Evidence review for role of neuroimaging prior to lumbar puncture

Included studies

Three cohort studies were included for this review: 2 prospective cohort studies (Glimaker 2018, Hasbun 2001) and 1 retrospective cohort study (Glimaker 2015). No relevant test and treat randomised controlled trials were identified.

The included studies are summarised in Table 2.

All studies compared lumbar puncture without prior computerised tomography (CT) with lumbar puncture after CT in adults only. Only 1 study provided adjustment for confounding factors for relevant outcomes (Glimaker 2018).

See the literature search strategy in appendix B and study selection flow chart in appendix C.

Excluded studies

Studies not included in this review are listed, and reasons for their exclusion are provided in appendix J.

Included studies

A single economic search was undertaken for all topics included in the scope of this guideline, but no economic studies were identified which were applicable to this review question. See the literature search strategy in appendix B and economic study selection flow chart in appendix G.

The outcomes that matter most

Brain herniation can occur following a lumbar puncture if there is raised intracranial pressure. As this is a potentially life-threatening complication, brain herniation and mortality were selected as critical outcomes due to the potential for neuroimaging to identify raised intracranial pressure and indirectly reduce the risk of brain herniation by impacting the decision to perform a lumbar puncture. Long-term neurological impairment was also prioritised as a critical outcome as it can be a complication of both brain herniation and meningitis itself.

Time between hospital admission and 1) lumbar puncture and 2) starting antibiotics were included as important outcomes due to concerns from the committee that the time required for neuroimaging can cause delays in both of these, which may have a detrimental effect on outcomes. As with long-term neurological impairment, functional impairment may be a complication of both brain herniation and meningitis itself, so it was selected as an important outcome in adults. However, severe developmental delay was included as an important outcome for children as this may be more commonly reported than neurological and functional impairment in trials of young children. Serious intervention-related adverse effects leading to death, disability or prolonged hospitalisation or that are life threatening or otherwise considered medically significant was also included as an important outcome to identify serious adverse effects other than brain herniation, for example, medically significant bleeding.

The quality of the evidence

The quality of the evidence was assessed with GRADE methodology and was rated as moderate quality for the absence of functional impairment outcome due to the population being indirect (including people who were immunocompromised), and low to very low quality for the remaining outcomes due to high risk of bias (arising from lack of adjustment for confounding factors), the population being indirect, and imprecision (due to wide confidence intervals and small number of events).

No evidence was found that reported on brain herniation, severe developmental delay or serious intervention-related adverse effects leading to death, disability or hospitalisation.

Benefits and harms

The committee considered the evidence comparing outcomes for people who underwent lumbar puncture without prior computerised tomography (CT) relative to those who underwent lumbar puncture with prior CT. Lumbar puncture without prior CT reduced mortality compared with lumbar puncture performed after neuroimaging. Lumbar puncture without prior CT was also associated with lower rates of neurological and/or hearing deficits and functional impairment, and a shorter time to antibiotic treatment (with or without corticosteroids), relative to lumbar puncture after CT. These findings were consistent with the clinical expertise of the committee, and they agreed that neuroimaging should not be routinely performed before lumbar puncture.

No evidence was identified for the critical outcome of brain herniation. Based on their clinical knowledge and experience, the committee recognised the potential for neuroimaging to identify raised intracranial pressure and indirectly reduce the risk of brain herniation by impacting the decision to perform a lumbar puncture. Based on evidence from the review on factors associated with brain herniation (see evidence review B4) and their knowledge and experience, the committee agreed that there are specific instances when imaging should be performed before lumbar puncture to mitigate the risk of brain herniation, namely when a person shows recognised signs of raised intracranial pressure (new focal neurological features, abnormal pupillary reactions, a Glasgow coma scale score of 9 or less, or a progressive and sustained or rapid fall in level of consciousness). The committee also agreed that imaging should be performed prior to lumbar puncture where the person has risk factors for an evolving space-occupying lesion. However, the committee discussed that neuroimaging should not cause a delay to management of bacterial meningitis and thus recommended that a blood sample should be taken, antibiotics should be started, and the person should be stabilised before imaging.

Cost effectiveness and resource use

No original economic analysis was undertaken for this review question and therefore the committee made a qualitative assessment of the likely cost-effectiveness of their recommendations.

The committee recognised that undertaking neuroimaging prior to lumbar puncture would increase costs by introducing a diagnostic investigation into the pathway which, as well as delaying optimal treatment for bacterial meningitis, would often provide little or no information that would improve subsequent management. Whilst the committee were aware that much of the evidence reviewed was of low quality, because of a high risk of bias, they did note that neuroimaging prior to lumbar puncture was associated with poorer outcomes having a significant impact on health-related quality of life. Therefore, the committee considered that it was reasonable to conclude that neuroimaging prior to LP was not cost-effective for people with suspected bacterial meningitis, where prompt appropriate antibiotics are critical to favourable outcomes.

Notwithstanding the limitations of the evidence, the committee were aware that their recommendations were consistent with other international guidance as well as recommendations made in previous NICE guidance. However, the committee noted that poor adherence with guidelines has been observed in the US, UK, the Netherlands, and Sweden (Salazar 2017, Ellis 2022, Costerus 2016, Glimaker 2018). A recent retrospective cohort study in the UK (Ellis 2022) in a population with community acquired meningitis, found that neuroimaging prior to lumbar puncture occurred in 94% of patients, even though the majority of these (83%) had no contraindication to lumbar puncture. Less than 1% of patients had lumbar puncture within the first hour after arrival at hospital and only 26% had lumbar puncture within the first 8 hours. The study authors remarked that delays in obtaining CSF is associated with worse pathogen detection, more exposure to unnecessary anti-infectives, increased hospital length of stay and increased mortality. They concluded that “in most cases, brain imaging is not indicated in adults with suspected community-acquired meningitis; however, in our cohort, a significant number of patients had unnecessary scans. Although complications following LP are rare, there may be an unfounded fear of cerebral herniation following LP, even in those with no clinical features of brain shift, which is leading to excessive use of imaging.”

In addition to overcaution with respect to brain herniation, the committee also discussed other factors as to why compliance with guidelines might be poor. They noted that clinical assessment often takes place in a busy emergency department which they thought would not always be conducive to lumbar puncture and where there may be incentives in terms of patient workflow management from providing CT first, as that can facilitate faster patient outflows from the emergency department. The committee noted that the numbers presenting was also quite small which could limit the opportunities to learn from experience. The committee believed that logistically it was easier for LP to be performed quickly on a medical ward and that the time from admission to lumbar puncture would be improved if patients could be moved from the emergency department to an acute medical ward more quickly. The committee recognised that bed capacity could be a major implementation obstacle to this. The committee considered that the availability of LP kits could also be a limiting factor. Equipment shortages and lack of training were also factors picked up in a questionnaire (Defres 2015) considering possible barriers to timely LP. The committee noted that generating requests for all the tests required on a cerebrospinal fluid sample is laborious, and that sometimes tests are inadvertently omitted. It was suggested that electronic order sets could improve practice in this regard.

The committee believed that their recommendations made it clearer that neuroimaging should not be routinely undertaken. Whilst the recommendations do not substantively change current guidance the committee recognised that current practice is varied and often suboptimal. The committee believed that widespread implementation of their recommendations has the potential to be cost saving to the NHS reducing unnecessary tests, ineffective treatments and hospital stay.

Excluded effectiveness studies

Table 6

Excluded studies and reasons for their exclusion.

Excluded economic studies

No economic evidence was identified for this review.