BACKGROUND: Cooling towers containing Legionella spp are a high-risk source of Legionnaires\' disease outbreaks. Manually locating cooling towers from aerial imagery during outbreak investigations requires expertise, is labour intensive, and can be prone to errors. We aimed to train a deep learning computer vision model to automatically detect cooling towers that are aerially visible.
METHODS: Between Jan 1 and 31, 2021, we extracted satellite view images of Philadelphia (PN, USA) and New York state (NY, USA) from Google Maps and annotated cooling towers to create training datasets. We augmented training data with synthetic data and model-assisted labelling of additional cities. Using 2051 images containing 7292 cooling towers, we trained a two-stage model using YOLOv5, a model that detects objects in images, and EfficientNet-b5, a model that classifies images. We assessed the primary outcomes of sensitivity and positive predictive value (PPV) of the model against manual labelling on test datasets of 548 images, including from two cities not seen in training (Boston [MA, USA] and Athens [GA, USA]). We compared the search speed of the model with that of manual searching by four epidemiologists.
RESULTS: The model identified visible cooling towers with 95·1% sensitivity (95% CI 94·0-96·1) and a PPV of 90·1% (95% CI 90·0-90·2) in New York City and Philadelphia. In Boston, sensitivity was 91·6% (89·2-93·7) and PPV was 80·8% (80·5-81·2). In Athens, sensitivity was 86·9% (75·8-94·2) and PPV was 85·5% (84·2-86·7). For an area of New York City encompassing 45 blocks (0·26 square miles), the model searched more than 600 times faster (7·6 s; 351 potential cooling towers identified) than did human investigators (mean 83·75 min [SD 29·5]; mean 310·8 cooling towers [42·2]).
CONCLUSIONS: The model could be used to accelerate investigation and source control during outbreaks of Legionnaires\' disease through the identification of cooling towers from aerial imagery, potentially preventing additional disease spread. The model has already been used by public health teams for outbreak investigations and to initialise cooling tower registries, which are considered best practice for preventing and responding to outbreaks of Legionnaires\' disease.
BACKGROUND: None.

BACKGROUND: In case of pneumonia, some biological findings are suggestive for Legionnaire\'s disease (LD) including C-reactive protein (CRP). A low level of CRP is predictive for negative Legionella Urinary-Antigen-Test (L-UAT).
METHODS: Observational retrospective study in Nord-Franche-Comté Hospital with external validation in Besançon University Hospital, France which included all adults with L-UAT performed during January 2018 to December 2022. The objective was to determine CRP optimal threshold to predict a L-UAT negative result.
RESULTS: URINELLA included 5051 patients (83 with positive L-UAT). CRP optimal threshold was 131.9 mg/L, with a negative predictive value (NPV) at 100%, sensitivity at 100% and specificity at 58.0%. The AUC of the ROC-Curve was at 88.7% (95% CI, 86.3-91.1). External validation in Besançon Hospital patients showed an AUC at 89.8% (95% CI, 85.5-94.1) and NPV, sensitivity and specificity was respectively 99.9%, 97.6% and 59.1% for a CRP threshold at 131.9 mg/L; after exclusion of immunosuppressed patients, index sensitivity and NPV reached also 100%.
CONCLUSIONS: In case of pneumonia suspicion with a CRP level under 130 mg/L (independently of the severity) L-UAT is useless in immunocompetent patients with a NPV at 100%. We must remain cautious in patients with symptoms onset less than 48 h before CRP dosage.

Background and Objectives: Legionnaires\' disease (LD) is an acute respiratory disease with increasing annual numbers of reported domestic and global cases. This study aimed to establish foundational data for the prevention and control of LD by investigating the occurrence and infection routes of reported and suspected cases of LD in Gyeonggi Province, Korea, from January 2016 to December 2022, and by and analyzing the risk factors for death. Materials and Methods: A sex-and-age standardization was performed on LD patients and suspected cases reported in Gyeonggi Province. The monthly average number of confirmed cases was visualized using graphs, and a survival analysis was performed using Kaplan-Meier survival curves. The mortality risk ratio was estimated using the Cox proportional hazards model. Results: The incidence of LD in Gyeonggi Province mirrored the national trend, peaking in July with the highest number of confirmed and suspected cases. While there was no significant difference in survival rates by age, the survival rate was higher for suspected cases when analyzed separately. Comparing the death ratio by infection route, nosocomial infections showed the highest death ratio, and intensive care unit (ICU) admission and the presence of coinfections were significantly correlated with mortality. Factors such as nosocomial infection, admission within 1 to 3 days following diagnosis, and the development of complications were factors contributing to a higher risk of death. Conclusions: The general characteristics of patients with LD were similar to those suggested by previous studies. The proportion of community-acquired infections was lower than in previous studies, but the length of hospital stay was similar for survivors and the deceased, and the mortality rate within 30 days after diagnosis was higher for nosocomial infections. In conclusion, nosocomial infection, a period of up to 3 days from admission to diagnosis, and complications were significantly related to the mortality rate of LD.

