60 / 199
670
B
ird
et al
.:
J
ournal of
AOAC I
nternational
V
ol
.
99, N
o
.
3, 2016
20 (dairy products) or 40 for all other foods. Enumerate plates after
24 ± 2 h of incubation (or 48 ± 3 h in the case of dairy powders,
including whey powder). 3MPetrifilmRAC Plates can be counted
using a standard colony counter with the use of a back-light or
an illuminated magnifier to assist with the estimated enumeration.
(i)
Enumerate all colonies regardless of size, color, or intensity.
(j)
The circular growth area is approximately 30 cm
2
. Plates
containing >300 colonies can be either estimated or recorded as
TNTC. Estimation can only be done by counting the number of
colonies in one or more representative squares and determining
the average number per square. The average number can be
multiplied by 30 to determine the estimated count per plate. If
a more accurate count is required, the sample may need to be
retested at higher dilutions.
(k)
Average the counts between the replicate plates. Report
final results as CFU per gram or milliliter (CFU/g or CFU/mL).
Note
: If there are two dilutions within the countable range, use
the following calculation to determine the final count:
N C 1.1
d
(
)
= Σ ×
where, N is the number of colonies per milliliter or per gram of
product, ΣC is the sum of all colonies on both plates, and
d
is the
dilution from which first counts were obtained.
(l)
Food samples may occasionally show interference on
the 3M Petrifilm RAC Plates for example: (
1
) Uniform blue
background color (often seen from the organisms used in cultured
products). These should not be counted as TNTC.
(
2
) Intense pinpoint blue specs (often seen with spices or
granulated products).
(m)
When necessary, colonies may be isolated for further
identification test using standard procedures. Lift the top film
and pick the colony from the gel.
Results of the Collaborative Study
In this collaborative study, the 3M Petrifilm RAC Plate was
compared with two reference methods for the enumeration
of aerobic bacteria: FDA BAM Chapter 3 (for raw easy-peel
shrimp) and SMEDP Chapter 6 (for instant NFDM). A total of
16 laboratories throughout the United States participated in the
evaluation (all 16 laboratories participated in the evaluation of
the raw easy-peel shrimp, with 15 laboratories participating in
the evaluation of the pasteurized skim milk and instant NFDM)
with 16 laboratories submitting data for raw easy-peel shrimp,
13 laboratories submitting data for the pasteurized skim milk,
and 15 laboratories submitting data for the instant NFDM as
presented in Table
2015.13C
. For the raw easy-peel shrimp and
dry milk power, all participating laboratories submitted data. For
pasteurized skim milk, two laboratories reported deviations from
the protocol and their data were not included in the statistical
analysis. After receipt of samples, Laboratory 6 indicated that
samples were stored at room temperature (24 ± 2°C) instead
of refrigeration temperature (2–8°C) for 24 h prior to testing.
Laboratory 7 reported that their samples were not received until
after 48 h of shipment. The temperature control indicated the
samples were at extremely elevated temperature (30°C). Both
laboratories proceeded with sample analysis, however all samples
analyzed produced results that were greater than the countable
range (TNTC) for all dilutions and no data were submitted.
The 3M Petrifilm RAC Plate results along with FDA BAM
and SMEDP results reported by each laboratory were converted
to logarithmic values for statistical analysis and plotted using
a Youden plot. The Log
10
individual laboratory results are
presented in Tables
2015.13D-G
. Figures 1–4 present theYouden
plots for each method for each matrix. Figures 5–8 present the
mean Youden plots for 3M Petrifilm RAC and the reference
methods. The transformed data were analyzed for outliers by
Cochran and Grubbs’ tests. No evidence of physical cause or
suspicion of cause was noted, so all identified outliers were
included in the statistical analysis. The difference of means and
the reverse-transformed difference of means (including 95%
CIs) were determined for each contamination level for each
matrix to determine whether a statistically significant difference
existed between the methods. s
r
and s
R
were determined for
each contamination level for both the 3M Petrifilm RAC
Plate and FDA BAM and SMEDP methods. The results of the
Figure 1. Youden plots for 3M Petrifilm RAC Plate and FDA BAM results for raw easy-peel shrimp evaluated at 32°C.