Real-world inflammatory bowel disease-modification outcomes: A narrative review

Abstract

Inflammatory bowel diseases (IBDs) culminate in disease progression in many patients. Characterized by chronic inflammation and gastrointestinal damage, IBD negatively impacts quality of life, often requires surgery, and can result in disability. Disease modification has been a target for treatment for other chronic inflammatory diseases, and disease-modification trials will help determine if this approach can prevent disease progression in IBD. The standard protocol items: Recommendations for interventional trials (SPIRIT) consensus has recommended disease-modification endpoints for use in clinical trials, but before such studies are conducted, real-world data can be informative. This review of real-world studies in patients with ulcerative colitis or Crohn’s disease treated with advanced treatments that employed the SPIRIT endpoints identified considerable gaps in the use of recommended disease-modification endpoints. Specifically, no studies used the recommended outcomes for midterm complications of fecal incontinence or bowel damage, or assessed macroscopic proximal disease extension, permanent stoma, or short-bowel symptoms. Long-term complications of cancer or mortality were not assessed over the recommended 5-year period. Additionally, studies have not assessed quality of life using the recommended composite of the 36-Item IBD Questionnaire and 36-Item Short-Form Health Survey. In conclusion, there remains a need to incorporate SPIRIT-recommended disease-modification endpoints in prospective real-world studies in patients with IBD receiving advanced treatments.

Key Words: Crohn’s disease; Disease modification; Inflammatory bowel disease; Real-world evidence; Standard protocol items: Recommendations for interventional trials consensus; Ulcerative colitis

Core Tip: The standard protocol items: Recommendations for interventional trials (SPIRIT) consensus has recommended disease-modification endpoints for use in clinical trials in patients with inflammatory bowel disease (IBD) to help determine if this approach can prevent disease progression. Before such studies are conducted, real-world evidence can be informative. This review of real-world studies utilizing the SPIRIT consensus endpoints identified considerable gaps in the use of these recommended disease-modification endpoints. Therefore, there remains a need to incorporate SPIRIT recommended disease-modification endpoints in prospective real-world studies in patients with IBD receiving advanced treatments.

INTRODUCTION

The inflammatory bowel diseases (IBDs) ulcerative colitis (UC) and Crohn’s disease (CD) are lifelong diseases that culminate in disease progression in many patients[1,2]. IBD, characterized by chronic inflammation and damage to the gastrointestinal tract, has a negative impact on health-related quality of life (HRQoL), often requires surgery, and can result in disability[1,3,4]. The estimated prevalence in Western countries has stabilized in recent years but exceeds 0.3% in Western countries, is > 0.7% in the United States, and is growing in newly industrialized African, Asian, and South American countries[5-7].

Biologic therapies investigated in clinical trials for patients with IBD include anti-tumor necrosis factor (TNF)-α therapies (e.g., infliximab, adalimumab, golimumab), anti-interleukin (IL)-12 and IL-23 (e.g., ustekinumab), anti-integrins (e.g., vedolizumab), Janus kinase inhibitors (e.g., tofacitinib, filgotinib, upadacitinib), and sphingosine-1-phosphate receptor modulators (e.g., ozanimod)[8].

Due to regulatory agency requirements, clinical trials typically evaluate short-term outcomes at 1 year, such as clinical and endoscopic remission, as recommended by the selecting therapeutic targets in IBD (STRIDE) II consensus (Table 1)[9]. Disease modification has been the focus of treatment for other progressive diseases (e.g., rheumatoid arthritis, multiple sclerosis, Alzheimer’s, osteoarthritis) and is defined as interventions or treatments targeting the underlying disease pathophysiology, with long-term beneficial outcomes for patients such as symptom improvement/reduced disease progression and disease modification to improve the disease course[10-15]. A “treat-to-target” approach in IBD is often used, involving regular assessment of disease activity and adjustment of treatment based on predefined therapeutic objectives to prevent long-term bowel damage and disability[16]. Future disease-modification trials will determine if this early disease control approach can prevent disease progression in IBD[8,16]. Disease-modification outcomes are not included in the STRIDE II consensus[9]; however, the standard protocol items: Recommendations for interventional trials (SPIRIT) consensus, based on expert opinion, has recommended disease-modification endpoints for use in clinical trials to assess the impact of treatment on quality of life, disability, disease complications, and the long-term risks of cancer and death (Table 1)[16]. There remains a need to validate these endpoints in prospective clinical trials assessing disease modification[16]. The ongoing VERDICT trial in patients with UC is evaluating optimized treatment targets for disease modification in UC, including the combination of symptomatic, endoscopic, and histologic remission known as disease clearance. Before such studies validate the evidence for disease modification, real-world data using these endpoints can be informative. We reviewed real-world studies employing disease-modification endpoints, as recommended by the SPIRIT consensus, in patients with UC and CD treated with advanced therapies.

Table 1 Outcome measures recommended by the selecting therapeutic targets in inflammatory bowel disease II and standard protocol items: Recommendations for interventional trials consensus for use in inflammatory bowel disease-modification trials.

