Post-marketing surveillance of tofacitinib in patients with ulcerative colitis in Japan: a post hoc analysis of safety and effectiveness in older (≥65 years) and younger (<65 years) patients
Article information
Abstract
Background/Aims
To assess, post hoc, tofacitinib safety/effectiveness in patients with ulcerative colitis (UC), stratified by age, using data from a 60-week post-marketing surveillance (PMS) study in Japan.
Methods
All patients with UC receiving tofacitinib in Japan were enrolled in a large PMS study. Incidence proportions of adverse events (AEs), incidence rates (IRs; unique patients with events/100 patient-years of exposure) of clinically important AEs, reasons for discontinuation, and partial Mayo score clinical remission, stratified by age (≥ 65 and < 65 years), were evaluated.
Results
The analysis included 212 older (≥ 65 years) and 1,770 younger (< 65 years) patients. Demographics and baseline disease characteristics were generally similar between groups; however, more older versus younger patients had cardiovascular disease (23.1% vs. 4.6%). Incidence proportions of AEs were comparable between groups, but IRs (95% confidence intervals) in older versus younger patients were numerically higher for herpes zoster (9.81 [5.72–15.71] vs. 5.44 [4.28–6.82]), and higher for serious infections (4.45 [1.92–8.76] vs. 1.14 [0.65–1.85]). More older versus younger patients discontinued due to AEs (28.6% vs. 17.6%); more younger versus older patients discontinued due to insufficient clinical responses (50.3% vs. 35.2%). Clinical remission rates through 60 weeks were generally similar between groups.
Conclusions
Older patients had higher IRs of herpes zoster and serious infection than younger patients, although tofacitinib effectiveness was similar between age groups. Discontinuation due to AEs was more common in older patients. Despite the smaller sample size of older versus younger patients, a focused evaluation of older patients is of benefit.
INTRODUCTION
Ulcerative colitis (UC) is a chronic inflammatory bowel disease (IBD) presenting with heterogeneous symptoms, including rectal bleeding and abdominal pain [1]. In 2019, UC had an age-standardized annual incidence rate (IR) of 12.7 per 100,000 person-years in Japan [2], which is an aging society with an increasing number of older patients with UC [3].
Tofacitinib is an oral Janus kinase inhibitor (JAKi) for the treatment of UC. Global phase 3 clinical trials have demonstrated the efficacy and safety of tofacitinib as induction (OCTAVE Induction 1&2) [4-6] and maintenance therapy (OCTAVE Sustain) [4-6] in patients with moderate to severe UC. The long-term efficacy and safety of tofacitinib have also been shown in a global, open-label, long-term extension study of up to 7.0 years (OCTAVE Open) [7,8], while the real-world effectiveness and safety of tofacitinib in Japan has been illustrated in a post-marketing surveillance (PMS) study [9].
ORAL Surveillance, a global post-authorization study in patients with rheumatoid arthritis aged ≥ 50 years who had ≥1 cardiovascular risk factor, reported an increased risk of major adverse cardiovascular events and malignancies (excluding non-melanoma skin cancer) with tofacitinib versus tumor necrosis factor inhibitors (TNFi) [10]. Post hoc analyses of ORAL Surveillance identified differentiating risk factors for tofacitinib versus TNFi, including a history of atherosclerotic cardiovascular disease (major adverse cardiovascular events only) [11] and long-time current/past smoking or age ≥ 65 years [12]. These findings should be taken into consideration for all JAKi and approved disease indications such as UC, where JAKi are administered at different doses [13], and in patients with different characteristics than in rheumatoid arthritis [14], as a precautionary approach, until further dedicated safety studies of sufficient size and duration indicate otherwise.
This post hoc analysis of the tofacitinib Japan PMS study in patients with UC sought to evaluate the safety and effectiveness of tofacitinib by age-stratified groups (≥ 65 or <65 years). These findings could potentially inform clinical decision-making and strategies for managing UC, particularly in the context of age-related differences in safety profiles and treatment responses.
