Comparative effectiveness of ustekinumab vs. vedolizumab for anti-TNF-naïve or anti-TNF-exposed Crohn's disease: a multicenter cohort study Authors: Hongsheng Yang, Zhaopeng Huang, Miao Li, Haiyan Zhang, Lingyu Fu, Xiaoling Wang, Qiaoqiao Yang, Yun He, Wenhong Wu, Taofeng Jiang, Zhimei Sun, Tao Zhang, Dongping Lai, Xixin Wu, Lishuo Shi, Qin Guo, Yanhui Wu, Jian Tang, Kang Chao, Xiang Gao Published: 2023 ABSTRACT SummaryBackgroundUstekinumab and vedolizumab are both effective for treating Crohn’s disease (CD). However, no head-to-head trials have been conducted thus far. We aimed to compare the effectiveness of ustekinumab and vedolizumab in CD patients either naïve or exposed to tumor necrosis factor-alpha inhibitors (TNFi).MethodsPatients treated with vedolizumab or ustekinumab for luminal CD were included from six centers in China from May 2020 to July 2023. Steroid-free remission, clinical remission, objective response, and remission at Weeks 26 and 52 were evaluated in a retrospective multicenter propensity score–weighted cohort.FindingsA total of 536 patients were included (386 ustekinumab, and 150 vedolizumab). After adjustment, ustekinumab showed higher rates of clinical remission (56.4% vs. 47.8%, P = 0.005), steroid-free remission (55.4% vs. 46.1%, P = 0.003), and objective response (67.8% vs. 42.7%, P < 0.001) than vedolizumab at Week 26. At Week 52, ustekinumab exhibited significantly higher rates of clinical remission (65.8% vs. 37.5%, P < 0.001), steroid-free remission (65.8% vs. 37.5%, P < 0.001), objective response (66.7% vs. 23.8%, P < 0.001), and objective remission (31.4% vs. 12.7%, P < 0.001). Subgroup analyses revealed that ustekinumab had higher rates of clinical remission, steroid-free remission, and objective response at Weeks 26 and 52, and objective remission at Week 52 in TNFi-exposed patients, while ustekinumab showed higher rates of objective response at Weeks 26 and 52 and clinical remission, steroid-free remission and objective remission at Week 52 in TNFi-naïve patients. Adverse event rates were similar between the groups (4.9% ustekinumab vs. 6.7% vedolizumab, P = 0.423).InterpretationUstekinumab showed superior clinical and objective outcomes compared to vedolizumab, with comparable safety outcomes. The therapeutic superiority was observed in both short-term and long-term phases in TNFi-exposed patients, and the long-term phase in TNFi-naïve patients.Funding10.13039/501100001809National Natural Science Foundation of China, 10.13039/501100021171Guangdong Basic and Applied Basic Research Foundation, Key Research Projects of the Sixth Affiliated Hospital, 10.13039/501100002402Sun Yat-sen University, the program of Guangdong Provincial Clinical Research Center for Digestive Diseases, and National Key Clinical Discipline. Research in contextEvidence before this studyWe searched PubMed for publications without language restriction, published before July 25, 2023, using search terms “Crohn’s disease”, “vedolizumab”, and “ustekinumab” with search terms found in abstract, title or MESH headings. There are no head-to-head trials comparing the effectiveness of ustekinumab and vedolizumab for treating Crohn’s disease. A few retrospective studies have evaluated the comparative effectiveness for patients with Crohn’s disease (CD) who were exposed to tumor necrosis factor-alpha inhibitors (TNFi), but results were controversial and constrained by limited sample sizes and incomplete evaluations. Data on TNFi-naïve patients were still missing.Added value of this studyWe carried out the largest real-world cohort study to date investigating the comparative effectiveness of these two drugs in CD patients either naïve or exposed to TNFi. Our study showed that ustekinumab was associated with superior clinical and objective effectiveness than vedolizumab, but safety outcomes were comparable. This therapeutic superiority was observed both at the short-term and long-term phase in TNFi-exposed patients, while only the long-term superiority was apparent in TNFi-naïve patients.Implications of all the available evidenceUstekinumab appeared be superior to vedolizumab for achieving long-term effectiveness in CD patients regardless of TNFi exposure. The findings may facilitate clinical decision-making regarding the management of CD. Research in context Evidence before this study We searched PubMed for publications without language restriction, published before July 25, 2023, using search terms “Crohn’s disease”, “vedolizumab”, and “ustekinumab” with search terms found in abstract, title or MESH headings. There are no head-to-head trials comparing the effectiveness of ustekinumab and vedolizumab for treating Crohn’s disease. A few retrospective studies have evaluated the comparative effectiveness for patients with Crohn’s disease (CD) who were exposed to tumor necrosis factor-alpha inhibitors (TNFi), but results were controversial and constrained by limited sample sizes and incomplete evaluations. Data on TNFi-naïve patients were still missing. Added value of this study We carried out the largest real-world cohort study to date investigating the comparative effectiveness of these two drugs in CD patients either naïve or exposed to TNFi. Our study showed that ustekinumab was associated with superior clinical and objective effectiveness than