Welcome to the ESOC 2021 Virtual Conference Calendar

Background And Aims

N-terminal pro B-type natriuretic peptide (NT-proBNP), a marker of ventricular dysfunction; and high-sensitivity Troponin T (hs-cTnT), a marker of cardiac injury, have been linked to subsequent strokes; suggesting the biomarkers may have a role in primary prevention. In older men free of stroke, myocardial infarction (MI), heart failure (HF) and atrial fibrillation (AF); we examined associations of each biomarker with incident all, fatal and non-fatal strokes; TIAs; and a combined outcome of stroke or TIA.

Methods

In the British Regional Heart Study, men (60-79y) provided history and clinical measurements at baseline and were followed-up for stroke events for 20y. Cox models estimated risk of each outcome across fourths (Q1-Q4) of NT-proBNP (n=2957) and hs-cTnT (n=3128); adjusting for cardiovascular risk factors; incident MI, HF, AF; and biomarkers mutually.

Results

Risk of all outcomes increased across fourths of NT-proBNP. Compared to Q1, HR (95% CI) of combined event was: 1.39 (1.07-1.80), 1.57 (1.20-2.05) and 2.00 (1.47-2.71); and of fatal strokes: 2.13 (1.07-4.26), 1.21 (0.55-2.63), 3.50 (1.66- 7.36); for Q2-Q4 respectively. In contrast, hs-cTnT was not associated with TIAs and predicted strokes weakly, mainly above the 75^th centile. HR of combined event: Q4 1.32 (1.01-1.72) and fatal strokes: Q4 1.65 (0.88–3.09). Adjusting for intervening cardiac events attenuated associations for hs-cTnT more, whereas adjusting for biomarkers mutually had little effect. C-index (combined event, 0.6437) was improved more by NT-proBNP (0.6544) than hs-cTnT (0.6477).

Conclusions

In older men free of stroke and overt cardiac disease, NT-proBNP appears to capture more information on long-term stroke risk than hs-cTnT.

Background And Aims

Peak width of skeletonized mean diffusivity (PSMD) is a new fully automated diffusion tensor imaging (DTI) marker that showed clinically relevant changes in cerebral small vessel disease (SVD) and was suggested to be one of the earliest markers of brain aging across the adult lifespan. We conducted the first genome-wide association studies of PSMD to decipher its genetic underpinnings.

Methods

We included 55,988 stroke-free participants from 19 population-based cohorts (91.3% European). PSMD was generated using the same automated algorithm across cohorts. We tested association of genetic variants with log (PSMD) in each cohort followed by inverse-variance-weighted meta-analysis (with and without adjustment for total intracranial volume). We performed secondary association analyses stratified on age (age <35, 36-65, >65 years) and searched for shared genetic variation with vascular and neurological phenotypes using linkage disequilibrium-score regression.

Results

We identified 14 independent genome-wide significant associations across all association models. Half of these were known genome-wide risk loci for white matter hyperintensity (WMH) volume, 43% were shared with other DTI metrics (mean diffusivity [MD] or fractional anisotropy [FA]), and 36% were specific for PSMD. Genetic correlation of PSMD across age strata was high, although genome-wide significant loci partly differed across age groups. We found significant positive correlation of PSMD with WMH, MD, FA, any and ischemic stroke, deep intracerebral hemorrhage, systolic/diastolic blood pressure, type-2 diabetes.

Conclusions

Identified genetic risk variants and patterns of genetic correlation support the hypothesis that PSMD is reflecting underlying SVD, with a pattern across age-strata suggesting long-term progression.

Trial Registration Number

“Not applicable"

Group Name

Blood Inflammatory Stroke Collaboration (BISC)

Background And Aims

Anti-inflammatory therapies reduce the risk of stroke in coronary disease but benefit after stroke is unknown. In a systematic review, existing studies reported conflicting findings for the association of blood inflammatory markers with vascular recurrence.

Methods

We investigated the association between high-sensitivity C-reactive protein (hsCRP) or interleukin-6 (IL-6) and risk of recurrent stroke/major vascular events in individual-participant data from 7,486 patients with ischaemic stroke/TIA from 9 prospective cohort studies. We did multivariate Cox regression analyses within each study and then combined adjusted hazard ratios (HR) by random-effects meta-analysis.

