ADVANCEMENT OF THE SCIENCE
TABLE 1
FIGURE 1
Occurrence of Presumptive Extended-Spectrum Beta-Lactamase (ESBL) E. coli by Assay Method
Box and Whisker Plot of Presumptive Positive Extended-Spectrum Beta- Lactamase (ESBL) E. coli Concentrations
Metric
Measurement
Membrane Filtration (CFU)
ESBL CBT (MPN)
GEL ESBL (CFU)
Geometric mean
CFU or MPN/100 ml
1.12 0.69
1.00 0.64
1.05 0.69
3
Log 10 CFU or MPN/100 ml
SD
Nondetect Minimum
%
62%
71%
70%
2
Log 10 CFU or MPN/100 ml Log 10 CFU or MPN/100 ml
-0.30
-0.40
-0.30
2.03
1.87
1.69
Maximum
1
Note. CBT = compartment bag test; CFU = colony forming unit; MPN = most probable number.
0
isolates were confirmed using this method. Lastly, isolates were evaluated for ESBL posi- tivity using the Tricycle protocol confirma- tion criteria. Using this technique, 64.4% of the isolates were found to be ESBL-resistant (scored as ESBL-positive). By method, 55.6% of the membrane filtration isolates were resis- tant, 76.9% of the ESBL CBT isolates were resistant, and 60.2% of the GEL ESBL isolates were resistant. Discussion The results of this study indicate that the three methods detect similar concentrations of ESBL E. coli in surface water. Based on a Freidman test analysis, there was no statisti- cally significant difference between any of the methods in detecting median concentrations. This result is similar to other evaluations of the E. coli CBTs (Stauber et al., 2014; Wang et al., 2017) and initial evaluations of the ESBL CBT method (Appling et al., 2023). In the surface water samples collected for our study, there were low concentrations of ESBL E. coli , ranging from nondetect- able amounts to approximately 100 CFU or MPN per 100 ml. The rate of nondetects was between 62% and 71% of each sample set depending on the assay type (Table 1). Other studies conducted on surface waters also showed high numbers of nondetects for ESBL E. coli (Blaak et al., 2014), indicating that input into these waters—including wastewa- ter, surface runoff, and animal waste—might not be consistent throughout the year and
could be dependent on temperature, weather, and other factors. In our confirmation analysis of the ESBL E. coli isolates, each of the three methods resulted in a similar number of positive iso- lates, specifically the membrane filtration method ( n = 117), ESBL CBT ( n = 91), and GEL ESBL ( n = 98). Each isolate was then subjected to a series of confirmation tests including a secondary streak plating on ESBL antibiotic-impregnated TBX agar, an indole test, and then antibiotic resistance testing according to the Tricycle protocol. Each method was able to detect a similar percent- age of positive ESBL-resistant E. coli , adjusted for the number of isolates analyzed. Of the 117 isolates detected from the mem- brane filtration method, 65 (55.6%) were con- firmed by the methods previously described and identified as ESBL-positive E. coli . For the ESBL CBT method, 70 of the 91 initial isolates identified (76.9%) were confirmed as ESBL E. coli. For the GEL ESBL method, 59 of the 98 isolates (60.2%) were confirmed as ESBL E. coli. These resistance percentages are higher than the percentages reported in the initial evaluation of the ESBL CBT (Appling et al., 2023); however, our results are similar to other evaluations of isolates in surface waters in North America (Haberecht et al., 2019). Although not directly considered in the methods comparison presented in our evalu- ation, previous published work has compared CBT to standard laboratory methods, includ- ing the membrane filtration method consid-
-1
MF CBT GEL
ered here (Bain et al., 2012). When factor- ing in the cost of agar and petri dishes, the estimated cost of the standard CBT method is approximately the same amount per sam- ple. The ease of use for the CBT, however, which allows a user to process the sample without the need for an incubator, special- ized pipette, or other laboratory equipment, greatly reduces the cost of this test itself. For limited-resource settings, field tests such as the CBT or GEL method offer an opportunity to test for pathogens that might not be con- sidered within standard monitoring practice due to a lack of available facilities or labora- tory equipment. The three methods we evaluated are com- parable in terms of the detection of ESBL E. coli concentrations and the overall confir- mation of ESBL E. coli isolates. There were several limitations. Specifically, while the ESBL CBT detects MPN concentrations and membrane filtration and GEL ESBL methods detect CFU concentrations, we compared Note. Lines denote median log 10 -transformed concentrations. Boxes denote first and third quartile log 10 concentrations. Whiskers denote 95% confidence limits for log 10 concentration values. Mean values are designated by the dashed line. CBT = ESBL compartment bag test; CFU = colony forming unit; GEL = GEL ESBL; MF = membrane filtration; MPN = most probable number.
10
Volume 87 • Number 4
Powered by FlippingBook