RAMPART and Exception to Informed Consent for Emergency Research

Purpose of this meeting Better Human Subjects Protection Cultivate Cooperation

Objectives Study Synopsis Tenets and Approach Teleological Pitch

RAMPART Synopsis Replace “Presenter's Name” with your name. Alternative title slides are available at the end of this presentation in the “alternatives” slide.

Rapid Anticonvulsant Medication Prior to Arrival Trial (RAMPART) Paramedic treatment of status epilepticus Standard treatment is IV benzodiazepine IV starts difficult / dangerous in the convulsing patient Best IV agent, lorazepam, impractical for EMS IM treatment is faster and easier Best IM agent, midazolam, is practical for EMS For those whose attention span is getting short. The first two slides are the “everything you need to know slides”. And yes, the study name is a tribute to Johnny and Roy, the inspirations for many young would-be emergency physicians.

Rapid Anticonvulsant Medication Prior to Arrival Trial (RAMPART) IM midazolam autoinjector v. IV lorazepam Double dummy blinded design Exception to consent for emergency research Outcome: termination of seizure prior to ED arrival Sample 800 patients (400 per group) Intention to treat, non-inferiority analysis

Status Epilepticus 120,000 to 200,000 cases / yr Mortality 22% at 30 days 55,000 deaths in the US 1st Yr cost $40,000 /patient Bassin S, et al. Crit Care 2002;6(2):137-42 Claassen J, et al. Neurology 2002;58(1):139-42 DeLorenzo RJ, et al. Neurology 1996;46(4):1029-35 Penberthy LT, et al. Seizure 2005;14(1):46-51 Wu YW, et al. Neurology 2002;58(7):1070-6

Pre-hospital care issues PHTSE trial proved EMS treatment effective Ideal agent and route remain unknown Convulsions can make IV placement challenging Lorazepam has stocking / cost concerns

Intramuscular midazolam Favorable pharmacology for treatment of SE Effectiveness Rapidity Better stability – lower cost Increasing acceptance by EMS

Midazolam levels near 80% of peak as early as 5 minutes after IM administration Human study on healthy volunteers given 0.08 mg/kg to a max of 5 mg Alfonzo-Echeverri, Anesth Prog 1990;37:277-281

IM midazolam stops seizures 4 times faster than IM diazepam (in mice) Raines, Epilepsia. 1990;31:313-7 A model system is described in which sustained clonic seizures are produced by a combination of phenytoin (PHT) and pentylenetetrazol (PTZ) in the mouse, the former agent preventing the terminal tonic spasms produced by the latter.

Brain midazolam concentration remains high even as serum concentration is dropping Megarbane, Toxicology Letters 2005;159:22–31 This is an rat model of MDZ overdose (160 mg/kg IV over 20 minutes), but it makes the point that interstitial brain MDZ concentrations are maintained much longer than serum concentrations.

Duration of seizure suppression with midazolam is hours, and similar to that of diazepam Towne, J Emerg Med 1999;17:323–328 This is based on a human study with 6 patients on continuous EEG monitoring after status epilepticus. It looked at the effect of midazolam versus diazepam versus saline at suppressing inter-ictal spikes after ictal control had been obtained.

Meta-analysis of IM/IN midazolam shows the same efficacy as IV diazepam Review: IV diazepam versus IM/IN midazolam for treatment of seizures Comparison: 01 Effectiveness of IM/IN MDZ as compared to IV DZP Outcome: 01 Termination of seizure IV Diazepam IM/IN Midazolam RR (fixed) Weight RR (fixed) Study n/N n/N 95% CI % 95% CI Chamberlain 8.99 0.92 [0.69, 1.21] 11/13 12/13 Lahat 17.23 1.04 [0.87, 1.25] 24/26 23/26 Rainbow 19.96 0.73 [0.47, 1.12] 23/62 23/45 Mahmoudian 15.73 1.33 [0.97, 1.83] 28/35 21/35 Shah 38.10 0.92 [0.80, 1.07] 54/65 45/50 Total (95% CI) 201 169 100.00 0.97 [0.86, 1.09] 0.5 0.7 1 1.5 2 Favors IM/IN MDZ Favors IV DZP Total events: 140 (IV Diazepam), 124 (IM/IN Midazolam) Test for heterogeneity: Chi² = 6.87, df = 4 (P = 0.14), I² = 41.8% Test for overall effect: Z = 0.54 (P = 0.59)

Meta-analysis of IM/IN midazolam shows more rapid termination of seizures compared to IV diazepam Review: IV diazepam versus IM/IN midazolam for treatment of seizures Comparison: 01 Effectiveness of IM/IN MDZ as compared to IV DZP Outcome: 02 Time to seizure control IV DZP IM/IN MDZ WMD (fixed) Weight WMD (fixed) Study N Mean (SD) N Mean (SD) 95% CI % 95% CI Chamberlain 11 11.20(3.60) 13 7.80(4.10) 3.51 3.40 [0.32, 6.48] Lahat 26 8.00(4.10) 26 6.10(3.60) 7.58 1.90 [-0.20, 4.00] Shah 65 4.20(2.30) 50 1.60(0.90) 88.91 2.60 [1.99, 3.21] Total (95% CI) 102 89 100.00 2.58 [2.00, 3.15] -10 -5 5 10 Favors IM/IN MDZ Favors IV DZP Test for heterogeneity: Chi² = 0.68, df = 2 (P = 0.71), I² = 0% Test for overall effect: Z = 8.74 (P < 0.00001)

