Evaluation of proposed FDA criteria for the evaluation of radiolabeled red cell recovery trials


  • Conflict of interest: The authors' research laboratory has been supported by participating sponsors listed in the manuscript and/or the authors have served as advisors to these sponsors.

Larry J. Dumont, MBA, PhD, Dartmouth-Hitchcock Medical Center, Department of Pathology, One Medical Center Drive, Lebanon, NH 03756-0001; e-mail: Larry.J.Dumont@hitchcock.org.


BACKGROUND: FDA requirements for recovery of radiolabeled red blood cells (RBCs) 24 hours after autologous reinfusion in normal subjects have evolved over time. This study defined the ability of currently available RBC systems to satisfy the most recently proposed criteria.

STUDY DESIGN METHODS: RBC recoveries were collected from US laboratories participating in clinical trials for RBC systems that have received FDA approval. Data were stratified for analysis into liquid-stored, gamma-irradiated, and frozen components. With statistical software, 24 individual samples were randomly selected with replacement from each stratum, repeating this for 5000 sample groups per stratum, to simulate experimental outcomes for each population. The percentage of sample groups that passed each and all of the proposed FDA criteria was determined. This procedure was repeated for recovery success thresholds of 75, 70, and 67 percent.

RESULTS: A total of 941 RBC recoveries were obtained from 11 laboratories and 34 studies performed between 1990 and 2006 for 12 sponsors. While the criterion for the mean was almost always satisfied, the standard deviation (SD) criterion was more problematic. Causing most failures was the success threshold definition. Changing the success threshold from 75 to 70 percent or 67 percent increased the likelihood of meeting the requirement for all RBC types.

CONCLUSION: The probability of passing the FDA-proposed criteria for currently FDA-approved products was poor. Changing the success threshold for an individual RBC recovery from 75 to 67 percent resulted in improved ability to meet this criterion for all three RBC types. This change had no affect on the pass rates based on the mean and SD criteria.