Effect of Levocarnitine vs Placebo as an Adjunctive Treatment for Septic Shock

Key Points Question Do 1 or more doses of levocarnitine reduce organ failure in septic shock at 48 hours, and, if so, what is the likelihood of success in a phase 3 trial? Findings In an adaptive randomized, blinded clinical trial of 250 adults, the most efficacious dose of levocarnitine (18 g) demonstrated a posterior probability of efficacy of 0.78, which did not reach the a priori threshold of 0.90. Meaning Levocarnitine did not meaningfully reduce organ failure at 48 hours in patients with septic shock.

International Conference on Harmonisation Guidelines. This study will be registered on clinicaltrials.gov prior to enrollment and in public communication will be referred to as RACE (Rapid Administration of Carnitine in sEpsis).

OBJECTIVES AND SPECIFIC AIMS
More humans die in the intensive care unit from sepsis than from any other cause. Death from sepsis is the culmination of widespread hypoperfusion, cellular hypoxia, and multiple organ failure. A growing body of evidence shows that early therapeutic intervention improves outcome for patients with sepsis. Novel targeted strategies that bolster a strong and durable systemic hemodynamic response have been proven to reduce or even reverse organ dysfunction in patients with sepsis. Lcarnitine provides the key elements of a novel therapy to ameliorate the adverse hemodynamic 5/P Version 1.3.6 Dated: 07/26/2016 effects of sepsis. Severe physiological stress forces energy metabolism to shift from primary fatty acid oxidation toward glycolysis and lactate oxidation. Prior work has shown that exogenous L-carnitine administration enhances glucose and lactate oxidation, attenuates fatty acid toxicity and improves endothelial-smooth muscle coupling and cardiac mechanical efficiency. The overall goal of this proposal is to investigate L-carnitine as a novel adjunctive treatment of septic shock; specifically to test the clinical efficacy of L-carnitine on reducing organ failure, a critical patient oriented outcome, and provide evidence as to whether stimulation of carbohydrate oxidation creates host resilience to severe sepsis.
Clinical Efficacy Hypothesis 1: Early adjunctive L-carnitine administration in vasopressor dependent septic shock will significantly reduce cumulative organ failure at 48 hours with an associated decrease in 28-day mortality suggesting the need for further phase III study.

Specific Aim 1: Test if intravenous L-carnitine reduces cumulative organ failure in septic
shock.
SA1a. Using an adaptive trial design, randomize up to 250 eligible patients with septic shock to receive low (6g), medium (12g) or high (18g) dose intravenous L-carnitine or placebo for 12 hours.
SA1b. Measure the first efficacy endpoint of reduction in cumulative organ failure, defined as a decrease in the sequential organ failure assessment (SOFA) score at 48 hours after treatment, and use the SOFA data to preferentially allocate subsequent patients to the L-carnitine dose that is most effective.
SA1c. Measure the second efficacy endpoint of mortality at 28 days, and determine in an ongoing manner as the trial progresses, the probability that the dose of L-carnitine associated with the largest decrease in SOFA score would demonstrate efficacy in a subsequent phase III, randomized, placebo-controlled trial.
Rationale: L-carnitine enhances lactate oxidation, which enhances cardiac inotropy and vascular smooth muscle tone, thus enhancing vital organ perfusion.
Expected results: Significantly more patients treated with the optimal dose of L-carnitine will demonstrate a decrease SOFA score as compared to placebo and this dose will have a substantial probability of success in a follow up phase III trial with a mortality endpoint.
Clinical Mechanism Hypothesis 2: L-carnitine will produce a significant improvement in microcirculatory blood flow in patients with vasopressor dependent septic shock. Rationale: L-carnitine should attenuate or reverse vasoplegia, one of the main causes of microcirculatory insufficiency in septic shock. The ΔMFI will provide insight into the clinical mechanism of L-carnitine benefit. Expected results: L-carnitine patients will have a significant improvement in MFI compared to placebo.

