Vedolizumab for stricturing Crohn’s disease of the terminal ileum – an international multicenter cohort study

Abstract

Objective

Crohn’s disease (CD) frequently progresses to stricturing complications. Anti-stricture therapies are lacking, and anti-inflammatory therapies are hence the mainstay of medical treatment. However, data on the efficacy of vedolizumab in stricturing CD are not available. We therefore aimed to assess outcomes of vedolizumab in this setting.

Design/method

We evaluated adult patients who received vedolizumab within 6 months of imaging-confirmed symptomatic ileal stricturing CD (CONSTRICT criteria) in a multicenter observational cohort study. Images were centrally read by an expert radiologist with experience in IBD imaging. Outcomes were rates of intervention (composite endpoint for endoscopic balloon dilation (EBD) or surgery during follow-up), EBD, surgery, obstructive symptoms, and vedolizumab persistence.

Results

Forty-four patients were included (61% female, median age 52 years); 64% had an anastomotic stricture; 84% had received prior biologics. Median follow-up was 17 months. Cumulative intervention rates were 23%, 44% and 50% of patients at 12, 24 and 48 months, respectively. 52% of patients reported symptom resolution, of whom 26% subsequently developed symptoms during follow-up. Vedolizumab persistence was 82%, 74% and 59% at 12, 24 and 48 months. Anastomotic stricture was inversely associated with intervention (HR0.21, P = 0.012) and surgery (HR0.11, P = 0.013). Higher obstructive symptom index (HR4.47, P = 0.040) and degree of prestenotic bowel dilatation (HR1.09, P = 0.030) were associated with recurrent obstructive symptoms. Elevated baseline CRP was associated with vedolizumab discontinuation (HR1.96, P = 0.014).

Conclusion

Vedolizumab appears to be a treatment option in CD patients with terminal ileal strictures, especially in patients with anastomotic strictures, without obstructive symptoms or prestenotic bowel dilatation. Validation in prospective controlled studies is needed.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12876-026-04815-4.

Introduction

Vedolizumab, a monoclonal antibody against α₄β₇ integrin is a safe and effective treatment for moderate to severe Crohn’s disease [1]. Vedolizumab is frequently used in luminal Crohn’s disease (CD), however there are limited data regarding outcomes in stricturing CD.

This is relevant, as a significant proportion of CD patients develop obstructive symptoms in their lifetime due to stricturing disease [2]. Currently, no selective anti-fibrotic therapies are available and patients often require endoscopic balloon dilation (EBD) or surgery [3]. Since strictures comprise both fibrotic and inflammatory components [4], initial treatment often involves anti-inflammatory therapy. Anti-tumor necrosis factor (anti-TNF) therapy has indeed shown efficacy in this setting, and its use is supported by a recent global consensus on the management of stricturing CD [4–7].

The Stenosis Therapy and Anti-Fibrosis Research (STAR)Consortium [8, 9], a global stakeholder group with the mission to build a pathway to testing anti-stricture therapies, generated radiologic definitions for stricturing terminal ileal (TI) CD. The CONSTRICT criteria [10] are now widely adopted and are included into clinical trials for fibrostenosing CD. (NCT05843578, NCT05013385).

To address the knowledge gap in the literature regarding vedolizumab in TI stricturing CD disease, we performed an international retrospective cohort study assessing progression rates to and risk factors for interventions, medication persistence or symptom recurrence in a well-defined population using accepted objective pre-specified radiologic stricture criteria. These data are important for clinical care and design of future clinical trials in stricturing CD.

Materials and methods

Study design

This was an international multicenter observational retrospective cohort study of patients with symptomatic TI stricturing CD. Participating centers were the Cleveland Clinic, Cleveland, OH; Cleveland Clinic, Abu Dhabi; Mayo Clinic, Rochester, MN and University of Calgary, Calgary, Canada. All consecutive patients with an established CD diagnosis who underwent cross-sectional imaging (Magnetic Resonance enterography (MRE), Computed tomography enterography (CTE) or intestinal ultrasound (IUS)) at the participating sites between February 2010 and December 2020 were identified through either a prospectively maintained IRB-approved radiology database or clinical databases as available in each site. To minimize selection bias, all consecutive patients who received vedolizumab according to its labelled dosing within 6 months after eligible baseline imaging were included. If multiple imaging studies were eligible within 6 months, the most recent study closest to treatment onset was selected. Surgical resection between imaging and vedolizumab start was not permitted. The timeframe of the study was chosen to allow adequate follow-up to meet the study objective and to capture patients post approval of vedolizumab. To ensure study rigor, cross-sectional imaging was centrally read at each site by an expert radiologist with experience in IBD imaging (M.B., A.M., N.B., J.F., A.I.) or a gastroenterologist with experience in IUS (C.L.) blinded to the outcome parameters.

Study population

Inclusion criteria comprised: (1) Adult patients with an established diagnosis of CD; (2) The presence of small bowel stricturing TI disease on cross-sectional imaging including CTE, MRE or IUS as defined by the CONSTRICT criteria (see ‘Definitions’); (3) Cross-sectional imaging performed according to commonly accepted technical parameters [11–13] with imaging quality being deemed sufficient by the central reader radiologist; (4) The presence of at least one obstructive symptom, as defined below (see ‘Definitions’); (5) Vedolizumab treatment initiated within 6 months after baseline imaging.

