Introduction

Keynote

Systemic effects include rigors, diarrhoea, muscle breakdown, pulse temperature dissociation, hyponatremia (low sodium in the blood) and other effects. It is very difficult to distinguish LD from other pneumonias. Even highly qualified clinicians missed 11% of LD patients in hospitals per an investigation. The wide spectrum of symptoms is complicated by immunosuppression factors which invite other infections and indeed even a superinfection may result, eg a LD plus TB co-infection.

LD is a very destructive pneumonia and many patients are not able to properly remodel their lungs afterwards; hence there may be ongoing dizziness, loss of energy, loss of memory, loss of consciousness, fatigue, weakness, myalgia (muscle pain), and even arthritis as a result.

Microbiology and Pathogenesis

M. Machner (Boston): Recruitment of protozoan vesicles by Legionella begins in only 15 minutes after invasion. There is full development in 14 hours and host destruction in 16 hours.

H. Shuman (New York): Legionella uniquely does the same set of tricks in mammalian cells (macrophages) as in protozoan amoebae.

K. Heuner (Germany): The flagellum facilitates mobility and assists biofilm formation. There is a link between flagella regulation and virulence. Regulation and gene expression depend on temperature, viscosity, osmolarity (ie osmotic pressure = fluid flow across membranes) and nutrition.

Activated macrophages are very efficient at killing Legionella, As the T-cells migrate to the site of infection they recognise the flagella and proteins involved signal activators to respond (T-cells regulate our immune response and signal our bacteria-infected cells to kill their intracellular parasites).

Mice, rather than guinea pigs, tend to be used for lab experiments.

C. Buchrieser ((France): Legionella uses a very high proportion of its proteins to target a protozoan invasion. It has versatility, an ability to exploit host functions, and extreme redundancy. For the L. pneumophila gene sequence see http://genolist.pasteur.fr/Legiolist/ . Genome projects are ongoing now for L. longbeachae and L. anisa. These tasks may be completed in 12 months.

Epidemiology

C. Joseph (UK): Over the last 10 years 37 countries have joined the European surveillance scheme. 27,250 cases have been reported including 2,250 deaths. France had 1,200 cases in 2004; similar number for Spain. Although there appears to be an 8% fatality rate, the real rate may be much higher. For more information see www.ewgli.org .

J. van Wijngaarden (Netherlands): The European aim is to bring water hygiene to the same standard as for food hygiene.>

C. Navarro (Spain): There was an outbreak in Murcia in July 2001 with 449 confirmed cases and total estimated cases were 696. There were only 6 deaths. Source considered to be a cooling tower called CT3 at a hospital called Hospital H. The tower was poorly maintained. Weather conditions: thermal inversion, low wind speed (under 9 kph), high ambient temperatures and RH. This was the largest outbreak that had occurred worldwide. The tower was replaced with air-cooled plant. Measures introduced include cooling tower registrations.

G. Tallis (Australia): Final statistics for the Melbourne Aquarium outbreak (April 2000) were 125 cases including 4 deaths. 25% of the cases were not hospitalised. Another outbreak at a Melbourne suburb called Thomastown in October 1998 involved 18 cases. Investigation of these and other smaller outbreaks has shown the median Legionella count to be 2,400 cfu/mL; most of the outbreaks involve small cooling towers for which the counts are invariably over 1,000 cfu/mL.

T. Harrison (UK): The Barrow outbreak involved 179 cases including 7 deaths. Legionella count in the cooling tower believed to be the source was 100,000 cfu/mL.

P. Aavitsland (Norway): The scrubber at Borregard was believed to be the source of an outbreak in May 2005 involving 55 cases including 10 deaths. Distance of aerosol travel was 20 km. The scrubber looked like an incubator! Water was 40 C. There was no disinfection at all. Water was recirculated at the rate of around 4 to 5 cu.m. per hr. (about 1 to 2 L/s).

N. Tran Minh (France): In Nov 2003 there was an outbreak in Harnes in northern France with 86 confirmed cases, including 18 deaths. Cases were up to 12 km away from the source which was considered to be a cooling tower although the waste water sludge basin associated with the tower also may have had a role. No workers at the plant contracted LD.

B. Diederen (Netherlands): At the Dutch flower show outbreak there were 188 patients. This outbreak showed that the sensitivity of the UAT depends greatly on the severity of the illness.

Microbiology and Taxonomy

M. Pedro-Botet (Spain): Copper-silver ionisation of hospital water since 1999 has effectively controlled Legionella.

A. Flieger (Germany): Proteins from Legionella cells attack molecules on the surface of host cells before invading and attacking the molecules inside. They then multiply and prepare the escape route. The key trigger is the PlaA enzyme. The process could be called adhesion followed by invasion.

H. Bruggemann (Germany): Before release from the protozoan host, Legionella is able to prime itself ready for the next invasion. Growth is enhanced at higher temperatures.

D. Raoult (France): If Legionella survives the amoeba then it will probably survive the mammalian macrophage which is not as strong. The Mimivirus discovered by T. Rowbotham in 1992 is a giant virus. It cycles through amoebae and is as large as several bacteria, is visible on the Gram stain, and mimics bacteria. Mimivirus = Mimicking Microbe. It has RNA and is found on ordinary lab hardware.

F. Cohan (Connecticut, USA): Different Legionella species appear to be on different evolutionary tracks. They are quite ecologically distinct. There is an evolutionary chain involving selection, purge of diversity, single survivor, and then diversity again. There are ecoclusters with some survival similarities. There are 492 mip genes in the total genus. (The mip protein is a virulence factor necessary for establishing infection and intracellular survival within macrophages as well as within its protozoan host. It is found in all Legionella strains. Mip = Macrophage Inflammatory Protein).

