key: cord-0322206-pz70xhfb
authors: Alayo, Q. A.; Khatiwada, A.; Patel, A.; Zulfiqar, M.; Gremida, A.; Gutierrez, A.; Rood, R. P.; Ciorba, M.; Christophi, G.; Deepak, P.
title: Effectiveness and Safety of Combining Tofacitinib with a Biologic in Patients with Refractory Inflammatory Bowel Diseases
date: 2020-10-21
journal: nan
DOI: 10.1101/2020.10.18.20214841
sha: 8d14ac92a6bf9e0e3ec32522239759a791a5bb79
doc_id: 322206
cord_uid: pz70xhfb

Background: One therapeutic option with limited data among patients with active moderate-to-severe ulcerative colitis (UC) and Crohn's disease (CD) despite biologic monotherapy is using a combination of a biologic with Tofacitinib (TBT). Our aim was to examine the effectiveness and safety of TBT in this subset of patients. Methods: Data of IBD patients at 2 referral centers on TBT were extracted. The primary outcome was clinical response (>50% reduction in symptoms) at week 8 and/or 16 determined by Physician Global Assessment. Secondary outcome was clinical remission (resolution of symptoms), corticosteroid-free clinical response and remission, normalization of CRP and endoscopic/radiographic response and remission. Adverse events (AEs) including any abnormal lipid profile or surgical complications were also assessed. Results: Thirty-five patients (25UC, 10CD) were included. Biologics combined with tofacitinib were vedolizumab (68.6%), ustekinumab (17.1%), and infliximab (14.3%) and the median follow-up duration was 4 months. A majority (57.2%) had failed at least two biologics prior to starting TBT. At weeks 8 and/or 16, 37.1% achieved clinical response with 5.7% in clinical remission. Among the 23 patients with endoscopically/radiographically active disease at baseline, 56.5% had endoscopic/radiographic response and 34.8% achieved remission. Three AEs occurred in 2 (5.7%) patients, with an IR of 20.5 (15.0-47.2)/100PYF. No VTE and herpes zoster was reported. Conclusions: TBT is effective at inducing endoscopic/radiographic response and a modest clinical response in UC and CD patients with active clinical symptoms despite prior biologic monotherapy. No new safety signals were detected beyond those reported with tofacitinib monotherapy.

CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)

One therapeutic option with limited data among patients with active moderate-to-severe ulcerative colitis (UC) and Crohn's disease (CD) despite biologic monotherapy is using a combination of a biologic with Tofacitinib (TBT). Our aim was to examine the effectiveness and safety of TBT in this subset of patients.

Data of IBD patients at 2 referral centers on TBT were extracted. The primary outcome was clinical response (>50% reduction in symptoms) at week 8 and/or 16 determined by Physician Global Assessment. Secondary outcome was clinical remission (resolution of symptoms), corticosteroid-free clinical response and remission, normalization of CRP and endoscopic/radiographic response and remission. Adverse events (AEs) including any abnormal lipid profile or surgical complications were also assessed.

Thirty-five patients (25UC, 10CD) were included. Biologics combined with tofacitinib were vedolizumab (68.6%), ustekinumab (17.1%), and infliximab (14.3%) and the median follow-up duration was 4 months. A majority (57.2%) had failed at least two biologics prior to starting TBT. At weeks 8 and/or 16, 37.1% achieved clinical response with 5.7% in clinical remission.

Among the 23 patients with endoscopically/radiographically active disease at baseline, 56.5% had endoscopic/radiographic response and 34.8% achieved remission. Three AEs occurred in 2 (5.7%) patients, with an IR of 20.5 (15.0-47.2)/100PYF. No VTE and herpes zoster was reported.

. CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted October 21, 2020 . . https://doi.org/10.1101 /2020 INTRODUCTION . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted October 21, 2020 . . https://doi.org/10.1101 /2020 Inflammatory bowel diseases (IBD), composed of Crohn's disease (CD) and ulcerative colitis (UC) , are a group of immune-mediated diseases with increasing prevalence and morbidity. 1 Over the last two decades, the discovery and development of biologics and small molecules has expanded the therapeutic options for patients with moderate to severe IBD. Despite these novel therapeutics, there remain a substantial proportion of patients who either fail to respond, lose response or develop antibodies to these therapeutics. 2 These patients often end up dependent on prolonged steroid use, having frequent disease flares and hospitalizations, undergoing multiple surgeries, and having a low quality of life. 3 There is therefore a need for alternative therapeutic approach to managing these patients.

