ImportanceApproximately 20% to 30% of patients admitted to an intensive care unit have sepsis. While fluid therapy typically begins in the emergency department, intravenous fluids in the intensive care unit are an essential component of therapy for sepsis.ObservationsFor patients with sepsis, intravenous fluid can increase cardiac output and blood pressure, maintain or increase intravascular fluid volume, and deliver medications. Fluid therapy can be conceptualized as 4 overlapping phases from early illness through resolution of sepsis: resuscitation (rapid fluid administered to restore perfusion); optimization (the risks and benefits of additional fluids to treat shock and ensure organ perfusion are evaluated); stabilization (fluid therapy is used only when there is a signal of fluid responsiveness); and evacuation (excess fluid accumulated during treatment of critical illness is eliminated). Among 3723 patients with sepsis who received 1 to 2 L of fluid, 3 randomized clinical trials (RCTs) reported that goal-directed therapy administering fluid boluses to attain a central venous pressure of 8 to 12 mm Hg, vasopressors to attain a mean arterial blood pressure of 65 to 90 mm Hg, and red blood cell transfusions or inotropes to attain a central venous oxygen saturation of at least 70% did not decrease mortality compared with unstructured clinical care (24.9% vs 25.4%; P = .68). Among 1563 patients with sepsis and hypotension who received 1 L of fluid, an RCT reported that favoring vasopressor treatment did not improve mortality compared with further fluid administration (14.0% vs 14.9%; P = .61). Another RCT reported that among 1554 patients in the intensive care unit with septic shock treated with at least 1 L of fluid compared with more liberal fluid administration, restricting fluid administration in the absence of severe hypoperfusion did not reduce mortality (42.3% vs 42.1%; P = .96). An RCT of 1000 patients with acute respiratory distress during the evacuation phase reported that limiting fluid administration and administering diuretics improved the number of days alive without mechanical ventilation compared with fluid treatment to attain higher intracardiac pressure (14.6 vs 12.1 days; P < .001), and it reported that hydroxyethyl starch significantly increased the incidence of kidney replacement therapy compared with saline (7.0% vs 5.8%; P = .04), Ringer lactate, or Ringer acetate.Conclusions and RelevanceFluids are an important component of treating patients who are critically ill with sepsis. Although optimal fluid management in patients with sepsis remains uncertain, clinicians should consider the risks and benefits of fluid administration in each phase of critical illness, avoid use of hydroxyethyl starch, and facilitate fluid removal for patients recovering from acute respiratory distress syndrome.