Background

Legionnaires’ disease (LD) is a severe form of pneumonia with a case fatality of 5–10% [1, 2]. The disease is caused by Gram-negative Legionella spp. bacteria, ubiquitously found in freshwater environments and soil. The bacterium is facultative intracellular and replication in amoeba is likely the predominant mechanism for its proliferation. This interaction with amoeba plays an important role in the persistence and release of Legionella spp. from its environmental reservoirs [3, 4]. Transmission to humans occurs through inhalation of aerosols or aspiration of water containing Legionella spp. In the lung, Legionella spp. is phagocytosed into alveolar macrophages, where it replicates intracellularly [5]. Human-to-human transmission is a rare exception [6].

In Switzerland, LD is notifiable to the Federal Office of Public Health (FOPH) [7]. Similar to trends observed in other European countries [8], notification rates for LD in Switzerland continue to rise. In 2021, the notification rates reached a new high of 7.8 cases per 100,000 population [2]. About 70–80% of all reported LD cases in Europe, including Switzerland, are community-acquired, about 15–20% of cases are travel-associated, and only about 5% are nosocomial acquired. [2, 9]. Additionally, the majority of LD cases are occurring sporadically, in contrast to outbreaks or clusters. To date, numerous sources including showerheads, dental units, cooling towers, and fountains have been linked to community-acquired Legionella spp. infections (CALD) [10, 11], yet little is known about their contribution to the overall disease burden [12, 13].

Estimating the impact of infection sources on the overall disease burden is difficult. In order to draw any significant conclusions on the contribution of a potential infection source to the disease burden, a sizable proportion of LD patients must be screened as LD remains relatively rareFootnote 1 [14]. Regional variability in notification rates additionally suggests that infection sources might differ between regions [15, 16], hampering the generalisability of results to different geographic areas. To link a LD case to an infection source, genomic comparison of Legionella spp. isolates recovered from patients with LD and from the environment is required. Such molecular epidemiological investigations are resource-intensive and challenging for multiple reasons: First, clinical Legionella spp. isolates are recovered from only about 5–10% of patients in routine surveillance [2, 8]. Second, the ubiquity of the Legionella bacteria and the variable incubation period of 2–14 days for LD [5] requires consideration of multiple potential infection sources for a single LD case [17]. The incubation time may also create an inherent delay of up to 14 days between the time a patient is infected and the time environmental samples can be collected. The delay between infection and environmental source investigation might be further prolonged based on reporting timelines set for case notifications by public health authoritiesFootnote 2 [7]. This prolongation of the period between the time of infection and the investigation of the environmental source may reduce the chances to successfully recover the disease-causing Legionella spp. strains from a suspected infection source. Finally, the recovery of Legionella spp. isolates may depend on the chosen sampling approaches (e.g. exact sampling location but also procedures) and sampling time points as the detachment of bacteria from biofilms and their release from amoeba may vary over time [18, 19]. Additionally, culture isolation of Legionella spp. from environmental samples is labour intensive. It requires careful selection of culture plates and pre-treatment conditions prior to plating (e.g. filter concentration, heat treatment) to optimise growth conditions for Legionella spp. and minimise overgrowth of plates by competing organisms. As a result, molecular source attribution of sporadic CALD are primarily reported in single case studies. From such studies is difficult to conclude on a source’s contribution to the overall disease burden [10, 11].

Conducting a combined case–control and molecular source attribution study allows to address some of the challenges outlined above. The case–control study design enables the exploration of various (including transient) host, behavioural, and environmental exposure risk factors for CALD [20,21,22]. Data obtained from case–control questionnaires can inform the sampling of potential environmental infection sources, in turn, facilitating molecular source attribution [23, 24]. For now, combined case–control and molecular source attribution studies in Europe focused primarily on urban settings or were part of outbreak investigations and, thus, did not investigate any regional variability of infection hazards [23,24,25]. The potential of combined molecular and epidemiological approaches to investigate infection sources for sporadic CALD cases at a national and general population level has not yet been realised.

