Methods
Study design and treatment
ORAL Sequel (NCT00413699; Study A3921024) was an open-label follow-up LTE study conducted in 414 centers across 43 countries (further details provided in Additional file
1: Table S1). Eligible patients had previously completed a prior qualifying index study of tofacitinib (Additional file
1: Table S1), which included two phase 1 studies, eight phase 2 studies, and six phase 3 studies.
Across qualifying index studies, tofacitinib was dosed at 1, 3, 5, 10, 15, and 30 mg BID, or 20 mg once daily, as monotherapy or in combination with background csDMARDs (mostly MTX).
The majority of enrolled patients from phase 2 qualifying index studies initiated open-label tofacitinib at 5 mg BID, and the majority of patients from phase 3 qualifying index studies initiated open-label tofacitinib at 10 mg BID, except for patients from China who initiated tofacitinib 5 mg BID as per the protocol. After LTE study baseline, the tofacitinib dose could be increased or decreased at the discretion of the investigator (e.g., increased in the case of inadequate control of RA symptoms [5 mg to 10 mg BID] or decreased in response to adverse events [AEs] or laboratory anomalies [10 mg to 5 mg BID]).
Patients receiving tofacitinib were also eligible for temporary discontinuation if deemed necessary by the investigator. In addition, patients participating in a vaccine sub-study (data not reported) underwent temporary withdrawal from tofacitinib treatment.
Patients were allowed to continue or add stable background arthritis therapy (including non-steroidal anti-inflammatory drugs, COX2 inhibitors, and opioids at ≤ 30 mg oral morphine/day potency), certain csDMARDs (MTX, leflunomide, sulfasalazine, anti-malarials, auranofin, and injectable gold preparations at approved doses), and corticosteroids (≤ 10 mg prednisone or equivalent/day), with adjustment permitted at the investigator’s discretion for reasons of inadequate efficacy, or tapering/discontinuation with disease improvement. Patients taking MTX must also have been taking folic acid (according to local standards). Prohibited concomitant medications included moderate/potent cytochrome P450 3A4 inhibitors or inducers.
This study was conducted in accordance with the International Ethical Guidelines for Biomedical Research Involving Human Subjects, the Declaration of Helsinki, and the Good Clinical Practice Guidelines, along with applicable local regulatory requirements and laws. The study protocol was approved by the Institutional Review Boards and/or Independent Ethics Committee at each study center. An independent Data Safety Monitoring Board external to the study sponsor reviewed unblinded safety data on a cumulative basis, and Safety Endpoint Adjudication Committees, blinded to treatment assignment of prior randomized double-blind index studies, supported with standardized safety endpoint assessment for selected events as described below. All patients provided written, informed consent.
Patients
Patients who met the criteria for the index studies, which included a diagnosis of RA based on the American College of Rheumatology (ACR) 1987 Revised Criteria, and had completed the index studies were eligible for enrollment in ORAL Sequel. For those enrolling > 14 days after completion of their prior qualifying index study, the investigator must have deemed that their RA disease activity warranted treatment with tofacitinib and that no evidence of active or inadequately treated Mycobacterium tuberculosis infection be present. Key exclusion criteria for this LTE study included the following for all patients enrolling from the index studies: current/recent history of uncontrolled clinically significant renal, hepatic, hematologic, gastrointestinal, endocrine, pulmonary, cardiac, or neurologic disease; lifetime history of lymphoproliferative disorder, lymphoma, leukemia, or signs and symptoms suggestive of lymphatic disease; history of recurrent herpes zoster infection, current human immunodeficiency virus or hepatitis B/C infection, or any infection requiring hospitalization (including herpes zoster), parenteral antimicrobial therapy, or judged to be opportunistic by the investigator within 6 months prior to the first study drug dose (including those that occurred during the prior qualifying index study); current or history of malignancy (with the exception of adequately treated or excised non-metastatic basal/squamous cell skin cancer or cervical carcinoma in situ); or use of prohibited concomitant medications. For patients enrolling > 14 days after completion of their prior qualifying index study, additional exclusion criteria included hemoglobin levels < 9 g/dL or hematocrit < 30%; absolute white blood cell count < 3.0 × 109/L, absolute neutrophil count (ANC) < 1.2 × 109/L, or absolute lymphocyte count (ALC) < 0.5 × 109/L (< 0.75 × 109/L for patients in Croatia, Czech Republic, Denmark, Ireland, Korea, Germany, Spain, Sweden, and the UK); thrombocytopenia (platelet count < 100 × 109/L); estimated creatinine clearance < 40 mL/min (Cockcroft-Gault calculation); and total bilirubin, aspartate aminotransferase (AST), or alanine aminotransferase (ALT) > 1.5 times the upper limit of normal (ULN).
