Efficacy and Safety of Tofacitinib in Subjects With Active Ankylosing Spondylitis (AS)

ASAS20 assess 4 domains: Patient Global Assessment of Disease (PGA) (assess disease activity on a scale of 0 \[not active\] to 10 \[very active\], high score=more disease activity), total back pain (scale of 0 \[no pain\] to 10 \[most severe pain\], high score=more severity), Function (Bath Ankylosing Spondylitis Functional Index \[BASFI\]; participant's level of ability on scale of 0 \[easy\] to 10 \[impossible\], low score= better functional health) and Inflammation (morning stiffness, Mean of Question \[Q\]5 and Q6 of Bath Ankylosing Spondylitis Disease Activity Index \[BASDAI\] defined as 6-item questionnaire measure disease activity on a scale of 0 \[none\] to 10 \[severe\], high score=more disease activity). ASAS20 response: greater than or equal to (\>=) 20 percent (%) improvement from baseline in disease activity and absolute change of \>=1 unit in \>=3 domains and no worsening of \>=20% and an absolute change of \>=1 unit in remaining domain.

Time frame: Week 16

Population: Full Analysis Set (FAS): included all participants who were randomized to the study and received at least one dose of the randomized investigational product (i.e., tofacitinib or placebo). Here, on-drug data was used and missing response (MR) was considered to be Non-response (NR) (MR=NR).

ASDAS(CRP) was derived using BASDAI (6-item questionnaire measures disease activity on a scale of 0 \[none\] to 10 \[severe\], higher score=more disease activity) and PGA (measure disease activity on a scale of 0 \[not active\] to 10 \[very active\], high score=more disease activity) and calculated by using the following formula, 0.121 x Back Pain (Q2 of BASDAI) + 0.058 x Duration of Morning Stiffness (Q6 of BASDAI) + 0.110 x PGA + 0.073 x Peripheral Pain/Swelling (Q3 of BASDAI) + 0.579 x Ln(high sensitivity \[hs\] CRP mg/Liter \[L\] + 1). If hsCRP values were smaller than 2 mg/L, they were set to 2 mg/L in the formula. The range of score was \>= 0.636 to no defined upper limit. A negative change from baseline value indicates decrease in disease activity; a positive change from baseline value indicates increase in disease activity.

Time frame: Baseline, Weeks 2, 4, 8, 12, 16, 24, 32, 40 and 48

Population: Full Analysis Set (FAS): included all participants who were randomized to the study and received at least one dose of the randomized investigational product (i.e., tofacitinib or placebo). Here, on-drug data was used and missing response was not imputed. Here Overall number of participants analyzed signifies participants evaluable for this outcome measure.

The ASQoL was an 18-item questionnaire assessed the amount of restriction participant experienced in daily activities, level of pain and fatigue, and the impact on the participant's emotional state. Each item was scored as 0 (no impact) or 1 (yes - impact). A total score was calculated by summing the items. The total score ranged from 0 (no impact) to 18 (yes-impact), with higher values indicated more impaired health-related quality of life.

Time frame: Baseline, Weeks 16 and 48

Population: FAS: included all participants who were randomized to the study and received at least one dose of the randomized investigational product (i.e., tofacitinib or placebo). Here, on-drug data was used and missing response was not imputed. Here Overall number of participants analyzed signifies participants evaluable for this outcome measure. Here, number analyzed signifies participants evaluable for this outcome measure for specified time point.

The BASMI was used to assess the axial status and spinal mobility (cervical, dorsal and lumbar spine, hips and pelvic soft tissue) and was analyzed using the linear function method. BASMI score composed of five clinical measures: lateral spinal flexion, tragus-to-wall distance, lumbar flexion (modified Schober), maximal intermalleolar distance, and cervical rotation. BASMI - Linear Method score was the average of 5 individual component scores mapped between 0 and 10 and thus the BASMI - Linear Method total score ranged from 0 (very good) to 10 (very poor), wherein higher scores indicated more impairment of axial status and spinal mobility; lower scores indicated better health status.

Time frame: Baseline, Weeks 2, 4, 8, 12, 16, 24, 32, 40 and 48

Population: FAS: included all participants who were randomized to the study and received at least one dose of the randomized investigational product (i.e., tofacitinib or placebo). Here, on-drug data was used and missing response was not imputed. Here Overall number of participants analyzed signifies participants evaluable for this outcome measure.

BASMI assess axial status and spinal mobility. It composed of 5 clinical measures: lateral spinal flexion, tragus-to-wall distance, lumbar flexion (modified Schober), maximal intermalleolar distance, cervical rotation. BASMI - Linear Method score was average of 5 individual component score mapped between 0 and 10, high score=more impairment of axial status, spinal mobility. For cervical rotation angle, participant sit straight on chair with chin level and hands on knees. Blinded assessor place goniometer at top of head in line with nose and ask participant to rotate neck maximally to left, follows with goniometer and record angle between sagittal plane and new plane after rotation. A second reading obtained and both readings recorded. Procedure repeated for right side. Better of two for each side was selected for scoring. Scoring done by calculating mean of left and right measurement and recorded in degrees (range: 0 to 90), higher cervical rotation value=better health status.

Time frame: Baseline, Weeks 2, 4, 8, 12, 16, 24, 32, 40 and 48

Population: FAS: included all participants who were randomized to the study and received at least one dose of the randomized investigational product (i.e., tofacitinib or placebo). Here, on-drug data was used and missing response was not imputed. Here Overall number of participants analyzed signifies participants evaluable for this outcome measure.

