BASDAI subject asessment of discomfort, pain and fatigue measured using a 100 millimeter Visual Analog Scale; range: 0=none to 100=very severe. Normalized net incremental area under the curve (AUC) = area between baseline and the BASDAI curve as a function of time (randomization to Week 12); computed using the linear trapezoidal method. All areas above baseline and under the curve are positive and all areas below baseline and above the curve are negative. Net incremental AUC = sum of these areas; this result was then divided by the study duration of the patients; negative value = improvement.

Modified Intent-To-Treat (mITT) population: includes all randomized subjects who received at least 1 dose of blinded study drug. Last observation carried forward (LOCF).

BASDAI subject assessment of discomfort, pain and fatigue was measured using a 100 millimeter Visual Analog Scale (VAS); range: 0=none to 100=very severe. BASDAI 50 response defined as at least a 50 percent (%) improvement (decrease) from baseline to observation (last observation carried forward) in the BASDAI. Baseline score minus score at observation divided by Baseline score * 100 = >=50%.

mITT; LOCF.

ASAS measures symptomatic improvement in Ankylosing Spondylitis (AS) patients; 4 domains: patient global assessment of disease activity, pain, function,and inflammation. ASAS 20 = 20% improvement (versus baseline) and an absolute change (net improvement) ≥ 10 millimeters (mm) on a 0-100 mm scale (100=high disease activity) for ≥ 3 domains, and no worsening (absence of deterioration by >=20% and by >= 10 mm) in remaining domain.

mITT; LOCF.

ASAS measures symptomatic improvement in Ankylosing Spondylitis (AS) patients; 4 domains: patient global assessment of disease activity, pain, function, and inflammation. ASAS 20 = 20% improvement (versus baseline) and an absolute change (net improvement) ≥ 10 units (millimeters) on a 0-100 scale (100=high disease activity) for ≥ 3 domains, and no worsening (absence of deterioration by >=20% and by >=10 mm) in remaining domain.

Open-label population: all subjects who received at least 1 dose of open test article (Etanercept). LOCF.

ASAS measures symptomatic improvement in Ankylosing Spondylitis (AS) patients; 4 domains: patient global assessment of disease activity, pain, function, inflammation. ASAS 50 = 50% improvement (versus baseline) and an absolute (net) change ≥ 20 units on a 0-100 scale (high disease activity) for ≥ 3 domains, and no worsening (absence of deterioration by >=20% and by >=10 mm) in remaining domain.

mITT; LOCF.

ASAS measures symptomatic improvement in Ankylosing Spondylitis (AS) patients; 4 domains: patient global assessment of disease activity, pain, function, inflammation. ASAS 50 = 50% improvement (vs. baseline) and an absolute (net) change ≥ 20 units on a 0-100 scale (high disease activity) for ≥ 3 domains, and no worsening (absence of deterioration by >=20% and by >=10 mm) in remaining domain.

Open-label population; LOCF.

ASAS measures symptomatic improvement in Ankylosing Spondylitis (AS) patients; 4 domains: patient global assessment of disease activity, pain, function, inflammation. ASAS 70 = 70% improvement (vs. baseline) and an absolute change ≥ 20 units on a 0-100 scale (high disease activity) for ≥ 3 domains, and no worsening (absence of deterioration by >=20% and by >= 10 mm) in remaining domain.

mITT; LOCF.

ASAS measures symptomatic improvement in Ankylosing Spondylitis (AS) patients; 4 domains: patient global assessment of disease activity, pain, function, inflammation. ASAS 70 = 70% improvement (vs. baseline) and an absolute change ≥ 20 units on a 0-100 scale (high disease activity) for ≥ 3 domains, and no worsening (absence of deterioration by >=20% and by >= 10 mm) in remaining domain.

Open-label population; LOCF.

Partial remission defined as a score of <20 millimeters (mm) on a scale of 0 to 100 mm in each of 4 domains: Visual Analog Scale (VAS) patient global assessment, VAS pain score (Total Back Pain), Bath Ankylosing Spondylitis Functional Index (BASFI) score, and Bath Ankylosing Spondylitis Disease Activities Index (BASDAI)-mean of two morning stiffness-related VAS scores. A negative score indicates an improvement in disease activity and a positive score indicates worsening.

mITT; LOCF.