Objectives and endpoints
The primary objective was to determine the long-term safety and tolerability of tofacitinib 5 mg and 10 mg BID, via evaluation of AE reports, clinical laboratory data, physical examinations, vital signs, and electrocardiogram (ECG) values. The key secondary objective was to evaluate the long-term persistence of efficacy with tofacitinib 5 mg and 10 mg BID, via endpoints including ACR20/50/70 response rates; observed mean score over time in Health Assessment Questionnaire-Disability Index (HAQ-DI) and proportion of patients reporting improvements ≥ the minimum clinically important difference (≥ 0.22) in HAQ-DI; observed mean score over time in Disease Activity Score in 28 joints using erythrocyte sedimentation rate (DAS28-4[ESR]), and proportions of patients achieving DAS28-4(ESR)-defined remission (scores < 2.6) and DAS28-4(ESR)-defined low disease activity (LDA; scores ≤ 3.2); and the observed proportions of patients achieving remission defined by Clinical (CDAI) and Simplified (SDAI) Disease Activity Indices (scores ≤ 2.8 and ≤ 3.3, respectively). Exploratory ACR20/50/70 response maintenance analyses and CDAI/SDAI remission maintenance analyses were also conducted.
Statistical methods
Statistical analyses of safety and efficacy data are descriptive in nature, and no formal comparisons between tofacitinib treatment groups were performed.
Baseline values for safety and efficacy endpoints were those of the qualifying index study (index baseline) for patients who enrolled in the LTE study within ≤ 14 days of index study completion, or the LTE baseline for patients who enrolled > 14 days after index study completion. The safety analysis set comprised all patients who received at least one dose of study medication.
Assignment to the tofacitinib 5 mg or 10 mg BID arm for analysis of efficacy and safety outcomes in this LTE study was based on the study average total daily dose (TDD) for each patient (i.e., sum of all doses received divided by number of days of treatment over the entire study duration for each patient): tofacitinib 5 mg BID if TDD < 15 mg and tofacitinib 10 mg BID if TDD ≥ 15 mg. Patients were assigned to “stay-on monotherapy” (received tofacitinib monotherapy) or “stay-on background csDMARDs” (received tofacitinib plus csDMARD combination therapy) subgroups for analysis of safety outcomes if they remained on their initial study-start therapy for the entire duration of the study (irrespective of tofacitinib or csDMARD dose adjustments), with the exception of a ≤ 28-day break in csDMARD use allowed for stay-on csDMARD patients. Patients who switched from csDMARD to monotherapy, or vice versa, were not included in these subgroup analyses and, therefore, they do not sum to total number of patients treated with tofacitinib. Equivalent subgroup analyses for efficacy outcomes, including data pooled from ORAL Sequel, have been previously published [
17].
Exposure-adjusted event rates of the number of patients with events per 100 patient-years (EAERs per 100 patient-years) were calculated for AEs, and incidence rates of the number of patients with events per 100 patient-years (IRs per 100 patient-years) and 95% confidence intervals (CI; calculated via the Exact Poisson method) were calculated for AEs of special interest. EAERs were based on the number of unique patients with events per 100 patient-years over all patients’ exposures between their first dose of tofacitinib and their last dose (excluding any temporary treatment breaks in between). IRs were based on the number of unique patients with events during the time between the first and last tofacitinib dose plus 28 days, divided by the time accruing during the risk period (i.e., between the first and last tofacitinib dose plus 28 days, or the time accruing to the first event, whichever occurred earlier). The recurrence rate of herpes zoster was also calculated for patients with at least one event within the risk period.
In relation to IRs for AEs of special interest, cardiovascular events were adjudicated from February 2009, opportunistic infections from February 2013, hepatic events from December 2012, gastrointestinal events from December 2014, and interstitial lung disease events from April 2014. Events prior to these dates were not adjudicated and were identified by clinical review of AEs. For malignancies, the central histopathological review of AEs was initiated in July 2009, with events adjudicated from February 2014. Events prior to this were subsequently reviewed and adjudicated.
The efficacy analyses were conducted for all patients who received at least one dose of study medication and had at least one post-index/LTE baseline efficacy measurement available. All analyses were based on observed data with no imputation for missing data.
