IntroductionInflammatory, autoimmune and metabolic disorders have been associated with alterations in osteopontin (OPN) serum levels. Furthermore, elevated serum levels of OPN were reported from a small cohort of patients with sepsis. We therefore analyzed OPN serum concentrations in a large cohort of critically ill medical patients.MethodsA total of 159 patients (114 with sepsis, 45 without sepsis) were studied prospectively upon admission to the medical intensive care unit (ICU) as well as after 3 days of ICU treatment and compared to 50 healthy controls. Clinical data, various laboratory parameters as well as investigational inflammatory cytokine profiles were assessed. Patients were followed for approximately 1 year.ResultsWe found significantly elevated serum levels of OPN at admission to the ICU and after 3 days of treatment in critically ill patients compared to healthy controls. OPN concentrations were related to disease severity and significantly correlated with established prognosis scores and classical as well as experimental markers of inflammation and multi-organ failure. In the total cohort, OPN levels decreased from admission to day 3 of ICU treatment. However, persistently elevated OPN levels at day 3 of ICU treatment were a strong independent predictor for an unfavorable prognosis, with similar or better diagnostic accuracy than routinely used markers of organ failure or prognostic scoring systems such as SAPS2 or APACHE II score.ConclusionsPersistently elevated OPN serum concentrations are associated with an unfavourable outcome in patients with critical illness, independent of the presence of sepsis. Besides a possible pathogenic role of OPN in critical illness, our study indicates a potential value for OPN as a prognostic biomarker in critically ill patients during the early course of ICU treatment.Electronic supplementary materialThe online version of this article (doi:10.1186/s13054-015-0988-4) contains supplementary material, which is available to authorized users.