Tumor necrosis factor inhibitors added to nonbiological immunosuppressants vs. nonbiological immunosuppressants alone: a different signal of cancer risk according to the condition. A disproportionality analysis in a nationwide pharmacovigilance database

We aimed at detecting a signal of an increased risk of cancer in patients treated with TNF inhibitor (TNFi) and nonbiological immunosuppressant (NBIS), compared with NBIS alone for autoimmune diseases. Secondly, we aimed at comparing this risk between the different TNFis. We conducted a disproportionality analysis (case/noncase study) from the French National PharmacoVigilance Database. We selected all the reports of serious adverse drug reactions from 2000 to 2010 in patients treated with NBIS for labeled indications of TNFi. Cases were all the reports of cancer that occurred after a minimal 3-month exposure to NBIS. Noncases were all the other reports. We searched for exposure to TNFi and calculated reporting odds ratios (RORs), stratified by condition and type of cancer and adjusted by age, gender, history of cancer, type of NBIS and year of reporting. Of the 1918 reports included in the study population, 217 were cases (135 solid and 82 blood cancers). A safety signal was found in rheumatoid arthritis (RA) (ROR: 5.43, 95% CI[3.52-8.38]) particularly for nonmelanoma skin cancer (NMSC) (20.17[2.49-163.36]), and in psoriasis/psoriatic arthritis (3.45[1.09-10.92]). No signal was found in inflammatory bowel diseases (IBD) and ankylosing spondylitis, whatever the type of cancer. There was no difference between TNFis. This study puts the argument of an increased risk of cancer (particularly NMSC) in patients with rheumatoid arthritis exposed to TNFi and NBIS compared with NBIS alone, but not in IBD and ankylosing spondylitis patients. No signal was detected for melanoma potentially related to the lack of power. The signal seems similar whatever the TNFi.