Discussion
The ORAL Sequel LTE study of tofacitinib is part of the largest clinical development program undertaken for any RA treatment to date. Up to 9.5 years of treatment with tofacitinib in more than 4000 patients worldwide with RA is represented in this study, with a combined tofacitinib exposure of > 16,000 patient-years.
Safety data up to 114 months for all tofacitinib, and efficacy data up to 96 months for tofacitinib 5 mg BID and 72 months for 10 mg BID, are reported, with low patient numbers limiting interpretation beyond these time points. Patient baseline demographic and disease characteristics were generally similar between the tofacitinib 5 mg and 10 mg BID treatment arms, although a greater proportion of patients enrolled from prior phase 3 studies compared with phase 2 studies and then initiated the LTE at 10 mg BID, per protocol (the exception to this was that all phase 3 patients from China initiated the LTE at 5 mg BID, also per protocol). Most (> 90%) patients were enrolled in the LTE study within ≤ 14 days of index study completion, and so their baseline index data were used, per protocol.
In all, the safety profile remained consistent with that observed in prior phase 2 [
2‐
6] or phase 3 [
7,
9‐
13] tofacitinib studies, or prior combined analyses including LTE data [
16,
18]. The IRs for AEs leading to discontinuation were comparable for all tofacitinib, tofacitinib 5 mg BID, and tofacitinib 10 mg BID (6.8, 6.7, and 6.8, respectively), as were the IRs for SAEs (9.0, 8.2, and 9.4, respectively). Overall in ORAL Sequel, 52% of patients discontinued by month 114 (24% due to AEs and 4% due to insufficient clinical response) and the incidence of AEs leading to discontinuation remained stable over time. These data are consistent with a recent integrated analysis of safety data over 8.5 years for tofacitinib (5 mg and 10 mg BID combined, with or without background csDMARDs) across 17 phase 1/2/3 studies and two LTE studies, including ORAL Sequel, in which the IR for discontinuation due to AEs was 7.5 and the IR for SAEs was 9.0, for all tofacitinib [
18]. Furthermore, in pooled LTE safety analyses from ORAL Sequel and Japanese Study A3921041 (tofacitinib 5 mg and 10 mg BID combined, with or without background DMARDs), 51% of patients discontinued by month 114 (24% due to AEs and 4% due to insufficient clinical response) and the IR for SAEs was 9.1 for all tofacitinib [
16].
Most malignancy-, cardiovascular-, mortality-, and infection-related events had IRs < 0.5; exceptions were herpes zoster, all serious infections, malignancies excluding NMSC, and NMSC. IRs for herpes zoster, all serious infections, and malignancies excluding NMSC remained generally stable over time. Recurrent herpes zoster was observed in 30 patients. IRs were higher (CI non-overlapping) with tofacitinib 10 mg BID versus 5 mg BID for herpes zoster and opportunistic infections excluding tuberculosis and numerically higher (CI marginally overlapping) for all serious infections; no dose-dependencies were indicated for other events.
Infections and infestations were anticipated to represent the most common class of overall AEs. IRs for serious infections were consistent between ORAL Sequel, the integrated analysis of safety data for tofacitinib [
18], and pooled ORAL Sequel and Study A3921041 data [
16], at 2.4, 2.7, and 2.5, respectively, as were IRs for herpes zoster between ORAL Sequel and the integrated analysis of safety data, at 3.4 and 3.9, respectively (data not reported for the pooled ORAL Sequel and Study A3921041 analysis; IR was 7.4 for herpes zoster in the Japanese population of Study A3921041 alone [
15]). With respect to tuberculosis, the IR for all tofacitinib (0.2) was slightly lower compared with the IR for other opportunistic infections excluding tuberculosis (0.4) in ORAL Sequel. In the integrated analysis of safety data for tofacitinib, IRs were similar between tuberculosis (0.2) and other opportunistic infections excluding tuberculosis (0.3), and concordant with ORAL Sequel data [
18]. Prior evaluation has supported the understanding that the risk of tuberculosis in patients receiving tofacitinib treatment directly varies with background tuberculosis prevalence in different geographic sub-populations [
19]. ORAL Sequel IR data for tuberculosis are comparable with data for biologic DMARDs [
20‐
22].