METHODS
1. Study Design and Patients
Comprehensive details of the Japan PMS study design (Supplementary Material) have been outlined previously [9]. Patients were stratified into 2 groups: those aged ≥ 65 years (older patients) and those aged <65 years (younger patients). Follow-up analyses focused on patient demographics, disease characteristics, and safety and effectiveness outcomes. Data were gathered using electronic or paper-based case report forms, which were later updated through inquiries, also known as data-clarification forms, to identify and rectify any inconsistencies or errors.
2. Safety Analysis
The safety analysis set included patients who received ≥1 dose of tofacitinib. Safety evaluations included adverse events (AEs), defined as unfavorable events that occurred during the observation period or week 60, whichever came first. This applied to all AEs, except malignancies, which were documented until the conclusion of the observation period, irrespective of the commencement date of administration (i.e., potentially >60 weeks). AEs were classified by Preferred Term and System Organ Class, using Medical Dictionary for Regulatory Activities (MedDRA [Japanese]) version 24.1. AE seriousness was categorized based on the investigators’ evaluation. In the calculation of the number of AE episodes, multiple instances of the same Preferred Term event in a single patient were counted as a single patient.
The risk management plan identified several AEs as clinically significant or potential risks. These included “serious infections” (including all infections classified as serious and coded as MedDRA System Organ Class “Infections and Infestations,” except herpes zoster [HZ]), “HZ,” “neutropenia (neutrophil count decreased), lymphopenia (lymphocyte count decreased), haemoglobin decreased,” “hepatic function disorder,” “reactivation of hepatitis B virus,” “gastrointestinal perforation,” “interstitial lung disease,” “venous thromboembolism,” “malignancy,” “rhabdomyolysis and myopathy,” and “cardiovascular events” (excluding hyperlipidaemia and lipid increase; including dyslipidaemia). Cardiovascular events were defined as ≥1 of the following Standardized MedDRA Queries: “cardiac death,” “central nervous system vascular disease (narrow zone),” “congestive heart failure,” “myocardial infarction (narrow zone),” “other ischaemic heart disease (narrow zone),” “pulmonary embolism,” and “sudden cardiac death.”
3. Effectiveness Analysis
The effectiveness analysis set included patients in the safety analysis set where effectiveness was observed. The primary effectiveness assessment was clinical remission based on the partial Mayo score, a disease activity measure (range 0–9 points), comprising 3 subscores (0–3 points): stool frequency; rectal bleeding; and Physician Global Assessment (PGA) [15]. Assessments were conducted at weeks 2, 4, 8, 12, 16, 24, 32, 40, 48, and 60. Clinical remission (partial Mayo score of ≤ 2) had subscores of ≤ 1 for stool frequency, rectal bleeding, and PGA.
4. Statistical Analysis
The safety analysis was conducted by evaluating the incidence proportion of clinically significant AEs based on age (the number of patients with clinically significant AEs in each age group [ ≥65 and <65 years]/the total number of patients in each age group). IRs of clinically significant AEs per duration of exposure (IR per 100 patient-years) in each group was also calculated as 100×number of patients with AEs/total duration of exposure. IR comparisons for AEs between age groups were described as higher or lower if 95% confidence intervals (CIs) did not overlap, and numerically higher or lower if 95% CIs overlapped.
The effectiveness analysis included the proportion of patients in partial Mayo score clinical remission and mean changes from baseline in the partial Mayo score. Data were evaluated with last observation carried forward (LOCF) for imputation of missing data and, specifically, non-responder imputation (NRI) if there were no data to impute (i.e., NRILOCF).
5. Ethical Considerations
Ethics committee approval was exempt as this was a PMS study. This study was an all-case surveillance, non-interventional study based on the local regulations in Japan. Patient consent was not required.
RESULTS
1. Baseline Demographics and Clinical Characteristics
Overall, 2,043 patients from 456 sites were enrolled in the study; of these, 2,003 had their case report form fixed. Data were excluded for 13 patients from 2 hospitals who did acquire patient consent and did not permit publication of these data. In total, 1,982 and 1,969 patients were included in the safety and effectiveness analysis sets, respectively.