Results

During 12,895 patient-years of follow-up there were 775 recurrent strokes (10.4% 95% confidence interval [CI] 9.7-11.1%; 730 ischemic, 34 intracerebral haemorrhage, and 11 unknown subtype). Median time from index event to phlebotomy was 1 day (interquartile range 0.25-6). On bivariate analysis hsCRP (HR 1.13, 95% CI 1.04-1.24; p=0.005) and IL-6 (HR 1.18, 95% CI 1.03-1.35; p=0.015) were associated with recurrent stroke (per 1-unit increase in marker log-concentration). After adjustment for age, sex, hypertension, smoking, hyperlipidaemia, diabetes mellitus, atrial fibrillation and coronary disease, baseline hsCRP was associated with stroke recurrence (HR 1.08, 95% CI 1.01-1.15; p=0.03) and a statistical trend was identified for IL-6 (HR 1.12, 95% CI 0.99-1.27; p=0.07). The adjusted HR for the composite outcome of any major vascular event was 1.12 (95% CI 1.03-1.22; p=0.011) for CRP and 1.16 (95% CI 1.00-1.34; p=0.05) for IL-6.

Conclusions

This new evidence suggests that inflammation is associated with recurrent stroke/vascular events and strengthens the rationale for randomized trials of anti-inflammatory therapies for prevention after ischemic stroke/TIA.

Trial Registration Number

Not applicable

Group Name

BIO-FAST investigators

Background And Aims

Several tools such as clinical scales or individual blood biomarkers are being used to identify LVO patients out of the hospital and shift them directly to thrombectomy centers. However, their accuracy is suboptimal. We aimed to validate a new panel of blood biomarkers in a rapid Point-of-Care (POCT) format to be used to triage LVO patients.

Methods

Patients with suspected stroke (< 6h) were enrolled (n=197) and blood samples collected at ED. A 10 biomarkers panel was measured by immunoassays to select combinations of biomarkers associated with LVO. Best final candidates with available commercial POCTs were FABP/NT-proBNP in combination with simple clinical data (age, NIHSS, etc). Those results were prospectively replicated in BIO-FAST study that recruited n=300 patients with stroke suspicion onset < 6h in an ambulance network serving two thrombectomy centers (Biomarkers for Initiating Onsite and Faster Ambulance Stroke Therapies, ClinicalTrials.gov identifier: NCT04612218).

Results

Median NIHSS =7 (3-15), ASPECTS =10 (9-10) and blood samples obtention since stroke onset =2.16 hours (1.08, 5.25). Final diagnosis percentage was ischemic stroke (65.9%), hemorrhagic stroke (17.0%) and mimics (17.0%). Of the whole cohort 34.1% fulfilled LVO strict definition. LVO-check was 95% specific and 70% sensitive for LVO. Even using more simple scales such as Cincinnati instead of NIHSS the test showed good accuracy (100% specificity & 50% sensitivity).

Conclusions

LVO check is a rapid (10 ́) blood test that with a very high specificity and high sensitivity might allow referring LVO patients to thrombectomy centers or even attempting direct shift to the angio-suite.

Trial Registration Number

ClinicalTrials.gov identifier: NCT04612218

Background And Aims

Lipoprotein(a) concentration is associated with cardiovascular events independently of LDL-C. Post hoc analyses of clinical trials show that PCSK9 inhibitors might lower lipoprotein(a), but whether this effect contributes to reductions in cardiovascular risk remains unknown. In the absence of available lipoprotein(a)-reducing therapies, we aimed to assess whether genetically proxied PCSK9 inhibition influences lipoprotein(a) levels, and whether any such effect could mediate part of the effects of PCSK9 inhibitors on coronary heart disease (CAD) and ischemic stroke (IS).

Methods

We identified 12 genetic variants in the PCSK9 gene to be used as proxies (instruments) for PCSK9 inhibition based on association with LDL-C (p<5x10^-8, N=188,577) and performed mendelian randomization analyses.

Results

In our analyses, we found a 6% decrease in lipoprotein(a) levels (beta for log-lipoprotein(a): -0.061, 95%CI: -0.085 to -0.037) per 1-SD decrement in LDL-C levels (-39.8 mg/dL). We then performed two-step mediation analyses based on data from the CardioGramPLUSC4D (60,801 CAD cases, 184,305 controls) and the MEGASTROKE Consortia (60,341 IS cases, 454,450 controls). We estimated a 0.8% reduction in the odds for CAD (OR: 0.992, 95%CI: 0.988-0.996) and a 0.9% reduction in the odds for large artery atherosclerotic stroke (OR: 0.991, 95%CI: 0.985-0.998) due to reductions in lipoprotein(a) levels through variation in the PCSK9 gene, corresponding to 1.8% and 3.3% of the total effects, respectively.

Conclusions

Genetically proxied PCSK9 inhibition is associated with lower lipoprotein(a) levels. However, this lipoprotein(a) reduction contributes only modestly to the total effects of genetic proxies for PCSK9 inhibitors on risk of CAD and IS.