BASMI assessed axial status and spinal mobility, using linear function. It composed of 5 clinical measures: lateral spinal flexion, tragus-to-wall distance, lumbar flexion (modified Schober), maximal intermalleolar distance, cervical rotation. BASMI - Linear Method score was average of 5 individual component scores mapped between 0 and 10, with higher scores indicating more impairment of axial status and spinal mobility. For the assessment of intermalleolar distances, participant should lie supine with the knees straight and feet/toes pointing straight up and asked to separate the legs as far as possible and the distance between the medial malleoli was measured (in Centimeters \[cm\] to the nearest 0.1 cm). Distance (in cm) was greater than or equal to 0, with no maximum defined range: higher intermalleolar distance value indicates better health status.

Time frame: Baseline, Weeks 2, 4, 8, 12, 16, 24, 32, 40 and 48

Population: FAS: included all participants who were randomized to the study and received at least one dose of the randomized investigational product (i.e., tofacitinib or placebo). Here, on-drug data was used and missing response was not imputed. Here Overall number of participants analyzed signifies participants evaluable for this outcome measure.

BASMI assess axial status, spinal mobility, using linear function. It composed of 5 clinical measures. BASMI - Linear Method score-average of 5 component scores mapped between 0 and 10, with high scores =more impairment of axial status, spinal mobility. For assessment of lateral spinal flexion: participant stand upright with head and back rest against wall as close as possible with shoulders level and feet 30 cm apart and feet parallel. At tip of middle finger, place a mark on thigh. This neutral position recorded. Participant bend sideways without bending knees or lifting heels while attempting to keep shoulders in same position (flexion position). Second mark placed, lateral flexion recorded (left or right as appropriate) using cm tape measure. Two tries for left, two tries for right measured. Result of two tries recorded for left and right separately in cm to nearest 0.1 cm. Distance (in cm) should be \>=0, with no maximum defined range: high value indicates better health status.

Time frame: Baseline, Weeks 2, 4, 8, 12, 16, 24, 32, 40 and 48

Population: FAS: included all participants who were randomized to the study and received at least one dose of the randomized investigational product (i.e., tofacitinib or placebo). Here, on-drug data was used and missing response was not imputed. Here Overall number of participants analyzed signifies participants evaluable for this outcome measure.

BASMI assessed axial status and spinal mobility. BASMI - Linear Method score was average of 5 individual component scores mapped between 0 and 10, with higher scores indicating more impairment of axial status and spinal mobility. For the assessment of lumbar flexion, With the participant standing erect and outer edges of feet 30 cm apart, a mark was placed in the midpoint of a line that joins the posterior superior iliac spines (baseline mark). A second mark (A) was placed 10 cm above the baseline mark and a third mark (B) 5 cm below the baseline mark. Then have the participant maximally bend forward, keeping the knees fully extended. With the participant's spine in full flexion, the distance between marks A and B (in cm to the nearest 0.1 cm) was re-measured. Distance (in cm) was greater than or equal to 0, with no maximum defined range. Higher value indicates better health status.

Time frame: Baseline, Weeks 2, 4, 8, 12, 16, 24, 32, 40 and 48

Population: FAS: included all participants who were randomized to the study and received at least one dose of the randomized investigational product (i.e., tofacitinib or placebo). Here, on-drug data was used and missing response was not imputed. Here Overall number of participants analyzed signifies participants evaluable for this outcome measure.

BASMI assessed axial status and spinal mobility, using linear function. It composed of 5 clinical measures: lateral spinal flexion, tragus-to-wall distance, lumbar flexion (modified Schober), maximal intermalleolar distance, cervical rotation. BASMI - Linear Method score was average of 5 individual component scores mapped between 0 and 10, with higher scores indicating more impairment of axial status and spinal mobility. For the assessment of tragus-to-wall distance, participant was placed standing with his/her back against the wall; knees straight; scapulae, buttocks, and heels against wall; and head in as neutral position as possible. The distance between the tragus and wall in cm was measured (to the nearest 0.1 cm) from both the right side and left side at the maximum effort to touch the head against the wall. Distance should be greater than or equal to 0 cm with no defined maximum value, lower tragus-to-wall value indicates better health status.

Time frame: Baseline, Weeks 2, 4, 8, 12, 16, 24, 32, 40 and 48

Population: FAS: included all participants who were randomized to the study and received at least one dose of the randomized investigational product (i.e., tofacitinib or placebo). Here, on-drug data was used and missing response was not imputed. Here Overall number of participants analyzed signifies participants evaluable for this outcome measure.

EQ-5D-3L, a health profile questionnaire was used to assess quality of life along 5 dimensions. Participants rated 5 aspects of health (mobility, self-care, usual activities, pain/discomfort, and anxiety/depression) by choosing from 3 answering options (1=no problems; 2=some problems; 3=extreme problems). The mean of the summed score ranged from 1 to 3 with 1 corresponding to no problems and 3 corresponding to severe problems in the 5 dimensions, where higher score indicates more severe problems.

Time frame: Baseline, Weeks 16 and 48

Population: FAS: included all participants who were randomized to the study and received at least one dose of the randomized investigational product (i.e., tofacitinib or placebo). Here, on-drug data was used and missing response was not imputed. Here Overall number of participants analyzed signifies participants evaluable for this outcome measure.

EQ-5D-3L, a health profile questionnaire was used to assess quality of life along 5 dimensions. Its second part included EQ-VAS. EQ-VAS recorded the participant's self-rated health on a VAS ranging from 0 mm (worst imaginable health state) to 100 mm (best imaginable health state), with higher scores indicating better health state.