Exclusion criteria included: (1) Internal penetrating disease including fistula, abscess or inflammatory mass at baseline (a blind-ending sinus was not excluded); (2) Prior strictureplasty at the stricture site; (3) Endoscopic stricturotomy within the last 12 months; (4) Strictures associated with an ileal pouch anal anastomosis or Kock pouch; (5) Proximal diverting loop ileostomy; (6) Stricture associated with an ileostomy; (7) Gastrointestinal malignancy; (8) Concurrent colonic stricture. The selection criteria are consistent with a CD population with small bowel strictures without associated complications.

Data collection and definitions

Patient demographics, as well as clinical, laboratory and radiological characteristics were collected from electronic health records and documented in a central REDCap database hosted at Cleveland Clinic, Ohio [14, 15]. Patients were both assessed at baseline prior to starting vedolizumab and during follow-up. Cross-sectional imaging and medical records post treatment were reviewed for any developing complications.

Strictures on cross-sectional imaging were defined according to the CONSTRICT criteria [10] and required 2 out of the 3 following criteria for diagnosis for both naive and anastomotic strictures: (1) 25% bowel wall thickness increase relative to adjacent nonaffected bowel, (2) lumen diameter reduction of at least 50% relative to non-affected adjacent bowel loop; (3) ≥ 3 cm prestenotic dilation. Baseline cross-sectional imaging was defined as the first imaging modality that met inclusion criteria, within 6 months of treatment onset.

Medical records were assessed for obstructive symptoms as defined by the CONSTRICT consensus [10]. An obstructive symptom index was calculated, as previously described [16] to quantify the degree of obstructive symptomatology and ranged from 0 (no symptoms) to 7 (maximum number of symptoms) (Supplementary Table 1). At least one documented obstructive symptom was required for inclusion.

Study outcomes

The following outcomes were pre-defined: (1)Occurrence of event (primary endpoint: composite endpoint of surgery or EBD) during follow-up and time to intervention; (2)Occurrence of EBD and time to EBD during follow-up; (3)Occurrence of surgery (including strictureplasty or resection performed at the stricture site) and time to surgery during follow-up; (4)Presence of obstructive symptoms after vedolizumab start and (5)Vedolizumab persistence (defined as the time from initiation to discontinuation of vedolizumab). Additional exploratory outcomes were the time to symptom recurrence for those initially becoming asymptomatic after vedolizumab start and the development of penetrating complications during follow-up. Follow-up for each patient started at the time of baseline imaging and was censored at the time of surgery or when follow up was lost, whatever came first.

Statistical analysis

Categorical factors were summarized with frequencies and percentages. Continuous measures were evaluated for normality and summarized with means and standard deviations, if normal, and medians and quartiles otherwise. Kaplan–Meier estimates of event frequencies at different timepoints were calculated with 95% confidence intervals, and plots through 3 years and 5 years were created. Predictors were assessed using Cox proportional hazards models. In univariable tables, factors with 0 events in one or more groups were analyzed using Firth’s penalized regression models. Hazard ratios and 95% confidence intervals from these models are presented. A separate analysis for predictors of each event was performed, adjusting for stricture length using Cox models. Analyses were performed using R software (version 4.0; Vienna, Austria), using the gtsummary package [17].

Results

Patient demographics and disease characteristics

Fourty-four consecutive patients starting on vedolizumab with cross-sectional imaging studies (14 MR, 23 CT and 7 IUS) with symptomatic stricturing CD were eligible (Table 1, Supplementary Fig. 1). Median time from baseline imaging to vedolizumab start was 1.63 [25th percentile (P25) 0.63, 75th percentile (P75) 3.58] months. A total of 61% were female; 84% were biologic-exposed. Among patients previously exposed to biologics, 68% had prior exposure to infliximab, 64% to adalimumab, 25% to certolizumab, and 11% to ustekinumab. Co-induction with corticosteroids occurred in 39% and 23% received concomitant immunomodulators. The median obstructive symptom score at baseline was 2 [1, 2] and the median stricture length was 10 [5, 18] cm.

Table 1.

