Explanation for the choice of comparators

Both treatment pathways reflect widely accepted, albeit usually implemented in a non-standardized manner, treatment approaches used in North American Institutions. As it is unknown which approach is most beneficial, there is equipoise. Both pathways have been reviewed and accepted by critical stakeholders at each institution (including emergency medicine, nephrology, gastroenterology, infectious diseases and hospital medicine) as reasonable and appropriate care options. Embedding protocolized management of either clinical approach (i.e., hyperhydration or conservative fluid management) may lead to improvements over current practice, reducing undesirable outcomes by avoiding dehydration, use of antibiotics or antimotility-agents, and by facilitating close follow up and early recognition of complications.

Intervention description Fluid protocol #1—hyperhydration

In this pathway, all eligible children are admitted for the administration of intravenous fluids. To rapidly and safely achieve intravascular volume expansion, based on existing evidence [19, 20, 28], the starting intravenous fluid volume will be 200% of maintenance fluids calculated using the Holliday–Segar formula [38] with the rate adjusted based on clinical parameters and targets.

The following specifics will form the basis of the fluid management protocol (Fig. 3):

Fig. 3

The hyperhydration care pathway recommends the infusion of 200% maintenance fluids until a hematocrit reduction of 20% and weight gain of 10% occur. Patients may be discharged if the following criteria are met even if the targets are not achieved: (1) platelet count is > 50 × 109/L (in absence of transfusion) AND they have increased by ≥ 5%* since the preceding test AND (2) absence of diarrhea (loose or watery stool) × 24 h AND (3) ≥ 5 days since the onset of diarrhea. *If > 10 days since the onset of diarrhea, up to a 5% decrease in platelet count since preceding test is acceptable

1)

Reversal of dehydration: initial rehydration strategies should focus on rapidly reversing dehydration.

2)

Infusion of 200% of maintenance fluids × 24 h provided, ideally, as a balanced crystalloid (e.g., PlasmaLyte™, Ringer’s Lactate) intravenous solution. Electrolytes and dextrose may be administered as deemed appropriate; customized solutions are permitted. Intravenous fluid solutions containing < 130 mEq/L sodium should be avoided as they increase the likelihood of developing hyponatremia and are less effective in achieving intravascular volume expansion.

3)

If the hematocrit reduction is < 20% from the initial value at 24 h, and the weight gain is < 10%, repeat step #2 (infusion of 200% maintenance intravenous fluids for 24 additional hours) unless the weight gain is ≥ 10% in which case move to step #5.

4)

Oral fluids permitted ad lib.

5)

If the target hematocrit reduction (20% decrease from baseline) AND a 10% weight gain are achieved, adjust total intravenous fluid volume to maintain targeted weight gain and prevent worsening fluid accumulation (i.e., replace insensible losses plus urine and stool output). If both targets are not achieved, then return to step #2.

◦ Insensible losses will be estimated to be 400 mL/m2 and these losses should be replaced, along with additional losses through output, every 4 to 6 h [39].

◦ Daily weights should be performed and be used to guide the assessment of fluid status with additional fluid adjustments made, as required, to maintain the targeted weight gain.

6)

Complete blood count, electrolytes, and kidney function should be repeated every 24 h (or more frequently as deemed clinically indicated).

7)

Consider nephrology consultation for all children.

Safety targets

Based on the use of this fluid protocol in children with HUS who experienced a mean weight gain of 12.5% without any serious adverse events (SAEs) [18], we have identified parameters ((1) weight gain of 10% AND (2) evidence of hemodilution (i.e., 20% reduction in hematocrit)) which will serve to indicate that intravascular volume expansion has been achieved. This dual parameter approach ensures that hemodilution does not simply reflect hemolysis due to the ongoing microangiopathic process while also identifying a maximal weight gain target to help limit the risk of volume overload.

Fluid protocol #2—conservative fluid management

The conservative fluid protocol has been designed to align and integrate into existing practice patterns. Clinicians can manage children as they normally would with a target of euvolemia. In this pathway, children will undergo a protocolized baseline evaluation that includes reversal of dehydration, if present, and close laboratory monitoring (Fig. 4). The need and approach to reverse dehydration, if present, is at the discretion of the clinical care team. In the absence of evidence of microangiopathy (i.e., normal urinalysis, lactate dehydrogenase, hemoglobin and platelet counts, and creatinine concentrations), the decision to admit or discharge the child is at the discretion of the clinical care team. If microangiopathy is present, admission for monitoring is recommended.