Time frame: Baseline, Weeks 16 and 48

Population: FAS: included all participants who were randomized to the study and received at least one dose of the randomized investigational product (i.e., tofacitinib or placebo). Here, on-drug data was used and missing response was not imputed. Here Overall number of participants analyzed signifies participants evaluable for this outcome measure. Here, number analyzed signifies participants evaluable for this outcome measure for specified rows.

FACIT-F is a 13-item (felt fatigued, felt weak all over, felt listless \[washed out\],felt tired, had energy, had trouble starting things as tired, had trouble finishing things as tired, was able to do usual activities, needed to sleep during day, too tired to eat, needed help doing my usual activities, frustrated by being too tired to do things wanted to do, had to limit my social activity because tired) questionnaire, with each item scored on a 5-point scale ranging from 0 (not at all) to 4 (very much). FACIT-F experience domain score was calculated by summing 5 items: I feel fatigued, I feel weak all over, I feel listless (washed out), I feel tired, and I have energy. FACIT-F total experience domain score ranged from 0 (not at all) to 20 (very much), with higher scores represented better (less) fatigue impact on daily functioning. In this outcome measure, change from baseline in FACIT-F experience domain score was reported.

Time frame: Baseline, Weeks 2, 4, 8, 12, 16, 24, 32, 40 and 48

Population: FAS: included all participants who were randomized to the study and received at least one dose of the randomized investigational product (i.e., tofacitinib or placebo). Here, on-drug data was used and missing response was not imputed. Here Overall number of participants analyzed signifies participants evaluable for this outcome measure.

FACIT-F is a 13-item (felt fatigued, felt weak all over, felt listless \[washed out\],felt tired, had energy, had trouble starting things as tired, had trouble finishing things as tired, was able to do usual activities, needed to sleep during day, too tired to eat, needed help doing my usual activities, frustrated by being too tired to do things wanted to do, had to limit my social activity because tired) questionnaire, with each item scored on a 5-point scale ranging from 0 (not at all) to 4 (very much). FACIT-F experience domain score calculated by summing 5 items : I feel fatigued, I feel weak all over, I feel listless, I feel tired, and I have energy, while FACIT-F impact domain score was calculated by summing the remaining 8 items. FACIT-F impact domain score ranged from 0 (not at all) to 32 (very much), with higher scores represented better (less) fatigue impact on daily functioning. In this outcome measure, change from baseline in FACIT-F impact domain score was reported.

Time frame: Baseline, Weeks 2, 4, 8, 12, 16, 24, 32, 40 and 48

Population: FAS: included all participants who were randomized to the study and received at least one dose of the randomized investigational product (i.e., tofacitinib or placebo). Here, on-drug data was used and missing response was not imputed. Here Overall number of participants analyzed signifies participants evaluable for this outcome measure.

FACIT-F is a 13-item questionnaire (felt fatigued, felt weak all over, felt listless \[washed out\],felt tired, had energy, had trouble starting things as tired, had trouble finishing things as tired, was able to do usual activities, needed to sleep during day, too tired to eat, needed help doing my usual activities, frustrated by being too tired to do things wanted to do, had to limit my social activity because tired), with each item scored on a 5-point scale ranging from 0 (not at all) to 4 (very much). Three type of scores were derived: change in FACIT-F total score, change in FACIT-F experience domain score, and change in FACIT-F impact domain score. FACIT-F total score was calculated by summing the 13 items (range 0 \[not at all\] to 52 \[very much\]); higher scores represent less fatigue status. In this outcome measure, change from baseline in FACIT-F total score was reported.

Time frame: Baseline, Weeks 2, 4, 8, 12, 16, 24, 32, 40 and 48

Population: FAS: included all participants who were randomized to the study and received at least one dose of the randomized investigational product (i.e., tofacitinib or placebo). Here, on-drug data was used and missing response was not imputed. Here Overall number of participants analyzed signifies participants evaluable for this outcome measure.

Blood samples were collected for analysis of hsCRP using an assay analyzed by central laboratory. hsCRP is an acute phase reactant, which was indicative of inflammation and of its severity.

Time frame: Baseline, Weeks 2, 4, 8, 12, 16, 24, 32, 40 and 48

Population: FAS: included all participants who were randomized to the study and received at least one dose of the randomized investigational product (i.e., tofacitinib or placebo). Here, on-drug data was used and missing response was not imputed. Here Overall number of participants analyzed signifies participants evaluable for this outcome measure.

BASFI was a functional index which included 10 items assessing ability of participants to perform normal daily activities. The first 8 questions/items consider activities related to functional anatomy. The final 2 questions/items assess the participants' ability to cope with everyday life. Each item was scored on a scale of 0=easy to 10=impossible. The BASFI total score was calculated as the average score of these 10 individual items. BASFI total score ranged from 0 (easy) to 10 (impossible), where higher scores indicated more severe disease activity.

Time frame: Baseline, Weeks 2, 4, 8, 12, 16, 24, 32, 40 and 48

Population: FAS: included all participants who were randomized to the study and received at least one dose of the randomized investigational product (i.e., tofacitinib or placebo). Here, on-drug data was used and missing response was not imputed. Here Overall number of participants analyzed signifies participants evaluable for this outcome measure.

