Evaluation of E. coli and Enterococcus Measurements Using Different Testing Methods at Recreational Beach  (1 to 12 Aug 2025)

By Ng Wun Jern
Emeritus Professor
School of Civil and Environmental Engineering
Nanyang Technological University, Singapore

Background

1Typically, there are several standardised testing protocols that are used to assess a particular physical, chemical or biological parameter of a water body. Standard Methods for the Examination of Water and Wastewater, jointly published by the American Public Health Association (APHA), is commonly used as a reference for water-quality testing and each sub-protocol may have slight differences procedures, culture media or reagents, incubation conditions, detection limits and quality-control requirements. The suitability and performance of a method may also depend on factors such as the water matrix, expected bacterial concentration, salinity, turbidity, suspended solids, background microbial populations, the presence of inhibitory substances, and sample storage and holding conditions. Accordingly, different water matrices, including coastal water, inland surface water, raw water and wastewater, may influence bacterial recovery and method performance.

2This article evaluates the results obtained for the testing of Escherichia coli (E. coli) and Enterococcus in recreational beach waters by three accredited laboratories, using the traditional culture-based method and the rapid commercial testing method. This further assesses the factors which could contribute to the differences in results, such as   methodological differences, inter-laboratory variability, or matrix effects specific to marine water environments, thereby informing evidence-based decisions for future water quality monitoring protocols at recreational beaches.

3Water samples were collected from eight sampling locations at a recreational beach between 1 and 12 Aug 2025. Each sample was tested for both E. Coli and Enterococcus using traditional culture-based methods (APHA 9222H for E. Coli and APHA 9230C for Enterococcus) and IDEXX methods (Colilert and Colilert-18 for E.Coli; Enterolert for Enterococcus). The testing was conducted by three labs as follows:

*Colilert-18 produces results in 18 hours compared to Colilert’s 24 hours. Colilert-18 is more resistant to interference from high salt content due to its optimised formulation of its reagent (proprietary info), making it more applicable for testing on marine water samples.

Methodology          

4For E. coli and Enterococcus, the following analyses were conducted to assess the comparability of culture-based method against the IDEXX methods, as well as the statistical difference between the same methods when used at different laboratories.

(a)   Analysis of differences between testing methods

For analysis comparing across 3 groups, the Kruskal-Wallis test was applied to determine if the results showed statistically different distributions.

i.     Culture vs Colilert 18 vs Colilert – If there is any difference in the testing methodologies

ii.    For Enterococcus, Culture (by Lab A) vs IDEXX Enterolert (by Lab B) were compared to check if there is any difference in the testing methodologies.

(b)   Analysis of differences between testing laboratories

For analyses comparing 2 groups i.e. same test method across 2 labs (Lab B and Lab C), Wilcoxon signed-rank test is applied for paired data analysis.

i.     Lab C vs Lab B – If there is any difference between accredited labs.

5Spike recovery tests were also conducted by Lab A and Lab C to evaluate the accuracy of the different methods for detecting E. coli in marine water. Sterilised (autoclaved) seawater was used the test matrix to eliminate background bacterial interference, after which known concentrations of E. coli (ATCC reference strains) were then added (spiked). The percentage of E. coli recovered, referred to as recovery, reflects how effectively each method quantifies E. coli under marine conditions.

Results

6Results on E. coli to understand the differences in testing methodologies or accredited laboratories in IDEXX testing.

i. Difference in testing methods (Culture vs Colilert 18 vs Colilert)

Across all three groups, the culture-based method and Colilert-18 (both Lab B and Lab C) consistently yielded lower counts compared to Colilert, and the small p-values (<0.05) indicated statistically significant differences among the methods. Pairwise comparisons further confirmed that the different test methods produced significantly different results for the same samples.

Culture (Lab A) vs IDEXX (Lab B)	Culture (Lab A) vs IDEXX (Lab C)
Pairwise Comparison p-value Significant (Yes/No) Lab A vs Lab B Colilert 3.5e-12 Yes Lab A vs Lab B Colilert-18 3.4e-12 Yes Lab B Colilert vs Lab B Colilert-18 3.5e-12 Yes	Pairwise Comparison p-value Significant (Yes/No) Lab A vs Lab C Colilert 3.6e-08 Yes Lab A vs Lab C Colilert-18 7.0e-08 Yes Lab C Colilert vs Lab C Colilert-18 3.6e-08 Yes

ii. Differences between testing laboratories in IDEXX testing (Colilert and Colilert-18)

When comparing IDEXX Colilert-18 and IDEXX Colilert results between the two laboratories (Lab B and Lab C), Lab C consistently reported higher E. coli counts than Lab B. For Colilert-18, Lab C’s counts ranged from 40–386 compared to 64–201 for Lab B, while for Colilert, Lab C’s counts ranged from 209–10,462 compared to 40–799 for Lab B. In both cases, the small p-values (<0.05) indicated that the differences between laboratories were statistically significant. We reviewed the laboratory procedures from the labs and verified that there is no difference in their protocols.

