How do you take a water sample for legionella?

A legionella result is only as good as the sample behind it, and sampling is where self-managed duty holders most often go wrong. The procedure is not difficult, but it is specific: the right bottle, the right outlets, a deliberate choice about flushing, honest labelling and prompt transport. Sampling follows BS 7592, the British Standard for sampling water for legionella, and this page turns it into a bottle-to-laboratory walkthrough.

Before you start: the kit and the plan

Order the sampling kit from a UKAS-accredited laboratory before you touch a tap. It contains sterile bottles dosed with sodium thiosulphate, which neutralises any chlorine or other disinfectant in the water so the bacteria are not silently killed off in transit, plus labels, paperwork and usually prepaid return packaging. Our testing laboratory directory shows which laboratories supply kits and what is included.

Have the sampling plan written down: which outlets, in what order, and whether each is pre-flush or post-flush. Take the outlets from the risk assessment, typically the sentinel hot outlets nearest and furthest from the calorifier, the calorifier drain, the cold tap nearest the storage tank, and any outlet the assessment flagged. Avoid sampling immediately after the system has been disinfected or after a weekly flush, unless the point of the exercise is to check that measure: you want the system in its normal working state.

Pre-flush versus post-flush

These are two different questions and the choice is not optional detail.

Pre-flush (first draw). You take the sample from the outlet without running it first, capturing the water that has been standing in the outlet and its branch. This is the worst-case picture: whatever has grown in that stagnant water ends up in the bottle. Cold outlets are normally sampled this way.

Post-flush. You run the outlet first, a hot tap until it reaches its normal operating temperature, and then sample. This captures the water a user is actually exposed to once the branch has turned over. Hot outlets are normally sampled this way, since hot water at the outlet should reach 50°C within a minute in normal use.

Some plans take both from the same key outlets: a pair of bottles that together show both the standing condition and the in-use condition. Whatever you take, write the flush status on the label and the paperwork, because a count of 800 cfu/L means something different in a first-draw sample than in a post-flush one.

The procedure, bottle to box

1. Wash your hands and put on clean gloves if you have them. Do not touch the inside of the bottle or cap, and do not rinse the bottle: the sodium thiosulphate inside is meant to be there.
2. Do not disinfect, flame or swab the tap. Drinking-water bacteriology often sterilises the outlet first; legionella sampling deliberately does not, because you want the sample to represent the outlet as it really is, including any biofilm the water passes over.
3. For a pre-flush sample, remove any aerator only if your plan says so, open the outlet gently to limit spray, and fill the bottle straight away to the indicated line. For a post-flush sample, run the outlet to temperature first, then fill.
4. Cap the bottle immediately, tighten it, and invert it a few times to mix the neutraliser.
5. Label the bottle at once: outlet identifier matching your records, date, time, pre-flush or post-flush, and the outlet temperature if you measured it. Complete the laboratory paperwork as you go.
6. Pack the bottles upright in the cool bag or insulated packaging with the coolant packs, and dispatch the same day where possible.

Take one bottle per outlet and keep them separate. A composite sample mixed from several outlets cannot be traced back to any one of them, so a high count would tell you there is a problem somewhere but not where, which is the one thing you needed to know.

Transport: the 24 to 48 hour rule

Once capped, the bottle is a tiny, changing ecosystem. The count the laboratory reports should reflect the outlet, not three warm days in the post, so samples need to reach the laboratory within 24 to 48 hours, kept cool and never frozen. Sample early in the week so bottles are not stranded over a weekend, use the prepaid courier or return packaging the laboratory supplies, and do not leave the box in a hot vehicle while you finish the rest of the rounds. If transport will be delayed, tell the laboratory when you book.

The mistakes that invalidate a sample

• Sampling right after disinfection or a flush, which flatters the result.
• Sterilising the tap, which strips out exactly the colonisation you are looking for.
• Unlabelled or swapped bottles, which turn a high count into a guessing game.
• Composite samples, which locate nothing.
• Slow, warm transport, which lets the count drift in the bottle.

When the report comes back, read it against the action levels in legionella CFU levels explained, file it in the log book, and record any action taken. If you have not sampled before, the wider walkthrough in how to test for legionella puts this procedure in the context of the whole exercise, and legionella sampling and action levels covers what the counts mean.