Of interest, the long-latency events MACE and lymphoma had IRs at 0.4 and 0.1, respectively. IRs for other AEs of interest reported in both ORAL Sequel and the integrated analysis of safety data for tofacitinib [
18] were consistent: 0.8 and 0.9 for malignancies excluding NMSC, 0.7 and 0.6 for NMSC, and both 0.1 for gastrointestinal perforation (which is also comparable with tumor necrosis factor inhibitors [TNFi] [
23]). Furthermore, the IRs for DVT and PE in patients receiving tofacitinib (both 0.1) were comparable to those reported in data pooled from phase 2 and phase 3 studies of tofacitinib as monotherapy or in combination with DMARDs (DVT IR of 0.1 for both tofacitinib 5 and 10 mg BID, and PE IR of 0.1 for tofacitinib 5 mg BID and 0.2 for tofacitinib 10 mg BID [
24]) and in the literature for patients with RA, including those treated with DMARDs (DVT IRs of 0.45 [
25] and 0.62 [
26], and PE IRs of 0.26 [
25] and 0.20 [
26]).
In support of the consistencies noted above, safety data from LTE studies and meta-analyses of biologic and targeted synthetic DMARDs are comparable with those reported for tofacitinib. Considering discontinuation from biologic DMARDs, 62% of patients discontinued by year 10 of treatment in an LTE study of adalimumab (23% due to AEs and 12% due to lack of efficacy/disease progression) [
27]; and 31% of patients discontinued by year 5 in an LTE study of subcutaneous abatacept (31% due to AEs and 7% due to lack of efficacy) [
28]. Considering safety events of special interest, previous LTE studies have reported IRs for serious infections of 3.1 and 1.7 events for adalimumab (across 10 years of treatment) and subcutaneous abatacept (across 5 years of treatment), respectively [
27,
28]. In the integrated analysis of safety data for baricitinib, event IRs were reported for serious infections (3.0), herpes zoster (3.3), tuberculosis (0.1), MACE (0.5), DVT (0.4), PE (0.2), malignancies excluding NMSC (0.8), lymphoma (0.08), NMSC (0.4), and gastrointestinal perforation (0.04) [
29,
30]. Meta-analysis data showed that in a dataset comprising RCTs and LTE studies, the IRs for tofacitinib (serious infection 2.74; malignancies excluding NMSC 0.89) were comparable to those reported for biologic and targeted synthetic DMARDs, including abatacept, adalimumab, baricitinib, certolizumab pegol, etanercept, golimumab, infliximab, rituximab, and tocilizumab (serious infection 3.04–7.59; malignancies excluding NMSC 0.75–1.06) [
31,
32], which in turn are concordant with data reported for ORAL Sequel. In addition, serious infection IRs for tofacitinib were observed to be congruous with IRs previously calculated in a 2017 meta-analysis of RCTs for biologic DMARDs, including abatacept (3.0), rituximab (3.5), tocilizumab (5.4), and TNFi (5.5), and the targeted synthetic DMARD baricitinib (4.8 for 2 mg, 3.7 for 4 mg) [
31]. Taking the findings of ORAL Sequel and these data together, this supports the work of other groups concluding that the rates of serious infections for tofacitinib are within the range of those reported for biologic DMARDs (up to 12 weeks only) [
33] and the targeted synthetic DMARD, baricitinib (versus placebo) [
29].
AE data in ORAL Sequel indicated some differences between tofacitinib monotherapy and combination therapy. IRs for AEs leading to discontinuation in ORAL Sequel were higher (CI non-overlapping) in patients receiving tofacitinib combination therapy (7.7) compared with patients receiving tofacitinib monotherapy (5.9); SAE IRs were also numerically different (CI marginally overlapping) between patients receiving combination therapy (9.5) and tofacitinib monotherapy (8.1). In terms of AEs of special interest, IRs were higher (CI non-overlapping) in patients receiving tofacitinib as combination therapy versus monotherapy for herpes zoster. Within the subgroup of patients receiving tofacitinib as combination therapy, IRs were higher (CI non-overlapping) with tofacitinib 10 mg BID versus 5 mg BID for all serious infections, herpes zoster, and opportunistic infections excluding tuberculosis.