Partial remission defined as a score of <20 millimeters (mm) on a scale of 0 to 100 mm in each of 4 domains: Visual Analog Scale (VAS) patient global assessment, VAS pain score (Total Back Pain), Bath Ankylosing Spondylitis Functional Index (BASFI) score, and Bath Ankylosing Spondylitis Disease Activities Index (BASDAI)-mean of two morning stiffness-related VAS scores. A negative score indicates an improvement in disease activity and a positive score indicates worsening.

Open-label population; LOCF.

PGA was measured using a 100 millimeter Visual Analog Scale (VAS) ranging from 0 = very good to 100 = very bad. Normalized net incremental area under the curve (AUC)=area between baseline and the Patient Global Asessment curve as a function of time (randomization to Week 12); computed using the linear trapezoidal method. All areas above baseline and under the curve are positive and all area below baseline and above the curve are negative. Net incremental AUC = sum of these areas; this result was then divided by the study duration of the patients; negative value = improvement.

mITT; LOCF. N = number of subjects with evaluable data.

Patient global assessment of all the ways ankylosing spondylitis affected them in the last 48 hours using a 100 millimeter (mm) Visual Analog Scale (VAS); range 0=very good to 100=very bad.

mITT; LOCF.

Patient global assessment of all the ways ankylosing spondylitis affected them in the last 48 hours using a 100 millimeter (mm) Visual Analog Scale (VAS); range 0=very good to 100=very bad.

Open-label population; LOCF.

PGA was measured using a 100 millimeter Visual Analog Scale (VAS) ranging from 0 = very good to 100 = very bad. Normalized net incremental area under the curve (AUC) = area between baseline and the Physician Global Assessment curve as a function of time (randomization to Week 12); computed using the linear trapezoidal method. All areas above baseline and under the curve are positive and all area below baseline and above the curve are negative. Net incremental AUC = sum of these areas; this result was then divided by the study duration of the patients; negative value = improvement.

mITT; LOCF. N= number of subjects with evaluable data.

Physician global assessment of all the ways ankylosing spondylitis has affected patient during the last 48 hours. 100 millimeter (mm) Visual Analog Scale (VAS); range: 0=very good to 100=very bad.

mITT; LOCF. N = number of subjects with evaluable data.

Physician global assessment of all the ways ankylosing spondylitis has affected patient during the last 48 hours. 100 millimeter (mm) Visual Analog Scale (VAS); range: 0=very good to 100=very bad.

Open-label population; LOCF.

Nocturnal back pain was measured using a 100 millimeter Visual Analog Scale (VAS); range: 0 = none to 100 = extreme. Normalized net incremental area under the curve (AUC) = area between baseline and the Nocturnal Back Pain curve as a function of time (randomization to Week 12); computed using the linear trapezoidal method. All areas above baseline and under the curve are positive and all area below baseline and above the curve are negative. Net incremental AUC = sum of these areas; this result was then divided by the study duration of the patients; negative value = improvement.

mITT; LOCF. N=number of subjects with evaluable data.

Subject assessment of nocturnal back pain due to ankylosing spondylitis during the last 48 hours using a 100 millimeter (mm) Visual Analog Scale (VAS); range: 0=none to 100=extreme.

mITT; LOCF.

Subject assessment of nocturnal back pain due to ankylosing spondylitis during the last 48 hours using a 100 millimeter (mm) Visual Analog Scale (VAS); range: 0=none to 100=extreme.

Open-label population; LOCF.

Total back pain was measured using a 100 millimeter Visual Analog Scale (VAS); range: 0 = none to 100 = extreme. Normalized net incremental area under the curve (AUC) = area between baseline and the Total Back Pain curve as a function of time (randomization to Week 12); computed using the linear trapezoidal method. All areas above baseline and under the curve are positive and all area below baseline and above the curve are negative. Net incremental AUC = sum of these areas; this result was then divided by the study duration of the patients; negative value = improvement.

mITT; LOCF. N=number of subjects with evaluable data.

