Objectives: To compare the prevalence and correlates of metabolic syndrome (MetS) stratified by body mass index (BMI) categories in rheumatoid arthritis (RA) and spondyloarthritis (SpA).Methods: The age- and sex-standardized prevalence of MetS was calculated by BMI categories and compared between RA and SpA patients before starting first biologic, and controls. The determinants of metabolic syndrome in patients without obesity were investigated.Results: MetS was observed in 28% of RA (21/75), 22.5% of SpA (18/80), 19% of controls (187/998). The age- and sex-standardized prevalence of MetS was not significantly different between RA 19% (95% CI: 11-27%), SpA 26% (95% CI: 16-36%) and controls 16% (95% CI: 14-18%). When stratified by BMI, the standardized prevalence of MetS was less frequent in obese RA patients (15%, 95% CI: 4-27%) compared to obese controls (48%, 95% CI: 40-55%) or to obese SpA (36%, 95% CI: 26-45%). In normal-weight RA patients, MetS standardized prevalence was 16% (95% CI: 7-25%) compared to 5% (95% CI: 0-11%) in SpA, and 6% (95% CI: 4-8%) in controls. In non-obese SpA, MetS was associated with abdominal obesity, visceral fat mass and cardiovascular risk. In non-obese RA patients with metabolic syndrome, body composition did not differ from metabolically healthy RA patients.Conclusions: MetS is not uniform among patients with similar BMI. In RA, MetS was less frequent in obese patients, and unlike SpA, was not associated with body fat composition in non-obese patients. Differences between RA and SpA for metabolic health suggest various pathophysiological mechanisms.