|
| Front page | | Contents | | Previous | | Next |
Coliform bacteria and E. coli in drinking water. Comparison of EU reference method with alternative methods
4 Danish equivalency study on TTC-Tergitol (EN ISO 9308-1:2000) compared to three other methods – natural samples
As concluded in chapter 3 this study on natural samples was based on the results from the spiked samples and included besides TTC-Tergitol two of the five tested alternative methods , supplemented with the Danish reference method until now (MPN, DS 2255). The natural samples in this part of the equivalency study were drinking water samples from public as well as private supplies.
4.1 Materials and Methods
4.1.1 Water samples
The water samples for this part of the study were the routine samples of drinking water sampled by the EPA reference laboratory from public as well as private water supplies for analysis.
The samples were also tested for aerobic colony counts at 22°C and for some samples at 37°C as well according to Danish legislation (DS/EN 6222).
In total 38 samples were analysed.
4.1.2 Methods
The methods chosen for comparison in this part were:
- M1: TCC-Tergitol (EN ISO 9308-1:2000)
- M4: Chromogenic medium (Oxoid CM 1046B)
- M6: Colilert
- M7: MPN with MacConkey broth (DS 2255, 2. ed., 2001).
In this part of the equivalency study verification of presumptive findings with all four methods were verified as described in EN ISO 9308-1:2000 regardless the method used. This should enable a direct comparison of verified findings.
4.2 Results and discussion
Results are shown in table 7 with place of sampling, aerobic colony counts and the results from the comparison.
E. coli was not detected in any of the samples and therefore these results are let out in table 7.
Table 7: Results of analyses of 38 natural samples from public and private supplies for -coliform bacteria (37°C). In none of the samples E. coli were detected.
| Sampling site |
Colony count pr. ml |
Coliform bacteria (37°C) per 100 ml |
| 22°C |
37°C |
DS 2255 |
TTC-Tergitol |
Chromogenic |
Colilert |
| Private supply |
18 |
- |
<1 |
<1 |
1 |
1 |
| Private supply |
47 |
- |
<1 |
<1 |
<1 |
<1 |
| Private supply |
11 |
- |
<1 |
5 |
6 |
4 |
| Public distribution system |
3 |
- |
<1 |
<1 |
<1 |
<1 |
| Public distribution system |
7 |
- |
<1 |
<1 |
<1 |
<1 |
| Boring |
4 |
<1 |
<1 |
<1 |
<1 |
<1 |
| Public distribution system |
1 |
- |
<1 |
<1 |
<1 |
<1 |
| Water plant supply |
<1 |
<1 |
<1 |
<1 |
<1 |
<1 |
| Public distribution system |
13 |
- |
<1 |
<1 |
<1 |
<1 |
| Water plant supply |
4 |
<1 |
<1 |
<1 |
1 |
2 |
| Public distribution system |
<1 |
<1 |
<1 |
<1 |
<1 |
<1 |
| Public distribution system |
240 |
- |
<1 |
39 |
Overgrown |
62 |
| Public distribution system |
8 |
- |
<1 |
<1 |
16 |
25 |
| Public distribution system |
1 |
- |
<1 |
1 |
2 |
<1 |
| Well for flow measuring |
1 |
- |
<1 |
<1 |
<1 |
<1 |
| Public distribution system |
<1 |
- |
<1 |
<1 |
<1 |
<1 |
| Well for flow measuring |
4 |
- |
<1 |
<1 |
<1 |
<1 |
| Water plant supply |
16 |
1 |
<1 |
11 |
17 |
25 |
| Water plant supply |
32 |
3 |
<1 |
<1 |
3 |
3 |
| Public distribution system |
13 |
- |
<1 |
<1 |
1 |
<1 |
| Public distribution system |
7 |
- |
<1 |
<1 |
<1 |
<1 |
| Public distribution system |
10 |
- |
<1 |
<1 |
<1 |
<1 |
| Public distribution system |
10 |
- |
<1 |
<1 |
1 |
<1 |
| Public distribution system |
6 |
- |
<1 |
<1 |
<1 |
1 |
| Industry |
53 |
- |
<1 |
<1 |
<1 |
<1 |
| Private supply |
>3000 |
- |
<1 |
<1 |
15 |
<1 |
| Water plant supply |
54 |
3 |
<1 |
<1 |
Uncountable due to iron |
<1 |
| Private supply |
41 |
- |
<1 |
<1 |
<1 |
<1 |
| Water plant supply |
4 |
<1 |
<1 |
<1 |
<1 |
<1 |
| Water plant supply |
260 |
35 |
<1 |
<1 |
Uncountable due to iron |
4 |
| Private supply |
56 |
- |
<1 |
<1 |
<1 |
<1 |
| Private supply |
<1 |
- |
<1 |
<1 |
<1 |
<1 |
| Public distribution system |
16 |
- |
<1 |
<1 |
<1 |
<1 |
| Public distribution system |
43 |
- |
<1 |
1 |
<1 |
<1 |
| Public distribution system |
9 |
- |
<1 |
<1 |
<1 |
<1 |
| Public distribution system |
6 |
- |
<1 |
<1 |
<1 |
<1 |
| Private supply |
8 |
- |
<1 |
<1 |
<1 |
<1 |
| Public distribution system |
7 |
- |
<1 |
<1 |
1 |
<1 |
Table 7 shows that the Danish reference method DS 2255 was not able to detect coliform bacteria in any of the samples. There was not found any presumptive positive findings of coliform bacteria (any colour changes at all) and therefore no E. coli. This means that it is the method it self that is too insensitive, it is not due to any verification steps.