Subject assessment of total back pain due to ankylosing spondylitis during the last 48 hours using a 100 millimeter (mm) Visual Analog Scale (VAS); range: 0=none to 100=extreme.

mITT; LOCF.

Subject assessment of total back pain due to ankylosing spondylitis during the last 48 hours using a 100 millimeter Visual Analog Scale (VAS); range: 0=none to 100=extreme.

Open-label population; LOCF.

BASFI was measured using a 100 millimeter Visual Analog Scale (VAS) ranging from 0 = easy to 100 = impossible. Normalized net incremental area under the curve (AUC) = area between baseline and the BASFI curve as a function of time (randomization to Week 12); computed using the linear trapezoidal method. All areas above baseline and under the curve are positive and all area below baseline and above the curve are negative. Net incremental AUC = sum of these areas; this result was then divided by the study duration of the patients; negative value = improvement.

mITT; LOCF. N=number of subjects with evaluable data.

BASFI is a validated self assessment tool that determines the degree of functional limitation in ankylosing spondylitis (AS) patients using a 100 millimeter (mm) Visual Analog Scale (VAS) measuring level of ability with activities in the last 48 hours; range: 0=easy to 100=impossible. Higher score = greater limitation.

mITT; LOCF. N = number of subjects with evaluable data.

Subject rating of last 48 hours, 100 millimeter Visual Analog Scale; range 0=easy to 100=impossible: 1)Putting on socks/tights without (w/o) help; 2)Bending forward from waist to pickup pen from floor w/o aid; 3)Reaching to high shelf w/o aid; 4)Getting out of armless dining room chair w/o using hands/other help; 5)Getting up off floor w/o help from lying on back; 6)Standing unsupported 10 minutes w/o discomfort; 7)Climbing 12-15 steps w/o handrail or walking aid; 8)Looking over shoulder w/o turning body; 9) Doing physically demanding activities; 10) Doing full days activities (home or work).

mITT; in the case of missing data, no replacement or imputation method was performed. Abbreviations: C = component (number), Act = Activities.

Subject rating of last 48 hours, 100 mm Visual Analog Scale; range 0=easy to 100=impossible: 1) Putting on socks/tights without (w/o) help; 2) Bending forward from waist to pickup pen from floor w/o aid; 3) Reaching to high shelf w/o aid; 4) Getting out of armless dining room chair w/o using hands/other help; 5) Getting up off floor w/o help from lying on back; 6) Standing unsupported 10 minutes w/o discomfort; 7) Climbing 12-15 steps w/o handrail or walking aid; 8) Looking over shoulder w/o turning body; 9) Doing physically demanding activities; 10) Doing full days activities (home or work).

Open-label population; in the case of missing data, no replacement or imputation method was performed. Abbreviations: C=component (number), Act=Activities.

BASDAI is a validated self assessment tool used to determine disease activity in patients with ankylosing spondylitis in the last 48 hours. Utilizing a Visual Analog Scale (VAS) of 0-10 (0=none and 10=very severe) patient's answered 6 questions measuring discomfort, pain and fatigue. The BASDAI final mean score was calculated taking all 6 VAS assessments.

mITT; LOCF.

BASDAI is a validated self assessment tool used to determine disease activity in patients with ankylosing spondylitis (AS) in the last 48 hours. Utilizing a Visual Analog Scale (VAS) of 0-10 (0=none and 10=very severe) patient's answered 6 questions measuring discomfort, pain and fatigue. The BASDAI final mean score was calculated taking all 6 VAS assessments.

Open-label population; LOCF.

