Table 1 Characteristics of included studies

Barrow et al. [10]

Inclusion criteria:

• Age ≥ 17 years

• GCS 14 or 15 at presentation

• Presenting to ED >4 h after injury

Exclusions:

• Patients with only facial injuries

Single-site: large central-east London teaching hospital

1st Jan. 2008–10th May 2009

Prospective observational study

‘Positive CT’: any traumatic finding related to presenting injury

Detailed patient demographics not reported

497 patients included:

• 193 presented 4–12 h

• 140 presented 12–24 h

• 62 presented 24–28 h

• 58 presented 24–168 h

• 44 presented ≥ 1 week

147/497 (29.6 %) had CT head; 64/147 presented 4–12 h, 50/147 12–24 h, 11/147 24–48 h, 21/147 48 h–1 week, 1/147 > 1 week

11/497 (2.21 %) positive CT scans; 1/11 presented 4–12 h, 3/11 12–24 h, 4/11 24–48 h, 3/11 48–168 h

4/497 (0.80 %) had neurosurgery; 3/4 (75 %) presented 12–48 h after injury, 1/4 (25 %) 48–168 h

1/497 died (0.20 %)—time since presentation not reported

69/497 (13.9 %) contactable at 2 weeks; 11/69 (15.9 %) symptomatic

Lower rates of intra-cranial injury compared to previous studies

Similar rates of CT and neurosurgical intervention compared to previous studies

Statistically significant predictors of intra-cranial injury: LOC, coagulopathy, evidence of injury above the clavicles, open or depressed skull fracture and acute alcohol/drug use

Prospective and contemporaneous review of notes—likely that most eligible cases were identified and included

NICE guidelines used for triage patients to CT head and discharge

2–4 week telephone interview follow-up for further treatment/deterioration

4 h is not a long delay

Data collection: Daily identification of cases from search of paper records and review of computerised discharges

Identification of clinical risk factors predictive of intra-cranial injury

No control or comparison group

Sampling biases: small numbers, young population, >50 % from Indian subcontinent

Small absolute rates of pathology, therefore prone to outlier bias

Very high loss to follow-up

Hemphill et al. [23]

Inclusion criteria:

• Any age

• GCS 15 at presentation

• Presenting to ED >12 h after injury

• Re-attenders included

Exclusions:

• None stated

Dual-site: academic Level I Trauma Centres (San Antonio, USA)

Jan.–Dec. 1996

Retrospective chart review

Searched 85,000 ED charts

‘Significant delayed injury’. Defined as ‘abnormal CT results such as: intracerebral bleeding, skull fracture, or subdural or epidural haematoma’

2900 patients with head injury

194 (6.69 %) presented >12 h:

• 112/194 (56.9 %) female

• 34 ± 24 years (mean ± SD)

• 21/194 (10.8 %) re-attenders

101/194 (52.1 %) patients had CT head; 9/21 (42.9 %) of re-attenders had CT head scan

6/194 (3.1 %) patients had abnormal CT scans:

• 2 infants (aged 1 and 5 months)

• 4 adults (29F, 46F, 60M, 74M)

• Note: one patient (74M) presented GCS 3 with large DSH at 25 h after normal CT acutely after injury—required neurosurgery and died

1 patient re-presented at 3 months with headache: chronic SDH—did not originally have CT head

Mean time to presentation:

• Overall: 73 ± 105 h

• If abnormal CT: 29.3 ± 10.7 h

Retrospective review—data may be missing

Comparisons between patients with/without CT

Exclusions not stated

Comparisons between hospitals

No formal follow-up of patients who did not have CT head scans

Sampling bias: small numbers

Small absolute rates of pathology, therefore prone to outlier bias

Includes re-attenders: a distinct and possibly higher-risk group than delayed (first-time) presenters

Borczuk et al. [28]

Abstract only

Inclusion criteria:

• Age > 16 years

• Presenting to ED >24 h after injury

• Blunt head injury

• Had CT head

Exclusions:

• Not stated

Single-site: MA, USA

Conducted ‘over a 2-year period’ (dates not reported)

Case series

Any abnormality on CT head

206 consecutive patients identified

GCS on presentation not stated

13/206 (6.3 %) had abnormality on CT head

No patient required neurosurgery

Time to presentation (mean ± SD):

• Positive CT findings: 5.03 ± 6.52 days

• Negative CT findings: 5.79 ± 7.09 days

Positive CT findings more likely if LOC or amnesia of the event reported

Case series—does not state which patients were excluded or how

Sampling biases: small numbers, only those who had CT head after injury were included

GCS/other indicators of injury severity not discussed

Abstract only—unable to contact authors for further information

‘Abnormality’ on CT not defined or explained further