Herein, we present the study design of a national case–control and molecular source attribution study (SwissLEGIO). The study aims at investigating risk factors and possible exposure sites for community-acquired, mainly sporadic, LD cases across Switzerland. Engaging with a network of 20 participating university and cantonal hospitals, and collaborating closely with the Swiss National Reference Centre for Legionella (NRCL), the project creates a framework that enables timely, nationwide recruitment of patients with LD, and facilitates the collection and processing of clinical Legionella spp. isolates. Together with the Legionella Control in Buildings (LeCo) research consortium, environmental source investigations are conducted within days after case detection.

Methods

Study design and objectives

This research comprises of a one-year, prospective, national case–control study applying whole genome sequencing (WGS) to link LD patients to potential exposure sources in Switzerland. Host, behavioural and exposure risk factors are investigated by conducting interviews with newly diagnosed LD patients (cases) and healthy control subjects. For patients, further parameters on the clinical, radiological and laboratory characteristics, the clinical case management, disease severity, and health outcomes are extracted from electronic medical records. For a subset of cases and controls, clinical and environmental Legionella spp. isolates are collected and sequenced (Fig. 1).

Fig. 1
figure 1

Study design for the national case–control and molecular source attribution study on Legionnaires’ disease in Switzerland, SwissLEGIO. The environmental sampling and sample analytics are conducted in collaboration with the LeCo Consortium

Full size image

The objectives of the SwissLEGIO study are: (i) To identify host, behavioural, and environmental risk factors for LD, (ii) To attribute infection sources to LD cases by comparing clinical and environmental Legionella spp. isolates using WGS, (iii) To assess the genome sequence of Legionella spp. differing in virulence and to identify potential traits of more virulent strains, (iv) To assess strain diversity and concentration of Legionella spp. in standard household and other environmental samples,Footnote 3 (v) To explore the illness experiences of patients with LD, their health-seeking and long-term quality of life, and (vi) To describe clinical, laboratory, and radiological characteristics of LD and the patient’s clinical case management.

The study involves multiple governmental and research stakeholders. Since its inception, the FOPH, the NRCL, and the Federal Food Safety and Veterinary Office (FSVO) are involved as advisory and strategic planning partners. For the implementation, we closely collaborate with a hospital network consisting of 20 university and cantonal hospitals, the NRCL, the Institute of Medical Microbiology (IMM) at the University of Zurich, and the LeCo consortium led by Eawag.Footnote 4

Study setting, recruitment process and participation eligibility

Cases of Legionnaires’ disease

The study includes newly diagnosed CALD patients from all of Switzerland over a one-year period to account for seasonal and meteorological impacts on infections. Patients are recruited through a hospital network representing a significant proportion of diagnosed LD patients (the network collectively reported about 55% of all LD cases between 2018 and 2020) (Fig. 2). The decision to recruit through hospitals and the selection of the participating hospital sites were informed by previous research: in Switzerland, diagnostic testing for Legionella spp. is mainly limited to the hospital setting and, therefore, most reported cases are identified at the hospital. In outpatient care, patients with pneumonia are primarily treated empirically or are referred to the hospital for further clinical and diagnostic evaluations [26]. An in-depth analysis of LD notification data in Switzerland was used to identify hospitals notifying the most LD cases and also showed that LD notification rates regionally differ across Switzerland [2], highlighting the importance to include patients from all seven greater regions (NUT-2 level).

Fig. 2
figure 2

SwissLEGIO operational flowchart for the recruitment and the collection of data from patients with Legionnaires’ disease: A Overview of the SwissLEGIO hospital network, B Data collection upon enrolment, C Collection and analytics for clinical samples as part of routine case management. BAL Bronchoalveolar lavage, NRCL National Reference Centre for Legionella, WGS Whole genome sequencing, UZH University Hospital Zurich

Full size image

For participating hospitals, individualised recruitment procedures were developed to ensure that the study is embedded optimally in each hospital’s existing workflows and in order to help minimise the risk of missing any admitted patients with LD and to optimise efforts to obtain clinical Legionella spp. isolates from LD patients as part of their routine clinical case management. In brief, the central study team at the Swiss Tropical and Public Health Institute (Swiss TPH) is immediately informed by the hospital’s clinical laboratory in case of a positive diagnostic Legionella spp. test result. The central study team then coordinates with the hospital’s appointed study physician or the attending physician on the pre-assessment of the patient’s eligibility for participation (the eligibility criteria are summarised in Table 1) and the subsequent enrolment. Written study-specific informed consent is obtained by the study physician, a study nurse or the central study team before the questionnaire-based interview is conducted (Fig. 2). The study does not interfere with the case management of enrolled patients.