BASDAI subject rated components over last 48 hours using 100 mm Visual Analog Scale; components 1-5: range 0=none to 100=very severe; component 6: range:0=0 hours to 100=2 hours or more. 1) Overall level of fatigue/tiredness experienced; 2) Overall level of AS neck,back or hip pain experienced; 3) Overall level of pain/swelling in joints other than neck, back or hips; 4) Overall level of discomfort from any areas tender to touch or pressure; 5) Overall level of morning stiffness from time of awakening; 6) Duration of morning stiffness from time of awakening, and morning stiffness subscale.

mITT; in case of mising data, no replacement or imputation method was performed. Abbreviations: C=component, AS=Ankylosing Spondylitis.

BASDAI subject rated components over last 48 hours using a 100 millimeter Visual Analog Scale; components 1-5: range 0=none to 100=very severe; component 6: range:0=0 hours to 100=2 hours or more. 1) Overall level of fatigue/tiredness experienced; 2) Overall level of AS neck,back or hip pain experienced; 3)Overall level of pain/swelling in joints other than neck,back or hips; 4) Overall level of discomfort from any areas tender to touch or pressure; 5) Overall level of morning stiffness from time of awakening; 6) Duration of morning stiffness from time of awakening.

Open-label population; in case of missing data, no replacement or imputation method was performed. Abbreviations: C=component, AS=Ankylosing Spondylitis.

BAS-G was measured using a 100 millimeter Visual Analog Scale (VAS) ranging from 0=none to 100=very important. Normalized net incremental area under the curve (AUC) = area between baseline and the BAS-G curve as a function of time (randomization to Week 12); computed using the linear trapezoidal method. All areas above baseline and under the curve are positive and all area below baseline and above the curve are negative. Net incremental AUC = sum of these areas; this result was then divided by the study duration of the patients; negative value = improvement.

mITT; LOCF. N=number of subjects with evaluable data.

Subject assessment of the effect of their disease on well-being over last 48 hours using a 100 millimeter (mm) Visual Analog Scale (VAS); range: 0=none to 100=very important.

mITT; LOCF. N=number of subjects with evaluable data.

Subject assessment of the effect of their disease on well-being over last 48 hours using a 100 millimeter Visual Analog Scale (VAS); range: 0=none to 100=very important.

Open-label population; LOCF.

Subject evaluation of the effect of their disease on well-being over the last week using a using a 100 millimeter Visual Anaog Scale; range: 0=none to 100=very important.

mITT; LOCF.

mITT. In the case of missing data, no replacement or imputation method was performed.

mITT; N=number of subjects with evaluable data. In the case of missing data, no replacement or imputation method was performed.

BASMI was composed of 5 measures; each measure scored 0-2 (0=normal mobility, 2=severe reduction); final score range: 0 to 10. Normalized net incremental area under the curve (AUC) = area between baseline and the BASMI curve as a function of time (randomization to Week 12); computed using the linear trapezoidal method. All areas above baseline and under the curve are positive and all area below baseline and above the curve are negative. Net incremental AUC = sum of these areas; this result was then divided by the study duration of the patients; negative value = improvement.

mITT; LOCF.

BASMI is an objective measure of spinal mobility. The BASMI score is composed of 5 measures: cervical rotation, intermalleolar distance, modified Schober test, lateral flexion and tragus to wall distance. Each measure was scored 0-2 (0=normal mobility, 2=severe reduction) to give a final score ranging 0 to 10. Higher score = greater reduction in spinal mobility.

mITT; LOCF.

BASMI is an objective measure of spinal mobility. The BASMI score is composed of 5 measures: cervical rotation, intermalleolar distance, modified Schober's test, lateral flexion and tragus to wall distance. Each measure was scored 0-2 (0=normal mobility, 2=severe reduction) to give a final score ranging 0 to 10. Higher score = greater reduction in spinal mobility.

Open-label population; LOCF.

Cervical rotation: measurement of degrees to which the subjects could turn their heads as far as possible to the right and then to the left. Mean of ordinal scores from 0: >= 85 degrees to 10: <= 8.5 degrees.

mITT; LOCF. In the case of missing data, no replacement or imputation method was performed.

Cervical rotation: measurement in degree to which the subjects could turn their heads as far as possible to the right and then to the left. Mean of ordinal scores from 0: >= 85 degrees to 10: <= 8.5 degrees.

