Prognostic value of the ABCD2 clinical prediction rule: A prospective cohort study
1Galvin
1HRB
R, 2Atanassova P, 1Motterlini N, 1Fahey T, 1Dimitrov BDD.
Centre for Primary Care Research, Department of General Practice, Royal College of Surgeons in Ireland, Dublin, Ireland
2Clinic of Cerebrovascular Diseases, Department of Neurology, Medical University, Plovdiv, Bulgaria.
Background
Early recognition and rapid treatment of patients who present with a transient ischaemic
attack (TIA) can significantly reduce the risk of subsequent stroke. The ABCD2 clinical
prediction rule (CPR) was derived in 2007 to assist clinicians with the timely and
appropriate management of individuals with TIA. The ABCD2 rule is a 7 point summation
of clinical factors independently predictive of stroke risk. These factors include age,
clinical features such as motor impairment and speech disturbance, duration of
symptoms, history of diabetes and hypertension. A summary of the rule is contained in
Figure 1. The developers originally identified three strata of stroke risk after TIA
according to ABCD2 score; low risk (0-3 points), moderate risk (4-5 points) and high risk
(6-7 points).
Multiple validations of the ABCD2 system have reported inconsistent results, ranging
from excellent predictive value to little better than chance. Therefore, further broad
validation studies by investigators from different specialties and healthcare settings are
necessary.
Aim
The aim of this study is to examine the predictive value of the ABCD2 rule in a Bulgarian
setting. This is the first broad validation study of the ABCD2 rule in Bulgaria.
Statistical analysis
Discrimination
A 2x2 table is used to calculate sensitivities and specificities at the different dichotomised
cut points of ≥3, ≥4 and ≥5, as recommended in different national clinical guidelines. A
Chi-squared test for trend is computed to determine if there is an increasing trend in risk
of stroke across the three ABCD2 risk categories.
Calibration
The initial derivation study of the ABCD2 rule is used as a predictive model against which
our validation study is compared. The number of strokes predicted across the three
strata of risk - low risk (score 0-3), moderate risk (score 4-5) and high risk (score 6-7) is
compared with the observed number of strokes in each of these risk categories in our
study.
Results
From January 2002 to December 2005, 89 patients enrolled in the study. The mean age
of patients is 63 years (+/- 12 years). Fifty-nine percent (n=53) of the cohort are male.
There are 13 subsequent strokes observed in the cohort of patients. There is no incident
of stroke within the first 7 or 90 days after TIA. Seven strokes occur within the first year
and six further strokes within three years. The 7 day, 90 day, 1 year and 3 year risk of
stroke stratified according to ABCD2 score are presented in Table 1.
The risk of stroke increases as the ABCD2 score increases (p<0.001). In terms of the
discriminative ability of the rule, the ABCD2 rule is more useful at ruling out stroke at one
year, with a greater summary sensitivity estimate than specificity at all cut points
examined. At three years following stroke, the pattern changes and the rule is more
useful at ruling in stroke, with a greater pooled specificity estimate than sensitivity across
all the points.
Methods
Study design and setting
We conducted a retrospective analysis of data collected from a cohort of consecutive
patients presenting to the emergency department with symptoms of TIA at the Medical
University Hospital of Plovdiv from 2000-2005. The STROBE standardized reporting
guidelines were followed to report the findings of the study.
Ethical approval
Ethical approval was received from the Medical University Hospital of Plovdiv ethics
committee.
Inclusion criteria
1) Patient population: adult patients (>40 years of age) with symptoms of a TIA
2) Study setting: Emergency department
5) Willing to give informed consent to participate in the study
Outcomes
Baseline data and the clinical examinations including the ABCD2 scores were
documented by the study neurologists. Determination of the ABCD2 risk score was
performed in a manner identical to that reported by the developers of the score. The
elapsed time from the episode to registry was less than 48 hours.
All strokes were assessed by MRI or CT scan and confirmed by a study neurologist.
Determination of the outcome endpoint was done blinded to the ABCD2 score.
When using the original derivation study as a predictive model, the ABCD2 rule
significantly over predicts risk of stroke at 7 days across all three risk strata: low risk, [RR
3.00, 95% CI (0.13 - 71.51)]; intermediate risk, [(RR 5.00, 95% CI (0.25 – 99.82)]; high risk,
[(RR 4.33, 95% CI (0.75 - 24.90)]. The results are broadly similar at 90 days across all three
risk strata.
Discussion
This is the first broad validation study to examine the predictive value of the ABCD2 score
in a Bulgarian cohort. This study demonstrates that the ABCD2 rule significantly over
predicts risk of stroke at 7 and 90 days across all three risk strata. However, our results
need to be interpreted with caution due to the low patient numbers included in the
validation study. The totality of evidence across a variety of clinical settings suggests that
rule correctly predicts risk of stroke at 7 days. There is a need for future large multicentre randomised controlled trials to examine the impact of applying the rule in
different clinical settings, particularly in the primary care setting, in terms of patient
outcome, clinician behaviour, cost effectiveness and resource use, or any combination of
these.
Conflict of Interest Statement: The authors declare no conflict of interest.
Funding: This research was supported by the Health Research Board of Ireland through
the HRB Centre for Primary Care Research.
Corresponding author: Rose Galvin, HRB Centre for Primary Care Research, RCSI Ireland,
rosegalvin@rcsi.ie