Fig. 4

The conservative fluid management pathway leaves the decision to admit the child to the discretion of the clinical care team. If microangiopathy is present admission is encouraged. For inpatients and outpatients, laboratory monitoring should be performed a minimum of every 24 h until discharge criteria are met. If intravenous fluids are administered, the maximum rate is 100% of maintenance and the target is euvolemia and maintaining weight gain below < 5%. The pathway should be discontinued once the same criteria are achieved as for the hyperhydration pathway (Fig. 3)

The hospitalization protocol includes 1)

Reversal of dehydration: initial rehydration strategies should focus on rapidly reversing dehydration.

2)

Targeting euvolemia: intravenous fluids should be administered at a maximum rate of 100% maintenance with the rate adjusted to take into consideration ongoing losses (e.g., vomiting, diarrhea, third spacing) and oral fluid intake.

◦ The intravenous fluid administered should be isotonic with the volume calculated employing the Holliday–Segar formula [38] with adjustments made as needed to target euvolemia.

◦ The fluids provided should be balanced crystalloid (e.g., PlasmaLyte™, Ringer’s lactate) with electrolytes and dextrose added as deemed appropriate; customized solutions are permitted. Intravenous fluid solutions containing < 130 mEq/L sodium should be avoided as they increase the likelihood of developing hyponatremia.

3)

Monitoring

◦ Daily weights should be performed and be used to guide the assessment of fluid status with additional fluid adjustments made, as required, to maintain euvolemia and weight gain < 5% above baseline.

◦ Complete blood count, electrolytes, and kidney function should be repeated every 24 h.

4)

Consider nephrology consultation for all children.

The outpatient management protocol includes 1)

Reversal of dehydration: initial rehydration strategies should focus on rapidly reversing dehydration, if present.

2)

Targeting euvolemia: oral fluids permitted ad lib.◦

3)

Monitoring:

◦ Complete blood count, electrolytes, and kidney function should be repeated every 24 h.

◦ Emergency department evaluation required if laboratory tests or clinical symptoms concerning for HUS. These include any of the following:

◾ New bleeding, bruising, petechial rash

◾ Severe abdominal pain

◾ Unusual/severe headache

◾ Tea-colored urine

◾ No urine output for 12 h

◾ Irritability/ altered mental status

◾ Edema

4)

Outpatient management otherwise as per the clinical care team.

Goal

Once dehydration has been reversed, fluid targets should aim to maintain euvolemia and avoid dehydration.

Criteria for discontinuing or modifying allocated interventions

Treatment decisions will be at the discretion of the responsible physician who may choose to deviate from to the allocated pathway. Clinical events and participant preference may lead to care modifications; these would not affect the conduct of study activities (e.g., follow-up). Specific documentation will be made regarding allocated pathway adherence, participant withdrawal, and study discontinuation.

End-of-fluid protocol criteria

The pathways will end when the at-risk window for developing HUS, or worsening HUS if present, has passed and this time point is defined by:

1)

Platelet count is > 50 × 109/L (in the absence of transfusion) AND

2)

Platelets have increased by ≥ 5%* (in the absence of transfusion) since preceding test AND

3)

Absence of diarrhea (loose or watery stool) × 24 h AND

4)

 ≥ 5 days since the onset of diarrhea

*If >10 days since the onset of diarrhea, up to a 5% decrease in platelet count since preceding test is acceptable.

In addition, the fluid pathways should no longer be followed in children meeting any of the following criteria:

1)

Decision to initiate RRT

2)

Anuria for ≥ 12 h

3)

Hypoxemia requiring supplemental oxygen

4)

Weight gain > 15%

5)

Hypertensive emergency

If any of these criteria are met, the treatment pathway will not resume once they are resolved; care tailored to each patient should be provided as directed by the responsible physician.

