[The concept of reactive arthritis].
More than fifty years after the clinical syndrome was described by Reiter, Fiessinger and Leroy, the diagnostic criteria of what has now been termed reactive arthritis are still under debate. Reactive arthritis can be defined as an aseptic inflammatory joint disease occurring either in a patient with a bacterial infection located in a distant organ or with a particular genetic predisposition. The lack of specific criteria has led to a certain degree of liberty in interpreting the three basic elements of reactive arthritis: clinical presentation, "distant" infection, genetic predisposition. It is generally accepted that limiting the diagnosis of reactive arthritis only to patients with the classic oculo-uretro-synovial manifestations is unsatisfactory. The question remains open as to whether only patients presenting both inflammatory joint disease and spondylarthropathy have reactive arthritis or whether any undetermined etiology of arthritis is sufficient if the patient also has a "distant" infection and a specific genetic predisposition (HLA B27). Two new elements further complicate the situation. The first is the recent demonstration that the synovial fluid in patients with reactive arthritis might contain microbial antigens. The second is that longterm antibiotics (more than 3 months) could have a beneficial effect. Asking when the diagnosis should be entertained is thus not a moot question. Although we still have no clear answer, it is reasonable to state that the diagnosis of reactive arthritis should be retained only in patients with both inflammatory joint disease and criteria compatible with spondylarthropathy. Clinical research is required to determine whether the germ is actually present or not within the joint and to assess the effect of long-term antibiotherapy.