Patient characteristics at baseline

Characteristic	N	N = 44
Age (years), Median [IQR]	43	52 [31, 63]
Gender, Female, n (%)	44	27 (61)
Disease duration, Median [IQR]	44	276 [107, 424]
Race, n (%)	43
American Indian		1 (2.3)
Asian		0 (0)
Black		1 (2.3)
Pacific Islander		0 (0)
White		41 (95)
Ethnicity, n (%)	44
Hispanic		2 (4.5)
Non-Hispanic		33 (75)
Unknown		9 (20)
BMI (kg/m2), Median [IQR]	32	23.6 [20.4, 26.6]
Smoking, n (%)	44
Active		9 (20)
Former		13 (30)
Never		22 (50)
Anastomotic stricture, n (%)	44	28 (64)
Months from Stricture Diagnosis to Imaging, Median [IQR]	44	8 [3, 42]
Obstructive Index, n (%)	44
1		15 (34)
2		21 (48)
3		5 (11)
4		2 (4.5)
5		1 (2.3)
Duration of Obstructive Symptoms at Baseline, Months, Median [IQR]	38	7 [5, 23]
Months from Imaging to Medication Start, Median [IQR]	44	1.63 [0.63, 3.58]
Combination Therapy, n (%)	44	10 (23)
Fecal calprotectin (mg/kg), Median [IQR]	9	764 [500, 1,865]
CRP (mg/dl), Median [IQR]	29	0.63 [0.31, 1.54]
Albumin (g/dL), Median [IQR]	40	3.80 [3.48, 4.10]
Co-induction of Medication with Steroids, n (%)	41	16 (39)
Prior Biologic Therapies, n (%)	44
0		7 (16)
1		11 (25)
2		14 (32)
3		12 (27)
EBD Prior to Baseline, n (%)	44
None		35 (80)
Within last 6 months		4 (9.1)
Prior to last 6 months		5 (11)
Narrowest luminal diameter within stricture (mm), Median [IQR]	44	2.00 [1.63, 3.00]
Length of stricture (cm), Median [IQR]	43	10 [5, 19]
Length of stricture > 3 cm, n (%)	43	37 (86)
Length of stricture > 5 cm, n (%)	43	30 (70)
Length of stricture > 7 cm, n (%)	43	25 (58)
Maximal diameter of proximal small bowel dilation (mm), Median [IQR]	44	30 [26, 36]
Presence of Sinus, n (%)	44	4 (9.1)
Maximal wall thickness of stricture (mm), Median [IQR]	44	8.60 [7.07, 10.00]

BMI body mass index, CRP C-reactive protein, EBD endoscopic balloon dilation, IBD inflammatory bowel disease, MRE magnetic resonance enterography

Event rates and factors associated with intervention, EBD, surgery and symptom occurrence

Intervention during follow-up

The primary endpoint was defined as the composite need for EBD or surgery during follow-up (whichever came first, termed ‘intervention’). The median follow-up was 13 [6, 19] months, during which 23%, 44% and 50% of patients progressed to an event at 12, 24 and 48 months, respectively (Table 2, Fig. 1A). The median time to intervention was 11 months. No predictors of intervention were found on univariate analysis (Supplementary Table 2). Anastomotic stricture was a negative predictor of intervention on multivariable analysis (HR 0.21 [0.06, 0.70]; P = 0.012) (Table 3).

Table 2.

Cumulative rates of events during follow-up

Characteristic	N	Event N	6 Months	12 Months	18 Months	24 Months	36 Months	48 Months	60 Months
Surgery	44	9	6.9% (0%, 14%)	13% (1.5%, 23%)	13% (1.5%, 23%)	22% (5.0%, 36%)	34% (10%, 51%)	34% (10%, 51%)	34% (10%, 51%)
Endoscopic Balloon Dilation	44	9	6.9% (0%, 14%)	10% (< 0.1%, 19%)	18% (3.2%, 31%)	29% (8.1%, 44%)	34% (11%, 52%)	34% (11%, 52%)	34% (11%, 52%)
Obstructive Symptoms	23	6	4.5% (0%, 13%)	20% (< 0.1%, 36%)	27% (2.8%, 46%)	38% (5.3%, 59%)	38% (5.3%, 59%)	38% (5.3%, 59%)	38% (5.3%, 59%)
Vedolizumab Persistence	44	14	95% (89%, 100%)	82% (70%, 95%)	78% (65%, 93%)	74% (60%, 91%)	64% (49%, 85%)	59% (42%, 82%)	39% (17%, 93%)
Intervention (Surgery or EBD)	44	15	14% (2.9%, 24%)	23% (8.3%, 35%)	30% (13%, 44%)	44% (22%, 60%)	50% (25%, 66%)	50% (25%, 66%)	50% (25%, 66%)

EBD, endoscopic balloon dilation

Fig. 1.

Time from baseline to A Intervention (defined as need for endoscopic balloon dilation or surgical resection, whichever came first). B Endoscopic balloon dilation. C Surgery. D Obstructive symptom recurrence. E Vedolizumab persistence

Table 3.

Predictors of intervention (surgery or EBD) on multivariable analysis

Characteristic	N	Event N	HR (95% CI)1	p-value
BMI (kg/m2)	31	12	0.93 (0.82, 1.06)	0.28
Smoking
Active	9	2	—
Former	12	1	0.37 (0.03, 4.24)	0.42
Never	22	11	2.58 (0.57, 11.7)	0.22
Anastomotic stricture
No	16	7	—
Yes	27	7	0.21 (0.06, 0.70)	0.012
Months from Stricture Diagnosis to Imaging	43	14	1.00 (0.98, 1.01)	0.71
Obstructive Index	43	14	1.47 (0.96, 2.25)	0.074
Duration of Obstructive Symptoms at Baseline, Months	37	13	1.00 (0.98, 1.02)	0.91
CRP (mg/dl)	29	9	1.28 (0.79, 2.09)	0.31
Albumin (g/dL)	39	13	0.76 (0.23, 2.53)	0.65
Co-induction of Medication with Steroids
No	24	7	—
Yes	16	6	2.19 (0.73, 6.59)	0.16
Prior Biologic Therapies	43	14	0.82 (0.44, 1.53)	0.53
EBD Prior to Baseline
None	34	11	—
Within last 6 months	4	1	0.66 (0.08, 5.26)	0.70
Prior to last 6 months	5	2	0.68 (0.14, 3.27)	0.63
Narrowest luminal diameter within stricture (mm)	43	14	1.09 (0.70, 1.70)	0.69
Maximal diameter of proximal small bowel dilation (mm)	43	14	1.02 (0.97, 1.06)	0.45
Maximal wall thickness of stricture (mm)	43	14	1.18 (0.90, 1.53)	0.23