Strategies to improve adherence to interventions

Site investigators and champions at each site, which include a pediatric emergency medicine physician, nephrologist, pediatric hospital medicine specialist, and at certain sites a gastroenterologist, and/or infectious disease specialist, will be responsible for educating all relevant individuals (at the participating site and in their catchment/referral area) including attending physicians and trainees regarding the treatment pathways. All pathway eligible children will be monitored in real-time, and support will be provided to the clinical care team regarding pathway adherence which has been facilitated through the building and implementation of clinical care order sets embedded into the electronic medical record (EMR) at participating sites.

For each patient, the EMS (emergency medical services) Data Center (EDC, Salt Lake City, UT) will classify the amount of fluid administered in relation to their target as defined by the pathway to which the site has been allocated. Adherence at an individual patient level will be reviewed during weekly leadership team meetings and those cases where participants met our study definition of non-adherence (> 110% and < 175% maintenance fluids in the conservative fluid management and hyperhydration groups, respectively) will undergo review to identify potential etiologies and solutions in an attempt to mitigate future such events.

Hyperhydration

Data related to all intravenous fluids administered will be extracted from the medical record. For the calculation of pathway adherence, the intravenous fluid administration period begins (i.e., time of study initiation) at the:

1)

Time of hospital admission plus 2 h (for participants asked to come to the hospital)* OR

2)

Time the participant met eligibility criteria plus 2 h (for participants already in hospital)* OR

3)

Actual time of initiation of intravenous fluid administration if occurs before the additional 2-h period has lapsed

*Some participants will be admitted into the hospital before eligibility, and some will become eligible before admission (i.e., positive STEC result in child who is at home). The later of the time points between options (1) and (2) will be used to determine the timing of study initiation if (3) does not apply.

The intravenous fluid administration period ends (i.e., time of study end for pathway adherence determination) when any fluid protocol termination criteria are met (i.e., weight/hematocrit targets, complication ensues, end-of fluid protocol criteria) or after 48 h of fluid administration—whichever happens first. The targeted volume of fluid is defined by the administration of ≥ 175% of maintenance fluids calculated according to the Holliday–Segar formula for the duration of time the participant was in the volume expansion phase of the protocol [38]. Thus, adherence is defined as the administration of fluids in excess of: 100% maintenance fluids × 1.75 × number of hours the participant was in the volume expansion phase of the protocol. Boluses will be included in the quantification of total fluids. Exact monitoring and recording of oral fluid intake are logistically prohibitive and will not be included in the fluid calculation.

The management of patients who receive < 175% of maintenance intravenous fluids in the hyperhydration arm will be classified as nonadherent. However, all patients will be included in the intent-to-treat (ITT) analyses regardless of the volume of fluids administered; secondary analyses will include a per protocol adherence analysis.

Conservative fluid management

Outpatient/emergency department: All these participants will be considered adherent.

Inpatient: Adherence is defined by the administration of < 110% of maintenance fluids calculated according to the Holliday–Segar formula [38] during the period of time when the participant was in the fluid administration phase of the pathway (defined as per hyperhydration). Thus, adherence to the conservative fluid management pathway is defined as the administration of fluids at or below the following threshold: 100% maintenance fluids × 1.1 × number of hours the participant was in the fluid administration phase. Boluses will be included in the quantification of total fluids. Exact monitoring and recording of oral fluid intake are logistically prohibitive and will not be included in the fluid calculation.

The management of patients who receive > 110% of maintenance intravenous fluids will be classified as nonadherent. However, all patients will be included in ITT analyses regardless of the volume of fluids administered; secondary analyses will include a per protocol adherence analysis.

Relevant concomitant care permitted or prohibited during the trial

We will permit the concomitant administration of all medications, supplements, complementary and alternative therapies, treatments, and/or procedures as deemed necessary by the responsible treating physicians. The study will not supply or recommend any specific rescue medication. As part of the overall site education strategies, we will discourage the use of antibiotics, opioids, antimotility agents, and nephrotoxic agents such as non-steroidal anti-inflammatory drugs. We will discourage the use of hypotonic intravenous fluids and will recommend balanced solutions instead of 0.9% saline. However, as this is a pragmatic embedded clinical trial, the treating clinicians, with input from patients and guardians will be responsible to make all treatment decisions.