vedolizumab, but safety outcomes were comparable. This therapeutic superiority was observed both at the short-term and long-term phase in TNFi-exposed patients, while only the long-term superiority was apparent in TNFi-naïve patients. Implications of all the available evidence Ustekinumab appeared be superior to vedolizumab for achieving long-term effectiveness in CD patients regardless of TNFi exposure. The findings may facilitate clinical decision-making regarding the management of CD. Introduction Crohn’s disease (CD) is a chronic and debilitating inflammatory disease affecting the gastrointestinal tract.1 Two decades after the introduction of tumor necrosis factor inhibitors (TNFi), two novel biologics, ustekinumab (UST) and vedolizumab (VDZ), have emerged as important therapeutic options for patients with inflammatory bowel disease and offer distinct mechanisms of action.2,3 UST is a fully humanized IgG1κ monoclonal antibody, which can specifically bind to the P40 subunit of IL-12/IL-23 and block its biological activity, thus reducing the intestinal inflammation. VDZ is an anti α4β7 integrin recombinant humanized IgG monoclonal antibody which can inhibit T lymphocyte homing to the intestine, thereby regulating intestinal inflammation. Both UST and VDZ are effective in induction and maintenance therapy in patients with CD, either naïve or exposed to TNFi.4,5 However, head-to-head trials comparing these two drugs are currently lacking, posing challenges for clinicians in real-world decision-making. Several indirect comparative observational studies have been conducted to evaluate the comparative effectiveness of UST and VDZ in TNFi-exposed CD patients, albeit yielding inconsistent conclusions.6, 7, 8, 9, 10, 11, 12 The studies were further constrained by limited sample sizes and incomplete evaluations. In clinical practice, UST and VDZ are frequently used as second-line therapies following TNFi exposure in patients with CD. They are also indicated as first-line therapies in patients who respond inadequately to conventional therapies. However, there remains a paucity of real-world evidence comparing these two drugs in TNFi-naïve CD patients. In 2020, both UST and VDZ have been approved in China for the treatment of CD patients who are refractory or intolerant to TNFi or conventional therapies. VDZ was reimbursed in March 2021, and UST was reimbursed in January 2022. Building upon three years of experience with these agents, we comprehensively evaluated the effectiveness of UST and VDZ in patients with CD, employing both clinical and objective assessment methods. Additionally, we evaluated the effectiveness of VDZ vs. UST in both TNFi-naïve and TNFi-exposed CD patients within this study. Methods Study design This retrospective observational real-world cohort study involved six IBD centers in China (The Sixth Affiliated Hospital, Sun Yat-sen University, Guangzhou, Guangdong; The Second Affiliated Hospital of Guangzhou University of Chinese Medicine, Guangzhou, Guangdong; The Eighth Affiliated Hospital, Sun Yat-sen University, Shenzhen, Guangdong; Liuzhou People's Hospital, Liuzhou, Guangxi; Meizhou People's Hospital, Meizhou, Guangdong; Ruikang Hospital Affiliated Guangxi University of TCM, Nanning, Guangxi). Patients with CD receiving at least one dose of either UST or VDZ from May 2020 to July 2023 were considered for inclusion. Data reporting was per the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) Statement.13 Inclusion and exclusion criteria The inclusion criteria were as follows: (1) patients with a confirmed diagnosis of CD according to current guidelines14,15; (2) patients who received at least one dose of either VDZ or UST for luminal disease; and (3) patients aged ≥18 years. The exclusion criteria were as follows: (1) prior exposure to VDZ and/or UST; (2) definitive or transitory stoma; (3) insufficient clinical data to evaluate effectiveness at baseline and Week 26 ± 4. Data collection Baseline and follow-up data of patients were collected from the electronic medical records. Two physicians at each center were responsible for data retrieval; one independently retrieved the data using a standardized data extraction form, and one verified the data. All the investigators communicated via online conferences. Disagreements were resolved through discussion to achieve a consensus. The following data were collected: sex, age, baseline active smoking status, course of disease, symptoms, history of bowel surgery, disease location and behavior classified according to the Montreal Classification, disease activity, medication history, endoscopic manifestations, ultrasound manifestations, computed tomography (CT) or magnetic resonance (MR) manifestations, therapeutic regimens, and adverse events. Investigated drugs Both UST and VDZ were given at standard doses according to the label. UST was initiated with a weight-based intravenous infusion (260 mg ≤ 55 kg, 390 mg between 55 kg and 85 kg, 520 mg > 85 kg) in the induction period, and the maintenance dosage is a subcutaneous 90 mg dose administered 8 weeks after the initial intravenous dose, then every 8 or 12 weeks thereafter. VDZ was administered intravenously (300 mg) at Weeks 0, 2, and 6 for induction and every 8 weeks for maintenance. Dose escalation during the maintenance period, as well