Pathogenic hypotheses for spondyloarthropathies are evolving. Several candidates have been added to the list of inciting microorganisms, and genes other than HLA-B27 are under scrutiny. Above all, the chiefly 'immunological' theory of spondyloarthropathies incriminating a cross-reaction between self-proteins and bacterial peptides is giving way to a more 'microbiological' concept in which latent bacteria residing within macrophagic or dendritic cells undergo reactivation through a process facilitated by HLA-B27. This molecule is prone to misfolding, which decreases the presentation of bacterial peptides to the immune system and stimulates the Nf-KB inflammation pathway within infected macrophages and/or dendritic cells. Migration of these infected cells from the mucous membranes to the tissues targeted by spondyloarthropathies, particularly the bone marrow located near entheses, may facilitate transient reactivation of dormant intracellular bacteria by creating a favorable cytokine environment. This environment may include high levels of TGFbeta and IL-10, noted also at other sites that enjoy immune privilege, such as the eye. The reactivation may be stopped by a local response of CD4+ and/or CD8+ T cells at the expense of local inflammation responsible for clinical manifestations. This scenario seems consistent with results from studies of murine models transgenic for the HLA-B27 antigen: exposure to bacteria is necessary to the development of spondyloarthropathy, but the disease occurs even when only the heavy chain of HLA-B27 is present (i.e., beta2-microgloblin is not indispensable). Improved understanding of the mechanisms that confer to some bacterial strains a strong potential for persisting within cells, including macrophagic cells, may open the way toward new treatment approaches capable of complementing antagonists of TNF-alpha and other monokines, which merely suspend the disease process, and antibiotic therapy, which fails to kill dormant bacteria located within cells. Pathogenic hypotheses for spondyloarthropathies are evolving. This review presents the most recent concepts. These concepts have not all received confirmation from experimental data. However, the high degree of consistency among them prompted us to consolidate them into a single picture. Although this approach may yield a motley composite of fact and speculation, it may open up new avenues of thought for rheumatologists interested in the links between chronic intracellular infections and inflammatory joint disease.