CDAI was derived as the sum of the following: tender joint count (TJC), swollen joint count (SJC), participant global assessment (PGA) of disease activity, and physician assessment of disease activity. TJC and SJC were taken as the number of tender and swollen joints, respectively, out of 28 assessed joints. PGA and physician assessment of disease activity were scored 0-100 millimeters (mm) and rounded to the nearest centimeter (cm) on a visual analog scale (VAS), where higher scores indicate greater perceived disease activity. The total CDAI score range was 0-76, where higher scores indicate increased disease activity. Change from baseline was averaged among all participants. Negative values indicate improvement/reduction in RA disease activity.

ITT Set; only those patients who provided data at baseline and at least one post-baseline assessment were analyzed.

DAS28-ESR was based on TJC, SJC, PGA of disease activity, and laboratory-derived ESR. TJC and SJC were taken as the number of tender and swollen joints, respectively, out of 28 assessed joints. PGA of disease activity was scored 0-100 mm on a VAS, where higher scores indicate greater perceived disease activity. The total DAS28-ESR score was transformed to a single score range of 0-10, where higher scores indicate increased disease activity. Change from baseline was averaged among all participants. Negative values indicate improvement/reduction in RA disease activity.

ITT Set; only those patients who provided data at baseline and at least one post-baseline assessment were analyzed.

ACR response was assessed on the basis of percent improvement (20% for ACR20, 50% for ACR50, 70% for ACR70) in both TJC and SJC as well as at least three of the following: physician assessment of disease activity, PGA of disease activity, PGA of pain, Health Assessment Questionnaire-Disability Index (HAQ-DI), and either ESR or C-reactive protein level. TJC and SJC were taken as the number of tender and swollen joints, out of 68 and 66 assessed joints, respectively. PGA and physician assessments were scored 0-100 mm on a VAS, where higher scores indicate greater perceived disease activity or pain. HAQ-DI was scored using participant responses to 20 questions assessing activities of daily living (ADLs), with total score scale of 0-3, where higher scores indicate increased functional disability. The percentage of participants meeting criteria for each level of ACR response was reported.

ITT Set

EULAR response was assessed by change from baseline and absolute DAS28-ESR score. EULAR response classification was as follows: Good (change >1.2 with absolute score </=3.2), Moderate (change >1.2 with absolute score >3.2 or change >0.6 with absolute score </=5.1), None (change </=0.6 or absolute score >5.1). DAS28-ESR was based on TJC, SJC, and PGA of disease activity, and laboratory-derived ESR. TJC and SJC were taken as the number of tender and swollen joints, respectively, out of 28 assessed joints. PGA of disease activity was scored 0-100 mm on a VAS, where higher scores indicate greater perceived disease activity. The total DAS28-ESR score was transformed to a single score range of 0-10, where higher scores indicate increased disease activity. The percentage of participants meeting criteria for each level of EULAR response was reported.

ITT Set

CDAI was derived as the sum of the following: TJC, SJC, PGA of disease activity, and physician assessment of disease activity. TJC and SJC were taken as the number of tender and swollen joints, respectively, out of 28 assessed joints. PGA and physician assessment of disease activity were scored 0-100 mm and rounded to the nearest cm on a VAS, where higher scores indicate greater perceived disease activity. The total CDAI score range was 0-76, where higher scores indicate increased disease activity. Change from baseline was averaged among all participants. Negative values indicate improvement/reduction in RA disease activity.

ITT Set; only those patients who provided data at baseline and at least one post-baseline assessment were analyzed.

SDAI was derived as the sum of the following: TJC, SJC, PGA of disease activity, physician assessment of disease activity, and laboratory-derived C-reactive protein level. TJC and SJC were taken as the number of tender and swollen joints, respectively, out of 28 assessed joints. PGA and physician assessment of disease activity were scored 0-100 mm and rounded to the nearest cm on a VAS, where higher scores indicate greater perceived disease activity. The total SDAI score range was 0-86, where higher scores indicate increased disease activity. Change from baseline was averaged among all participants. Negative values indicate improvement/reduction in RA disease activity.

ITT Set; only those patients who provided data at baseline and at least one post-baseline assessment were analyzed.

TJC was taken as the number of tender joints out of 28 assessed joints.

ITT Set; only those patients who provided data at baseline and at least one post-baseline assessment were analyzed.

SJC was taken as the number of swollen joints out of 28 assessed joints.

ITT Set; only those patients who provided data at baseline and at least one post-baseline assessment were analyzed.

The percentage of participants with at least one adverse event leading to dose/frequency reduction or temporary dose hold was reported.

ITT Set

Participants were evaluated for the presence of anti-tocilizumab antibodies. Confirmatory assays were performed in the case of a positive screen assay result.

ITT Set

Tocilizumab concentration was determined, averaged among all participants, and expressed in micrograms per milliliter (mcg/mL).

ITT Set. The Analysis was conducted on those participants with quantifiable tocilizumab concentration at the specified assessment.

sIL-6R concentration was determined, averaged among all participants, and expressed in nanograms per milliliter (ng/mL).

ITT Set

PGA of disease activity was scored 0-100 mm on a VAS, where higher scores indicate greater perceived disease activity. Change from baseline was averaged among all participants. Negative values indicate improvement/reduction in RA disease activity.

ITT Set; only those patients who provided data at baseline and at least one post-baseline assessment were analyzed.

PGA of RA-related pain was scored 0-100 mm on a VAS, where higher scores indicate greater perceived pain. Change from baseline was averaged among all participants. Negative values indicate improvement/reduction in RA-related pain.

ITT Set; only those patients who provided data at baseline and at least one post-baseline assessment were analyzed.

HAQ-DI consisted of 20 questions assessing ADLs in 8 domains (dress/groom, arise, eat, walk, reach, grip, hygiene) with each item rated 0 (no difficulty) to 3 (unable to do). The highest score recorded for any question in a domain determined the score for that domain, unless assistance was required. The total HAQ-DI score was the sum of domain scores divided by the number of domains answered/scored, for a single score range of 0-3, where higher scores indicate increased functional disability. Change from baseline was averaged among all participants. Negative values indicate improvement in ability to perform ADLs.

ITT Set; only those patients who provided data at baseline and at least one post-baseline assessment were analyzed.

Participants were provided with diary cards to record home injections. Compliance with treatment was calculated individually for each participant as the actual number of injections as a percentage of the planned number of injections (up to the point of discontinuation for those who discontinued study treatment prematurely) and then averaged among all participants.

ITT Set

FACIT-F consisted of 40 questions/statements assessing chronic illness therapy with special emphasis on fatigue over the past 7 days, with each item rated 0 (not at all) to 4 (very much). During score calculations, negatively-worded item scales (e.g., "I have a lack of energy") were reversed so that higher scores indicated more favorable conditions. The total FACIT-F score was the sum of all item scores and ranged 0-160, and the brief FACIT-F score was the sum of 13 item scores and ranged 0-52, where higher scores indicate greater well-being. Change from baseline was averaged among all participants. Positive values indicate improvement in well-being.

ITT Set; only those patients who provided data at baseline and at least one post-baseline assessment were analyzed.