as concomitant steroid and immunosuppressant use, were at the discretion of the physician. Patients who had an inadequate response or loss of response to the standard dose of VDZ or UST could undergo dose escalation. Outcomes and definitions In our routine clinical practice, patients were scheduled for visits at Weeks 0, 26 and 52, during which pre-defined clinical and objective outcome measures were assessed. The Crohn’s disease activity index (CDAI) was documented, and objective methods including, endoscopy, bowel ultrasound, and CT/MR enteroclysis, were performed at Weeks 0, 26 and 52. Week 0 was defined as the day of the first infusion of the biologics. In the absence of clinical data at Week 26 or 52, the closest medical report (within 4 weeks) was used. The primary outcome was steroid-free remission at Weeks 26 and 52. Secondary outcomes were clinical remission at Weeks 26 and 52, objective response and remission at Weeks 26 and 52, and the occurrence of adverse events during follow-up. CDAI was used to evaluate the clinical effectiveness. The disease was considered mildly active if the CDAI was 150–220, moderately active if the CDAI was 221–450, and severely active if the CDAI was >450.16 Clinical remission was defined by CDAI ≤150, and steroid-free remission by CDAI ≤150 with steroids below the equivalent of prednisolone 10 mg/day.16 Objective response was defined by endoscopic or radiographic CD improvement, as assessed by ileocolonoscopy, contrast-enhanced ultrasound or CT/MR enterography. Objective remission was defined by endoscopic remission or normalization of radiographic appearance on bowel ultrasound or CT/MR enterography. These definitions were used in previous cohort studies.12,17,18 Endoscopic response was defined by the reduction of SES-CD ≥50%, and endoscopic remission was defined by SES-CD ≤2.19 Mucosal healing was defined as the absence of ulcers in the intestinal mucosa.20 The ultrasound response was defined as a reduction in bowel wall thickness (BWT) as evaluated by bowel ultrasound, and ultrasonic remission was defined as a normal BWT (≤3 mm for the small bowel and ≤4 mm for the colon) without any detectable complications.21 Radiological response was defined by an improvement in BWT, inflammatory fat, mural blood flow, and hyperenhancement evaluated by CT/MR enteroclysis, and radiological remission was defined as the normalization of inflammatory parameters on CT/MR enteroclysis.22,23 C-reactive protein (CRP) normalization was defined as a serum CRP level of <5 mg/L.24 Adverse events were recorded, including injection reactions, acute or delayed allergic reactions, acute or chronic infections, hematological adverse reactions, and tumors. Statistical analysis Patient characteristics are described as mean and standard deviation (SD) or median and interquartile range (IQR) for continuous variables and as number of cases and proportions for categorical variables. The χ2 or Fisher’s test was used to compare the proportions. Student’s t-test or its non-parametric counterparts were used to compare quantitative variables. To minimize the effect of confounding variables, propensity score with inverse probability of treatment weighting (IPTW) was used.25, 26, 27 We included the potential confounding variables for the propensity score adjustment based on their clinical relevance as shown in previous studies.6, 7, 8, 9,11 The following variables were included: age, sex, active smoking status, disease location and behavior as classified by the Montreal classification, active intestinal fistula at baseline, perianal disease at baseline, previous steroid, extensive enteral nutrition (EEN) or immunomodulator use, concomitant immunomodulators, EEN or steroid at baseline, CDAI at baseline, proportion of CDAI >150 at baseline, previous TNFi exposure, serum CRP level, and serum albumin at baseline. Firstly, the propensity score was estimated by logistic regression models as the predicted probability of initiating treatment with UST. Secondly, we calculated a stabilized weight. For the UST group, stabilized weight = 1/π; and for the VDZ group, stabilized weight = 1/(1 − π), in which π is the propensity score. These weights were further used in the analysis models. The level of balance between the VDZ and UST groups in the unweighted sample (original cohort) and in the weighted sample (pseudo-cohort obtained by weighting) for all variables included in the IPTW analysis was verified by computing the standardized mean differences (SMD), and SMD ≤0.1 has been taken to indicate a negligible difference between treatment groups. Results from primary outcomes by using IPTW analyses were replicated by using propensity score matching (PSM) as sensitivity analysis with the “MatchIt” package. In the propensity score matching step, UST-treated group and VDZ-treated group were matched in a 2:1 ratio by using a greedy nearest-neighbor matching algorithm without replacement, with a caliper of 0.15 standard deviation of the logit of the propensity score. The propensity score that predicted probability of UST use was derived from a logistic regression model that was the same as IPTW. We then computed odds ratios (OR) and their 95% confidence intervals (CI). There were only 4 missing values for baseline characteristics, including 1 for baseline CRP and 