With the other three methods coliform bacteria were detected and verified in five samples with EN ISO 9308-1:2000, in 11 samples with Chromogenic agar and in 9 samples with Colilert.
In total 15 out of the 38 samples were found positive for coliforms although in eight of the 15 samples the counts were only 1 – 2 cfu pr. 100 ml meaning that there is a large uncertainty on these results and also on other lower count.
Nevertheless there was a natural pattern in the findings with more of the positive samples (eight of 15) positive with two or three methods. For the other seven positive samples cfu was as low as 1 in five of the seven samples. This meant that only two samples were in reality found positive with only one method.
Verification data showed (raw data not shown here) that of 77 presumptive colonies picked from Chromogenic agar only 31 were oxidase negative. This is a verification rate of 40%, where as 100% of the 45 tested positive wells from Colilert were oxidase negative, i.e. verified coliform bacteria.
According to ISO/DIS 17994:2002 the relative differences (RDi%) were calculated for TTC-Tergitol compared to Chromogenic agar (see table 8) and for TTC-Tergitol compared to Colilert (see table 9). Finally the two alternative methods are compared in table 10.
The results are evaluated as a one-sided evaluation according to ISO/DIS 17994:2002 as it is decided to accept an alternative method whenever its average performance is either quantitatively equivalent (D = 10) or higher than the reference method. In fact the conclusion for the study described in this chapter will be the same regardless of the use of one-sided or two-sided evaluation (details for the two-sided evaluations not shown). According to ISO/DIS 17994:2002 the number of paired data with regular counts – not zero counts – should be higher than produced in this study. As this Danish study however is meant to verify other results it is accepted to use the data from natural contaminated samples and therefore with a higher percentage of very low counts.
Table 8: Results of the comparative study of TTC-Tergitol and Chromogenic agar according to ISO 17994.
| TTC-Tergitol per 100 ml |
Chromogenic per 100 ml |
Relative Difference % |
Remarks |
| <1 |
1 |
69,31 |
b |
| <1 |
<1 |
- |
a |
| 5 |
6 |
18,23 |
|
| <1 |
<1 |
- |
a |
| <1 |
<1 |
- |
a |
| <1 |
<1 |
- |
a |
| <1 |
<1 |
- |
a |
| <1 |
<1 |
- |
a |
| <1 |
<1 |
- |
a |
| <1 |
1 |
69,31 |
b |
| <1 |
<1 |
- |
a |
| 39 |
Overgrown |
- |
a |
| <1 |
16 |
283,32 |
b |
| 1 |
2 |
69,31 |
|
| <1 |
<1 |
- |
a |
| <1 |
<1 |
- |
a |
| <1 |
<1 |
- |
a |
| 11 |
17 |
43,53 |
|
| <1 |
3 |
138,63 |
b |
| <1 |
1 |
69,31 |
b |
| <1 |
<1 |
- |
a |
| <1 |
<1 |
- |
a |
| <1 |
1 |
69,31 |
b |
| <1 |
<1 |
- |
a |
| <1 |
<1 |
- |
a |
| <1 |
15 |
277,26 |
b |
| <1 |
Uncountable due to iron |
- |
a |
| <1 |
<1 |
- |
a |
| <1 |
<1 |
- |
a |
| <1 |
Uncountable due to iron |
- |
a |
| <1 |
<1 |
- |
a |
| <1 |
<1 |
- |
a |
| <1 |
<1 |
- |
a |
| 1 |
<1 |
-69,31 |
b |
| <1 |
<1 |
- |
a |
| <1 |
<1 |
- |
a |
| <1 |
<1 |
- |
a |
| <1 |
1 |
69,31 |
b |
a) Deleted according to ISO 17994, 6.1 as the counts with both methods was zero resulting in <1 cfu per 100 ml or as one or both methods gave results other than a count, e.g. “overgrown”.