Trial Registration Number

Not applicable

Group Name

on behalf of the CHARGE consortium

Background And Aims

Dilated perivascular space (dPVS) burden is an emerging MRI-marker of cerebral small vessel disease (cSVD). We conducted a genome-wide association study of dPVS burden in 40,095 participants from 18 middle-aged to older population-based cohorts to decipher genetic underpinnings.

Methods

We tested association of genetic variants with the top quartile of white matter (WM), basal ganglia (BG), and hippocampal (HIP) dPVS burden distribution in each cohort followed by sample-size weighted meta-analysis. We explored shared genetic variation with other MRI-markers, putative risk factors, and neurological traits and conducted extensive functional exploration of identified dPVS loci using innovative bioinformatics approaches.

Results

We identified 24 genome-wide significant loci associated with extensive dPVS burden in older community-persons, primarily in WM. A weighted genetic risk score for WM dPVS burden was significantly associated with WM dPVS in young adults in their twenties. BG and HIP dPVS showed significant genetic correlation with high blood pressure, stroke (ischemic and hemorrhagic), and other MRI-cSVD markers. While some pleiotropy with vascular risk factors was observed, 2/3 of dPVS risk loci point to novel biological pathways, involving extracellular matrix, membrane transport, and developmental processes. Genome-wide significant dPVS loci were enriched in genes harboring causal mutations for monogenic white matter disease and genes expressed in brain endothelial cells. We used transcriptome-wide association studies to seek evidence for causal implication of specific genes underlying observed associations.

Conclusions

These findings shed new light on the biology underlying dPVS across the lifespan, with distinct mechanisms and clinical significance according to dPVS location.

Trial Registration Number

Not applicable

Group Name

for the MR CLEAN Investigators

Background And Aims

Collateral score is a strong predictor of outcome after endovascular treatment for ischemic stroke. We investigate whether an Artificial Intelligence approach performs similar to human observers in collateral scoring.

Methods

A neural network has been developed, trained and assessed for collateral scoring, using a 4-point scale, in CTA images of patients with an intracranial large vessel occlusion. The network model consists of a VoxResNet CNN trained in a Siamese fashion using each of the brain hemispheres as input for the Siamese branches. The feature maps obtained are used for collateral score prediction. The model architecture and hyper-parameters were optimized using cross-validation on the training data (278 CTA images). The results were validated on a validation set (69 CTA images, MR Clean Registry) and on an independent test set (39 CTA images, MR Clean Trial) against human observers. Scores were dichotomized using the 1-2 cutpoint. Inter-observer variation was computed as well.

Results

For the validation set, the confusion matrices for the automated scoring and for the inter-observer variation, and ROC curve for dichotomized collateral scoring are respectively in Image 1-3. These include collateral scoring accuracy, and dichotomized accuracy, sensitivity and specificity.

On the independent test set, three cases could not be scored due to preprocessing errors. Results when ignoring these cases are in below table.

Conclusions

Collateral scoring can be done accurately and within the inter-observer variation with a dedicated neural network trained end-to-end on a moderately sized dataset.

Trial Registration Number

Not applicable

Background And Aims

Atrial fibrillation (AF) increases the risk of ischemic stroke in asymptomatic individuals and may be the underlying cause of a high amount of cryptogenic stroke cases. Detection of AF would change primary and secondary stroke prevention strategies by recommending oral anticoagulation, but this arrhythmia is usually underdiagnosed. We aimed to test the usefulness of several blood-biomarkers to detect AF in two prospective cohorts.

Methods

492 patients from two studies were included: 274 subjects aged 65-75 years with hypertension and diabetes from the AFRICAT cohort, and 218 cryptogenic stroke patients aged >55 years from the CRYPTOFA cohort. AF was assessed by 4 weeks of monitoring with a wearable Holter device (Nuubo^TM). Blood was collected before monitoring started. 10 candidate biomarkers were identified by automated immunoassays (Roche, Penzberg) in the plasma of all patients. Univariate and logistic regression analyses were performed in each cohort separately.

Results

AF prevalence was 12.4% and 22.9% in each cohort. 4 biomarkers were significantly increased in asymptomatic individuals with AF (NT-proBNP, Ang-2, Troponin-T, and ESM-1) and 7 biomarkers showed significantly higher levels in cryptogenic stroke patients with AF detection (GDF-15, IL-6, Troponin-T, Ang-2, BMP-10, DKK-3, and NT-proBNP). The combination of Ang-2 and NT-proBNP showed the best performance in both cohorts, with an AUC of 0.772(0.692-0.865) to detect AF in high-risk asymptomatic individuals, and 0.747(0.671-0.823) in cryptogenic stroke patients.

Conclusions

Blood-biomarkers, in particular, NT-proBNP and Ang-2, were increased in AF patients and could be useful in AF screening strategies in the primary care setting, and for searching AF after cryptogenic stroke.