Legionella testing

Legionella Testing Occurrences and Trends
G. Pinna B App Sc (Med Tech) B Sc MASM MAIRAH
Manager of Biotech Laboratories Pty Ltd, Brisbane, Queensland.


Statistical Data
The routine testing for Legionella bacteria in water samples from air-conditioning systems has been ongoing in Australia for well over a decade. My laboratory has been collecting data since April 1992. In that time we have seen the Australian Standard method for detection improved to use 5 agar plates and incorporate acid treatment and heat treatment of the sample to try to improve the inhibition of non-Legionella bacteria, increased regulatory control, guidelines and handbooks published, public, medical and industry awareness increased and yet we still have seen a bucket load of Legionnaires' Disease outbreaks in Australia and overseas.

If we look at the period from January 2000 to the end of September 2003, our laboratory processed over 25,000 water samples for Legionella. There is a seasonal variation in the percentage of samples that have detectable levels of Legionella. As you would expect in the summer months the percentages increase, although there is often a slight lag period where will still get a comparatively high isolation rate into early autumn. During that 45-month period we had 2,818 Legionella positive samples, i.e. an isolation rate of 11%. That represents about 14 samples per week that had Legionella at a level of 10 CFU/mL or above. CFU stands for colony forming unit, a scientific term that is a measure of the growth or density of bacterial colonies formed and can best be replaced by the word "bacteria".

Of those 2,818 positive samples the break down of the concentration of Legionella found and whether they were Legionella pneumophila or another species of Legionella is shown in the following table.

 Table 1. Breakdown of 2,818 positive Legionella samples in the period 1/1/2000 to 30/9/2003

Positive Leg

This means that on average once a week we isolated Legionella at a level exceeding 1,000 CFU/mL. This level is generally accepted as the highest risk requiring immediate shutdown and disinfection of the system.


The seasonal variation of the isolation rate of Legionella is shown in the following graph.

  Isolate Leg

Regarding the sampling technique, I would recommend that the free-flowing water be used whenever possible. This is the potential source of aerosols that can carry the bacteria and infect people and therefore should be tested rather than water in dead legs, slime, pipes or condensers where Legionella can be found in higher levels. However, if you had an ongoing problem with Legionella contamination, then sampling at these other sites may give you the source of the problem.
Despite all this data, the value of Legionella testing of water samples is one of the most controversial health issues of building air quality. There is scientific data suggesting that the routine testing of samples from water-cooled air conditioning systems can be a total waste of time. At the same time there is the laboratory data showing detectable levels of Legionella from cooling towers responsible for Legionnaires' Disease outbreaks in Australia and overseas.  

Let's now look at a few of the issues from both sides of the debate.


Arguments against Legionella Testing

A negative result means nothing. I do not agree with this statement in total. By that I mean that I do agree that if you are sampling the free flowing water, as I have recommended for routine testing, a negative result does not indicate that Legionella bacteria are absent (or more accurately, below detectable levels) in the whole water system. Legionella can found in higher levels in dead legs and biofilm/slime. But as it is the free flowing water that is the source of aerosols, then a negative result from these samples does mean something. You have to remember that this is not a chemical test, bacteria are not perfectly even distributed in the water sample and also that we actually culture a total of only 0.23mL of the sample submitted. It's unrealistic to expect that results are consistently representative of a whole water-cooling system.

A positive result means nothing. This is a very emotional issue. If you look at Queensland we get our fair share of Legionella contaminated cooling towers. Since January 2002, my laboratory percentage of water samples positive for Legionella ranged from 2.8% to 20.9% per month with an average of 1 in 10 samples yielding detectable levels of Legionella. And of the positive samples 3% to 15% had Legionella levels in excess of 1,000 CFU/mL. Yet, there has never been a reported outbreak of Legionnaires' Disease linked to a water-cooling system in Queensland. Of course the reverse side of the coin is that Victoria also has Legionella isolations routinely, but has had more than their fair share of Legionnaires' Disease outbreaks. Then there is the fact that other factors affect the likelihood of Legionella in the water causing disease. These include the efficiency of drift eliminators, system position, humidity, air temperature, wind velocity and possibly the presence of amoeba in the water. 

If you maintain a system to AS/NZS 3666 you won't get Legionella in it. Totally untrue. In fact, you would expect to find Legionella to be isolated eventually from any system.

There are no Australian Standards for interpretation of results. With the exception to cooling towers that cannot be shut down this is true. Some states have specific guidelines, others do not. What is required is that a national standard to cover result interpretation. My personal belief is that the acceptable level for Legionella should be increased from less than 10 CFU/mL to less than 100 CFU/mL.
Laboratory results are inaccurate. True, a laboratory can test the same water sample 5 times and get 5 different results. The values can easily be within 1 log phase without us being the slightest bit concerned. Whereas, for the building industry, the difference between a Legionella level of less than 10 (i.e. not detected) and 100 CFU/mL or 400 and 1,200 CFU/mL is highly significant. As stated before as we sample a total of 0.23 mL (0.1mL onto two agar plates and 0.01mL onto three agar plates) and knowing that bacteria are not evenly distributed in water, it is easy to see how varying results will occur.