The MASES is an index used to measure the severity of enthesitis. Enthesitis is the inflammation of enthuses (heels). The MASES assesses 13 sites for enthesitis. Sites assessed included 1st costochondral joint (left \[l\]/right \[r\]), 7th costochondral joint (l/r), posterior superior iliac spine (l/r), posterior anterior iliac spine (l/r), iliac crest (l/r), proximal insertion of Achilles tendon (l/r) and 5th lumbar spinous process. Each site was graded for the presence (1) and absence (0) of tenderness yielding total MASES score ranging from 0 (no tenderness) to 13 (worst possible score) with higher score indicated more severe tenderness.

Time frame: Baseline, Weeks 4, 8, 12, 16, 24, 32, 40 and 48

Population: FAS: included all participants who were randomized to the study and received at least one dose of the randomized investigational product (i.e., tofacitinib or placebo). Analysis included only participants with baseline MASES \> 0. Here, on-drug data was used and missing response was not imputed. Here Overall number of participants analyzed signifies participants evaluable for this outcome measure.

Participants marked their level of nocturnal spinal pain on a NRS ranged from 0 (no pain) to 10 (most severe pain), with higher scores indicated more severe pain.

Time frame: Baseline, Weeks 2, 4, 8, 12, 16, 24, 32, 40 and 48

Population: FAS: included all participants who were randomized to the study and received at least one dose of the randomized investigational product (i.e., tofacitinib or placebo). Here, on-drug data was used and missing response was not imputed. Here Overall number of participants analyzed signifies participants evaluable for this outcome measure.

Participants marked their level of total back pain on a numerical rating scale (NRS) ranged from 0 (no pain) to 10 (most severe pain), with higher scores indicated more severe pain.

Time frame: Baseline, Weeks 2, 4, 8, 12, 16, 24, 32, 40 and 48

Population: FAS: included all participants who were randomized to the study and received at least one dose of the randomized investigational product (i.e., tofacitinib or placebo). Here, on-drug data was used and missing response was not imputed. Here Overall number of participants analyzed signifies participants evaluable for this outcome measure.

Participants answered the question, How active was your spondylitis on average during the last week?. Participant's response was recorded using a numerical rating scale ranged from 0 (Not Active) to 10 (Very Active), with higher scores indicated more severe disease.

Time frame: Baseline, Weeks 2, 4, 8, 12, 16, 24, 32, 40 and 48

Population: FAS: included all participants who were randomized to the study and received at least one dose of the randomized investigational product (i.e., tofacitinib or placebo). Here, on-drug data was used and missing response was not imputed. Here Overall number of participants analyzed signifies participants evaluable for this outcome measure.

SF-36 v.2 (Acute): 36-item generic health status measure. It measured 8 general health domains: physical functioning, role limitations due to physical health, bodily pain, general health perceptions, vitality, social functioning, role limitations due to emotional problems, mental health. These domains were aggregated into 2 summary scores - physical component summary (PCS), mental component summary (MCS). Four domains comprised PCS score (physical functioning, role-physical, bodily pain, general health) and remaining 4 domains comprised MCS score (vitality, social functioning, role-emotional, mental health). Normalized domain scores, PCS, MCS scores are used in analyses. Component and domain scores were scored by using United States 1998 general population norm. Resulting norm-based T-scores for both the SF36 version 2 and SF36 health domain scale and component summary measures had means of 50 and standard deviations of 10. Higher PCS/MCS/domain score represent better health status.

Time frame: Baseline, Weeks 16 and 48

Population: FAS: included all participants who were randomized to the study and received at least one dose of the randomized investigational product (i.e., tofacitinib or placebo). Here, on-drug data was used and missing response was not imputed. Here Overall number of participants analyzed signifies participants evaluable for this outcome measure.

Chest expansion (measured in centimeters (cm), is defined as the difference in the thoracic circumference during full expiration versus full inspiration. This was measured at the 4th intercostal space. The difference between maximal inspiration and expiration of the two attempts was recorded. The better of the two attempts was used to calculate chest expansion which was defined to be greater than or equal to 0 cm with no defined maximum/upper limit. Greater chest circumference corresponds to higher score indicated more spinal mobility/better health status (measured as Chest Expansion in cm).

Time frame: Baseline, Weeks 2, 4, 8, 12, 16, 24, 32, 40 and 48

Population: FAS: included all participants who were randomized to the study and received at least one dose of the randomized investigational product (i.e., tofacitinib or placebo). Here, on-drug data was used and missing response was not imputed. Here Overall number of participants analyzed signifies participants evaluable for this outcome measure.

Swollen joint count was an assessment on 44 joints (sternoclaviculars, acromioclaviculars, shoulders, elbows, wrists, metacarpophalangeals, thumb interphalangeal, proximal interphalangeals, knees, ankles, and metatarsophalangeals). Each joint was assessed for swelling as: Present or Absent. Artificial joints were not assessed

Time frame: Baseline, Weeks 2, 4, 8, 12, 16, 24, 32, 40 and 48

Population: FAS: included all participants who were randomized to the study and received at least one dose of the randomized investigational product (i.e., tofacitinib or placebo). Analysis included only participants with baseline SJC(44) \> 0. Here, on-drug data was used and missing response was not imputed. Here Overall number of participants analyzed signifies participants evaluable for this outcome measure.

The BASDAI was a validated questionnaire that consisted of 6 questions pertaining to the 5 major symptoms of AS: fatigue; spinal pain; peripheral arthritis; enthesitis, intensity of morning stiffness and duration of morning stiffness. Each question was rated using a numerical rating scale from 0 (none) to 10 (very severe), higher score=high disease activity. The BASDAI score was calculated by computing the mean of questions 5 and 6 and adding it to the sum of questions 1 to 4. This score was then divided by 5. The total BASDAI score was ranged from 0= none to 10= very severe, where higher score indicated high disease activity. BASDAI inflammation score was derived by taking mean of the responses of question 5 and 6 and ranged from 0 (none) to 10 (very severe), where higher score indicated more inflammation (morning stiffness).