BMI, body mass index, CRP, C-reactive protein, EBD endoscopic balloon dilation, IBD inflammatory bowel disease, MRE magnetic resonance enterography

¹HR Hazard Ratio, CI Confidence Interval

EBD during follow-up

The median follow-up was 14 [7, 20] months, during which 10%, 29% and 34% of patients required EBD at 12, 24 and 48 months, respectively (Table 2, Fig. 1B). Median time to EBD was 14 months. No predictors for EBD were found on univariate or multivariable analysis (Supplementary Table 3 and Table 4).

Table 4.

Predictors of EBD on multivariable analysis

Characteristic	N	Event N	HR (95% CI)1	p-value
BMI (kg/m2)	31	6	1.00 (0.86, 1.15)	0.95
Smoking2
Active	9	0	—
Former	12	0	0.66 (0.003, 126.4)	0.84
Never	22	8	7.21 (0.88, 935.8)	0.070
Anastomotic stricture
No	16	3	—
Yes	27	5	0.60 (0.13, 2.79)	0.52
Months from Stricture Diagnosis to Imaging	43	8	1.00 (0.98, 1.02)	0.86
Obstructive Index	43	8	1.57 (0.89, 2.77)	0.12
Duration of Obstructive Symptoms at Baseline, Months	37	7	1.01 (0.99, 1.03)	0.57
CRP (mg/dl)	29	4	1.16 (0.45, 2.98)	0.77
Albumin (g/dL)	39	7	0.67 (0.12, 3.72)	0.65
Co-induction of Medication with Steroids
No	24	4	—
Yes	16	3	1.52 (0.34, 6.89)	0.59
Prior Biologic Therapies	43	8	0.44 (0.17, 1.15)	0.092
EBD Prior to Baseline
None	34	5	—
Within last 6 months	4	1	1.34 (0.14, 12.5)	0.80
Prior to last 6 months	5	2	1.65 (0.30, 9.16)	0.57
Narrowest luminal diameter within stricture (mm)	43	8	0.89 (0.45, 1.77)	0.74
Maximal diameter of proximal small bowel dilation (mm)	43	8	1.00 (0.95, 1.06)	0.87
Maximal wall thickness of stricture (mm)	43	8	1.19 (0.83, 1.71)	0.33

BMI body mass index, CRP C-reactive protein, EBD endoscopic balloon dilation, IBD inflammatory bowel disease, MRE magnetic resonance enterography

¹HR Hazard Ratio, CI Confidence Interval

² Firth’s penalized regression

Surgery during follow-up

The median follow-up was 17 [8, 21] months, during which 13%, 22% and 34% of patients required surgery at 12, 24 and 48 months, respectively (Table 2, Fig. 1C). Median time to surgery was 11 months. The presence of an anastomotic stricture was negatively associated with surgery on univariate analysis (HR 0.23 [0.05, 0.98]; P = 0.047) (Supplementary Table 4) as well as on multivariable analysis (HR 0.11 [0.02, 0.63]; P = 0.013) (Table 5). The type of surgery in all cases was a small bowel resection.

Table 5.

Predictors of surgery on multivariable analysis

Characteristic	N	Event N	HR (95% CI)1	p-value
BMI (kg/m2)	31	8	0.88 (0.73, 1.08)	0.22
Smoking
Active	9	2	—
Former	12	1	0.37 (0.03, 4.36)	0.43
Never	22	6	1.18 (0.24, 5.88)	0.84
Anastomotic stricture
No	16	5	—
Yes	27	4	0.11 (0.02, 0.63)	0.013
Months from Stricture Diagnosis to Imaging	43	9	0.99 (0.96, 1.01)	0.30
Obstructive Index	43	9	1.38 (0.79, 2.41)	0.26
Duration of Obstructive Symptoms at Baseline, Months	37	9	0.98 (0.95, 1.02)	0.36
CRP (mg/dl)	29	6	1.28 (0.72, 2.29)	0.40
Albumin (g/dL)	39	8	0.63 (0.13, 2.98)	0.56
Co-induction of Medication with Steroids
No	24	5	—
Yes	16	4	1.89 (0.50, 7.11)	0.35
Prior Biologic Therapies	43	9	1.09 (0.49, 2.42)	0.83
EBD Prior to Baseline2
None	34	9	—
Within last 6 months	4	0	0.41 (0.003, 3.15)	0.47
Prior to last 6 months	5	0	0.13 (0.001, 1.12)	0.067
Narrowest luminal diameter within stricture (mm)	43	9	1.23 (0.73, 2.05)	0.44
Maximal diameter of proximal small bowel dilation (mm)	43	9	1.00 (0.95, 1.05)	> 0.99
Maximal wall thickness of stricture (mm)	43	9	1.07 (0.76, 1.49)	0.71

