---
title: "Cooling tower Legionella sampling: frequency, action levels and what to do at each result"
source_url: https://legionella.io/articles/cooling-tower-legionella-sampling-action-levels/
canonical_url: https://legionella.io/articles/cooling-tower-legionella-sampling-action-levels/
pillar: "Monitoring, Flushing & Sampling"
summary: "What to do when a cooling tower Legionella count lands at each action level, how often to sample, and when a result means your system has lost control"
primary_keyword: "cooling tower Legionella sampling"
date_published: 2026-06-27
date_reviewed: 2026-06-27
author: "Legionella.io editorial team (REMOTE TECH LTD)"
reviewed_against: "HSE L8 and HSG274 guidance"
region: "United Kingdom"
license: "CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). Quote, adapt or republish with attribution to REMOTE TECH LTD and a link to source_url."
license_url: https://creativecommons.org/licenses/by/4.0/
---

# Cooling tower Legionella sampling: frequency, action levels and what to do at each result

A dip slide tells you the tower is broadly clean and a biocide reading tells you the dosing pump is working. Neither tells you whether *Legionella* is multiplying in the water you are aerosolising over a car park or a high street. Only the Legionella sample does that, and it is the one result with a tiered response written against it.

Most responsible persons can read a service sheet. Far fewer know, before the email lands, exactly what they are obliged to do when the count comes back at 80, at 600, or at 4,000 colony-forming units per litre. That gap is where loss of control hides.

This piece sets out how often a cooling tower should be sampled for Legionella and the graded action expected at each result band, so the decision is made before the laboratory report arrives, not improvised after it.

## Why the Legionella sample is the trigger, not the thermometer

A cooling tower runs warm, holds nutrients and throws aerosol. Your daily and weekly checks, biocide residual, conductivity, dip-slide total viable counts, are leading indicators that the control regime is functioning. They are fast and cheap, and you should act on them. But they are proxies.

The Legionella sample is the direct measure. It is cultured by an accredited laboratory and reported as Legionella in colony-forming units per litre (cfu/L). Because culture takes time to grow, the result describes the tower as it was on sampling day, not today, which is precisely why routine sampling has to be frequent enough to catch a drift, and why a bad result demands action even though it is already several days old.

## How often to sample a running tower

For an evaporative cooling system in normal operation, published guidance commonly points to routine Legionella sampling at least quarterly, with the exact frequency set by your risk assessment and written scheme of control [1][2]. Quarterly is a floor for a stable, well-controlled tower, not a ceiling.

Sample more often than the routine interval when the system is new or recently recommissioned, after any cleaning and disinfection, when the treatment programme has been changed, when daily or weekly checks have been out of specification, or whenever there is any reason to doubt control [1]. Verify the precise cadence and triggers against HSG274 Part 1 and your own scheme; treat the figures here as typical guidance rather than fixed law.

Collect the sample to a recognised technique so the result is valid, representative point on the system, correct bottle with neutraliser, proper labelling and prompt delivery to the lab within its stated window [3]. A mishandled sample produces a number you cannot trust in either direction.

## The action levels and what each one means

The widely cited Legionella action-level bands for cooling towers are expressed in cfu/L and carry a graded response. The bands below follow the structure in HSG274 Part 1; confirm the exact thresholds and wording against the current guidance, as they are the reference point an enforcing inspector will use [1].

- **Up to around 100 cfu/L**, the system is generally regarded as under control. Continue the regime, record the result and review at the next routine sample.
- **More than about 100 and up to around 1,000 cfu/L**, the result calls for review. Resample to confirm, check the whole control regime, and investigate why Legionella has appeared. If the count is confirmed and the cause is not obvious, the programme needs corrective action.
- **More than about 1,000 cfu/L**, the system may be out of control. This is the band that triggers immediate action: resampling, shot-dosing or increasing the biocide regime, reassessing the whole control programme, and checking that corrective action worked [1].

Two readings can sit in the same band and still mean different things: a one-off mid-band count on a previously clean tower reads differently from the same count on a tower that has been creeping upward for two quarters. The trend matters as much as the single figure, which is the argument for keeping the results together rather than reading each email in isolation.

