-three children did not have an immediate lumbar puncture. The main reason for delaying lumbar puncture was severe obtundation, usually with a Glasgow Coma Scale score of 7 or less. Seventeen children had a lumbar puncture later. In seven children the lumbar puncture was delayed due to suspected raised intracranial pressure. A lumbar puncture was performed after a cranial computed tomography (CT) scan showed no abnormalities. Three children in the delayed lumbar puncture group had bacterial meningitis. / Country: Australia
Funding: One author funded by the Helen McMillan Bequest
Kneen et al., 2002 87 / Retrospective survey
Evidence level: 3 / Aim: to describe practice at study hospital and identify contribution of lumbar puncture to diagnosis and management of care.
Review of medical records at a paediatric secondary and tertiary referral hospital January 1st to April 30th 2000. / Total N=95
Suspected central nervous system (CNS) infections n=52
Suspected meningococcal septicaemia n=43 / Children admitted to hospital with suspected CNS infection or suspected meningococcal septicaemia.
Median age 14 months, range: 9 days to 16 years.
Included: Suspected CNS infection: children with febrile illness and at least one of the following: neck stiffness, bulging fontanelle, photophobia, severe headache, irritability, reduced level of consciousness, focal neurological signs or convulsions.
Also included all sick infants less than 6 months of age who had no obvious focus for infection.
Excluded: Patients with long term medical problems. / Lumbar puncture vs. no lumbar puncture – survey of usual practice and subsequent management. / Outcomes:
Appropriate use of LP (where no contraindication noted as defined by authors); diagnosis; appropriate antibiotic therapy. / Suspected CNS infection: 22 suspected meningitis and no rash; 13 suspected meningitis with meningococcal rash; 17 infants non-specifically unwell and febrile.
LP was contraindicated in 5 children with suspected meningitis with meningococcal rash. Of remaining 47 children with suspected CNS infection and no contraindications 25 (53%) had an LP performed.
43 children had suspected meningococcal septicaemia without CNS involvement, none of this group had LP performed.
CSF analysis was abnormal in 7 of 25 children who had an LP performed (28%). CSF culture was positive in 3 of these children, all with negative blood cultures.
Sterile CSF cultures at 48 hours enabled 15 of these 25 (60%) children to have antibiotics discontinued early. 3 of the 22 children who did not have an LP had antibiotics discontinued early (p<0.001). / Country: UK
Funding: Not reported
Rennick et al., 1993 88 / Retrospective survey
Evidence level: 3 / Aim: To investigate whether incidence of cerebral herniation is increased immediately after LP in children with bacterial meningitis [and whether any children with herniation have normal results on computed tomography].
Review of medical records at a paediatric referral centre 1984-1989 (N=445). / N=19 children with cerebral herniation. / Children over 30 days old admitted to hospital and later diagnosed with bacterial meningitis .
Mean age: 2.0 years (range 4 months to 15 years). / Lumbar puncture. / Outcomes: Timing of herniation in relation to LP.
Herniation defined as having occurred when it was either found on post-mortem examination or when at least 3 of the following 5 clinical criteria were met:
Depressed conscious state – difficult to rouse or unrousable;
One or both pupils dilated or unresponsive to light;
Hyperventilation, irregular respiration, apnoeas, or respiratory arrest;
Decorticate posture, decerebrate posture or complete flaccidity;
Endotracheal intubation and mechanical ventilation instituted s soon as possible after recognition of these findings.
Follow up: Timing of follow up not stated but does include testing of hearing and neurological impairment after discharge. / 19 children diagnosed as having cerebral herniation (21 episodes – 2 children had 2 episodes of herniation).
8 of the 21 episodes of herniation occurred within 3 hours of LP being performed.
3 episodes of herniation occurred between 18.5 and 40.5 hours after LP.
4 episodes occurred between 3 and 12 hours of LP.
6 episodes of herniation occurred before LP or in a child who did not undergo LP.
At the time of lumbar puncture three children were unresponsive to pain, three were drowsy but rousable, one had a purpuric rash and clonus of the right ankle, another had neck stiffness, and one had decerebrate posturing and a rash. Outcomes for children who had cerebral herniation were very poor. Fourteen of the children died; two had no long-term sequelae reported, one had hearing loss and behavioural problems noted on follow-up (timing not noted) and two were discharged with serious neurological impairment. / Country: Australia
Funding: Not reported.
Note: this observational study is not able to demonstrate cause and effect.
Benjamin et al., 1988 89 / Retrospective case control study
Evidence level: 2– / Aim: To identify features of meningitis associated with cerebral herniation and death.
Retrospective review of medical records from 2 children’s tertiary referral centres 1974-1985. / N=38 children / Included: Children who were diagnosed with meningitis and who had had a lumbar puncture.
