127 / 129
M
ozola
et al
.:
J
ournal of
AOAC I
nternational
V
ol
.
97, N
o
. 3, 2014
835
vulgaris
produced colonies on XLT-4 agar. The remaining cases
of no growth appeared to be random with respect to strain and
medium. A total of 691 analyses were performed on exclusive
strains. Collaborator 16 reported positive results on six of seven
plates streaked with the
E. cloacae
culture. The remaining
agar, TSA, was reported to have no growth. Collaborator 16
reported that the six plates all contained growth with colonies
of a
Salmonella
-like appearance. It is concluded that this
culture became contaminated at some point during preparation
or analysis and therefore these data were eliminated from the
statistical analysis. Of 685 remaining analyses, 661 produced
negative results for accuracy with exclusive strains of 96.5%.
Asummary of results by agar medium is shown in Table 4. The
percentage of correct results was very similar for all seven media,
ranging from 97.6 to 98.9%.
Discussion
In this multilaboratory evaluation of the ANSR
Salmonella
test for identification of presumptive
Salmonella
spp. isolates
from agar media, the method exhibited exceptional accuracy
with inclusive strains and a high degree of exclusivity with non-
salmonellae. Of the 18 laboratories participating in the study,
15 reported results with overall accuracy of 99 to 100%. There
was only a single false-negative result out of 756
Salmonella
spp. colonies tested. Excluding data generated from a suspected
contaminated slant culture, there were 24 false-positive results
on non-
Salmonella
spp. colonies out of 685 colonies tested. All
but seven of these aberrant results occurred in three laboratories.
Laboratory 16 reported six false-positive results in addition
to those linked to the contaminated slant culture. No further
information is available for these samples, except that all six
ANSR fluorescence curves were very strong, typical of true
positive results. Laboratory 2 reported six false-positive results;
four of these occurred in a singleANSR assay run of 15 samples.
All but one of the false-positive results showed atypical, weak
fluorescence curves, suggestive of cross-contamination during
performance of the ANSR assay. Laboratory 13 reported five
false-positive results. Again, all but one of these results showed
atypical, weak fluorescence curves. Additionally, raw data
received from this laboratory indicated that one assay run was
repeated in total due to extreme aberrant results (i.e., invalid
assays), suggesting that the technician was experiencing
difficulty in performing the assay correctly.
Including data from all 18 laboratories (with the exclusion
of the six suspected contaminated samples from laboratory 16),
accuracy on inclusive and exclusive strains was 99.9 and 96.5%,
respectively. Considering only data from the 15 laboratories
without clusters of aberrant results, accuracy on exclusive strains
was 98.8%.
Recommendations
The ANSR
Salmonella
test was adopted as Official First
Action status for use as a rapid, accurate adjunct or alternative to
biochemical testing for identification of presumptive
Salmonella
spp. isolates.
Acknowledgments
We thank the following collaborators for their participation in
this study:
Dorn Clark and Hondo Dammann, Marshfield Food Safety
(Marshfield, WI)
Jessica Dyszel and Matthew Vross, Richter International
(Columbus, OH)
Nicole Cuthbert and Brian Kupski, Silliker (Crete, IL)
Joe Benzinger, Megan Boyle, and Jonathan Flannery,
Q Laboratories (Cincinnati, OH)
Eric S. Adams, John B. Barrett, Mark E. Berrang, Douglas
E. Cosby, Nelson A. Cox, Jonathan G. Frye, Lari M. Hiott,
Charlene R. Jackson, Steven W. Knapp, and Luanne L. Rigsby,
U.S. Department of Agriculture, Agricultural Research Service
(Athens, GA)
Robert Fuller and Jarrod Van Brunt, Tyson Foods (Springdale,
AR)
Hamoud Alnughaymishi and Andrew Scollon, Michigan State
University, Department of Food Science and Human Nutrition
(East Lansing, MI)
Melanie Corebello and Erika Sai, Unilever U.S. (Englewood
Cliffs, NJ)
Lisa Kuepfer and Jill Stepnitz, Covance (Battle Creek, MI)
Michael Hudgens andWalter Jones, NPAnalytical Laboratories
(St. Louis, MO)
DouglasWaltman and SelenaYork, Georgia Poultry Laboratory
(Oakwood, GA)
Jake Cannon, Benjamin Howard, and Neil Rogman, Certified
Laboratories of the Midwest (Bolingbrook, IL)
Chad Pidgeon and Amy Quenneville, Ben & Jerry’s
(Burlington, VT)
Vikas Gill and Hua Wang, United States Food and Drug
Administration, Center for Food Safety and Applied Nutrition
(College Park, MD)
Cori Flores and Priyanwada Kulkarni, Henningsen Foods
(Omaha, NE)
Table 4. Results by agar medium
Medium
a
Correct
Misidentified Total
BGS
Inclusive
108
0
108
Exclusive
98
5
103
BS
Inclusive
108
0
108
Exclusive
99
4
103
DMLIA
Inclusive
108
0
108
Exclusive
93
3
96
HE
Inclusive
107
1
108
Exclusive
100
4
104
TSA
Inclusive
108
0
108
Exclusive
101
3
104
XLD
Inclusive
108
0
108
Exclusive
101
3
104
XLT-4
Inclusive
108
0
108
Exclusive
69
2
71
Total
Inclusive
755
1
756
Exclusive
661
24
685
a
BGS = brilliant green sulfa agar; BS = bismuth sulfite agar;
DMLIA = double-modified lysine iron agar; HE = Hektoen enteric agar;
TSA = tryptic soy agar; XLD = xylose lysine deoxycholate agar;
XLT-4 = xylose lysine tergitol agar.
118