BACKGROUND AND SIGNIFICANCE
Sepsis is a significant public health problem. Sepsis is common. It is estimated that severe sepsis occurs at an incidence rate of 3 cases per 1,000 persons/year, resulting in approximately 750,000 affected persons annually in the U.S. Of those affected, 500,000 (67%) require intensive care unit (ICU) services. 1 Sepsis is growing. Recent estimates indicate that the rate of severe sepsis hospitalizations doubled during the last decade, and its prevalence exceeds that of other high profile diseases in the U.S. During 2002, for example, the number of cases per 100,000 population of severe sepsis far exceeded the number of cases of acquired immunodeficiency syndrome (AIDS), venous thromboembolism (VTE) and both colon and lung cancer. [1][2][3][4] Furthermore, the incidence of severe sepsis is projected to increase by 1.5% per annum, resulting in more than 1 million cases per year by 2020. 1 Sepsis kills. Sepsis ranks as the tenth leading cause of death in the U.S., resulting in 215,000 deaths annually. 1;5 In the U.S., with a hospital mortality rate of 30%, mortality rates due to sepsis are strikingly higher than other high profile diseases such as stroke, myocardial infarction, and trauma. 1;6-8 Moreover, the highest mortality rate from sepsis occurs among our nation's elderly, those who are 65 years or older, and age-adjusted population based mortality is increasing. 1 Carnitine is a naturally occurring nutrient required in mammalian energy metabolism. It has been shown to facilitate long-chain fatty acid entry into cellular mitochondria, therefore delivering a substrate for oxidation and subsequent energy production. L-carnitine refers to the levo isomer of the compound, which is biological active. It is obtained primarily from dietary intake of protein, though it can be synthesized in the liver and kidneys from the amino acids lysine and methionine. Although it is present throughout the body in mitochondria, L-carnitine is concentrated in muscle tissue, including the heart, skeletal and smooth muscle, as well as the endothelial cell layer.
Healthy heart muscle uses a combination of free fatty acids (FFA), glucose, and lactate for energy, with FFAs providing 50-70% of the energy for the myocardium. 10;11 As depicted in Figure 1, FFAs must be esterified to Lcarnitine to be capable of crossing the mitochondrial membranes via the tandem action of L-carnitine palmatoyl transferases I and II (CPT-I, II). 12 Key to the study hypothesis, L-carnitine also modulates activity of the  phosphorylation of its active site by a PDH kinase, which is stimulated by an increased acetyl-CoA/CoA ratio. The E1 is activated by a Ca ++ -and insulin-stimulated phosphatase that dephosphorylates serine and threonine residues in its active site. L-carnitine accelerates pyruvate oxidation (and the forward flow of lactate disposal) by esterifying and exporting acetyl-CoA out of the mitochondrial matrix, thereby decreasing the acetyl-CoA/CoA ratio. 39;13 Through this mechanism Lcarnitine infusion has been demonstrated not only to improve FFA but also carbohydrate metabolism. 14,15 These roles as an essential component at rate-limiting steps of combustion of multiple fuel substrates make L-carnitine an attractive therapeutic target for diseases that feature derangements of metabolism.
During stress conditions, such as sepsis, the myocardium drastically alters its primary source of energy from FFAs to lactate, as illustrated in Figure 2. 10;11 Septic shock also induces profound inefficiencies in myocardial work not explained by cellular or mitochondrial hypoxia. 16;17 Although the mechanisms of this inefficiency are not completely understood, research has shown that the cardiac isozyme of CPT-I (the key mediator of FFA transport) is significantly down-regulated in sepsis, contributing to decreased FFA metabolism. [18][19][20][21] Simultaneously, excessive acetyl-CoA production stimulated by sepsis alters the acetyl-CoA/CoA ratio and inhibits PDH. 13;22 This imbalance leads to decoupling of the stoichiometry between glycolysis and pyruvate oxidation. Consequently, the cytosol accumulates lactate, FFA metabolites, and intracellular protons, which together exert a tremendous intracellular cost in terms of chemical energy required to remove them. 15;23 Inhibition of both of these enzymes in sepsis decreases the primary supplies of energy to the failing heart, and likely plays a role in sepsisinduced cardiac dysfunction.
During septic shock the cardiovascular system experiences a net loss of tissue L-carnitine, both in the myocardium 24 and endothelium. 25 L-carnitine is redistributed from the tissue to the peripheral blood, resulting in transiently elevated plasma L-carnitine levels. 26 The L-carnitine resorption mechanism in the kidney, which demonstrates saturation kinetics 27 and is highly efficient (90-99%) in healthy individuals, is overwhelmed by the increased plasma levels of L-carnitine in sepsis, resulting in increased urinary excretion, which in turn ultimately leads to whole body L-carnitine loss. 28;29 Importantly, L-carnitine deficiency has been demonstrated to result in impaired cardiovascular function in both septic and non-septic conditions. 39;30;31 Exogenous infusion of L-carnitine effectively increases L-carnitine levels in depleted tissues in animal models of sepsis 32;33 and enhances cardiac contractility in depressed hearts primarily by stimulating pyruvate oxidation, recoupling glycolysis and oxidation of pyruvate, leading to more efficient heart muscle contraction in shock. [34][35][36] Improvement in carbohydrate and lactate metabolism, which are more efficient than FFAs in energy production, is a primary reason for L-carnitine's ability to improve cardiac efficiency in states of stress. 23 Therefore, as opposed to vasoactive agents, which increase the amount of cardiac work,