BMI body mass index, CRP C-reactive protein, EBD endoscopic balloon dilation, IBD inflammatory bowel disease, MRE, magnetic resonance enterography

¹HR Hazard Ratio, CI Confidence Interval

² Firth’s penalized regression

Occurrence of symptoms during follow-up

At baseline, all patients had at least one obstructive symptom with a median duration of 7 [5, 22] months prior to baseline. At follow-up, 52% of patients reported initial symptom resolution after vedolizumab start. However, 26% of those patients subsequently developed obstructive symptoms. Among patients whose symptoms recurred, 20% and 38% developed recurrent obstructive symptoms at 12 and 24 months, respectively. (Table 2, Fig. 1D). Predictors of obstructive symptom recurrence after initial symptom resolution on univariate analysis included a higher obstructive index at baseline (HR 2.78 [1.01,7.68]; P = 0.048) and prestenotic maximal bowel diameter (HR 1.06 [1.00,1.13]; P = 0.044; Supplementary Table 5). On multivariable analysis, considering stricture length in the model, the obstructive index (HR 4.47 [1.07, 18.6]; P = 0.040), as well as prestenotic maximal bowel diameter (HR 1.09 [1.01, 1.17]; P = 0.030), (Table 6) remained predictors of symptom recurrence.

Table 6.

Predictors of obstructive symptom recurrence on multivariable analysis

Characteristic	N	Event N	HR (95% CI)1	p-value
BMI (kg/m2)	18	6	0.85 (0.61, 1.18)	0.33
Smoking
Active	4	1	—
Former	8	2	0.44 (0.02, 8.20)	0.58
Never	11	3	1.33 (0.13, 13.7)	0.81
Anastomotic stricture
No	9	4	—
Yes	14	2	0.22 (0.04, 1.24)	0.086
Months from Stricture Diagnosis to Imaging	23	6	0.95 (0.88, 1.02)	0.15
Obstructive Index	23	6	4.47 (1.07, 18.6)	0.040
Duration of Obstructive Symptoms at Baseline, Months	20	6	0.97 (0.91, 1.03)	0.36
CRP (mg/dl)	15	3	6.93 (0.49, 98.2)	0.15
Albumin (g/dL)	21	6	1.80 (0.20, 16.2)	0.60
Co-induction of Medication with Steroids
No	15	3	—
Yes	5	2	5.66 (0.25, 130)	0.28
Prior Biologic Therapies	23	6	1.18 (0.47, 2.99)	0.72
EBD Prior to Baseline2
None	17	4	—
Within last 6 months	3	2	2.04 (0.35, 9.33)	0.39
Prior to last 6 months	3	0	0.27 (0.002, 2.77)	0.32
Narrowest luminal diameter within stricture (mm)	23	6	1.26 (0.63, 2.50)	0.52
Maximal diameter of proximal small bowel dilation (mm)	23	6	1.09 (1.01, 1.17)	0.030
Maximal wall thickness of stricture (mm)	23	6	0.94 (0.60, 1.48)	0.79

BMI body mass index, CRP C-reactive protein, EBD endoscopic balloon dilation, IBD inflammatory bowel disease, MRE magnetic resonance enterography

¹HR Hazard Ratio, CI Confidence Interval

² Firth’s penalized regression

Vedolizumab persistence during follow-up

Patients were censored at time of surgery or at end of follow-up. The median overall follow-up was 17 [10, 23] months, during which 82%, 74% and 59% of patients were still on vedolizumab at 12, 24 and 48 months, respectively (Table 2, Fig. 1E). An elevated baseline CRP was associated with vedolizumab discontinuation on univariate analysis (HR 1.92 [1.14, 3.25]; P = 0.014), whereas the presence of anastomotic stricture was inversely associated with vedolizumab discontinuation (HR 0.31 [0.10, 0.99]; P = 0.049) (Supplementary Table 6). On multivariable analysis, CRP was the only predictor of vedolizumab discontinuation (HR1.96 [1.15, 3.35]; P = 0.014) (Table 7). During follow-up, escalation of vedolizumab to 4-weekly dosing occurred in 44% over a median time of 9 months [8, 24]. Among the 17 patients who discontinued vedolizumab, 47% did so due to lack of efficacy, 24% due to perceived adverse effects, 12% due to cost, and 18% for unspecified reasons.

Table 7.

