Pragmatic Pediatric Trial of Balanced Versus Normal Saline Fluid in Sepsis: A Pilot Feasibility Study

Approximately 5,000 children die from septic shock each year in the US and thousands more die
worldwide. Despite widespread implementation of resuscitation protocols, contemporary studies
still report 2-6% mortality for children with septic shock treated in the pediatric emergency
department (ED). In the investigators' recent survey of the Pediatric Emergency Care Applied
Research Network (PECARN), 45% of physicians had treated a child for septic shock in the ED
who subsequently died in the hospital in the past two years.

Fluid resuscitation is the cornerstone of resuscitation for hypovolemia and shock, and
intravenous fluids are among the most commonly used therapies worldwide. Yet, there remains
uncertainty as to the most appropriate fluid type to restore effective blood volume and
optimize organ perfusion. In the absence of a clear role for the early use colloids,
administration of crystalloid fluids is generally preferred (except in cases of hemorrhage).
For septic shock, in particular, crystalloid fluids have long been the standard resuscitative
fluid. Crystalloid fluids can be categorized as non-buffered (most commonly 0.9% normal
saline [NS]) or buffered/balanced (in the US, this is most commonly lactated Ringer's [LR])
solutions. NS and LR are inexpensive, stable at room temperature, and nearly universally
available with identical storage volumes and dosing strategies. Notably, both are also of
proven clinical benefit in septic shock and have extensive clinical experience for use in
fluid resuscitation of critically ill patients. However, while NS is currently used in 80-95%
of cases of septic shock, an increasing body of data now suggest that LR resuscitation may
have superior efficacy and safety. Buffered crystalloids, including LR, have demonstrated a
1-4% absolute mortality reduction and up to a 50% lower odds of dialysis compared to NS in
observational and non-randomized interventional studies in adult sepsis. Nevertheless,
because definitive conclusions have not been able to be drawn from existing observational and
non-randomized studies, NS overwhelmingly remains the most commonly used fluid based on
historical precedent while controversy remains.

To definitively test the comparative effectiveness of NS and LR, a well-powered randomized
controlled trial (RCT) is necessary. A large pragmatic randomized trial embedded within
everyday clinical practice provides a cost-efficient and generalizable approach to inform
clinicians about best comparative effectiveness of common therapies. Unlike explanatory RCTs,
pragmatic trials need heterogeneity in patients, non-study therapies, and settings. To
accomplish this, these trials must be large enough to detect small effects and simple enough
to incorporate into routine clinical practice. The characteristics of LR and NS provide the
ideal scenario for a large pragmatic trial.18 An ED-based trial is necessary to enroll
patients at initiation of resuscitation. While any benefit is expected to be small, even a
1-2% absolute reduction in mortality that is in line with prior adult studies would be a
clinically important difference by saving the lives of 50-100 children in the US (and many
more worldwide) each year. This overall public health impact is commensurate with changing
from NS to LR because such a practice change is a simple, cost-neutral shift from largely
using NS to largely using LR.

However, before embarking on a large, pragmatic randomized trial that will determine the
comparative effectiveness and safety of NS and LR, several concerns regarding feasibility of
such a trial need to be addressed including a) ensuring adequate compliance with study fluid
administration within each randomized arm using the proposed pragmatic study design, b)
determining that a sufficient proportion of patients can be enrolled using the proposed
pragmatic study design that will be embedded within routine clinical practice rather than use
of a dedicated study team, and c) demonstrating that the study can feasibly be performed
using EFIC when enrolling critically ill infants, children, and adolescents into this
clinical trial. Demonstrating these feasibility criteria at a single site will strongly
support success in a larger, multicenter study that will enroll several thousand patients
across the 18 sites comprising Pediatric Emergency Care Applied Research Network (PECARN) to
test morbidity and mortality outcomes.

Inclusion Criteria:

1. Males or females age >6 months to <18 years

2. Clinician concern for septic shock, operationalized as:

1. a "positive" ED sepsis alert confirmed at the physician-led "sepsis huddle" OR

2. a physician diagnosis of suspected septic shock requiring parenteral antibiotics
and fluid resuscitation as per the ED sepsis management pathway

3. administration of at least 20 mL/kg IV/ intraosseous (IO) fluid resuscitation

4. Receipt of ≤40 mL/kg IV/IO crystalloid fluid prior to randomization

5. Additional fluid deemed likely to be necessary to treat poor perfusion, defined as
either hypotension or abnormal (either "flash" or >2 second) capillary refill (as
determined by clinician's judgment)10

6. Parental/guardian permission (informed consent) if time permits; otherwise, EFIC
criteria met

Exclusion Criteria:

1. Clinician judgement that patient's condition deems it unsafe to administer either NS
or LR (since patients will be equally likely to receive NS or LR at time of study
enrollment), including (but not limited to):

1. Clinical suspicion for impending brain herniation based on data available at or
before patient meets criteria for study enrollment

2. Known hyperkalemia, defined as non-hemolyzed whole blood or plasma/serum
potassium > 6 mEq/L, based on data available at or before patient meets criteria
for study enrollment

3. Known hypercalcemia, defined as plasma/serum total calcium >12 mg/dL or whole
blood ionized calcium > 1.35 mmol/L, based on data available at or before patient
meets criteria for study enrollment

4. Known acute fulminant hepatic failure, defined as plasma/serum alanine
aminotransferase (ALT) >10,000 U/L or total bilirubin >12.0 mg/dL, based on data
available at or before patient meets criteria for study enrollment

5. Known history of severe hepatic impairment, defined as diagnosis of cirrhosis,
"liver failure", or active listing for liver transplant

6. Known history of severe renal impairment, defined as current dependency on
peritoneal dialysis or hemodialysis

7. Known metabolic disorder, inborn error of metabolism, or primary mineralcorticoid
deficiency (e.g., mitochondrial disorder, urea cycle disorder, amino acidemia,
fatty acid oxidation disorder, glycogen storage disorder, congenital adrenal
hypoplasia, Addison's disease) as reported by subject, LAR or accompanying
caregiver, or as listed in the medical record

2. Known pregnancy determined by routine clinical history disclosed by patient and/or
legally authorized representative (LAR) (or other accompanying acquaintance)

3. Known prisoner as determined by routine social history disclosed by patient and/or LAR
(or other accompanying acquaintance)

4. Known allergy to either normal saline or lactated Ringer's as determined by routine
allergy history disclosed by patient and/or LAR (or other accompanying acquaintance)
or as indicated in the medical record

5. Indication of prior declined consent to participate based on presence of "PRoMPT BOLUS
Opt-Out" bracelet