OBJECTIVE: The effectiveness of monotherapy and combination therapy with quinolones and macrolides for treating Legionnaires\' disease remains uncertain; this study aimed to assess the comparative effectiveness of three treatment approaches.
METHODS: Using a nationwide inpatient database, we analyzed 3560 eligible patients hospitalized for Legionnaires\' disease between April 1, 2014, and March 31, 2021; patients were divided into combination therapy, quinolone monotherapy, and macrolide monotherapy groups according to the antibiotics administered within 2 days of admission. We compared in-hospital mortality, total hospitalization costs, and length of stay across these groups using multiple propensity score analysis with inverse probability of treatment weighting.
RESULTS: Of the 3560 patients, there were 564 (15.8%), 2221 (62.4%), and 775 (21.8%) patients in the combination therapy, quinolone monotherapy, and macrolide monotherapy groups, respectively. No significant differences were observed in in-hospital mortality between combination therapy and quinolone monotherapy groups, and between combination therapy and macrolide monotherapy groups. There were no significant differences in total hospitalization costs or length of stay among the three groups.
CONCLUSIONS: The study suggests that there may not be a significant advantage in using a combination of quinolones and macrolides over monotherapy for the treatment of Legionnaires\' disease. Given the potential for increased side effects, careful consideration is advised when choosing this combination therapy.

BACKGROUND: Legionnaires disease (LD) is a rare, life-threatening opportunistic bacterial infection that poses a significant risk to patients with impaired cell-mediated immunity such as solid organ transplant recipients. However, the epidemiologic features, clinical presentation, and outcomes of LD in this population are poorly described.
OBJECTIVE: What are the clinical manifestations, radiologic presentation, risk factors for severity, treatment, and outcome of LD in solid organ transplant recipients?
METHODS: In this 10-year multicenter retrospective cohort study in France, where LD notification is mandatory, patients were identified by hospital discharge databases. Diagnosis of LD relied on positive culture findings from any respiratory sample, positive urinary antigen test (UAT) results, positive specific serologic findings, or a combination thereof. Severe LD was defined as admission to the ICU.
RESULTS: One hundred one patients from 51 transplantation centers were eligible; 64 patients (63.4%) were kidney transplant recipients. Median time between transplantation and LD was 5.6 years (interquartile range, 1.5-12 years). UAT results were positive in 92% of patients (89/97). Among 31 patients with positive culture findings in respiratory samples, Legionella pneumophila serogroup 1 was identified in 90%. Chest CT imaging showed alveolar consolidation in 98% of patients (54 of 57), ground-glass opacity in 63% of patients (36 of 57), macronodules in 21% of patients (12 of 57), and cavitation in 8.8% of patients (5 of 57). Fifty-seven patients (56%) were hospitalized in the ICU. In multivariate analysis, severe LD was associated with negative UAT findings at presentation (P = .047), lymphopenia (P = .014), respiratory symptoms (P = .010), and pleural effusion (P = .039). The 30-day and 12-month mortality rates were 8% (8 of 101) and 20% (19 of 97), respectively. In multivariate analysis, diabetes mellitus was the only factor associated with 12-month mortality (hazard ratio, 3.2; 95% OR, 1.19-8.64; P = .022).
CONCLUSIONS: LD is a late and severe complication occurring in solid organ transplant recipients that may present as pulmonary nodules on which diabetes impacts its long-term prognosis.

Legionnaires\' disease is a severe form of pneumonia caused by the inhalation or aspiration of water droplets contaminated with Legionella pneumophila and other Legionella species. These bacteria are commonly found in natural habitats and man-made water systems. Legionnaires\' disease is a significant public health problem, especially in healthcare settings where patients may be exposed to contaminated environmental sources. Nosocomial outbreaks have been reported worldwide, leading to high morbidity and mortality rates, and increased healthcare costs. This study aimed to compare, the clonal relationship of clinical L. pneumophila strains from two different hospitals with L. pneumophila strains isolated from the water supply.
In the period from 2019 to 2021, clinical and environmental strains involved in three cases of legionellosis were compared by means of pulsed field gel electrophoresis and sequence based typing techniques.
Our findings highlight the persistence of clonally distinct strains within each hospital examined. Furthermore, the L. pneumophila strains detected from hospital environmental sources were related to the clinical strains isolated, demonstrating the nosocomial origin of these cases.
Therefore, it is important to implement more accurate surveillance systems both for epidemiological studies and to check the effectiveness of remediation procedures. (www.actabiomedica.it).

