Bloodstream infections comprise a wide variety of pathogens and clinical syndromes with considerable overlap with similar syndromes of non-bacteraemic infections and diverse risk factors, therapeutic implications and outcomes. Yet, this heterogeneous 'entity' has the advantage to be pathogen-defined compared with the broad and even more heterogeneous entity 'sepsis', and so has become helpful for clinicians and epidemiologists for research and surveillance purposes. The increasing availability of population-based and large multicentre well-defined cohort studies should allow us to assess with much confidence and in detail its burden, the significance of antimicrobial resistance, and areas of uncertainty regarding further epidemiological evolution and optimized treatment regimens.