• 
• 

Three Part Question

In [adult patients presenting to hospital with a first seizure] is [CT head] necessary [to rule out intracranial pathology that would require management in the ED]

Clinical Scenario

A 19-year-old male presents to your ED after a generalised tonic-clonic seizure. He has never had a seizure before and is otherwise fit and well. You conduct a full neurological examination and request several bedside investigations. You wonder whether you should also request a CT head scan. More broadly, you wonder whether your department should be scanning all patients who present with a first seizure.

Search Strategy

EMBASE and MEDLINE databases were searched with no filters (1974-2024) utilising the Elsevier interface and the following keyword strategy:

ct:ab,ti AND 'first seizure':ab,ti OR (ct:ab,ti AND 'first fit':ab,ti) OR (ct:ab,ti AND 'first epilep*':ab,ti) OR (ct:ab,ti AND 'initial seizure':ab,ti) OR (ct:ab,ti AND 'initial fit':ab,ti) OR (ct:ab,ti AND 'initial epilep*':ab,ti) OR (ct:ab,ti AND 'new* seizure':ab,ti) OR (ct:ab,ti AND 'new* fit':ab,ti) OR (ct:ab,ti AND 'new* epilep*':ab,ti) OR (computed AND tomography:ab,ti AND 'first seizure':ab,ti) [...] OR (computed AND tomography:ab,ti AND 'new* epilep*':ab,ti)
A supplementary search of GoogleScholar and Cochrane databases was then conducted using the same search terms. Reference lists of relevant papers were screened for studies missed by our keywords.

Search Outcome

Our search strategy identified 337 papers. 277 papers were excluded on title and abstract review as they were not relevant to our three-part question.
The remaining 60 studies underwent full text review. Eighteen were excluded because they were conference presentations (n=11), review articles (n=6), or case reports (n=1). Twenty-five were excluded because they were conducted in outpatient or ward settings (n=16) or did not report CT findings (n=9). Two papers appeared to address our three-part question, but had an unclear methodology that we were unable to clarify despite contacting the corresponding authors.
Eight papers were excluded for high risk of selection bias. These were retrospective studies of patients who underwent CT imaging for a first seizure. Missing from these analyses were patients who were not believed to require an urgent scan in ED. We were interested in whether all adults presenting with first seizures would benefit from CT, including low-acuity patients lacking typical indications for neuroimaging. It is our experience that these patients create the most uncertainty in clinical practice.
Accordingly, we restricted our review to seven observational studies in which patients were either recruited prospectively and scanned[Isenberg, Zarmehri] or selected retrospectively from centres where all first seizure patients are scanned[Kotisaaria, Pathan, Tardy, Schoenenberger, Henneman].

All patients were recruited from urban hospitals and the average ages ranged from 35 to 47.

The results are summarised in Table 1.
The diagnostic yield of CT imaging varied significantly. Between 5.8% and 41% of patients had actionable findings – i.e. CT abnormalities that would affect their management in ED. Common reported pathologies [Isenberg, Zarmehri, Kotisaari, Pathan, Tardy] were intracranial haemorrhage, tumours, and infarctions.

Relevant Paper(s)

Author, date and country	Patient group	Study type (level of evidence)	Outcomes	Key results	Study Weaknesses
Isenberg et al 2023 USA	First focal or generalised motor seizure n=242 Mean age 45	Multi-centre prospective study	CT findings	Normal: 88.1% Actionable: 5.8% Other: 6.1%	Low rate of CT abnormalities may be attributable to high proportion (44%) of patients with clear toxicological aetiology for seizure (intoxicated or withdrawing) in sample. Data collection reliant upon potentially incomplete patient notes. Non-motor seizures (e.g. absence seizures) not included in study.
Zarmehri et al 2020 Iran Urban Hospital	First seizure n=100 Mean age 39	Single centre prospective study	CT findings	Normal: 73% Actionable: 13% Other: 14%	No definition of ‘seizure’ provided for patient selection. Pregnant women excluded from study with no rationale described. Unclear whether consecutive sampling was used in recruitment. Unclear whether individuals interpreting CT scans were aware of study objectives.
Kotisaaria et al 2017 Finland	First witnessed convulsive seizure not associated with head injury n=449 Mean age 47 Mixture of contrast and non-contrast CT used	Single centre retrospective study	CT findings	Normal: 47% Actionable: 12% Other: 41%	Exclusion of seizures that were unwitnessed or occurring in non-independent elderly patients may limit generalisability of study findings to uncertain or complex cases. Retrospective data collection reliant on potentially incomplete patient notes. Interpretation of CT scans conducted by clinician aware of study objectives.
Pathan et al 2014 Qatar Urban Hospital	First seizure not associated with head injury n=439 Mean age 35	Single centre retrospective study	CT findings	Normal: 64.7% Actionable: 14.9% Other: 20.4%	Sample may have included a small number of 14-15 year-old patients. High levels of neurocysticercosis (9.2%) in sample unlikely to be replicated outside endemic areas. Retrospective data collection reliant on potentially incomplete patient notes
Tardy et al 1995 France Urban hospital	First witnessed seizure with generalised tonic-clonic movements and post-ictal confusion n=247	Single centre retrospective study	CT findings	Normal: 47% Actionable: 34% Other: 18%	Exclusion of patients with focal, non-motor, or unwitnessed seizures may limit generalisability of study findings to uncertain cases. Retrospective data collection reliant on potentially incomplete patient notes. Use of now-outdated CT technology. Unclear whether individuals interpreting CT scans were aware of study objectives.
Schoenenberger et al 1994 Switzerland Urban Hospital	First generalised seizure <1 hour ago n=119 Mean age 46 Mixture of contrast and non-contrast CT used	Single centre retrospective study	CT findings	Normal: 42.9% Actionable: 33.6% Other: 23.5%	Exclusion of patients presenting to ED >1 hour after seizure may have prevented participation in patients with significant prehospital needs, or difficulty accessing ambulance services. Retrospective data collection reliant on potentially incomplete patient notes. Use of now-outdated CT technology. Unclear whether individuals interpreting CT scans were aware of study objectives
Henneman et al 1994 USA Urban Hospital	First seizure not associated with head injuryn=333 Mean age 39	Single centre retrospective study	CT findings	Normal: 48% Actionable: 41% Other: 11%	Exclusion of intoxicated patients may limit generalisability of findings to complex cases. High levels of neurocysticercosis (12%) in sample unlikely to be replicated outside endemic areas. Retrospective collection of data via review of notes Use of now-outdated CT technology. Unclear whether individuals interpreting CT scans were aware of study objectives.

