Blood C-reactive protein concentration with ABCD² is a better prognostic tool than ABCD² alone

Increased C-reactive protein (CRP) is a known predictor of vascular events in asymptomatic individuals and stroke patients. Only a few studies included transient ischaemic attack (TIA) patients. We assessed CRP levels in addition to traditional risk factors in a cohort of patients with TIA to examine the relationship of these parameters to the occurrence of ischaemic stroke.

This is a prospective, longitudinal clinical evaluation of the efficacy of CRP as a prognostic indicator. CRP levels were measured in 194 TIA patients and in 1,024 asymptomatic individuals (recruited from a project on stroke prevention, the PrATO, which was ongoing at the same time in the Aosta Valley). A clinical risk score was determined using the ABCD² score in TIA patients. The area under the receiver operating characteristic curve (AUC) was used to evaluate the significance of the markers as predictors. Two models were evaluated: model 1 used the ABCD² score and model 2 used serum CRP levels in addition to the ABCD²) score. The primary outcome was an ischaemic stroke.

Within 2 years ischaemic strokes occurred in 33/194 patients. The Cox proportional hazards models, after adjustments for conventional risk factors, identified CRP levels ≥3 mg/l and ABCD² scores ≥4 as independent predictors of stroke. The corresponding AUCs were 0.565 and 0.636, based on model 1 and model 2, respectively; this represented a statistically significant difference (p = 0.043). The absolute integrated discrimination improvement was 0.0249 (p = 0.007) and the relative integrated discrimination improvement was 2.3710. The net benefit became significant from a predicted probability ≥10% and was 0.077 when based on model 1 and 0.087 when based on model 2.

Routine CRP measurements in the acute phase might be a useful tool for identifying TIA patients who are at a higher risk of ischaemic stroke. The additional use of CRP levels for the risk assessment in TIA patients improves risk definition in terms of the ABCD² score alone.