Incidence rate (number of participants with event per 100 participant year [PY]) was defined as the total number of participants with admissible events divided by the total (for all qualifying participants) time at risk for the cohort/treatment group of interest. Malignancy events, excluding NMSC were adjudicated by a steering committee. The risk period (RP) was the last contact date. The last contact date was the maximum of (AE start date, AE stop date, last study visit date, withdrawal date, telephone contact date). If a participant died, last contact date was the death date. First events were counted within the RP. If a participant did not have an event or had an event but outside the risk period, the participant was censored at the end of RP.

SAS included all participants who were randomized in the study and received at least one dose of the randomized investigational drug (tofacitinib or TNFi).

MACE included the cardiovascular death, non-fatal myocardial infarction (MI) and non-fatal stroke of any classification, including reversible focal neurologic defects with imaging evidence of a new cerebral lesion consistent with ischemia or hemorrhage. Incidence rate was defined as the total number of participants with admissible events divided by the total (for all qualifying participants) time at risk for the cohort/treatment group of interest. The risk period (RP) was the minimum of last contact date or last study treatment dose date + 60 days. The last contact date was the maximum of (AE start date, AE stop date, last study visit date, withdrawal date, telephone contact date). If a participant died, last contact date was the death date. First events were counted within the RP. If a participant did not have an event or had an event but outside the risk period, the participant was censored at the end of RP.

SAS included all participants who were randomized in the study and received at least one dose of the randomized investigational drug (tofacitinib or TNFi).

Non-fatal stroke included reversible focal neurologic defects with imaging evidence of a new cerebral lesion consistent with ischemia or hemorrhage. Incidence rate was defined as the total number of participants with admissible events divided by the total (for all qualifying participants) time at risk for the cohort/treatment group of interest. The risk period (RP) was the minimum of last contact date or last study treatment dose date + 60 days. The last contact date was the maximum of (AE start date, AE stop date, last study visit date, withdrawal date, telephone contact date). If a participant died, last contact date was the death date. First events were counted within the RP. If a participant did not have an event or had an event but outside the risk period, the participant was censored at the end of RP.

SAS included all participants who were randomized in the study and received at least one dose of the randomized investigational drug (tofacitinib or TNFi).

Incidence rate was defined as the total number of participants with admissible events divided by the total (for all qualifying participants) time at risk for the cohort/treatment group of interest. The risk period (RP) was the minimum of last contact date or last study treatment dose date + 60 days. The last contact date was the maximum of (AE start date, AE stop date, last study visit date, withdrawal date, telephone contact date). If a participant died, last contact date was the death date. First events were counted within the RP. If a participant did not have an event or had an event but outside the risk period, the participant was censored at the end of RP.

SAS included all participants who were randomized in the study and received at least one dose of the randomized investigational drug (tofacitinib or TNFi).

Opportunistic infections (OI) were reviewed and adjudicated by the opportunistic infection review committee (OIRC). Incidence rate was defined as the total number of participants with admissible events divided by the total (for all qualifying participants) time at risk for the cohort/treatment group of interest. The risk period (RP) was the minimum of last contact date or last study treatment dose date + 28 days. The last contact date was the maximum of (AE start date, AE stop date, last study visit date, withdrawal date, telephone contact date). If a participant died, last contact date was the death date. First events were counted within the RP. If a participant did not have an event or had an event but outside the risk period, the participant was censored at the end of RP.

SAS included all participants who were randomized in the study and received at least one dose of the randomized investigational drug (tofacitinib or TNFi).

Hepatic events (adjudicated) included drug-induced liver injury (DILI) - probable, highly likely and definite, DILI - listed separately, DILI - cases meeting classification and severity, participants with elevations of transaminase levels greater than (>) 1* upper limit of normal (ULN), greater than or equal to (>=) 3*ULN, >=5*ULN (based on laboratory values). Incidence rate was the total number of participants with admissible events divided by total (for all qualifying participants) time at risk for the cohort/treatment group of interest. The risk period (RP) was minimum of The risk period (RP) was the minimum of last contact date or last study treatment dose date + 28 days. Last contact date was maximum of (AE start date, AE stop date, last study visit date, withdrawal date, telephone contact date). In case of death, last contact date was death date. First events counted within RP. Participant did not have an event or had an event outside risk period were censored at end of RP.

SAS included all participants who were randomized in the study and received at least one dose of the randomized investigational drug (tofacitinib or TNFi).