L-carnitine functions as a metabolic therapy, improving the efficiency of cardiac work.
Like the rest of the cardiovascular system, the vascular endothelium utilizes L-carnitine-dependent FFA oxidation. 25;37 Elevated levels of serum FFA, a common finding in sepsis, disrupts the chemical signaling between the endothelial cell layer and smooth muscle cells, contributing to vasoplegia. 38;39 The endothelium appears to be particularly vulnerable to the intermittent ischemia associated with altered microcirculatory flow, and leads to preferential loss of low molecular weight compounds, including L-carnitine. 25;40 Animals deficient in L-carnitine demonstrate exaggerated vasoplegic response to a LPS challenge. 30 Infusion of L-carnitine can lead to restoration of flow regulation after periods of intermittent ischemia, an effect that may be related to changes in endothelial permeability. 37;41 Additionally, L-carnitine has been shown to improve endothelial function through and antioxidant effect of decreasing superoxide anion production and increasing nitric oxide participation. 42 As microcirculatory dysfunction is a key mechanism in the evolution of septic shock, and L-carnitine improves endothelial function, directly imaging the microcirculation via intra-vital video-microscope technology will provide us an opportunity to explore a key mechanism of a potential therapy for sepsis-induced organ failure.
Preliminary human data suggests that L-carnitine infusion in septic shock improves patient oriented outcomes. One previous study examined the effect of L-carnitine infusion on patients with septic shock. 43  systolic and mean arterial pressures, lower right atrial pressure, and higher arterial partial pressure of oxygen and hemoglobin oxygen saturation at the end of the L-carnitine infusion as compared to placebo treated patients. These results suggest the potential for L-carnitine to improve dysfunction of multiple organ systems in septic shock and provide an important basis for the use of change in cumulative organ failure as a primary outcome measure.
Microcirculatory dysfunction is a pivotal element in the pathogenesis of severe sepsis and septic shock. Microcirculatory dysfunction is a pivotal element in the pathogenesis of severe sepsis and septic shock. [44][45][46][47] Using intra-vital video-microscopy, experimental models of sepsis have demonstrated decreased microcirculatory flow velocity, "stopped-flow" microvessels, increased heterogeneity of regional perfusion, and decreased density of perfused capillaries. 48-50;50;51 These derangements can cause marked alterations of oxygen transport including impaired tissue oxygen delivery and consumption leading to anaerobic cellular metabolism. 52 With the advent of new imaging modalities such as sidestream dark field (SDF) video-microscopy, it is now possible to visualize the microcirculatory network in human subjects. Microcirculatory "failure" is one of the critical pathogenic events in sepsis that is associated with acute multi-organ dysfunction and mortality. [44][45][46] Although septic shock research is classically focused on macrocirculatory hemodynamics that reflect the distribution of blood flow globally throughout the body, a functioning microcirculation is another critical component of the cardiovascular system that is essential for effective oxygen delivery to tissues.
The SDF microcirculatory imaging technique utilizes optical filtration of polarized light that is absorbed by hemoglobin so that red blood cells appear dark. This permits direct visualization of blood flow in the sublingual microcirculatory network in human subjects in a non-invasive fashion using a hand-held video-microscope (Figure 3). The technique has been validated in both experimental and human studies. [53][54][55] Several investigators have demonstrated that impaired   58;59 In the proposed randomized controlled trial, we will use SDF videomicroscopy to determine whether a novel metabolic therapy for sepsis can augment microcirculatory flow, potential providing a mechanistic investigation of any perceived clinical effect.

Organ failure causes mortality and is a critical patient oriented research outcome measure.
Acute multi-organ dysfunction is a critical event in the pathogenesis of sepsis, and is closely linked with mortality. 1;64 Early evidence of organ failure is an especially strong mortality predictor. 64 Although the use of surrogate outcomes is standard in phase II studies, we recognize their lesser clinical significance compared with a mortality endpoint. As has been noted in many previous sepsis trials, positive findings using surrogate outcomes in phase II studies have failed to translate into 12/P Version 1.3.6 Dated: 07/26/2016 meaningful outcomes (i.e. mortality rates). The SOFA score represents a strong independent predictor of 28-day mortality. As such it will be used to preferentially allocate patients to the Lcarnitine dose that is most effective during the trial. The definition of success at the conclusion of the currently proposed phase II trial requires BOTH demonstration of improvement in SOFA score in the optimal dose intervention subjects AND the predicted probability of success of >30% in a subsequent phase III trial of the optimal dose versus placebo investigating 28-day mortality.
Using this methodology, the optimal dose will be identified and only carried to phase III if the results are promising.