All relevant data will be extracted daily during chart reviews to enable clarification and confirmation in real-time, if required, and will include concomitant medications administered during the patient’s current illness and any antibiotics taken in the past 14 days. By collecting this information and incorporating it into study analyses, we will be able to identify the independent effects of administered concomitant medications and study interventions.

Provisions for post-trial care

As study participation involves only data and biospecimen collection, there are no anticipated harms associated with study participation. All participants will receive standard of care as per their hospital’s routine care of children with STEC infection.

Outcomes Primary outcome: MAKE30 (major adverse kidney events by 30 days)

Because clinical research evaluating the prevention and treatment of AKI has been hampered by the unclear relationship between acute changes in kidney function and longer-term outcomes, the National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK) [40] workgroup on Clinical Trials in AKI recently recommended the use of the MAKE30, a composite endpoint, for phase III trials related to AKI [25]. It is defined by:

1.

Death OR

2.

Provision of dialysis OR

3.

Sustained loss of kidney function (at 30 days) reflected by a 100% [41] increase (i.e., doubling) of serum creatinine from baseline

This outcome will be classified following the completion of data extraction and the 30-day follow-up assessment (time 0, as defined above, is used to calculate the 30-day interval). It will be adjudicated based on laboratory values, chart documentation, and 30-day laboratory results. As baseline creatinine values are rarely available for children, all participants will be assigned a value which will be calculated employing a standardized baseline glomerular filtration rate (GFR = 120 ml/min/1.73m2) [42].

Outcome justification

The rationale for selection of the MAKE30 as the primary outcome is twofold: (1) the kidneys are the major target organ through which STEC infections cause adverse outcomes and (2) the kidneys are the organ system most likely to benefit from the proposed intervention. Several prior cohort studies demonstrated that RRT use, a component of the MAKE30, is reduced with the use of aggressive intravascular fluid administration in STEC-infected children [18,19,20]. In addition, the MAKE30 is a widely used and easily assigned composite endpoint of AKI as it includes variables related to kidney prognosis and is based on objective endpoints (i.e., death, provision of RRT, sustained loss of kidney function at 30 days) and thus is unlikely to be influenced by knowledge of study group allocation. The MAKE30 outcome has been used in several recent high profile acute care trials focusing on AKI [41, 43]. In addition, pediatric sepsis survivors who experience the MAKE30 outcome have a three- to fivefold higher rate of developing chronic kidney disease [44].

Secondary outcomes

These will focus on quantifying other measures of efficacy and safety. They have been selected to provide supportive information about the effect of volume expansion on the primary outcome and the natural history of disease (see Table 1).

Participant timeline (Table 2)Table 2 Schedule of activitiesSample sizeTarget study sample size

We estimate the overall study sample size required is 1040. The number to be screened will likely be 25% higher given that some children will be ineligible, and some will decline consent. Preliminary data indicate that the primary outcome event rate in the conservative fluid management arm will be approximately 10%. The study will be powered to detect a 5% absolute reduction (risk ratio 0.5) in the rate of MAKE30 in the hyperhydration arm. Using a conservative estimate of 5–15 subjects per year per site, we used simulated data (R software) to generate a power table (Table 3) for a 20- or 26-site cluster-crossover design under varying underlying intra-class correlation (ICC) values. The table shows that power is not highly dependent upon ICC and that 20 and 26 sites are required for 80% and 90% power, respectively, at the ICC estimated from preliminary data (i.e., 0.106) [16]. We expect very low levels of missing data, as participants with elevated creatinine values at the time of discharge will have laboratory follow-up as part of their standard clinical care. Patients who deteriorate and develop the MAKE30 outcome will be readily identified. Thus, the impact of loss to follow-up on power is expected to be minimal.

Table 3 Power calculations for the primary outcome

The anticipated HUS rate in the conservative fluid management arm is 17.5%, and the rate of extrarenal complications is 7.5%. Detectable absolute differences for these two outcomes with 80% and 90% power, at approximately the same ICC as estimated for the primary outcome, are provided (Table 4). Here, a conservative Bonferroni correction (alpha = 0.025) is used.