Cardiovascular events (adjudicated) were death (coronary and non-coronary), MI, all coronary revascularization, unstable angina, new ischemic heart disease, stroke (fatal and non-fatal), transient ischemic attack (TIA), congestive heart failure (CHF), peripheral arterial vascular disease (PAVD), deep vein thrombosis, pulmonary embolism, arterial embolism, arterial thrombosis. Incidence rate was total number of participants with admissible events divided by total (for all qualifying participants) time at risk for cohort/treatment group of interest. Risk period (RP) was the minimum of last contact date or last study treatment dose date + 28 days. Last contact date was maximum of (AE start date, AE stop date, last study visit date, withdrawal date, telephone contact date). In case of death, last contact date was death date. First events counted within RP. Participant did not have an event or had an event outside risk period were censored at end of RP.

SAS included all participants who were randomized in the study and received at least one dose of the randomized investigational drug (tofacitinib or TNFi).

AE was any untoward medical occurrence post treatment; event need not necessarily had causal relationship with treatment or usage. SAE: any untoward medical occurrence at any dose: resulted in death, life-threatening, required inpatient hospitalization or prolongation of existing hospitalization, resulted in persistent or significant disability/incapacity, resulted in congenital anomaly. TEAE: event that occurred for first time during effective duration of treatment and not seen prior to start of treatment or event seen prior to start of treatment but increase in severity during treatment. Risk period (RP) for AE: minimum of last contact date or last study treatment dose date+28 days. RP for SAEs: last contact date. Last contact date was maximum: AE start, AE stop, last study visit, withdrawal and telephone contact. In case of death, last contact was death date. First events counted within RP. Participant did not have event or had event outside risk period were censored at end of RP.

SAS included all participants who were randomized in the study and received at least one dose of the randomized investigational drug (tofacitinib or TNFi).

Clinically significant laboratory abnormalities: Hematology: hemoglobin, hematocrit, erythrocytes with primary criteria as less than [<] 0.8* lower limit of normal [LLN]), platelets (<0.5* LLN; >1.75* ULN), leukocytes (<0.6*LLN; >1.5*ULN), lymphocytes, lymphocytes/leukocytes, neutrophils, neutrophils/leukocytes (<0.8*LLN; >1.2*ULN), eosinophils, eosinophils/leukocytes, monocytes, monocytes/leukocytes (>1.2*ULN); urinalysis: urine glucose, urine protein, urine hemoglobin, and leukocyte esterase (>=1); chemistry: bilirubin, indirect bilirubin (>1.5*ULN) aspartate aminotransferase, alanine aminotransferase (>3.0*ULN), creatinine, triglycerides, cholesterol (>1.3*ULN) and HDL cholesterol (<0.8*LLN). Risk period (RP) was minimum of last contact date or last study treatment dose date+28 days. Last contact date was (date of death or maximum of dates: AE start, AE stop, last study visit, withdrawal, telephone contact). Participants without event or event outside RP were censored at end of RP.

Analysis population included all participants from SAS with an abnormal baseline with at least one observation of the given laboratory test while on study treatment or during lag time.

All-cause death was defined as the death due to any cause during the course of study. Incidence rate was defined as the total number of participants with admissible events divided by the total (for all qualifying participants) time at risk for the cohort/treatment group of interest. Incidence rate of all-cause deaths (adjudicated by Adjudication Committee) was reported in this outcome measure. The risk period (RP) was the minimum of last contact date or last study treatment dose date + 28 days. The last contact date was the maximum of (AE start date, AE stop date, last study visit date, withdrawal date, telephone contact date). If a participant died, last contact date was the death date. First events were counted within the RP. If a participant did not have an event or had an event but outside the risk period, the participant was censored at the end of RP.

SAS included all participants who were randomized in the study and received at least one dose of the randomized investigational drug (tofacitinib or TNFi).

Number of participants who permanent or temporary discontinued study medication due to any AE, treatment related AEs, Coronavirus disease 2019 (COVID 19) related AEs, and herpes zoster were reported. The risk period (RP) was the minimum of last contact date or last study treatment dose date + 28 days. The last contact date was the maximum of (AE start date, AE stop date, last study visit date, withdrawal date, telephone contact date). If a participant died, last contact date was the death date. First events were counted within the RP. If a participant did not have an event or had an event but outside the risk period, the participant was censored at the end of RP.

SAS included all participants who were randomized in the study and received at least one dose of the randomized investigational drug (tofacitinib or TNFi).