Laboratory variables of interest, including cholesterol, LDL, ALT, AST, and serum creatinine, remained generally stable over time
. ALC levels gradually declined until month 48, whereafter the level stabilized. Overall, 40.6% of patients experienced moderate lymphopenia, and 30.1% of patients experienced severe lymphopenia; these events did not meet with study discontinuation criteria, per protocol. Approximately 1% of patients had a drop in lymphocyte count to < 500 cells/mm
3 (potentially life-threatening lymphopenia); however, nearly all recovered to above 500 cells/mm
3 upon treatment discontinuation. Additionally, five of these patients experienced a serious infection during the study; two occurred within 30 days of the patients’ lowest lymphocyte count. Although the proportion of patients who developed a serious infection within 30 days of the lowest lymphocyte count was highest among patients with potentially life-threatening lymphopenia, these results should be interpreted with caution due to the low number of patients with ALC < 500 cells/mm
3. As per the tofacitinib prescribing information, routine clinical practice should involve evaluation of lymphocyte count at baseline, monitoring every 3 months during tofacitinib treatment, and consideration of the following actions: discontinuation if ALC reaches < 500 cells/mm
3 (a level associated with increased risk of serious infections [
18,
34]), dose reduction or interruption if ALC reaches 500–750 cells/mm
3, and dose maintenance if ALC is ≥ 750 cells/mm
3 [
35,
36].
Clinically meaningful improvements in the signs and symptoms of RA as measured by ACR response rates, DAS28-4(ESR) improvements from baseline, and CDAI- and SDAI-defined remission rates, as well as improvements in physical functioning as measured by HAQ-DI improvements from baseline, were also maintained over time in patients who remained on tofacitinib treatment. Efficacy data for patients receiving tofacitinib monotherapy or tofacitinib plus csDMARD combination therapy were not available here, but have been previously published [
17]; in this earlier analysis of data pooled from ORAL Sequel and Study A3921041, efficacy was maintained through to month 72, regardless of treatment regimen.
It is acknowledged that LTE studies are limited by enrolling eligible patients who completed preceding index studies and excluding those who developed tofacitinib treatment-related SAEs and so were discontinued from preceding index studies. As per the protocol, significant opportunistic infections or serious infection events, and certain malignancies, precluded enrollment to ORAL Sequel. LTE study populations, therefore, represent patients in whom the study drug is known to be efficacious and well tolerated, restricting full evaluation of the benefit to risk profile. It is important to note that a greater exposure (patient-years) occurred with tofacitinib 10 mg BID compared with 5 mg BID, as a result of the majority of patients enrolling from large phase 3 studies and initiating the LTE study at 10 mg BID per protocol (except for patients from China, bringing a geographic skew). Caution should be taken when interpreting results for patients with the longest tofacitinib exposure due to the relatively small patient numbers at later months. Furthermore, tofacitinib dose assignment based on using average TDD does not account for cumulative treatment exposure or dose changes over time. It also does not provide the actual dose when the event occurred; therefore, although 77% of patients remained on their study-start dose, this may have attenuated any between-dose differences for the 5 mg versus 10 mg BID tofacitinib dose. We also recognize that there may be limitations to the interpretation of efficacy results, given that all analyses were conducted “as observed” and patients not achieving an effect on treatment in the qualifying study were more likely to not enroll in the LTE study or discontinue. To this point, it should be noted that only < 4% of patients discontinued from the LTE study due to insufficient clinical response. The lack of a placebo or comparator arm also restricts benefit to risk profile evaluation. Finally, comparisons of data for combination versus monotherapy should be treated with caution because each prior qualifying index study defined the regimen for all patients within that study, and patients were not randomized to dose for the LTE study or to one regimen versus the other (monotherapy versus combination therapy). The monotherapy subgroup was also smaller (n = 1298) than the combination therapy subgroup (n = 2464).
Nevertheless, open-label LTE studies provide valuable information. A controlled setting with rigorous safety reporting, including independent adjudication of events, is facilitated. Data for drug exposure are collected over months or years, yielding a long-term safety profile. Indeed, additional long-term studies of tofacitinib are ongoing: Study A3921133 (NCT02092467), an ongoing post-marketing surveillance study, was initiated to compare the safety of tofacitinib versus TNFi in patients with RA in an open-label design. With final enrollment estimated at ~ 4400 patients, outcomes are event-driven, with malignancies excluding NMSC and the incidence of MACE being the primary outcome measures. Study A3921133 is, therefore, anticipated to provide further insights that will contribute further to the evaluation of the benefit to risk profile of tofacitinib. Furthermore, a recent publication on post-marketing surveillance experience with tofacitinib showed that no new safety risks were revealed in this real-world setting compared with the safety profile identified in the RA clinical development program [
37]. As of December 2017, it is estimated that over 115,000 patients with RA worldwide have received tofacitinib (Pfizer data on file).