Search

It is not known whether rainfall increases the risk of sporadic cases of Legionella pneumonia. We sought to test this hypothesis in a prospective observational cohort study of non-immunosuppressed adults hospitalized for community-acquired pneumonia (1995-2011). Cases with Legionella pneumonia were compared with those with non-Legionella pneumonia. Using daily rainfall data obtained from the regional meteorological service we examined patterns of rainfall over the days prior to admission in each study group. Of 4168 patients, 231 (5.5%) had Legionella pneumonia. The diagnosis was based on one or more of the following: sputum (41 cases), antigenuria (206) and serology (98). Daily rainfall average was 0.556 liters/m(2) in the Legionella pneumonia group vs. 0.328 liters/m(2) for non-Legionella pneumonia cases (p = 0.04). A ROC curve was plotted to compare the incidence of Legionella pneumonia and the weighted median rainfall. The cut-off point was 0.42 (AUC 0.54). Patients who were admitted to hospital with a prior weighted median rainfall higher than 0.42 were more likely to have Legionella pneumonia (OR 1.35; 95% CI 1.02-1.78; p = .03). Spearman Rho correlations revealed a relationship between Legionella pneumonia and rainfall average during each two-week reporting period (0.14; p = 0.003). No relationship was found between rainfall average and non-Legionella pneumonia cases (-0.06; p = 0.24). As a conclusion, rainfall is a significant risk factor for sporadic Legionella pneumonia. Physicians should carefully consider Legionella pneumonia when selecting diagnostic tests and antimicrobial therapy for patients presenting with CAP after periods of rainfall.

Background: To describe the prevalence, clinical characteristics, impact of systemic steroids exposure and outcomes of delayed cerebral vasculopathy (DCV) in a cohort of adult patients with pneumococcal meningitis (PM). Methods: Observational retrospective multicenter study including all episodes of PM from January 2002 to December 2015. DCV was defined as proven/probable/possible based upon clinical criteria and pathological-radiological findings. DCV-patients and non-DCV-patients were compared by univariate analysis. Results: 162 PM episodes were included. Seventeen (10.5%) DCV-patients were identified (15 possible, 2 probable). At admission, DCV-patients had a longer duration of symptoms (>2 days in 58% vs. 25.5% (p 0.04)), more coma (52.9% vs. 21.4% (p 0.03)), lower median CSF WBC-count (243 cells/uL vs. 2673 cells/uL (p 0.001)) and a higher proportion of positive CSF Gram stain (94.1% vs. 71% (p 0.07)). Median length of stay was 49 vs. 15 days (p 0.001), ICU admission was 85.7% vs. 49.5% (p 0.01) and unfavorable outcome was found in 70.6% vs. 23.8% (p 0.001). DCV appeared 1-8 days after having completed adjunctive dexamethasone treatment (median 2,5, IQR = 1.5-5). Conclusions: One tenth of the PM developed DCV. DCV-patients had a longer duration of illness, were more severely ill, had a higher bacterial load at admission and had a more complicated course. Less than one third of cases recovered without disabilities. The role of corticosteroids in DCV remains to be established.