DAS28 is a measure of disease activity in participants with rheumatoid arthritis based on a 28-joint assessment. DAS28-4 (CRP) was calculated from number of painful joints out of 28 joints (TJC28) and number of swollen joints out of 28 joints (SJC28), CRP (milligrams per liter [mg/L]) and patient's global assessment of disease activity (PtGA) on a 100 mm Visual Analog Scale (VAS) (scores ranging from 0 millimeter [mm] [very well] to 100 mm [worst], higher scores indicated worse health condition). Total DAS28-4 (CRP) score range: 0 to 9.4, higher score indicated more disease activity. DAS28-4 (CRP) <= 3.2 indicates low disease activity and > 3.2 to <=5.1 indicates moderate disease activity, >5.1 indicates high disease activity, and DAS28-4 (CRP) < 2.6 indicates remission. DAS28-4 (CRP) = 0.56*sqrt(TJC28) + 0.28*sqrt(SJC28) + 0.36*ln(CRP in mg/L +1) + 0.014*PtGA in mm+ 0.96; ln = natural logarithm, sqrt = square root.

Full Analysis Set (FAS): all participants randomized and received at least one dose of drug. Mixed model for repeated measures used without imputation for missing values. Analyses included data while participants were receiving drug. Only visits with N > 50 participants in each treatment group included in analyses. Number of Participants Analyzed: number of participants included in mixed model for repeated measures. Number Analyzed: number of participants evaluable for each specified time point.

SDAI is the numerical sum of five outcome parameters: TJC and SJC both based on a 28-joint assessment, PtGA and physician's global assessment of health (PhyGA) both assessed on a 0 to 100 mm VAS (higher scores indicate greater affection due to disease activity), and CRP (mg/L). SDAI total score ranges from 0 to 86 with higher score indicating greater disease activity. SDAI <=3.3 indicates disease remission, >3.4 to 11 indicates low disease activity >11 to 26 indicates moderate disease activity, and >26 indicates high disease activity. SDAI = (28TJC) + (28SJC) + PhyGA/10 + PtGA/10 + CRP/10.

FAS: all participants randomized and received at least one dose of drug. Mixed model for repeated measures used without imputation for missing values. Analyses included data while participants were receiving drug. Only visits with N > 50 participants in each treatment group included in analyses. Number of Participants Analyzed: number of participants included in mixed model for repeated measures. Number Analyzed: number of participants evaluable for each specified time point.

CDAI is the numerical sum of four outcome parameters: TJC and SJC both based on a 28-joint assessment, PtGA and PhyGA both assessed on a 0 to 100 mm VAS (higher scores indicate greater affection due to disease activity). CDAI total score ranges from 0 to 76 with higher score indicating greater disease activity. CDAI <=2.8 indicates disease remission, >2.8 to 10 indicates low disease activity, >10 to 22 indicates moderate disease activity, and >22 indicates high disease activity. CDAI = (28TJC) + (28SJC) + PhyGA/10 + PtGA/10.

FAS: all participants randomized and received at least one dose of drug. Mixed model for repeated measures used without imputation for missing values. Analyses included data while participants were receiving drug. Only visits with N > 50 participants in each treatment group included in analyses. Number of Participants Analyzed: number of participants included in mixed model for repeated measures. Number Analyzed: number of participants evaluable for each specified time point.

ACR-EULAR Boolean-based definition of remission participant must satisfy all of the following: TJC28 <=1, SJC28 <=1, CRP <=10 mg/L, PtGA on a 0-100 mm scale, higher scores indicate greater affection due to disease activity.

FAS included all participants randomized in study and received at least one dose of randomized investigational drug. Analyses included data while participants were receiving the study drug. 'Number Analyzed' signifies number of participants evaluable for each specified time point. Last Observation Carried Forward (LOCF) applied for missing components and composite binary outcome was calculated. The percentages were calculated using numbers of participants evaluable at each visit as denominators.

SDAI is the numerical sum of five outcome parameters: TJC and SJC both based on a 28-joint assessment, PtGA and PhyGA both assessed on a 0 to 100 mm VAS (higher scores indicate greater affection due to disease activity), and CRP (mg/L). SDAI total score ranges from 0 to 86 with higher score indicating greater disease activity. SDAI <=3.3 indicates disease remission, >3.4 to 11 indicates low disease activity >11 to 26 indicates moderate disease activity, and >26 indicates high disease activity. SDAI = (28TJC) + (28SJC) + PhyGA/10 + PtGA/10 + CRP/10.

