About one third of early deaths and poor outcomes after acute stroke are caused by potentially preventable stroke-associated complications, especially infections. Early identification of patients at high risk of infections and poor prognosis with biomarkers might help to initiate adequate therapies and guide treatment decisions. Acute injury of the central nervous system, including stroke, disturbs the normally well-balanced interplay between the sympathetic nervous system and the immune system, thereby impairing the antibacterial immune response in stroke patients. Changes in immune and stress markers, for example a reduction in HLA-DR expression on monocytes or an increase in serum catecholamine levels, occur very early after stroke onset, explain the high susceptibility of stroke patients to bacterial infections, and are predictive of infectious complications occurring up to 2 weeks after stroke. Outcome prediction is of utmost importance for decision-making in stroke units as well as in neurological intensive care units. However, to date the accuracy of outcome prediction by physicians and clinical scoring systems is only moderate. So far, only two blood-based biomarkers have been identified as independent predictors of outcome and mortality after stroke: the stress marker copeptin and midregional pro-atrial natriuretic peptide. Careful evaluation of prognostic markers is needed to prevent the occurrence of self-fulfilling prophecy.