Predictors of vedolizumab discontinuation on multivariable analysis

Characteristic	N	Event N	HR (95% CI)1	p-value
BMI (kg/m2)	31	11	1.04 (0.93, 1.16)	0.53
Smoking
Active	9	4	—
Former	12	3	0.35 (0.04, 2.79)	0.32
Never	22	6	0.89 (0.20, 3.86)	0.87
Anastomotic stricture
No	16	6	—
Yes	27	7	0.31 (0.09, 1.06)	0.063
Months from Stricture Diagnosis to Imaging	43	13	1.00 (0.99, 1.02)	0.96
Obstructive Index	43	13	0.76 (0.39, 1.48)	0.41
Duration of Obstructive Symptoms at Baseline, Months	37	13	0.99 (0.97, 1.02)	0.63
CRP (mg/dl)	29	9	1.96 (1.15, 3.35)	0.014
Albumin (g/dL)	39	13	0.50 (0.17, 1.49)	0.21
Co-induction of Medication with Steroids
No	24	6	—
Yes	16	5	0.76 (0.19, 2.99)	0.69
Prior Biologic Therapies	43	13	1.24 (0.64, 2.39)	0.52
EBD Prior to Baseline
None	34	10	—
Within last 6 months	4	2	1.36 (0.21, 8.62)	0.75
Prior to last 6 months	5	1	0.47 (0.06, 3.94)	0.49
Narrowest luminal diameter within stricture (mm)	43	13	1.14 (0.79, 1.65)	0.50
Maximal diameter of proximal small bowel dilation (mm)	43	13	1.00 (0.95, 1.05)	0.98
Maximal wall thickness of stricture (mm)	43	13	1.09 (0.82, 1.44)	0.55

BMI body mass index, CRP C-reactive protein, EBD endoscopic balloon dilation, IBD inflammatory bowel disease, MRE magnetic resonance enterography

¹HR Hazard Ratio, CI Confidence Interval

Penetrating complications

Two patients (4.5%) developed internal penetrating disease during follow-up. None of them had a pre-existing sinus on baseline imaging.

Discussion

In this cohort of patients with symptomatic stricturing TI CD as confirmed radiologically, per the CONSTRICT criteria [10] and treated with vedolizumab, 44% of patients underwent intervention (either EBD or surgery) within 24 months. Median time to intervention was 11 months. A total of 52% of patients reported initial symptom resolution, of which 26% subsequently re-developed obstructive symptoms. Vedolizumab persistence was 74% at 24 months.

A recent global expert consensus for the treatment of stricturing CD concluded that anti-inflammatory therapy in patients with signs of active inflammatory disease is appropriate.⁴ In principle, the type of anti-inflammatory medication largely follows treatment algorithms for luminal CD. In patients with intestinal obstruction, after initial successful conservative therapy with steroids and decompression, azathioprine showed superiority over mesalamine for longer time to re-hospitalization [24]. The CREOLE [6] and STRIDENT [7] prospective studies evaluating adalimumab in symptomatic small bowel CD [7] demonstrated that anti-TNF biologics may be effective in patients with symptomatic CD strictures. This is consistent with a recent systematic review [18] and the TREAT registry [25]. Data are limited on non-anti-TNF treatments [26]. In fact, patients with stricturing disease are excluded from clinical registration trials. Risankizumab and upadacitinib led to improvement in luminal narrowing on endoscopy [27] from a post-hoc analysis of the registration trials [22, 28]. However, the luminal narrowing component of the Simple Endoscopic Score (SES-CD), which was used in this investigation as the endpoint, lacks reliability and is highly operator-dependent [29]. Using rigorous CONSTRICT inclusion criteria, outcomes for the anti-IL12/23 ustekinumab in patients with ileal CD have been reported [26]. In this study which also included a predominantly anti-TNF-exposed population, 45% of patients underwent intervention (EBD or surgery), within 24 months [26]. Our vedolizumab cohort demonstrated comparable outcomes, with 44% of patients necessitating intervention (EBD or surgery) within 24 months.

Prior to our investigation, no data was available on vedolizumab outcomes in patients with purely stricturing CD. This is surprising as vedolizumab has shown efficacy in moderate to several luminal CD¹ and is approved for this indication. In a retrospective cohort study assessing the real-world efficacy of vedolizumab in CD, patients with stricturing disease were included [19]. However, the stricturing CD subgroup comprised patients with penetrating disease and stricture-related outcomes were not assessed [19]. A recent retrospective study assessing the efficacy of vedolizumab and ustekinumab in CD patients refractory to anti-TNFs included 22 patients with stricturing CD [30]. However, no specific stricture-related outcomes were described in that subgroup of patients. In addition, characteristics such as stricture definition, stricture location and presence of concomitant penetrating disease were not available.

The outcome of vedolizumab in symptomatic stricturing CD is in line with what has been reported for ustekinumab [26], and vedolizumab, in our investigation, indicated to be efficacious in a population that was largely anti-TNF experienced. There is a belief in the IBD community that vedolizumab is not an optimal choice for CD patients with small bowel involvement and/or after they failed an anti-TNF, but our observational data indicate that vedolizumab may be an option in stricturing TI CD. The outcomes are largely comparable to an observational real-world cohort study in patients with TI strictures by our group in which any therapies were permitted [16].