FAS included all participants randomized in the study and received at least one dose of investigational drug. Analyses included the data while the participants were receiving the study drug. Here, 'Number Analyzed' signifies number of participants evaluable for each specified time point. LOCF was applied for missing components, and then composite binary outcome was calculated. The percentages were calculated using the numbers of participants evaluable at each visit as the denominators.

CDAI is the numerical sum of four outcome parameters: TJC and SJC both based on a 28-joint assessment, PtGA and PhyGA both assessed on a 0 to 100 mm VAS (higher scores indicate greater affection due to disease activity). CDAI total score ranges from 0 to 76 with higher score indicating greater disease activity. CDAI <=2.8 indicates disease remission, >2.8 to 10 indicates low disease activity, >10 to 22 indicates moderate disease activity, and >22 indicates high disease activity. CDAI = (28TJC) + (28SJC) + PhyGA/10 + PtGA/10.

FAS included all participants randomized in the study and received at least one dose of investigational drug. Analyses included the data while the participants were receiving the study drug. Here, 'Number Analyzed' signifies number of participants evaluable for each specified time point. LOCF was applied for missing components, and then composite binary outcome was calculated. The percentages were calculated using the numbers of participants evaluable at each visit as the denominators.

SDAI is the numerical sum of five outcome parameters: TJC and SJC both based on a 28-joint assessment, PtGA and PhyGA both assessed on a 0 to 100 mm VAS (higher scores indicate greater affection due to disease activity), and CRP (mg/L). SDAI total score ranges from 0 to 86 with higher score indicating greater disease activity. SDAI <=3.3 indicates disease remission, >3.4 to 11 indicates low disease activity >11 to 26 indicates moderate disease activity, and >26 indicates high disease activity. SDAI = (28TJC) + (28SJC) + PhyGA/10 + PtGA/10 + CRP/10.

FAS included all participants randomized in the study and received at least one dose of investigational drug. Analyses included the data while the participants were receiving the study drug. Here, 'Number Analyzed' signifies number of participants evaluable for each specified time point. LOCF was applied for missing components, and then composite binary outcome was calculated. The percentages were calculated using the numbers of participants evaluable at each visit as the denominators.

CDAI is the numerical sum of four outcome parameters: TJC and SJC both based on a 28-joint assessment, PtGA and PhyGA both assessed on a 0 to 100 mm VAS (higher scores indicate greater affection due to disease activity). CDAI total score ranges from 0 to 76 with higher score indicating greater disease activity. CDAI <=2.8 indicates disease remission, >2.8 to 10 indicates low disease activity, >10 to 22 indicates moderate disease activity, and >22 indicates high disease activity. CDAI = (28TJC) + (28SJC) + PhyGA/10 + PtGA/10.

FAS included all participants randomized in the study and received at least one dose of investigational drug. Analyses included the data while the participants were receiving the study drug. Here, 'Number Analyzed' signifies number of participants evaluable for each specified time point. LOCF was applied for missing components, and then composite binary outcome was calculated. The percentages were calculated using the numbers of participants evaluable at each visit as the denominators.

DAS28 is a measure of disease activity in participants with rheumatoid arthritis based on a 28-joint assessment. DAS28-4 (CRP) was calculated from number of painful joints out of 28 joints (TJC28) and number of swollen joints out of 28 joints (SJC28), CRP (mg/L) and PtGA on a 100 mm VAS (VAS: scores ranging from 0 mm [very well] to 100 mm [worst], higher scores indicated worse health condition). Total DAS28-4 (CRP) score range: 0 to 9.4, higher score indicated more disease activity. DAS28-4 (CRP) <= 3.2 indicates low disease activity and > 3.2 to <=5.1 indicates moderate disease activity, >5.1 indicates high disease activity, and DAS28-4 (CRP) < 2.6 indicates remission. DAS28-4 (CRP) = 0.56*sqrt(TJC28) + 0.28*sqrt(SJC28) + 0.36*ln(CRP in mg/L +1) + 0.014*PtGA in mm+ 0.96.

FAS included all participants randomized in the study and received at least one dose of investigational drug. Analyses included the data while the participants were receiving the study drug. Here, 'Number Analyzed' signifies number of participants evaluable for each specified time point. LOCF was applied for missing components, and then composite binary outcome was calculated. The percentages were calculated using the numbers of participants evaluable at each visit as the denominators.

