dumoffii /em and anti- em L. A definition of Pontiac fever was devised based on clinical symptoms explained in epidemic investigations and on their timing after the exposure event. The association between incidence of Pontiac fever and shower contamination levels was evaluated to test the relevance of this definition. Results The proposed definition of Pontiac fever associated the following criteria: Anguizole occurrence of at least one symptom among headache, myalgia, fever and shivers, possibly associated with other ‘minor’ symptoms, within three days after a shower contaminated by em Legionella /em , during a maximum of 8 days (minimum 2 days). 23 such cases occurred during the study (incidence rate: 0.125 cases per person-year [95% CI: 0.122C0.127]). A concentration of em Legionella /em in water equal to or greater than 104.L-1 (FISH method) was associated with a significant increase of incidence of Pontiac fever (p = 0.04). Conclusion Once validated in other settings, the proposed definition of Pontiac fever might be used to develop epidemiological surveillance and help draw attention on sources of em Legionella /em Anguizole TSPAN12 . Background Pontiac fever (PF) is the moderate form that takes contamination by em Legionella /em . It usually appears on an epidemic mode and is not associated with pneumonia . Like for Legionaires’ disease, contamination stems from inhalation of an aerosol contaminated by em Legionella /em . Clinically, Pontiac fever’s symptoms mimic influenza, with fever, asthenia, myalgia, arthralgia, headache, cough, nausea and sore throat [1,3]; other symptoms such as dyspnea, thoracic aches and pains [2-4] vomiting and diarrhoea [3,5,6] have also been explained. Patients recover in two to five days, without treatment [7-9]. Because of its benignity and lack of specificity, the occurrence of PF is often undiagnosed and is therefore less reported than Legionnaires’ disease. Epidemiologically, PF is usually characterized by a short incubation period (typically 30 to 90 hours, with an average of 36 h), a high Anguizole attack rate (70 to 90%) , and absence of fatalities or long term complications . Age, gender and smoking do not seem to be risk factors [10,11]. Rather, PF seems to impact preferentially young subjects: the age of cases was 36 to 39 years in the original Pontiac episode [12,13], and age medians during different documented epidemics were 29 [4,10], 30  and 32 years . Pathogenesis of the PF is usually poorly known. To date, there is no consensus around the duration of the incubation period, on its clinical symptoms, nor around the causal species of em Legionella /em . Different serogroups (SGs) of em Legionella pneumophila /em ( em Lp /em ) (1, 6 and 7) [14-16], as well as em L. feeleii /em [12,17], em L. micdadei /em [2,11,12], em L. anisa /em  can cause PF. In terms of diagnosis, according to some authors, PF develops the same serological characteristics as Legionnaires’disease [10,18]. Others claim on the contrary that serology during of a PF is usually inconstantly positive . Presence of urinary antigen is not systematic either, even for epidemics connected to em Lp /em SG 1 [10,20]. Detection of PF is a marker of an environmental contamination by em Legionella /em and should thereby call for prevention measures. Efforts to standardize the Anguizole definition of PF may facilitate comparison of risk levels and help draw attention on sources of em Legionella /em . In this article, based on data from your Legion’Air project, we propose an operational definition of PF for the purpose of surveillance and epidemiological studies. Methods The objectives of the Legion’Air project are: 1) to assess the exposure of elderly people residing in nursing homes to em Legionella /em through aerosols generated by hot-water during showers, and 2) to evaluate the risk that is associated with this exposure. Nursing homes solicited to participate in the Legion’Air project were located in the Lorraine region, north east of Anguizole France. The selection process was based on the capacity of the nursing homes and on practical considerations (it should be located not too far from the study centre); no concern was given to prior knowledge of contamination of the hot water system, in order to prevent selection bias. This epidemiological study is a retrospective follow-up study. A populace of 560 elderly volunteers (informed consent was obtained from patients or guardians) have been followed during 4 months. A set of predefined symptoms were registered daily by the auxiliary nursing.