Time frame: Baseline, Weeks 2, 4, 8, 12, 16, 24, 32, 40 and 48

Population: FAS: included all participants who were randomized to the study and received at least one dose of the randomized investigational product (i.e., tofacitinib or placebo). Here, on-drug data was used and missing response was not imputed. Here Overall number of participants analyzed signifies participants evaluable for this outcome measure.

The BASDAI was a validated questionnaire that consisted of 6 questions pertaining to the 5 major symptoms of ankylosing spondylitis: fatigue; spinal pain; peripheral arthritis; enthesitis, intensity of morning stiffness and duration of morning stiffness. Each question was rated using a numerical rating scale from 0 (none) to 10 (very severe), higher score=high disease activity. The BASDAI score was calculated by computing the mean of questions 5 and 6 and adding it to the sum of questions 1 to 4. This score was then divided by 5. The total BASDAI score was ranged from 0= none to 10= very severe, where higher score indicated high disease activity.

Time frame: Baseline, Weeks 2, 4, 8, 12, 16, 24, 32, 40 and 48

Population: FAS: included all participants who were randomized to the study and received at least one dose of the randomized investigational product (i.e., tofacitinib or placebo). Here, on-drug data was used and missing response was not imputed. Here Overall number of participants analyzed signifies participants evaluable for this outcome measure.

The WPAI assesses work productivity and impairment. It is a 6-item questionnaire used to assess the degree to which Ankylosing Spondylitis affected work productivity and regular activities over the past 7 days. The questions are as follows: Q1 = currently employed; Q2 = hours missed due to health problems; Q3 = hours missed due to other reasons; Q4 = hours actually worked; Q5 = degree health affected productivity while working (0-10 scale, with higher numbers indicating less productivity); Q6 = degree health affected regular activities (0-10 scale, with higher numbers indicating greater impairment of regular activities). Percent activity impairment due to health problem was a subscale and calculated as: Q6/10 for all respondents. Subscale score was expressed as an impairment percentage (range: 0-100%) where higher numbers indicate greater impairment.

Time frame: Baseline, Weeks 16 and 48

Population: FAS: included all participants who were randomized to the study and received at least one dose of the randomized investigational product (i.e., tofacitinib or placebo). Here, on-drug data was used and missing response was not imputed. Here Overall number of participants analyzed signifies participants evaluable for this outcome measure.

The WPAI assesses work productivity and impairment. It is a 6-item questionnaire used to assess the degree to which Ankylosing Spondylitis affected work productivity and regular activities over the past 7 days. The questions are as follows: Q1 = currently employed; Q2 = hours missed due to health problems; Q3 = hours missed due to other reasons; Q4 = hours actually worked; Q5 = degree health affected productivity while working (0-10 scale, with higher numbers indicating less productivity); Q6 = degree health affected regular activities (0-10 scale, with higher numbers indicating greater impairment of regular activities). Percent Impairment while Working due to Health Problem was a subscale and calculated as: Q5/10 for those who were currently employed and actually worked in the past 7 days. Subscale score was expressed as an impairment percentage (range: 0-100%) where higher numbers indicate greater impairment and less productivity.

Time frame: Baseline, Weeks 16 and 48

Population: FAS: included all participants who were randomized to the study and received at least one dose of the randomized investigational product (i.e., tofacitinib or placebo). Here, on-drug data was used and missing response was not imputed. Here Overall number of participants analyzed signifies participants evaluable for this outcome measure. Here, number analyzed signifies participants evaluable for this outcome measure for specified time point.

The WPAI assesses work productivity and impairment. It is a 6-item questionnaire used to assess the degree to which Ankylosing Spondylitis affected work productivity and regular activities over the past 7 days. The questions are as follows: Q1 = currently employed; Q2 = hours missed due to health problems; Q3 = hours missed due to other reasons; Q4 = hours actually worked; Q5 = degree health affected productivity while working (0-10 scale, with higher numbers indicating less productivity); Q6 = degree health affected regular activities (0-10 scale, with higher numbers indicating greater impairment of regular activities). Percent overall work impairment due to health problem was a subscale and calculated as: Q2/(Q2+Q4)+\[(1- Q2/(Q2+Q4))×(Q5/10)\] for those who were currently employed. Subscale score was expressed as an impairment percentage (range: 0-100%) where higher numbers indicate greater impairment.

Time frame: Baseline, Weeks 16 and 48

Population: FAS: included all participants who were randomized to the study and received at least one dose of the randomized investigational product (i.e., tofacitinib or placebo). Here, on-drug data was used and missing response was not imputed. Here Overall number of participants analyzed signifies participants evaluable for this outcome measure. Here, number analyzed signifies participants evaluable for this outcome measure for specified time point.

The WPAI assesses work productivity and impairment. It is a 6-item questionnaire used to assess the degree to which Ankylosing Spondylitis affected work productivity and regular activities over the past 7 days. The questions are as follows: Q1 = currently employed; Q2 = hours missed due to health problems; Q3 = hours missed due to other reasons; Q4 = hours actually worked; Q5 = degree health affected productivity while working (0-10 scale, with higher numbers indicating less productivity); Q6 = degree health affected regular activities (0-10 scale, with higher numbers indicating greater impairment of regular activities). Percent work time missed due to health problem was a subscale and calculated as: Q2/(Q2+Q4) for those who were currently employed. Subscale score was expressed as an impairment percentage (range: 0-100%) where higher numbers indicate greater impairment and less productivity.