R. Ratcliff (Australia): There are 50 formally recognised species but it is now harder to differentiate new ones as what works when there are only 10 species is insufficiently precise when there are 50 and new definitions are needed. There are at least 30 and probably 40 more waiting for full characterisation.

Environmental Microbiology

R. Bentham (Australia): L.pneumophila survival in sea water is best at 15C and non-existent at 35C but the sea water had low salt levels due to outfalls.

S. Surman (UK): Bromine found to be ineffective in spa pools; chlorine at 2ppm is needed. When there is an outbreak there is always a reason for Legionella to be out of control.

B. Fields (CDC, Atlanta): Characterisation in biofilms:

• Legionella persists in building water systems by colonising biofilms
• Legionella can’t form biofilms without other microorganisms
• The bacteria may be released from biofilm to cause disease
• Some outbreaks are caused by massive descalement of biofilm
• The mechanism for biofilm colonisation by Legionella is unique
• The natural mechanism of Legionella colonisation and release from biofilm is not understood

Legionella responds to signals from other microorganisms comprising the biofilm. Dead Legionella released in a spa pool after bromine decontamination overnight caused an outbreak of Pontiac fever.

S. Berk (Tennessee): A field study looked at 40 natural waters and 39 cooling towers as well as 21 other types of water-based mechanical plant, hot tubs, spas, etc to see if amoebae present were infected with Legionella. It was found that 21 of the cooling towers were positive (including 5 novel strains; only 2 towers were L. pneumophila), but only 3 of the 40 natural waters and 3 of the 21 industrial samples were positive.

G. Allestam (Sweden): In Dec 2004 there was an outbreak caused by an aeration pond at a pulp and paper plant in Sweden. Count was 1,000,000 cfu/mL of Lp1 but some other Lp serogroups were present and other species too. Cases were many kms away but some came within 100 metres of the plant. Environmental sampling showed 3 of 12 cooling towers were positive at the plant and that 24 of 42 aeration ponds were positive as well as 28 of 38 biosludge samples which also presented lots of foam. The researchers concluded that:

• Legionella growth is stimulated by multiple factors in Biological Treatment Plants (BTP)
• Counts vary over time
• Simple but sensitive sampling methods are needed
• Cooling towers had a limited impact on the BTP findings and counts were low.

Risk Assesment and Prevention

R. Bentham (Australia): There are special problems presented in risk management as:

• Legionella is not normally distributed within water systems as are other heterotrophs
• Populations in environmental waters don’t follow lab growth patterns
• Log increases in populations are common, and insignificant
• Perceived “high” numbers are abnormal events (numbers may fluctuate rapidly)
• Complexity of the “representative” sample is an issue.

There are also problems associated with dose-response.
At outbreaks, characteristics vary with the routes of infection and types of sources, eg

• cooling towers – explosive, high numbers, wide area
• spa pools – explosive, moderate numbers, geographically contained
• HWS – prolonged, low numbers, may be extremely localised.

J. Lee (UK): On biocides for HWS. Copper-silver ionisation works with HWS and is best in soft water but the data is much less scientific than for experiments with cooling towers. It is not always acceptable to add ionisation to a particular system.

For Chlorine dioxide there are claims but not much supporting evidence; such a system failed at the SS Ocean Monarch cruise ship where Legionella was detected even when the Chlorine dioxide was at a concentration of over 100 mg/L and caused skin to be burnt. Provisional guidelines are underway for this biocide and the concentration needs to be around 1 to 2 mg/L but must be ongoing (continuous). The recommended concentration of only 0.5 mg/L is insufficient even after superchlorination.

Use of monochloramination failed to stop the outbreaks that have occurred with Thames Water in use. There was a case of LD at Rushcliffe Golf Club arising from an instantaneous HWS although it arose from a long spur pipe from the heater. The heater was at the Male toilet and the spur pipe ran to the Ladies toilet where the case occurred.

On materials of construction, latex flexible hoses encourage growth, but copper has been shown to be biocidal. These plastic hoses were involved in a serious outbreak at a hospital in the UK.

J. Stout (Pittsburgh): Should use evidence-based medicine, ie let the evidence be our guide. The Wadsworth Veterans hospital has abandoned use of hyperchlorination. Ionisation is the best for hospital HWS systems per a long-term survey of 16 hospitals. Chlorine dioxide has been found to be inadequate even after 2 years of application.

The first water sample from testing is usually a good indicator of the system condition but a swab is useful too.

W. McCoy (Illinois): In risk management we should be employing HACCP (Hazard Analysis and Critical Control Points) as it is process-driven. Environmental pathogens are controlled by process. Eliminating biofilm is useful but the key is control. The process includes use of operational and maintenance procedures. The seven principles of HACCP are always qualitative and are:

1. Systematically analyse the hazards using flow diagrams.
2. Identify the critical control points.
3. Establish limits to CCP, eg receiving, storing, heating, disinfection.
4. Establish control and monitoring procedures (distribution, waste).
5. Establish the criteria for action.
6. Record keeping.
7. Validation, i.e. data (quantitative evidence) that shows the control limits have prevented, eliminated, or at least reduced the hazard. Then verification is needed as an independent confirmation that the RMP is being implemented correctly and is periodically reassessed. Validation is highly important.

Next Conference

Paris. October 2009. Contact cbuch@pasteur.fr .