## Working a result: the decision flow

When a report lands, run it through the same sequence every time. The standout below is that flow, laminate it next to where the reports get opened.

| Result band (cfu/L, hedged) | First read | Confirm / check | Action | Escalate when |
|---|---|---|---|---|
| Up to ~100 | Under control | Confirm sample was valid and representative | Continue regime; record; review at next routine sample | If trend is rising across samples, tighten frequency |
| ~100 to ~1,000 | Review needed | Resample; audit biocide, conductivity, dip-slide history and recent works | Investigate cause; correct the programme; increase monitoring | If confirmed and cause unclear, treat as loss of control |
| Over ~1,000 | Possible loss of control | Resample; immediate review of all control measures | Shot-dose or increase biocide; investigate cause; take corrective and preventive action | Consider notification duties and specialist/UKHSA advice if illness is linked [4] |

The flow is deliberately blunt. The moment a result sits above the lower band, the question stops being "is this a problem?" and becomes "what is the cause, and what have I done about it today?" An out-of-control tower is an active aerosol source, so the clock on the high-band response is short.

Resampling is not a way to make a bad number go away. It confirms whether the first result was real and whether your corrective action worked, it is a check, not a delay tactic. Keep treating and investigating while you wait for the repeat.

## Reading results when you outsource treatment

If a water-treatment contractor takes your samples, the *legal* duty to act on the result stays with the duty holder and responsible person. The contractor reports the number; you own the response. Read every Legionella report yourself, hold it against the action bands, and confirm in writing what was done at each above-target result. A count that arrives, gets filed and is never acted on is one of the more common ways a tower drifts out of control without anyone noticing.

This is general guidance, not a substitute for your site's control scheme. The sampling frequency, the action thresholds applied to your tower, and the response to any given result must come from a competent, site-specific Legionella risk assessment and written scheme of control, the bands here are the common reference, but your assessor sets what applies to your system and your circumstances [2].

## FAQ

### Is a Legionella culture the same test as my dip slides?

No. A dip slide gives a rapid, on-site total viable count of general aerobic bacteria and is a housekeeping indicator of overall cleanliness and biocide performance. The Legionella sample is a laboratory culture specific to Legionella, reported in cfu/L against action levels. A clean dip slide does not prove the tower is Legionella-free, and the two results answer different questions.

### Where on the system should the cooling tower sample be taken?

Take it from a point representative of the recirculating water the tower is actually aerosolising, following a recognised sampling technique with a neutralised bottle so any residual biocide does not skew the count [3]. Your scheme should name the sample point. Sampling somewhere unrepresentative, or letting the bottle sit warm before delivery, can produce a misleading low or high result.

### A result came back high but the tower seems fine, do I wait for a resample before acting?

Treat a high result as real until proven otherwise. Resample to confirm, but begin the review and corrective action straight away rather than pausing on the strength of the tower "looking fine", the aerosol risk does not wait for the repeat. The resample tells you whether the first reading held and whether your action worked [1].

### Does a count over the top band always mean an outbreak?

No. A high Legionella count means the system may be out of control and is a clear trigger to disinfect and investigate, not evidence that anyone has been made ill. The two are separate. If a clinical case is ever linked to a site, that is handled through the public-health route, and you should take specialist and UKHSA advice rather than relying on environmental counts alone [4].

## What to do next

Find your last three Legionella reports for the tower and lay them side by side. Mark each against the action bands and, for any result above the lowest band, confirm there is a dated record of what was done in response. If a result is filed with no recorded action, that is the gap to close first, before the next sample is even taken.

## Related reading

- [Cooling tower biocide programme](https://legionella.io/articles/cooling-tower-biocide-programme/)
- [A worked cooling tower control scheme](https://legionella.io/articles/cooling-tower-control-scheme-worked-example/)
- [Who controls the cooling tower?](https://legionella.io/articles/cooling-tower-responsible-person-vs-water-treatment-contractor/)
- [Legionella sampling 101](https://legionella.io/articles/legionella-sampling-101-how-and-why-to-test-your-water/)

## Sources

[1] HSE, HSG274 Part 1 (2024), “Monitoring water quality and understanding water treatment analytical reports”, p.42. https://www.hse.gov.uk/pubns/books/hsg274.htm
[2] HSE, ACoP L8 (2013), “Review of control measures: monitoring and routine inspection”, p.18. https://www.hse.gov.uk/pubns/books/l8.htm
[3] BSI, BS 7592:2022, clause 7 (sampling procedures), p.14. https://www.bsigroup.com/
[4] UK Health Security Agency, "Legionnaires' disease: guidance, data and analysis". https://www.gov.uk/government/collections/legionnaires-disease-guidance-data-and-analysis