3 for baseline albumin. Due to the small number, the missing data were handled by filling in the median during the application of IPTW or PSM. A priori subgroup analyses were performed to compare the effectiveness of UST in TNFi-naïve and TNFi-exposed patients separately. These subgroup comparisons were conducted based on prior evidence suggesting a potential reduction in the effectiveness of biologics in TNFi-exposed patients with CD.28 We conducted sensitivity analyses as following: (1) replicating results from IPTW analysis by using PSM; (2) comparing effectiveness by regarding dose escalation as treatment failure; (3) comparing effectiveness in a subset of patients who started receiving medications after the policy of reimbursement; (4) comparing clinical effectiveness in a subset of patients who received at least one objective evaluation; (5) comparing clinical effectiveness at Week 52 by applying multiple imputation to handle missing data due to loss to follow-up. Predictive factors were first evaluated by univariate logistic regression using unweighted data. The factors that were significant with a P < 0.1 were included in a multivariable model to identify factors independently associated with steroid-free remission at Week 26 in patients treated with UST and VDZ, respectively. All statistical tests were two-sided, and a P-value of less than 0.05 was considered statistically significant. The analyses were performed using R software (version 4.4.2). Sample size In this observational nonrandomized cohort study, we used the observed event rates in the control group (VDZ-treated CD patients) to determine power calculations for unweighted comparisons of safety and effectiveness outcomes using the methodology described by Chow S et al.29 The sample size was determined beforehand to evaluate the feasibility of the project based on an expected steroid-free remission rate of 36% at Week 26 with VDZ, a two-sided α level of 0.05 with a power of 80% and sampling ratio of 0.5 (VDZ-group/UST-group).30 A sample size of 375 patients (125 for VDZ and 250 for UST) was needed to find an absolute difference of 15% between the two groups. Ethical considerations Institutional review board (IRB) approval was obtained at each institution (Supplementary Table S1). Informed consent was not required per IRB as the study was retrospective, and data collected was de-identified. Role of the funding source All authors had access to the dataset and accepted responsibility for the decision to submit for publication. The funding source had no access to the dataset, had no role in study design, data collection, analysis and interpretation of the data, writing of the manuscript, and decision to submit the manuscript for publication. Results Study population A total of 654 patients with CD received UST or VDZ at one of the six centers between May 2020 and July 2023. Among them, 536 met the inclusion criteria, including 386 who received UST and 150 who received VDZ (Fig. 1). The baseline patient characteristics are shown in Table 1. Before adjustment, patient characteristics were similar between the two groups, except for sex (male: 74.6% UST vs. 63.3% VDZ, P = 0.013), course of disease (UST, 53.2 ± 52.4 months; VDZ, 71.7 ± 57.3 months, P < 0.001), previous intestinal resection (UST, 26.4%; VDZ, 36.0%, P = 0.037), disease behavior (P < 0.001), previous EEN use (UST, 40.7%; VDZ, 52.0%, P = 0.023), previous steroid use (UST, 28.5%; VDZ, 44.0%, P = 0.023), previous immunomodulator use (UST, 38.6%; VDZ, 58.7%, P < 0.001), and proportion of baseline CRP >5 mg/L (UST, 52.8%; VDZ, 42.0%, P = 0.031). After IPTW adjustment, no statistically significant difference between the two groups was observed, and the SMDs for all included variables were less than 0.1.Fig. 1Study flowchart.Table 1Baseline characteristics of the overall population before and after IPTW.Before IPTWAfter IPTWUSTVDZPUSTVDZSMDN386150Male (n, %)288 (74.6)95 (63.3)0.013a71.270.00.028Age (years, mean (SD))33.3 (11.6)34.6 (10.9)0.26433.7 (11.6)33.7 (10.7)0.007Course of disease (months, mean (SD))53.2 (52.4)71.7 (57.3)<0.001a58.1 (55.7)58.7 (51.9)0.011Active smokers (n, %)48 (12.4)15 (10.0)0.52411.711.30.013Previous bowel resection (n, %)102 (26.4)54 (36.0)0.037a29.831.50.037Disease location (n, %)0.144 L150 (13.0)22 (14.7)13.512.00.049 L27 (1.8)7 (4.7)2.22.5 L3329 (85.2)121 (80.7)84.385.5 L421 (5.4)9 (6.0)0.9655.86.10.010Disease behavior (n, %)<0.001a0.079 B1209 (54.1)63 (42.0)50.951.3 B299 (25.6)32 (21.3)23.720.7 B378 (20.2)55 (36.7)25.528.0Active intestinal fistula (n, %)66 (17.1)32 (21.3)0.31017.415.90.040Perianal diseases (n, %)201 (52.1)82 (54.7)0.65752.452.80.008TNFi naïve (n, %)192 (49.7)66 (44.0)0.27248.048.40.009TNFi exposure (n, %)194 (50.3)84 (56.0)52.051.6Previous EEN (n, %)157 (40.7)78 (52.0)0.023a44.447.70.067Previous steroids (n, %)110 (28.5)66 (44.0)0.001a32.231.50.016Previous immunomodulator (n, %)149 (38.6)88 (58.7)<0.001a44.244.80.012Basic CDAI (mean (SD))185.1 (82.3)192.4 (108.2)0.403186.6 (86.2)185.4 (101.5)0.012Baseline CDAI >150 (n, %)265 (68.7)94 (62.7)0.22266.864.40.049EEN at baseline (n, %)45 (11.7)20 (13.3)0.69912.411.90.016Steroids at baseline (n, %)23 (6.0)10 (6.7)0.9166.26.10.002Immunomodulator at baseline (n, %)7 (1.8)2 (1.3)0.9891.61.40.019Baseline CRP (mg/L, mean (SD))15.3 (23.3)14.4 (26.9)0.70614.8 (21.2)14.4 (25.0)0.019Baseline CRP >5 mg/L (n, %)204 (52.8)63 (42.0)0.031a48.645.60.061Baseline albumin (g/L, mean (SD))38.7 (5.2)38.4 (4.6)0.49238.6 (5.2)38.6 (4.6)0.003Baseline albumin <35 g/L (n, %)83 (21.5)30 (20.0)0.79121.021.00.001IPTW, inverse probability of treatment weighting; UST, ustekinumab; VDZ, vedolizumab; TNFi, tumor necrosis factor alpha inhibitor; EEN, extensive enteral nutrition; CDAI, Crohn’s disease activity index; CRP, C-reactive protein.aP < 0.05 indicates statistical significance. Study flowchart. Table 1Baseline characteristics of the overall population before and after IPTW.Before IPTWAfter IPTWUSTVDZPUSTVDZSMDN386150Male (n, %)288 (74.6)95 (63.3)0.013a71.270.00.028Age (years, mean (SD))33.3 (11.6)34.6 (10.9)0.26433.7 (11.6)33.7 (10.7)0.007Course of disease (months, mean (SD))53.2 (52.4)71.7 (57.3)<0.001a58.1 (55.7)58.7 (51.9)0.011Active smokers (n, %)48 (12.4)15 (10.0)0.52411.711.30.013Previous bowel resection (n, %)102 (26.4)54 (36.0)0.037a29.831.50.037Disease location (n, %)0.144 L150 (13.0)22 (14.7)13.512.00.049 L27 (1.8)7 (4.7)2.22.5 L3329 (85.2)121 (80.7)84.385.5 L421 (5.4)9 (6.0)0.9655.86.10.010Disease behavior (n, %)<0.001a0.079 B1209 (54.1)63 (42.0)50.951.3 B299 (25.6)32 (21.3)23.720.7 B378 (20.2)55 (36.7)25.528.0Active intestinal fistula (n, %)66 (17.1)32 (21.3)0.31017.415.90.040Perianal diseases (n, %)201 (52.1)82 (54.7)0.65752.452.80.008TNFi naïve (n, %)192 (49.7)66 (44.0)0.27248.048.40.009TNFi exposure (n, %)194 (50.3)84 (56.0)52.051.6Previous EEN (n, %)157 (40.7)78 (52.0)0.023a44.447.70.067Previous steroids (n, %)110 (28.5)66 (44.0)0.001a32.231.50.016Previous immunomodulator (n, %)149 (38.6)88 (58.7)<0.001a44.244.80.012Basic CDAI (mean (SD))185.1 (82.3)192.4 (108.2)0.403186.6 (86.2)185.4 (101.5)0.012Baseline CDAI >150 (n, %)265 (68.7)94 (62.7)0.22266.864.40.049EEN at baseline (n, %)45 (11.7)20 (13.3)0.69912.411.90.016Steroids at baseline (n, %)23 (6.0)10 (6.7)0.9166.26.10.002Immunomodulator at baseline (n, %)7 (1.8)2 (1.3)0.9891.61.40.019Baseline CRP (mg/L, mean (SD))15.3 (23.3)14.4 (26.9)0.70614.8 (21.2)14.4 (25.0)0.019Baseline CRP >5 mg/L (n, %)204 (52.8)63 (42.0)0.031a48.645.60.061Baseline albumin (g/L, mean (SD))38.7 (5.2)38.4 (4.6)0.49238.6 (5.2)38.6 (4.6)0.003Baseline albumin <35 g/L (n, %)83 (21.5)30 (20.0)0.79121.021.00.001IPTW, inverse probability of treatment weighting; UST, ustekinumab; VDZ, vedolizumab; TNFi, tumor necrosis factor alpha inhibitor; EEN, extensive enteral nutrition; CDAI, Crohn’s disease activity index; CRP, C-reactive protein.aP < 0.05 indicates statistical significance. Baseline characteristics of the overall population before and after IPTW. IPTW, inverse probability of treatment weighting; UST, ustekinumab; VDZ, vedolizumab; TNFi, tumor necrosis factor alpha inhibitor; EEN, extensive enteral nutrition; CDAI, Crohn’s disease activity index; CRP, C-reactive protein. P < 0.05 indicates statistical significance. Clinical effectiveness At Week 26, in the overall adjusted population, UST was superior to VDZ in achieving steroid-free remission (55.4% vs. 46.1%, P = 0.003, OR = 1.451, 95% CI 1.140–1.850) and clinical remission (56.4% vs. 47.8%, P = 0.005, OR = 1.413, 95% CI 1.109–1.800; Fig. 2). The same results were replicated by using propensity score matching (Supplementary Table S2 and Figure S1). Among TNFi-naïve patients, there was no significant difference in steroid-free remission (65.5% UST vs. 59.6% VDZ, P = 0.166, OR = 1.289, 95% CI 0.900–1.848) and clinical remission (66.5% UST vs. 59.6% VDZ, P = 0.105, OR = 1.348, 95% CI 0.940–1.935) between the two groups. Nevertheless, among the TNFi-exposed patients, UST was superior to VDZ in achieving steroid-free remission (46.0% UST vs. 33.4% VDZ, P = 0.003, OR = 1.698, 95% CI 1.203–2.404) and clinical remission (47.0% UST vs. 36.7% VDZ, P = 0.015, OR = 1.531, 95% CI 1.089–2.158).Fig. 2Comparison of clinical effectiveness between UST and VDZ. (A) Comparison of steroid-free remission at Weeks 26 and 52 in the total, TNFi-naïve, and TNFi-exposed patients. (B) Comparison of clinical remission at Weeks 26 and 52 in the total, TNFi-naive and TNFi-exposed patients. Comparison of clinical effectiveness between UST and VDZ. (A) Comparison of steroid-free remission at Weeks 26 and 52 in the total, TNFi-naïve, and TNFi-exposed patients. (B) Comparison of clinical remission at Weeks 26 and 52 in the total, TNFi-naive and TNFi-exposed patients. At Week 52, in the overall adjusted population, UST was better than VDZ in achieving steroid-free remission (65.8% UST vs. 37.5% VDZ, P < 0.001, OR = 3.207, 95% CI 2.364–4.369; Fig. 2, Supplementary Table S3). The same results were replicated by using propensity score matching (Supplementary Table S4 and Figure S1). Subgroup analyses showed that UST showed superior effectiveness to VDZ in achieving steroid-free remission in both TNFi-naïve (79.4% UST vs. 49.0% VDZ, P < 0.001, OR = 4.029, 95% CI 2.437–6.780) and TNFi-exposed patients (55.6% UST vs. 29.7% VDZ, P < 0.001, OR = 2.965, 95% CI 1.985–4.464). Similar results were obtained for clinical remission. The sensitivity analysis applying multiple imputation to handle missing data due to loss to follow-up yielded similar results (Supplementary Figure S1). Sensitivity analyses limited to patients who started receiving medications after the