One major obstacle in the field of CD was the absence of a unified definition for stricturing CD. This fostered significant heterogeneity in existing investigations [5, 6, 20, 21, 31–33], with for instance the inclusion of patients with internal penetrating disease [6, 21, 31, 34]. This impacts results, as penetrating complications are an indication for surgery, regardless of underlying stricturing disease [35]. Of note, patients with TI stricturing CD without internal penetrating disease have been found to be the most appropriate clinical trial population for fibrostenosing CD [10]. In addition, in the available literature, stricture location is heterogenous, including the colon and upper GI tract [6, 21, 31, 34, 36, 37]. In studies using radiology as inclusion and/or outcome parameter, no pre-specified scoring conventions were used or no central reading was performed. To address this shortcoming, the STAR Consortium [8, 9] generated radiologic definitions for TI strictures- the “CONSTRICT” criteria [10] and patient outcomes in TI stricturing CD per CONSTRICT have been described [16]. In this study, identical eligibility criteria and pre-specified scoring conventions were used, and central re-reading by an experienced radiologist or IUS-trained gastroenterologist at each site was performed, in order to achieve rigor, a homogenous population and ensure external validity. Patients with colonic strictures and/or internal penetrating disease were excluded [6, 37–39].This was done as most CD strictures are located in the TI [40] and this population reflects the population selected in current or future clinical trials for stricturing Crohn’s disease (NCT05013385, NCT05843578).

Surgical rates in this study were comparable to previously published anti-TNF-treated stricture cohorts. Surgical rates were 13% in our cohort compared with 16% at 1 year in STRIDENT [7], and were 22% at 24 months compared with 29% in our own previous observational cohort [16] and consistent as well with the pooled surgical rate of 27.6% at 23 months in a systematic review assessing anti-TNF outcomes [5]. This suggests that vedolizumab may be an option in stricturing CD, including in a highly biologic-refractory population, with 84% being biologic-exposed.

Anastomotic strictures were associated with a reduced risk of surgery and intervention. This may be explained by anastomotic strictures usually being shorter and hence potentially easier to treat. Stricture length was a risk factor for intervention in an observational cohort study [16]. In addition, pre-stricture bowel diameter was linked with symptom recurrence during follow-up. Pre-stricture dilation is an established risk factor for surgery in CD TI strictures [41]. In addition, intentional avoidance of repeat surgery in patients with prior resection may have possibly contributed to this association.

EBD rates seen in our vedolizumab cohort are in line with rates seen in other stricturing CD cohorts. In a study assessing ustekinumab for ileal stricturing CD, 8.1% of patients required EBD at 12 months, compared with 10% in our vedolizumab cohort [26]. In the CREOLE adalimumab-treated cohort, 2% underwent EBD by week 24 compared to 6.9% in our cohort [6]. Of note, no factors were independently associated with EBD, which may reflect either a true lack of predictors or may possibly be due to the limited sample size and number of events.

Using a definition of biologic persistence that is comparable to other real world cohorts [23, 42], vedolizumab persistence rates were overall close to vedolizumab persistence rates in general CD cohorts not limited to stricturing CD [43], although higher compared to the anti-TNF persistence rates in the stricturing CD literature [6]. While one may interpret this as better efficacy, this is unlikely: dose-escalation of vedolizumab to 4-weekly infusions occurred in 44% of patients over a median time of 9 months. As opposed to other studies [6], we did not prohibit steroid use or EBD during follow-up. Our findings however may closer reflect clinical practice in which patients may receive intermittent steroids or undergo EBD for symptom control, while continuing their therapy. A higher CRP was linked with vedolizumab discontinuation, suggesting that increased inflammation severity was less likely to respond to vedolizumab long term.

More than half of symptomatic TI stricture patients experienced initial symptom resolution upon starting vedolizumab, with more than a quarter of them showing symptom recurrence by 24 months. Co-induction with steroids may have contributed to initial symptomatic resolution in some patients, independently of vedolizumab. In addition, other factors may have contributed to symptom recurrence, including secondary loss of response to vedolizumab (which occurs in a significant number of patients [44]), but also, in theory, due to progression of fibrotic disease over time. We found that a higher obstructive symptom index and larger pre-stricture dilation at baseline predicted a higher symptom recurrence rate. This may suggest that in our refractory patient population, with high pre-stricture dilation rates and severe symptoms, vedolizumab may not be sufficient to lead to complete symptom relief for most patients.

The frequency of internal penetrating complications during follow-up was low and comparable to prior observational studies [16], confirming that progression to fistulizing disease in established strictures is low. One may wonder, if this is due to overall limited observation time or due to stricturing with internal penetrating disease being a distinct phenotype from stricturing CD altogether.

Our study has several limitations: it is not a randomized controlled prospective intervention but rather is observational in nature. The number of included patients is modest. However, the size of our population is comparable to multiple other investigations of this kind [16, 21, 31–34, 38] and this represents the first study assessing outcomes of vedolizumab in purely stricturing CD, as defined by the CONSTRICT criteria. In addition, our patients were very well phenotyped and characterized. Outcomes observed reflect real-world practice in a quaternary care IBD unit. We could not ascertain compliance to vedolizumab as we did not have access to infusion records but had to rely on medical records to analyze vedolizumab persistence. With clinical trial endpoints now being available through the STAR consortium [9, 10] and those endpoints being used in clinical trials (NCT05013385, NCT05843578), we believe that our data will support design and sample size determinations for future clinical trials.