BACKGROUND: Severe community-acquired pneumonia (SCAP) is commonly treated with an empiric combination therapy, including a macrolide, or a quinolone and a β-lactam. However, the risk of Legionella pneumonia may lead to a prolonged combination therapy even after negative urinary antigen tests (UAT).
METHODS: We conducted a retrospective cohort study in a French intensive care unit (ICU) over 6 years and included all the patients admitted with documented SCAP. All patients received an empirical combination therapy with a β-lactam plus a macrolide or quinolone, and a Legionella UAT was performed. Macrolide or quinolone were discontinued when the UAT was confirmed negative. We examined the clinical and epidemiological features of SCAP and analysed the independent factors associated with ICU mortality.
RESULTS: Among the 856 patients with documented SCAP, 26 patients had atypical pneumonia: 18 Legionella pneumophila (LP) serogroup 1, 3 Mycoplasma pneumonia (MP), and 5 Chlamydia psittaci (CP). UAT diagnosed 16 (89%) Legionella pneumonia and PCR confirmed the diagnosis for the other atypical pneumonia. No atypical pneumonia was found by culture only. Type of pathogen was not associated with a higher ICU mortality in the multivariate analysis.
CONCLUSIONS: Legionella pneumophila UAT proved to be highly effective in detecting the majority of cases, with only a negligible percentage of patients being missed, but is not sufficient to diagnose atypical pneumonia, and culture did not provide any supplementary information. These results suggest that the discontinuation of macrolides or quinolones may be a safe option when Legionella UAT is negative in countries with a low incidence of Legionella pneumonia.

The number of reported cases of Legionnaires\' disease (LD) has risen markedly in Switzerland (6.5/100,000 inhabitants in 2021) and abroad over the last decade. Legionella, the causative agent of LD, are ubiquitous in the environment. Therefore, environmental changes can affect the incidence of LD, for example by increasing bacterial concentrations in the environment or by facilitating transmission.
The aim of this study is to understand the environmental determinants, in particular weather conditions, for the regional and seasonal distribution of LD in Switzerland.
We conducted a series of analyses based on the Swiss LD notification data from 2017 to 2021. First, we used a descriptive and hotspot analysis to map LD cases and identify regional clusters. Second, we applied an ecological model to identify environmental determinants on case frequency at the district level. Third, we applied a case-crossover design using distributed lag non-linear models to identify short-term associations between seven weather variables and LD occurrence. Lastly, we performed a sensitivity analysis for the case-crossover design including NO2 levels available for the year 2019.
Canton Ticino in southern Switzerland was identified as a hotspot in the cluster analysis, with a standardised notification rate of 14.3 cases/100,000 inhabitants (CI: 12.6, 16.0). The strongest association with LD frequency in the ecological model was found for large-scale factors such as weather and air pollution. The case-crossover study confirmed the strong association of elevated daily mean temperature (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 LD occurrence.
Our analyses showed an influence of weather with a specific temporal pattern before the onset of LD, which may provide insights into the effect mechanism. The relationship between air pollution and LD and the interplay with weather should be further investigated.

The evidence for the incubation period of Legionnaires\' disease is based on data from a small number of outbreaks. An incubation period of 2-10 days is commonly used for the definition and investigation of cases. In the German LeTriWa study, we collaborated with public health departments to identify evidence-based sources of exposure among cases of Legionnaires\' disease within 1-14 days before symptom onset. For each individual, we assigned weights to the numbered days of exposure before symptom onset, giving the highest weight to exposure days of cases with only one possible day of exposure. We then calculated an incubation period distribution where the median was 5 days and the mode was 6 days. The cumulative distribution reached 89% by the 10th day before symptom onset. One case-patient with immunosuppression had a single day of exposure to the likely infection source only 1 day before symptom onset. Overall, our results support the 2- to 10-day incubation period used in case definition, investigation, and surveillance of cases with Legionnaires\' disease.

Objective Recommendations on the timing of Legionella urinary antigen tests for community-acquired pneumonia patients differ among guidelines in Japan, the United States, and European nations. We therefore evaluated the association between the timing of urinary antigen tests and in-hospital mortality in patients with Legionella pneumonia. Methods We conducted a retrospective cohort study using the Diagnosis Procedure Combination database, a nationwide database of acute care inpatients in Japan. Patients who underwent Legionella urinary antigen tests on the day of admission formed the tested group. Patients who were tested on day 2 of admission or later or were unexamined formed the control group. We performed a propensity score matching analysis to compare in-hospital mortality, length of hospital stay and duration of antibiotics use between the two groups. Results Of the 9,254 eligible patients, 6,933 were included in the tested group. One-to-one propensity score matching generated 1,945 pairs. The tested group had a significantly lower 30-day in-hospital mortality than the control group (5.7 vs. 7.7%; odds ratio, 0.72; 95% confidence intervals, 0.55-0.95; p=0.020). The tested group also showed a significantly shorter length of stay and duration of antibiotics use than the control group. Conclusion Urine antigen testing upon admission was associated with better outcomes in patients with Legionella pneumonia. Urine antigen tests upon admission may be recommended for all patients with severe community-acquired pneumonia.