Article Text

Download PDFPDF
Review: 3 prediction rules, particularly ABCD, identify ED patients who can be discharged with low risk of stroke after TIA

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.


In patients with transient ischaemic attack (TIA), can physicians in the emergency department (ED) reliably use a clinical prediction rule to determine which patients can be discharged because their risk of stroke is low (1–2%)?

Review scope

Included studies reported either the derivation or validation of clinical prediction rules for determining risk of stroke within 7 days after TIA in adults. Exclusion criteria were non-acute care setting and sample of patients with known history of TIA. Outcome was risk of stroke in relation to prediction rule scores.

Review methods

PubMed, EMBASE/Excerpta Medica, and DARE were searched for English-language studies. 5 studies were included. 2 studies were derivation cohorts (California rule and ABCD2 rule), 2 were validation studies (ABCD rule), and 1 was a derivation and validation study (ABCD rule). Levels of evidence were assigned (1  =  high to 4  =  low) for each rule: ABCD  =  2, California  =  3, and ABCD2  =  4.

Main results

All prediction rules showed increasing risk of stroke with higher scores. In 5 validation cohorts, patients with ABCD scores <4 had 0–1.7% risk of stroke at days 2 and 7. With ABCD scores ⩾4, risk of stroke increased to ⩾3.4% (mean 5.4%) at 2 days and 0–36% {means ranging from 6% to 13%}* at 7 days. In 1 validation cohort, California scores <3 were associated with 0–1.8% risk of stroke at 2 days and 0–2.9% at 7 days. Risk of stroke was 4.6–7.1% at 2 days and 7.5–13% at 7 days for California scores ⩾3. In 1 study, ABCD2 scores <4 were associated with 0–1.3% risk of stroke at 2 days and 0–1.5% at 7 days. Risk of stroke was 3.8–6.3% at 2 days and 5.5–11% at 7 days for ABCD2 scores ⩾4.


In patients presenting to the emergency department after transient ischaemic attack, the ABCD prediction rule can identify patients whose risk for stroke is low enough (1–2%) to discharge them home. The California and ABCD2 rules perform similarly, although more studies are needed to validate these rules.

*Data provided by author.

Abstract and commentary also appear in ACP Journal Club.

Abstracted from

Shah KH, Metz HA, Edlow JA. Clinical prediction rules to stratify short-term risk of stroke among patients diagnosed in the emergency department with a transient ischemic attack. Ann Emerg Med 2009;53:662–73.

Clinical Impact Ratings: GP/FP/Primary care 6/7; Emergency medicine 6/7; Internal medicine 5/7; Neurology 5/7


The objective of this systematic review was to compare different rules rather than to summate the findings as a meta-analysis. Shah et al present data indicating that risk of stroke within 7 days is proportional to the number of risk factors. This is clinically intuitive and supported by the literature. Based on a number of validation studies (1 prospective and 2 retrospective) and citing a <2 % stroke miss rate, the authors also concluded that the ABCD rule is clinically useful for identifying which patients can be safely discharged from the ED.

Prospective validation is the litmus test of clinical decision/prediction rules, and an important prospective validation study was excluded from this review because the primary outcome was an event at 90 days.1 This study reported stroke within 2 days (1/88 with ABCD score <4). However, the authors admit that they probably missed some strokes within the first 24 hours after TIA. This would indicate that a 1–2% stroke miss rate is an optimistic estimate, even if clinically acceptable.

Until an accurate, prospectively validated (conducted on ED patients in the ED), and clinically acceptable rule becomes available, management in the ED should focus on early neurology consultation to confirm the diagnosis of TIA, rapid diagnostic assessment, and implementation of aetiology-specific preventive measures. These primarily include identifying patients with carotid artery stenosis amenable to revascularisation and those with a cardioembolic source requiring anticoagulation. The decision to discharge a patient after TIA should be based on the availability of the necessary outpatient services. As Shah et al point out, “It is not the hospitalization per se that confers this risk reduction, but the rapidity of the evaluation; in most hospitals, this rapid evaluation could occur only with an admission.”


View Abstract


  • Source of funding: no external funding.