ObjectiveTo assess the incidence of orthopaedic surgery (OS) (including total hip arthroplasty (THA), total knee arthroplasty, upper limb arthroplasty, arthrodesis and spinal surgery) and associated trends in patients with spondyloarthritis (SpA) over a long period (17 years).MethodsAn observational, retrospective, population-based, serial cross-sectional study was conducted. All hospital admissions of patients with SpA reported between 1999 and 2015 were analysed, and a control group was selected and matched by age, sex and year of admission. Incidence rates for OS (and subtypes) were calculated. Generalised linear models were used to analyse trends; unconditional logistic regression models were used to calculate crude and adjusted ORs (aORs) with the aim of evaluating the association between OS and SpA.ResultsThe study database contained data on 214 280 hospital admissions (SpA/non-SpA 1:1 ratio). In the SpA cohort, 5 382 admissions (5.02%) had undergone OS compared with 3 533 in the non-SpA cohort (3.29%) (AOR 1.64; 95% CI 1.57 to 1.72). OS rates increased for both cohorts (+4.92% per year vs +8.41%). The trend in OS, THA, arthrodesis and spinal surgery decreased or stabilised in patients under age 60 in the SpA cohort, while the non-SpA cohort remained stable. In the SpA cohort, the mean age was 53.68 years (SD 13.65) in 1999, increasing to 62.76 years (SD 12.74) in 2015. In the non-SpA cohort, the mean age remained stable at around 63 years.ConclusionsA 9-year difference in the age of patients undergoing OS was observed in patients with SpA. The incidence of OS, THA and arthrodesis decreased in patients under age 60, and the incidence of spinal surgery decreased in patients under age 40. Our findings suggest that these patients are increasingly able to defer surgical interventions.