Title: Evaluation of Proposed FDA Criteria for Evaluation of Radiolabeled Red Cell Recovery Trials Larry J. Dumont, MBA, PhD James P. AuBuchon, MD for the Biomedical Excellence for Safer Transfusion (BEST) Collaborative
1Evaluation of Proposed FDA Criteria
forEvaluation of Radiolabeled Red Cell Recovery
TrialsLarry J. Dumont, MBA, PhDJames P.
AuBuchon, MDfor the Biomedical Excellence for
Safer Transfusion (BEST) Collaborative
Blood Products Advisory Committee Rockville, MD
May 1, 2008
TRANSFUSION online 22-Feb-2008 doi
10.1111/j.1537-2995.2008.01642.x.
2Conflicts of Interest
- Consultant
- Gambro BCT
- bioMérieux
- BCSI
- Verax Biomedical
- Research support DHMC
- Cerus
- Fenwal
- Gambro BCT
- Haemonetics
- Immunetics
- Navigant Biotechnologies
- Verax Biomedical
- Travel Support
- FDA
3FDA - RBC Performance Criteria
- In vitro e.g., hemolysis and ATP
- In vivo - autologous 24 hour recovery
- In vivo - clinical outcome (safety trials)
4FDA - RBC Performance Criteria
- In vitro e.g., hemolysis and ATP
- In vivo - autologous 24 hour recovery
- In vivo - clinical outcome (safety trials)
5Background FDA Requirements
Mean 24h recovery 70 (Ross et al. JCI
194726687-703)
6FDA Requirements
Mean recovery ? 75 - and SD ? 9 - and
LCL95 for population proportion of successes (?
75) gt 70
BPAC, July, 2004
Success Threshold ?21 successes out of 24 ?18
successes out of 20
7IT WILL BE OK
CAUTION
8Radiolabeled Red Cell Recovery circa 1970
9Radiolabeled Red Cell Recovery circa 1985
10Radiolabeled Red Cell Recovery circa late 1990s
11Radiolabeled Red Cell Recovery circa 2004
12Implication
Test RBC must have ? 90.3 success for an 80
chance of passing an in vivo recovery trial
- What is the clinical evidence?
- What is the capability of current RBC products?
13Our Conclusion after review of the evidence
A success threshold of 67-70 will provide a
reasonable probability of passing the
FDA-proposed criteria for RBC products in current
use in the United States
14Objective
- Define the ability of currently available Red
Blood Cell (RBC) collection and storage systems
to satisfy new RBC in vivo recovery criteria
proposed by the FDA for approval of RBC systems.
15Methods
Data review and cleaning
Approved/cleared methods 1990-2006
Central Database
Data verification by Sponsors n12
16Methods
Sample with replacement n24 N5000
Liquid Stored (42 days)
Central Database
Gamma Irradiated (28 days post)
Sample with replacement n24 N5000
Frozen, deglycerolized (15-30 days)
Sample with replacement n24 N5000
FDA/CBER
17Evaluable RBC Recoveries 1990 - 2006
n34 studies Leukocyte- reduced Non-leukoctye- reduced
Automated Collection
Liquid stored 205 9 214
Gamma Irradiated 67 0 67
subtotal 272 9 281
Manual Collection
Liquid stored 291 136 427
Gamma Irradiated 24 32 56
subtotal 315 168 483
Frozen 140 37 177
TOTAL 727 214 941
18BINOMIAL EXPANSION
N24 Probability (Number of successes 21 )
0.693
N Min 5 Median Frequency of Recovery lt 75 Frequency of Recovery lt 70 Frequency of Recovery lt 67 Mean SD
641 36 70.7 82.4 11.7 4.5 1.9 82.1 6.71
19BINOMIAL EXPANSION
N24 Probability (Number of successes 21 )
0.035
N Min 5 Median Frequency of Recovery lt 75 Frequency of Recovery lt 70 Frequency of Recovery lt 67 Mean () SD ()
123 47.1 63.6 79.3 30.9 18.7 8.9 77.9 8.36
20BINOMIAL EXPANSION
N24 Probability (Number of successes 21 )
0.957
N Min 5 Median Frequency of Recovery lt 75 Frequency of Recovery lt 70 Frequency of Recovery lt 67 Mean SD