A. Trial Design:
This trial will be a randomized, double-blind, placebo-controlled, dose/efficacy-finding trial that utilizes a Bayesian adaptive approach, incorporating a flexible dose-response model, frequent interim analyses to allow the efficient identification of a clinically-important efficacy or trial futility, and a response-adaptive randomization to increase the scientific efficiency of the trial.

B. Trial Sites:
University

C. Study Setting and Population:
Patients will be identified in the ED or ICU. Potential participants will be identified as adult patients who are treated with a quantitative resuscitation protocol for septic shock initiated in the ED or ICU within 24 hours of enrollment. The following inclusion and exclusion criteria will apply: F. Treatment Arms: The trial will include one control and three active treatment arms. The control (placebo) arm is a normal saline solution. The treatment arms are L-carnitine in three different doses: low (6 grams), medium (12 grams) and high (18 grams), each administered intravenously over a 12-hour period.
Neither the investigators nor the research participants will know to which arm the participant has been assigned. The pharmacist who mixes the treatment solution will not be blinded; however, this individual will have no study-related contact with the investigators or the study participants.
Selection bias should not be an issue given that the investigator is blinded from group assignments and from the randomization procedure itself. The research pharmacists and staff are the only study personnel who have access to the group assignment given to each subject. To ensure blinding, the pharmacist will prepare either L-carnitine or placebo in identical polypropylene infusion bags and label the bags with identical labels that will include the patient's study ID number, patient name, medical record number, and infusion rate. For each dose of L-carnitine, 33% of the total dose will be administered as a bolus over 2-3 minutes followed by a fixed rate continuous infusion over the next 12 hours. For example in the medium dose (12 gram) arm the patient would receive a 4 gram bolus injection over 2-3 minutes followed by 8 gram infusion (8 grams in 1000 ml of 0.9% normal saline) over the following 12 hours (83 ml/hr). The placebo groups will have the identical number and appearance of polypropylene infusion bags filled with normal saline and run at the same rates as the active treatment arms. The bags will not otherwise be identified. The study solution will be transported to the patient's care area where it will be administered by the patient's clinical care nurse and supervised by the site PI. The study solution will be administered via an existing central venous catheter, placed as a part of each institution's early quantitative resuscitation protocol, and will be administered using FDA-approved medical equipment (IV tubing, IV pumps, etc.) that is routinely available at study sites. Sigma-Tau Pharmaceuticals, Inc. will provide the study drug, which will be maintained by the pharmacy staff. A standard operating procedure will be used to track lot numbers of L-carnitine administered. The pharmacy will keep a drug accountability log.

G. Risk/Benefit Ratio:
Patients eligible for enrollment have at least a 40%-50% probability of experiencing death in the hospital and as much as a 25% chance of experiencing a serious side effect from vasopressors (supraventricular or ventricular arrhythmias, acute coronary event, limb or penile necrosis, stroke, CNS bleeding), representing substantial risk of morbidity or mortality from the disease process. L-carnitine has a very favorable safety profile, with no known significant or serious side effects. All adverse effects that have previously been reported are mild such as abdominal cramping and body odor or moderate such as increase seizure risk. If L-carnitine infusion is able to reduce the amount of organ dysfunction, this might directly result in improved chance for survival and decreased hospital length of stay and its potential complications.

H. Study Procedures:
1) Screening and Enrollment: All of the trial sites have an established protocolized quantitative resuscitation strategy for septic shock. Site PIs will assure that research personnel are always available to cover this study 24 hours a day, 7 days a week. The inclusion criteria require enrollment within 24 hours of identification and initiation of quantitative resuscitation for septic shock and the use of high dose vasopressors for 4 or more hours. These criteria allow a reasonable window for enrollment but still satisfy the aim of the study of early administration of L-carnitine therapy. A timeline for study enrollment is shown in Figure 5. Eligible patients (or their surrogate in the case of mental incapacitation) will be approached for consent.  be required at 5 time points. Using an SOP, plasma will be collected, centrifuged and stored. All stored will be batched and shipped to the coordinating center for storage.
Genomic variations: As genomic variations in the carnitine transporter may affect the results of our study, we will collect a portion of the 45 cc of blood drawn at the T0 time point outlined above in a Paxgene tube to allow DNA and RNA recovery. This will allow for subsequent analysis of sequence variations, expression profiles, and allow for correlation to observed phenotypes, including responders versus non-responders. Images will be obtained at T0 (up to 4 hours prior to carnitine administration) and a second time, at T12 (± 4 hours) after carnitine administration is initiated. The video microscope is non-invasive and is not harmful in any way other than some potential mild discomfort in the mouth.

