FDA Presentation to BPAC: NMRC RESUS Protocol Using HBOC201

1
FDA Presentation to BPACNMRC RESUS Protocol
Using HBOC-201

• Toby Silverman MD, LTC, USAR (Ret)
• Branch Chief, Clinical Review Branch
• Division of Hematology, Office of Blood
• December 14, 2006

2
Oxygen Therapeutics in Trauma

• FDA recognizes the important role that oxygen
  therapeutic agents might play in improving
  outcomes in traumatic hemorrhagic shock, and
  supports the development of safe and effective
  agents for use in resuscitation
• FDA recognizes the unmet military and civilian
  need for improved outcomes in trauma

3
RESUS Trial Protocol Overview

• HBOC-201 vs lactated Ringers solution (LR) for
  Rx of life-threatening post-traumatic hemorrhage
  in the urban-ambulance setting
• Pre-hospital use only
• Waiver from requirements for informed consent (21
  CFR 50.24)
• Powered to detect 15 relative reduction in
  all-cause mortality at 28 days from estimated
  58.1 to 49.4 (1130 subjects, p 0.045)
• Phase 2, 50-subject study to assess
• Feasibility and ability of study to answer
  efficacy and safety outcome questions
• Appropriateness of entry criteria to target the
  desired population

4
RESUS Trial Clinical Hold

• 1. Safety signals arising out of previous phase
  2/3 studies
• Excess clinically significant AEs (adverse
  events) in all analyses
• 2. Dosing and administration
• Lack of preclinical/clinical dose response
  studies
• 3. Mortality estimate
• Wide variability in projected mortality for
  individual subjects
• 4. Magnitude of treatment effect
• Cannot be derived from animal data

5
RESUS Trial Clinical Hold (2)

• 5. Benefit Risk
• SAEs (serious adverse events) observed in
  previous trials, uncertainty of the treatment
  effect, and wide variability in expected
  mortality for individual subjects preclude
  determination of a positive benefitrisk ratio.
• 6. Risk mitigation strategies proposed by NMRC
• Monitoring and therapeutic interventions may
  not suffice to offset risks associated with use
  of HBOC-201

6
Additional Concerns (Non-hold Items)

• Restriction on age (exclusion of subjects 70
  years old)
• Generalizability of data from RESUS to routine
  prehospital emergency care
• Complexity of RESUS trial
• Requirement for specialized EMT training
• Requirement for specialized training of
  in-hospital personnel
• Practicality of calculating RTS (Revised Trauma
  Score) under field conditions
• Identification of patients for whom use of
  product may be appropriate

7
1. Safety Data Limitations

• Complete Review (CR) letter of July 30, 2003 to
  Biopure documented numerous deficiencies in the
  conduct of pivotal trial HEM-0115
• Good Clinical Practice
• Data quality/completeness
• Difficulties assessing/verifying seriousness and
  frequency of AEs
• Laboratory database issues due to co-mingling
  central laboratory and site information

8
1. Safety Data Limitations (2)

• Because of the limitations of the databases, the
  dataset provided represents a minimum estimate of
  adverse event information
• For purposes of discussion, FDA will be
  presenting information on AEs derived from a
  consensus safety database
• FDA and Biopure differ on adjudication of a few
  cases which are highlighted in the FDA tables.

9
FDA Safety Analysis

• Assumed a priori that a rigorous statistical
  assessment of differences between HBOC-201 and
  control for a particular AE would not be possible
  because of small sample size, even for HEM-0115
• AEs/SAEs were expected to occur with low
  frequency
• Data were pooled to achieve a larger sample size
  from which to estimate frequency of low-incidence
  events

10
1. FDA Safety Analysis (2)

• Subjects in previous studies, including HEM-0115,
  were
• Stable
• Medically cleared
• Judged not to be at particular cardiovascular
  risk
• Monitored and treated according to standard care
• Study designs were generally similar
• Administration of HBOC-201 vs control (LR, HES
  (Hespan), RBC)
• Hb lt threshhold level with/without other
  signs/symptoms of anemia, or after fixed volume
  blood loss
• RBC, crystalloid, and colloid available as needed
• Safety and tolerability vs control
• Effect on allogeneic RBC usage

