Introduction: Tumor necrosis factor (TNF) dysregulation plays a central role in the aetiology of several inflammatory diseases. Currently there are five approved TNF-inhibitors: infliximab, etanercept, adalimumab, certolizumab pegol and golimumab. Anti-TNF drugs are usually well tolerated, however, there are some concerns regarding their safety profile, particularly for their potential tumorigenic effect [1, 2]. Aim: The aim of this study is to perform a disproportionality analysis using the Italian National Network of Pharmacovigilance (RNF) database of the Italian Medicines Agency to assess the occurrence of a signal of risk for the drug event pairs TNF inhibitors-lymphoma and TNF inhibitors-skin cancers. Methods: This study included all reports recorded in the RNF database from 1988 up to October 31st, 2017. Cases were defined as all reports with MedDRA Preferred Term (PT) codes of lymphoma, melanoma and non-melanoma skin cancer (NMSC). Index reports included all reports in which at least one anti-TNF drug (infliximab, etanercept, adalimumab, golimumab and certolizumab pegol) was reported as the suspected causative drug. Reference reports included all cases related to non-anti-TNF drugs. Bayesian (information component, IC) and frequentist (adjusted reporting odds ratio, aROR) methods were used to calculate disproportional distribution of drug- ADR pairs. ROR was adjusted for sex, age quartiles, reporting year, number of suspected drugs (1 or C 2) and number of concomitant drugs (0 or C 1). Results: A total of 482,479 drug-event pairs were recorded in the RNF database; among them, 74,324 were attributed to vaccines and 279 to allergens. After exclusion of these records, we included 407,876 pairs, with 139 lymphoma, 137 melanoma and 280 NMSC cases. The most represented indication of use was rheumatoid arthritis with 56 reports. Anti-TNF drugs yielded an information component (IC) of 4.06 (CrI, credible intervals, 3.59–4.52) and an aROR of 37.1 (CI 95, 95% confidence intervals, 25.5–53.9) for lymphoma; an IC of 3.95 (3.47–4.43) and an aROR of 31.1 (21.1–45.7) for melanoma and an IC of 3.52 (3.11–3.93) and an aROR of 17.2 (12.7–23.3) for NMSC. The highest IC, based on Bayesian confidence propagation, was found for adalimumab, for both lymphoma (IC: 3.97 [3.32–4.63]), skin cancers (IC 3.62 [2.89–4.36]) and melanoma, IC 3.59 [2.97–4.20] for NMSC). Conclusions: Several previous reports from the RNF suggested a possible signal of risk for anti-TNF drugs and the development of some types of cancer. Notably, investigating on the risk of such cancers in patients exposed to anti-TNF drugs is complicated by the intrinsic risk associated with the underlying diseases [3–5]. For example, for patients with rheumatoid arthritis the risk for lymphoma and skin cancer is two-fold increased [5]. Well-designed observational studies are warranted to correctly assess the potential tumorigenic effects associated with anti-TNF drugs.
The Anti-TNF Drugs and the Risk of Lymphoma, Skin Cancer and Melanoma: A Signal Detection Analysis
Leonardi, L;Fornili, M;Convertino, I;Lucenteforte, E;Ferraro, S;Blandizzi, C;Baglietto, L;
2019-01-01
Abstract
Introduction: Tumor necrosis factor (TNF) dysregulation plays a central role in the aetiology of several inflammatory diseases. Currently there are five approved TNF-inhibitors: infliximab, etanercept, adalimumab, certolizumab pegol and golimumab. Anti-TNF drugs are usually well tolerated, however, there are some concerns regarding their safety profile, particularly for their potential tumorigenic effect [1, 2]. Aim: The aim of this study is to perform a disproportionality analysis using the Italian National Network of Pharmacovigilance (RNF) database of the Italian Medicines Agency to assess the occurrence of a signal of risk for the drug event pairs TNF inhibitors-lymphoma and TNF inhibitors-skin cancers. Methods: This study included all reports recorded in the RNF database from 1988 up to October 31st, 2017. Cases were defined as all reports with MedDRA Preferred Term (PT) codes of lymphoma, melanoma and non-melanoma skin cancer (NMSC). Index reports included all reports in which at least one anti-TNF drug (infliximab, etanercept, adalimumab, golimumab and certolizumab pegol) was reported as the suspected causative drug. Reference reports included all cases related to non-anti-TNF drugs. Bayesian (information component, IC) and frequentist (adjusted reporting odds ratio, aROR) methods were used to calculate disproportional distribution of drug- ADR pairs. ROR was adjusted for sex, age quartiles, reporting year, number of suspected drugs (1 or C 2) and number of concomitant drugs (0 or C 1). Results: A total of 482,479 drug-event pairs were recorded in the RNF database; among them, 74,324 were attributed to vaccines and 279 to allergens. After exclusion of these records, we included 407,876 pairs, with 139 lymphoma, 137 melanoma and 280 NMSC cases. The most represented indication of use was rheumatoid arthritis with 56 reports. Anti-TNF drugs yielded an information component (IC) of 4.06 (CrI, credible intervals, 3.59–4.52) and an aROR of 37.1 (CI 95, 95% confidence intervals, 25.5–53.9) for lymphoma; an IC of 3.95 (3.47–4.43) and an aROR of 31.1 (21.1–45.7) for melanoma and an IC of 3.52 (3.11–3.93) and an aROR of 17.2 (12.7–23.3) for NMSC. The highest IC, based on Bayesian confidence propagation, was found for adalimumab, for both lymphoma (IC: 3.97 [3.32–4.63]), skin cancers (IC 3.62 [2.89–4.36]) and melanoma, IC 3.59 [2.97–4.20] for NMSC). Conclusions: Several previous reports from the RNF suggested a possible signal of risk for anti-TNF drugs and the development of some types of cancer. Notably, investigating on the risk of such cancers in patients exposed to anti-TNF drugs is complicated by the intrinsic risk associated with the underlying diseases [3–5]. For example, for patients with rheumatoid arthritis the risk for lymphoma and skin cancer is two-fold increased [5]. Well-designed observational studies are warranted to correctly assess the potential tumorigenic effects associated with anti-TNF drugs.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
