Fluid Management in Sepsis

1Division of Allergy, Pulmonary and Critical Care Medicine, Vanderbilt University Medical Center, Nashville, TN

1Division of Allergy, Pulmonary and Critical Care Medicine, Vanderbilt University Medical Center, Nashville, TN

Authors contributions: Drafting of the manuscript: R.M.B.; Critical revision of the manuscript for important intellectual content: R.M.B., M.W.S.

Among critically ill adults, sepsis remains both common and lethal. In addition to antibiotics and source control, fluid resuscitation is a fundamental sepsis therapy. The physiology of fluid resuscitation for sepsis, however, is complex. A landmark trial found early goal-directed sepsis resuscitation reduced mortality, but three recent multicenter trials did not confirm this benefit. Multiple trials in resource-limited settings have found increased mortality with early fluid bolus administration in sepsis, and the optimal approach to early sepsis resuscitation across settings remains unknown. After initial resuscitation, excessive fluid administration may contribute to edema and organ dysfunction. Using dynamic variables such as passive leg raise testing can predict a patient’s hemodynamic response to fluid administration better than static variables such as central venous pressure. Whether using measures of “fluid responsiveness” to guide fluid administration improves patient outcomes, however, remains unknown. New evidence suggests improved patient outcomes with use of balanced crystalloids compared to saline in sepsis. Albumin may be beneficial in septic shock, but other colloids such as starches, dextrans and gelatins appear to increase the risk of death and acute kidney injury. For the clinician caring for sepsis patients today, the initial administration of 20 mL/kg of intravenous balanced crystalloid, followed by consideration of the risks and benefits of subsequent fluid administration represents a reasonable approach. Additional research is urgently needed to define the optimal dose, rate, and composition of intravenous fluid during the management of patients with sepsis and septic shock.

Sepsis, a dysregulated host response to severe infection, accounts for 2–6% of all hospital admissions and carries an in-hospital mortality of up to 15%.1–3 Mortality is even greater when sepsis is accompanied by hypotension and hypoperfusion (septic shock).3 Guidelines for sepsis management recommend early administration of antibiotics and intravenous (IV) fluid in addition to source control. Despite multiple recent clinical trials examining fluid management in sepsis, fundamental questions about which intravenous fluid to administer and in what amount remain unanswered. This article summarizes the physiologic principles and scientific evidence currently available to help clinicians make decisions regarding fluid management for patients with sepsis.

Patients with sepsis experience altered oxygen delivery and extraction, in part due to varying degrees of actual and relative intravascular volume depletion from decreased oral intake, increased insensible losses, sepsis-induced vasodilation, increased venous capacitance, and capillary leakage. The classic understanding is that during early sepsis most patients experience “relative hypovolemia” and the administration of intravenous fluid increases preload, which increases cardiac output, resulting in improved oxygen delivery to organs experiencing tissue hypoxia (Figure 1). This classic understanding is increasingly recognized to be overly simplistic. There are many factors that influence tissue oxygen delivery and extraction other than hemodynamics. In addition, the hemodynamic response to intravenous fluid is determined by an intricate interaction of mean systemic filling pressure, right atrial pressure, venous resistance, ventricular compliance, and afterload.4 Fluid administration may affect many of these components, some of them deleteriously (e.g., fluid administration may decrease venous return by increasing right atrial pressure).5,6 The complexity of patients’ responses to fluid administration in sepsis is evidenced by numerous studies reporting that approximately half of patients with sepsis do not experience hemodynamic improvement after fluid bolus administration, and that right atrial pressure poorly predicts hemodynamic improvement with intravenous fluid administration.7–9 Moreover, the century-old Starling model conceptualizing maintenance of vascular volume as the balance of hydrostatic and oncotic pressure gradients between the vessel lumen and interstitial space has been challenged by the recent recognition of the importance of the endothelial glycocalyx.10 Because it is a primary determinant of membrane permeability, damage to the glycocalyx during sepsis may alter patients’ response to fluid resuscitation. Although the clinical implications of these findings are not yet fully understood, they argue against an overly simplified approach to understanding the effects of fluid composition and dose in sepsis.

The same increase in preload (ΔP) can either cause a significant increase in stroke volume (ΔSV) (a), or no significant increase in stroke volume (b, c) depending on the initial preload (starting point on the curve) (b) or the shape of the curve, such as a flattened curve from decreased contractility (c). Unfortunately, this simple view fails to capture all the ways in which intravenous fluid may affect stroke volume. P, preload; SV, stroke volume.

Fluid administration is considered a fundamental part of early sepsis treatment.1 In the landmark Early Goal-Directed Therapy (EGDT) trial,11 Rivers and colleagues compared usual care to a protocolized approach to sepsis resuscitation using intravenous fluids, vasopressors, and blood transfusion among 263 patients in a single emergency department. In the usual care group, patients received arterial and central venous catheterization and were administered IV fluid to maintain a central venous pressure (CVP) of 8–12 mm Hg, and vasopressors to maintain mean arterial pressure (MAP) ≥65 mm Hg. The EGDT group used the same hemodynamic targets, but additionally received continuous monitoring of central venous oxygen saturation, with blood transfusion for a hematocrit less than 30% and dobutamine administration to achieve a central venous saturation ≥ 70%. During the 6 hours of intervention, EGDT patients received more IV fluid (mean 5.0 vs 3.5 L; P < 0.001), blood transfusions (64.1% vs 18.5%; P