Accessibility Skip to Global Navigation Skip to Local Navigation Skip to Content Skip to Search Skip to Site Map Menu

Optimising the use of lumbar puncture in patients with suspected sub-arachnoid haemorrhage

A 2018/2019 Summer Studentship research project

This research may may lead to a reduction in unnecessary use of unpleasant lumbar punctures, which are time consuming and occasionally harmful.

Student: Annie Vincent
Supervisors: Dr Scott Pearson (Consultant Emergency Physician, Emergency Department, Christchurch Hospital), Dr Martin Than (Consultant Emergency Physician, Emergency Department), Dr James Weaver (Consultant Emergency Physician, Emergency Department), Associate Professor John Pickering (Acute Medicine Research Fellow, Emergency Care Foundation, Christchurch)
Sponsor: TBC

Project brief


Atraumatic  sub-arachnoid haemorrhage (SAH) is a type of brain haemorrhage that results from blood leaking from an aneurysm in 90% cases. The remaining 10% are due to perimesencephalic bleeds which are non aneurysmal. SAH can be catastrophic and lead to death.

In many patients the first SAH is a sudden-onset headache which results from a small, initial, bleed from the aneurysm. Usually such “warning headaches” are self-limiting and the headache can gradually subside. However, these “warning bleeds” are often an indicator of a catastrophic brain haemorrhage to come. Consequently, patients with symptoms of warning bleeding undergo thorough investigation to identify if they need an urgent procedure to prevent the subsequent catastrophic bleed.

Most patients presenting with symptoms of possible SAH do not have this pathology, Yet, understandably, such is the concern about missing such an important diagnosis that clinicians often investigate patients with even the slightest suspicion that this problem exists.

The key initial test is a brain CT-scan.  Historically studies have suggested that this is very accurate, detecting ~ 99% of SAH. However, the exact figure is not clear. It is usual clinical practice for patients undergoing a CT for SAH to also undergo a lumbar puncture (LP) which can detect the presence of blood products in the spinal fluid which are a very clear indicator of haemorrhage having taken place. Unfortunately, LP is a) unpleasant for the patient b) time consuming in a busy emergency department, and c) can lead to complications such as ongoing severe headache or local bleeding at the needle puncture site. It can also be a technically difficult procedure in certain patients. However it is also understood that the sensitivity of CT for detecting subarachnoid blood declines with time from index headache.

As the technology of CT-scanners has improved clinicians have increasingly questioned whether the performing of LP is still needed on all patients and if it now only provides minimal incremental gain in the context of the disadvantages of performing it.

If it were demonstrated that the current generation of CT-scanners could detect a very high percentage of SAH (e.g. 99.5%), then it is likely that most clinicians and patients would believe a lumbar puncture to be unnecessary or at least only used in a certain subset of those who present in a delayed fashion after their index headache.


To establish the accuracy (sensitivity) of current generation CT-scanning to identify SAH.


The student will:

Identify patients diagnosed with SAH at Christchurch Hospital in the past 10-years through ICD10 coding. As a double check, the neurosurgical database may be checked for consistency.

Identify patients that had a CT-scan prior to the SAH diagnosis. Natural language processing will be used to identify all scans which were clearly reported as negative. Review all “non-clearly negative” scans.

Determine the primary outcome: This will be the proportion of all patients with SAH that had a positive CT-scan result (sensitivity).

Assess secondary outcomes:

  1. Descriptive analysis of the use of other investigations including LPs, MRI scans and CT angiograms
  2. Thematic analysis of the clinical presentation pattern of patients that have false negative results including age, sex, and ethnicity.