Safety and Efficacy of Tofacitinib in Combination With Biologic Therapy for Refractory Crohn’s Disease

Abstract

Background

The majority of patients with Crohn’s disease (CD) will not achieve endoscopic remission on current therapy. Addition of tofacitinib to biologics may improve remission rates.

Methods

We retrospectively assessed safety and clinical and endoscopic effectiveness of off-label tofacitinib and biologics for CD.

Results

We identified 19 patients treated with tofacitinib and a biologic for refractory CD between 2017 and 2019. Tofacitinib was added for luminal disease on colonoscopy (n = 13), luminal disease on capsule (n = 3), and pyoderma gangrenosum (n = 3). The mean age was 41.2 years (28–62), mean disease was duration 16.9 years (6–36), and prior exposure to biologics was a median of 4 (1–6). Mean treatment duration was 9.6 months (SD, 3.3). Adverse events (AEs) were reported in 36.8% of patients, most commonly minor infection or CD flare, and no patients had a serious AE; 80.0% (n = 8) achieved clinical response, and 60.0% (n = 6) achieved clinical remission based on Harvey-Bradshaw Index. Endoscopic improvement occurred in 54.5% (n = 6), endoscopic remission in 18.2% (n = 2), and endoscopic healing in 18.2% (n = 2) of patients. Mean Simple Endoscopic Score in CD significantly improved from 13.6 ± 5.2 to 6.5 ± 4.0 after treatment (P < .01).

Conclusions

In patients treated with tofacitinib in combination with a biologic, no new safety signals were observed. Combination tofacitinib and a biologic was effective in achieving clinical and endoscopic improvement in some patients with severe, refractory CD, although a larger sample size is needed to further assess the efficacy and long-term safety of this treatment strategy.

Introduction

Inadequately controlled Crohn’s disease (CD) is associated with significant morbidity and risk of penetrating and stricturing disease. Biologic therapies improve outcomes, but up to one-third of patients have nonresponse to their first biologic, which predicts lower response rates to subsequent agents. Secondary loss of response to biologics is common. Currently, no single medication results in endoscopic remission in >30% of patients; and ongoing active CD increases risks of strictures, abscesses, and fistulas, which require surgical intervention and increases risk of short gut syndrome, malnutrition, and chronic symptoms.¹ Additionally, in patients with refractory CD who achieve endoscopic improvement (response) but not remission, treating providers are hesitant to change biologics due to immunogenicity risk and concern that subsequent agents may not result in a superior response.

Tofacitinib (Xeljanz, Pfizer Inc.) is a FDA-approved Janus kinase (JAK) 1 and 3 inhibitor for ulcerative colitis (UC).² Two randomized trials evaluating tofacitinib in CD did not demonstrate clinical response³; however, findings were limited by high placebo response rate, and endoscopic response was not evaluated. Tofacitinib CD trials reported statistically significant improvement in C-reactive protein (CRP) in patients treated with 10 mg of tofacitinib orally twice daily compared with placebo (P < .0001), suggesting that tofacitinib may improve objective disease activity measures. Given the potential benefit of tofacitinib in CD and significant risk of disease-related complications with refractory CD, we hypothesized that off-label, compassionate use of tofacitinib may be effective in this population. We further hypothesized that the addition of tofacitinib to a patient’s current biologic therapy when they have not achieved remission would be effective at reducing disease activity and would be safe in the short-term. We theorized that 2 agents with distinct mechanisms of action may synergistically decrease the inflammatory burden.

Pyoderma gangrenosum (PG) is a dermatologic extraintestinal manifestation of CD characterized by skin ulceration and neutrophilic infiltrate on histopathology that is treated with immunosuppressive therapy. It has been theorized that the JAK/signal transducer and activator of transcription (STAT) pathway drives inflammation in PG, suggesting that JAK inhibitors may be effective treatments. Several small studies suggest tofacitinib may be effective for PG.^4–6

Historically, combination therapies for CD have been used; for example, azathioprine in combination with infliximab is more effective than azathioprine or infliximab alone.⁷ A trial evaluating natalizumab in combination with infliximab suggested a trend towards better efficacy compared with infliximab alone.⁸ Dual biologic therapy or combination of tofacitinib with a biologic (CTB) may benefit patients with refractory CD, although current data are limited to retrospective studies.^9,10

Although tofacitinib is associated with adverse events (AEs) of interest (eg, venous thromboembolic events [VTEs], 5% annual risk of shingles), biologics have not been as strongly associated with these. Thus, we hypothesized there would not be compounded risk of these specific adverse events in patients treated with CTB. It remains unknown if CTB will increase infection or malignancy risk. Given the high risk associated with ongoing severe CD activity, we felt the risk:benefit ratio of CTB in a patient with severe refractory CD was reasonable.

