Physiologic Effects of Steroids in Cardiac Arrest

BACKGROUND AND RATIONALE Despite recent improvements in the quality of care, in-hospital cardiac arrest is still associated with a high probability of poor outcome. Patients resuscitated from vasopressor-requiring cardiac arrest frequently exhibit early postresuscitation hemodynamic instability that is poorly responsive to hemodynamic support with high vasopressor doses and intravenous fluids titrated to cardiac filling pressures of at least 12 mmHg. Furthermore, preceding studies indicate that postresuscitation disease is a "sepsis-like" syndrome characterized by plasma cytokine elevation, endotoxemia, coagulopathy and adrenal insufficiency contributing to postresuscitation shock. Steroids are currently being used for improving hemodynamics in septic patients, however, their effect on postresuscitation hemodynamics has not been thoroughly elucidated yet.

OBJECTIVES To determine whether stress-dose steroid supplementation during and after cardiopulmonary resuscitation (CPR) improves the hemodynamic parameters (arterial blood pressure, cardiac output, cerebral blood flow) in patients with in-hospital cardiac arrest. Furthermore to study the effects of steroid administration on the inflammatory response and organ failures, and to determine potential, corticosteroid-associated complications such as hyperglycemia, infections, bleeding peptic ulcers and paresis. In summary, we aim to directly assess the physiological effects and safety of steroids during and after CPR. The possible clinical usefulness of steroids during and after CPR - in the context of early postresuscitation hemodynamic support - currently corresponds to an important knowledge gap, as recently acknowledged by Guidelines Evidence Reviewers.

METHODS SETTING Intensive/coronary care units, (ICUs/CCUs) of the Evaggelismos Hospital, Athens, Greece (1,200 beds) and of the Larissa University Hospital, Larissa, Greece (700 beds).

PATIENTS

Inclusion criteria:

Patients who have experienced an in-hospital, vasopressor-requiring cardiac arrest, according to guidelines for resuscitation from 2015.

Exclusion criteria are reported in the dedicated subsection.

ETHICS AND INFORMED CONSENT The study will be conducted in concordance with European Union Clinical Trials Regulation No 536/2014 and the Helsinki Declaration. Due to the emergency situation, consent will not be requested for steroid supplementation during CPR. The patients' families and patients who regain consciousness and communication ability during follow-up will be informed about the study as soon as possible, and any objection will result in exclusion of the patient data from any subsequent analyses. Informed, written next-of-kin consent and non-written patient consent (whenever feasible) will be requested as soon as possible for stress-dose hydrocortisone in postresuscitation shock and continued participation in the study. If consent cannot obtained before patient death, the patient's next of kin will be informed of the study and their permission for inclusion of the patient data in the subsequent analyses will be requested. All consent procedures will additionally be documented on the patient's medical record.

The original protocol version has been approved by the Institutional Review Board (IRB) of Evaggelismos General Hospital on July 14 2016 (Approval No. 126/16-6-2016), and by the IRB of Larissa University Hospital on October 10, 2016 (Approval No. 46113/11-10-2016 - IRB Discussion No. 13/10-10-2016 Θ.6). Subsequent protocol amendments have been approved by the Evaggelismos IRB on January 24, 2017 (Approval No. 8/26-1-2017), and this has been communicated to the IRB of Larissa University Hospital.

STUDY DESIGN We propose a prospective, randomized, double-blind, placebo-controlled, parallel- group clinical trial.

RANDOMIZATION Research Randomizer version 4 (https://www.randomizer.org/) will be used by the study statistician for group allocation. For each study center, random numbers (range, 1-100) will be generated in sets of 4. Each random number of each set will be unique and correspond to 1 of the consecutively enrolled patients. In each set, an odd or even first number will result in assignment of the corresponding patient to the control or steroids group, respectively. In each study center, the group allocation rule will be known solely by the pharmacists who will prepare the study drugs.

CPR AND POSTRESUSCITATION INTERVENTIONS We will enrol adult in-patients with cardiac arrest due to ventricular fibrillation/pulseless tachycardia not responsive to three direct current countershocks, or asystole, or pulseless electrical activity. Study treatments will be administered during the first CPR cycle postenrollment. Patients will be randomized to receive either methylprednisolone 40 mg (Steroids group) or normal saline placebo (Control group) on the first, postenrollment CPR cycle. Otherwise, advanced life support will be conducted according to the 2015 guidelines for resuscitation. After resuscitation, patients will be treated with either stress-dose hydrocortisone of 240 mg daily for 7 days maximum (Steroids group), or saline placebo (Control group). More specifically, at 4 hours after ROSC, patients will receive 100 mL/day (average pump infusion rate ~ 4.2 mL/h) of normal saline that will either contain the stress-dose of hydrocortisone (Steroids group) or solely saline placebo (Control group) for a maximum of 9 days. On days 8 and 9 the hydrocortisone dose of the Steroids group will be tapered to 120 mg and 60 mg, respectively, and finally discontinued on day 10 postrandomization. On ICU/CCU admission, patients will receive a central venous line, and an arterial line, either standard or as part of pulsatility index continuous cardiac output monitoring. Patients with a standard arterial line will also receive a pulmonary artery catheter, provided that attending physicians also agree to this.