Time frame: Baseline, Weeks 16 and 48

Population: FAS: included all participants who were randomized to the study and received at least one dose of the randomized investigational product (i.e., tofacitinib or placebo). Here, on-drug data was used and missing response was not imputed. Here Overall number of participants analyzed signifies participants evaluable for this outcome measure. Here, number analyzed signifies participants evaluable for this outcome measure for specified time point.

Complete physical examination: included general appearance, skin (presence of rash), heent (head, eyes, ears, nose and throat), lungs (auscultation), heart (auscultation for presence of murmurs, gallops, rubs), lower extremities (presence of peripheral edema), abdominal (palpation and auscultation), neurologic (mental status, station, gait, reflexes, motor and sensory function, coordination) and lymph nodes. Abnormalities in physical examination was based on investigator's discretion/clinical judgement.

Time frame: Screening, Week 16, and Week 48

Population: Safety analysis set: included all participants who were randomized and received at least one dose of the investigational product (i.e., tofacitinib or placebo). Here, number analyzed signifies participants evaluable for this outcome measure for specified rows.

Twelve-lead electrocardiograms (ECGs) were obtained for all participants. Criteria for ECG abnormality: PR interval \>=300 and a percent change from baseline of \>=25 or 50%; QRS duration \>=140 and a percent change from baseline of \>=50%; QT interval \>=500; QTCB, QTCF interval \<480 or \>=450, \<500 or \>=480, \>=500, change from baseline of \<60 and \>=30, and change from baseline of \>=60.

Time frame: Baseline up to Week 16, Baseline up to Week 48

Population: Safety analysis set: included all participants who were randomized and received at least one dose of the investigational product (i.e., tofacitinib or placebo). Here Overall number of participants analyzed signifies participants evaluable for this outcome measure and number analyzed signifies participants evaluable for this outcome measure for specified rows.

Hematology (Hemoglobin, Hematocrit, Erythrocyte, Lymphocyte/Leukocyte, Neutrophil/Leukocyte \<0.8\*Lower limit of normal (LLN), Reticulocyte \>1.5\*Upper limit of normal (ULN), Erythrocyte Mean Corpuscular Volume, Erythrocyte Mean Corpuscular Hemoglobin, Erythrocyte Mean Corpuscular HGB Concentration \<0.9\*LLN, \>1.1\*ULN, Reticulocyte/Erythrocyte, Leukocyte \>1.5\*ULN, Lymphocyte, Neutrophil \<0.8\*LLN and \>1.2\*ULN, Basophil, Basophil/Leukocyte, Eosinophil, Eosinophil/Leukocyte, Monocyte, Monocyte/Leukocyte \>1.2\*ULN); Clinical Chemistry (Bilirubin, Glucose \>1.5\*ULN, AST, ALT, Gamma Glutamyl Transferase \>3.0\*ULN, Urea, Creatinine, Triglyceride, Cholesterol \>1.3\*ULN, LDL Cholesterol\>1.2\*ULN, Potassium, C Reactive Protein \>1.1\*ULN, Bicarbonate \<0.9\*LLN, Creatine Kinase \>2.0\*ULN, HDL Cholesterol \<0.8\*LLN), Urinalysis (Specific Gravity \>1.035, pH \>8, Glucose, Ketones, Protein, Hemoglobin \>=1, Erythrocyte, Leukocyte \>=20, Granular Cast, Hyaline Cast\>1.

Time frame: Baseline up to Week 16 and Baseline up to Week 48

Population: Safety analysis set: included all participants who were randomized and received at least one dose of the investigational product (i.e., tofacitinib or placebo). Here Overall number of participants analyzed signifies participants evaluable for this outcome measure.

An AE: any untoward medical occurrence in a participant who received study drug without regard to possibility of causal relationship. As per National Cancer Institute - Common Terminology Criteria for Adverse Events (NCI-CTCAE) version 4.03, severity was graded as Grade 1: asymptomatic/mild symptoms, clinical or diagnostic observations only, intervention not indicated; Grade 2: moderate, minimal, local or noninvasive intervention indicated, limiting age-appropriate instrumental activities of daily living (ADL); Grade 3: severe or medically significant but not immediately life-threatening, hospitalization or prolongation of existing hospitalization indicated, disabling, limiting self-care ADL; Grade 4: life-threatening consequence, urgent intervention indicated; Grade 5: death related to AE. Treatment-emergent were events between first dose of study drug and up to 48 weeks that were absent before treatment or that worsened relative to pretreatment state.

Time frame: Baseline up to Week 16 and Baseline up to Week 48

Population: Safety analysis set: included all participants who were randomized and received at least one dose of the investigational product (i.e., tofacitinib or placebo).

AE: any untoward medical occurrence in subject who receive study drug without regard to possibility of causal relationship. Treatment-emergent AEs were events that occurred between first dose of study drug and up to 48 weeks that were absent before treatment or that worsened relative to pretreatment state. The severity grades (mild, moderate and severe) were defined as - mild: did not interfere with participant's usual function, moderate: Interfered to some extent with participant's usual function and severe: Interfered significantly with participant's usual function.

Time frame: Baseline up to Week 16 and Baseline up to Week 48

Population: Safety analysis set: included all participants who were randomized and received at least one dose of the investigational product (i.e., tofacitinib or placebo).