Open-label population; in the case of missing data, no replacement or imputation method was performed.

Measurement in centimeters (cm) of distance between the tragus and wall from right and left side while subject is standing with back against the wall; knees straight; scapulae, buttocks, and heels against the wall; with head in neutral position. Measurement of two attempts on right and left sides. Mean of ordinal scores from 0: <= 10 cm to 10: >= 37 cm.

mITT. In the case of missing data, no replacement or imputation method was performed.

Measurement in centimeters (cm) of distance between the tragus and wall from right and left side while subject is standing with back against the wall; knees straight; scapulae, buttocks, and heels against the wall; with head in neutral position. Measurement of two tries on right and left sides. Mean of ordinal scores from 0: <= 10 cm to 10: >= 37 cm.

Open-label population; in the case of missing data, no replacement or imputation method was performed.

Measurement in centimters (cm) of distance between subject's middle fingertip and the floor after bending sideways, without bending knees or lifting heels, while attemting to keep shoulders in same place (flexion position). Measurement of two attempts on each side (right and left). Mean of ordinal scores from 0: >=20 cm to 10: <=1.2 cm.

mITT. In the case of missing data, no replacement or imputation method was performed.

Measurement in centimters (cm) of distance between subject's middle fingertip and the floor after bending sideways, without bending knees or lifting heels, while attempting to keep shoulders in same place (flexion position). Measurement of two attempts on each side (right and left). Mean of ordinal scores from 0: >=20 cm to 10: <=1.2 cm.

Open-label population; in the case of missing data, no replacement or imputation method was performed.

Measurement in centimeters of the distance between marks originally placed while the subject was standing erect 10 centimeters (cm) above and 5 cm below the midpoint of a line that joints the posterior superior iliac spines. Distance between marks was remeasured with subject maximally bend forward, knees fully extended, with spine in full flexion. Measurement of two attempts. Mean of ordinal scores from 1: 5.7 to 6.3 cm, to 10: <=0.7 cm.

mITT. In the case of missing data, no replacement or imputation method was performed.

Measurement in centimeters of the distance between marks originally placed while the subject was standing erect 10 centimeters (cm) above and 5 cm below the midpoint of a line that joints the posterior superior iliac spines. Distance between marks was remeasured with subject maximally bend forward, knees fully extended, with spine in full flexion. Measurement of two attempts. Mean of ordinal scores from 1: 5.7 to 6.3 cm, to 10: <=0.7 cm.

Open-label population; in the case of missing data, no replacement or imputation method was performed.

Measurement in centimeters (cm) of the distance between the medial malleoli when subject is lying supine with knees and feet straight up with legs separated as far as possible. Measurement of two attempts. Mean of ordinal scores from 0: >= 120 cm, to 9: 30 to 39.9 cm.

mITT. In the case of missing data, no replacement or imputation method was performed.

Measurement in centimeters (cm) of the distance between the medial malleoli when subject is lying supine with knees and feet straight up with legs separated as far as possible. Measurement of two attempts. Mean of ordinal scores from 0: >= 120 cm, to 9: 30 to 39.9 cm.

Open-label population; in the case of missing data, no replacement or imputation method was performed.

Normalized net incremental area under the curve (AUC) = area between baseline and the Chest Expansion Test curve as a function of time (randomization to Week 12); computed using the linear trapezoidal method. All areas above baseline and under the curve are positive and all areas below baseline and above the curve are negative. Net incremental AUC = sum of these areas; this result was then divided by the study duration of the patients; negative value = improvement.

mITT; LOCF. N=number of subjects with evaluable data.

Measurement and remeasurement of standing maximal inspiration; chest circumference at nipple line or at 4th intercostal space in centimeters.

mITT; LOCF. N=number of subjects with evaluable data.

Measurement and remeasurement of standing maximal inspiration; chest circumference at nipple line or at 4th intercostal space in centimeters (cm).

Open-label population; LOCF.