Despite the available safety and efficacy data on tofacitinib, it remains unknown if CTB is a safe and effective therapy for patients with CD. We report preliminary data on the safety and efficacy of CTB for the treatment of severe, refractory CD.

Methods

Patient Population and Ethics

Institutional review board approval was obtained for this retrospective chart review. Patients were included if they received care between 2017 and 2019 at an academic center or county hospital in Washington State, were prescribed tofacitinib for CD in addition to a biologic, were 18 years or older, and had at least 1 outpatient encounter after starting tofacitinib. Administrative databases identified 11,047 patients; 19 were eligible.

Patient demographics, disease characteristics (eg, sex, age, disease duration, location, behavior, age at diagnosis, tobacco use, concomitant medications, and prior CD therapy), and medical/surgical history were collected from medical records. Harvey-Bradshaw Index (HBI), Short Inflammatory Bowel Disease Questionnaire (SIBDQ), C-reactive protein (CRP), hematocrit, and albumin were collected from the medical record prior to and after 3 to 12 months of therapy. Simple Endoscopic Activity Score in CD (SESCD) was collected prior to and after 3 to 24 months of therapy. The SESCD is scored at the procedure time per standard of practice. Two patients had missing SESCD scores in the chart but also had available color photographs of the terminal ileum, ascending, transverse, descending, and sigmoid colon, and rectum; the SESCD for these 2 patients was retrospectively scored by an IBD specialist. Adverse events reported during therapy until 8 weeks after discontinuation or date of last data capture were included.

Approach to Treatment

Combination of tofacitinib with a biologic was used in the context of off-label, compassionate use in a refractory population of patients with CD who had not responded to other FDA-approved therapies. Combination of tofacitinib with a biologic was used in patients with endoscopic, clinical, or biochemical response to an optimized biologic regimen—but with ongoing luminal or extraintestinal disease activity thought to increase risk of disease-related complications. Tofacitinib was added to the biologic to avoid interruption of biologic therapy, given the risk of worsened disease activity and antibody formation with biologic interruption, and to use 2 agents with distinct mechanisms of action to synergistically decrease the inflammatory burden. All patients initiated 10 mg of tofacitinib orally twice daily for at least 8 weeks and then reduced to 5 mg orally twice daily if clinical remission was achieved, per the prescribing provider.

End Points

End point definitions were predefined: active clinical disease (baseline HBI ≥5); clinical response (≥3 point decrease in HBI); clinical remission (HBI ≤4); active endoscopic disease (baseline SESCD ≥6, or ≥4 for isolated ileal disease); endoscopic improvement (≥50% decrease in SESCD); endoscopic remission (SESCD ≤3); and endoscopic healing (SESCD = 0).

Statistical Analysis

Statistical analysis (Stata Statistical Software: Release 16, StataCorp LLC, College Station, TX) included descriptive statistics (numbers, frequencies, means with standard deviations) for demographics/characteristics, paired t test for changes in means, proportions for outcome rates and adverse events, and χ ² test to assess differences between proportions. When multiple values were available, the lowest score was used. Scores were excluded from analysis if they were missing, if surgical anatomy invalidated the score (HBI), or if the patient did not meet the predefined active clinical or endoscopic disease score at baseline. Surgically absent bowel was scored as zero.

Results

Baseline Characteristics and Treatment Duration

Nineteen patients with CD were included, with a mean disease duration of 16.9 years (range, 6–36), 47.3% had penetrating or stricturing disease, 41.1% had history of perianal fistula, 57.9% had history of gastrointestinal surgery, and 94.7% had prior failure of ≥2 biologics. All patients had prior failure of ≥1 tumor necrosis factor antagonist therapy (Table 1).

Table 1.

Patient Demographics & Disease Characteristics

Characteristics	Tofacitinib + biologic n = 19
Female sex, n (%)	10 (52.6)
Mean age; years (range)	41.2 (28–62)
Mean disease duration; years (range)	16.9 (6–36)
Mean treatment duration; months (SD)	9.6 (3.3)
Disease Location, n (%)
Ileal	1 (5.3)
Colonic	6 (31.6)
Ileocolonic	12 (63.2)
Disease behavior; n (%)
Nonstricturing, nonpenetrating	10 (52.6)
Stricturing	4 (21.1)
Penetrating	1 (5.3)
Stricturing and penetrating	4 (21.1)
Perianal fistulizing disease
Age at diagnosis; n (%)	8 (42.1)
Less than 16 years old	5 (26.3)
17–40 years old	13 (68.4)
Over 40 years old	1 (5.3)
History of GI surgery
Prior biologic treatment; n (%)	11 (57.9)
Biologic naive	0
1 biologic	1 (5.3)
2 biologics	2 (10.5)
3 biologics	6 (31.6)
4 biologics	7 (36.8)
5 biologics	2 (10.5)
6 biologics	1 (5.3)
Prior failure of tumor necrosis factor antagonist; n (%)	19 (100)
Tobacco use; n (%)
Never	14 (73.7)
Former	4 (21.1)
Current	0 (0)
Unknown	1 (5.3)
Concomitant corticosteroids at baseline; n (%)	5 (26.3)
Concomitant Biologic; n (%)
Ustekinumab	11 (57.9)
Vedolizumab	7 (36.8)
Certolizumab pegol	1 (5.3)
Received shingles vaccine	6 (31.6)