11
1. Outcome of Safety Analysis

• Trends seen in the pooled database were also seen
  in the individual studies and across different
  types of studies (e.g. by control)
• The pooled analysis
• Showed safety signals already noted in the
  various individual phase 2 studies leading up to
  pivotal HEM-0115
• Identified new concerns for further analysis
  (e.g. MI, renal failure requiring dialysis, CVA)

12
Data Pooling Hypertension (Example)
13
1. Safety AEs All Clinical Trials
FDA Study 0115-5405 and study 0107-0403 Study
0101-725 ?Study 0115-4308
14
1. Safety AEs by Type of Control Solution
FDA Study 0115-5405 and study 0107-0403 Study
0101-725 ?Study 0115-4308
15
1. Safety AEs Stratified by Age
FDA Study 0115-5405 and study 0107-0403 Study
0101-725 ?Study 0115-4308
16
Pivotal Trial HEM-0115 (48 of BLA)

• Design Multicenter, randomized, single-blind,
  RBC-controlled, parallel-group
• Population 693 subjects undergoing non-emergent
  orthopedic surgery
• Randomization at first transfusion decision
• HBOC-201 given as 60 g (2 x 30 g bags) to a
  maximum of 300 g (10 units) for low Hb (gt6.5 but
  lt10 g/dL) at least one additional sign or
  symptom of anemia
• Primary Endpoint Avoidance of RBC transfusion
  during 6-week study period

17
1.Safety Clinically Important AEs in HEM-0115
Subject 5405 and 4308
18
1. Safety AEs in HEM-0115

• Biopure hypothesized that
• Inaccurate dosing guidelines led to over-infusion
  of the product (e.g., pulmonary edema)
• More test subjects than control subjects had a
  history of cardiac disease
• However, within HBOC-201 cohort, post hoc
  stratification presence/absence of history of
  heart disease ? no difference in incidence of AEs

19
Safety AEs in HEM-0115 (2)

• HH HBOC-201 subjects who did not receive RBC
• HR HBOC-201 subjects who also received RBC
• R- RBC subjects who received 3 units
• R RBC subjects who received gt 3 units

20
1. Safety AEs in HEM-0115 (2)

• Biopure hypothesized that
• Total Hb lower in HBOC-201 arm - ? risk of
  ischemia
• However,
• Mean 1.23 g/dL difference between HBOC-201 and
  RBC in total Hb for lowest recorded value
  probably does not explain the excess of AEs for
  HBOC-201
• Within HBOC-201 cohort, post hoc stratification
  by nadir total Hb lt 8 g/dL or 8 g/dL ? no
  difference in incidence of high frequency AEs
  such as hypertension, elevated troponin levels,
  or oliguria

21
1. Safety Conclusions

• Excess adverse events are consistently associated
  with use of HBOC-201
• RBC-controlled surgery studies
• Crystalloid/colloid controlled surgery studies
• Age stratification in surgery studies
• Total Hb and history of heart disease do not
  appear to be independent predictors of adverse
  event imbalances when assessed in post hoc
  stratification of HBOC-201 cohort in HEM-0115
• FDA considers these AEs important to consider
  when thinking about RESUS

22
2. Dosing and Administration

• Default administration rate 50 mL/min (actual
  rate determined by judgment of EMS provider)
• Preclinical animal studies of hemorrhagic shock
• Range gravity infusion (not otherwise
  quantified) to 10 mL/kg/min
• No dose-ranging studies
• Limited clinical data
• Infusion rate 3.8 mL/min in Phase 2
  crystalloid/colloid studies
• Mean infusion rate 5.5 mL/min in HEM-0115
• 4/ 353 subjects at rates 40 mL/min
• No dose-ranging studies

23
2. Dosing and Administration (2)

• Limited safety data for product administration at
  higher dosing rates and doses is a principal
  concern given the known intrinsic properties of
  HBOCs (vasoactivity and vascular injury) and the
  AE profile of HBOC-201 in previous trials

