Current Controversies in Sepsis Management

Stephanie R. Moss, MD; Hallie C. Prescott, MD, MSc


Semin Respir Crit Care Med. 2019;40(5):594-603. 

In This Article

Fluid Heavy Versus Early Vasopressor Resuscitation Strategy

Historically, fluid resuscitation has been a cornerstone of management in septic shock, predating the advent of antibiotics.[85] In the United States in particular, a liberal fluid strategy prevails. The SSC guidelines and CMS SEP-1 measure both promote an initial 30 mL/kg fluid bolus in septic patients with shock or elevated lactate.[8,86] However, this standard practice has been driven largely by expert opinion, as many studies evaluating fluid resuscitation in sepsis are observational, and results are mixed.[87]

The physiological argument for fluid resuscitation in septic shock is to correct intravascular volume depletion which occurs as a result of capillary endothelial dysfunction and decreased systemic vascular resistance.[88] Increasing intravascular volume can increase cardiac preload and stroke volume, which in turn increases tissue perfusion.[89] The physiological argument for restricting fluids in sepsis shock is that fluids only transiently increase intravascular volume, because they do not correct endothelial dysfunction, and ultimately lead to pathologic edema which can result in organ dysfunction and worse functional outcomes.[90,91] Meanwhile, vasopressors increase systemic vascular resistance, which increases cardiac preload without edema;[92] however, vasopressors increase the risk for several complications, including tachyarrhythmias, myocardial ischemia, mesenteric hypoperfusion, and skin necrosis.[93–96]

In animal models of sepsis, crystalloid bolus improves perfusion and survival,[97,98] but also results in a paradoxical increase in vasopressor requirements.[99] Extrapolating these findings to humans is difficult. Whereas humans tend to develop hyperdynamic shock with increased cardiac output and decreased systemic vascular resistance, many animal models develop hypodynamic shock.[85]

In a seminal RCT, Rivers et al evaluated "early goal-directed therapy" (EGDT)—a septic shock resuscitation strategy which included a 500-mL crystalloid bolus every 30 minutes to achieve central venous pressure of 8 to 12 mm Hg.[100] Patients randomized to EGDT developed less severe organ dysfunction and had improved survival. Thus, this strategy was recommended by the 2004 SSC guidelines and widely adopted into practice.[41] However, as EGDT is a multicomponent protocol, it is impossible to tease out the independent impact of early fluid bolus.[85]

Three multicenter RCTs (ProMISe, ARISE, and ProCESS)[101–103] subsequently evaluated EGDT versus usual care (and also "protocol-based usual care" in the ProCESS study). In all three RCTs, mortality was indistinguishable between patients randomized to EGDT versus usual care.[101–103] A patient-level meta-analysis likewise found no difference in mortality, but did show an increase in ICU resource utilization.[104] Importantly, however, the median fluid bolus prior to randomization was 2 L (just over 27 mL/kg), and total resuscitation in the first 6 hours was 4 to 5 L for patients in both treatment arms.

The 2016 SSC guidelines recommend that at least 30 mL/kg of crystalloid fluids be given within the first 3 hours of presentation for patients with sepsis-induced hypotension (strong recommendation, low-quality evidence).[8] While RCT data to support an early 30 mL/kg bolus are lacking, the amount was viewed as standard practice based on the prerandomization fluid resuscitation in ProMISe, ARISE, and ProCESS.[101–103]

However, as the pendulum has swung toward fluid-heavy resuscitation, there have been growing concerns among the critical care community about the potential harms of fluid boluses.[105] In RCTs of postresuscitation fluid management, fluid-heavy strategies are associated with increased ventilator days, increased ICU length of stay, and increased severity of organ dysfunction.[106,107] Multiple observational studies suggest that a positive fluid balance is associated with increased mortality, even after adjustment for potential confounders.[108–112] Finally, RCTs in lower resource settings—where there is less ability to manage potential negative sequelae of fluid resuscitation—have shown increased mortality with fluid resuscitation. The FEAST study of pediatric patients in Africa was stopped early due to increased mortality in the treatment arm.[113] Similarly, a RCT of an early resuscitation protocol in Zambia revealed increased mortality in the treatment arm, with a number needed to harm of just seven patients.[114] Taken together, these studies have led many clinicians to question the necessity of early fluid bolus in patients with sepsis-induced hypotension.[85]