3) Endpoints and Data Collection:
The following data for determination of endpoints will be collected. Additional data to be collected can be found on the complete enrollment collection templates and case report forms.

Primary Clinical Endpoints
Reduction in organ failure, defined as a decrease in sequential organ failure assessment (SOFA) score from T0 to T48. Survival by severity of illness measure at enrollment (e.g., APACHE II score)

Mechanistic Endpoints
Change in microcirculatory flow index from T0 to T12.

4) Adverse Event Reporting:
Safety assessments will consist of monitoring and reporting adverse events (AEs) and serious adverse events (SAEs) that occur, their relatedness to Lcarnitine, all events of death, and any study specific issue of concern.
Adverse Events An AE is any unfavorable and unintended sign, symptom, or disease temporally associated with the use of L-carnitine in this study. The study-specific definition of "temporally associated" includes up to seven days after discontinuation of L-carnitine infusion.
This includes the following: AEs not previously observed in the subject that emerge during the protocol-specified AE reporting period.
Preexisting medical conditions (other than the condition being studied) judged by the investigator to have worsened in severity or frequency or changed in character during the protocol-specified AE reporting period. Abnormal laboratory values that fall into an abnormal range based upon the hospital's laboratory standards, the abnormality was not preexisting prior to enrollment, and the abnormality leads to a new treatment or ongoing treatment within the AE time frame.

Serious Adverse Events
An AE should be classified as an SAE if: It results in death (i.e., the AE actually causes or leads to death).
It is life threatening (i.e., the AE, in the view of the investigator, places the subject at immediate risk of death. It does not include an AE that, had it occurred in a more severe form, might have caused death).

It requires or prolongs inpatient hospitalization.
It results in persistent or significant disability/incapacity (i.e., the AE results in substantial disruption of the subject's ability to conduct normal life functions).
It results in a congenital anomaly/birth defect in a neonate/infant born to a mother exposed to the investigational product.
It is considered a significant medical event by the investigator based on medical judgment (e.g., may jeopardize the subject or may require medical/surgical intervention to prevent one of the outcomes listed above).

Methods and Timing for Assessing and Recording Safety Variables
The investigator is responsible for ensuring that all AEs and SAEs that are observed or reported during the study are collected and reported to the FDA and appropriate IRB(s) in accordance with CFR 312.32 (IND Safety Reports).

Adverse Event Reporting Period
The reporting period for all study-defined adverse events will begin at the time of L-carnitine infusion and for 7 days after infusion. Patients will be followed for SAEs until hospital discharge and/or the 28-day follow up phone call. Every reasonable attempt will be made to follow laboratory abnormalities until return to normalization or baseline.

Assessment of Adverse Events
All AEs whether volunteered by the subject, discovered by study personnel during questioning, or detected through physical examination, laboratory test, or other means will be reported up until 7 days after L-carnitine infusion completion. Abnormal laboratory values will be considered an AE if the value falls into an abnormal range based upon the hospital's laboratory standards, the abnormality was not preexisting prior to enrollment, and the abnormality leads to a new treatment or ongoing treatment within the AE time frame.
Each reported SAE will be described by its duration (i.e., start and end dates), regulatory seriousness criteria if applicable, suspected relationship to L-carnitine and actions taken.
To ensure consistency of SAE causality assessments, investigators will apply the following general guideline: Yes -There is a plausible temporal relationship between the onset of the AE and administration of the L-carnitine and the AE cannot be readily explained by the subject's clinical state, intercurrent illness, or concomitant therapies; and/or the AE follows a known pattern of response to L-carnitine or the AE abates or resolves upon discontinuation of L-carnitine.
No -Evidence exists that the AE has an etiology other than the L-carnitine (e.g., preexisting medical condition, underlying disease, intercurrent illness, or concomitant medication); and/or the AE has no plausible temporal relationship to L-carnitine.

