Commentary: Circulating cytokines and risk stratification of stroke incidence--will we do better in future?

Development and complications of atherosclerotic cardiovascular disease are characterized by an elevated inflammatory state. Pro-inflammatory cytokines play a central role in the initiation and perpetuation of inflammatory activity. Since inflammation had been identified as crucial to the atherosclerotic process,1 multiple systemic inflammatory biomarkers have been investigated in relation to outcome in cross-sectional and longitudinal studies. For robust down-stream markers of the inflammatory cascade (i.e. C-reactive protein), positive associations have consistently been demonstrated for coronary endpoints. Stroke, however, is a more diverse phenotype. Confirmed relations of inflammatory markers and incident stroke starting with fibrinogen in the 1970s have been less strong compared with coronary events. Despite the reproducible association of inflammatory biomarkers with incident and recurrent cardiovascular disease, the additional risk information gained from the measurement of a single marker or a panel of biomarkers does not relevantly enhance risk prediction across risk groups based on traditional risk factors.2 In this context, the article by Stott et al.3 in this issue examined the relation of the adipocytokine adiponectin and the two cytokines interleukin(IL)-18 and tumour necrosis factor alpha (TNFα) over a mean follow-up of 3.2 years in a nested case–control design of the PROspective Study of Pravastatin in the Elderly at Risk (PROSPER) trial, a sample at high risk for cardiovascular disease. Baseline adiponectin concentrations were inversely correlated with the component endpoint4 ischaemic stroke only, but did not reach significance in the treatment subgroups, indicating a relatively weak association. The difference in baseline concentrations was modest and of questionable clinical value. About 26% of stroke cases remained uncategorized in the primary trial analyses, indicating the heterogeneity and difficulty of clear classification of the end-point. In contrast to the original trial, transient ischaemic attack was not examined in this article.