reimbursement policy confirmed similar findings in the overall population (Supplementary Figure S1). Objective effectiveness A total of 400 patients (291 UST, 109 VDZ) underwent at least one objective assessments from Weeks 0 to Week 26. Among them, 63.8% (255/400) underwent endoscopy, 77.3% (309/400) underwent bowel ultrasound, and 74.3% (297/400) underwent CT/MR enteroclysis. Notably, 38.8% (155/400) of the patients underwent all three objective assessments, and 37.3% (149/400) underwent two of the methods. At Week 26, in the overall adjusted population, UST was better than VDZ in achieving an objective response at Week 26 (67.8% UST vs. 42.7% VDZ, P < 0.001, OR = 2.819, 95% CI 2.111–3.777; Fig. 3). The difference was significant for both TNF-naïve (81.8% UST vs. 54.3% VDZ, P < 0.001, OR = 3.771, 95% CI 2.400–6.024) and TNFi-exposed patients (54.4% UST vs. 31.0% VDZ, P < 0.001, OR = 2.656, 95% CI 1.766–4.026). Nevertheless, no statistically significant difference was observed in objective remission in CD patients (overall: 18.7% UST vs. 14.4% VDZ, P = 0.099, OR = 1.376, 95% CI 0.943–2.017), regardless of whether they were TNFi-naïve (26.8% UST vs. 18.7% VDZ, P = 0.057, OR = 1.594, 95% CI 0.990–2.589) or TNFi-exposed (11.0% UST vs. 10.0% VDZ, P = 0.731, OR = 1.120, 95% CI 0.589–2.150).Fig. 3Comparison of objective effectiveness between UST and VDZ. (A) Comparison of objective response at Week 26 and 52 in total, TNFi-naïve, and TNFi-exposed patients. (B) Comparison of objective remission at Weeks 26 and 52 in total, TNFi-naïve and TNFi-exposed patients. Comparison of objective effectiveness between UST and VDZ. (A) Comparison of objective response at Week 26 and 52 in total, TNFi-naïve, and TNFi-exposed patients. (B) Comparison of objective remission at Weeks 26 and 52 in total, TNFi-naïve and TNFi-exposed patients. From Week 0 to Week 52, a total of 215 patients (121 UST, 94 VDZ) underwent at least one objective assessment. Among them, 58.1% (125/215) underwent endoscopy, 90.2% (194/215) underwent bowel ultrasound, and 82.8% (178/215) underwent CT/MR enteroclysis. Notably, 47.4% (102/215) of the patients underwent all three objective assessments, and 36.3% (78/215) underwent two methods. In the adjusted overall population, UST was superior to VDZ in achieving objective response (66.7% UST vs. 23.8% VDZ, P < 0.001, OR = 6.418, 95% CI 4.213–9.913) and objective remission (31.4% UST vs. 12.7% VDZ, P < 0.001, OR = 3.146, 95% CI 1.928–5.255; Fig. 3). The differences were also significant for both TNF-naïve and TNF-exposed patients. The subgroup analyses based on endoscopy, bowel ultrasound, and CT/MR enteroclysis confirmed the significant findings (Supplementary Table S5). In addition, the sensitivity analysis evaluating the clinical outcomes from this subset of patients who received at least one objective evaluation showed that the results were consistent with those observed in the overall population (Supplementary Figure S1). Sensitivity analyses limited to patients who started receiving medications after the reimbursement policy also confirmed similar findings in the overall population (Supplementary Figure S1). CRP normalization rate in the weighted cohort was similar between UST-treated and VDZ-treated patients at Week 26 (UST 59.1% vs. VDZ 56.7%, P = 0.426), but CRP normalization rate was higher in UST-treated patients at Week 52 (UST 65.0% vs. VDZ 47.5%, P < 0.001). Dose escalation More patients in the UST group received dose escalation than those in the VDZ group from Week 0 to Week 52 (30.8% UST vs. 12.0% VDZ, P < 0.001). A total of 119 (30.8%) patients treated with UST received dose escalation, with 110 (92.4%) received intravenous more than once, 2 (1.7%) received interval reduction every 4 weeks, and 7 (5.9%) received both intravenous reinduction and interval reduction. A total of 18 (12.0%) patients treated with VDZ received dose escalation, with 14 (77.8%) receiving interval reduction every 4 weeks, and 4 (22.2%) receiving interval reduction every 6 weeks. Sensitivity analysis comparing the effectiveness between UST and VDZ regarding dose escalation as a failure confirmed the therapeutic superiority of UST over VDZ in patients with TNFi-naïve or TNFi-exposed CD (Supplementary Figure S1). Predictors of steroid-free remission at Week 26 In multivariate logistic regression analyses, a higher CDAI at baseline (P < 0.001, OR = 1.010, 95% CI 1.006–1.014), previous TNFi exposure (P = 0.045, OR = 1.665, 95% CI 1.012–2.739) and CRP >5 mg/L (P < 0.001, OR = 2.711, 95% CI 1.608–4.570) were identified as independent risk factors for failure to achieve UST-induced steroid-free remission at Week 26 (Supplementary Table S6). In multivariate logistic regression analyses, active intestinal fistula at baseline (P = 0.001, OR = 15.577, 95% CI 3.042–79.764), previous TNFi-exposure (P = 0.009, OR = 3.503, 95% CI 1.369–8.963) and higher CDAI at baseline (P < 0.001, OR = 1.017, 95% CI 1.010–1.025) were identified as independent risk factors for failure to achieve VDZ-induced steroid-free remission at Week 26 (Supplementary Table S7). Safety There was no significant difference in the adverse event rate between the two drugs (4.9% UST vs. 6.7% VDZ, P = 0.423). Among UST-treated patients, 1 (0.3%) developed symptomatic enteritis associated with Clostridium