What happened next… Outbreak investigation team drawn up. Enhanced prospective surveillance to identify other LD cases in the area who had some contact with the hospital. Extensive environmental sampling in the hospital - assess the extent of colonisation. Resample from the assisted baths used by the two cases. Domestic samples taken from the home of the patient. Hospital H site visit Any clinical isolates to be sent to Reference Laboratory - further characterisation & molecular typing. Very localised colonisation (shower in one particular bay of RS ward). Used point of use filters in high risk areas which were legionella positive and all shower/bath fittings. Arjo baths out of use. RS Ward closed except for emergency. Further pasteurisation of main distribution. Hyperchorinated HCWTanks. Flushing of outlets. Hyperchlorinated the system. Clean descale and disinfect showerheads. All water fountains taken out of use.

Enhanced prospective surveillance All positive urine antigen results were reviewed. All admissions with pneumonia were specifically interrogated about hospital contact. GP's were asked to screen for Legionella in patients presenting with flu like illness or pneumonia -who reported hospital contact in the 10 days before they became ill.

Interpretation of environmental legionella results The assisted bath is colonised and the temperature of the water is ideal breeding ground for bacterial growth. Some of the hot tap temperature is not reaching the target minimum of 50°C Introduce control measures whilst the HWS is being investigated to rectify the hot water temperature. Some of the cold water is gaining heat- Look for insulation of pipework, look for plumbing faults. Highlight the importance of monitoring (sampling & temp profile) conventional separate hot and cold taps. This is a very select number of outlets not representing the whole hospital, sampling strategy is to undertake wider sampling. For areas where high Legionella positives have been reported, need to undertake a root cause analysis for that outlet. Showerhead, hoses, TMV-dismantle, descale, clean & disinfect. Replace flexible hoses with solid copper pipe. Resample.

Emergency control measures Disinfection of the water system. Flushing of outlets. Point of Use filters in high risk areas. Monitoring - environmental sampling. Risk assessments: Clinical & Environmental.

Hospital H Site visit To review policies, systems and procedures for Legionella prevention control on site. To review the results of bacteriology, other water quality, and temperature monitoring systems. To interview key personnel regarding current approaches to Legionella prevention and control. To view the hospital site and relevant facilities. Walk round - temperature spot checks at outlets and on flow and return pipes on wards.

Extent of legionella colonisation Only 10 of the initial 155 samples collected contained legionellae. Seven of these were in the immediate area of ward S where Case 1 was located including the shower and a geriatric bath used by the patient.

Microbiological monitoring for L. pneumophila sg.1 Aug Sep Oct Nov Dec Jan Feb Mar <100 19 33 25 23 27 11 8 22 100 - 1,000 13 37 14 10 16 10,000 7 12 9 3 5 4 >=10,000 2 1 ND 375 610 619 604 589 196 244 312 200 300 400 500 600 700 15 20 30 35 40 No. of L. pneumophila positive samples This graph shows results from all samples which were positive for L. pneumophila sg 1. The right hand axis shows the number of ND Results are sorted according to the range of CFU by month Initial samples were focussed within the vicinity of where the patient had stayed and the facilities she may have used. Over the months wider sampling to establish the extent of colonisation. Gradual decline of the No. of samples with high bacterial load. There are still a small number of sample points with >10,000cfu/L recovery of Lpn sg1. Most of the legionella positives were <100cfu/L. Despite the fact that this graph does not show non pneumophila data, there were very few sample points which recovered non-pneumophila Legionella. Looking at the data we have, The proportion of Lpn sg1 positives is small between 10-15%. Recent US paper (Allen et al. 2012) The proposed 30% positivity metric has 59% sensitivity and 74% specificity (ie, a 41% false-negative rate and a 26% false-positive rate). These notable error rates could have significant implications, given that 16 peer-reviewed articles and 6 government guidance documents that referenced the 30% positivity metric as a risk assessment tool. However, the possible consequence of using a percent positivity metric with low sensitivity and specificity is that many hospitals might fail to mitigate when a true risk is present, or might unnecessarily allocate limited resources to deal with a negligible risk. 9.9% 11.2% 8.9% 7.4% 7.7% 15.2% 9.3% 10.3% 7

Environmental investigation Extensive microbiological sampling: L. pneumophila sg1 mAb Benidorm ST42 found in many water samples. Assisted baths used by two patients were found to be colonised with legionella. Site visit to the hospital-look for any obvious risks (from a microbiological point of view) within the system. Risk assessment - engineering investigations were carried out in the hospital immediately following the diagnosis of the index case.

Facilitator notes: Acknowledgements The creation of this training material was commissioned in 2010 by ECDC to Health Protection Agency (UK) and the University of Chester (UK) with the direct involvement of Louise Brown, Janice Gidman, Emma Gilgunn-Jones, Ian Hall (on behalf of the ECDC Legionnaires Disease Outbreak Toolbox Development Group), Tim Harrison, Rob Johnston, Carol Joseph, Sandra Lai, John Lee, Falguni Naik, Nick Phin, Michelle Rivett, and Susanne Surman-Lee. The revision and update of this training material was commissioned in 2017 by ECDC to Transmissible (NL) with the direct involvement of Arnold Bosman and Kassiani Mellou.