In conclusion, vedolizumab may represent a treatment option for patients with symptomatic stricturing CD with more than half achieving initial symptom resolution and acceptable long-term outcomes and medication persistence. Future prospective studies are needed to confirm these results. These data will serve the community in clinical care and development of clinical trials.

Abbreviations

BMI

Body mass index

CD

Crohn’s disease

CI

Confidence interval

CTE

Computed tomography enterography

DWI

Diffusion-weighted imaging

EBD

Endoscopic balloon dilation

HR

Hazard ratio

IBD

Inflammatory bowel disease

IUS

Intestinal ultrasound

MRE

Magnetic resonance imaging

PROs

Patient-reported outcomes

SES-CD

Simple endoscopic score for Crohn’s disease

TNF

Tumor necrosis factor

VEDO

Vedolizumab

Authors’ contributions

SE, FR and MB conceived and designed the study. SE, CL, MB, AI, JGF, NB, RR, SS, AM, DHB, RO, SG and IG collected the data. JB performed the statistical analysis. SE, MB, FR, JB, TQ, BC, SDH, DHB, KF, JS, MR, JGF, NSG, IG and CL interpreted the data. SE drafted the manuscript. All authors critically revised the manuscript for important intellectual content. All authors read and approved the final manuscript.

Funding

This study was supported in part by the Helmsley Charitable Trust through the Stenosis Therapy and Anti-Fibrotic Research (STAR) Consortium, National Institutes of Health [R01DK123233 & R01DK132038] to F.R.

Data availability

The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.

Declarations

Ethics approval and consent to participate

This study was conducted in accordance with the Declaration of Helsinki. The Institutional Review Board at Cleveland Clinic (20–257), Mayo Clinic (20–008419), Cleveland Clinic Abu Dhabi (A-2021–008) and University of Calgary (19–2155) approved the study. The need for informed consent was waived as there was no direct contact with patients for the purpose of the study.

Consent for publication

Not applicable.

Competing interests

Conflicts of interests and potential competing interests: SE has received speaker fees from Janssen, Abbvie, Takeda, Eli Lilly and has been on the advisory board for Eli Lilly and Janssen MB has disclosed grants to institution from Helmsley Foundation and Pfizer CL has received speaker fees from Abbvie and Janssen. CL has received advisory board fees from AbbVie, Celltrion, Janssen, Lilly, Pfizer, Takeda, Pendopharm, and Ferring. AM has received royalties from Elsevier (author contributor). DHB has been on an advisory board for Janssen and Johnson & Johnson. JGF has disclosed grants to institution: The Leona M. and Harry B. Helmsley Charitable Trust, Siemens Healthineers, Alimentiv, Pfizer; Co-investigator: Motilent; Consulting with funds to institution: Janssen, Alimentiv, Genentech, Boehringer Ingelhem, GSK, Re X Pharma Takeda AI: has no conflict of interest. MR receives the following financial support: Advisory Boards and Consultant (both) for Abbvie, Janssen, UCB, Takeda, Pfizer, BMS, Organon,Amgen, Genentech, Gilead, Salix, Prometheus, Lilly, Celgene, Boehringer Ingelheim Pharmaceuticals Inc. (BIPI), Celltrion, Roche RKO has no conflict of interest. SS has been on an advisory board for Abbvie. RR has no conflict of interest. JB has no conflict of interest. TQ has received the following financial support: Advisory Board and Speakers fee for AbbVie, Celgene-Bristol Meyers Squibb, Janssen, Prometheus Biosciences. Educational Grants: AbbVie, Pfizer, Eli Lilly and Janssen SDH has received speaker fees from Takeda, and research funding from the Crohn’s & Colitis Foundation, and the American Society of Colon & Rectal Surgeons. SG has no conflict of interest. NSG has no conflict of interest. KF receives grant funding from the Crohn’s and Colitis Foundation (916943) and Pfizer (90146787). She served on an advisory panel for Janssen and is providing content for Takeda IBDIQ. JS has no conflict of interest. BLC receives the following financial support: Advisory Boards and Consultant for Abbvie, Celgene-Bristol Myers Squibb, Johnson & Johnson Innovative Medicine, Lilly, Pfizer, Sublimity Therapeutics, Takeda, TARGET RWE; DSMB: Emmes Biopharma Services LLC; Speaking: Abbvie and Takeda IG has no conflict of interest. NB has no conflict of interest. FR is on the advisory board or a consultant for Adiso, Adnovate, Agomab, Allergan, AbbVie, Arena, Astra Zeneca, Bausch & Lomb, Boehringer-Ingelheim, Celgene/BMS, Celltrion, CDISC, Celsius, Cowen, Eugit, Ferring, Galapagos, Galmed, Genentech, Gilead, Gossamer, Granite, Guidepoint, Helmsley, Horizon Therapeutics, Image Analysis Limited, Index Pharma, Landos, Jannsen, Koutif, Mestag, Metacrine, Mirum, Mopac, Morphic, Myka Labs, Organovo, Origo, Palisade, Pfizer, Pliant, Prometheus Biosciences, Receptos, RedX, Roche, Samsung, Sanofi, Surmodics, Surrozen, Takeda, Techlab, Teva, Theravance, Thetis, Trix Bio, UCB, Ysios, 89Bio.