Comment(s)

Our search returned seven observational studies reporting a high yield of actionable CT findings in patients presenting to ED with a first seizure. Methodological weaknesses are summarised in Table 1. There was significant variability in results. We attribute this in part to patient selection. All studies included generalised seizures but some excluded focal [Tardy, Schoenenberger], unwitnessed [Kotisaari], and non-motor [Isenberg, Zarmehri] seizures. Inclusion of unwitnessed events and atypical symptoms might have lowered the proportion of abnormal scans – because many patients who present to ED in this way have not had a seizure. The lowest rates of intracranial pathology were found in the prospective studies [Isenberg, Zarmehri] that used pragmatic inclusion criteria. Chronological bias may also account for some of the variability seen. There was an inverse linear relationship between year of publication and CT pick-up rate. This was particularly evident in the oldest studies, which reported striking rates (33.6% to 41%) of intracranial abnormalities – most commonly brain tumours [Tardy, Schoenenberger, Henneman]. It is likely that our results reflect the changing demographics of patients attending ED, in addition to improvements in the early detection of cancer. Focusing on the two studies of highest methodological quality [Isenberg, Zarmehri] : for ED clinicians, the number-needed-to-scan for first seizures appears to be between 8 and 17. We believe that this pick-up rate would justify the financial costs of imaging all patients with this presentation. It is equivalent to what has been reported in MRI for suspected cauda equina syndrome [Metcalfe]. A CT head scan is generally £100-200 in the UK. The radiation risk is relatively low: for an 18-year-old, the added lifetime cancer risk is between 1/2492 and 1/4101 [X-Ray Risk Online]. Future research should extend our clinical question as follows: in first-seizure patients with no other indications for a CT head, what is the likelihood that imaging will affect management in ED? None of the studies in our review estimated the proportion of their sample who would likely have received a CT head regardless of local protocols for first-time seizures. How many patients with intracranial bleeds described red-flag headaches? How many with brain tumours demonstrated obvious focal neurology? Questions like these would be best addressed in a prospective cohort study with standardised data collection on clinical findings. However, an appropriately designed retrospective study with a sub-group analysis would be a helpful first step.

Clinical Bottom Line

Adults presenting with a first seizure are a high-yield group for CT head scan, with significant rates of intracranial haemorrhage, infarctions, and tumours. The number-needed-to-scan is sufficiently low to justify the routine use of neuroimaging when these patients present to the ED.

References

1. Isenberg D, Gunchenko M, Hameier A. First-time Seizure Patients Have High Rate of Abnormalities on Emergency Department Performed Brain Computed Tomography: A Prospective Study. The Journal of Emergency Medicine. 2023 65 (5). Pages e432-e437.
2. Zarmehri B, Teimouri A, Ebrahimipour N et al. Brain CT findings in patients with first-onset seizure visiting the Emergency Department in Mashhad, Iran. Open access emergency medicine. 2020 Jun 10:159-62.
3. Kotisaari K, Virtanen P, Forss N et al. Emergency computed tomography in patients with first seizure. Seizure. 2017 May 1;48:89-93.
4. Pathan SA, Abosalah S, Nadeem S et al. Computed tomography abnormalities and epidemiology of adult patients presenting with first seizure to the emergency department in Qatar. Academic Emergency Medicine 2014 Nov;21(11):1264-8.
5. Tardy B, Lafond P, Convers P et al. Adult first generalized seizure: etiology, biological tests, EEG, CT scan. The American journal of emergency medicine. 1995 Jan 1;13(1):1-5.
6. Schoenenberger RA, Heim SM. Indication for computed tomography of the brain in patients with first uncomplicated generalised seizure. BMJ 1994 Oct 15;309(6960):986-9.
7. Metcalfe D, Hoeritzauer I, Angus M, et al. Diagnosis of cauda equina syndrome in the emergency department Emergency Medicine Journal. 2023 Nov 1;40(11):787-93.
8. X-Ray Risk.[Online] Calculate Your Risk: Brain CT (Standard) [Internet]. X-ray Risk. 2024 2024
9. Henneman PL, DeRoos F, Lewis RJ. Determining the need for admission in patients with new-onset seizures. Annals of emergency medicine. 1994 Dec 1;24(6):1108-14.