The safety analysis set (n=1,982) included 212 (10.7%) older and 1,770 (89.3%) younger patients (Table 1). Median age was 70.0 years for older and 39.0 years for younger patients; most patients were male (older patients: 57.1%; younger patients: 60.9%). Mean disease duration was 9.4 years for older patients and 7.7 years for younger patients. A higher proportion of older patients had complications versus younger patients, including cardiovascular disease (23.1% vs. 4.6%, respectively), diabetes mellitus (11.3% vs. 2.4%, respectively), hepatic function disorder (5.2% vs. 2.8%, respectively), and metabolic abnormalities (18.4% vs. 6.9%, respectively). Similarly, a higher proportion of older versus younger patients had a history of infections (6.1% vs. 2.7%, respectively), malignancy (8.0% vs. 1.0%, respectively), cardiovascular disease (6.6% vs. 2.0%, respectively), and hepatic function disorder (5.7% vs. 1.1%, respectively). For older and younger patients, respectively, infliximab had been used previously by 56.8% and 60.3%, adalimumab by 34.5% and 43.6%, and golimumab by 43.9% and 40.4%. Mean partial Mayo scores were 5.4 in each age group (Table 1).
Patient Demographics and Clinical Characteristics at Baseline in the PMS Study, Stratified by Age (Safety Analysis Set)
2. Safety Analysis
Overall, 661 patients reported AEs in the safety analysis set. AEs were reported in 83 (39.2%) older patients and 578 (32.7%) younger patients. Serious AEs were reported in 17 (8.0%) older patients and 86 (4.9%) younger patients (Table 2).
For the entire observation period, HZ was the most common clinically important AE for older patients, reported in 17 patients (8.02%), and the second most common in younger patients, reported in 75 patients (4.24%). The HZ IR was numerically higher in older patients (IR, 9.81; 95% CI, 5.72–15.71) versus younger patients (IR, 5.44; 95% CI, 4.28–6.82). The highest HZ IR in both age groups was observed at ≤15 weeks of treatment (older patients: IR, 13.80; 95% CI, 5.55–28.42; younger patients: IR, 6.50; 95% CI, 4.32–9.39) (Table 2).
Dyslipidaemia, the second most common clinically important AE for older patients, was reported in 10 patients (4.72%: IR, 5.73; 95% CI, 2.75–10.54), while it was the most common in younger patients, reported in 80 patients (4.52%: IR, 5.92; 95% CI, 4.69–7.37). Neutropenia/lymphopenia/hemoglobin decreased was reported in 8 older patients (3.77%: IR, 4.51; 95% CI, 1.95–8.88), and 46 younger patients (2.60%: IR, 3.32; 95% CI, 2.43–4.43) (Table 2).
Incidence of serious infections was higher in older versus younger patients (8 patients, 3.77%; IR, 4.45; 95% CI, 1.92–8.76 vs. 16 patients, 0.90%; IR, 1.14; 95% CI, 0.65–1.85, respectively). For both age groups, the highest IR was also observed at ≤ 15 weeks of treatment (older patients: IR, 9.78; 95% CI, 3.17–22.82; younger patients: IR, 2.08; 95% CI, 0.95–3.94) (Table 2).
Malignancies were reported in 3 older patients (1.42%: IR, 0.95; 95% CI, 0.20–2.78) and 14 younger patients (0.79%: IR, 0.58; 95% CI, 0.32–0.97). Venous thromboembolisms were reported in 1 older patient (0.47%: IR, 0.55; 95% CI, 0.01–3.06) and 4 younger patients (0.23%: IR, 0.28; 95% CI, 0.08–0.73). Cardiovascular events were reported in 2 younger patients (0.11%: IR, 0.14; 95% CI, 0.02–0.51); these included 1 case of subarachnoid hemorrhage and 1 case of subdural hematoma. No cardiovascular events were reported in older patients (Table 2 and Supplementary Results section).
3. Discontinuations
Overall, 1,038 out of 1,982 patients (52.4%) discontinued treatment during the observation period, including 91 out of 212 (42.9%) older patients and 947 out of 1,770 (53.5%) younger patients (Table 3). For older patients, the most common reason for discontinuation was insufficient clinical response (32 patients, 35.2%), followed by AEs (26 patients, 28.6%). In younger patients, insufficient clinical response (476 patients, 50.3%) was also the most common reason for discontinuation, followed by AEs (167 patients, 17.6%) (Table 3).