difficile, 1 (0.3%) developed secondary tuberculosis infection and suspended treatment; 8 (2.1%) got rashes that resolved after anti-allergic treatment, 5 (1.3%) got joint pain, and 4 (1.0%) got upper respiratory tract infections. Among VDZ-treated patients, 3 (2.0%) developed symptomatic enteritis associated with C. difficile, 4 (2.7%) got rashes, 1 (0.7%) got joint pain, and 2 (1.3%) got fatigue. Discussion To the best of our knowledge, the present study is the largest real-world cohort study to date investigating the comparative effectiveness of UST and VDZ in patients with CD, both TNFi-naïve and TNFi-exposed. In our cohort of 536 patients with CD from six IBD centers, we observed superior effectiveness of UST over VDZ in achieving both clinical and objective outcomes; however, there was no significant difference between the two drugs in terms of the risk of adverse events. Subgroup analyses suggested that UST was superior to VDZ in achieving clinical remission, steroid-free remission, and objective response at Weeks 26 and 52 in TNFi-exposed patients, whereas the therapeutic superiority of UST over VDZ was apparent only at Week 52 in TNFi-naïve patients. Although several studies have indirectly compared the effectiveness of UST and VDZ in TNFi-exposed patients with CD, the results were highly controversial.6, 7, 8, 9, 10, 11, 12 In cohorts from the Netherlands (85 UST, 128 VDZ)7 and France (224 UST, 88 VDZ; 107 UST, 132 VDZ),6,10 UST was associated with better clinical outcomes (steroid-free remission or clinical remission) in the long-term phase (Weeks 48–54); however, no statistically significant differences were observed in the short-term phase (Weeks 14–26). Conversely, two other retrospective studies from the UK (275 UST, 118 VDZ; 45 UST, 85 VDZ) demonstrated superiority of UST over VDZ in achieving better clinical outcomes in the short-term phase (Months 2–14), but the difference was not significant in the long-term follow-up of one year.8,11 In our study, UST showed better clinical effectiveness than VDZ in both the short-term and long-term phases. These inconsistent results may be due to the widespread use of additional VDZ dosing at Week 10 and dose escalation during the maintenance phase in most of the aforementioned studies. In addition, there was a more frequent use of concomitant medications at baseline in other studies (steroids: 6–6.7% in our study vs. 15–49% in others; immunosuppressants: 1.3%–1.7% in our study vs. 14%–46% in others). These two strategies are thought to enhance or regain the therapeutic effectiveness of biologics,31,32 which may narrow the gap between the two drugs, especially in the short-term phase. Most studies were limited by a lack of objective evaluations. To address this issue, we reported objective outcomes evaluated using endoscopy, CT/MR enteroclysis, and small-bowel ultrasound. In our cohort, UST was associated with a higher objective response rate at Weeks 26 and 52, and a higher objective remission rate at Week 52 in TNFi-exposed patients with CD. These objective data are consistent with the clinical data obtained in our study. Interestingly, a recent Italian cohort with clinical and objective evaluations showed data contrary to those of our study and others.12 Onali et al. reported no differences in clinical response, remission, and steroid-free remission at Week 26, whereas VDZ-treated patients had higher clinical remission and steroid-free remission rates at Week 52. UST and VDZ achieved similar objective response and remission at Week 52. The authors highlighted that the discrepancies between their study and previous studies could be attributed to differences in the study population, specifically a lower proportion of refractory disease and a potentially more favorable response to VDZ in their cohort. In contrast, most TNFi-exposed patients with CD in our cohort received either UST or VDZ early after their approval and availability in the market, reflecting a population with relatively more refractory disease. In addition, in the Italian cohort, more patients, particularly those treated with VDZ, received concomitant steroids at baseline (31% UST vs. 49% VDZ, P < 0.0001). In routine clinical practice, we often use UST or VDZ as monotherapy. Therefore, only a small number of patients in our study were concomitant with steroids or immunosuppressants at baseline. This indicates that our results might be less biased by concomitant medications compared to previous studies. Few studies have compared these two drugs in TNFi-naïve patients with CD. Two network meta-analyses showed no significant difference between these two drugs in inducing a clinical response or remission in biologic-naïve patients with moderate-to-severe CD; however, UST was found to have a slightly higher ranking than VDZ.33,34 Consistent with these results, our study showed no differences between UST and VDZ in achieving clinical or steroid-free remission in the short-term phase; however, UST was associated with a higher objective response rate at Week 26. Regarding the comparative effectiveness in maintaining remission, network meta-analyses provided data on a mixed population of biologic-naïve and biologic-exposed patients. Our study further showed that UST was superior to VDZ in terms of clinical remission, steroid-free