AEs or SAEs. Avoid colloquialisms and abbreviations.Diagnosis vs. Signs
and Symptoms: If known at the time of reporting, a diagnosis will be reported rather than individual signs and symptoms. However, if a constellation of signs and/or symptoms cannot be medically characterized as a single diagnosis or syndrome at the time of reporting, it will be reported based upon the information that is currently available. If a diagnosis is subsequently established, it will be reported as follow-up information.
Deaths: All deaths that occur during the SAE reporting period, regardless of attribution, will be reported to the appropriate parties. When recording a death, the event or condition that caused or contributed to the fatal outcome will be reported as the single medical concept. If the cause of death is unknown and cannot be ascertained at the time of reporting, report "Unexplained Death".
Preexisting Medical Conditions: A preexisting medical condition is one that is present at the start of the study. Such conditions will be recorded as medical and surgical history, where the case report form allows. A preexisting medical condition will be re-assessed throughout the trial and reported as an AE or SAE only if the frequency, severity, or character of the condition worsens during the study. When reporting such events, it is important to convey the concept that the preexisting condition has changed by including applicable descriptors (e.g., "more frequent headaches").
Hospitalizations for Medical or Surgical Procedures: Any AE that results in prolonged hospitalization will be documented and reported as an SAE. If a subject is hospitalized to undergo a medical or surgical procedure as a result of an AE, the event responsible for the procedure, not the procedure itself, will be reported as the SAE. Hospitalizations for the following reasons do not require reporting: Prolonged hospitalization for diagnostic or elective surgical procedures for preexisting conditions, or hospitalization or prolonged hospitalization for scheduled therapy of the target disease of the study. Post-Study Adverse Events: The investigator will expeditiously report any SAE occurring after a subject has completed or discontinued study participation if attributed to prior L-carnitine. If the investigator becomes aware of the development of cancer or a congenital anomaly in a subsequently conceived offspring of a female subject who participated in the study, this will be reported as an SAE.
SAE Reporting: Investigators will report all SAEs to the local IRB, data coordinating center, and the FDA (as applicable).