4. Effectiveness Analysis
At week 60, 69.81% (n=148) of older patients were in clinical remission compared with 57.09% (n=1,003) of younger patients (Fig. 1A). The mean change from baseline in partial Mayo score was –3.6 for older and –2.9 for younger patients at week 60 (Fig. 1B).
Effectiveness analysis. (A) Proportion of patients in clinical remission by partial Mayo score and (B) mean changes from baseline in partial Mayo score in the PMS study, by age and up to week 60 (NRI-LOCF). Effectiveness analysis set, which included 1,969 patients. CI, confidence interval; PMS, post-marketing surveillance; NRI-LOCF, non-responder imputation-last observation carried forward.
DISCUSSION
This post hoc analysis assessed the safety and effectiveness of tofacitinib in patients with UC, stratified by age, using data from a 60-week PMS study in Japan. Japan is an aging country with an increasing number of older patients with UC [3]. Due to the increased number of complications in older patients, it is crucial that a variety of treatment options are available. Tofacitinib safety and effectiveness in this study were generally similar to those observed in the overall population, regardless of age [9]. However, HZ IR was numerically higher, and the IR of serious infections was higher, in older versus younger patients, emphasizing the need for vigilant monitoring in older patients. IRs for neutropenia/lymphopenia/hemoglobin decreased were also numerically higher in older versus younger patients. Discontinuation due to AEs was more frequent in older versus younger patients, while more younger versus older patients discontinued due to insufficient clinical responses. Tofacitinib effectiveness was demonstrated across both age groups, with generally similar clinical remission rates between groups through 60 weeks.
Similar proportions of older and younger patients experienced AEs (39.2% and 32.7%, respectively). This is consistent with results from an analysis of pooled data from the global tofacitinib UC program, where similar percentages of patients aged ≥ 65 and <65 years reported treatment-emergent AEs [16], indicating that tofacitinib was generally well tolerated, regardless of age, in global clinical settings and in real-world settings in Japan.
However, some differences in tofacitinib safety were identified in this analysis. Specifically, the HZ IR was numerically higher in older versus younger patients (IR, 9.81; 95% CI, 5.72–15.71 and IR, 5.44; 95% CI, 4.28–6.82, respectively), and the IR of serious infections in older patients (IR, 4.45; 95% CI, 1.92–8.76) was higher versus younger patients (IR, 1.14; 95% CI, 0.65–1.85). These results align with a prospective study of patients with IBD in Japan, where age (≥ 50 years) was identified as a risk factor for opportunistic infections [17]. Similarly, in an analysis of data from patients with ≤ 6.8 years’ tofacitinib treatment in the global tofacitinib UC program, higher HZ IRs were observed in patients aged ≥ 65 versus <65 years, and age was found to be a significant predictor of HZ [16]. However, IRs of serious infections were similar between age groups in the tofacitinib UC program and age was not a significant predictor [16]. Separate real world studies of patients with IBD receiving TNFi therapy in France and Italy have also described associations between older age (≥65 and >65 years, respectively) and incidence of serious and/or opportunistic infections [18,19]. Collectively, these studies illustrate a higher rate of certain AEs with increasing age in patients with UC receiving advanced therapy. However, a causal relationship of advanced treatment with AEs should be interpreted with caution, given that the risk of specific safety events such as opportunistic infections and HZ increase with age in patients with IBD and in the general population in Japan, respectively [17,20]. Indeed, at study baseline of the analysis presented here, a higher proportion of older versus younger patients had a medical history of infection. Still, with proper management of serious infections and HZ, including prior vaccination, patient education, and early treatment, tofacitinib can be an effective therapy for older patients with UC.