Subject evaluation of level of difficulty to perform daily physical activities and/or kinesitherapy for ankylosing spondylitis (AS) because of AS using a 100 millimeter Visual Analog Scale (VAS) ranging from 0=easy to 100=impossible. Subgroup of subjects who responded that they were able to perform daily physical activities and/or kinesitherapy for their ankylosing spondylitis at each visit.

mITT; LOCF. N = subgroup of subjects who reported performing daily physical activities and/or kinesitherapy for their ankylosing spondylitis at each visit (Week 2 through Week 12).

Subject evaluation of level of difficulty to perform daily physical activities and/or kinesitherapy for ankylosing spondylitis (AS) due to AS using a 100 millimeter Visual Analog Scale (VAS) ranging from 0=easy to 100=impossible. Subgroup of subjects who responded that they were able to perform daily physical activities and/or kinesitherapy for their ankylosing spondylitis at each visit.

Open-label population; LOCF.

Normalized net incremental area under the curve (AUC) = area between baseline and the Erythrocyte Sedimentation Rate curve as a function of time (randomization to Week 12); computed using the linear trapezoidal method. All areas above baseline and under the curve are positive and all areas below baseline and above the curve are negative. Net incremental AUC = sum of these areas; this result was then divided by the study duration of the patients; negative value = improvement.

mITT; LOCF. N=number of subjects with evaluable data.

ESR is a laboratory test that provides a non-specific measure of inflammation. The test assesses the rate at which red blood cells fall in a test tube. A higher rate is consistent with inflammation.

mITT; LOCF. N = number of subjects with evaluable data.

ESR is a laboratory test that provides a non-specific measure of inflammation. The test assesses the rate at which red blood cells fall in a test tube. A higher rate is consistent with inflammation.

Open-label population; LOCF.

Normalized net incremental area under the curve (AUC) = area between baseline and the C-reactive Protein curve as a function of time (randomization to Week 12); computed using the linear trapezoidal method. All areas above baseline and under the curve are positive and all areas below baseline and above the curve are negative. Net incremental AUC = sum of these areas; this result was then divided by the study duration of the patients; negative value = improvement.

mITT; LOCF.

CRP is a marker of inflammation. A higher level is consistent with inflammation.

mITT. LOCF.

Change from baseline in C-reactive Protein (CRP).

Open-label population; LOCF.

Percent of participants with normal (<6 milligrams per liter) and abnormal (>= 6 milligrams per liter) C-Reactive Protein.

Open-label population; LOCF.

Percent of subjects reporting acceptable symptom state: acceptance to remain for the rest of their lives with the level of pain they had during the last 48 hours; and unacceptable symptom state: not able to remain for the rest of their lives with the level of pain they had during the last 48 hours. In the case of missing data, subjects who withdrew from the study because of inefficacy or toxicity were considered "unacceptable".

mITT. In the case of missing data, no replacement or imputation method was performed.

Percent of subjects reporting acceptable symptom state: acceptance to remain for the rest of their lives with the level of pain they had during the last 48 hours; and unacceptable symptom state: not able to remain for the rest of their lives with the level of pain they had during the last 48 hours. In the case of missing data, subjects who withdrew from the study because of inefficacy or toxicity were considered "unacceptable".

Open-label population. In the case of missing data, no replacement or imputation method was performed.

Subjects were asked how their pain had been during the last 48 hours compared to baseline. Those subjects that reported improvement assessed how important this improvement was to them; range: very important, moderately important, slightly important, or not at all important. Binary response options: 1=improved very important, or improved moderately important; 2=slightly important, not at all important, no change, or worse-more pain.

mITT; in the case of missing data, no replacement or imputation method was performed. Abbreviation: Mod = moderately.

Subjects were asked how their pain had been during the last 48 hours compared to baseline. Those subjects that reported improvement assessed how important this improvement was to them; range: very important, moderately important, slightly important, or not at all important. Binary response options: 1=improved very important, or improved moderately important; 2=slightly important, not at all important, no change, or worse-more pain.

Open-label population; in the case of missing data, no replacement or imputation method was performed. Abbreviation: Mod = moderately.