The indication for CTB was luminal disease on colonoscopy (n = 13), luminal disease on capsule endoscopy (n = 3), and pyoderma gangrenosum (n = 3). Mean treatment duration was 9.6 months (SD, 3.3). The concomitant biologic therapy was ustekinumab (n = 11), vedolizumab (n = 7), and certolizumab pegol (n = 1). At the end of data collection, 26.3% (n = 5) of patients discontinued tofacitinib because of insurance issues (n = 2), nonresponse (n = 1), inadequate response (n = 1), and simply to evaluate if response could be maintained on biologic alone (n = 1).

Safety

Among patients on CTB, 36.8% had ≥1 AE, most commonly minor infection and worsened CD symptoms. We observed 1 case of basal cell carcinoma. No severe adverse events occurred. We observed no cases of shingles, cardiovascular, or venous thromboembolic events (VTEs; supplemental material, Table 2). Three patients underwent elective surgery during the study period. One patient had ileitis, duodenitis, enterocutaneous fistula, and ileal stricture at baseline. This patient achieved remission with CTB, then underwent laparotomy with enterolysis, terminal ileal resection, and fistula repair. A second patient had severe colitis and known colonic stricture at baseline. They improved with CTB, were felt to be medically optimized, and then underwent laparoscopic-assisted segmental colectomy. A third patient with longstanding severe perianal fistulizing disease was started on CTB due to refractory PG. With treatment, PG resolved. Subsequently the patient developed acute on chronic worsening of perianal fistulizing disease and underwent planned unroofing skin-to-skin tracts with chronic granulation tissue right groin and perineal wound debridement.

Effectiveness

After treatment with CTB, the mean HBI decreased from 9.6 ± 4.5 to 6.3 ± 5.5 (P = .12), 80% (n = 8) achieved clinical response, and 60.0% (n = 6) clinical remission. Mean SESCD significantly decreased from 13.6 ± 5.2 to 6.5 ± 4.0 (P < .01), 54.5% (n = 6) achieved endoscopic improvement, 18.2% (n = 2) achieved endoscopic remission, and 18.2% (n = 2) achieved endoscopic healing. Mean CRP decreased from 42.0 ± 14.0 mg/L to 5.1 ± 5.5 (reference range [RR], 0–10 mg/L; P = .03) and 75% (n = 3) normalized CRP. Mean albumin increased from 3.2 ± 0.2 to 3.9 ± 0.8 (RR, 3.5–5.2, not significant) and 66.7% (n = 2) normalized albumin. All patients (n = 3) normalized hematocrit (supplemental information).

Pyoderma Grangrenosum

Tofacitinib was added for 3 patients in endoscopic remission on a biologic but with ongoing PG. A fourth patient initiated CTB for luminal disease but also had PG. All 4 patients had resolution of PG after CTB per visual assessment (Figure 1).

Figure 1.

Pyoderma gangrenosum response in patient 1 prior to (A) and 1 month after (B) adding tofacitinib to biologic therapy, and in patient 2 prior to (C), 2 months after (D), and 1 year after (E) adding tofacitinib to biologic therapy.

Discussion

Patients with refractory CD are at high risk for strictures, fistulas, and abscesses that require surgery and can decrease quality of life. Novel therapies or combinations of therapies may improve outcomes in patients with refractory disease who do not achieve remission with standard FDA-approved medications. Our study reports that CTB was well-tolerated and resulted in clinical and endoscopic improvement in some patients with refractory CD. These results build the growing body of literature demonstrating the potential role of CTB in CD^10,11 and are important for treating providers, as the majority of patients fail to achieve endoscopic remission on current therapy.