Regardless of age, time-dependent analyses indicated that the HZ IR was highest ≤15 weeks after tofacitinib initiation (≥ 65 years: 13.80 [95% CI, 5.55–28.42]; <65 years: 6.50 [95% CI, 4.32–9.39]). The highest IR for serious infections was also observed after ≤15 weeks of tofacitinib treatment (≥ 65 years: 9.78 [95% CI, 3.17–22.82]; <65 years: 2.08 [95% CI, 0.95–3.94]). At baseline, a similar proportion of patients in each age group had used biologic therapies before tofacitinib initiation; therefore, the results of the time-dependent analyses may have been impacted by a lasting effect of prior biologic use. In an analysis of the global tofacitinib UC clinical program, prior TNFi failure was identified as an independent risk factor for HZ in patients with UC receiving tofacitinib [21]. Similarly, higher HZ IRs were observed in patients receiving tofacitinib 10 versus 5 mg twice daily [21]; therefore, use of higher tofacitinib doses during treatment initiation could be a contributing factor to HZ incidence during this treatment phase. Nevertheless, the lasting effect of prior biologics and the impact of tofacitinib dose on the risk of HZ warrants further investigation.
A limiting factor of this PMS study was the lack of reporting of concomitant therapies, particularly corticosteroids, which may have influenced our safety analyses. A previous analysis of the global tofacitinib rheumatoid arthritis clinical program showed that HZ risk was greater in patients who received tofacitinib in combination with corticosteroids than in those who received tofacitinib monotherapy [22], a finding which may also be applicable to patients with UC receiving tofacitinib with concomitant corticosteroids.
Although insufficient clinical response was the most common reason for tofacitinib discontinuation, regardless of age, a higher proportion of older versus younger patients discontinued due to AEs, while a higher proportion of younger versus older patients discontinued due to insufficient clinical responses. This may be reflective of the fact that older patients had higher IRs of specific AEs such as HZ and serious infection, versus younger patients.
In an analysis of pooled data from the global tofacitinib UC program, tofacitinib efficacy versus placebo at week 52, assessed as clinical remission (total Mayo score ≤2, with no individual subscore >1), was demonstrated across all age groups [16]. These findings are consistent with the results reported here, where clinical remission rates (partial Mayo score ≤ 2, with subscores of ≤ 1 for stool frequency, rectal bleeding, and PGA) were generally similar between age groups through 60 weeks. However, as this PMS study did not have a comparator arm, and a low number of patients had evaluable endoscopic data to calculate the total Mayo score, comparison with the data reported for the tofacitinib UC program is limited. Additionally, in a longitudinal cohort study of patients with UC in Spain who were in clinical remission (partial Mayo score ≤ 2) regardless of treatment (5-aminosalicylates, azathioprine, biologics, or no treatment) for at least 3 months, age was not significantly associated with an increased risk of disease relapse [23]. In contrast to our results, in an analysis of a phase 3 study of the TNFi golimumab in patients with UC in Japan, a multivariate regression analysis illustrated that increasing age, particularly age >60 years, was a limiting factor in achieving mucosal healing [24]. This suggests that the effectiveness of different therapeutic classes for UC may be age dependent, a notion that warrants further investigation.
This post hoc analysis had some limitations. Firstly, this was an observational study, and tofacitinib doses were adjusted at the healthcare professional’s discretion, which limited the ability to definitively establish a correlation between tofacitinib doses and AEs. An observation period of 60 weeks may not be sufficient to fully investigate the incidence of cardiovascular events and malignancies. Safety events were also not centrally adjudicated in this PMS study. Due to the post hoc nature of the analysis, no statistical testing of the age group differences was performed, limiting the interpretation of the results. The sample size was significantly lower in the older versus younger group, affecting the generalizability of the results to older patients. The sample size difference also limited power to conduct statistical testing between groups, which may have led to potential misinterpretation in evaluating differences in safety risks and effectiveness. Concomitant corticosteroid or immunomodulator use was not reported for these patients, which could have influenced treatment safety and effectiveness; these may also be potential confounders in the statistical analyses. Lastly, missing endoscopic data prevented the calculation of the total Mayo score for all patients, which could have provided a more comprehensive assessment of disease severity at baseline and the effectiveness of tofacitinib treatment.
In conclusion, in this post hoc analysis of the tofacitinib PMS study in patients with UC in Japan, older patients had numerically higher HZ IRs and higher IRs of serious infections compared with younger patients. A higher percentage of older patients discontinued due to AEs, while more younger patients discontinued due to insufficient clinical responses. Clinical remission rates through 60 weeks were similar between age groups. These findings suggest that while tofacitinib is equally effective in treating UC across different age groups, the risks and benefits of tofacitinib treatment should be carefully evaluated in older patients. The lack of concomitant medication data limited the analysis; however, despite the smaller sample size of older versus younger patients, a focused evaluation of older patients is of benefit.