Criteria for abnormalities in vital signs: Pulse rate \<40 beats per minute (bpm) to \>120 bpm, Sitting Diastolic blood pressure \< 50 millimeter of mercury (mmHg), increase and decrease in change from baseline of \>= 20mmHg, sitting systolic blood pressure \< 90 mmHg, increase and decrease in change from baseline of \>= 30mmHg.

Time frame: Baseline up to Week 16 and Baseline up to Week 48

Population: Safety analysis set: included all participants who were randomized and received at least one dose of the investigational product (i.e., tofacitinib or placebo). Here Overall number of participants analyzed signifies participants evaluable for this outcome measure.

ASAS40 assessed 4 domains: the PGA (assess disease activity on a scale of 0 \[not active\] to 10 \[very active\], higher score=more disease activity), total back pain (on a scale of 0 \[no pain\] to 10 \[most severe pain\], higher score=more severity), Function (from BASFI: assess participant's level of ability on a scale of 0 \[easy\] to 10 \[impossible\], lower scores= better functional health) and Inflammation (morning stiffness, Mean of Q5 and Q6 of BASDAI defined as 6 item questionnaire: measures disease activity on a scale of 0 \[none\] to 10 \[severe\], higher score=more disease activity). ASAS40 response: \>=40% and \>=2 units improvement in \>=3 domains and no worsening at all in the remaining domain.

Time frame: Week 16

Population: FAS: included all participants who were randomized to the study and received at least one dose of the randomized investigational product (i.e., tofacitinib or placebo). Here, on-drug data was used and missing response (MR) was considered to be Non-response (NR) (MR=NR).

ASAS20 assess 4 domains: PGA of Disease (assess disease activity on a scale of 0 \[not active\] to 10 \[very active\], high score=more disease activity), total back pain (scale of 0 \[no pain\] to 10 \[most severe pain\], high score=more severity), Function (BASFI; participant's level of ability on scale of 0 \[easy\] to 10 \[impossible\], low score= better functional health) and Inflammation (morning stiffness, Mean of Q5 and Q6 of BASDAI defined as 6-item questionnaire measure disease activity on a scale of 0 \[none\] to 10 \[severe\], high score=more disease activity). ASAS20 response: \>= 20% improvement from baseline in disease activity and absolute change of \>=1 unit in \>=3 domains and no worsening of \>=20% and an absolute change of \>=1 unit in remaining domain.

Time frame: Weeks 2, 4, 8, 12, 24, 32, 40 and 48

Population: FAS: included all participants who were randomized to the study and received at least one dose of the randomized investigational product (i.e., tofacitinib or placebo). Here, on-drug data was used and missing response (MR) was considered to be Non-response (NR) (MR=NR).

ASAS40 assessed 4 domains: the PGA (assess disease activity on a scale of 0 \[not active\] to 10 \[very active\], higher score=more disease activity), total back pain (on a scale of 0 \[no pain\] to 10 \[most severe pain\], higher score=more severity), Function (from BASFI: assess participant's level of ability on a scale of 0 \[easy\] to 10 \[impossible\], lower scores= better functional health) and Inflammation (morning stiffness, Mean of Q5 and Q6 of BASDAI defined as 6 item questionnaire: measures disease activity on a scale of 0 \[none\] to 10 \[severe\], higher score=more disease activity). ASAS40 response: \>=40% and \>=2 units improvement in \>=3 domains and no worsening at all in the remaining domain.

Time frame: Weeks 2, 4, 8, 12, 24, 32, 40 and 48

Population: FAS: included all participants who were randomized to the study and received at least one dose of the randomized investigational product (i.e., tofacitinib or placebo). Here, on-drug data was used and missing response (MR) was considered to be Non-response (NR) (MR=NR).

ASAS 5/6 consists of 6 domains: 4 used in ASAS20 - PGA (assess disease activity on a scale of 0 \[not active\] to 10 \[very active\], higher score=more disease activity), Spinal Pain (total back pain) (on a scale of 0 \[no pain\] to 10 \[most severe pain\], higher score=more severity), Function (using BASFI which assess participant's level of ability on a scale of 0 \[easy\] to 10 \[impossible\], lower scores= better functional health) and Inflammation (using BASDAI, mean of Q 5 and 6, which assess disease activity on a scale of 0 \[none\] to 10 \[severe\], higher score=more disease activity), CRP (was measured in mg per liter) and Spinal mobility was measured in centimeter and calculated as mean of right and left measurements of lateral spinal flexion from BASMI. ASAS 5/6: defined as \>=20% improvement in at least 5 domains.

Time frame: Weeks 2, 4, 8, 12, 16, 24, 32, 40 and 48

Population: FAS: included all participants who were randomized to the study and received at least one dose of the randomized investigational product (i.e., tofacitinib or placebo). Here, on-drug data was used and missing response (MR) was considered to be Non-response (NR) (MR=NR).

Partial remission was defined as a score of 2 or less (on a scale of 0-10, where 0=no disease activity and 10=high disease activity) in each of the 4 domains in ASAS. These 4 domains included: PGA (assess disease activity on a scale of 0 \[not active\] to 10 \[very active\], higher score=more disease activity), total back pain (on a scale of 0 \[no pain\] to 10 \[most severe pain\], higher score=more severity), Function (using BASFI which assess participant's level of ability on a scale of 0 \[easy\] to 10 \[impossible\], lower scores= better functional health) and Inflammation (using BASDAI, mean of Q 5 and 6, which assess disease activity on a scale of 0 \[none\] to 10 \[severe\], higher score=more disease activity).

