<<The>> epidemiology of legionnaires' disease in Switzerland: a re-emerging disease

For almost two decades, the number of Legionnaires’ disease cases, a severe pneumonia caused by the bacterium Legionella spp., has been increasing in Switzerland. An increase in the number of cases has also been observed in many other countries for which case estimates are available, such as countries in Europe, the United States of America and Canada. After the discovery of Legionnaires’ disease in 1976 and a subsequent period of low observed case numbers and little public health attention, the disease has now been described as a re-emerging infectious disease: In 2021, about seven times as many cases were reported in Switzerland as in 2000. However, as with many (re-)emerging infectious diseases, little is known about its epidemiology, specifically the various risk factors and sources of infection. The latter is a particular conundrum because Legionella spp. are ubiquitous in our everyday environment and suspected infection sources are numerous and diverse. The lack of knowledge on the aetiology and population health dynamics of Legionnaires’ disease poses severe challenges for evidence-based prevention and control measures – and to research itself.
The overall aim of this thesis was to provide a comprehensive description of the epidemiology of Legionnaires’ disease in Switzerland to guide and support future research, and to provide a basis for evidence-based decision-making.
The analysis of Legionnaires’ disease national notification data regarding temporal and spatial patterns and data quality marked the starting point of this research. To better understand and interpret this data, we then investigated the processes involved in the diagnosis and reporting of cases. The positivity rate, which relates the number of positive Legionella findings to the number of diagnostic tests performed in Swiss medical laboratories, was determined for a ten-year period. Using a qualitative approach, we further explored physicians’ decision-making pathways and case management of community-acquired pneumonia. These studies were complemented by a comprehensive review of existing recommendations, guidelines and legislation on or in the management of Legionnaires’ disease and Legionella. We looked at four different topics: environmental prevention and control, clinical case management, disease surveillance, and outbreak management. Furthermore, we explored large- and small-scale risk factors for population exposure. For this purpose, we used an ecological model to investigate spatial and environmental infection determinants at the district level and a case-crossover design to identify the short-term associations between the onset of Legionnaires’ disease and the preceding weather and air pollution levels.
The development of a research study to investigate small-scale risk factors and exposure sites for community-acquired and sporadic Legionnaires’ disease is particularly challenging. Legionnaires’ disease remains comparatively rare even today. Rare diseases are usually investigated with case-control studies; however, case-control studies must rely on self-reporting by study participants to evaluate exposure. The challenges and limitations of working with self-reported data are exacerbated by the variable and long incubation period of Legionnaires’ disease and the often-poor health status of patients, resulting in data being collected long after the relevant exposure. Beyond the case-control study, comparative genomics, in conjunction with epidemiological data, provides the most conclusive evidence of a source of infection. It involves comparing Legionella isolates from the lower respiratory tract of patients and the suspected source of infection using whole genome sequencing. However, clinical isolates are rarely available, as lower respiratory tract samples are often not collected or tested for Legionella. The procedure for collecting and analysing environmental samples presents further difficulties. Due to the ecology of Legionella (e.g. their intracellular persistence in amoebae), they cannot always be detected consistently in the environment. In addition, Legionella are difficult to culture. Therefore, the design and implementation of a national prospective case-control and molecular source-attribution study to identify host, behavioural and environmental risk factors and individual exposure sites concludes this thesis.
This thesis synthesizes existing knowledge and generates new evidence on the epidemiology of Legionnaires’ disease in Switzerland. The analysis of the national notification data showed that between 2000 and 2020 the annual crude notification rate for legionellosis cases increased from 1.1 cases (confidence interval (CI): 0.9–1.4) to 5.6 cases (CI: 5.1–6.1) per 100,000 inhabitants. Despite the overall increase, the case numbers in 2020 have been slightly lower than in previous years. Fewer clinical reports were sent in in 2020, likely due to an overburdening of reporting physicians during the COVID-19 pandemic. The lack of clinical notifications may have led to an underestimation of cases, as the case definition only classifies cases with clinically proven pneumonia as Legionnaires’ disease. Additionally, in 2020 we observed a short-term 35% decline in reported cases, which was associated with COVID-19 containment measures, such as travel restrictions and/or related behavioural changes. In 2021, the number of cases increased again; the latest reports from the Federal Office of Public Health show a notification rate of Legionnaires’ disease of now 6.5 per 100,000 inhabitants - one of the highest in Europe.
Apart from the long-term temporal development, Legionnaires’ disease in Switzerland is subject to a pronounced seasonality, with 37% of all cases occurring between June and August. This contrasts with the number of diagnostic tests for Legionnaires’ disease, which generally peaks in winter. The overall number of diagnostic tests more than doubled between 2007 and 2016. The urinary antigen test has been reported as the most widely used test, continuously reflected in over 80% of all reported case diagnoses.
