Critically ill patients frequently suffer from sepsis or localized infections. Diagnosing sepsis can be a challenge since several of its signs overlap with those found with other inflammatory states. Recognition of localized infections can at times be difficult too. While microbiological cultures of blood or other specimens are frequently used to distinguish infection from non-infectious conditions, this diagnostic technique lacks sensitivity and specificity. In addition, there is often a considerable time delay since bacterial cultures may require 24-48 h for analysis, which may be too long for a treatment decision in critically ill patients. Also, the reliability of microbiological cultures decreases in case of prior antimicrobial therapy. Use of biologic markers such as procalcitonin (PCT) or soluble triggering receptor expressed on myeloid cells-1 (sTREM-1) have been suggested to improve recognition of patients with true infection and facilitate decisions of whether or not to treat. Unfortunately, neither PCT nor sTREM-1 fulfill all expectations. Data on the diagnostic value, in particular of sTREM-1, are contradicting. The combination of systemic PCT and local and/or systemic sTREM-1 could be useful in distinguishing patients with infection from those with non-infectious illness, though. Results from several randomized intervention studies on PCT-guided antimicrobial therapy in sepsis or lower respiratory tract infections show the superiority of PCT in clinical decision making. At present, randomized intervention studies on the potential antimicrobial stewardship of sTREM-1 are lacking.