Beyond the Surviving Sepsis Campaign Guidelines: a systematic review of interventions affecting mortality in sepsis

Introduction: Sepsis-related mortality is decreasing over time after the introduction of "Surviving Sepsis Campaign" Guidelines in 2004. The last Guidelines version collects 93 recommendations, but several interventions supported by randomized evidence of mortality reduction are not included. Evidence acquisition: We performed a systematic review of all randomized controlled trials reporting a statistically significant mortality reduction in septic patients and compared the identified studies to the Surviving Sepsis Campaign Guidelines 2021 to highlight discrepancies. Evidence synthesis: We identified 83 randomized controlled trials (58 interventions) influencing mortality in sepsis. Only 9/58 of these interventions were included in the Guidelines: lactate measurement and lactate-guided hemodynamic management, procalcitonin-guided antibiotics discontinuation, balanced crystalloids as first choice fluids, albumin infusion, avoidance of starches, noradrenaline as first line vasopressor, vasopressin as an adjunctive vasopressor to noradrenaline, neuromuscular blocking agents in moderate-severe sepsis-associated acute respiratory distress syndrome, and corticosteroids use. Only 11/93 Guidelines recommendations were supported by randomized evidence with mortality difference. Five of the interventions with survival benefit in literature (vitamin C, terlipressin, polymyxin B, liberal transfusion strategy and immunoglobulins) were recommended to avoid in the Guidelines, while 44 interventions were not mentioned, including three interventions (esmolol, omega 3, and external warming) with at least two randomized controlled trials with a documented survival benefit. Conclusions: Several discrepancies exist between the randomized controlled trials with mortality difference in septic patients and the latest Surviving Sepsis Campaign Guidelines. This systematic review can be of help for improving future guidelines and may guide research on specific promising topics.