Numerous blood biomarkers have been suggested for diagnosis and outcome prediction of acute ischemic stroke. However, none of them has yet found its way into routine clinical practice. Serum S-100B, a protein expressed primarily by astrocytes, is considered a marker of blood-brain-barrier (BBB) damage and therefore of particular interest as a potential marker of hemorrhagic transformation as well as edema after ischemic stroke. In guidelines for mild traumatic brain injury, S-100B is already suggested to guide imaging (1). First studies reporting an association of S-100B levels with stroke severity and stroke outcome have been published more than two decades ago, and several studies have shown an association of increased S-100B levels with an increased risk of symptomatic intracerebral hemorrhage (sICH) (2) as well as development of malignant edema (3) after stroke. However, previous studies were limited by small sample size and mostly monocentric design.

In the current issue of the European Stroke Journal, Honegger and colleagues report results from the prospective multicenter BIOSIGNAL study on the predictive value of acute serum S-100B levels for the aforementioned complications of acute stroke (4). In this study S-100B levels were measured within 24h of symptom onset (median 4.5h) in 1749 acute stroke patients. sICH occurred in 2.6% and symptomatic edema in 5.2%. In multivariate analysis including further known predictors, higher acute serum S-100B levels were significantly associated with both sICH (together with reperfusion treatment) and symptomatic edema (together with higher NIHSS on admission and lower age). Adding serum S-100B levels to existing sICH prediction scores resulted in a small but significant increase in model performance.

The findings of this study validate earlier observations and confirm serum S-100B as a blood biomarker of interest for acute stroke management. These results are also not surprising, given the known association of serum S-100B with infarct size (5), and the fact that lesion size represents a key predictor of both sICH and space-occupying edema. Thus, the most interesting finding of this analysis from the BIOSIGNAL study is the particularly strong association of elevated serum S-100B values with sICH in patients with mild stroke (NIHSS £5). This interaction with stroke severity might point towards a role of BBB damage captured by increases in serum S-100B values in the prediction of sICH.

Prediction models in the study rely on single S-100B measurements which were taken at a rather early time-point of stroke management. Serial measurements, which were not available in BIOSIGNAL, might further improve the identification of patients at risk of either sICH or edema expansion. As with cardiac troponin, where early dynamics are considered for clinical decision making, more complex decision algorithms incorporating the dynamics of serum S-100B are also conceivable and could be the subject of further research.

In conclusion, this study provides new insights in the field of stroke biomarker research and further highlights the potential of blood biomarkers to be a piece of the puzzle for risk stratification and early treatment of stroke.

Defining the etiology of ischemic stroke is paramount for optimal secondary prevention. Despite wide screening, up to 50% of ischemic stroke cases the diagnostic work-up turns out negative, with no cardioembolic, atherothrombotic or specific cause fully explaining the vascular event¹.

Finding a cardioembolic origin of an ischemic stroke opens the way to anticoagulation, which has clear advantage in the prevention of embolism compared to antiplatelets. Beyond atrial fibrillation, a leading cause of cardioembolic stroke, other sources of cardiac embolism exist, including cardiac thrombi². In this Issue of the European Stroke Journal, Leon Rinkel and the Amsterdam UMC team led by Jonathan Coutinho report the results from the Mind the Heart study, a prospective cohort study investigating the yield of cardiac CT in acute ischemic stroke³. Overall, 452 consecutive patients underwent ECG-gated cardiac CT during hyperacute stroke imaging, with 38 being found with cardiac thrombi (8.4%). Location of the thrombus spanned left atrial appendage (82%), left atrium (5%) and left ventricle (18%). Interestingly, trans-thoracic echocardiogram was performed in 350 patients and only identified two cases of cardiac thrombi (vs 38 with ECG-gated cardiac CT).

People with cardiac thrombus had a high frequency of atrial fibrillation (40% of cases), and had sub-therapeutic INR levels (<2) in 44% of cases. People with cardiac thrombus received similar reperfusion strategies compared to patients without cardiac thrombi detected, but the presence of cardiac thrombus was an independent predictor of poor functional outcome (OR for shift analysis 2.22, ranging 1.16 to 4.23), and carried a 3% absolute higher risk of recurrent stroke³.

Overall, three main themes emerge from this brilliant research. First, as cardiac thrombi are more frequently detected in patients with atrial fibrillation, the relationship between atrial fibrillation and a lower rate of recovery might have finally come to light. Indeed, if echocardiography fails to detect 36/38 cases of thrombus in a cardiac chamber, one might wonder how many can be missed during standard diagnostic work-up in relation to timing of ascertainment and specificity. Therefore – second point – one might argue in favour of implementing a time-sparing diagnostic technique in the clinical routine, at least for patients at high risk of cardiac thrombus. Finally, immersed as we are in over-diagnosis and over-treatment, we should note that identifying a cardiac thrombus has direct impact on management, with anticoagulation to be started early on. As the standard diagnostic work-up with transthoracic echocardiography would likely result unremarkable, cardiac CT seems worth of implementation to test its yield on a larger scale.