Predisposing factors of secukinumab discontinuation at 1 year (multivariate analysis)
| Predictive factor | Modality* (N) | Multivariate Cox regression | ||
| HR (95% CI) | P value† | P type III | ||
| At least one objective sign of inflammation | No (N=165)* | |||
| Yes (N=711) | 1.44 (1.08 to 1.93) | 0.014 | ||
| Secukinumab treatment line | First line (N=68)* | 0.084 | ||
| Second line (N=132) | 1.53 (0.91 to 2.57) | 0.107 | ||
| ≥Third line (N=676) | 1.67 (1.06 to 2.62) | 0.028 | ||
| Past or present history of IBD | No (N=854)* | |||
| Yes (N=22) | 1.76 (1.01 to 3.07) | 0.047 | ||
| History of depression or anti-depressive concomitant treatment | No (N=716)* | |||
| Yes (N=160) | 1.25 (0.97 to 1.60) | 0.090 | ||
*The modality given in the first row (eg, ‘No’ for ‘At least one objective sign of inflammation’) defined the reference in the Cox model.
†P value calculated against the reference value. Factors with >20% missing data such as age, gender, BMI, smoking status, duration of disease, axial involvement, past or present history of synovitis, psoriasis, uveitis, and fibromyalgia, initial secukinumab dose, concomitant treatment with csDMARDs, oral corticosteroids, HLA-B27 positivity and sacroiliitis were not included in the multivariate Cox analysis.
BMI, body mass index; csDMARDs, conventional synthetic disease-modifying anti-rheumatic drugs; HLA-B27, human leucocyte antigen B-27; IBD, inflammatory bowel disease; N, size of the population.