24
3. Mortality Estimate Challenges

• Wide variability in projected mortality for
  individuals based on proposed RESUS entry
  criteria
• Proportion of trauma population who can
  potentially benefit from any life-saving therapy
  is a very small subset of the total trauma
  population
• RESUS represents lt1 of total trauma population
• Information on proportion of serious trauma
  patients alive at the scene who expire before
  reaching the ER is not readily available

25
3. Basis for Mortality Estimate in RESUS

• Enrollment of subjects at higher risk of dying
  from hemorrhagic shock
• SBP lt90 mm Hg
• Weighted Revised Trauma Score (RTS, full range
  0-7.84)
• Enrollment criteria ranging from 1 to lt 5
• Exclusion of subjects for whom blood is readily
  available within 10-15 minutes

26
3. Mortality Estimate

• The Revised Trauma Score is calculated based on
  three parameters Glasgow Coma Score, systolic
  blood pressure, and respiratory rate
• GCS has 3 components- eye opening, verbal
  response, motor response

RTS 0.9368 GCS 0.7326 SBP 0.2908 RR
27
Revised Trauma Score Issues

• RTS cannot be computed unless data from all three
  components captured
• Glasgow Coma Score (GCS) heavily confounded by
  intubation, severe facial injury, intoxication,
  etc.
• No consensus in literature for allocating verbal
  response scores for intubated or
  pharmacologically paralyzed patients
• Studies report a loss of cases for analysis of
  3-28
• Usually it is the GCS that is missing
• Difficulty in coding GCS portion of RTS can lead
  to large variation in the RTS

28
3. Mortality Estimate (2)
www. Trauma .org
29
3. Mortality Estimate RTS Issues

• Number of subjects and number of deaths are not
  equally distributed throughout the range of RTS
  scores
• Greatest potential benefit to offset risk is
  distributed predominantly to those with lower
  proposed RTS scores
• Least potential benefit to offset risk is
  distributed to those with higher proposed RTS
  scores
• Small imbalances in RTS scores can have greater
  effect on outcome than the therapeutic
  intervention

30
3. Mortality Estimate NTDB
31
3. Mortality Estimate NTDB
Mortality and N in hypotensive subjects lt 70 y/o
admitted to US trauma centers with/without TBI
(NTDB Hospital Arrival Data)
N2441
1168
516
367
390
32
3. Mortality Estimate UAB/UMD
33
Mortality Estimate UAB/UMD
N234
44
95
55
40
34
3. Mortality Estimate Conclusions

• The patient population is likely to be
  heterogeneous
• While ranges for mortality differ in the three
  available databases, all indicate a very wide
  range of survival probabilities
• While the overall average mortality rate is
  58, many subjects will have a probability of
  death that is much lower than the average

35
4. Treatment Effect

• There is no clinical or preclinical basis to
  allow the numerical estimate of treatment effect
  for HBOC-201 in pre-hospital trauma resuscitation
• Not possible to estimate the potential magnitude
  of the treatment effect from clinical trials
  using HBOC-201 in elective surgery
• No prospective, randomized, controlled Phase 2
  studies in consenting trauma subjects
  with/without TBI have been conducted/completed
  with HBOC-201

36
4. Treatment Effect (2)

• Sponsor bases its estimate of treatment effect
  and its assessment of likely safety for RESUS on
  results of a subset of preclinical animal models
  of trauma and hemorrhagic shock

37
4. Treatment Effect Limitations of Models

• Limitations inherent to the animal models that
  preclude direct extrapolation of results to
  humans.
• Models of hemorrhagic shock
• Basic physiology vs survival
• Controlled vs uncontrolled hemorrhage
• Vascular vs parenchymal organ hemorrhage
• With/without TBI
• Resuscitation strategies
• Fixed volume vs fixed BP vs BP/HR-controlled
• Periods of observation
• Hypotensive vs normotensive resuscitation
• Volume ratios of resuscitation fluids
• Short vs long transit times

38
4. Treatment Effect (3)

