Multiple biomarkers covering several pathways for the prediction of depression after ischemic stroke

BACKGROUND To assess the potential incremental utility of multiple biomarkers reflecting several pathological pathways for the risk prediction of depression after stroke. METHODS We used data from the China Antihypertensive Trial in Acute Ischemic Stroke, and a panel of 13 circulating biomarkers were measured. The study outcome was depression (24-item Hamilton Depression Rating Scale score≥8) at 3 months after ischemic stroke. Logistic regression models were performed to evaluate the risk of depression associated with multiple biomarkers. Discrimination and risk reclassification for depression were analyzed. RESULTS Among 631 included ischemic stroke patients, elevated growth differentiation factor-15, anticardiolipin antibodies, antiphosphatidylserine antibodies and matrix metalloproteinase-9 were individually associated with increased risks of depression after stroke. The multiple biomarker analysis showed a clear gradient in the risk of depression with increasing numbers of elevated biomarkers, and multivariate adjusted odds ratio (95% confidence interval) of patients with 4 elevated biomarkers was 6.52 (2.24-18.95) compared with those without elevation in any of 4 biomarkers. The simultaneous inclusion of all 4 biomarkers to the conventional model significantly improved discrimination (C statistic increased from 0.702 to 0.748, P=0.004) and risk reclassification (net reclassification improvement 45.0%; integrated discrimination improvement 6.2%; both P<0.001) for depression after stroke. LIMITATIONS We selected biomarkers that had previously been reported to be promising predictors of depression after stroke, while other novel biomarkers not tested might have additional predictive value. CONCLUSIONS Simultaneously adding multiple biomarkers from several pathophysiological pathways to traditional risk factors provided substantial incremental utility of the risk stratification for depression after stroke.