Association between rheumatic diseases and cancer: results from a clinical practice cohort study.

The association between cancer and immune-mediated rheumatic conditions is controversial, especially as far as polymyalgia rheumatica (PMR) is concerned. Furthermore, no clinical feature has been shown to be suggestive of a paraneoplastic rheumatic syndrome. With the present study, we aim to address both these issues. The study population comprised N = 1750 patients, including N = 100 with PMR, who attended our tertiary immuno-rheumatology clinic between January 1, 2005 and November 30, 2012. A rheumatic disease was deemed paraneoplastic if cancer had been diagnosed in the 2 years preceding or following its onset. The probability of a significant association between a specific rheumatic disease and cancer was evaluated by computing the odds ratio (OR): N = 702 patients with osteoarthritis serving as controls. Furthermore, clinical features distinguishing paraneoplastic rheumatic diseases were searched for by univariate and multivariate analysis. Sjogren's syndrome (SS) [OR 3.6 (CI 95% 1.7-7.5)], PMR (OR 5.1 CI 95% 2.9-8.9), dermatomyositis/polymyositis [OR 12.09 (CI 95% 2.6-55.8)] and vasculitis [OR 3.70 (CI 95% 1.81-7.52)] are associated with cancer. At multivariate analysis, older age is associated with cancer among SS patients (p = 0.03), while in the PMR group, older age, male gender, and ≥6 tender joints are independent predictors of paraneoplastic PMR (p < 0.0004). Cancer frequently either heralds or follows rheumatic manifestations, including PMR. Older age, male gender and a more extensive joint involvement should be considered red flags for paraneoplastic PMR.