Controlling risks from Legionella in our water systems is essential in any workplace – to comply with health and safety law, but more importantly, to protect users of those water systems from potentially fatal infections. When we apply control measures to control these risks, the law requires that checks are conducted to ensure that those measures are effective and that they remain so. As many will know, these checks include those that seek to ensure that water temperatures are maintained, as far as possible, outside of the recognised growth temperature range of the bacterium, that water systems are kept clean, and that water is kept moving through systems to avoid creating conditions that would encourage proliferation to dangerous levels. These checks are, by and large, quite simple to perform and the results of those checks, if performed effectively, can be relied upon.
In addition to these physical checks on the condition and operating characteristics of water systems, it is common practice, in many workplace settings, to take water samples for laboratory analysis to see whether the bacteria can be detected. There are advantages to this type of testing, not least that if a positive test is returned from the lab, it provides compelling evidence that the controls that are being applied have not been effective and that more needs to be done. However, what happens if the test is negative? Does this mean that there are no legionellae in the water system and that it is safe?
The current, most commonly used method used by laboratories to detect legionellae is BS EN ISO 11731:2017, often cited, erroneously, in my opinion, as the ‘gold standard’. I say this because the method described in the standard allows variations in the procedures that can dramatically affect the limit of detection – sometimes, by orders of magnitude.
Some variations include filtering or centrifuging the original sample to concentrate any legionellae that might be present. When this is done, it is inevitable that some legionellae will remain adhered to the filter or the centrifuge tube, resulting in fewer bacteria available for onward processing. The standard also allows pre-treating the water sample with heat and/or acid to reduce the number of other faster growing bacteria that are also likely to be present alongside the legionellae and which, if the sample was not treated, would overgrow the agar plates before any legionellae would have a chance to form colonies that can be seen and counted. Whilst legionellae tend to be more tolerant to high temperatures and low pH compared to their commensal counterparts, they are not entirely immune from their effects. This means that the number of viable legionellae that end up on an agar plate that could potentially form colonies is reduced in relation to the number in the original sample. Indeed, studies have shown that, when compared to the highly sensitive polymerase chain reaction (PCR) testing method, recovery rates for legionellae are around 50% and can be as low as 16%.
And it’s not all about the testing method, the way we sample water systems can affect the likelihood of detection. Conventionally, sample volumes of between 0.5 and 1.0 L of water are taken from water systems for testing, which may represent only a tiny fraction of the total amount of water within the system. In addition, if you consider that the vast majority of aquatic bacteria, including legionellae, tend to be associated with biofilms fixed to surfaces, then only a relatively few of these will be in a planktonic state in the water and available for capture during the sampling process. All in all, the way we usually sample water systems tends to reduce the likelihood of detecting the bacteria once the sample has been sent to the laboratory.
There is no denying that sampling and testing for legionellae is an important part of many monitoring programmes, but care is needed in the interpretation of results. If Legionella are detected in a laboratory test, the number reported is likely to be an underestimate of the number that were originally present in the water sample, and almost always an underestimate of the number that were present in the water system sampled. If no Legionella are detected, this is by no means a guarantee that the bacteria weren’t present in the sample or system.
Back to top