Notes
Funding Source
This study was sponsored by Pfizer. Medical writing support was funded by Pfizer.
Conflict of Interest
Matsuoka K reports research grants from AbbVie, EA Pharma, JIMRO, Kissei, Mitsubishi Tanabe, Mochida, Nippon Kayaku, and Zeria, and speaker fees from AbbVie, EA Pharma, Eli Lilly, Gilead Sciences, Janssen, JIMRO, Kissei, Kyorin, Mitsubishi Tanabe, Mochida, Nippon Kayaku, Pfizer Inc., and Zeria. Yamamoto T reports lecture fees from AbbVie, Celltrion, EA Pharma, Gilead Sciences, Janssen, JIMRO, Kissei, Kyorin, Mitsubishi Tanabe, Mochida, Nippon Kayaku, Pfizer Inc., and Takeda. Matsuura M reports lecture fees from AbbVie, Janssen, JIMRO, Kissei, Kyorin, Mochida, Nippon Kayaku, Pfizer Inc., and Takeda. Fujii T reports research grants from AbbVie, Alfresa, Boehringer Ingelheim, Bristol Myers Squibb, Celgene, EA Pharma, Eisai, Eli Lilly, Gilead Sciences, Janssen, Kissei, Mebix, Sanofi, and Takeda, and speaker fees from AbbVie, Ajinomoto Pharma, Boehringer Ingelheim, Bristol Myers Squibb, Daiichi Sankyo, EA Pharma, Eisai, Janssen, Kissei, Kyorin, Kyowa Hakko Kirin, Mitsubishi Tanabe, Mochida, Nichi-Iko, Nippon Kayaku, Pfizer Inc., Taiho, Takeda, and Zeria. Arai S is an employee and stockholder of Pfizer Japan. Endo Y is an employee and stockholder of Pfizer R&D Japan. Sato K is an employee of Pfizer R&D Japan. Yuasa H is an employee and stockholder of Pfizer R&D Japan. Mizuno Y is an employee and stockholder of Pfizer Japan. Kobayashi Y is an employee and stockholder of Pfizer Japan. Hisamatsu T reports grants from AbbVie, Boston Scientific, EA Pharma, JIMRO, Kissei, Kyorin, Mitsubishi Tanabe, Mochida, Nippon Kayaku, Pfizer Inc., Takeda, and Zeria, consulting fees from AbbVie, Bristol Myers Squibb, EA Pharma, Eli Lilly, Gilead Sciences, Janssen, Mitsubishi Tanabe, NichiIko, and Pfizer Inc., and lecture fees from AbbVie, EA Pharma, Janssen, JIMRO, Kissei, Kyorin, Mitsubishi Tanabe, Mochida, Pfizer Inc., and Takeda. Except for that, no potential conflict of interest relevant to this article was reported.
Data Availability Statement
Upon request, and subject to review, Pfizer will provide the data that support the findings of this study. Subject to certain criteria, conditions, and exceptions, Pfizer may also provide access to the related individual de-identified participant data. See https://www.pfizer.com/science/clinical-trials/trial-data-and-results for more information.
Author Contributions
Conceptualization: all authors. Data curation; Formal analysis; Investigation: all authors. Writing - original draft: all authors. Writing - review & editing: all authors. Approval of final manuscript: all authors.
Additional Contributions
Medical writing support, under the direction of the authors, was provided by Lewis C. Rodgers, PhD, CMC Connect, a division of IPG Health Medical Communications, and was funded by Pfizer, New York, NY, USA, in accordance with Good Publication Practice (GPP 2022) guidelines (Ann Intern Med. 2022;175:1298–304). Pfizer’s generative artificial intelligence (AI) assisted technology was used to assist production of the first draft of this manuscript. After using this tool/service, the authors reviewed and edited the content as needed and take full responsibility for the content of the publication.
The authors would like to thank the patients, investigators, and study teams involved in the PMS study.
Supplementary Material
Supplementary materials are available at the Intestinal Research website (https://www.irjournal.org).