Tofacitinib as Salvage Therapy for 55 Patients Hospitalised With Refractory Severe Ulcerative Colitis


Mathieu Uzzan; Clément Bresteau; David Laharie; Carmen Stefanescu; Christophe Bellanger; Franck Carbonnel; Mélanie Serrero; Stéphanie Viennot; Maria Nachury; Aurélien Amiot; Romain Altwegg; Laurence Picon; Stéphane Nahon; Lucine Vuitton; Philippe Ah Soune; Julien Kirchgesner; Laurent Peyrin-Biroulet; Yoram Bouhnik


Aliment Pharmacol Ther. 2021;54(3):312-319. 

In This Article


Baseline Characteristics

We included 55 patients from 14 GETAID centres. 31 (56.4%) retrospectively and 24 (43.6%) prospectively from June 2020. Their general baseline characteristics are displayed in Table 1. To summarise, 45.5% were female, median age was 27.6 years old (IQR [21.8–42.5]), including 11 (20%) patients older than 50 and 7 (12.7%) over 60 years. The median disease duration at inclusion was 4.4 years (IQR [2.4–7.1]).

Regarding prior drug exposure, 49 (89.1%) patients have received infliximab and 19 (34.5%) ciclosporin. Patients were previously exposed to a median of 2.5 lines of biologic treatment before tofacitinib, including 54 (98.2%) patients to at least one anti-TNFα agent, 38 (69.1%) to vedolizumab and 6 (10.9%) to ustekinumab. One patient, who was previously exposed to ciclosporin, vedolizumab and ustekinumab but not to anti-TNFα, due to active uncontrolled systemic lupus erythematosus, was included.

Regarding treatments received during hospitalization for the current flare, intravenous steroids were given to 29 (52.7%) patients. Infliximab was administered to two patients (3.6%) and intravenous ciclosporin to 8 (14.5%) prior to tofacitinib. Median length of in-hospital stay was 14 days (IQR [8.0–26.5]). Median Lichtiger at inclusion was 12 (IQR [9.5–13]) and median C-reactive protein was 17.2 mg/L (IQR [7.3–66.3]). 41 patients (74.5%) had a Lichtiger of 10 or more. Mayo endoscopic subscore was at 2 for 18 (32.7%) patients and was at 3 for 37 patients (67.2%). Deep ulcerations were seen during flexible sigmoidoscopy in 9 (16.4%) patients.

Tofacitinib was initiated after a median of 3 days (IQR [1–6]) from hospital admission. 36 (65.5%) patients were receiving steroids at tofacitinib initiation.

Colectomy-free Survival and Factors Associated With Colectomy

Within a median follow-up duration of 6.5 ((IQR) [3–12.3]) months, 15 patients (27.3%) underwent colectomy. The colectomy-free survival was estimated at 85.2% (95 CI [76.3–95.2]), 78.9% (95 CI [68.5–90.9]) and 73.6% (95 CI [61.9–87.3]) at 1, 3 and 6 months, respectively (Figure 1).

Figure 1.

Kaplan-Meier analysis showing colectomy free-survival. x-axis representing time in months. The coloured area shows 95% confident intervals

We did not find any clinical or biological factors associated with colectomy on univariate analysis (Table 2). Pancolitis (hazard ratio [HR] = 2.52, 95 CI [0.71–8.96]), Lichtiger (HR = 1.14 for an increase of 1, 95 CI [0.95–1.37), CRP level (HR = 1.01 for an increase of 1 mg/L, 95 CI [1.00–1.01]), and the presence of deep ulcerations at the flexible sigmoidoscopy (HR = 2.77, 95 CI [0.86–8.92]) tended to be associated with the risk of colectomy with a P value below 0.2. The multivariate Cox proportional hazards regression model including the latter 4 variables revealed no trend and no significant association with the risk of colectomy (Table 2).

Clinical Response and Remission Rates at Week 6 and 14

At week 6, clinical response and remission rates of all 55 patients were assessed. Patients still on tofacitinib (43 of 55 patients) were all treated with a 10 mg twice daily regimen. Of the 55 patients, 33 (60%), 25 (45.5%), and 20 (37.5%) achieved clinical response, clinical remission and clinical steroid-free remission, respectively.

Out of the 31 patients still on tofacitinib at week 14, 24 patients (77.4%) were receiving a dose of 10 mg twice daily while 7 (22.6%) of them were treated with 5 mg twice daily. At week 14, 23 (41.8%), 19 (34.5%) and 18 (32.7%) of 55 patients were in clinical response, clinical remission and in steroid-free clinical remission, respectively (Table 3 and Figure 2).

Figure 2.

Histograms representing percentages of response and remission. Clinical response: decrease of the partial clinical Mayo score of at least 3 points, with a decrease of at least 30% and a reduction of at least 1 point on the bleeding subscore with a subscore of 0 or 1, as compared with inclusion. Clinical remission: partial clinical Mayo score of less than 3, without any of the subscores with a value higher than 1

Among patients still on tofacitinib at week 6, 36 patients had data available for CRP. CRP was at a median of 4 mg/L (IQR [1.5–9.5]) versus 15 mg/L (IQR [9.5–44.8]) at baseline (P = 0.047, paired Student's t test). At week 14, 22 patients still on tofacitinib were evaluable for CRP with a significant decrease from baseline levels (3 mg/L [IQR] [2.3–5]) versus 14.5 mg/L (IQR [6.9–50.3]), P = 0.01, paired student t test).

Tofacitinib Persistence

At the end of follow-up, 28 patients (50.9%) were still treated with tofacitinib. The survival without tofacitinib discontinuation was estimated at 81.8% (95 CI [72.2–92.7), 68.1% (95 CI [56.6–81.9]) and 44.9% (95 CI [32.6–61.8]) at 1, 3, and 6 months (Figure 3).

Figure 3.

Kaplan-Meier survival estimates for discontinuation in patients receiving tofacitinib for severe flare of acute colitis. The coloured area shows 95% confident intervals

Reasons for discontinuation are shown in Table 4. Primary nonresponse within 3 months of initiation was the main cause of treatment withdrawal.


Adverse events occurred in six patients (10.3%) including four severe adverse events, leading to discontinuation in three of them. Two patients (3.7%) aged 60 and 62 years old, experienced an episode of herpes zoster. One of them had extensive cutaneous bullous lesion of the trunk that led to cessation of tofacitinib. A patient stopped the treatment at day 3 because of unusual abdominal pain, nausea and vomiting. The third withdrawal secondary to adverse events was due to recurrent viral pneumonia in a patient with medical history of chronic lung disease (bronchiectasis and prior recurrent infections). Additionally, one patient experienced alopecia and another one had an episode of C. difficile infection. Both patients continued tofacitinib. No venous thrombotic or major adverse cardiovascular events occurred in the cohort.

In the cohort, one patient died. An 81-year-old male with an history of chronic obstructive pulmonary disease and diabetes mellitus died of infectious complications 17 days after colectomy. He was treated with tofacitinib, for 7 days, which was stopped 19 days before colectomy. His death was considered not directly related with tofacitinib.