Progesterone Treatment of Blunt Traumatic Brain Injury

Title: Progesterone Treatment of Blunt Traumatic Brain Injury Principal Investigator(s): Arthur Kellermann, M.D., M.P.H. Institution: Emory University Biostatistician: Vicki Hertzberg, Ph.D. Institution: Emory University Project phase or primary methodology: Phase II pilot, double blind, 4:1 randomized controlled, clinical trial Target disease: Blunt traumatic brain injury Intervention: Intravenous progesterone infused over three days Number of patients to be enrolled: 100 Number of clinical centers: One (Grady Memorial Hospital) Primary Hypothesis: Dose: Intravenous infusion of progesterone at a dose of 0.5 mg/kg/hr produces a steady state of 450 +/- 100 nmol/L in male and female subjects. Safety: Compared to study subjects who receive placebo, subjects who receive IV progesterone do not have an increased risk of mortality or an increased incidence of adverse effects. Efficacy: Administering IV progesterone shortly following TBI produces any or all of the following effects, including: * Reduction of the intracranial pressure therapeutic intensity level (ICP-TIL). * Decreased duration of coma. * Decreased inpatient mortality. * Improved neurological outcome one month post-injury. Secondary Hypotheses: Benefit: The benefits of IV progesterone for TBI are evident across a range of clinically relevant treatment subgroups, regardless of: * Time from injury to drug administration: (i.e., \<2 hrs. vs. 2-4 hrs. vs. 4-6 hrs.). * Victim gender. * Presence or absence of severe extracranial trauma. * Presence or absence of premorbid confounders of outcome (e.g., alcohol intoxication, alcoholism, stroke, etc.). * Incidence or lack of confounding neurological insults (e.g., hypoxemia, hypotension). * Severity of the traumatic brain injury - moderate (index GCS 9-12) versus severe (index GCS 4-8). * Severity of non-brain injuries (AIS, ISS scores). Description of methodology and design: Double blind randomized controlled, clinical trial. Block randomization will be used to insure that moderate and severely injured patients of both genders and races (black and nonblack) have an equal chance of allocation to either treatment group. Patient selection criteria: Age \> 18; blunt traumatic brain injury; moderate to severe brain injury (iGCS 4-12); arrival \< 11 hours after Injury Description of pre-randomization procedures or process: Potential study subjects will be identified by Emergency Medical Service (EMS) personnel and Emergency Medicine attending physicians and residents in the ECC of GMH. EMS (ambulance) personnel will notify ECC personnel that they are bringing a blunt trauma patient with a Glasgow Coma Score (iGCS) of 12 or less. On arrival, the attending emergency physician on duty will conduct a primary and secondary trauma survey and note the patient's post initial resuscitation or i GCS. If the patient has a iGCS of 4 - 12 and meets criteria for presumptive eligibility, the ECC attending will page the investigator on-call (IOC) and ask Hospital Social Services to initiate aggressive efforts to contact the patient's family. To insure prompt administration of progesterone/placebo, a "door to needle time" of ≤ 2 hours has been established as our performance goal. To meet this deadline, a member of the research team will immediately respond to the ECC and determine if the patient meets the inclusion and exclusion criteria. Time from injury will be established as clearly as possible from the EMS report, police report or witness interview. Block randomization will be used to insure that moderate and severely injured patients of both genders and races (black and nonblack) have an equal chance of allocation to either treatment group. To accomplish this goal, the GMH pharmacy will be given block sets of randomly ordered packets with sequential study numbers for subjects with moderate TBI (presenting GCS 9-12) and severe TBI (presenting GCS 4-8). Inside each packet will be instructions to prepare an infusion of progesterone or a comparable volume of placebo. When the pharmacist receives a copy of the randomization request, he/she will open the next packet in sequence based on the patient's severity of brain injury (moderate versus severe), gender, and race. Following treatment group assignment, the pharmacist will record the patient's name, medical record number, dose, and time the medication was prepared in a confidential log for use by the Safety Committee. Precise treatment dose description for protocol: Our goal is to initiate treatment within 2 hours of injury as documented from an EMS report, police report or witness interview. We will accept patients up to 11 hours post-injury. We will attempt to achieve serum levels of progesterone in human subjects that are similar to those normally observed in the third trimester of pregnancy (i.e., 450 +/- 100 nmol/L). To determine the infusion rate needed to achieve a SSSPC of progesterone of 450 +/- 100nmol/L, we will conduct a dose-escalation study. The first 12 subjects randomized to progesterone will receive a loading dose of 0.714 mg/kg over 60 minutes followed by a continuous infusion of 0.50 mg/kg/hr for three days. Because study personnel will be blinded to group assignment, an unblinded pharmacologist will analyze the SSSPC data from these 12 patients. If the target concentration of 450 +/- 100nmol/L is not achieved in 90% of these subjects, the infusion rate will be modified and applied to the next ten subjects randomized to progesterone therapy. This will be performed a third time if necessary. If pharmacokinetic variability precludes the achievement of the target concentration in 90% of subjects using a fixed infusion, we will adopt an individualized dosing strategy based on real-time serum concentration data. If this is required, adjustments