Professional Comment

Legionella Sampling: Part 3 – Testing Methods & What Next?

By Roy Sullivan / Authorising Engineer [Water], Water Hygiene Centre Ltd (

Methods used for testing Legionella and other water samples
Whilst there are many diagnostics available to a test laboratory – dependent upon the level of analyses required and the type of sample being processed, we will focus on the traditional microbiological methodologies for processing water samples and the associated benefits and limitations of these tests.

With this in mind, new (rapid) diagnostic technologies will also be highlighted and the associated benefits of using such tech.

Traditional microbiology typically refers to the process of using selective agar to determine the presence of bacteria in a given sample.

Therefore, agar may be defined as a combination of nutrients required for a particular organism to grow.

That said, some microorganisms (such as Legionella) are considered fastidious (fussy growers) even when attempted to be cultured on Legionella-selective agar – buffered charcoal yeast extract (BCYE) with cysteine, thus giving them the best chance to grow. This is a problem as other microorganisms (such as Pseudomonas aeruginosa) which are less fussy growers may inhibit the growth of Legionella, which therefore means that additional provisions are required to encourage the growth of the ‘target organism’, which is the one you want to grow!

Such provisions require the use of glycine, vancomycin, polymixin & cyclohexamide (GVPC) agar as these ‘selective agents’ inhibit the growth of non-target organisms, thus assisting with the growth of Legionella bacteria (in accordance with ISO 11731). ISO 11731 references the use of the x3 plate method for the detection and enumeration (counting) of Legionella bacteria (this x3 plate method ensures the best chance of recovering this fastidious pathogen from water samples).

The x3 plate method requires samples to be; directly plated onto selective agar, which are either untreated, treated with heat or treated with acid. These additional (heat and acid) treatments further inhibit the growth of non-target organisms thus helping to ensure a more accurate final result. For this reason it’s important that the test laboratory of choice are able to process such samples in accordance with this methodology.

What Happens Next?
Once samples have been plated, they are placed in an incubator at 36 +/- 2°C and checked 3 times with the final plate reading being on the 10th day of incubation. Therefore, from the day the water sample is processed by the test laboratory, it may take up to x2 weeks to receive a ‘confirmed Legionella result’, this time taken to report highlights the main limitation with this test method.

You can now appreciate why it takes time for the laboratory to report your final results!

A look in to the future…
As we have a duty of care to ensure that individuals are protected from such waterborne pathogens, this has given rise to new technologies which aim to deliver faster results than those obtained through traditional methods. That said, many new technologies have been centred on using a molecular process known as polymerase chain reaction (PCR) and whilst it’s unequivocal that these diagnostics may provide results in a fraction of the time taken by traditional tests, they’re still not without limitations.

The additional benefits of using rapid diagnostics – that utilise PCR for example are that these diagnostics often have a ‘negative predictive value’ (NPV) of 100% which means that this test will correctly identify a negative test result 100% of the time.

Therefore due to the high sensitivity of this test method, there is no chance of receiving a false negative test result. This however can be considered a ‘double edged sword’ as whilst it’s very appealing to be using a test that will always report a negative result correctly, PCR-based methods are traditionally poor at discriminating between living and dead bacterial cells, so whilst the NPV may be 100%, the chance of reporting false positive results may be quite high from previously affected water systems and especially from systems that have been recently disinfected – following chemical, thermal disinfection etc.

Some PCR-based tech utilise pre-filtration stages which aim to mitigate this by only capturing viable cells that will be used in the molecular stage of the test but there are still concerns about whether PCR actually reports viable but non-culturable bacterial cells or whether the tech is still reporting a combination of living and dead cells due to its historical limitations with discriminating between the two!

The other issue with utilising PCR-based technology is that test results are reported in ‘genomic units’ (GU) rather than colony forming units (CFU) (as detailed in HSG274 part 2). Therefore as the action limits for organisms like Legionella are reported in CFU/L, then one must question the relevance of counting GU/L (known as quantitative PCR or qPCR) in the absence of logarithmic data that converts one unit of measure to another.
In summary, whilst PCR-based technology undoubtedly has its limitations at present, the speed and sensitivity of test results generated (in respect of selecting for pathogens like Legionella) should not be ignored and whilst qPCR may also have limited value at present, PCR can still be used as a useful ‘screening tool’ on previously unaffected systems – where detecting the ‘presence’ of an organism is good enough and where many samples require processing quickly.

Following such screening, traditional test methods (plating and colony counting) may then be used to complete ‘confirmatory’ works – where samples have tested ‘positive’ for the target organism! Therefore, using a combination of PCR-based methods and traditional testing may form the ‘blueprint’ for the best way forward with regards to microbiological test work.