Table 4 Power calculations for secondary outcomesRecruitment Source of participants

Research teams will seek to be notified and obtain consent to approach all eligible high-risk, STEC-infected children within their catchment area. Study sites are all leading academic tertiary care centers and leads will disseminate their allocated pathway and a desire to serve as a clinical care site for all STEC-infected children. They will connect with satellite institutions and laboratories to be informed of all potentially eligible children. At their local institutions, through mechanisms put in place with their laboratory and all relevant clinical care teams (e.g., ED, infectious disease, nephrology, gastroenterology, pediatrics), the research team will be notified of potentially eligible children. Study teams will take advantage of their electronic medical records to set up reports that inform them of all positive stool tests and relevant diagnoses (i.e., bloody diarrhea, STEC, HUS). The cluster design promotes the initiation of the allocated treatment at any time and does not require research team involvement of participant research consent. The latter can be obtained following treatment implementation. In addition, consent can be obtained for data collection electronically, and as such, children meeting all eligibility can be enrolled without having to come to the study institution (i.e., they can be treated in accordance with the allocated pathway either at a satellite institution or by a care provider who is willing and able to adhere to the pathway).

Recruitment venues

Recruitment into the randomly allocated intervention (i.e., hyperhydration vs. conservative fluid management) will occur in the following venues:

1)

Emergency departments

2)

Inpatient units: general pediatrics, nephrology, gastroenterology, infectious diseases

3)

Medical microbiology laboratories: partnering laboratories will develop a process to provide the research team with the contact information for the ordering physician so that the allocated pathway can be shared to guide care and to request permission to contact the patient/caregivers of children with positive STEC specimens to obtain research consent. This approach will enable participating sites to recruit participants from their broader geographic catchment area.

How potential participants will be identified and approached

Participant identification/approach will vary between sites, but will be based on the following key concepts:

1)

Clinician identification: once a child is identified as having a high-risk STEC infection, clinicians will provide patients/caregivers with an STEC information sheet which clearly states that the site is participating in this study, the institution’s treatment approach within the study, and the pros/cons of the two treatment protocols. Clinicians at all sites will be trained to contact the appropriate research team member to determine research eligibility and to obtain permission for the research team to approach the caregiver to request informed consent for data extraction and biospecimen collection.

2)

Research team identification:

A)

Daily screening: reports from the ED and inpatient units will be created to identify all children with potentially eligible chief complaints (e.g., bloody diarrhea, abnormal laboratory result) and discharge diagnoses (e.g., STEC, HUS). Patients infected with a high-risk STEC (evidence of an E. coli O157:H7, detection of Stx 2 or a gene encoding this toxin, bloody diarrhea, and presence of Stx or a Shiga toxin gene not otherwise specified, or evidence of early HUS) will be identified (Fig. 5). The research team will approach the clinical team caring for such patients to review research eligibility and to obtain permission to approach the patient for research consent.

3)

Medical microbiology laboratory identification: reports identifying all STEC positive stool specimens will be set up. For all children identified potentially eligible children, contact will be initiated by the research team with the responsible physician who will be informed of the sites allocated pathway and who will be asked to obtain consent for the research team to approach the family.

Fig. 5

Approach to optimize the enrollment of high-risk STEC infected participants

For all discharged and outpatients, the child will be asked to go to the ED (if required for the hyperhydration arm) for screening (i.e., baseline laboratory testing) and treatment per the institution allocated pathway. If in the conservative fluid management arm, at the discretion of the responsible physician, the child may be managed as an outpatient and consented for research electronically.

Targeted outreach activities

To optimize awareness of the embedded pathways at participating sites to optimize both pathway adherence and study awareness to promote recruitment, we will conduct knowledge dissemination activities aimed at the following groups:

1)

Physicians (particularly pediatricians, hospitalists, emergency physicians, nephrologists, gastroenterologists, infectious disease) and nurses at the participating center. These strategies will include emails, lectures, seminars, section/department meetings, creation of order sets, and clinical care guidelines.

2)

Physicians providing primary care, hospital-based care, and emergency department care in the institution catchment area—through emails, lectures, seminars, section/department meetings, and the sharing of order sets and clinical care guidelines.

3)

Medical microbiology laboratories—through direct communication with laboratory directors to optimize diagnostic testing approaches and mechanisms to become notified of positive test results.