The 15th Anniversary of

Clinical trials of drugs specifically targeting the sepsis response have frequently produced negative or inconclusive results. This has largely been due to the broad heterogeneity of enrolled patient populations, particularly when inclusion was based on the presence of the non-specific systemic inflammatory response syndrome (SIRS) criteria. The heterogeneity may have diluted any possible treatment effect: while some patients may have benefited from the intervention under investigation, others will have been harmed, resulting in an overall null-effect. Furthermore, an underlying infection is not always required for immune-modulating interventions to be effective; for example, patients with severe acute pancreatitis, but no infection, may still benefit from such therapies. There is therefore a need for better patient stratification or subphenotyping to identify those most likely to benefit from a particular therapy. Several trials have already adopted this approach using prognostic and/or predictive enrichment strategies. For example, measurements of triggering receptor expressed on myeloid cells (TREM) could be used to identify candidates most likely to respond to anti-TREM therapies, or patients with coagulopathy or specific inflammatory patterns could be selected for treatments like thrombomodulin or activated protein C. However, challenges remain, including the need for more rapid tools that can be used at the bedside to inform real-time treatment decisions given the rapidly evolving nature of the sepsis response. Nevertheless, the era of broad-spectrum "sepsis drugs" is now giving way to more selective, personalized interventions tailored to individual biological profiles, offering a more promising pathway for future therapeutic development-even in the absence of infection.