Table 2

Efficacy of tofacitinib 5 mg two times per day versus placebo at week 16: type I error-controlled primary and secondary endpoints†

Tofacitinib 5 mg two times per day (N=133)Placebo (N=136)p value
Global type I error-controlled endpoints at week 16, tested in the sequence below
 ASAS20 response,‡ n (%)75 (56.4)40 (29.4)<0.0001***§
 ASAS40 response,‡ n (%)54 (40.6)17 (12.5)<0.0001***§
 ∆ASDAS,¶ LSM (SE) (N1)−1.36 (0.07) (129)−0.39 (0.07) (131)<0.0001***§
 ∆hsCRP (mg/dL),¶ LSM (SE) (N1)−1.05 (0.10) (129)−0.09 (0.10) (131)<0.0001***§
 ∆ASQoL,** LSM (SE) (N1)−4.03 (0.40) (129)−2.01 (0.41) (130)0.0001***§
 ∆SF-36v2 PCS score,** LSM (SE) (N1)6.69 (0.59) (129)3.14 (0.59) (130)<0.0001***§
 ∆BASMI,¶ LSM (SE) (N1)−0.63 (0.06) (129)−0.11 (0.06) (131)<0.0001***§
 ∆FACIT-F total score,¶ LSM (SE) (N1)6.54 (0.80) (129)3.12 (0.79) (131)0.0008***§
Type I error-controlled ∆ASAS components at week 16¶§§ tested in the sequence below
 ∆PtGA (NRS 0–10), LSM (SE) (N1)−2.47 (0.20) (129)−0.91 (0.20) (131)<0.0001***††
 ∆Total back pain (NRS 0–10), LSM (SE) (N1)−2.57 (0.19) (129)−0.96 (0.19) (131)<0.0001***††
 ∆BASFI (NRS 0–10), LSM (SE) (N1)−2.05 (0.17) (129)−0.82 (0.17) (131)<0.0001***††
 ∆Morning stiffness (inflammation, NRS 0–10),‡‡ LSM (SE) (N1)−2.69 (0.19) (129)−0.97 (0.19) (131)<0.0001***††
  • Data are from the week 16 analysis: data cut-off 19 December 2019; data snapshot 29 January 2020.

  • ***p<0.001 for comparing tofacitinib 5 mg two times per day versus placebo.

  • †In each family of type I error-controlled endpoints, statistical significance could be declared only if the prior endpoint in the sequence met the requirements for significance (p≤0.05).

  • ‡Normal approximation adjusting for the stratification factor (bDMARD treatment history: bDMARD-naïve versus TNFi-IR or prior bDMARD use without IR) derived from the clinical database via the Cochran-Mantel-Haenszel approach was used. Missing response was considered as non-response.

  • §p≤0.05 for comparing tofacitinib 5 mg two times per day versus placebo, according to the prespecified step-down testing procedure for global type I error control.

  • ¶Mixed model for repeated measures included fixed effects of treatment group, visit, treatment group by visit interaction, stratification factor derived from the clinical database, stratification factor by visit interaction, baseline value, and baseline value by visit interaction. The model used a common unstructured variance–covariance matrix, without imputation for missing values.

  • **Analysis of covariance model included fixed effects of treatment group, stratification factor derived from the clinical database and baseline value. Missing values were not imputed.

  • ††p≤0.05 for comparing tofacitinib 5 mg two times per day versus placebo, according to the prespecified step-down testing procedure for type I error control of ASAS components.

  • ‡‡Morning stiffness (inflammation) assessed as mean of questions 5 and 6 of the BASDAI.

  • §§Endpoints were tested in sequence after ASAS20 response at week 16 met the requirements for significance (p≤0.05).

  • ∆, change from baseline; ASAS20, ASAS ≥20% improvement; ASAS40, ASAS ≥40% improvement; ASAS, Assessment of SpondyloArthritis international Society; ASDAS, Ankylosing Spondylitis Disease Activity Score using hsCRP; ASQoL, Ankylosing Spondylitis Quality of Life; BASDAI, Bath Ankylosing Spondylitis Disease Activity Index; BASFI, Bath Ankylosing Spondylitis Functional Index; BASMI, Bath Ankylosing Spondylitis Metrology Index; bDMARD, biologic disease-modifying antirheumatic drug; FACIT-F, Functional Assessment of Chronic Illness Therapy-Fatigue; hsCRP, high-sensitivity C-reactive protein; IR, inadequate response or intolerance; LSM, least squares mean; n, number of patients with response; N, number of patients in full analysis set; N1, number of patients with observation at week 16; NRS, Numerical Rating Scale; PtGA, Patient Global Assessment of Disease Activity; SF-36v2, Short Form-36 Health Survey Version 2 Physical Component Summary; TNFi, tumour necrosis factor inhibitor.