Colilert-18 (Lab B vs Lab C)	Colilert (Lab B vs Lab C)

7Results on Enterococcus were examined to understand differences in testing methodologies and accredited laboratories. Comparing Lab A’s culture-based method with Lab B’s IDEXX Enterolert, the Enterolert method consistently reported higher Enterococcus counts. Lab A values ranged from <1 to 33, whereas Lab B values ranged from <10 to 226, with most results at <10 or 10. Although paired tests yielded a small p-value (p < 0.05), indicating statistical significance, the actual differences were minor in magnitude. Overall, the two methods produced results that largely converged, suggesting that the observed statistical significance reflects methodological sensitivity rather than a substantial divergence in Enterococcus levels.

iii. Spike Test Analysis

8The percentage of E. coli recovered, referred to as recovery, reflects how effectively each method quantifies E. coli under marine conditions. As shown in the plot below, all recoveries were below the 1:1 line (which denotes 100% recovery), indicating there was no 100% recovery at any of the spiked concentration. Different marker colours denote the 3 different test methods. In general, Colilert-18 (i.e. green markers) consistently achieved higher recovery (35-85%), the culture-based method (i.e. blue markers) yielded the lowest recovery (15-32%) across most spiked concentrations, while Colilert (i.e. orange markers) showed intermediate recoveries (25–59%). A control test in sterile deionised water (indicated as reverse triangles) showed that Colilert-18 still performed best (59% recovery), while Colilert (9%) and culture (25%) were less accurate, indicating that marine water matrix effects alone do not account for the poor performance of the latter methods.

Discussion

9Variability in E. coli detection in marine waters is influenced both by the choice of methodology and by differences in laboratory and sample handling practices. The marine environment itself adds complexity  ^{[1], [2]}: high salinity, sunlight, and background microbial flora all stress E. coli cells, often driving them into a viable but non-culturable (VBNC) state. In such cases, culture-based methods systematically undercount, as VBNC cells fail to proliferate on selective media despite remaining viable. In contrast, enzyme-based methods such as IDEXX Colilert-18 and Colilert detect β-glucuronidase activity, enabling the enumeration of metabolically active but non-culturable cells. However, this advantage comes with the drawback of cross-reactivity, as halophilic marine bacteria such as Vibrio spp and non-halophilic bacteria such as Providencia sp. can mimic the targeted enzymatic activity ^[3][4][5], thereby inflating counts. Consequently, IDEXX assays may overestimate E. coli, whereas culture methods may underestimate them.

10Comparisons between Colilert-18 and Colilert have further highlighted sources of uncertainty. Both assays are based on proprietary substrates that contain similar components, but differences in incubation period (18 versus 24 hours) and performance with marine matrices have been noted. Whether these findings extend to tropical marine waters remains unclear, underscoring the need for context-specific validation. This distinction also highlights a broader methodological contrast that IDEXX assays are designed around substrates that actively “select for” target organisms, while traditional culture methods rely on broader substrates and chemical inhibitors to suppress non-targets. This fundamental difference may explain the consistent tendency of IDEXX to generate higher counts relative to culture-based enumeration.

11Inter-laboratory variation introduces another critical dimension of variability. Even when using the same methodology, results can diverge due to operational practices. Sample handling factors, including time between collection and analysis, adherence to cold chain, storage conditions, dilution practices, warming prior to incubation, and the condition of equipment, consistent mixing protocols (e.g., adequately shaking sample bottles before sub-sampling), can significantly influence outcomes. Manpower experience and laboratory workflow may also lead to systematic biases, where a given laboratory consistently produces higher or lower results than its peers. Such patterns were evident in the current dataset, where IDEXX consistently yielded higher counts than culture, and one laboratory’s results were systematically elevated relative to the other despite both employing the same kit. This suggests that laboratory-specific handling, preparation, and process rigor are non-trivial contributors to data variability.

Conclusion

12Taken together, our findings underscore the importance of harmonising methodologies and strengthening quality assurance in the enumeration of E. coli and Enterococcus in marine waters. Statistically significant differences across methods and laboratories, coupled with variable recovery rates from spike experiments, illustrate how both assay design and laboratory practices can shape results. Performing replicate analyses, conducting inter-laboratory proficiency testing, and enforcing rigorous sample handling protocols are critical steps to minimise random and systematic error. At the same time, greater transparency on the performance characteristics of proprietary assays, particularly Colilert-18 and Colilert in tropical marine water matrices, remains necessary to resolve persistent uncertainties. Ultimately, acknowledging the inherent limitations of each method while working towards methodological standardisation will be essential to ensure that marine water quality data are robust, comparable, and meaningful for public health protection.

[1] Byappanahalli, M. N., Nevers, M. B., Korajkic, A., Staley, Z. R., & Harwood, V. J. (2012). Enterococci in the environment. Microbiology and Molecular Biology Reviews, 76(4), 685–706.
[2] Anderson, K. L., Whitlock, J. E., & Harwood, V. J. (2005). Persistence and differential survival of fecal indicator bacteria in subtropical waters and sediments. Applied and Environmental Microbiology, 71(6), 3041–3048.
[3] Davies, C. M., S. C. Apte, S. M. Peterson, and J. L. Stauber. 1995. Possible interference of lactose-fermenting marine vibrios in coliform β-d-galatosidase assays. J. Appl. Bacteriol. 78:287-393.
[4] Geissler, K., M. Manafi, I. Amoros, and J. L. Alonso. 2000. Quantitative determination of coliforms and Escherichia coli in marine waters with chromogenic and fluorogenic media. J. Appl. Microbiol. 88:280-285.
[5] Pisciotta, J. M., Rath, D. F., Stanek, P. A., Flanery, D. M., & Harwood, V. J. (2002). Marine bacteria cause false-positive results in the Colilert-18 rapid identification test for Escherichia coli in Florida waters. Applied and environmental microbiology, 68(2), 539–544.