The therapeutic approach to JIA is sometimes very troublesome and progression to erosive polyarthritis may occur in all JIA categories. Only Methotrexate has shown efficacy and safety in a large controlled trial. Nevertheless, in many cases, drug resistance or intolerance has led to try other therapeutic options, with still debatable results. Therefore, there has been space, in the last few years, for new therapies as the TNF-inhibitors. This therapeutic approach has shown a dramatic clinical benefit in active polyarticular refractory JIA: the rate and rapidity of response have exceeded those of all other studied DMARDs. Preliminary data show that they are efficacious also for other pediatric rheumatic disease (spondyloarthropathies, autoimmune uveitis, dermatomyositis, Kawasaki syndrome and some autoinflammatory diseases). TNF-inhibitors in JIA have demonstrated a favourable benefit-to-risk profile. However, as their use has increased worldwide, some unusual, usually not serious, adverse events have emerged. Severe infections, including TB, and deaths have been reported. Long-lasting active disease, systemic disease, concurrent and previous immunosuppressive therapies, all contribute to risk of infection and other serious AEs. Given the evidence that TNF has a primary role in the pathogenesis of JIA, particularly in joint destruction, neutralizing this cytokine early, within the window of opportunity, could halt or delay progression of joint damage and debilitating consequences of the disease. Thus, for JIA patients whose disease is not quickly controlled with MTX, TNF blockers may be considered as first-line treatment, although long-term safety data still need to be established.