Early ficolin-1 is a sensitive prognostic marker for functional outcome in ischemic stroke

Abstract

Background: Several lines of evidence support the involvement of the lectin pathway of complement (LP) in the pathogenesis of acute ischemic stroke. The aim of this multicenter observational study was to assess the prognostic value of different circulating LP initiators in acute stroke.

Methods: Plasma levels of the LP initiators ficolin-1, -2, and -3 and mannose-binding lectin (MBL) were measured in 80 stroke patients at 6 h only and in 85 patients at 48 h and later. Sixty-one age- and sex-matched healthy individuals served as controls. Stroke severity was measured on admission using the National Institutes of Health Stroke Scale (NIHSS). The outcome was measured at 90 days by the modified Rankin Scale (mRS).

Results: Ficolin-1 was decreased in patients compared with controls measured at 6 h (median 0.13 vs 0.33 μg/ml, respectively, p < 0.0001). At 48 h, ficolin-1 was significantly higher (0.45 μg/ml, p < 0.0001) compared to the 6 h samples and to controls. Likewise, ficolin-2 was decreased at 6 h (2.70 vs 4.40 μg/ml, p < 0.0001) but not at 48 h. Ficolin-3 was decreased both at 6 and 48 h (17.3 and 18.23 vs 21.5 μg/ml, p < 0.001 and <0.05, respectively). For MBL no difference was detected between patients and controls or within patients at the different time points. In multivariate analysis, early ficolin-1 was independently associated with unfavorable mRS outcome (adjusted odds ratio (OR): 2.21, confidence interval (CI) 95 % 1.11-4.39, p = 0.023). Early ficolin-1 improved the discriminating ability of an outcome model including NIHSS and age (area under the curve (AUC) 0.95, CI 95 % 0.90-0.99, p = 0.0001).

Conclusions: The ficolins are consumed within 6 h after stroke implicating activation of the LP. Early ficolin-1 is selectively related to 3-month unfavorable outcome.

Fig. 1

Patient flowchart

Fig. 2

LP initiators at early and late time points. Plasma concentrations of ficolin-1 (a), ficolin-2 (b), ficolin-3 (c), and MBL (d) in controls (n = 61) and in two different group of stroke patients sampled at different time points after stroke (n = 80 patients within 6 h and n = 85 within 48 h). In the latter group, blood samples were collected at baseline, at days 3–5 (n = 78) and at 1 month (n = 60) after stroke. The dotted red line indicates the median value in controls. Data are expressed as median with interquartile range. p values *<0.05, **<0.01, ***<0.001, ****<0.0001 versus control and among groups, Kruskal-Wallis test with Dunn post hoc test

Fig. 3

Diagnostic accuracy of ficolin-1 and ficolin-3 in discriminating stroke patients from controls. ROC curve of early ficolin-1 levels (6 h) demonstrating sensitivity as a function of 1-specificity for discriminating case/control status at the early time point (a). ROC analysis data, including optimal cut-offs, of ficolin-1 and ficolin-3 for discriminating case/control status at early and late time points (b). The AUC and exact p value for asymptotic significance are reported. SE sensitivity, SP specificity, AUC area under the curve, CI 95 % 95 % confidence interval, OR odds ratio. ^aThe odds ratio corresponds to a unit change in the explanatory categorical variables

Fig. 4

Prognostic accuracies of early ficolin-1 as predictor of unfavorable outcome. ROC curve of early ficolin-1 levels (6 h) demonstrating sensitivity as a function of 1-specificity for predicting functional outcome at 3 months, based on the logistic model incorporating the relative contribution of each predictor in the combined model (ficolin-1 adjusted predictive values for NIHSS and age (a). The AUC and exact p value for asymptotic significance are reported (b). AUC area under the curve, CI 95 % 95 % confidence interval, NIHSS National Institutes of Health Stroke Scale

Fig. 5

MPO levels at early and late time points in stroke patients and controls. Plasma concentrations of MPO in controls and patients with stroke (sampled within 6 h and 48 h, respectively). Data are expressed as median with interquartile range. p values *<0.05 versus control and among groups, Kruskal-Wallis test with Dunn post hoc test (a). Spearman’s rho for ficolin-1 versus MPO (b). MPO myeloperoxidase