REBEL Cast Ep116: The CLOVERS Trial – Restrictive vs Liberal Fluids in Sepsis-Induced Hypotension

Background: IV fluids are part of the standard resuscitation bundle in septic shock, however it is unclear if they provide a significant benefit. These patients can have a vasodilated vascular bed and the initial use of fluids is hypothesized to serve two purposes: Augmenting the macrovascular system (stroke volume and cardiac output) and augmenting the microvascular perfusion (capillary blood flow). However, IV fluids are not a benign intervention as they can cause coagulopathy and fluid overload.

The CENSER Trial (Link is HERE) was a randomized, double-blind, placebo-controlled clinical trial evaluating the use of early low-dose norepinephrine vs standard care (placebo) in adults with septic shock.  The goal of the trial was to see if early vasopressors improved shock control by 6 hours. Patients were started on low-dose norepinephrine at a median time from ED arrival of 93 minutes compared to 192minutes in the standard care arm.  This resulted in better shock control by 6hrs (76.1% vs 48.4%). The secondary outcome of mortality was not statistically different however numerically favored the early use of norepinephrine.

Physiologically it makes sense to initially use lower volumes of fluids and earlier initiation of vasopressors. However, there is a lack of robust evidence to help guide what fluid and vasopressor strategy is best in patients with septic shock.

REBEL Cast Ep116:  The CLOVERS Trial – Restrictive vs Liberal Fluids in Sepsis-Induced Hypotension

Paper: The National Heart, Lung and Blood Institute Prevention and Early Treatment of Acute Lung Injury Clinical Trials Network. Early Restrictive or Liberal Fluid Management for Sepsis-Induced Hypotension. NEJM 2023. PMID: 36688507 [Access on Read by QxMD]

Clinical Question: In patients with sepsis-induced hypotension, after an initial 1 to 3L of IV crystalloids, does using a restrictive fluid strategy improve 90d all-cause mortality compared to a liberal fluid strategy?

What They Did:

• Crystalloid Liberal or Vasopressor Early Resuscitation in Sepsis (CLOVERS)
• Multicenter, unblinded, randomized, superiority clinical trial conducted at 60 US centers
• Randomization occurred within 4 hours after patients met the criteria for sepsis-induced hypotension refractory to initial treatment with 1 to 3L of IV fluid
• Patients randomized 1:1 to one of two arms for a 24 hour period:
  • Restrictive Fluid Strategy: Prioritizing vasopressors and lower IV fluid volumes; Rescue fluids were allowed
  • Liberal Fluid Strategy: Prioritizing higher volumes of IV fluids before vasopressor use; Recommended initial 2000cc of isotonic crystalloid followed by fluid boluses based on clinical triggers
  • In October 2019 a protocol amendment was made allowing for limited initial infusion of crystalloids to 1000cc if patients’ blood pressure and heart rate had been stabilized (SBP ≥110mmHg or MAP ≥70mmHg and HR <90BPM) and patient was volume replete (unlikely to benefit from additional IV fluids)

Outcomes:

• Primary: All -cause mortality before discharge home by 90d
• Secondary:
  • Number of days free from ventilator at 28d
  • Days free from renal-replacement therapy at 28d
  • Days free from vasopressor use at 28d
  • Days of the ICU at 28d
  • Days out of the hospital at 28d
• Safety:
  • Initiation of mechanical ventilation
  • New onset atrial and ventricular arrhythmias
  • Complications related to peripheral and central venous catheter use

Inclusion:

• Adult (≥18 years of age) patients
• Suspected or confirmed infection
• Sepsis-induced hypotension (SBP <100mmHg after administration of ≥1000mL of IV fluids)

 Exclusion:

• >4hrs since meeting criteria for hypotension refractor to IV fluids of at least 1000cc
• >24hrs since presentation to the hospital
• Previous receipt of >3000cc of IV fluids
• Presence of fluid overload
• Presence of severe volume depletion from nonsepsis causes

Results:

• 1563 patients enrolled
• Resuscitation therapies administered prior to randomization were similar between two groups:
  • Restrictive: Median 2050mL (Range: 1500 to 2457mL)
  • Liberal: Median 2050mL (Range: 1371 to 2442mL)
• Median time from meeting trial eligibility criteria to randomization: 61min (No real difference between groups)
• Majority of patients were randomized from the ED: 91.9% of enrolled patients
• Resuscitation therapies administered during the 24hr protocol period differed between the two groups:
  • Restrictive fluid strategy group received less fluids than the liberal fluid strategy group (Median difference: -2134mL; 95% CI -2318 to -1949)
  • Restrictive fluid strategy group received earlier, more prevalent, and longer duration of vasopressors vs the liberal fluid strategy group
• Median Volume of Fluid During 1^st 6 Hours After Randomization:
  • Restrictive: 500mL (Range 130 to 1097mL)
  • Liberal: 2300mL (Range 2000 to 3000mL)
  • Difference: -1800mL
  • This difference remained for the 1^st 24hrs after randomization as well (1267mL vs 3400mL)
  • The difference also remained for the 1^st 24hrs after randomization when including fluids administered pre-enrollment: 3300mL vs 5400mL
  • Most common fluid type used was Lactated Ringer’s Solution
• Vasopressor Use in the 1^st 24hrs After Randomization:
  • Restrictive: 59%
  • Liberal: 37%
  • Also initiated earlier (mean difference -1.4hrs) and used for longer (mean difference 4.2hrs)
• 90d All-Cause Mortality:
  • Restrictive: 14.0%
  • Liberal: 14.9%
  • Estimated Difference: -0.9%; 95% CI 04.4 to 2.6; p = 0.61
  • Prespecified subgroup analyses did not show differences in treatment effects (Included SBP <90mmHg, Chronic Heart Failure, ESRD, Pneumonia as Cause of Sepsis)
• Number of reported serious adverse events were similar between groups