b) Calculated according to ISO 17994, 6.2.2 as one of the two methods gave count zero. Therefore the constant one (1) is added to these counts before calculation of natural logarithm (ln). Optimal at least 75% of the samples should contain regular count data, which has not been possible here with only 25%.
From the results in table 8 (n = 12) the following values can be calculated:
| Mean |
92,30 |
Accord. to ISO 17994, 6.3 |
| Standard deviation |
100,00 |
Accord. to ISO 17994, 6.4 |
| U |
57,735 |
Accord. to ISO 17994, 6.4 |
| LO |
34,56 |
Accord. to ISO 17994, 6.4 |
| HI |
150,03 |
Accord. to ISO 17994, 6.4 |
From the results in table 8, the calculations above and ISO 17994, 7.3.2 (LO > 0) it is shown that TTC-Tergitol and Chromogenic agar are found to be different in this equivalency study with Chromogenic agar giving the highest counts. From these data Chromogenic agar should then be at least as reliable as TTC-Tergitol. Nevertheless three results for Chromogenic agar were excluded before calculating the equivalency, as colony counts could not be read on this agar due to overgrowth (one sample) and iron residues in the water sample (two samples). Overall it is therefore concluded that Chromogenic agar might give reading problems as also shown with the spiked samples (chapter 3) and Chromogenic agar will therefore not be approved from this Danish study.
Table 9: Results of the comparative study of TTC-Tergitol and Colilert according to ISO 17994.
| TTC-Tergitol per 100 ml |
Colilert per 100 ml |
Relative Difference % |
Remarks |
| <1 |
1 |
69,31 |
b |
| <1 |
<1 |
- |
a |
| 5 |
4 |
-22,31 |
|
| <1 |
<1 |
- |
a |
| <1 |
<1 |
- |
a |
| <1 |
<1 |
- |
a |
| <1 |
<1 |
- |
a |
| <1 |
<1 |
- |
a |
| <1 |
<1 |
- |
a |
| <1 |
2 |
109,86 |
b |
| <1 |
<1 |
- |
a |
| 39 |
62 |
46,36 |
|
| <1 |
25 |
325,81 |
b |
| 1 |
<1 |
-69,31 |
b |
| <1 |
<1 |
- |
a |
| <1 |
<1 |
- |
a |
| <1 |
<1 |
- |
a |
| 11 |
25 |
82,10 |
|
| <1 |
3 |
138,63 |
b |
| <1 |
<1 |
- |
a |
| <1 |
<1 |
- |
a |
| <1 |
<1 |
- |
a |
| <1 |
<1 |
- |
a |
| <1 |
1 |
69,31 |
b |
| <1 |
<1 |
- |
a |
| <1 |
<1 |
- |
a |
| <1 |
<1 |
- |
a |
| <1 |
<1 |
- |
a |
| <1 |
<1 |
- |
a |
| <1 |
4 |
160,94 |
b |
| <1 |
<1 |
- |
a |
| <1 |
<1 |
- |
a |
| <1 |
<1 |
- |
a |
| 1 |
<1 |
-69,31 |
b |
| <1 |
<1 |
- |
a |
| <1 |
<1 |
- |
a |
| <1 |
<1 |
- |
a |
| <1 |
<1 |
- |
a |
a) Deleted according to ISO 17994, 6.1 as the counts with both methods was zero resulting in <1 cfu per 100 ml or as one or both methods gave results other than a count, e.g. “overgrown”.
b) Calculated according to ISO 17994, 6.2.2 as one of the two methods gave count zero. Therefore the constant one (1) is added to these counts before calculation of natural logarithm (ln). Optimal at least 75% of the samples should contain regular count data, which has not been possible here with only 27%.