177 52.2 74.2 88.0 5.6 2.8 1.2 86.6 7.41
21One Sample
- Replicate LAB ASC Rec lt70
- 257 1 1 64.20 1
- 257 1 1 71.60 0
- 257 1 1 70.20 0
- 257 1 1 79.30 0
- 257 1 1 76.70 0
- 257 1 1 93.50 0
- 257 1 1 73.40 0
- 257 1 1 85.53 0
- 257 1 1 71.14 0
- 257 2 1 81.04 0
- 257 2 1 66.85 1
- 257 2 1 77.13 0
- 257 2 1 84.04 0
- 257 2 1 81.85 0
- 257 2 1 77.68 0
- 257 2 1 80.49 0
- 257 2 1 58.09 1
- 257 2 1 82.71 0
Mean74.9 X SD10.5 X 4/24 lt 70
X
22RESAMPLING n24 N5000
Failures
Number of successes (75) 21 67.3
Mean 75 100
SD 9 95.2
23RESAMPLING n24 N5000
Failures
Number of successes (75) 21 3.5
Mean 75 95.5
SD 9 71.4
24RESAMPLING n24 N5000
Failures
Number of successes (75) 21 95.4
Mean 75 100
SD 9 83.9
25Preliminary Conclusion
- FDA-proposed success threshold of gt75 for
individual recovery is unacceptable - The general clinical performance for these
products is adequate as proved over years of
clinical practice, and certainly represents the
state-of-the-art
26Sensitivity to Success Thresholdn24
27Chance of current, state-of-the-art RBC products
passing FDA-proposed criteria
RBC Mean 75 SD 9 Success Recovery gt 75 Success Recovery gt 70 Success Recovery gt 67
42d Liquid Stored 100 95.2 69.3 (58) 97.9 (94) 99.9 (99)
Gamma 95.5 71.4 3.5 (3.0) 31.8 (25) 84.0 (74)
Frozen 100 83.7 95.7 (90) 99.6 (98) 100 (99)
18 successes out of 20 trials
28Other Key Observations
- Current RBC components are not different than the
study population - There are differences between laboratories and/or
study subjects
29Current Methods are not different than the study
population
lt 75 11.7 lt 70 4.5
4 of 36 lt 75 (11) 1 of 36 lt 70 (2.8)
30There are differences between laboratories and/or
study subjects
31Conclusions
- The FDA-proposed success threshold of gt75 is not
validated against currently approved RBC products
available in the US - Based on actual in vivo recovery performance, a
success threshold of 67-70 will provide a
reasonable probability of passing the
FDA-proposed criteria - LCL95 for population proportion of successes (?
67) gt 70 - The mean and SD criteria may be applied as
general guidance, but used with caution as
distributions do not meet normality assumptions.
32Our Conclusion after review of the evidence
- A success threshold of 67-70 will provide a
reasonable probability of passing the
FDA-proposed criteria for RBC products in current
use in the United States - We should not make it unnecessarily burdensome
for new innovations to enter the market
33Acknowledgements
Study Laboratories American Red Cross,
Louisville, KY Blood Center of Wisconsin,
Milwaukee, WI Dartmouth-Hitchcock Medical Center,
Lebanon, NH Hoxworth Blood Center, Cincinnati,
OH Mayo Clinic, Rochester, NY American Red Cross,
Norfolk, VA University of North Carolina, Chapel
Hill, NC University of New Mexico, Albuquerque,
NM University of Virginia, Charlottesville,
VA Walter Reed Army Institute of Research Yale
University, New Haven, CN
Study Sponsors American Red Cross Baxter Baxter/A
rmy Cutter / Pall Gambro BCT, Inc. Haemonetics /
Transfusion Technologies Hemasure MacoPharma /
UnitedPharma Terumo Vitex
Individuals Lauren Clark Jaime
Houghton Sherrie Sawyer Jose A.
Cancelas Mike McAteer Yariv Sivan Tammy
Corda Jeff Miripol Edward
Snyder M. Dean Elfath Ed Nelson
Tania VandenBroeke John Hess Leslie
Rose Pamela Whitley Stein Holme
Neeta Rugg