What a clean legionella sample actually proves (less than you think)

A clean legionella sample proves that one outlet, on one day, at one moment, grew fewer legionella than the lab's detection limit. That is worth having. It is also a good deal less than most people read into it. A negative result is a snapshot from a single point in time, and the thing keeping your building safe is not the sample, it is the temperature regime. HSE's position is plain: temperature and the control scheme are the primary defence, and routine sampling is not required for most well-managed hot and cold water systems. Sampling verifies that control. It does not create it.

The reason a single clean result proves so little is in how legionella behaves at an outlet. Shedding is intermittent and patchy, so the count varies from outlet to outlet and from one date to the next, and the culture method carries a real false-negative rate on top of that. You can sample a colonised outlet on a good day and get nothing back. Two samples from the same tap a fortnight apart can disagree, and neither is wrong, because they caught the outlet on different days. Which is why a negative is fair verification of a working control scheme, and near worthless as a substitute for one.

Pre-flush and post-flush, and why the order matters

The sampling code of practice is BS 7592:2022, and the lab culture behind it runs to ISO 11731. The distinction that trips people up is when you take the sample relative to running the tap.

A pre-flush sample is taken first, before any flushing or disinfection, from the outlet as you found it, unmixed where the fitting allows. Unmixed matters because a blending valve dilutes what the outlet is really holding, so on a mixer tap you want the hot or the cold on its own rather than the tempered blend. It shows the outlet's actual colonisation and is representative of what reaches the person using that tap, which is why it is the sample for routine monitoring. A post-flush sample, taken after running the outlet, tells you about the system feeding it rather than the outlet itself. BS 7592 does not recommend routine post-flush sampling without disinfection, for a simple reason. Once you have flushed, you can no longer tell outlet contamination apart from system contamination, and you have washed away the very thing you came to measure.

So the order is the finding. Sample first, then flush. Reverse it out of habit and the result answers a question you were not asking.

What makes a sample valid

A sample is only as good as the handling, and this is where results go wrong long before the lab ever sees them.

• The bottle is sterile and dosed with sodium thiosulphate, which neutralises any residual disinfectant in the water. Skip that and the biocide keeps killing legionella inside the bottle on the way to the lab, and you collect a false negative by post.
• The sampler is trained and competent in aseptic technique. Contaminate the sample while taking it and the number means nothing in either direction.
• The chain of custody is documented: sample points, times, temperatures and unique bottle IDs, with a copy of the paperwork travelling with the samples into a UKAS-accredited lab.
• The samples travel in insulated containers, kept cool but not frozen, and are analysed within 24 hours, and no more than 48.

Then you wait. A culture result takes around ten days, commonly anywhere from seven to fourteen, because you are growing organisms rather than reading a meter. Use a UKAS-accredited lab, which will work to a detection limit of around 100 legionella per litre or lower.

On sample volume the sources genuinely differ, so treat any single figure with care. UK practice commonly runs to something around a litre for culture, because a larger volume lets the lab concentrate the sample and reach that detection limit, while some routine guidance treats 250 ml as enough and reserves the full litre for outbreak and investigation work. Take it as roughly a litre for routine monitoring, more when you are investigating, rather than a fixed number.

Notice that temperature sits in that chain-of-custody list. It is there because a microbiology result means little without the thermal reading taken beside it. A positive at an outlet you also recorded running properly hot is a different problem from a positive at an outlet sitting lukewarm, and the second is usually a control failure the sample merely confirmed. The sample and the temperature log are read together, or neither says very much.

So when is sampling worth the trouble

If temperature control is the real defence, sampling earns its place as the check on that defence, not the defence itself. It is worth doing where you cannot rely on temperature alone, after remedial work when you want to confirm the fix took, in higher-risk settings and around vulnerable users, and when you are investigating a suspected problem. What it is not is a monthly ritual standing in for keeping the hot water hot and the cold water cold. Sampling everything on a calendar while the temperature regime slips is effort spent proving the wrong thing.

Used the right way, a positive result is not a filing exercise. It is a trigger. A count coming back above the action level means the control scheme has a gap, and the response that follows, from a resample to a possible disinfection, is where the sample earns its cost.

That trigger is the part BlueWave is built to catch. A sample gets logged against the specific outlet it came from, the result attaches to that record when the lab reports back, and a positive is flagged for the reactive response rather than sitting in an inbox. BlueWave records the sample and its result and raises the flag; the UKAS lab does the analysis, and the engineer does the sampling and the remediation. It stores the proof and makes the positive impossible to miss.

The record side matters as much as the science, because a sample result is a compliance record like any other and has to be authenticated to count, the same three tests any digital record has to pass. A positive then runs the same escalation path as a failed temperature reading, and both sit inside the wider monitoring cadence that the risk assessment sets.

Here is the blunt version to keep in your head. A clean sample tells you an outlet was clean enough on the morning you visited. It tells you nothing about tomorrow, and nothing about the outlet down the corridor. The hot water being hot tells you about both.