Reactive arthritis (ReA) is definitely caused by an infection. Several observations suggest that the triggering microbe may persist in the tissues of the patient for a prolonged time. The obvious conclusion is to consider antibacterial treatment. In two instances antibacterial agents are of definite value: in the primary and secondary prevention of rheumatic fever and for early eradication of Borrelia burgdorferi in order to prevent development of the arthritis associated with Lyme disease. Altogether, clinical and experimental data exist to indicate that if antibacterial treatment of ReA can be started very early during the pathogenetic process, the disease can be prevented or the prognosis improved. In fully developed ReA, the value of antibacterial agents is less certain. All available evidence indicates that short term antibacterial treatment has no effect on the prognosis and final outcome of ReA, and the results with long term administration of antibacterials are also overall poor. In some instances sulfasalazine appears useful, rather as a result of its antirheumatic effect or influence on an underlying inflammatory bowel disease than its action as an antibacterial agent. Tetracyclines have also been found to have an effect on ReA, but again, this is probably due to their anti-inflammatory action rather than any antibacterial effect.