Types of evidence

The outbreak investigation will usually involve conducting and interpreting epidemiological, microbiological and environmental information.

Epidemiological investigations

In an outbreak, epidemiological investigations can provide evidence that may link cases with specific sources of exposure. Background information on this can be found using the following links:

• Descriptive epidemiology

• Analytical epidemiological studies

• Geographical information systems

Environmental investigations

The environmental investigation assesses the source system's risk of causing legionellosis. This includes assessment on the likelihood of contamination, multiplication and dissemination of Legionella. The specifics of the investigation depend on the risk of the source (see here for further information).

Microbiological evidence obtained from environmental investigations on its own does not give strong evidence to associate a source with the cases of the outbreak, without clinical microbiology. However, relevant findings from the environmental investigations would add weight to the epidemiological evidence in the absence of microbiological evidence.

Microbiological investigations

Microbiological investigations are carried out on the cases and potential source(s) of legionella, and should be carried out by an accredited laboratory. A microbiological link between cases and source would provide strong evidence to implicate that source (see here for further information).

Combining evidence sources

The strength of association between cases and exposure can be classified according to the information provided by clinical and environmental isolates. Practical and theoretical examples are provided here.

A consensus Sequence-Based Typing (SBT) epidemiological typing scheme for clinical and environmental isolates of Legionella pneumophila has been developed by members of the European Working Group for Legionella Infections (EWGLI) and evaluated for implementation in the investigation of outbreaks of legionellosis caused by L. pneumophila; for more infromation see here.

The following criteria have been described to help focus an investigation [1]:

High assurance linking exposure to case(s)

Epidemiological links in time and space, plus matching clinical and environmental strains, support the acquisition of infection from a common source

Obtaining matching species from microbiological and environmental samples in an outbreak is strong evidence to link the source with the cases, particularly when a link can be established from descriptive epidemiology that cases were exposed to the source in the two weeks prior to symptom onset. Microbiological samples are not necessarily required from all cases and, as long as samples are obtained from at least one case, that the outbreak control team feel is representative of the other cases, the link can be established.

Over 80% of environmental samples are found to be serogroup 1, so some caution is required when associating the source with the cases in outbreaks involving serogroup 1. It is always better, if possible, to use additional tests to distinguish different species and subspecies of Legionella to increase the level of confidence that there is a potential link.

Low assurance linking exposure to cases(s)

Epidemiological links in time and space and where either clinical or environmental isolates, but not both, have been obtained. The environmental isolates may point to a common source of infection but there are no clinical isolates available for strain matching, or a cooling tower may be highly suspicious as the source of infection but has been shut down and cannot be sampled and compared with any clinical isolates that have been obtained.

It is not unusual to be unable to obtain specimens for microbiological analysis. For example, cleaning may have already taken place at the source, depriving the investigating team of contemporaneous samples, or samples taken at the time of environmental investigation may not reflect the true risk at the source. Alternatively, it may be considered unsafe to obtain samples from a site. Clinical samples may also be unavailable from cases. In non-serogroup 1 cases, it may not be possible to culture the organism and it may not be diagnosed by serology or urinary antigen testing. Without a microbiological link, association of a source with case(s) will be reliant upon robust epidemiological and environmental investigations.

In outbreaks where a microbiological link is not found, it is important to obtain strong epidemiological evidence. However, it may not be possible or practicable, to perform an analytical epidemiological study for all outbreaks. Even when an analytical study is carried out, it is important to design the study robustly so as to avoid introducing bias, which can be a difficulty in case control studies. In absence of a microbiological link, combining epidemiological and environmental investigation may provide evidence of strong association. In this situation, it may be more appropriate to think of the strength of evidence as a spectrum.