TRIAL STRATEGY AND STATISTICAL ANALYSIS
We propose an adaptive, phase II trial to efficiently identify the dose of L-carnitine that provides the greatest improvement in SOFA score and, simultaneously, to assess the probability of success if that dose of L-carnitine was evaluated in a confirmatory, phase III trial with a 28-day mortality endpoint.
To ensure we identify the best dose, a broad range will be considered initially (6 g, 12 g, and 18 g).
Because L-carnitine is likely to have its most direct effect on SOFA score, the change in SOFA score at time 48 hour post treatment will be used to guide the preferential randomized allocation of subjects to the doses that are the most promising. This approach allows the initial consideration of a wide dose range, but avoids the inefficiency of balanced randomization which would continue to allocate patients to non-promising doses even late in the trial. Because our overall goal is to develop a treatment that decreases the 28-day mortality associated with septic shock, the criteria for assessing the futility or success of the phase II trial will be based on the 28-day mortality benefit seen with the most-promising L-carnitine dose. Specifically, the trial will utilize frequent interim analyses, with ongoing predictive-probability-based assessments of both futility and success, so the trial may be stopped as soon as either: (1) a dose of L-carnitine has been identified that is sufficiently promising to warrant investigation in a phase III trial; or (2) it is clear that none of the three doses is sufficiently effective to warrant further investigation. The trial will enroll a maximum of 250 subjects allocated to four treatment arms (control, 6 g, 12 g, 18 g). The adaptive trial will utilize both the change in SOFA score at time 48 hour post treatment and 28 day mortality as efficacy endpoints. The allocation of subjects will be determined only by the observed change in SOFA score; however, interpretation of the trial results will be based on both the change in SOFA score and 28-day mortality. For determining the probability of success in a phase III trial, we assume the phase III trial would utilize standard, frequentist approaches and enroll up to a maximum of 2,000 subjects, with half in a control arm and half receiving the selected dose of Lcarnitine. A phase III trial of that size would have a power of 95% to detect a reduction of 28-day mortality from 40% to 32% or a power of 78% to detect a reduction from 40% to 34% with L-carnitine.
A normal dynamic linear model (NDLM) is utilized to improve the efficiency in estimating the dose response across adjacent dose levels. During an initial "burn in" period of 40 patients, subjects are allocated equally among the treatment arms. From that point on, an interim analysis is conducted after every 12 subjects and the randomization proportions among the arms are adjusted so that the probability of assignment to each active treatment arm is proportional to the probability that that arm leads to the greatest improvement in SOFA score. A blocked randomization approach is used to ensure that approximately one-third of subjects are allocated to the control arm throughout the trial.
This helps to ensure that the result is resistant to confounding due to secular trends in outcome.
The criteria for stopping the trial at each of the interim analyses are based on Bayesian posterior probability calculations. Specifically, the trial is stopped for futility if there is less than a 40% probability that the most promising L-carnitine dose leads to an improvement in SOFA at time 48 hour post treatment. The trial may also be stopped early for success, if there is a greater than 90% probability that the most promising dose of L-carnitine improves SOFA at time 48 hour post treatment and the probability of success in a subsequent phase III trial is greater than 70%. These criteria for defining futility and success at the interim analyses were selected empirically to yield desirable operating characteristics (see Trial Operating Characteristics section).
To estimate the true relationship between the change in SOFA score and 28-day mortality, SOFA score and mortality data from more than 250 subjects with inclusion criteria identical to the proposed trial were obtained from a multicenter observational study. All patients in the source study were treated according to the Surviving Sepsis Campaign Guidelines for the management of septic shock.
The observed odds ratio for 28-day mortality, for each decrease in SOFA of one point at 48  we simulated thousands of subjects and their corresponding SOFA data. Each simulated subjects' SOFA result was used to determine his or her predicted 28-day mortality, and then a 28-day mortality outcome was randomly assigned with that probability. All trial design work and simulations were completed using the Fixed and Adaptive Clinical Trial Software (FACTS) from Berry Consultants, LLC. Moreover, it can be seen that, under this scenario, the trial efficiently allocates to patients to both the control and highest dose arms, maximizing the information gained per subject (Figure 6). Table 1 also demonstrates the performance if there is a very strong treatment effect, in which case the power of the trial is even higher and the required sample size smaller.
At the end of the trial, the trial is considered negative if either there is less than a 90% probability that the most promising dose of L-carnitine improves SOFA at 48 hours post treatment or less than a 30% probability that the most promising dose of L-carnitine would be successful in a subsequent phase III trial. The trial is defined as positive if there is a greater than 90% chance that the selected dose of Lcarnitine improves SOFA at 48 hours post treatment and a greater than 30% chance of success in a subsequent phase III trial. These criteria for defining futility and success at the interim analyses were selected empirically to yield desirable operating characteristics. Collaboration with the Center for Biostatistics and Berry Consultants as required for the adaptive trial design to communicate blinded interim data for updating centralized randomization schedule and DSMB reports; Staffing for data project management, database administration and data entry training; Electronic case report form design and data entry resources; Secure online data entry creation and maintenance; Design, implementation and maintenance of centralized, web-based, randomization; Data queries and custom report generation.
The Biomedical Informatics Group is currently involved in designing, building, and implementing an enterprise data warehouse for research and quality to provide infrastructure to support research informatics. Epic will be deployed as the electronic medical record at UMMC hospitals and clinics on April 1, 2012. Epic will be a major source of data for the enterprise data warehouse (EDW). The EDW will also house data from approximately thirty legacy systems that will be replaced by Epic and sunsetted. Research databases and datasets will also reside on the EDW. The EDW will be governed by an executive governance committee and several focused working groups.
The Biomedical Informatics group has a close working relationship with the Center for Biostatistics. All data management services, including reports to the Center for Biostatistics for safety monitoring, will be produced with assurance of shielding the clinical management arm from access to unblinded data during performance of the trial. The data management center will also produce performance reports of the sites, which will be communicated to the clinical management center and clinical trial monitor. At the end of the trial, the data management center will clean the data and produce a final locked data set for the PI and participate in the final data analysis and manuscript development.
The data management center will inevitably also handle sensitive data, including information used to monitor the performance and enrollment at each site, and other data as dictated by the needs of the network. Thus the following security procedures will be used.
Standard processes are established to track the flow and quality control of case report forms and data. Data collection will be directly into the electronic database from source documents at the clinical sites. Data entry will be done via REDCap. Quality control will be done during and post-entry. Discrepancies identified through data review processes will result in query generation sent to clinical sites for resolution. Manual data review may also result in manual queries. Any query will be sent to sites via a data change form (DCF) signed by PI and Coordinator. Changes will then be made to REDCap. The DCF will then be signed-off by the site PI after completion.
Additionally, the Biomedical Informatics group will develop the data management plan to include (but not limited to) the following: o General CRF completion rules; o Data entry and processing plans; o QC -completeness of dates / data ranges for numeric variables / data logic / self-evident corrections; o Data receipt procedures; o Conduct web training to sites for CRF completion and provide CRF completion guidelines.
All data are managed on four virtual Linux servers. These servers are managed by a full time database administrator, and are user-id and password protected. These machines are members of a domain server system that are also user id and password protected. Equipment is housed in a building with a security alarm system, which is also patrolled at night by security regarding corrective action plans are maintained. Institutional policies also provide guidance for appropriate corrective action. In the unlikely event of a security breach, a formal written report will be presented to the Study Director and the IRB within 7 days of discovery. The data management center project manager will then work with the Study Director, the Steering Committee, and the IRB to implement an appropriate corrective plan.