Based on the prior studies demonstarting that a combination of a biologic with immunomodulator is more effective than immunomodulator alone, there has been an increasing attempt to explore the therapeutic potential of combining mechanistically different biologics and/or small molecules in refractory IBD patiennts. 4 However while the efficacy and safety of each individual biologics in IBD is well established, the efficacy and safety of a combinatorial therapy of two biologics or a small molecule and a biologic is yet to be shown in randomized control trials. Recently, there have been an increasing number of retrospective case series demonstrating that a dual biologic therapy (DBT) combining mechanistically different biologics may be a potential therapeutic option for refractory IBD patients, 5-8 there is however paucity of data on combination of biologic with a small molecule.

One small molecule that has gained increased interest as a therapeutic option for severe inflammatory diseases including IBD is tofacitinib. Tofacitinib is an oral, small-molecule Janus . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)

The copyright holder for this preprint this version posted October 21, 2020. . Kinase (JAK) inhibitor, preferentially inhibits JAK1 and JAK3 which regulate signaling for multiple immune-relevant mediators implicated in the pathogenesis of inflammatory bowel diseases. 9, 10 Tofacitinib has been approved for the treatment of moderate to severe UC based on safety and efficacy data from multiple clinical trials. 11, 12 Although tofacitinib has not been approved for CD due to lack of demonstrable clinical efficacy in two phase II clinical trials, 13, 14 a post-hoc analysis of these studies has shown some clinical benefit and biologic activity of tofacitinib in moderate to severe CD without new safety signal. 15 Compared to tofacitinib, biologics such as the tumour necrosis factor-α (TNF) antagonists (infliximab, adalimumab and certolizumab pegol), antibodies inhibiting interleukins (IL)-12 and -23 (ustekinumab) and α4 β 7 integrins on gut-seeking leukocytes (vedolizumab) have mechanistically different targets in the IBD inflammatory pathway. Therefore combination of these biologics with tofacitinib may provide a synergistic effect on efficacy.

Here, we present the effectiveness and safety of combining tofacitinib, an oral, small-molecule Janus Kinase (JAK) inhibitor, with various biologics in patients with refractory IBD.

We conducted a retrospective review of all IBD patients with Tofacitinib added to other Biologic Therapy (TBT) at 2 IBD referal centers in the US (Washington University in St. Louis School of Medicine, St. Louis, MO, and Brooke Army Medical Center, Fort Sam Houston, TX).

. CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)

The copyright holder for this preprint this version posted October 21, 2020 . . https://doi.org/10.1101 /2020 Patients were included if they were prescribed and used tofacitinib concomitantly with any other biologic for the treatment of UC or CD, and had at least one follow up visit. All patients had failed at least one biologic and were started on TBT because of active bowel-related clinical symptoms that had not responded to prior conventional therapy. Patients who were started on TBT primarily for an active extraintestinal manifestation were excluded. Data were collected retrospectively at each site between December 2018 and May 2020.

Data collection was performed using a standardized data collection form on REDCap (Research Electronic Data Capture, version 7.3.5: REDCap Consortium, Vanderbilt University, Nashville, TN, U.S.A.), using pre-specified definitions and criteria for coding. Baseline demographic, clinical, endoscopic, radiographic and laboratory data were collected prior to initiation of TBT. Outcomes . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted October 21, 2020. . https://doi.org/10. 1101 /2020 The primary effectiveness outcome was clinical response (>50% reduction in symptoms or remission) at week 8 and/or 16. The secondary effectiveness outcomes icluded corticosteroidfree clinical response, clinical remission (resolution of symptoms), corticosteroid-free clinical remission, endoscopic/radiographic response or remission, and normalization of CRP.