remission and objective remission during the long-term phase in TNFi-naïve patients. It is unlikely that this difference could solely account for the slower onset of action of VDZ, as the significant difference was more apparent at Week 52. There is a lack of widely accepted standards for current practice for minimal clinically important difference (MCID) in clinical trials of Crohn’s disease. Based on a systematic review of the published clinical data, a proposed MCID of 8.6% was presented for comparing UST vs. placebo in inducing clinical remission.35 Considering that our study compared two active biologics, the differences of 9.3% in steroid-free remission rate and 8.6% in clinical remission rate at Week 26 in the overall population were considered clinically significant. Notably, these differences were more apparent in the TNF inhibitor-exposed patients (12.6% in TNFi-exposed patients vs. 10.3% in TNFi-naïve patients). Moreover, the differences in the steroid-free and clinical remission rates at Week 52 in the overall population even reached 28.3%. Similar results were observed for objective outcomes. We believe that these data revealed that UST exhibited superior real-world effectiveness compared to VDZ in patients with CD, especially in the long-term phase. Similar to other studies, our study showed that the rates of adverse events associated with UST and VDZ were low.36,37 No serious adverse events were observed in this cohort. However, owing to the retrospective nature of our study, the incidence of minor adverse events that did not require medical intervention may have been underestimated. It was noted that a few more cases of enteritis associated with C. difficile were observed in VDZ-treated patients. However, the association requires further investigation. Both UST and VDZ were safe during the follow-up period in our cohort. Our study has several strengths. To the best of our knowledge, this is the first real-world cohort study to compare the effectiveness of UST and VDZ in TNFi-naïve patients with CD. Detailed patient data and clinical and objective evaluations allowed us to reliably assess the comparative effectiveness of these two drugs in a large cohort in real-life settings. However, this study has some limitations. First, the retrospective observational nature of this study had inherent limitations. As there are currently no ongoing or planned head-to-head trials, well-designed observational studies have yielded the best available evidence to guide clinical decision-making. Secondly, unknown confounders may have influenced the results. However, we adjusted for potential confounding factors that are widely recognized to be associated with refractory disease phenotypes or disease severity. Thirdly, UST dose escalation was more commonly performed in the UST population. However, it is unlikely that the difference in the therapeutic efficacy between the two drugs could solely account for UST dose escalation, as the sensitive analysis regarding dose escalation as treatment failure confirmed the significant difference. Fourthly, therapeutic drug monitoring was not routinely performed during this study. Trough levels were not routinely assessed in our study because of the lack of widely accepted threshold concentrations for UST and VDZ.38 The anti-drug antibodies have not yet been identified. However, low immunogenicity did not appear to be the main reason for the lack of therapeutic efficacy of these two drugs.38 Finally, caution should be exercised when interpreting comparisons of secondary outcomes and subgroup analyses given the potential for selection bias. In conclusion, our study demonstrated that UST exhibited superior clinical and objective effectiveness outcomes compared to VDZ in patients with CD, while safety outcomes were comparable between the two drugs. Subgroup analyses suggested that UST may be superior to VDZ in achieving short- and long-term effectiveness in TNFi-exposed patients, whereas the therapeutic superiority of UST over VDZ was apparent only in the long-term phase in TNFi-naïve patients. Randomized, head-to-head studies are required to validate these observations. Contributors Hongsheng Yang contributed to conceptualisation, data curation, formal analysis, funding acquisition, investigation, methodology, project administration, and writing—original draft; Zhaopeng Huang contributed to data curation, formal analysis, and writing—original draft; Miao Li, Haiyan Zhang, Lingyu Fu, Xiaoling Wang, Qiaoqiao Yang, Yun He, Wenhong Wu, Taofeng Jiang, Zhimei Sun, Tao Zhang, Dongping Lai, Qin Guo and Yanhui Wu contributed to data curation; Xixin Wu contributed to methodology; Lishuo Shi contributed to statistical analyses; Jiang Tang and Kang Chao contributed to data curation and writing—review & editing; Xiang Gao was responsible for supervision and writing—review & editing. The corresponding author had the final responsibility to submit for publication. All authors agree to be accountable for all aspects of the work and gave final approval for the version to be published. Data sharing statement The analysis code is available on github (https://github.com/Tim-FRT/UST-vs-VDZ-code). Individual participant data will not be shared due to privacy/ethical restrictions. Declaration of interests There is no conflict of interest.