Time frame: Weeks 2, 4, 8, 12, 16, 24, 32, 40 and 48

Population: FAS: included all participants who were randomized to the study and received at least one dose of the randomized investigational product (i.e., tofacitinib or placebo). Here, on-drug data was used and missing response (MR) was considered to be Non-response (NR) (MR=NR).

BASDAI was a validated questionnaire that consisted of 6 questions pertaining to the 5 major symptoms of AS: fatigue; spinal pain; peripheral arthritis; enthesitis, intensity of morning stiffness and duration of morning stiffness. Each question was rated using a numerical rating scale from 0 (none) to 10 (very severe), higher score=high disease activity. BASDAI score was calculated by computing mean of questions 5 and 6 and adding it to sum of questions 1 to 4. This score was then divided by 5. The total BASDAI score was ranged from 0= none to 10= very severe, where higher score indicated high disease activity. BASDAI50 response was defined as decrease of \>=50% from Baseline in BASDAI score at specified time points. Percentage of participants with BASDAI 50 response at specified weeks are reported.

Time frame: Weeks 2, 4, 8, 12, 16, 24, 32, 40 and 48

Population: FAS: included all participants who were randomized to the study and received at least one dose of the randomized investigational product (i.e., tofacitinib or placebo). Here, on-drug data was used and missing response (MR) was considered to be Non-response (NR) (MR=NR).

ASDAS(CRP) was derived using BASDAI (6-item questionnaire measures disease activity on a scale of 0 \[none\] to 10 \[severe\], higher score=more disease activity) and PGA (measure disease activity on a scale of 0 \[not active\] to 10 \[very active\], high score=more disease activity) and by using the following formula, 0.121 x Back Pain (Q2 of BASDAI) + 0.058 x Duration of Morning Stiffness (Q6 of BASDAI) + 0.110 x PGA + 0.073 x Peripheral Pain/Swelling (Q3 of BASDAI) + 0.579 x Ln(hsCRP mg/L + 1). If hsCRP values were smaller than 2 mg/L, they were set to 2 mg/L in the formula. The range of score was \>= 0.636 to no defined upper limit. A negative change from baseline value indicates decrease in disease activity; a positive change from baseline value indicates increase in disease activity. ASDAS(CRP) inactive disease is defined as a response if actual ASDAS(CRP) was \<1.3 units.

Time frame: Weeks 2, 4, 8, 12, 16, 24, 32, 40 and 48

Population: FAS: included all participants who were randomized to the study and received at least one dose of the randomized investigational product (i.e., tofacitinib or placebo). Analysis included only participants with baseline ASDAS(CRP) \>= 1.3 units. Here, on-drug data was used and missing response (MR) was considered to be Non-response (NR) (MR=NR).

ASDAS(CRP) was derived using BASDAI (6-item questionnaire measures disease activity on a scale of 0 \[none\] to 10 \[severe\], higher score=more disease activity) and PGA (measure disease activity on a scale of 0 \[not active\] to 10 \[very active\], high score=more disease activity) and by using the following formula, 0.121 x Back Pain (Q2 of BASDAI) + 0.058 x Duration of Morning Stiffness (Q6 of BASDAI) + 0.110 x PGA + 0.073 x Peripheral Pain/Swelling (Q3 of BASDAI) + 0.579 x Ln(hsCRP mg/L + 1). If hsCRP values were smaller than 2 mg/L, they were set to 2 mg/L in the formula. The range of score was \>= 0.636 to no defined upper limit. A negative change from baseline value indicates decrease in disease activity; a positive change from baseline value indicates increase in disease activity. ASDAS(CRP) major improvement was defined as a response if improvement (decrease) from Baseline in ASDAS(CRP) of \>=2.0 units.

Time frame: Weeks 2, 4, 8, 12, 16, 24, 32, 40 and 48

Population: FAS: included all participants who were randomized to the study and received at least one dose of the randomized investigational product (i.e., tofacitinib or placebo). Analysis included only participants with baseline ASDAS(CRP) \>= 2.636 units. Here, on-drug data was used and missing response (MR) was considered to be Non-response (NR) (MR=NR). Here Overall number of participants analyzed signifies participants evaluable for this outcome measure.

ASDAS(CRP) was derived using BASDAI (6-item questionnaire measures disease activity on a scale of 0 \[none\] to 10 \[severe\], higher score=more disease activity) and PGA (measure disease activity on a scale of 0 \[not active\] to 10 \[very active\], high score=more disease activity) and by using the following formula, 0.121 x Back Pain (Q2 of BASDAI) + 0.058 x Duration of Morning Stiffness (Q6 of BASDAI) + 0.110 x PGA + 0.073 x Peripheral Pain/Swelling (Q3 of BASDAI) + 0.579 x Ln(hsCRP mg/L + 1). If hsCRP values were smaller than 2 mg/L, they were set to 2 mg/L in the formula. The range of score was \>= 0.636 to no defined upper limit. A negative change from baseline value indicates decrease in disease activity; a positive change from baseline value indicates increase in disease activity. ASDAS(CRP) clinically important improvement was defined as decrease from Baseline of \>=1.1 units in ASDAS(CRP) score.

Time frame: Weeks 2, 4, 8, 12, 16, 24, 32, 40 and 48

Population: FAS: included all participants who were randomized to the study and received at least one dose of the randomized investigational product (i.e., tofacitinib or placebo). Analysis included only participants with baseline ASDAS(CRP) \>= 1.736 units. Here, on-drug data was used and missing response (MR) was considered to be Non-response (NR) (MR=NR). Here Overall number of participants analyzed signifies participants evaluable for this outcome measure.