will be made to the placebo infusions as well to maintain blinding of the study investigators. The total dose given to any patient over the three day dosing period should not exceed 7.5 grams. Plan for Follow-up: The GOS, DRS, and GOAT will be assessed at one month post-injury. At the time of hospital discharge, each patient's disposition will be noted (morgue, nursing home, rehabilitation facility, another acute care hospital, patient's home, relative's home), so plans can be made for a follow-up exam one month post-injury. Subjects who are still hospitalized at GMH, an inpatient rehabilitation facility, a skilled nursing home, or an acute care hospital other than GMH one month post-injury will be tested on site. We will visit patients who are homebound at the time of follow-up. All other subjects will be asked to come to a study clinic at GMH for testing. Patients who are too severely impaired to test at one month post-injury will be classified as "not testable". This category will be considered a surrogate marker for a poor outcome. We will utilize reliability codes to record all possible reasons for the non-administration of a particular measure, such as physical impairment (e.g., hemiparesis) cognitive impairment (e.g., did not understand instructions), or intoxication. During the follow-up visit, the study subject (or a proxy respondent) will be asked questions about their physical and mental health, any medical complications (e.g., seizure disorder, and headaches), their functional status, their occupational status, their level of sexual function, and their current living situation. Information on other variables that influence outcome, such as access to rehabilitation services, will be collected as well. To encourage participation, subjects will be reimbursed $25 for the follow-up visit, plus additional money for travel-related expenses. Extent and type of blinding/masking: Hospital pharmacy personnel will know each patient's group assignment, but this information will be withheld from the patient, the study team, and the clinical team responsible for the patient's care. Both study drug and placebo will be suspended in a lipid base, and will be indistinguishable by color or injection characteristics. Endpoints and outcomes: Four measures will be used to monitor each patient's clinical response to treatment. The first three will be tracked daily by our study nurses. The fourth will be assessed one month post-injury. These four measures are: 1. Reduction in ICP, reflected by a decline in the patient's therapeutic intensity level (ICP-TIL) a measure of the intensity of treatment needed to reduce increased intracranial pressure; 2. Duration of coma (i.e., total hours from injury to awakening); 3. Death prior to hospital discharge; 4. Neurological outcome at one month post -injury, as measured by the Glasgow Outcome Scale (GOS), the Disability Rating Scale (DRS), and the Galveston Orientation and Amnesia Test (GOAT). Statistical design and sample size calculations: Our data analysis strategy is designed to accomplish our three primary and one secondary goals: 1) achieve and maintain a predetermined steady state concentration of progesterone in 90% of our study patients, 2) confirm the safety of the study drug, 3) test the hypothesis that administration of progesterone within 6 hours of TBI improves clinical and neurological outcomes, and the secondary; determine if the benefit of IV progesterone is equally evident across a range of clinically relevant treatment subgroups. In the process of achieving these goals, we will quantify our ability to identify and enroll TBI patients, administer the study drug within strict time restraints, and adhere to the study protocol. To assess the efficacy of progesterone for treatment of TBI, we will employ a systematic statistical analysis strategy. Our first analytic step will be to compile descriptive statistics (e.g., mean, median, standard deviation, range, proportion) of patient demographics. This activity, while not directly addressing any of the study hypotheses, will serve to describe the central tendencies and variability of the outcome variables and covariables. The second step will be to compare treatment groups with respect to suspected important covariates, specifically patient age, gender, cause and type of brain injury, presence or absence of intracranial mass lesions, presence/absence of extracranial trauma, presence/absence of premorbid confounders (e.g., alcohol intoxication, alcoholism, stroke, etc.), presence/absence of confounding neurological insults (e.g., hypoxemia, hypotension), severity of TBI, and time to initial drug administration. This will assess the balance achieved through randomization. The third step will be to analyze the data according to our specific aims. To estimate the sample size needed to establish efficacy, we needed to specify the effect size to be detected by our outcome measures. We do not have any human data to definitively establish the effect size expected as a result of treatment. Based on animal models, we conservatively posit that progesterone will improve neurological outcome (i.e., dichotimized GOS, DRS, GOAT at one month post-injury) by 25%. We expect to enroll 100 subjects at a 4:1 randomization (80 treated and 20 controls) into our pilot study. With respect to duration of coma, the number of patients that would allow us to achieve 80% power to detect a difference of at least 0.6 standard deviation between treatment groups at the two-sided 0.05 significance level was 50 treated and 50 controls. For this outcome, a difference of 0.6 standard deviation units translates into a difference of + 1.1 days, or 29%, based on the findings of Stambrook, et al.6 However, since we have changed our protocol to a 4:1 randomization, we no longer