Study group n=19 children who died.
Matched control group n=19 children who recovered.
Children matched for: year of admission, gender, age and infecting micro-organism.
Excluded: Children with malformations of the central nervous system; preterm or low birthweight neonates. / Features associated with death as an outcome. / Signs associated with raised intracranial pressure:
Fits on admission;
Glasgow coma score;
Conscious level reported by parents prior to admission.
Follow up: In-patient and post-mortem data only. / Fits on admission: 5/17 vs. 0/17
RR 7.08 [95% CI 2.2 to 22.1]; p=0.02.
Glasgow coma scale score <8:
10/17 vs. 4/17 RR 4.6 [95% CI 1.06 to 35.8]; p=0.03.
Conscious level prior to admission:
Waking spontaneously: 13/18 vs. 16/17
RR 6.9 [95% CO 0.74 to 66.6]; p=0.09.
Could not be woken at home: 4/17 vs. 2/17
RR 2.3 [95% CI 0.36 to 35.7]; p=0.24.
Difficult to wake at home:
14/17 vs. 13/16
RR 1.1 [95% CI 0.19 to 6.3]; p=0.34. / Country: UK
Funding: Not reported
Akpede et al.,
2000 209 / Retrospective and prospective survey.
Evidence level: 3 / Aim: To determine the frequency and outcomes of possible cerebral herniation in relation to LP.
Review of medical records from a teaching hospital, 52 retrospectively (1993-1995) and 71 prospectively (mid 1995 to mid 1996). / N=123 children / Children with pyogenic meningitis. Mean age 4.07 years, range 6 weeks to 15 years.
2 risk categories compared. Risk defined by score obtained for clinical features of the Herson and Todd’s scoring system. A weighted score is given which includes: unrousable coma (3 points), hypothermia (2 points), convulsions (2 points), shock (1 point), age < 12 months (1 point) and symptoms persisting > 3 days (0.5 point). Cut-off score of >=4.5 points.
High risk: Severly ill, cut off score > 4.5 points (n=25).
Low risk: Mild to moderately ill, cut off score of < 4.0 (n=98). / Lumbar puncture vs. no lumbar puncture. High risk vs. low risk patients. / Outcome: Incidence and timing of cerebral herniation in high risk vs. low risk patients.
Herniation defined as having occurred when at least 2 of the following 4 clinical criteria were met:
Depressed conscious state –unrousable;
One or both pupils dilated or unresponsive to light;
Hyperventilation, irregular respiration, apnoeas, or respiratory arrest;
Decorticate posture, decerebrate posture or complete flaccidity. / LP performed on presentation in 112 patients (91%) and deferred in 11.
Timing of herniation:
None pre or post LP:
6/18 high risk vs. 86/94 low risk
RR 0.4 [95% CI 0.2 to 0.7]; p<0.0001].
Pre plus post LP:
4/18 vs. 0/94; p=0.0004.
Pre LP only:
7/18 vs. 0/94; p<0.0001
Post LP only:
1/18 vs. 8/94
RR 0.6 [95% CI 0.1 to 4.9]; p=1.0.
17 children who had an LP on presentation died, including 7 within 24 hours.
8 children who had deferred LP died, 7 within 24 hours of the procedure. 7 children in this group were classified as high risk, 6 of whom had cerebral herniation. / Country: Nigeria (Note: this setting is described as a low-income setting with no neurological support available).
Funding not reported.
Kanegaye et al., 2001 91 / Retrospective survey
EL=3 / Review of medical records.
Data extracted by one reviewer and checked by a second.
Inclusion criteria: Positive CSF culture; CSF pleocytosis with positive CSF Gram stain alone or in conjunction with positive antigen test; CSF pleocytosis with positive blood culture alone or in conjunction with a positive CSF Gram stain, antigen test or both; by CSF indices alone.
Exclusion criteria: positive CSF viral study in the absence of a known bacterial pathogen; neural tube defect; CSF shunt or catheter; penetrating cranial injury; neurosurgical procedure other than LP; medial record unavailable or incomplete. / N=128 included records.
Neonatal sub-group with Group B streptococcus infection n=21. / Overall age range 1 day to 16 years, median age 8 months.
Of these a sub-group with Group B streptococcus infection were predominantly neonates: median age 21 days (interquartile range: 9 days, 31 days). / Administration of parenteral antibiotics / 72+ hours
Sterility of CSF. / Results reported for neonates where CSF not sterile prior to administration of parenteral antibiotics (n=15).
0 – 1.0 hours: 0/3
5.1 – 6.0 hours: 0/1
6.1 – 24 hours: 0/1
24.1 – 48.0 hours: 2/2
48.1 – 72.0 hours: 1/2
72.1: 6/6 / Country: USA
Funding: Not reported.