Table 1 Summary of eligibility criteria (inclusion and exclusion) for participation in SwissLEGIO
Full size table

Controls

One control per enrolled case is recruited from the general Swiss population. Controls are selected from a dataset based on the national census list, which comprises a random population sample. The dataset is provided by the Swiss Federal Statistical Office [27]. A control matched to a LD case for age (± 5 years), sex, and location of residence (district level/“Bezirksebene”) is chosen and contacted by e-mail or postal mail as soon as a case has been enrolled. Following the written invitation, the study team assesses the control’s eligibility (Table 1), ability and willingness to participate in the study by phone. Informed consent from controls is obtained prior to the interview.

Sample size calculation

Calculation of the sample size was performed using Epi Info™ 7 (Centers for Disease Control and Prevention, USA). We consider the ubiquitous nature of Legionella spp. in the environment and assume that 60% of controls are exposed to a risk factor during the period of potential risk exposure [23, 28]. Therefore, a sample size of 205 cases and 205 controls is required to detect an odds ratio (OR) of 2 with 90% power and alpha = 0.05. We adjusted the size of the hospital network to reach the required sample size within one year.

Data collection and piloting

For cases, data and biological samples are collected at three different time points: (i) the treating physician obtains clinical samples suitable for Legionella-specific culturing prior to enrolment of the patient in the study, (ii) after informed consent is obtained, a questionnaire-based interview on potential risk exposures is conducted and electronic medical records are reviewed, and (iii) thereupon environmental samples from potential risk exposure sites are collected for a subset of cases. Preference for the environmental sampling is given to cases from whom clinical Legionella spp. isolates are available. For controls, the questionnaires-based interview is conducted and environmental samples are collected from a subset of controls matched to a LD case (for whom environmental samples were also collected) (Fig. 1).

Data and biological sample collection were carefully piloted in a two-step approach: in a first step, between October 2020 and October 2021, direct recruitment of patients with LD through the hospital was tested in collaboration with the University Hospital Basel. Moreover, by interviewing newly diagnosed LD patients, the manageability and comprehensibility of the questionnaire-based case–control interview were assessed. In a second piloting step, from March to June 2022, the participant invitation process, the electronic data collection tools, the coordination of the case–control interview, and the subsequent environmental sample collection, shipment, and analysis were pretested with healthy volunteers. Additionally, data collectors and laboratory staff of the central study team were trained on the data collection and laboratory processing of environmental samples. Finally, the manageability and comprehensibility of the interview completion guidelines and study-specific standard operating procedures (SOPs) on participant recruitment, sampling and sample analysis were tested.

Clinical samples

Clinical samples suitable for Legionella-specific culturing (such as sputum, Bronchoalveolar lavage (BAL), tracheal or bronchial secretion or pleural fluid in case of pleural effusion) are collected by the treating hospital physician as part of the routine clinical management of the patient and, hence, prior to patient enrolment. To enhance and promote the isolation of Legionella spp. from LD patients, the participating hospitals reviewed and, if required, refined their standardised diagnostic procedures to ensure clinical samples for Legionella-specific culture are collected promptly for suspected or confirmed LD cases; the collection of clinical samples can either be triggered by a positive Urinary Antigen Test (UAT) for Legionella or may occur prior to the LD diagnosis. After LD is confirmed, hospitals’ clinical laboratories initiate a Legionella spp.-specific culture on charcoal-based agar. The isolates obtained are subsequently sent to the NRCL as described in the guidelines for notifiable infectious diseases from the FOPH [7]. If a clinical laboratory cannot perform Legionella-specific culturing, clinical samples can be sent directly to the NRCL for further processing (i.e. culturing and serotyping). Upon enrolment, the study team enquires if Legionella spp. could be isolated from clinical samples and obtains written informed consent from the patient for the use of the Legionella spp. isolates in the study. Legionella spp. isolates are sent from the NRCL to the IMM for WGS analysis (Fig. 2).

To maximise the number of clinical Legionella spp. isolates that can be obtained in the hospitals’ routine assessments, treating physicians are encouraged to collect clinical samples for Legionella-specific culture from every patient, irrespective of whether antibiotic therapy has already been initiated. We expect to obtain clinical samples suitable for Legionella-specific culturing for about 50–60% of all patients (since LD is often associated with dry cough, we assume insufficient sputum for some patients) and to successfully isolate Legionella spp. from about half of these samples [5]. We anticipate to analyse one Legionella spp. isolate per patient [29].