Site Monitoring and Quality Assurance
The protection of human subjects is of paramount importance, and the Coordinating Center (CC) will coordinate the human subjects protection plan, data safety, and monitoring the project.
An individual from the CC who is qualified to perform the role of monitor will inspect records of the study periodically. The study monitor will be given access to all source records and data needed to verify the entries on the case report forms (CRFs) to original entries in source files.
In addition to the original medical record, the study will employ data collection templates as a source document. These forms contain the data fields that historically have been difficult to abstract from the medical record. Only site personnel designated on a "Delegation of Authority Log" will be authorized to enter data onto the CRF. The investigators agree to cooperate with the study monitor to ensure that any problems detected in the course of monitoring visits are resolved. The monitor will inspect the CRF entries on a predetermined percentage (up to 30%) of subject records to assess the completeness, correctness and accuracy of CRF entries and to verify adherence to the protocol and applicable regulations. The project research coordinators at the sites will maintain binders containing pertinent documents and information including IRB approval letters, documentation of GCP training, licenses, human subjects research ethics training certificates, curriculum vitae, and correspondence.
Sites will be monitored to ensure their compliance with local IRB approvals and reporting requirements, both prior to initial certification and as part of ongoing performance assessment and quality improvement. All operations will be immediately suspended at any site where ongoing review or approval lapses.
Adverse events will be reported to the CC. The CC will disseminate the information back to both sites for reporting to local IRBs, the sponsor, FDA when necessary, and the data safety monitoring board. This centralized reporting will ensure that unscheduled meetings of the DSMB are called when needed, that there is adherence to previously defined stopping rules, that events are handled in a consistent and timely manner, and that centralized records are available for review at all times.
All serious, study drug-related adverse events will be reported to the CC. SAEs will be reported to the CC by online SAE reporting form that when completed will generate an email to the project manager and Study Director at UMMC. Fatal or life threatening events categorized as "probably related" will be immediately followed up by the Study Director by telephone call to the study site submitting the SAE. The Study Director will notify the FDA within 7 days of learning of the event.
The study requires surveillance for SAEs on a daily basis for seven days after enrollment, and then again at hospital discharge. To determine SAEs at hospital discharge we will record all non-exempt (see below) SAEs that are documented in the hospital discharge summary for the index visit. An SAE will be defined in accordance with ICH-GCP guidelines (any unexpected illness or injury that poses a threat to life or body function and causes medical or surgical intervention or prolongation of hospitalization). Electronic forms will record an estimate of the relationship to the treatment as assessed by the site PI. The study associate will require an email reply from the project manager at the CC that indicates receipt of the SAE. The project manager will disseminate the SAE to both study sites.

DSMB Plan
The data safety monitoring board (DSMB) will follow a charter approved by the sponsor. The DSMB will follow the adaptive rules as outlined in the application unless there is a compelling, unanticipated reason to deviate from these rules. The DSMB may recommend stopping the trial early for one of three reasons, following these guidelines: [Trial Futility] Stop the accrual of subjects because there is less than a 40% probability that the most promising L-carnitine dose leads to an improvement in SOFA at time 48 hour post treatment.
[Trial Efficacy] Stop the accrual of new subjects because there is a greater than 90% probability that the most promising dose of L-carnitine improves SOFA at time 48 hour post treatment and the probability of success in a subsequent phase III trial is greater than 70%.
[Patient Safety] A recommendation to suspend or stop for harm would occur if an interim analysis showed strong evidence that the rate of IRB-reported SAEs was significantly higher in the experimental group than in the placebo group. Because the underlying rate of SAEs will be difficult to predict, it is not possible to provide a table of critical values at each interim meeting that could be used for stopping criteria.
The DSMB will operate in accordance with guidelines established by the FDA in "Guidance for

Clinical Trial Sponsors: Establishment and Operation of Clinical Trial Data Monitoring
Committee" jointly published by the CBER, CDER and CDRH of the FDA, OMB#0910-0581.
The key elements include: 1. Expertise and independence: The members must have expertise in the area of study, not be study authors and have no financial conflict of interest.

2.
Members will sign a confidentiality agreement.

3.
It is preferred that the DSMB have a face-to-face initial meeting, but can meet by telephone thereafter more frequently.

4.
All meetings must follow standard operating procedures (SOPs).
Minutes of all meetings will be filed with the Office of Sponsored Studies at the CC and a copy sent to the other relevant site administrative offices and NIH.
If the DSMB recommends closure of the study for any of the reasons above, then the DSMB Chair will email the Study Director and officials at NIH. Officials at NIH would then decide the action to be taken on the trial.
If the trial is stopped for patient safety and a repairable cause of unacceptably high adverse events can be identified, the action must be reviewed and approved by the DSMB and by each site IRB prior to resuming enrollment. For example, if one additional exclusion criterion is discovered that would account for the majority of unexpected events, this new exclusion criterion could be added, and the study could proceed.

Miscellaneous
Every attempt will be made to report adverse events using the most current version of MedDRA.
Subjects who withdraw from study therapy early will be followed in the same manner as other subjects in the study for safety (AEs and SAEs). In that case, other study procedures will not be discontinued unless instructed by the study participant/ LAR.