• Preclinical tests are not intended to supplant
  data derived from adequate and well-controlled
  trials in humans, nor is safety information
  derived from animal studies intended to supplant
  safety data derived from clinical trials
  performed in humans
• Results of preclinical studies
• do not establish a quantitative estimate of
  treatment effect and
• do not negate safety findings in completed
  clinical trials in humans

48 FR 26720 IND Regulation Rewrite See preamble
39
4. Conclusions About Treatment Effect

• Proof of concept that HBOC-201 might sustain life
  in trauma has been shown by animal studies in
  narrowly defined models of lethal hemorrhagic
  shock
• Nevertheless, preclinical data could potentially
  support studies of HBOC-201 in settings where an
  extremely high mortality rate is expected (e.g.
  massive hemorrhage with/without prolonged delay
  to definitive care, TBI)

40
5. Potential Concerns Using HBOC-201 for
Uncontrolled Hemorrhage in the Ambulance

• Risk of fluid under-resuscitation
• Limitations inherent to the ambulance setting
• Risk of increased bleeding or re-bleeding due to
  hypertension- of concern in all trauma patients
  but especially those with head trauma

41
5. Potenial Concerns Using HBOC-201 for
Uncontrolled Hemorrhage in the Ambulance

• Fluid under-resuscitation
• BP used as a surrogate for perfusion
• BP gt 100 mm Hg using vasoactive HBOC-201 could
  mislead healthcare providers to withhold needed
  crystalloid, resulting in tissue underperfusion
• Unclear how to interpret classic signs of occult
  shock (e.g. thready pulse, cool extremities) when
  using HBOC-201

42
5. Limitations of the Ambulance
43
5. Potential Concerns Using HBOC-201 for
Uncontrolled Hemorrhage in the Ambulance

• Increased bleeding and re-bleeding
• Although thrombus after an arterial injury is
  formed almost immediatelyit is initially soft
  and jelly-like Transformation to a more rigid
  hemostatic plug requires at least 20-30 minutes
  following injury
• Resuscitation strategies which cause abrupt
  increase in blood pressure and flow may increase
  hemorrhage volume.

• Stern, S. Low-volume fluid resuscitation for
  presumed hemorrhagic shock helpful or harmful?
  Curr Opin Crit Care 2001 7 422-430

44
5. Vasoactivity of HBOC-201
45
5. Percent of Hypotensive (SBP 90 mm Hg)
HEM-0115 Subjects With SBP Responses gt 130 mm Hg
9
4
3
2
2
0
1
1
46
5. Potential Safety Concerns in the Ambulance
Summary

• HBOC-201 is a vasoactive product with a duration
  of action lasting hours
• BP can continue to increase after HBOC-201 is
  stopped at 120 mm Hg
• BP elevations cannot be treated in the ambulance
• Abrupt BP elevations can increase bleeding
• Increased bleeding is problematic for all
  subjects with uncontrolled bleeding in the
  ambulance, but especially for subjects with TBI

47
6. Benefit and Risk

• 1999- Workshop on Safety and Efficacy evaluation
  of oxygen therapeutics when used as red blood
  cell substitutes and as resuscitation fluids
• 2004- Draft guidance
• Hierarchical approach to evaluation of safety
• Initial evaluation in situations where AEs
  expected to be uncommon to facilitate detection
  of potential safety problems
• Subjects medically cleared, carefully monitored,
  medically managed according to in-hospital
  standard of care guidelines
• RBC control
• Demonstration of adequate safety profile when
  compared with RBC allows evaluation in less
  stable trauma subjects able to provide consent,
  or unstable trauma subjects unable to provide
  consent

48
6. Benefit and Risk (2)

• In field setting, oxygen therapeutic should have
  superior survival outcome when compared to an
  asanguinous solution
• Possible for oxygen therapeutic to have an
  inferior safety profile when compared with blood,
  and yet reduce mortality in trauma in the field
  when compared with asanguinous solutions
• Very difficult to design clinical trial
• Not easy to weigh relative importance of observed
  safety signals and AEs vs. potential benefit in
  terms of lives saved, particularly if findings
  suggesting clinical benefit have not been
  observed in other settings

