Every physician is wedded to the notion of ‘primum non nocere’. The NHS in the UK has made patient safety and optimal clinical outcomes as quality indicators of service provision.1 Consequently, it is vital to learn from adverse clinical outcomes and revisit and challenge accepted clinical practice in the light of personal experience. Recently, we saw a child with acute onset worsening headaches and neck stiffness who had a lumbar puncture for suspected bacterial meningitis, confirming the diagnosis. There were no contraindications such as signs of raised intracranial pressure (ICP). One hour later, she abruptly deteriorated. Brain magnetic resonance imaging 12 hours later demonstrated tonsillar descent through the foramen magnum with patchy ischaemic infarction of the brainstem, cerebellum, and high cervical cord. Despite emergency foramen magnum decompression, bilateral facial, bulbar, respiratory, and 4-limb flaccid paralysis with absent sensory and autonomic function below mid-pons developed, with modest recovery subsequently. The National Institute for Health and Care Excellence (NICE) advocates lumbar puncture prior to antibiotic therapy in suspected cases of bacterial meningitis, unless there are signs suggesting raised ICP; shock; extensive/spreading purpura; coagulation abnormalities; convulsions (until stabilized); respiratory insufficiency; and local superficial infection at the lumbar puncture site. In such cases, it recommends delaying the procedure, and using clinical assessment and not brain computed tomography to determine when it is safe to proceed.2 The guideline does not place any emphasis on a history suggestive of raised ICP. Among other features, acute onset worsening headaches has been recognized as a potential symptom of significant raised ICP.3 Also, it makes no recommendation to routinely measure opening pressure at lumbar puncture. The spectre of ‘coning’ leading to death or secondary brain damage in patients with raised ICP is a haunting one. Raised ICP almost always accompanies bacterial meningitis and was reported in 33 out of 35 children with pyogenic meningitis.4 Brain herniation is not common in acute bacterial meningitis, reported in 5% of cases and contributing to 30% mortality.5, 6 Brain herniation has been implicated temporally with performance of lumbar puncture for diagnosis of suspected bacterial meningitis in many reports, usually occurring within 3 to 12 hours of the procedure,5, 6 but can occur spontaneously.7 While it is ideal to aspire to microbiological diagnosis by cerebrospinal fluid studies,8 can lumbar puncture justifiably be delayed or omitted altogether in some circumstances? Given the recent establishment of rapid, reliable polymerase chain reaction (PCR) testing for common causes of meningitis (which remain sensitive for up to 96 hours after antibiotic administration),2, 9 I wonder how justifiable protocolized lumbar puncture is, given its potentially catastrophic sequelae in the face of significant raised ICP. Additionally, are we paying too much attention to clinical examination2, 8 or investigations suggesting raised ICP at the sacrifice of pertinent features in the history, which could inform the decision for lumbar puncture at an earlier stage? The sick child with suspected bacterial meningitis should be assessed by an experienced clinician.10 More attention must be paid to features in the history suggesting the presence of significant raised ICP, which is where caution and judgement should be exercised, rather than just following protocol. Best practice from patient safety and outcome perspective would be to defer lumbar puncture, if clinically considered appropriate and as the guideline recommends, obtain blood for real time PCR and cultures, and treat with appropriate antimicrobial therapy.2, 9 In addition, immediate measures to monitor and reduce ICP should be initiated.6 Bacterial PCR on cerebrospinal fluid can be performed at a later time.2, 9 A systematic national surveillance study on mortality and morbidity of ‘coning’ post lumbar puncture through either the Royal College of Paediatrics and Child Health or the British Paediatric Neurology Association would greatly assist in determining the risk–benefit of this potentially risky procedure.