have the power to detect this difference. Similarly, in the subgroup of patients monitored for differences in the ICP-TIL, in the original 1:1 randomization scheme we would have been able to achieve 80% power to detect a difference of at least 1.0 standard deviation between treatment groups at the two-sided 0.05 significance level, but with the 4:1 scheme no longer have the power to detect a difference. With respect to mortality, our pilot will have insufficient power (\<50%) to definitively detect an effect size equal to 25% reduction in the rate of death. Similarly there is insufficient power to detect a "good" neurological outcome at hospital discharge or at one month post-injury using a dichotomized GOS (good or moderate recovery versus severe disability), with the expectation of 50% with good outcome in the placebo group, based on the findings of Levin, et al. However, this study should allow us to perceive trends in the data as well as determine the range and prevalence of outcomes that we might anticipate for a multi-center study of this therapy. This will allow us to refine our future power calculations appropriately. Initially, there will be a dose escalation component within the progesterone group, but the differences in the level of circulating progesterone between the subgroups will be minimal compared to the control group. Hence, all progesterone treated patients are assumed to respond similarly to the treatment. (2) At the completion of the pilot phase at least 100 patients will be randomized into this study. GMH's trauma registry indicates that at least 145 potentially eligible TBI patients (age ≥ 18, GCS 4-12, blunt mechanism, and \<11 hours from injury) will come to GMH during the study interval. Every effort will be made to obtain informed consent from a proxy in order to administer the study drug within the specified time limit (2 hours). Proposed safety and efficacy monitoring boundaries: Data Safety and Monitoring Board (DSMB): In accordance with standard procedures for NIH sponsored clinical trials, the DSMB chair and members have been selected by the NIH-NINDS. This committee will be informed of each subject's group assignment (A versus B), but will be blinded with respect to which group received the study drug. This committee will have three responsibilities: 1) monitor identification, enrollment, and randomization procedures to detect any evidence of bias; 2) monitor adverse events to determine if they occur disproportionately in one group, and 3) monitor key clinical outcomes to determine if one group does significantly better than the other. To assist the DSMB in its work, our statistician, Vicki Hertzberg, will monitor the progress of the study on a monthly basis. The DSMB will have no role in the day-to-day conduct of the trial. The study statistician will report to the DSMB at quarterly meetings. Safety Monitoring: A study nurse and/or the research program coordinator will round on subjects daily to track their clinical progress and the occurrence of any serious adverse events (SAE) or adverse events (AE). All SAEs will be reported to the IRB, DSMB, and FDA within 24 hours. All other adverse events will be reported to the PI, IRB, and FDA on a weekly basis. Specific events potentially related to the study drug, based on prior knowledge of the pharmacology of progesterone, combination agents that include progesterone, and the Intralipid carrier, will be closely monitored by the study team. These events include: a) phlebitis at the injection site; b) sedation unexplained by administration of other CNS medications; c) an increase in liver enzymes; d) unexplained hyperglycemia; e) hypotension or hypertension not known to be associated with increased intracranial pressure; f) unexplained fever; g) thromboembolic disease; and h) sepsis. Baseline metabolic parameters will be determined from the initial blood draw prior to study drug administration. Blood sampling will be performed at least once per day during drug administration for monitoring. In addition, hourly vital sign measurements and other parameters will be obtained from ICU charting forms. Patients will be examined daily by team members for any signs of adverse reactions. Team members will also review charting, laboratory reports and treating physician notes daily. In addition, all trauma deaths are reviewed by an interdisciplinary Trauma Outcome Review Committee (TORC) to determine if the death was non-preventable, potentially preventable, or preventable. A member of our study team will attend all of these meetings. If/when TORC reviews a death involving a participant in the study, their findings will be shared with the DSMB as well as the P.I. and co-P.I. The DSMB will closely monitor the incidence of death and other serious adverse events such as sepsis or thromboembolic disease. If they have concerns about the incidence of any of these events as linked to the study drug, they will immediately report these concerns to the National Institute for Neurological Disorders and Stroke (NINDS), as well as to the Emory IRB. Furthermore, all occurrences of death (regardless of cause) and unexpected serious adverse events will be reported to the Food and Drug Administration (FDA). Efficacy: To assess the efficacy of the drug on an interim basis, there will be a comparison of the study groups with respect to the endpoints (mortality, duration of coma, ICP-TIL, and neurological status indicators at one month) at the time when 12, 24, and 48 patients have been entered into the trial. Comparisons at these interim points which exceed a z-score of 2.782 (associated p-value of 0.0054) will be considered significant, using the O'Brien-Fleming procedure. To determine if one treatment group does better than the other, the team's study nurses will obtain the following