Project Summary/Abstract
Legionnaires’ disease (LD) is a severe bacterial pneumonia with a high case fatality rate (~10%). Recent public
health concern and funding has been directed towards understanding the increasing incidence of LD. Over the
past two decades, incidence has increased from <0.5 cases/100,000 population to almost 2.5 cases/100,000
population. There are multiple hypotheses for the increase, including an increase in the proportion of the
population that is highly susceptible to LD. While the burden of disease is likely due to a combination of factors
(e.g. diagnostic testing intensity, weather), we believe that an increase in the proportion of the population that
is highly susceptible may play a critical, and understudied, role in the increasing incidence. The highly
susceptible population of interest includes those who are taking immunosuppressing medications. TNF
inhibitors (TNFi) (e.g. adalimumab, etanercept, infliximab) are commonly prescribed for rheumatoid arthritis
and other autoimmune disorders and have been associated with increased susceptibility to all-cause
respiratory infections. In 2011, the FDA updated the Boxed Warning for TNFi to include potential risk of
infection due to Listeria and Legionella. This advisory was in response to a limited number of case reports and
observational studies and lacked formal analyses able to adjust for risk attributed to the chronic underlying
illness (e.g. rheumatoid arthritis). An analysis quantifying the relationship between LD and a modifiable risk
factor, such as TNFi, would provide needed information to potentially curb the increase in disease. Importantly,
uptake of TNFi increased dramatically during the same time period as the increase in reported LD. We
hypothesize that 1) TNF inhibitors place individuals at greater risk for Legionnaires’ disease compared to
conventional disease modifying anti-rheumatic drugs (cDMARDs; e.g. methotrexate) and that 2) there is a
temporal association between increasing uptake of TNFi and Legionnaires’ disease incidence. To robustly
estimate these potential associations, we will take advantage of robust Danish national register data. Danish
national medical register data for approximately 25,000 people indicated for TNFis or cDMARDs between 2000
and 2020 will be explored. Information on inpatient and outpatient visits, as well as pharmaceutical
prescriptions will be included which offers the unique ability to identify hazard windows, adjust for underlying
disease, and detect longitudinal associations between prescription uptake and LD incidence. Previous studies
of the association between TNFi and LD have not been able to adjust for underlying disease severity, adding
needed novelty to our study design. Secondary analyses will assess other infectious disease outcomes that
may be associated with TNFi use (e.g. Salmonella, Listeria, all-cause pneumonia). The results of these studies
will 1) influence Legionnaires’ disease surveillance and prevention measures, 2) inform clinical decision-
making regarding TNFi, and 3) provide insight on understudied immune-mediated risk factors for LD.