Journal of Southern Medical University ›› 2021, Vol. 41 ›› Issue (5): 633-639.doi: 10.12122/j.issn.1673-4254.2021.05.01

    Next Articles

Relapse of ankylosing spondylitis and its predictors after withdrawal of tumor necrosis factor-α inhibitors: a 52-week follow-up study

  

  • Online:2021-05-20 Published:2021-06-11

Abstract: Objective To investigate the recurrence of ankylosing spondylitis (AS) that has been relieved by standard-dose adalimumab (ADA) after dose reduction or withdrawal of tumor necrosis factor-α inhibitor (TNFi) and explore the factors that predict AS occurrence. Methods This study was conducted among 63 patients with AS who reduced the dose of or discontinued TNFi after completing at least 12 treatment cycles with ADA (40 mg/2 weeks) to achieve ASAS20 improvement with a BATH disease activity index (BASDAI) <4 for more than 8 weeks. The patients were followed up every 12 weeks for a total of 52 weeks, and the recurrence of AS, changes of BASDAI, C-reactive protein (CRP)-based disease activity score (ASDAS-CRP), low back pain (LBP) score, Bath Ankylosing Spondylitis Metrology Index (BASMI), CRP and ESR were recorded and analyzed. Cox regression model and ROC curve analyses were performed to analyze the risk factors of AS relapse after dose reduction or discontinuation of TNFi. Results Of the 63 patients enrolled, 57 completed the follow-up study, among whom 22 (38.6%) patients experienced AS relapse within 52 weeks, with a median clinical recurrence time of 31 weeks. The recurrence rate of AS was significantly higher in patients with complete withdrawal of medications (89.0%) than in those with TNFi dose reduction and TNFi discontinuation (P<0.001), and did not differ significantly between the latter two groups of patients (χ2= 0.071, P=0.791). The Cox regression model showed that a high baseline LBP score (HR=1.438, P=0.027) and a high BASMI score (HR=1.29, P=0.049) were the risk factors for AS recurrence after TNFi dose reduction or discontinuation, while maintenance of medication during follow-up was a protective factor (HR=0.209, P=0.001). ROC curve analysis showed that the combination of baseline LBP score, BASMI and medication during follow-up had a good predictive value for AS relapse (AUC=0.819) with a sensitivity of 0.772 and a specificity of 0.718. Conclusion Dose reduction or discontinuation of TNFi is associated with a high recurrence rate of AS that has been relieved by TNFi treatment. A high LBP score, a high BASMI score and discontinuation of maintenance medication are the risk factors for AS recurrence in patients after dose reduction or withdrawal of TNFi.

Key words: ankylosing spondylitis; tumor necrosis factor-α inhibitor; recurrence; risk factors