From a practical standpoint, clinician reliance on fluid resuscitation has also been influenced by a reluctance to administer vasopressors prior to obtaining central IV access, which takes time to establish. Central administration is preferred due to risk of tissue injury from extravasation, although the frequency of this complication has been poorly quantified.[115] Moreover, newer studies suggest vasopressors can be given peripherally,[115–117] with lower rates of complication than previously described for central venous catheters.[118]

One of the main goals of fluid resuscitation or vasopressor support is to maintain adequate blood pressure. The SSC guidelines recommend targeting a mean arterial pressure (MAP) of 65 mm Hg for initial resuscitation (strong recommendation, moderate quality of evidence).[8] RCTs that have randomized patients to a target MAP of 65 mm Hg versus 85 mm Hg have found similar lactate levels, renal function and urine output, and oxygen delivery.[119,120] However, in some RCTs, higher MAP targets have been associated with higher rates of new-onset atrial fibrillation,[121] and increased mortality among patients 75+ years of age.[122]

Given concerns about excessive fluid resuscitation in early septic shock management, several ongoing RCTs are assessing early resuscitation practices.

The REFRESH pilot RCT was a multicenter, unblinded trial which enrolled 99 patients with suspected infection and systolic blood pressures of <100 mm Hg after administration of 1000 mL of balanced crystalloid fluid from Australian emergency departments.[123] Patients were randomized to protocolized standard care versus restricted volume management. Standard care included a second 1,000 mL bolus, followed by additional 500 mL boluses if deemed necessary based on perfusion parameters until euvolemia was judged to have been achieved, followed by initiation of norepinephrine if MAP was <65 despite indicators of euvolemia. Patients in the experimental arm received norepinephrine immediately if MAP was <65. Fluid boluses of 250 mL each hour up to 1,000 mL total and IV fluids at maintenance rate (maximum 150 mL/h) could be given at physician discretion. In both groups, norepinephrine was administered peripherally until central access could be obtained. The primary outcome was total fluid administered at 6 hours.[123] Patients in the restricted volume arm received less IV fluid compared with standard care (2,387 vs. 3,000 mL at 6 hours, p < 0.001; 3,543 vs. 4,250 mL at 24 hours, p = 0.005) and had earlier initiation of vasopressors, but no difference in the total time spent on vasopressors.[124] A larger study with goal enrollment of 3,000 patients is being planned.[125]

Meanwhile, in the United States, the CLOVERS trial sponsored by the PETAL Network is underway.[89] CLOVERS is a multicenter, randomized, unblinded trial of liberal fluid management (additional 2 L of fluid prior to consideration of vasopressors) versus restrictive fluid management (immediate initiation of vasopressors) in patients with persistent sepsis-induced hypotension after 1 to 3 L of IV fluids.[89] The primary outcome is 90-day mortality.

Finally, in the United Kingdom, the 65 trial, assessing the clinical effectiveness of permissive hypotension (MAP target 60–65 mm Hg during vasopressor therapy) versus usual care in critically ill patients aged 65+ years, recently finished recruiting patients in March 2019.[126] The primary outcome is all-cause 90-day mortality.

While we await results from the above studies, we favor an initial 30 mL/kg fluid bolus in patients with sepsis-induced hypotension or lactic acidosis, followed by prompt initiation of vasopressors for patients who are still hypotensive (via peripheral access until central access can be established). We generally target a MAP goal of 65, except for patients who are known to have a lower baseline MAP.