Endoscopic/radiographic response or remission were determined using the aforementioned scores as previously described. [16] [17] [18] Endoscopic/radiographic response was defined as a ≥1grade reduction in Mayo endoscopic subscore (UC), or > 50% reduction in SES-CD or ≥ 50% reduction in the Global MaRIAs score. We defined endoscopic/radiographic remission as follow up Mayo subscore of 0 or 1 (UC), Simplified SES-CD of 0-2 (CD) or Global MaRIAs score of < 6 (CD). Adverse events (AEs) were defined as serious if life-threatening or resulting in a hospitalization, disability or discontinuation of therapy. Abnormal lipid profile was defined as the presence of any one of the following: total cholesterol ≥200, low density lipoprotein (LDL) ≥130, high density lipoprotein (HDL) <40 or Triglycerides ≥150. 19 Surgical complications of abdominal surgery were graded using the Clavien-Dindo Classification and Comprehensive Complication Index. 20

We presented descriptive statistics as medians with interquartile range (IQR) or range for continuous variables and frequencies and percentages for categorical variables. Non-parametric continuous and categorical variables were compared using Wilcoxon rank-sum and Fisher's exact tests respectively. Incidence rates (IRs) were calculated based on the unique number of patients with events per 100 patient-years of exposure. Exact Poisson CIs (adjusted for patientyears) are provided. All data were analyzed based on observed values, with no imputation for . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted October 21, 2020. . https://doi.org/10.1101/2020.10.18.20214841 doi: medRxiv preprint missing values. Statistical tests were two-sided and p-value of <0.05 was considered statistically significant. Analyses and graphically representations were done using Stata version 16.0 (StataCorp, College Station, TX) and GraphPad Prism version 8.3.0 (GraphPad Software; La Jolla, California, USA) respectively.

The study was approved by Institutional Review Boards (IRB) of both institutions and was conducted in accordance with the Declaration of Helsinki. This study is reported according to the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement for cohort studies. 21

Thirty-five IBD patients (25UC, 10CD) were included in this study with a median follow-up of 4 months (IQR 2.6-5.9). The median age at initiation of TBT was 32 years (IQR 26 -40). Majority were female (n=18, 51.4%), non-Hispanic white (23, 65.7%), and never smokers (26, 76.5%) and the median BMI of the cohort at baseline was 24.9 (22 -29) kg/m 2 . Disease extent in the UC cohort was most frequently left-sided or distal (17/25, 69%), and disease location among the CD cohort was mostly ileocolonic (8/10, 80%) with 50% (5/10) having a non-stricturing/nonpenetrating Montreal phenotype. More than half (20, 57.2%) had failed 2 or more biologics prior to starting TBT, and all patients received an induction dose of 10mg BID tofacitinib. Of the 30 patients who went on maintenance tofacitinib therapy, most (20, 66.7%) were on 5mg BID or 11mg XR daily dosing. The biologics combined with tofacitinib were vedolizumab (VDZ-TBT, . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted October 21, 2020. . https://doi.org/10.1101/2020.10.18.20214841 doi: medRxiv preprint n=24, 68.6%), infliximab (IFX-TBT, 6, 17.1%) and ustekinumab (UST-TBT, 5, 14.3%). A minority of patients (2, 5.7%) used concurrent immunomodulator at the time of initiation of TBT while 46.1% (16) were on concurrent corticosteroids (Table 1) 

Physician global assessment data was available for all patients at 8/16 weeks from initiation of TBT with 13 patients (37.1%) achieving clinical response and 2 patients (5.7%) in clinical remission. At weeks 8/16, corticosteroid-free clinical response was achieved in 25.7% (9) and corticosteroid-free clinical remission was seen in 5.7% (2) others, p=0.01) ( Table 2) . TBT was discontinued in 20 patients (57.5%), most commonly due to no response (15, 75%). Other reasons TBT was discontinued were loss of response (1, 5%), deescalation to monotherapy following clinical response on TBT (1, 5%), developing antibodies to . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted October 21, 2020. . https://doi.org/10.1101/2020.10.18.20214841 doi: medRxiv preprint the biologic (1, 5%), patient's concern about susceptibility to COVID-19 with TBT (1, 5%), and loss to follow up ater 3 months of combinations therapy (1, 5%). The median persistence on tofacitinib was about 4.4 months (18.7 weeks) (Figure 1d ).

A total of 32 patients had a pre-TBT endoscopic/radiographic assessments ( Table 1) . Of these, 28 patients had pre-and on-TBT Mayo endoscopic subscore (21 UC patients), simple endoscopic score for CD (3 CD patients) and simplified magnetic resonance index of activity score (4 CD patients) . Of the 23 patients who had endoscopically or radiographically active disease at baseline, 13 (56.5%) had an endoscopic/radiographic response while on TBT (Figure 1a) , with 8 (34.8%) patients in endoscopic/radiographic remission. Endoscopic response by IBD type and combination biologic type is shown in figures 1a and 1b, limited to those who had endoscopically or radiographically active disease at baseline. Similar to clinical response, the number of prior failed biologics was not predictive of endoscopic/radiographic response. Not using concurrent systemic steroid at the start of TBT (61.5% vs. 10%, p=0.02) and a higher albumin level at baseline were predictive of an endoscopic/radiographic response (4.1g/dl vs.

