Intra-individual reproducibility of early and late C-reactive protein and interleukin-6 in patients with non-severe ischaemic stroke and carotid atherosclerosis

Introduction: Acute and late inflammatory markers including high-sensitivity C-reactive protein (hsCRP) and interleukin-6 (IL-6) are associated with future vascular events after stroke. However, few longitudinal studies exist examining the intra-individual reproducibility of inflammatory biomarker measures at different time-points after atherosclerotic stroke. We sought to examine the reproducibility of hsCRP and IL-6 in a cohort of patients with minor stroke or transient ischaemic attack (TIA) caused by ipsilateral carotid atherosclerosis.

Methods: Two observational cohort studies (DUCASS and BIOVASC) were pooled. Included patients had non-severe ischaemic stroke and ipsilateral internal carotid artery stenosis (≥50%). Patients had bloods drawn within 2 weeks of their index stroke/TIA event which was stored for later analysis. All patients included were followed up at 5 years and repeat phlebotomy was performed. Bloods were analysed for hsCRP and IL-6 using high-throughput immunochemiluminescence. Difference between baseline and follow-up blood levels and intraclass correlation (ICC) were calculated.

Results: 95 participants were included, median age 69 (IQR 63-77), and 51 (53.7%) had TIA as their presenting event. When biomarkers were dichotomised, (for hsCRP <2mg/L or ≥2mg/L, and for IL-6 <7.5pg/ml (median) or ≥7.5pg/ml) 68.4% (IL-6) and 65.2% (hsCRP) of participants remained in the same risk-category (high or low) over time. However, when analysed as a continuous variable, intra-class correlation coefficients were low: ICC for IL-6 0.14 (95% CI -0.06 - 0.33), ICC for hsCRP 0.05 (95% CI -0.14 - 0.25). ICC increased after removing outliers. Clinical characteristics and treatment were not associated with observed variability.

Conclusion: Our results suggest that concordance between early and late-phase inflammatory marker risk categories is modest, and absolute levels are not highly-correlated at early and late timepoints, despite associations at both times with future vascular risk. Investigators should standardise timing of phlebotomy and analysis protocols in future studies of inflammatory biomarkers.