Strengths:

• Asked a clinically important question
• Primary outcome was objective (i.e. mortality)
• Monitored protocol adherence in the first 300 patients and in a 10% random sample of patients throughout the rest of the trial
• Groups had similar baseline characteristics and treatment before randomization
• Audited protocol adherence was high in both groups with overall adherence of 97% in the restrictive group and 96% in the liberal group. Adherence was sustained throughout the duration of the trial
• Authors achieved separation between groups in terms of the amount of fluid given between both strategies

Limitations:

• Convenience sample of enrolled patients mostly during weekdays may have caused a selection bias (Only enrolled ≈1500 patients from 60 hospitals over a near 4 year period).
• Also ≈3300 eligible patients screened of which many not enrolled. 900 patients could not get informed consent, ≈880 patients refused consent, ≈880 patients the MD refused to enroll, and ≈350 not excluded but not enrolled
• Excluded patients with extremes of volume overload or volume depletion which makes drawing conclusions in these populations not possible
• Although there was high adherence to the protocol, some patients in the restrictive fluid group received more fluid than was intended and vasopressors were started later than intended. Additionally, some patients in the liberal fluid group received less fluid than was intended and vasopressors were started earlier than was intended.  Both of these protocol violations could have diluted the difference between groups
• Unblinded trial: Clinical team was aware of which group patients were randomized to. This could have influenced care the patients received as well as reporting of adverse events (i.e. higher reporting of volume overload in the liberal fluid group or lesser reporting of volume overload in the restrictive fluid group)
• Goals of care were driven by blood pressure and lactic acid targets. It is unclear if outcomes would have differed with different targets or allowing for lower blood pressure values
• Trial protocol was only for 24 hours after randomization. It is unclear if a longer treatment period would have produced differing results
• Study terminated early which did not allow authors to achieve the appropriate statistical power required to find a between group difference of 4.5% in the incidence of death before discharge home by day 90. Even if the authors had completed the trial with the remaining patients having good outcomes in the restrictive strategy group, this trial still may not have reached statistical significance.
• The patients enrolled didn’t seem that sick (Baseline SBP 93.5mmHg and Lactic Acid of 2.9) with only ≈60% of patients in the restrictive group requiring vasopressors at all. Additionally, mortality rate was ≈14% in this trial (Compared to ProCESS 19%, ARISE 18.5%, and ProMISe ≈29%).

Discussion:

• Authors were looking for an absolute between-group difference of 4.5% in the incidence of death before discharge home by day 90 (Primary outcome) assuming death would occur in 15% of patients in the liberal fluid group and 10.5% in the restrictive fluid group. This would require a sample of 2320 patients to be enrolled to achieve a 90% power (Trial did not reach this number as the trial was terminated for futility after 2/3rds of patients enrolled)
• 3 instances of vasopressor extravasation among 500 patients who received peripherally administered vasopressors between randomization and 72hrs; All 3 events resolved without intervention and did not have any residual clinical consequences
• My biggest issue not only with this study, but even previous trials is lumping sepsis patients into one group as if they are all the same. Sepsis is a heterogenous process with different causes, and different responses by patients. Clinical judgement is still alive and well…one protocol will not work for a heterogenous disease process.

Author Conclusion: “Among patients with sepsis-induced hypotension, the restrictive fluid strategy that was used in this trial did not result in significantly lower (or higher) mortality before discharge home by day 90 than the liberal fluid strategy.”

Clinical Take Home Point: In a group of patients with septic shock (sepsis-induced hypotension), receiving initial treatment with 1 to 3 liters of IV fluids, a restrictive fluid strategy (with early vasopressor use) did not result in a significantly lower (or higher) mortality before discharge home by day 90 compared to a liberal fluid strategy. Basically, in a heterogenous disease process like sepsis, use your clinical judgement to determine how much fluid any particular patient needs.

References:

1. The National Heart, Lung and Blood Institute Prevention and Early Treatment of Acute Lung Injury Clinical Trials Network. Early Restrictive or Liberal Fluid Management for Sepsis-Induced Hypotension. NEJM 2023. PMID: 36688507 [Access on Read by QxMD]

For More Thoughts on This Topic Checkout:

• The Bottom Line: CLOVERS
• First10EM: The CLOVERS Trial – Does Nothing Matter in Sepsis
• FOAMcast: CLOVERS Trial – Restricted vs Liberal Fluids in Sepsis-Induced Hypotension
• REBEL EM: The CENSER Trial – Early Norepinephrine in Septic Shock
• EMCrit: EMCrit 345 – I Guess We Need to Talk About CLOVERS and Fluids in Sepsis (Hopefully for the Last Time Ever)

Post Peer Reviewed By: Anand Swaminathan, MD (Twitter: @EMSwami)