3.5g/dl, p=0.001). Other baseline characteristics examined were not predictive of endoscopic/radiographic response (Supplementary table S2). Figure 3 shows representative pre-and on-TBT MRE and endoscopic images of selected responders.

. CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted October 21, 2020. . https://doi.org/10.1101/2020.10.18.20214841 doi: medRxiv preprint C-Reactive protein data at baseline and at week 8/16 were available for 20 patients in our cohort.

Eleven (55%) of these patients had abnormal CRP at baseline. Normalization of CRP was achieved in 27.3% (n=8) of these patients at week 8/16. Biochemical response based on IBD type and combination biologic type is shown in figures 1a and 1b, limited to those with an elevated baseline CRP. Number of failed pre-TBT biologics did not predict biochemical response in this cohort (Figure 1c ).

Three AEs occurred in 2 patients (5.7%) in our cohort with an incidence rate of 20.5 [95% Confidence interval (CI), 15.0-47.2] per 100 patient years follow-up (PYF) ( Table 3) Combination therapy was not discontinued for any of these AEs and overall, no patient discontinued TBT due to adverse events. The patient who was hospitalized for Clostridium difficile infection later discontinued TBT 2 months after the infection due to loss of response. No DVT or herpes zoster reactivation or MACE was reported.

. CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted October 21, 2020. . https://doi.org/10. 1101 /2020 Paired lipid data at baseline and at 8/16 weeks was available for 21 patients. Of these, 9 (42.9%) had abnormal LP at baseline. One of 12 patients (7.14%) with normal lipid profile at baseline developed an abnormal profile at week 8/16 while 3 out of 9 patients (33.3%) with abnormal baseline lipids reverted to a normal profile at 8/16 weeks.

Seven UC patients (20% of cohort) underwent total proctocolectomy plus IPAA and diverting ileostomy while on TBT ( Table 4 ). All surgeries were due to failure of medical therapy and 2 (29%) were urgent or emergent. All surgical patients used tofacitinib within 2 weeks of surgery.

At least one Clavien-Dindo grade complication was reported in 4 (33.3%) patients. Two patients had both a surgical site infection (SSI) and were on post-operative TPN. One patient had an SSI only and the last one required post-operative TPN. None had a Grade 3 complication. No patient had readmission or re-operation within 30 days of surgery. In addition to this, two other patients had a IBD-related hospitalization while on TBT. One was hospitalized for a Crohn's disease flare and the other was hospitalized for clostridium difficile infection.

Here, we report the effectiveness and safety of combining tofacitinib, a small molecule JAK inhibitor with biologics (TBT) in IBD patients who were refractory to prior biologic therapy. We showed that in a significantly refractory subset of IBD patients with a median disease duration of 9 years and more than half exposed to 2 or more prior biologics, 37.1% achieved clinical response on TBT at week 8/16, with 5.7% in corticosteroid-free clinical remission. We also showed among those with active disease on pre-TBT endoscopy/radiographic assessment, 56.5% . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted October 21, 2020. The rate of clinical response (28.0%) and clinical remission (4.0%) following induction in the 25 UC patients in our cohort were substantially lower than the corresponding rates reported in the prior induction clinal trials with 10mg BID of tofacitinib monotherapy in UC (average of OCTAVE induction 1 and 2: 57.5% clinical response,17.6% clinical remission). Similarly, the corresponding rates in the CD subgroup in our study (60% response and 10% remission) were . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted October 21, 2020. . Additionally, the poor correlation between clinical disease activity and endoscopic disease activity scores have been previously reported in the post-hoc analyses of the SONIC trial among patients with CD. 27 Our study also showed that patients who were off systemic corticosteroids or . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted October 21, 2020. . https: //doi.org/10.1101 //doi.org/10. /2020 those with higher albumin levels at baseline were more likely to have an endoscopic/radiographic response. These findings may suggest a lower disease activity among TBT responders compared to non-responders. Further prospective studies are needed to assess these and other baseline factors as predictive of a response to TBT.