Case–control questionnaire and patient records

Upon completion of the informed consent, the central study team or a study nurse conducts questionnaire-based interviews with LD patients and controls. The questionnaire is based on information from published LD data collection tools (either for routine assessment [30, 31] or other case–control studies [21, 23]) and current literature on risk factors and exposure sites for LD. Swiss federal stakeholders in Legionella spp. control namely the FOPH, the FSVO, the Federal Office of Energy (SFOE) and collaborating researchers from the LeCo consortium were consulted for inputs on the questionnaire design.

The questionnaire consists of 20 sections and focuses on the 14 days before the onset of illness for cases and the same (matched) time period for controls. The questionnaire covers potential predisposing host risk factors for LD (e.g. age, sex, co-morbidities), potential behavioural risk factors for LD (e.g. regularly showering at sports facilities, gardening habits) and investigates exposure to potential environmental infection sources (e.g. housing water installation, public artificial water sources, natural water sources). For cases, the questionnaire further covers the illness experiences of patients with LD and their health-seeking (Table 2). Pretested interviews lasted approximately 60 min and were well received by patients and healthy control volunteers in content, length, and flow.

Table 2 Structure of the SwissLEGIO case–control questionnaire for Legionnaires’ disease
Full size table

For cases, parameters on the patients’ clinical case management and the disease severity are extracted from the electronic medical records. The parameters include the medical history, the timespan between onset of symptoms and admission to the hospital, length of hospital stay, CURB-65 parameters [32], radiological findings (e.g. consolidation, crazy paving, bronchial wall thickening, pleural effusions), ICU admission and length of ICU stay, disease progression within 48 h after admission, performed diagnostics, laboratory parameters and prescribed treatment.

Environmental samples

Environmental source investigation is triggered if either a clinical Legionella spp. isolate is available or the Legionella-specific culturing is confirmed to be ongoing at the time of case enrolment. Environmental samples and information on residential building and water installation are collected for these cases and their matched controls.

Up to six environmental samples from approximately 75 cases within 14 days following the questionnaire-based interview are collected (Fig. 1). Up to five of these samples are standardised water samples collected in the patient’s home: (i) first flush (1 L, mix of cold and warm water) kitchen tap water (ii) first flush (1 L, mix of cold and warm water) shower head water from the most-used shower, (iii) sequential sample from the hot water line collected at the most-used shower/bathtub,Footnote 5 (iv) sequential sample from the cold water line collected at the shower/ bathtubFootnote 6 of the most-used shower, and (v) first flush (1 L, mix of cold and warm water) from the second-used shower. The detailed procedures for water sampling are found in the supplementary file). Prior to this sampling, the participant is instructed to refrain from using the taps for four hours. In addition to the standard samples, up to two samples are collected from other likely environmental risk exposure sites reported in the patient interview. These exposure sites are sampled from private locations (e.g. garden hose, water dispenser or humidifier) or public locations (e.g. spas, car wash facilities, decorative fountains, air-conditioners or cooling towers of hotels or supermarkets in proximity of patient’s residency, permissions provided). For a subset of approximately 50 healthy controls, only standard household water samples are collected (Fig. 1).

Environmental samples are cultured and the number of colony-forming units (cfu) of Legionella spp. in the original water sample are estimated according to ISO 11731 guidelines. For the ISO culturing of standard household water samples, we prepare three plates per sample using the following pre-treatment conditions: filtration only, filtration plus heat treatment, and filtration plus acid treatment. Culture plates are regularly checked for growth for one week and up to three suspected Legionella spp. colonies of each morphology are culture-confirmed by direct plating of the colony on charcoal agar plates with and without L-cysteine. Isolates showing no growth on plates without L-cysteine are considered Legionella spp. (Fig. 3a). All culture-confirmed Legionella spp. isolates will be characterised by MALDI-TOF mass spectrometry (MS) (to differentiate strains) and agglutination tests (to differentiate serogroups of L. pneumophila). All Legionella spp. isolates recovered from potential infection sources of cases that are matching the strain and/ or serogroup of the case’s clinical Legionella spp. isolate are sent to the IMM for further strain characterisation using WGS (Fig. 3b). Finally, flow cytometry (for total cell count), IDEXX Legiolert (quantification of culturable L. pneumophila), and digital PCR (quantification of the ssrA gene for Legionella spp., mip gene for L. pneumophila and the wzm gene specific to L. pneumophila serogroup 1) are performed for quality control (Fig. 3c). The digital PCR protocol was developed by adapting the qPCR assay from Benitez and Winchell [33].