Hypotheses Primary IM midazolam is no less effective as IV lorazepam at stopping convulsions prior to ED arrival Secondary Convulsions stop more rapidly with treatment with IM midazolam versus IV lorazepam There is no difference in safety between the two treatments

Inclusion criteria Continuous or repeated convulsive seizure activity for > 5 minutes Patient is still seizing Estimated weight > 13 kg Subject to be taken to participating hospital

Exclusion criteria Major trauma precipitating seizure Hypoglycemia Known allergy to midazolam or lorazepam Cardiac arrest or heart rate <40 beats/minute Medic alert tag with “RAMPART declined” Prior treatment of this seizure in another study Known pregnancy Prisoner

double-dummy design All subjects get active treatment by either IM or IV route IM Active Treatment IV Active Treatment Randomized to: or Autoinjector midazolam Autoinjector placebo IM Route IV syringe placebo IV syringe lorazepam IV Route

Intervention - Dose Smaller Child (13- 39 kg) Lorazepam 2 mg or Midazolam 5 mg Adolescent or Adult (40 kg and up) Lorazepam 4 mg or Midazolam 10 mg

RAMPART Datalogger

Intervention Medic arrives on scene and evaluates patient Ask bystanders duration of seizure and trauma Look for medic alert jewelry Check glucose and vital signs For small children, check estimated weight If criteria are met, study box is opened to enroll Medic states that entry criteria are met Select dose bundle Give IM medication and verbalize Continued….

Intervention (continued) Start IV, give IV med, and verbalize Monitor vital sings and transport Verbalize if convulsions stop At 10 minute after treatment, provide “rescue” meds per local protocol if still seizing en route, verbalize tha med was given At ED arrival, verbailze whether patient is still seizing or not

ED and inpatient treatment Attempt standardized post-intervention care For further seizures in the ED or secondary treatment of prior status… Lorazepam 0.5-0.1 mg/kg plus Phenytoin or Fosphenytoin 18-20 mg/kg

ED and inpatient treatment If seizures continue then… Intubate and ventilate, keep ≤ 37°C Consider vecuronium 0.1 mg/kg Then add: Midazolam 0.2 mg/kg then 1.2 ug/kg/min or Propofol 1 mg/kg then 1-5 mg/kg/hr or Pentobarbital 5-15 mg/kg over 1 hr, then 0.5-5 mg/kg/hr Admit to ICU, early EEG monitoring

Study Activity and Data Collection Study team activated on ED arrival of subject Investigator or coordinator in ED Collect the data logger Complete as many CRF items as possible Approach subject or family to notify and seek consent to continue to collect and use data Restock ambulance with new study kit Follow patient in hospital for AE’s for 24 hours Collect remaining data at discharge

Primary outcome Proportion of subjects with termination of clinically evident seizure determined at arrival in the Emergency Department (ED) after a single dose of study medication. Non-inferiority analysis designed to detect greater than 10% absolute difference in proportion with termination at ED arrival.

Secondary outcomes Rapidity of seizure termination Frequency of subsequent tracheal intubation Frequency and duration of ICU and hospital stay

Sample Size Non-inferiority margin of 10% Power of 0.80 Significance at 0.025 Inflation for data loss and recidivists at 15% N = 800 (400 per group)

Enrollment 800 subjects over 36 months 16 subjects per hub per year If each hub recruits using 14 ambulances the rate is 0.10 subjects/ambulance*month By comparison the PHTSE trial enrolled just over 0.20 subjects/ambulance*month and did not enroll children

Exception from Informed Consent in RAMPART

Does it qualify? Life threatening? Available therapy inadequate or unproven? Informed consent not feasible? Prospect of direct benefit? Trial couldn’t be carried out with EFIC?

Tenets and Approach Flexibility in methodology Local decision making Centralized resources Communication and cooperation

Definitions of Community Geographic Condition-specific

Teleological Pitch If we knew why we are doing something… We can do it better We can judge how well we’ve done it

Possible answer… These are the ways that good doctors and researchers act. In fact, that is why we do informed consent when it is possible.

Built on the ideas of others… Nir Eyal is Instructor in Social Medicine (Division of Medical Ethics) at the Harvard Medical School, with a primary appointment at the university-wide Program in Ethics and Health. Carl Schneider, Chauncy Stillman Professor for Ethics, Morality, and the Practice of Law and a professor of internal medicine at the University of Michigan

People like to be asked more than they like to decide. Even more, they want their doctor/researcher to be the kind of person who would ask.

How do we remain “that sort of doctor” when the emergency situation prevents us from asking? What core values and acts are embodied in the proper moral agent? I’ll suggest 3 (or 4?)

We need to be… Heedful Respectful Transparent Humble

We need to be… Heedful  Comm. Consult. Respectful  Comm. Consult. Transparent  Public Disclosure Humble  Comm. Consult.

We need to be… Heedful  Listen and consider Respectful  Go to the people Transparent  Keep nothing concealed Humble  Do it yourself

These underlying ethical purposes mean that…. Focus is on due diligence of the investigator The investigator’s behavior is the metric

The community’s impact on the investigator matters more than the investigator’s impact on the community It is more important to be transparent then it is to be seen

nett.umich.edu Our website address

Open Discussion

Discussion Stimuli Collection of data after further participation is declined What to do when IRBs don’t agree? Can investigators decide the time to notify is not right? Will existing IRBs cover new EMS FWAs? What happens when emergency research is sustained and no longer novel?

Looking Forward Best format for more cooperative discussion? Electronic forums? Video conference? List-serv? Meetings like this? For each EFIC trial? Annual? As expertise builds, can this lead to a co-op cIRB?

Thank You !