Background: Infections which complicate rheumatic diseases such as Rheumatoid Arthritis (RA) and Spondyloarthropathy (SpA) (Psoriatic Arthritis [PA] and Ankylosing Spondylitis [AS]), may cause significant morbidity and mortality. However, among the studies on the incidence rate (IR) of infections in such patients, very few have involved controls and the results have been controversial, probably due to methodological difficulties.To estimate infection rates in RA and SpA patients under disease-modifying anti-rheumatic drugs (DMARDs), corticosteroids (CS) and tumor necrosis factor (TNF)α antagonists, alone or combined, a single-centre retrospective observational cohort study has been performed.Patients and methods: Incidence rates/100 patient-years of any infections were evaluated in RA and SpA outpatients observed in the period November 1, 2003 through December 31, 2009 and stratified according to therapy. Infection incidence rate ratios (IRR) were calculated using Poisson regression models which adjusted for demographic/clinical characteristics of the patients.Results: Three hundred and thirtyone infections [318 (96.1%) non-serious and 13 (3.9%) serious] have been registered among 176 of the 341 patients (52%). The IR/100 patient-years of all infections was 36.3 ranging from 12.4 (DMARDs + CS) to 62.7 (anti-TNFα + CS). The most frequent infection site was respiratory tract, and bacteria were responsible for three quarters of all infections. In the multivariate analysis, adding anti-TNFα to DMARDs doubled the IRR compared to DMARDs alone, anti-TNFα + CS significantly tripled it, whereas anti-TNFα + CS + DMARDs only increased the risk 2.5 times. The degree of disease activity was strongly and significantly associated with the infection risk (severe or moderate versus mild, IRR = 4). Female sex was significantly associated with increased infection risk, while duration of disease and anti-influenza vaccination were protective, the latter even for cutaneous/soft-tissue (mainly herpetic) infections.Conclusion: The combination anti-TNFα with CS was found to be the most pro-infective treatment, whereas DMARDs alone were relatively safe. Physicians, therefore, should be aware that there may be an increased risk of infection when using anti-TNFα and CS therapy together. Anti-influenza vaccination appears to provide broad protection, adding evidence to support its use in these patients, and deserves further study. © 2014 Germano et al.; licensee BioMed Central Ltd.