Fig. 3
figure 3

Overview of the laboratory analytics pipeline for the isolation and characterisation of environmental Legionella spp. strains from standard household and other environmental samples: A Summarises the isolation and enumeration of Legionella spp. according to ISO 11731, B Culture-confirmed Legionella spp. isolates are characterised by MALDI-TOF MS and agglutination tests and are selected for WGS, C For quality control, flow cytometry, digital PCR and Legiolert are performed for all samples. cfu colony-forming units, L-Cys L-Cysteine, MS mass spectrometry, WGS whole genome sequencing, SG serogroup, mpn most probable number

Full size image

For additional samples from potential exposure sources of patients, sample processing approaches are assessed on a case-by-case basis in consultation with researchers and external research partners from the LeCo consortium. The characterisation pipeline for such isolates will remain the same as for the standard household water samples. Based on rough estimates from the literature [23, 34], we expect to sequence a total of 750 environmental Legionella spp. isolates recovered from potential exposure sources of cases (Fig. 1).

Data management

Data are collected on standardised electronic Case Report Forms (eCRF) using the data collection software Open Data Kit (ODK, getodk.org). Forms are identified by subject IDs. Automated validation tools in the eCRF check for data completeness and plausibility during data entry. During data collection, the data collected is continuously checked by the research team for completeness, plausibility, and accuracy. Additionally, random source data verification is performed. Data are stored on a secured network drive accessible only to authorised study team members. Data on the network drive are backed up regularly, according to Swiss TPH institutional policy. Radiological images are coded and securely shared between hospitals via a secure data exchange platform, and the images are securely stored on two password protected hard drives.

Quality control measures for the analysis and storage of biological samples are performed according to the study laboratories’ routine standard operation procedures.

Statistical methods and analysis

Epidemiological analysis

Cases and controls are characterised in terms of demographics, illness experience (only cases), health-seeking (only cases), and co-morbidities. Crude OR for LD will be calculated by running univariable logistic regressions on single risk factors. Based on results of the univariable logistic regression and biological or epidemiological plausibility, variables will be subsequently selected for a multivariable (unconditional) regression to calculate adjusted OR (aOR). The population attributable fraction (PAF) is calculated for each statistically significant risk factor of the multivariable model as the difference of observed cases and expected cases in absence of the risk factor. The analysis will be conducted with the statistical software R [35]. Potential exposure sites of cases will be geocoded using geographic information systems (GIS) to assess regional distributions of LD cases and to identify clustering of potential infection sources.

Additionally, analysis of radiological imaging is performed independently by at least two experienced radiologists. All chest X-rays or CT scans are evaluated blinded from clinical and microbiological information for lung involvement, distribution (e.g. upper versus lower lobes, uni- versus multilobar), radiological patterns (e.g. consolidation, ground glass opacities, cavitation, nodules, crazy paving, bronchial wall thickening), pleural effusion, and lymphadenopathy.

Analysis of biological samples

WGS is performed using the IMM’s internal ISO accredited (ISO/IEC 17,025) sequencing workflow and analytical pipeline for the characterisation of the Legionella spp. strains. This workflow and the analytical pipeline are already in use for the characterisations of Legionella spp. isolates currently performed for the NRCL [7]. For library preparation, the Illumina NextFlex assay is used and sequencing is performed batch wise at a NextSeq 1000i with 150nt paired end sequencing. After sequencing, the data is quality controlled, raw reads are assembled, and analysed using SeqSphere (Ridom) and the CLC workbench. Only genomes with an average minimal coverage of 40-fold or more will be further evaluated. The sequence type (ST) of all isolates will be determined and the genomic relatedness will be visualised with a series of phylogenetic tools such as core genome multilocus sequence typing (cgMLST) e.g. as neighbour joining tree or as a single nucleotide polymorphism (SNP) tree. Figure 4 illustrates such visualisation for a Legionella spp. outbreak investigation conducted in Basel: allelic differences between L. pneumophila isolates shown in the figure are based on a published cgMLST scheme for L. pneumophila using 1′521 allelic loci [36].