49
6. Benefit and Risk (3)

• Wide variability in projected mortality for
  individual subjects means that benefits to offset
  risks are not evenly distributed
• Magnitude of treatment effect cannot be estimated
  from animal studies
• Prior studies in humans do not provide a basis
  for estimating treatment effect in trauma

50
6. Benefit and Risk- Challenges

• HBOC-201 was associated with 50 increase (1
  absolute excess) deaths due to SAEs
• Additional deaths due to SAEs will offset
  potential benefit in terms of lives saved with
  HBOC-201
• Reduce power of study to detect a beneficial
  effect

51
6. Benefit and Risk Challenges

• Victims of trauma, even young ones, may not have
  a lower risk for AEs, but rather, may have a
  higher risk than older, medically cleared
  subjects undergoing elective surgeries
• Excess AEs noted in HBOC-201 treatment arm in
  elective surgery could potentially be greater in
  critically ill trauma subjects
• Transport times in urban ambulance setting are
  short and window of opportunity for benefit is
  small

52
6. Benefit and Risk Conclusions

• FDA performed an extensive sensitivity analysis,
  varying assumptions for deaths due to SAEs,
  effect size, and the underlying mortality rate
• FDA found that the trial, as designed, is very
  sensitive to small fluctuations in the
  assumptions and is not robust

53
RiskBenefit Analysis 3 by NMRC

• The detailed methods and other information
  underlying the NMRC risk assessment in the Issue
  Summary have not been submitted to FDA and
  therefore could not be reviewed adequately.
• However, FDA has preliminary concerns with the
  model as presented
• Documentation on validation of the model is
  lacking
• Sensitivity analyses performed are incomplete and
  do not take power of the study for each scenario
  into account
• Analysis using retrospective post hoc subsetting
  of patient groups in HEM-0115 is problematic and
  may not be valid

54
7. Risk Mitigation Restrictions on Age

• FDAs position is that the trial should not have
  an upper age limit
• Fastest growing segment of trauma population is
  population gt 50 years old and more particularly,
  gt 70 years old (www.cdc.gov/ncipc/wisqars/default/
  htm)
• Distinguishing subjects above or below a
  particular age, under field conditions, likely to
  be difficult
• Older subjects may be at greater risk of ischemic
  consequences of severe hemorrhage and therefore
  might also potentially benefit more from
  administration of an oxygen-carrying
  resuscitation fluid

55
Summary

• FDA has a number of concerns about RESUS
• 1. Excess clinically significant AEs in all
  analyses
• 2. Lack of preclinical/clinical dose response
  studies to support proposed dosing
• 3. Wide variability in projected mortality for
  individual subjects
• 4. Magnitude of treatment effect cannot be
  derived from animal data

56
Summary

• 5. SAEs observed in previous trials, uncertainty
  of the treatment effect, and wide variability in
  expected mortality for individual subjects
  preclude determination of a positive benefitrisk
  ratio.
• 6. Monitoring and therapeutic interventions may
  not suffice to offset risks associated with use
  of HBOC-201
• 7. Excluding the elderly may not reduce risks
  associated with use of HBOC-201

57
Summary

• FDA asks the Advisory Committee to consider the
  following questions

58
Question 1

• Please discuss the following safety concerns
  raised by FDA
• Safety signals and adverse events in previous
  clinical studies
• Demonstrated vasoactivity of the product
• Limited safety data for higher doses and rates of
  administration

59
Question 2

• Please discuss whether the available preclinical
  and clinical data are sufficient to estimate a
  treatment benefit for all-cause mortality at 28
  days in the proposed RESUS trial

60
Question 3

• After considering all of the available data, do
  the benefits outweigh the risks for individual
  subjects in the RESUS trial?

61
Question 4

• Are there additional data that could help inform
  an assessment of benefitrisk in the RESUS trial?

62
Question 5

• Please comment on any modifications to the study
  design that might improve the benefitrisk ratio
  in the RESUS trial
• For example, a trial targeting a group with
  higher predicted mortality.