INTRODUCTION: Spondyloarthritis (SpA), an interrelated group of rheumatic diseases, has been suggested to be triggered by bacterial infections prior to the development of an autoimmune response that cause inflammation of the spinal and peripheral joints. Since human heat shock protein (HSP) 60, recently named HSPD1, and bacterial HSP60 are highly homologous, immunological cross-reactivity has been proposed as a mechanism of disease initiation. However, previous investigations of the humoral immune response to HSP60 in SpA patients have lacked determination of Immunoglobulin G (IgG) subclasses and patient follow-up. In this study we have focused on these parameters in a cohort group of axial SpA patients with a well-established set of clinical characteristics including MRI changes and human leukocyte antigen (HLA) B27. METHODS: IgG subclass antibodies (IgG1, IgG2, IgG3 and IgG4) against recombinant HSP60 of three reactive arthritis-related bacteria; human HSP60; and against the microorganisms C. trachomatis and Chlamydia pneumoniae were determined by ELISA. Serum samples collected in 2004-06 and in 2010-11 from 39 axial SpA patients were analysed and compared with 39 healthy controls. Mann Witney U-test and Wilcoxon matched pairs test were used to compare the antibody levels in different and paired groups, respectively. Probabilities <0.01 were considered as significant. Spearman nonparametric correlation was used to determine correlation between antibody levels, and between antibody levels and the disease parameters. RESULTS: Elevated levels of IgG1 and IgG3 to human HSP60 and IgG1 to HSP60 of Salmonella enterica Enteritidis were observed in SpA patients compared with healthy controls at both time points. The antibody levels were almost constant over time for IgG1, while high levels of IgG3 to human HSP60 tended to decrease over time. The antibody response to human HSP60 was predominantly of the IgG3 subclass, and patients with high levels of IgG3 to this antigen had low levels of IgG1 indicating an inverse association. Different IgG subclasses were produced against bacterial and human HSP60 in the same serum sample, IgG1 and IgG3 respectively, indicating that there was no cross-reaction. CONCLUSIONS: Significant association was observed between axial SpA and the presence of IgG1/IgG3 antibodies to human HSP60 and IgG1 to S. enterica Enteritidis and C. trachomatis. Generation of antibodies to human HSP60 was independent of the presence of antibodies to bacterial HSP60. No association was observed between clinical and MRI changes with antibodies over time. Altogether, such antibodies do not reflect the disease activity in these patients.