Fig. 4
figure 4

Environmental source of L. pneumophila. Isolates labelled with A and B are from two air conditioning cooling towers. Isolates with spikes are human isolates from the same seasons. Circles with numbers are human isolates from previous years. The small numbers between the circles indicate the number of allelic differences between two isolates (data from Egli-lab, IMM)

Full size image

All successfully sequenced strains will be contextualised with previously sequenced isolates from the Swiss database and with global available sequences from public data repositories such as the National Center for Biotechnology Information (NCBI). In addition, we will compare the human and environmental isolates using a bacterial genome wide association study approach. We will link clinical phenotypes (e.g. LD, disease severity) to potential enriched sequence types, genes, k-meres, or SNP in isolates causing invasive disease. We aim to identify genes and annotate potential functionalities linked to the clinical phenotypes. All sequenced genomes are shared according to FAIR principles [37] on the Swiss Pathogen Surveillance Platform (http://www.spsp.ch). The usage of the platform further eases data exchange between the different research partners [38].

Strengths and limitations

The SwissLEGIO study enrolls patients with LD from all seven greater regions of Switzerland. To our knowledge, this is the first national case–control and molecular source attribution study that is conducted outside of an outbreak setting in Switzerland. The study design innovates in (i) minimising the timespan between LD symptom onset and enrolment of cases in the study, (ii) enhancing and promoting the collection of clinical Legionella spp. isolates, and (iii) ensuring a high sensitivity of the molecular source attribution approach. All three aspects are crucial to ensure a successful linking of clinical isolates to an infection source. The direct recruitment of cases through an established hospital network instead of the Swiss National Notification System of Infectious Diseases (NNSID)—as in previous studies [20, 21]—significantly reduces the timespan between the patient’s diagnosis and enrolment in the study. When piloting the recruiting process, the time between patient’s hospitalisation and case notification to the study team averaged at three days. This is significantly shorter than the Swiss legal requirements to report an LD case within seven days to the NNSID [7]. The recovery of clinical Legionella spp. isolates was a major limitation in environmental source investigations in previous studies [17, 39]. By directly recruiting LD patients through the hospital network, and by exchanging closely with hospital partners on an ongoing basis, we believe to address this challenge by promoting and facilitating the collection of clinical Legionella spp. isolates.

We ensure a high sensitivity of the molecular source attribution approach by applying WGS, which has a strong discriminatory power between different Legionella spp. strains and, therefore, allows a direct comparison of environmental and clinical isolates [39]. During LD outbreak investigations, WGS analysis was successfully used to trace clinical Legionella spp. strains in the environment [25, 34, 39]. Yet, such outbreak investigations also highlighted the complexity of developing sensitive environmental sampling, culturing and isolate selection strategies to account for high Legionella spp. strain diversities in environmental samples [25, 34]. The challenges of obtaining and subsequently selecting appropriate environmental Legionella spp. isolates for WGS in a streamlined manner and the applicability of WGS for the investigations of sporadic LD cases remain largely unexplored. For the SwissLEGIO study we developed streamlined processes for the collection and processing of standard environmental household and other environmental samples and carefully implemented Legionella spp. isolates characterisation measures that will inform the selection of isolates for WGS analysis. Finally, the genomics of Legionellae is very complex: genetically highly similar Legionella spp. strains from different parts of the world have been described without apparent epidemiological link and studies have shown that also recombination events must be considered when interpreting similarities of WGS data between different clinical and environmental Legionella spp. strains [39, 40]. Epidemiological context information is, therefore, essential for a reliable interpretation of observed similarities between different environmental and clinical Legionella spp. strains and, hence, for infection source attribution. A key strength of the SwissLEGIO study is that it combines both epidemiological metadata and WGS data.