The variety of results from different laboratories are so scary, you have to think these guys are making them up. True, well not their making the results up part, at least I hope not, but the variety of results of the same water sample sent to multiple labs certainly can and will vary. This is despite that fact that the water samples are going to nationally accredited laboratories all using the same Australian Standard method. I have already mentioned how natural variation will occur, but remember that transport time and temperature can also play a major role in effective final laboratory results. But let's add to that the variables that can occur within the laboratory. The major ones are staff competency and quality of culture media. 
Staff Competency. You can train an inexperienced person to set up Legionella assays unsupervised in about a two weeks. You can train a scientist to read a legionella culture in about twice that time based on reading about a hundred cultures a day. But true expertise only comes with years of experience. When it gets busy, its easy for processing staff, for example, not to check that the pipette has been filled properly perform adding the water to agar plates or for scientific staff to miss suspect colonies or be too lazy to subculture the correct colonies. This will lead to false negative results. A false positive result is realistically impossible. The use of improperly trained, immature, inexperienced and/or poor quality staff will consistently lead to false negative results. 

Culture Media Quality. Just as you can buy most products from a variety of sources and get varying quality based on how much you are prepared to pay, culture media used for Legionella detection is the same. There are the quality products that cost a premium or the super-cheap culture media that are exactly that, super cheap in every way. The poor quality media will grow less Legionella, again resulting in false negative or lower count results. Even with quality media some species of Legionella seem to grow better on one type of agar rather than another.


Arguments for Legionella Testing
Legionella has been detected from cooling towers causing outbreaks.. It can be said that a positive result does not mean there is a necessary health risk. But at the same time, the bacteria responsible for many outbreaks of Legionnaires Disease can be traced back to a single water source through Legionella testing. The obvious statement would be: If that particular tower was tested prior to the outbreak occurring and Legionella was isolated then the outbreak would have been prevented. Legionella levels can jump up fairly rapidly and may also stay at detectable levels for months on end. The answer to the statement is that it is possible that routine monthly testing of towers may have detected the Legionella and prevented the outbreak. However, let's not forget that it is almost certainly the fault of poor maintenance and control of the system that caused the high levels of Legionella and therefore the outbreak and not the absence of testing.

If you don't look you don't find. This is certainly true. Since there are action levels that have been recommended and the absence of detectable Legionella in a system is what you are paying your water treatment company to achieve, then you should routinely monitor for compliance. 

Showing all due diligence and care. What if an outbreak was to occur from a system under your control and monthly testing was not performed? It could be argued in court that the industry standard was to perform monthly Legionella tests and that you could have prevented the contamination from reaching high levels by monthly monitoring. It is even feasible that building owners could take legal action against maintenance personnel (whether internal staff or external contractor) because they had not recommended monthly testing. I would strongly recommend all maintenance people to document a letter to the building owner strongly recommending routine monthly Legionella and heterotrophic plate count testing. If the owners refuse this option, get it in writing and sleep with it under your pillow.

What you can't measure, you can't mange . It is impossible for management to state that they are taking all reasonable precautions to prevent a Legionnaires' Disease outbreak if they have no idea of the microbial water quality.


How to Select a Laboratory for Testing
Obviously the laboratory must be NATA registered and should also be AS/NZS ISO 9002 accredited for their management system. But please don't tell me that being NATA accredited ensures quality of results. If you look at the results of the national NATA Legionella proficiency program I can assure you that a wide variety of acceptable and unacceptable results are obtained. It is even feasible for laboratories using cheap poor quality media to keep a few packs of the "good media" to be used when the proficiency testing samples come around, not to mention doing the NATA sample in triplicate or, as was the case recently in Queensland, sending isolates to the State Health laboratory to get them to perform speciation.  

My recommendation to you is to find out from your laboratory how many years experience each of the scientific officers reading Legionella cultures have and what proof of staff competency is performed. You should also have on file a copy of their in-house method for all tests that they are performing for you; this must include a statement of all validated deviations from the standard method. When you are shopping around for prices, it may not be always true, but if a laboratory stands out from all others as a discount king you are probably looking at someone cutting some corners to make ends meet. This could be by staff wages (usually associated with a high staff turnover and decreased experienced staff), cheaper media and less quality control. We once lost a large customer to a competitor because we found Legionella in more samples than they did! The bottom line is that you may save money but is the risk worth it? If you want quality you may well need to pay for it!