Objectives To evaluate the reliability of consensus-based ultrasound (US) definitions of elementary components of enthesitis in spondyloarthritis (SpA) and psoriatic arthritis (PsA) and to evaluate which of them had the highest contribution to defining and scoring enthesitis. Methods Eleven sonographers evaluated 40 entheses from five patients with SpA/PsA at four bilateral sites. Nine US elementary lesions were binary-scored: Hypoechogenicity, thickened insertion, enthesophytes, calcifications, erosions, bone irregularities, bursitis and Doppler signal inside and around enthesis. Kappa statistics were used to evaluate reliability. Sonographers were also asked to state which lesions can be considered as inflammatory or structural and should be included in the final definition of enthesitis. Only the lesions, scored as present in at least 75% of the entheses considered as having an enthesitis, were included in the final definition. Results The prevalence of detected lesions was quite low except for enthesophytes (55%) and bone irregularities (54%). Reliability ranged from poor to good (the lowest for thickened enthesis (kappa 0.1 (95% CI 0 to 0.7)) and the highest for enthesophytes (kappa 0.6 (95% CI 0.5 to 0.7)). When adjusted for low prevalence, kappa values increased for all lesions, with the best result observed for detecting Doppler signal at insertion (0.9) and for bursitis (0.8). The US components included in the final definition were hypoechogenicity, increased thickness at enthesis, erosions and calcifications/enthesophytes and Doppler signal at insertion. Conclusion By using a consensus-based stepwise approach, a final reliable US score and definition of enthesitis in SpA/PsA were produced. Further studies are sought for implementing this score in clinical trials and practice.