Clinical guidelines for the diagnosis and treatment of community-acquired pneumonia do not recommend aetiological testing of pneumonia in outpatient settings. Hence, the largest proportion of reported Legionnaires’ disease cases stems from hospitals and the hospitalisation rate for notified cases is generally high (89.9%). In our qualitative study, physicians working in hospitals indicated a high level of awareness of Legionnaires’ disease and its diagnostic and treatment approaches. In contrast, general practitioners indicated lower levels of awareness, reflecting the fact that they treat pneumonia cases empirically without identifying the causative agent. Thus, clinical guidelines and shape physicians’ level of awareness. Furthermore, physicians reported concerns about the urinary antigen test’s sensitivity and coverage limited to the detection of Legionella pneumophila serogroup 1. The availability of diagnostic tests and the physicians’ perception of the reliability of the testing procedures also influenced their preference for targeted treatment approaches with antibiotics or the use of broad-spectrum antibiotics. Finally, external constraints such as financial and time considerations also affected physicians’ testing and treatment preferences.
The extent and public health relevance of undetected, mild Legionnaires’ disease cases and the proportion of avoidable severe cases through earlier detection remains unknown. The case fatality rate of Legionnaires’ disease decreased between 2000 and 2020 from 7.7% to 3.6%. A comparison with hospital statistics, however, shows that the case fatality rate is underestimated by 30% on average. Despite this underestimation, the ‘true’ case fatality rate of about 5.1% seems to be somewhat lower than the European average of 8%.
Regarding regional distribution within Switzerland, the canton of Ticino in southern Switzerland consistently reported higher per capita Legionnaires’ disease case numbers than the rest of Switzerland, with a standardised notification rate of 14.3 cases per 100,000 inhabitants (CI: 12.6-16.0). It was also identified as a statistically significant regional hotspot for 2017-2021. However, in recent years, case numbers have decreased in Ticino and increased in all other regions of Switzerland.
We argue that the overall increase in Legionnaires’ disease case numbers is at least partly due to changes in the underlying disease incidence and does not represent only a surveillance artefact. The clinical guidelines for aetiological testing of pneumonia cases, which affect case detection and thereby the observed number of cases, have long been standardised for hospitalised pneumonia patients. Similarly, diagnostic test methods remained largely unchanged. Another hypothesis explaining the increase as a surveillance artefact has been that increasing physician awareness of Legionella as a cause of pneumonia may have led to increased case detection. However, it is reasonable to assume that the testing protocols from the clinical guidelines have been followed in the past, even when the level of awareness of the disease was not as high as it is today. Furthermore, the influence of growing awareness among physicians should diminish over time and lead to a plateau of notified cases. However, after 20 years of sustained increase, the increase in the number of Legionnaires’ disease cases shows no sign of slowing down.
Despite an improved understanding of the Legionnaires’ disease burden in Switzerland, the drivers for infection and causes of regional heterogeneity remained unclear. Using two different methodologies (an ecological regression model and a case-crossover study), we found evidence for the short-term association of elevated daily mean temperature (odds ratio (OR): 2.83; CI: 1.70-4.70) and mean daily vapour pressure (OR: 1.52; CI: 1.15-2.01) 6-14 days before Legionnaires’ disease onset. In the ecological model, we also found a strong association between Legionnaires’ disease incidence and air pollution levels, but no significant results in the case-crossover study. However, as the ecological model can be subject to ecological bias and the case-crossover study was limited in power, future studies are needed to further investigate the association. Knowledge of these large-scale risk factors, such as the impact of weather conditions and air pollution on the occurrence of Legionnaires’ disease, is essential. It contributes to the understanding of regional differences, provides information on the vulnerability of certain at-risk populations/regions and ultimately helps to anticipate disease trends.
The investigation of small-scale risk factors and exposure sites is of central importance for targeted prevention and control measures. However, due to the disease dynamics and the role of water supply systems in the transmission of Legionella, the investigation is complex and requires suitable research methodologies and a broad range of expertise. The studies summarised in this thesis have informed the design of a prospective one-year national case-control and molecular source attribution study. The study set-up includes the establishment of a network of 20 university and cantonal hospitals to facilitate and expedite recruitment of patients with Legionnaires’ disease and promote the sampling of material from the lower respiratory tract to obtain clinical Legionella isolates. In a subset of cases and controls (from the general population), water samples are collected from the shower and kitchen tap, which are then analysed and processed to obtain isolates of Legionella from the environment. In a last step, the clinical and environmental isolates are genetically matched using whole genome sequencing to support infection source attribution. The environmental component of this study has been developed and implemented jointly with experts in water hygiene in buildings and environmental microbiology. The study, thus, provides the framework for a wide range of research on Legionnaires’ disease and Legionella, including clinical aspects, such as long-term health effects, as well as the identification of household characteristics conducive to Legionella contamination. The implementation of this national research project strengthens intersectoral and multidisciplinary collaboration and capacity building to address the ongoing increase in Legionnaires’ disease case numbers.
In light of climate change, and demographic changes, the number of observed cases of Legionnaires’ disease is expected to increase further in Switzerland and abroad. To stop this trend, comprehensive research is needed to allow targeted and evidence-based action. Although Switzerland benefits from strong government support to combat this disease, data gaps remain an obstacle and, in many other countries, the data gap is even larger. The lack of data and, thus, estimates on the disease burden does not translate into the absence of a public health problem and efforts should be made to investigate the attributable Legionnaires’ disease burden globally. In the context of climate change and urbanisation, public health should advocate for healthy (built) environments to curb Legionnaires’ disease and other (re-)emerging infectious diseases.