DOCUMENTATION AND PATIENT FOLLOW-UP CPR attempts will be documented according to the Utstein style. Hemodynamics and gas-exchange, electrolytes, glucose, central body temperature, lactate and administered fluids and vasopressor/inotropic support will be determined/recorded during CPR, and at ~20 min and ~4 hours as well as at 24, 48 and 72 hours after the return of spontaneous circulation (ROSC); ROSC will be defined as sustained presence of a palpable arterial pulse for at least 20 min. Postresuscitation cardiac output will be monitored for at least 72 hours post-ROSC, and postresuscitation cardiac function will be assessed by ultrasonography within the first hour after ICU admission and at 72 hours post-ROSC. Central-venous blood gas analysis will also be performed at the aforementioned time points and blood samples will be taken for the determination of cytokines at approximately 20 min and 4, 24, 48, and 72 hours post-ROSC.

Follow-up during the first 10 days postrandomization will include 1) Determination/recording of hemodynamics and hemodynamic support, gas-exchange, fluid balance of the preceding 24 hours, and arterial blood lactate and central venous oxygen saturation at 9 a.m.; 2) Daily determinations of serum pro-inflammatory cytokines, and 3) Daily recording (within 8-9 a.m.) of laboratory data, and prescribed medication.

The results of 4 daily determinations (1 every 6 hours) of blood glucose will also be recorded to subsequently analyze the incidence of hyperglycemia (blood glucose exceeding 200 mg/dL -11.1 mmol/L). Follow-up to day 60 post-ROSC will include organ failures, and ventilator-free days. Morbidity/complications throughout ICU/CCU and hospital stay, and times to ICU/CCU and hospital discharge will also be recorded.

Study outcomes are reported in the dedicated subsection.

PROTOCOL AMENDMENTS [Approved, January 24, 2017]

Amendment: 1. Part A: Postresuscitation Cardiac Output (CO) has been moved from the primary to the secondary study outcomes (see also Outcome Measures' subsection). Reason for amendment: We ultimately anticipate to be able to collect such data in <50% of the patients. Part B: Furthermore, the measurement time point of "8 hours postresuscitation" will no longer be part of the protocol and this will apply for the total of the physiological measurements. This aims to limit attending investigator workload.

Amendment 2: Further clarification regarding the "acceptable" time frame for ICU / CCU admission of successfully resuscitated patients. This is a very important logistical issue and will include measures (e.g. monitoring and medical service support and oversight) to optimize the quality of care until ICU admission and the specification of the "maximum acceptable time to ICU admission". In the VSE 1 and 2 studies, this time has been 12 hours. During 2016, the Evaggelismos Department of Intensive Care Medicine has experienced a reduction of 5 ICU beds (i.e. from 30 to 25) and 8 HDU (high-dependency unit) beds from 20 to 12. Therefore, regarding cardiac arrest patients, we have decided to extend the target time limit for admission to 24 hours. Regarding CORTICA, any further delay is to be reported as an "unpreventable" protocol breach, and in the case of a possible extreme circumstance (e.g. a severe flu outbreak), any patients with a projected ICU admission time of more than 48 hours will be excluded. Hence, "a projected ICU admission time of more than 48 hours" will constitute an additional exclusion criterion.

Amendment 3: Further clarification of the "terminal illness" exclusion criterion - life expectancy of no more than 6 weeks): (i) Patients with metastatic cancer [with confirmed bone and/or brain metastases], and/or primary / metastatic disease causing respiratory failure with/without additional organ-system failures [as defined by a corresponding Sequential Organ Dysfunction Assessment (SOFA) subscore of 3 or 4]; (ii) Patients with a pre-arrest worst SOFA score of >=15 (this corresponds to a probability of death of at least 90%); and (ii) Patients with immunosuppression and a new, hospital-acquired septic complication).

Amendment 4: Addition of the following Exclusion Criterion: "Any deviation from the hospital's standard resuscitative procedure" (e.g. poor adherence to the standard Advanced Life Support algorithm such as an epinephrine dosing error or use of atropine, or "unjustified" interruption of chest compressions compromising CPR quality).

Amendment 5: Addition of the following Exclusion Criterion: Pre-arrest diagnosis of an "active" peptic ulcer; that is, either preceding gastroscopic confirmation of a peptic ulcer, or clinical evidence of acute, pre-arrest gastrointestinal bleeding, attributable to peptic ulcer disease.

Amendment 6: Follow-up during the first 10 days: The time of determination of cytokines will be 20 min, 4, 24, 48, and 72 hours post ROSC [as specified in the corresponding outcome measure], and 7 days post-ROSC (that is, the determinations of day 4, 5, 6, 8, 9, and 10 will be cancelled); Reason for change: Study cost reduction.

STATISTICS Data will be reported as mean±standard deviation, or median (interquartile range), or number (percentage), unless otherwise specified. Distribution normality will be tested by Kolmogorov-Smirnov test. Dichotomous and categorical variables will be compared by two-sided chi-square or Fisher's exact test. Continuous variables will be compared by two-tailed, independent samples t test or Mann-Whitney exact U test. P- values of multiple t-test comparisons will be subjected to the Bonferroni correction. We will use mixed model analysis to compare repeatedly measured variables between the two groups. Survival data will be analyzed by a previously employed methodology of multivariable Cox regression. Based on previously published data on the mean arterial pressure at 24 hours postresuscitation, to detect an effect size d of 0.761 with an α error probability of 0.015 and power 0.80, we need to enroll a total of 88 patients (44 in each group). A target enrollment of 100 patients with ROSC for at least 20 min will likely adequately compensate for possible dropouts or missing data. This target enrollment actually corresponds to an alpha value of 0.008 and a power of 0.80 according to the aforementioned, predicted effect size d of 0.761. Expected results pertain to a steroid-associated benefit with respect to the primary outcomes, consistent with results on secondary outcomes.