Evaluation of Tofacitinib in Primary Sclerosing Cholangitis and Associated Colitis: A Multicenter, Retrospective Study

Abstract

Primary sclerosing cholangitis (PSC) is frequently associated with inflammatory bowel disease (IBD).1Karlsen T.H. et al.J Hepatol. 2017; 67: 1298-1323Abstract Full Text Full Text PDF PubMed Scopus (330) Google Scholar The type of IBD (ulcerative colitis or Crohn’s disease) was shown to associate with PSC outcome, but a clear understanding of the pathogenetic link between colitis activity and PSC is lacking.2Palmela C. et al.Gut Liver. 2018; 12: 17-29Crossref PubMed Scopus (0) Google Scholar, 3Weismüller T.J. et al.Gastroenterology. 2017; 152: 1975-1984.e8Abstract Full Text Full Text PDF PubMed Scopus (0) Google Scholar, 4Ravikumar R. et al.J Hepatol. 2015; 63: 1139-1146Abstract Full Text Full Text PDF PubMed Scopus (0) Google Scholar Tofacitinib, mainly inhibiting the Janus-Kinase (JAK) 1/3 pathway, has been approved for the treatment of ulcerative colitis.5Sandborn W.J. et al.N Engl J Med. 2012; 367: 616-624Crossref PubMed Scopus (586) Google Scholar The aim of this study was to assess the safety and efficacy of tofacitinib in patients with PSC-IBD, including liver and bowel disease outcomes. We obtained retrospective data on non-transplant patients with large-duct PSC6European Association for the Study of the Liver.J Hepatol. 2022; 77: 761-806PubMed Scopus (0) Google Scholar who had received tofacitinib for at least 4 weeks, at the time of tofacitinib start, after approximately 3 and 12 months follow-up and at last follow-up available. In total, data was collected on 43 patients from 13 centers from Europe, North America, and the Middle East. One patient was excluded because tofacitinib was not given for PSC-IBD. Forty-two patients (29 of 42 male) with a median age at diagnosis of 28 years (interquartile range [IQR], 24.25 years) were eligible for inclusion. Sixty-three percent of patients were treated with ursodeoxycholic acid, and 98% were biologic experienced, thereby reflecting a difficult to treat population: Six of 42 patients received 1 biologic prior to tofacitinib; 35 of 42 received at least 2 biologics prior to tofacitinib. The median overall treatment period with tofacitinib was 13.1 months (IQR, 16.7 months). Those who remained on treatment with tofacitinib until last follow-up (n = 22) had a median treatment period of 20.4 months (IQR, 14.4 months). In 48% (20 of 42), treatment with tofacitinib was discontinued, mostly due to lack of efficacy (se Supplementary Table 1 for clinical characteristics of study cohort). Tofacitinib appeared to be well-tolerated and safe (Supplementary Table 2). In only 2 cases, treatment was stopped due to adverse events other than worsening of colitis activity. In 17% (7 of 41), adverse events were noted during ongoing treatment with tofacitinib that did not lead to treatment discontinuation. However, none of the known serious adverse events under treatment with tofacitinib such as thrombosis or thromboembolism, reactivation of viral disease, or development of lymphoproliferative disease were observed. Only 1 patient was reported to have an episode of bacterial cholangitis under treatment, and 1 had severe salmonella-associated colitis. Two patients showed histological low-grade colon dysplasia at baseline, of which 1 progressed to metastatic colorectal carcinoma at follow-up; the other remained at low grade dysplasia. Of those with follow-up colonoscopy (n = 37; median time between colonoscopies, 12 months [IQR, 11.7 months]), 3 had new evidence of low-grade dysplasia and 1 of high-grade dysplasia. Overall, in 87% of patients (32 of 37), there were no signs of new dysplasia. Two patients died due to their malignancy (gallbladder cancer, colorectal carcinoma) during follow-up. Furthermore, we assessed outcome on liver and bowel disease. Colitis activity significantly improved in the majority of patients (58%) assessed by the full Mayo Score (P = .003; n = 27 with data available) as well as by mucosal appearance using the Mayo Endoscopic Subscore (P = .024; n = 33 with data available) (Figure 1). At baseline, only 1 patient (n = 33 with data available) and, at follow-up, 10 patients (n = 30 with data available) were in remission with a Mayo Score ≤ 1. A significant decrease in IBD activity was only seen in those who received ongoing treatment with tofacitinib (Mayo Endoscopic Subscore, P = .047; n = 17; full Mayo Score, P = .004; n = 15). Overall, fecal calprotectin dropped from a median of 818 ug/g (IQR, 663.9 ug/g) at baseline to 359.7 ug/g (IQR, 567.5 ug/g) at follow up (P = .041; n = 14; median time between measurements, 6.3 months [IQR, 7.5 months]).Whereas 21 patients underwent biliary interventions before start of tofacitinib, only 2 patients had their first endoscopic biliary intervention during treatment with tofacitinib after a median observational follow-up of time of 11.1 months (IQR, 13.3 months); both discontinued treatment due to lack of efficacy. Liver stiffness remained stable in those with measurements before and under tofacitinib treatment (median, 5.0 kPA [IQR, 5.4 kPA] vs 6.1 kPA [IQR, 5.6 kPa]; P = .278; n = 13; median follow-up time, 16.3 months [IQR, 12.6 months]). Only 1 of 38 patients (3%) developed new signs of cirrhosis at follow-up, and 1 patient had newly diagnosed esophageal varices. Regarding the effect of tofacitinib on serum alkaline phosphatase (ALP), a longer treatment time with tofacitinib was associated with lower ALP ratio (baseline/last follow-up under treatment with tofacitinib) (r = −0.423; P = .011; n = 35). Mixed linear modeling with random intercept for patients allowing autocorrection between time points showed a significant difference, adjusted for baseline ALP and time between follow-ups, between those who discontinued and those who continued treatment (mean difference, 106 U/L; 95% confidence interval [CI], 14−198 U/L; P = .025). Thus, with a baseline ALP of 276 U/L, an estimated marginal mean ALP difference (baseline − follow-up) is modeled at 9.6 months: treatment continued, −68 U/L (95% CI, −132 to −4 U/L) vs discontinued, +39 U/L (95% CI, −27 to 104 U/L). Of note, IBD response to tofacitinib significantly correlated with ALP changes, with those showing an improvement in their IBD activity having a decline of ALP serum levels (r = −0.445; P = .012; n = 31) (Figure 1). We here present the first study on the safety and efficacy of tofacitinib in patients with PSC-IBD.7Lamb C.A. et al.Gut. 2019; 68: S1-S106Crossref PubMed Scopus (0) Google Scholar A reassuring conclusion of this study is that treatment with tofacitinib was well tolerated in patients with PSC and associated colitis. Specifically, there was no liver-related safety signal, taking into account the limitations of the retrospective nature of this case series. Tofacitinib led to an improvement in colitis activity in the majority of patients in this difficult-to-treat patient cohort. Of note, our study showed that ALP serum levels significantly declined in patients with ongoing treatment with tofacitinib. ALP is being regarded as one of the surrogate biomarkers for PSC disease activity and prognosis,8Ponsioen C.Y. et al.Hepatology. 2016; 63: 1357-1367Crossref PubMed Scopus (0) Google Scholar but it should be kept in mind that there are considerable natural variations in ALP levels in PSC and that even advanced disease can present with normal ALP levels.9Trivedi P.J. et al.Clin Gastroenterol Hepatol. 2021; 19: 1248-1257Abstract Full Text Full Text PDF PubMed Google Scholar,10Bakhshi Z. et al.J Gastroenterol. 2020; 55: 523-532Crossref PubMed Scopus (0) Google Scholar Given the potential involvement of JAK/STAT signaling in PSC pathogenesis, these findings warrant further studies investigating JAK inhibition for the treatment of PSC. Part of this work was presented at The International Liver Congress (European Association for Study of the Liver, June 2021) and at the Annual Meeting 2022 of the German Association for the Study of the Liver, January 2022. The authors thank all patients who allowed use of their clinical data. The authors thank Elaine Hussey for language editing and Susanne Lezius for statistical consultation. The authors thank those who contributed data and supported us in the editing process on behalf of the International PSC Study Group: John Eaton; Cynthia Levy; Olivier Chazouilleres; Tobias Müller; Jeremy Nayagam; Deepak Joshi; Ehud Zigmond; Oren Shibolet; Joost PH Drenth; Frank Hoentjen; Anja Geerts; Tobias J. Weismüller; and Taotao Zhou. Data on patients with large duct primary sclerosing cholangitis (PSC) and associated colitis who received tofacitinib for at least 4 weeks were collected retrospectively via the International PSC Study Group (IPSCSG) and the European Reference Network on Hepatological Diseases (ERN RARE-LIVER). Diagnosis of large duct PSC was made according to international guidelines.6European Association for the Study of the Liver.J Hepatol. 