Data collected during the questionnaire-based interviews might be subject to reporting biases. The selective memory effect of participants (recall bias) was accounted for by the design of the questionnaire: during the interview, the participant is guided in a structured manner through a wide range of potential exposure sources. To improve recall, participants are interviewed as soon as possible after the LD diagnosis is confirmed. Using an established hospital network, the time between diagnosis, interview and environmental sampling of risk exposures can be kept short. Additionally, any cases for which the diagnostic assessment occurred more than seven days before the study team is notified are excluded. The number of environmental samples collected and isolates analysed is limited due to financial and other resource constraints. Some non-standardised environmental samples may not be collected if access is difficult and/or permissions cannot be obtained.

Future perspectives and impact on policy

Large-scale studies combining molecular and epidemiological methods are sorely needed to investigate predominant infection sources of CALD, to explore their risk magnitude and to inform Legionella infection prevention and control measures in Switzerland. The current knowledge gaps regarding predominant infection sources for CALD is reflected in the Swiss Legionella spp. control guidelines. The guidelines mainly target shower and bathing water defining thresholds for Legionella spp. contamination for portable water and recommendations of hygienically optimally operating water temperatures [13]. However, the relative importance of potable water as infectious source has not been conclusively clarified. In contrast, for cooling towers, which have previously been described as infection sources for LD [34], no register exists in Switzerland as yet. The SwissLEGIO study investigates potential infection sources for Switzerland and aims to provide a foundation for evidence-based and targeted Legionella spp. prevention and control. The study draws on a unique range of scientific and policy expertise on Legionella spp. and LD in Switzerland. Through close collaboration with national partners in aquatic science and building-technology research, we harness expertise in environmental sampling and sample analysis, bridging the gap between human health and environmental exposure. The recruitment through the hospital network and the close collaboration with the NRCL and IMM aligns the study with Legionella spp. surveillance activities enacted by the FOPH [7]. The support from and regular consultations with the Federal Offices (FOPH and FSVO) further ensures that the project aligns with governmental research needs, facilitating research uptake and policy improvement.

The study contributes to capacity building for future national Legionella surveillance and LD case management. Through the hospital network, the study raises awareness for LD and promotes the collection and analysis of clinical Legionella spp. isolates as part of the routine surveillance. Clinical, radiological and laboratory data from the SwissLEGIO study will be used in satellite projects to validate the Legionella score developed by Fiumefreddo et al. [41] and to systematically assess radiological characteristics of pneumonia caused by Legionella spp. in a large LD patient population that is representing a significant proportion of reported LD cases in Switzerland. For both of these assessments we will use a control group of suspected pneumonia cases tested negative with the Legionella UAT. This analysis of clinical, laboratory and radiological characteristics of LD together with data that is collected on patients’ health-seeking and recovery from LD may inform revisions of current pneumonia management guidelines. Experiences gained from processes established during this study will also aid the effort to introduce a nationally standardised questionnaire for future case and outbreak investigations. As of today, Switzerland lacks such a comprehensive LD outbreak investigation toolbox [13]. This results in procedures applied to address LD clusters or outbreaks within Switzerland being heterogeneous and, hence, the responsibilities of different stakeholders not being well defined [13]. In turn, this hampers a successful and timely detection of the cluster’s infection source. Lastly, the study will, as part of the LeCo consortium’s research portfolio, play an important role in assessing and informing stakeholders and authorities of the applicability of WGS for single case and cluster investigations during routine surveillance activities.

The SwissLEGIO study also provides a unique platform for future research on Legionella and LD, including a more in-depth exploration of the bacteria’s complex ecology, of virulence factors, of antimicrobial susceptibility of clinical and environmental Legionella spp. strain, of clinical and laboratory characteristics and also on disease progression and long-term sequelae of LD. Additionally, the SwissLEGIO study data is contributing towards the establishment of a nationally centralised biobank for clinical and environmental Legionella spp. strains and associated epidemiological metadata on the spsp.ch platform. Similar to EpiPulse, which is currently implemented by the European Centre for Disease Prevention and Control (ECDC) [42], the spsp.ch platform will allow researchers and policy makers to exchange epidemiological and genomic data on LD. Such platforms promote research on LD to be conducted in inter-and transdisciplinary collaborations that are highly needed to address the complex pathway from environmental exposure to Legionella spp. to the clinical presentation of LD. Finally, the experiences gained conducting this study, and the data foundation SwissLEGIO is providing on LD may provide an opportunity to link Switzerland through LD on a scientific level to the international data sharing initiatives and EpiPulse, which connect European research and public health community.