How is water sampled for legionella and what do the results mean?

Legionella sampling means collecting water in a sterile bottle and sending it to a laboratory that grows, or cultures, any legionella bacteria present. It is not a routine activity for most well-controlled water systems. It is used where temperature control cannot be assured, in higher-risk systems, or to investigate after a problem.

This page goes deeper on the sampling procedure and on reading the numbers than our general legionella testing page. If you only want the difference between temperature monitoring and laboratory work, start there. If you have a sample to arrange or a laboratory report in front of you, read on.

When sampling is actually needed

For the majority of domestic-scale hot and cold water systems, temperature is the control that keeps legionella in check, and monitoring that temperature is how you demonstrate control. See legionella temperature control for the regime. Routine sampling adds cost without adding much, because a healthy temperature record already shows the water is being kept out of the growth range of roughly 20°C to 45°C.

Sampling earns its place in narrower circumstances: where the system cannot reliably hold hot water above 50°C or cold below 20°C; where a treatment such as chlorine dioxide or copper-silver ionisation is the primary control and needs verifying; in higher-risk premises with vulnerable occupants; and to confirm that remediation has worked after a positive result or an incident. The risk assessment, not habit, decides whether and where to sample.

How a legionella sample is taken

Each sample comes from one specific, documented outlet. Composite samples mixed from several taps are avoided because a positive result cannot then be traced back to a source. The volume is at least 1 litre, collected in a sterile bottle that contains a small amount of sodium thiosulphate. The thiosulphate neutralises any residual disinfectant so the count reflects the water at the moment it was sampled, rather than continued kill inside the bottle on the way to the lab.

Two kinds of sample answer two different questions, and a competent sampler often takes both:

• A pre-flush, or first-draw, sample is taken the instant the outlet is opened, before any flushing or disinfection. It captures the water sitting in the fitting and the pipework nearest to it, so it tells you about the outlet itself.
• A post-flush sample is taken after the outlet has been run and the fitting surface-disinfected, representing water drawn from the wider system rather than the tap. Hot water is sampled once it has reached its normal running temperature.

Comparing the two shows whether a problem is local to that outlet, for example a contaminated shower head or spray tap, or systemic across the network. The sealed samples are kept cool, out of sunlight and not frozen, and reach the laboratory within 24 to 48 hours. There they are cultured by the ISO 11731 method and incubated for up to ten days, which is why results take roughly a fortnight to come back. The bacteria have to grow before they can be counted.

What the action levels mean

The laboratory reports a colony-forming unit count per litre (CFU/litre) for each legionella species found, or "not detected". For hot and cold water systems, HSG274 Part 2 sets three response bands:

• Under 100 CFU/litre: consistent with a system under good control. Continue the monitoring and control scheme.
• 100 to 1,000 CFU/litre: review the control scheme and resample. If only a minority of outlets are positive the cause may be a single local issue; if most are positive the system may be colonised at a low level and needs a wider review.
• Over 1,000 CFU/litre: take remedial action, resample straight away, review the risk assessment, and consider whether disinfection is needed and whether occupants should be informed.

Source: HSE HSG274 Part 2. These bands apply to hot and cold water systems only. Cooling towers (HSG274 Part 1) and healthcare premises (HTM 04-01) are assessed more stringently.

The practical steps that follow an exceedance are set out in what to do if legionella is detected.

A result is a snapshot, not a certificate

This is the point most pages skip, and it matters. A sample measures a single litre of water, from a single outlet, at a single moment. Legionella numbers rise and fall with temperature, use, and season, so a clean result taken on a good day does not prove the system was safe last week or will be next week. Equally, one high count is not proof of an outbreak. It is a flag to investigate.

Because of that, sampling verifies control. It does not create it. The backbone stays the same: keep hot water hot and cold water cold, keep the little-used outlets moving through flushing, and record it all in the written scheme of control. A laboratory result is one data point read alongside that record. On its own it neither condemns nor clears a system.

Cooling towers and healthcare are different

Do not read the hot and cold water action levels across to every system. Cooling towers follow HSG274 Part 1, are sampled far more often, typically at least quarterly and frequently monthly, and any legionella is treated seriously because a tower disperses aerosol outdoors over distance. Healthcare premises follow HTM 04-01, and in augmented care areas, where patients are most vulnerable, any detection is significant rather than tolerated up to a threshold. The numbers on your report only mean what they mean for the type of system they came from.

Whatever the system, the outlets you sample, the method, and the frequency should all flow from your risk assessment and be recorded in the log book. That programme sits inside the duties framed by ACOP L8 and the detail of HSG274. Sampling is a tool those documents point you to in defined circumstances, not a substitute for the control scheme they require.