2022; 77: 761-806PubMed Scopus (0) Google Scholar Patients after liver transplantation and with an overall follow-up of less than 3 months were excluded. Pseudonymized data was analyzed centrally at the University Medical Center Hamburg-Eppendorf. The study was approved by the local ethics committee (PV4081, Ethics Committee of the Hamburg Medical Association). Data was obtained retrospectively at the time of tofacitinib start, after approximately 3 and 12-months follow-up and at last follow-up available. Prior immunomodulatory medications, as well as tofacitinib treatment schedule and changes of immunosuppressive medication after start of treatment with tofacitinib, were recorded. Clinical diagnosis of cirrhosis including imaging and liver histology was captured. Mayo Score/Disease Activity Index and results of transient elastography using FibroScan were used to stage colitis and liver disease, respectively. Alkaline phosphatase (ALP) and bilirubin levels were analyzed as surrogate biomarkers for PSC activity. New diagnosis of cirrhosis, development of varices or variceal bleeding, bacterial cholangitis, new hepatobiliary neoplasia, and need for endoscopic interventions were recorded as clinical outcome parameters. To assess the effect of tofacitinib on inflammatory bowel disease activity, endoscopic findings were registered with the Mayo Endoscopic Subscore and, if available, with fecal calprotectin levels. Where available, full Mayo Score/Disease Activity Index was provided. Response to treatment was classified as either improved (at least 1 point improvement), unchanged or worsened (at least 1 point deterioration). Wilcoxon signed-rank test was used to compare paired variables over time. Differences between independent variables were tested via the Mann-Whitney U test. The Spearman rank test was performed to test for correlation. Changes in biochemistry were measured by the difference and ratio between the value at the last available follow-up under tofacitinib treatment and the baseline value. To test for group differences in ALP changes between those that continued and those that discontinued treatment, mixed linear modeling with random intercept for patients allowing for autocorrection between time points was performed and adjusted for baseline ALP and time between follow-ups. IBM SPSS statistics version 23 was used as a statistical software, and GraphPad Prism 9 was used for visualization.Supplementary Table 1Clinical Characteristics of Study Cohort (n = 42 unless otherwise specified)CharacteristicValueMale29 (69)Age at PSC diagnosis, y28 (24.25)Previous treatment with ≥ 2 biologicals35 (83)Cirrhosis at baseline9 (21)Esophageal varices at baseline2/35 (6)Signs of colon dysplasia at baseline (n = 40) No signs of dysplasia38 (95) Low-grade dysplasia2 (5)Hepatobiliary neoplasia at baseline (n = 42)1 (2) (gallbladder cancer)Patients with prior endoscopic biliary intervention at baseline (n = 41)12 (30)Patients with ≥ 1 prior episode of cholangitis at baseline5 (20)UDCA at least 6 months prior to tofacitinib (n = 41)26 (63) Dose, mg/kg/day14.1 (6.3)Overall treatment duration with tofacitinib (n = 42), months13.1 (16.7) Duration of tofacitinib in those who discontinued treatment (n = 20), months4.9 (9.7) Duration of tofacitinib in those with ongoing treatment (n = 22), months20.4 (14.4)Note: Data are presented as median (interquartile range) or number (%).PSC, Primary sclerosing cholangitis; UDCA, ursodeoxycholic acid. Open table in a new tab Supplementary Table 2Tolerability and Safety of TofacitinibVariableValuePatients who discontinued tofacitinib (n = 42)20 (48)Reasons for tofacitinib discontinuation (n = 20) Adverse event2 (11) Lack of efficacy15 (75) Colectomy5 (25) Other3 (15)Initial dose of tofacitinib, mg/d20 (.0)Maintenance dose of tofacitinib, mg/d10 (10)Patients with adverse events during treatment (n = 41)7 (17)Patients with severe infection during treatment (n = 41)1 (2)Endoscopic signs of colon dysplasia at follow-up (n = 37) No new dysplasia32 (87) Low-grade dysplasia3 (8) High-grade dysplasia1 (3) Carcinoma1 (3)Note: Data are presented as median (interquartile range) or number (%). Open table in a new tab Note: Data are presented as median (interquartile range) or number (%). PSC, Primary sclerosing cholangitis; UDCA, ursodeoxycholic acid. Note: Data are presented as median (interquartile range) or number (%).

Topics & Concepts