1. Volume 6, Issue 3, Pages 183-195 (June 2004)
Comparison of automated culture systems with a CFR/USP-compliant method for sterility testing of cell-therapy products 
H.M. Khuu, F. Stock, M. McGann, C.S. Carter, J.W. Atkins, P.R. Murray, E.J. Read 
Cytotherapy 
Volume 6, Issue 3, Pages (June 2004)
DOI: /
Copyright © 2004 International Society for Cellular Therapy Terms and Conditions

2. Figure 1
Schema for sample preparation, aliquotting, and seeding of organisms into representative cell-therapy products. This diagram shows processing of one MNC apheresis product for one replicate of one organism on 1 day. For each replicate, MNCs were suspended in six different background media [plasma, infusion medium, freeze mix, RPMI, XVIVO-20, and RPMI+ multiple antibiotics (ABX)] and seeded with two organism dilutions (50 CFU and 10 CFU). Final product volumes were 0.5mL containinga targeted content of10-50×106 cells. Each seeded test sample was inoculated into aerobic bottles (open symbols) and anaerobic bottles (shaded symbols) of each of the three culture methods: CFR/ USP (squares), BacT/Alert (hexagons), and Bactec (circles), for a total of 72 culture bottles inoculated per replicate. Ten organisms were tested, and each organism was tested in triplicate for a total of 216 bottles inoculated per organism.
Cytotherapy 2004 6, DOI: ( / )
Copyright © 2004 International Society for Cellular Therapy Terms and Conditions

3. Figure 2
Detection of positives by the CFR/USP, BacT/Alert, and Bactec culture methods in representative MNC products that differed only in background medium. Each of the bars represents n= 60 (10 organisms×two organism dilutions×three replicates). With the exception of RPMI containing multiple antibiotics (RPMI+ABX), the product medium did not have an effect on the ability of any of the three methods to support growth of seeded organisms. For the other five product media, the rate of positive cultures was higher for BacT/Alert and Bactec than for the CFR/ USP method. Inhibition of organism growth in RPMI+ABX occurred despite the presence of penicillinase in CFR/ USP bottles and antibiotic-binding substances in BacT/Alert and Bactec bottles.
Cytotherapy 2004 6, DOI: ( / )
Copyright © 2004 International Society for Cellular Therapy Terms and Conditions

4. Figure 3
Time to detection of positive cultures by CFR/ USP, BacT/ Alert, and Bactec culture methods. For each of the 10 organisms, the time to detection is displayed as the mean (□) and range, i.e., minimum (+) and maximum (×) values. As there was poor growth of certain organisms in the medium containing multiple antibiotics (RPMI-j- ABX), values for those experiments were excluded. Therefore, for each organism, each mean and range of values represents a maximum n=30 (five product media×two organism dilutions×three replicates). Values were further excluded if positive cultures were not detected for one or more of the methods, to ensure an equal number when comparing the three culture methods for a given organism. Therefore, n=29 for S. aureus, 2 for M. luteus, 30 for K. pneumoniae, 30 for P. aeruginosa, 30 for Y. enterocolitica, 21 for B. subtilis, 30 for C. sporogenes, 29 for P. acnes, 20 for C. albicans, and 2 for A. niger. Time to detection for all organisms except P. acnes was longer for the CFR/ USP method than for either of the two automated methods. For six of the 10 organisms (S. aureus, K. pneumoniae, P. aeruginosa, Y. enterocolitica, B. subtilis, and C. sporogenes), the mean time to detection was ≤ 24h in both automated systems, but the CFR/ USP method was able to consistently detect only two of the same organisms (K. pneumoniae and C. sporogenes) by 24 h, and only an additional two (S. aureus and Y. enterocolitica) within 48 h. Three organisms (M. luteus, C. albicans, and A. niger) required > 7 days for detection in the CFR/ USP system.
Cytotherapy 2004 6, DOI: ( / )
Copyright © 2004 International Society for Cellular Therapy Terms and Conditions

5. Figure 4
Summary data for detection of positive cultures (expressed as percentage of positive cultures, y-axis on left) and time to detection (expressed as mean and range in hours,y-axis on right) by the CFR/USP, BacT/Alert, and Bactec culture methods, for 10 organisms tested. For each of the 10 organisms, the data include values for the five types of product media, but exclude data on products containing multiple antibiotics (RPMI-j-ABX) because of inconsistent growth in that medium. Therefore, for each organism, values represent a maximum n=30 (five product media×two organism dilutions×three replicates). Compared with the two automated methods, the CFR/ USP method had lower rates of detection for M. luteus, B. subtilis, C. albicans, and A. niger, and times to detection for cultures that did grow were substantially longer. Overall, the BacT/ Alert method failed to detectP. acnes, but for all other organisms, when growth was present, mean time to detection was within 24h for six organisms and within 48 h for the other three. The Bactec method was also able to detect all organisms, includingP. acnes, and had shorter times to detection than the CFR/ USP method.
Cytotherapy 2004 6, DOI: ( / )
Copyright © 2004 International Society for Cellular Therapy Terms and Conditions