The association between longstanding or recent onset rheumatoid arthritis and 4-year incidence of cancer.

<p>Values are odds ratios and <i>p</i>-values from multivariable logistic regression analyses for the development of lymphoproliferative cancer and solid tumors for the risk groups: No previous rheumatoid arthritis, Longstanding rheumatoid arthritis, and Recent onset rheumatoid arthritis. Odds ratios are adjusted in a stepwise manner: <a href="http://www.plosone.org/article/info:doi/10.1371/journal.pone.0099388#pone.0099388-Hemminki1" target="_blank">[1]</a> sex, age (quadratic), year and month; <a href="http://www.plosone.org/article/info:doi/10.1371/journal.pone.0099388#pone.0099388-Chakravarty1" target="_blank">[2]</a> previous and present eosinophilia; <a href="http://www.plosone.org/article/info:doi/10.1371/journal.pone.0099388#pone.0099388-Khurana1" target="_blank">[3]</a> Charlson's Comorbidity Index, and <a href="http://www.plosone.org/article/info:doi/10.1371/journal.pone.0099388#pone.0099388-Mercer1" target="_blank">[4]</a> C-reactive protein. <sup>†</sup><i>P</i>-value for the likelihood-ratio test of all categories of eosinophilia simultaneously. * Hodgkin's lymphoma, non-Hodgkin lymphoma and chronic lymphocytic leukemia.</p>