A pooled analysis of seven DBT studies showed a higher proportion of AEs (38.9%) compared to our study (5.7%). 28 Two other recent cases series of IBD patients on DBT by Yang et al. and

Kwapicz et al. also reported higher rates of AEs (13% and 26.7% respectively) in their cohorts compared to our report, suggesting that TBT may be safer compared to DBT. Surprisingly, the proportion of patients with AE in our cohort was significantly lower than what was reported in prior clinical trials with tofacitinib monotherapy in UC (induction-54.9%, maintenance-79.6%) and CD (induction-54.5%, maintenance-79.6%). 11, 13 Furthermore, the incidence rate of AE (20.7/100 PYF) reported here is lower than what we reported in our recently published realworld study of tofacitinib monotherapy in UC (IR: 27.2 (95% CI, 24.4-30.7) /100 PYF. 29 Similarly, the proportion of patients with SAE in our TBT cohort is also lower compared to these previous tofacitinib monotherapy studies. This unexpected finding is likely due to differences in the definition and monitoring of AE between clinical trials and real-world data in our study as well as the relatively shorter duration of follow-up in the current study. However, head-to-head comparisons of TBT versus DBT in longer RCTs are necessary to fully assess the safety of TBT in comparison to DBT.

Tofacitinib has been associated with a dose-dependent risk of HZ in IBD patients and the FDA recently warned of increased risk of VTE and death when tofacitinib10mg BID is used in . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted October 21, 2020. . patients with rheumatoid arthritis (RA) and IBD. 30, 31 However, none of our patients developed either of these complications while on TBT. We also did not find any case of MACE, new signal of lipid profile abnormalities and none of the 7 patients who underwent surgeries while on TBT experienced a severe Clavien-Dindo (grade 3 or higher) complication. Although, larger prospective studies with longer follow-up duration are needed to confirm these findings, our study adds to the growing body of evidence showing that DBT or combination of small molecule and biologics are well tolerated in IBD patients refractory to conventional biologic or small molecule monotherapy.

The strength of our study is that it is the largest cohort study to date from 2 large US IBD referral centers demonstrating the effectiveness and safety of TBT for luminal disease in a refractory IBD population. However, the retrospective study design, the assessment of clinical response using only the PGA, the possibility of missing AEs not adequately captured in the treating clinician documentation, lack of endoscopic follow-up in a small subset of patients, lack of a tofacitinib monotherapy control arm, and the short follow-up duration are notable limitations of our study. Despite these limitations, our results are similar to other reports on DBT or combination of small molecule and biologics in refractory IBD patients.

In conclusion, these results suggest that the combination of tofacitinib with a biologic agent induces a modest clinical response and significant endoscopic/radiographic response without any new safety signals in a subset of patients with IBD with active clinical symptoms despite biologic monotherapy.

. CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted October 21, 2020. . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted October 21, 2020. . https://www.fda.gov/drugs/drug-safety-and-availability/fda-approves-boxed-warning-aboutincreased-risk-blood-clots-and-death-higher-dose-arthritis-and. Accessed September 30, 2020. . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted October 21, 2020 . . https://doi.org/10.1101 /2020 . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted October 21, 2020 . . https://doi.org/10.1101 /2020 Charlson comorbidity index, median (IQR) 0 (0-0)

Ulcerative colitis 25 ( 

. CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted October 21, 2020. . . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted October 21, 2020. . . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted October 21, 2020. . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted October 21, 2020. . https://doi.org/10.1101/2020.10.18.20214841 doi: medRxiv preprint . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)

The copyright holder for this preprint this version posted October 21, 2020. CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)

The copyright holder for this preprint this version posted October 21, 2020. . https://doi.org/10.1101/2020.10.18.20214841 doi: medRxiv preprint . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)

The copyright holder for this preprint this version posted October 21, 2020. . https://doi.org/10.1101/2020.10.18.20214841 doi: medRxiv preprint CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)

The copyright holder for this preprint this version posted October 21, 2020. . https://doi.org/10. 1101 /2020 

Concurrent steroid use at start of tofacitinib, n (%)

Baseline hemoglobin (g/dL) median (IQR

Baseline albumin (g/dL), median (IQR)

Baseline CRP (mg/dL), median (IQR)