outcome measures: 1) ICP-TIL, 2) duration of coma, 3) in-hospital mortality, 4) Glasgow Outcome Score, 5) Disability Rating Scale, and 6) Galveston Orientation and Amnesia Test. The DSMB will monitor these outcomes by treatment group (A versus B) in order to promptly identify any between-group differences. If significant differences are noted, the code will be broken and our findings will be immediately reported to the Emory IRB, the FDA, the NINDS and DSMB. Patient accrual plan: We will enroll the first 100 eligible patients over a two year enrollment period. Ethical and consent considerations: Since our study involves treatment of patients with acute neurological impairment, they are not competent to give informed consent. We will therefore seek proxy consent from the patient's family or next-of-kin or legally authorized representative (LAR). If an LAR is subsequently located that objects to the study, the process will be halted and the patient's participation will be terminated. Only the IOCs, or the study coordinator, will be permitted to approach family members to request consent. A written form that complies with the polices and procedures of Emory University's IRB has been developed and approved. This form includes the following elements: title of the protocol, the name of the PI, study objectives and purpose, a detailed description of the procedure and interventions, explanation of the responsibilities of the subject and of the family member(s) who act as proxy respondents during follow-up interviews, any and all foreseeable risks, anticipated benefits, available alternatives, an explicit statement of confidentiality, non-compensation for participation, the right to withdraw at any time, a signature section, and a number to contact the PI or a member of the study team with any questions. Given the sudden, unanticipated nature of traumatic brain injury, and the potential time delays involved in getting a family member to the hospital, we have been authorized by the IRB to obtain proxy consent for participation over the telephone. When this is necessary, the on-call investigator will contact the family member, provide a brief explanation of the situation, read the consent form verbatim, and request verbal consent over the telephone. A third party will listen in to confirm that the family member understands and has provided proxy consent. On the signature line of the form, the witness will print the name of the consenting proxy, write their own name immediately below this name, note the time and date of the telephone call, and check a box labeled "telephone consent". Before recruiting the first patient in this study, we will undertake a comprehensive process to notify and consult the target communities that use GMH as their trauma center. The Emory University News and Information office has pledged its support. On the advice of Emory Healthcare marketing, we have allocated $40,000 to support marketing and community outreach activity during the first 6 months of the study. Lesser amounts have been allocated to keep the community informed in years 2 and 3. An independent interdisciplinary Safety Committee has been created to provide additional monitoring for study safety. In addition to designating a Safety Committee, we will convene an independent citizen advisory committee. Two influential Atlantans have agreed to co-chair this group. One is Rev. Gerald Durley, Pastor of Providence Missionary Baptist Church, one of Atlanta's largest African-American congregations. Rev. Durley is also a key member of Concerned Black Clergy, a local activist group. The other co-chair is Sandy Teepen, a TBI survivor, safety advocate, and wife of a national newspaper columnist and former editorial page editor for the Atlanta Constitution, the largest circulation newspaper in the Southeastern United States. Progesterone has been used for many years to treat a variety of medical conditions. Side effects ascribed to its use alone are minimal. Minor adverse reactions include temperature elevations of 1 degree Fahrenheit, mood changes (depression following drug withdrawal), vaginal, cervical and uterine lining changes, irregular vaginal bleeding or amenorrhea, hyperventilation, mild fluid retention, a transient increase in low density lipoprotein (LDL), a transient decrease in high density lipoprotein (HDL), interference with lactation during breast feeding, nausea, abdominal cramping, dizziness, headaches, breast pain, and a delay in fertility. Most of these effects are associated with prolonged administration of progesterone (i.e., weeks to months). They are unlikely to be experienced during short-course administration (i.e., 3 days). Throughout the pilot trial, subjects will be monitored for potential side effects of drug therapy (i.e. hyperglycemia, hypertension unexplained by increases in ICP, unexplained fever, local phlebitis, thromboembolic disease, abnormal vaginal bleeding). The Safety Committee will meet quarterly to review group-specific clinical outcomes and side effects. If they are concerned that one group is doing significantly better (or worse) than the other, they have the authority to break the code and order premature termination of the trial. Participating pharmaceutical or device manufacturing company: None Proposed sponsor or funding agencies: NINDS NIH Significance of the Research and Impact on Medical Care: If the therapeutic benefits observed in animals are replicated in humans, administration of intravenous progesterone should produce several benefits, including: a) reduction of intracerebral pressure; b) decreased duration of coma; c) decreased mortality; and d) improved neurological function at one month post-injury. If these findings are verified, it will represent a major advance in the treatment of traumatic brain injury.