ACR20 response is a >= 20% improvement in TJC (28) and SJC (28) and >=20% improvement in 3 of the 5 remaining ACR-core criteria: 1) PGA of arthritis, 2) PtGA of arthritis, 3) participant's assessment of arthritis pain, 4) participant's assessment of functional disability by HAQ-DI, and 5) CRP (mg/L) at each visit. PGA: physician's global assessment of arthritis on VAS, 0 (very well) to 100 mm (worst arthritis), higher scores=worse condition. PtGA: participant's global assessment of arthritis on VAS, 0 mm (very well) to 100 mm (worst arthritis condition), higher scores = worse condition. Participant's assessment of arthritis pain: assessed on VAS, 0 mm (no pain) to 100 mm (most severe pain), higher score = more pain. HAQ-DI: functional disability evaluation, score: 0 (no difficulty) to 3 (unable to do), higher score=more disability.

FAS included all participants randomized in the study and received at least one dose of investigational drug. Analyses included the data while the participants were receiving the study drug. Here, 'Number Analyzed' signifies number of participants evaluable for each specified time point. LOCF was applied for missing components, and then composite binary outcome was calculated.

ACR50 response is a >= 50% improvement in TJC (28) and SJC (28) and >=50% improvement in 3 of the 5 remaining ACR-core criteria: 1) PGA of arthritis, 2) PtGA of arthritis, 3) participant's assessment of arthritis pain, 4) participant's assessment of functional disability by HAQ-DI, and 5) CRP (mg/L) at each visit. PGA: physician's global assessment of arthritis on VAS, 0 (very well) to 100 mm (worst arthritis), higher scores=worse condition. PtGA: participant's global assessment of arthritis on VAS, 0 mm (very well) to 100 mm (worst arthritis condition), higher scores = worse condition. Participant's assessment of arthritis pain: assessed on VAS, 0 mm (no pain) to 100 mm (most severe pain), higher score = more pain. HAQ-DI: functional disability evaluation, score: 0 (no difficulty) to 3 (unable to do), higher score=more disability.

FAS included all participants randomized in the study and received at least one dose of investigational drug. Analyses included the data while the participants were receiving the study drug. Here, 'Number Analyzed' signifies number of participants evaluable for each specified time point. LOCF was applied for missing components, and then composite binary outcome was calculated.

ACR70 response is a >= 70% improvement in TJC (28) and SJC (28) and >=70% improvement in 3 of the 5 remaining ACR-core criteria: 1) PGA of arthritis, 2) PtGA of arthritis, 3) participant's assessment of arthritis pain, 4) participant's assessment of functional disability by HAQ-DI, and 5) CRP (mg/L) at each visit. PGA: physician's global assessment of arthritis on VAS, 0 (very well) to 100 mm (worst arthritis), higher scores=worse condition. PtGA: participant's global assessment of arthritis on VAS, 0 mm (very well) to 100 mm (worst arthritis condition), higher scores = worse condition. Participant's assessment of arthritis pain: assessed on VAS, 0 mm (no pain) to 100 mm (most severe pain), higher score = more pain. HAQ-DI: functional disability evaluation, score: 0 (no difficulty) to 3 (unable to do), higher score=more disability.

FAS included all participants randomized in the study and received at least one dose of investigational drug. Analyses included the data while the participants were receiving the study drug. Here, 'Number Analyzed' signifies number of participants evaluable for each specified time point. LOCF was applied for missing components, and then composite binary outcome was calculated.

HAQ-DI assesses the degree of difficulty a participant has experienced during the past week in 8 domains of daily living activities: dressing/grooming; arising; eating; walking; reach; grip; hygiene; and other activities. There were total of 30 items distributed in these 8 domains. Each item was scored on a 4-point scale from 0 to 3: 0= no difficulty; 1= some difficulty; 2= much difficulty; 3= unable to do. Overall score was computed as the sum of domain scores and divided by the number of domains answered. Total possible score range 0 (least difficulty) and 3 (extreme difficulty), where higher scores indicate more difficulty while performing daily living activities.

FAS: all participants randomized and received at least one dose of drug. Mixed model for repeated measures used without imputation for missing values. Analyses included data while participants were receiving drug. Only visits with N > 50 participants in each treatment group included in analyses. Number of Participants Analyzed: number of participants included in mixed model for repeated measures. Number Analyzed: number of participants evaluable for each specified time point.