From the results in table 9 (n = 11) the following values can be calculated:
| Mean |
76,49 |
Accord. to ISO 17994, 6.3 |
| Standard deviation |
112,96 |
Accord. to ISO 17994, 6.4 |
| U |
68,117 |
Accord. to ISO 17994, 6.4 |
| LO |
8,37 |
Accord. to ISO 17994, 6.4 |
| HI |
144,61 |
Accord. to ISO 17994, 6.4 |
From the results in table 9, the calculations above and ISO 17994, 7.2.2 (LO > 0) it is shown that TTC-Tergitol and Colilert are found to be different in this equivalency study, but with Colilert giving the highest counts. This means that Colilert shows results that are at least as reliable as those found with TTC-Tergitol and therefore Colilert can be regarded equivalent to the reference method.
Table 10: Results of the comparative study of Chromogenic agar and Colilert according to ISO 17994.
| Chromogenic per 100 ml |
Colilert per 100 ml |
Relative Difference % |
Remarks |
| 1 |
1 |
0,00 |
|
| <1 |
<1 |
|
a |
| 6 |
4 |
-40,55 |
|
| <1 |
<1 |
|
a |
| <1 |
<1 |
|
a |
| <1 |
<1 |
|
a |
| <1 |
<1 |
|
a |
| <1 |
<1 |
|
a |
| <1 |
<1 |
|
a |
| 1 |
2 |
69,31 |
|
| <1 |
<1 |
|
a |
| Overgrown |
62 |
|
a |
| 16 |
25 |
44,63 |
|
| 2 |
<1 |
-109,86 |
b |
| <1 |
<1 |
|
a |
| <1 |
<1 |
|
a |
| <1 |
<1 |
|
a |
| 17 |
25 |
38,57 |
|
| 3 |
3 |
0,00 |
|
| 1 |
<1 |
-69,31 |
b |
| <1 |
<1 |
|
a |
| <1 |
<1 |
|
a |
| 1 |
<1 |
-69,31 |
b |
| <1 |
1 |
69,31 |
b |
| <1 |
<1 |
|
a |
| 15 |
<1 |
-277,26 |
b |
| Uncountable due to iron |
<1 |
|
a |
| <1 |
<1 |
|
a |
| <1 |
<1 |
|
a |
| Uncountable due to iron |
4 |
|
a |
| <1 |
<1 |
|
a |
| <1 |
<1 |
|
a |
| <1 |
<1 |
|
a |
| <1 |
<1 |
|
a |
| <1 |
<1 |
|
a |
| <1 |
<1 |
|
a |
| <1 |
<1 |
|
a |
| 1 |
<1 |
-69,31 |
b |
a) Deleted according to ISO 17994, 6.1 as the counts with both methods was zero resulting in <1 cfu per 100 ml or as one or both methods gave results other than a count, e.g. “overgrown”.
b) Calculated according to ISO 17994, 6.2.2 as one of the two methods gave count zero. Therefore the constant one (1) is added to these counts before calculation of natural logarithm (ln). Optimal at least 75% of the samples should contain regular count data, which has not been possible here with only 50%.
From the results in table 10 (n = 12) the following values can be calculated:
| Mean |
-34,48 |
Accord. to ISO 17994, 6.3 |
| Standard deviation |
97,15 |
Accord. to ISO 17994, 6.4 |
| U |
56,090 |
Accord. to ISO 17994, 6.4 |
| LO |
-90,57 |
Accord. to ISO 17994, 6.4 |
| HI |
21,61 |
Accord. to ISO 17994, 6.4 |
From the results in table 10, the calculations above and ISO 17994, 7.3.4 it is shown that the results from Chromogenic agar compared with Colilert are found to be inconclusive and more samples should be examined before equivalency could be decided. No further studies will though be made as it was not the aim of this study to compare the two alternative methods with each other.
4.3 Conclusion on natural samples
Chromogenic agar will not be approved as an alternative method to the EU Reference method even though the Chromogenic agar generally gave higher counts than the TTC-Tergitol. It was shown that Chromogenic agar gave some of the same problems with reading the plates as was seen with the spiked samples in chapter 3 and this is the reason for not approving the method.
TTC-Tergitol and Colilert was found to be different in this equivalency study with Colilert giving the highest counts for coliform bacteria. This means that Colilert should be considered at least as reliable as TTC-Tergitol and therefore the methods are regarded as equivalent.
Chromogenic agar compared with Colilert was found to be inconclusive for detection of coliform bacteria in this equivalency study and more samples should be examined before equivalency could be decided.
As E. coli was not detected in any of the samples of natural drinking water conclusion cannot be drawn on this parameter from this part of the practical studies. In the final conclusion results from the spiked samples will therefore be included for this parameter.
| Front page | | Contents | | Previous | | Next | | Top |
Version 1.0 February 2007, © Danish Environmental Protection Agency
|