Our study reports the short-term safety of CTB in CD is not different from the risk profile previously reported in trials of monotherapy tofacitinib in CD or UC.^2,3 Although adverse events were reported, there were no new, previously undescribed adverse events. There was not a higher rate of adverse events compared with previous safety data on tofacitinib or biologic therapy. Safety concerns attributed to tofacitinib include increased risk of VTE, malignancy, and shingles.^2,3,12,13 We observed no VTE or cases of shingles during our study—though 31.6% received shingles vaccination prior to initiating CTB. One patient had basal cell carcinoma of the skin, but this is a known risk associated with tofacitinib and biologic therapy when used as monotherapy. The majority of patients receiving CTB were treated with either vedolizumab or ustekinumab biologics with lower risk of infection compared with tumor necrosis factor antagonists. The low known adverse event rate with vedolizumab or ustekinumab may explain why we observed no significant increase in adverse events that one might expect by combining medications. Our study population had long duration of refractory disease, with a mean disease duration of 16.9 years and >94% having previously failed ≥2 biologics. We do not advocate this regimen for first-line treatment among patients who are naïve to advanced therapy or with additional available options; refractory disease was a primary factor in recommending off-label tofacitinib. Even in the setting of refractory disease, 80.0% of patients in our study achieved clinical response. This clinical response rate is similar to the tofacitinib induction trials (69.8%)³ and higher than a recently published retrospective study (46.6%) that also included a refractory population.¹¹ In our study, 54.5% of patients achieved endoscopic improvement on tofacitinib. These preliminary results suggest that CTB may be effective at reducing clinical and endoscopic disease activity in some patients not in remission that are currently on biologic therapy. In addition, we report on 3 patients with PG as the primary indication for initiating CTB. Our cases are consistent with and support other literature that the JAK/STAT pathway influences PG and tofacitinib may be an effective therapy.

This study was limited by retrospective design, inherent bias in clinical decision-making, small sample size, lack of comparator, and difficulty in blinding data interpretation. Strengths include assessment of patient-reported symptoms and objective evaluation of CRP and endoscopic response to therapy. Further large randomized-controlled blinded trials are indicated to support the efficacy and safety of tofacitinib and other JAK inhibitors in CD in combination with biologics. One consideration for patients treated with CTB is whether both agents are needed long-term or whether disease control could be maintained on 1 agent after achieving remission. Another limitation to our study is the duration of follow-up. Longer duration of follow-up is needed to detect any other AEs of interest.

Finally, we acknowledge that utilization of CTB clinically is limited by lack of FDA approval of tofacitinib for CD. However, there are ongoing clinical trials investigating other JAK inhibitors for CD. If these do receive FDA approval for CD, this would make combining biologics with JAK inhibitors more feasible in clinical practice.

Conclusions

Our small retrospective study reports no new safety signals with CTB. Additionally, CTB appears effective in achieving clinical and endoscopic improvement in some patients with severe refractory CD with response to a biologic but not remission. Further research is needed to determine if the safety and efficacy profile in this population is generalizable to other populations. The long-term safety also remains unknown and warrants further research. In patients who achieved remission on CTB, further studies are needed to understand the implications of discontinuing either biologic or tofacitinib to determine if a single therapy could maintain remission, given the inherent risks of each therapy and the unknown risk of combining therapies in the long term. Our data suggest that CTB may be a well-tolerated and viable option to improve outcomes in patients who are refractory to standard therapy and who are at risk of disease complications and further surgery. Larger studies of longer duration are necessary to understand the true benefits and risks of these treatment regimens in CD.

Author Contributions

All authors approved the final version of the article, including the authorship list.

Funding

This study did not receive specific funding.

Conflicts of Interest

S.L. receives grant and research support from the following: AbbVie Pharmaceuticals, UCB Pharma, Janssen Pharmaceuticals, Inc., Salix Pharmaceuticals, Takeda Pharmaceuticals, Inc., Celgene Pharmaceuticals, Inc., Pfizer Pharmaceuticals, Inc., Atlantic Pharmaceuticals, Ltd., Gilead Sciences, Inc., Tetherex Pharmaceuticals, Arena Pharmaceuticals, Shield Therapeutics PLC; S.L. is a consultant for UCB Pharma, Mesoblast, Cornerstones, Janssen Pharmaceuticals, Inc., Takeda Pharmaceuticals, Inc., Arena Pharmaceuticals, Eli Lilly and Company, Celgene Pharmaceuticals, Inc., Celltrion Healthcare Co, Ltd, Pfizer Pharmaceuticals, Salix Pharmaceuticals. K.C.S. has been a consultant for Pfizer Pharmaceuticals and Bristol Myers Squibb. K.J.K. is supported, in part, by the National Institutes of Health, National Institute of Diabetes and Digestive and Kidney Diseases Program, at the University of Washington (Grant Nr. T32DK007742). For the remaining authors, no conflicts of interest were declared.