Regulatory endpoints: STRIDE II	Disease modification endpoints: SPIRIT
Intermediate targets	Outcome measures
	Midterm complications
Clinical response [at least 50% decrease in PRO2 (CD: Abdominal pain and stool frequency; UC: Rectal bleeding and stool frequency)]	Bowel damage in patients with CD (Lemann Index; 12-24 months)
Clinical remission [CD: PRO2 (abdominal pain ≤ 1 and stool frequency ≤ 3) or HBI < 5; UC: PRO2 (rectal bleeding = 0 and stool frequency = 0) or partial Mayo < 3 and no sub-score > 1]	IBD-related surgery [any colectomy (UC), any CD-related surgery, endoscopic balloon dilation, perianal surgery (CD); 24-36 months]
Normalization of C-reactive protein (values under the upper limit of normal) and fecal calprotectin (to 100-250 μg/g)	IBD-related hospitalizations (composite of number of hospitalizations and cumulative length of hospital stay; 12-24 months)
Long-term targets	Disease extension in UC (excluding patients with pancolitis) (macroscopic proximal disease extension; 2-5 years after diagnosis)
Restoration of normal growth in children	Extraintestinal manifestations (12-36 months)
Endoscopic healing (CD: SES-CD < 3 points or absence of ulcerations; UC: Mayo endoscopic sub-score = 0 points or UCEIS ≤ 1 point)	Permanent stoma
Absence of disability and normalization of HRQoL	Short-bowel syndrome (small-bowel length < 200 cm)
Adjunctive targets to endoscopic remission	Long-term complications
Histological remission	Dysplasia or cancer (5 years)
Transmural healing	Mortality (both IBD-related and non-IBD related mortality; 5 years)
	Impact on patient’s life
	Health-related quality of life (tool: Composite of IBDQ-36 and SF-36; every 6-12 months)
	Disability (IBD Disability Index; every 6-12 months)
	Fecal incontinence (excluding patients with isolated liminal ileal CD) (Cleveland score or Jorge-Wexner score; every 6-12 months)

STRIDE: Selecting therapeutic targets in inflammatory bowel disease; SPIRIT: Standard protocol items: Recommendations for interventional trials; PRO2: Patient-reported outcomes 2; CD: Crohn’s disease; UC: Ulcerative colitis; HBI: Harvey-Bradshaw index; SES-CD: Simple endoscopic score for Crohn’s disease; UCEIS: Ulcerative colitis endoscopic index of severity; HRQoL: Health-related quality of life; IBD: Inflammatory bowel disease; IBDQ-36: 36-Item Inflammatory Bowel Disease Questionnaire; SF-36: 36-Item Short-Form Health Survey.

METHODOLOGY

For this narrative review, PubMed was searched between April 2021 (publication of the SPIRIT consensus) and June 2024 to identify real-world studies of patients with UC or CD receiving advanced therapies that used disease-modification endpoints as recommended by the SPIRIT consensus (Table 1). The key search terms used were “inflammatory bowel disease”, “IBD”, “ulcerative colitis”, “UC”, “Crohn’s disease”, “CD”, and the search was limited to articles published in the English language. Abstracts of identified articles were screened by two reviewers, and real-world, observational studies that reported data on the SPIRIT recommended disease-modification endpoints in patients with UC or CD treated with advanced therapies were included. Clinical trials, case reports/series, review articles, and studies of patients not receiving advanced therapies were excluded.

NARRATIVE REVIEW

Results from the literature search of recent real-world studies are shown in Table 2. Endpoints used in real-world studies deviated significantly from those in the SPIRIT consensus, with alternative definitions using different timeframes, shorter questionnaires and not using composite endpoints. No studies used the Lemann Index for bowel damage in patients with CD, macroscopic proximal disease extension in UC, permanent stoma, or short-bowel syndrome (Table 3).

Table 2 Studies using standard protocol items: Recommendations for interventional trials disease-modification endpoints found in initial literature search[16-34,37,39,41-49,51-54,56,75].

Ref.	Treatment	Patient population	Endpoints	Key results
Midterm complications
IBD-related hospitalizations
D’Amico et al[17]	Not restricted (included infliximab, adalimumab, vedolizumab)	All patients with UC (n = 156); LOT not mentioned	IBD-related hospitalizations (median follow-up, 30.5 months)	Significantly more patients with histological disease activity were hospitalized than those with histological remission (36.0% vs 7.1%)
Ferrante et al[18]	Golimumab	Patients with UC (n = 109); biologic naive (n = 83) and biologic experienced (second-line biologic; n = 26)	UC-related hospitalizations (54 weeks)	UC-related hospitalizations (through week 54): 3.7% (4/109). Missing information: 61.5% (67/109)
Meyer et al[75]	Vedolizumab vs ustekinumab	Patients with UC (n = 218); anti-TNF experienced	IBD-related hospitalizations (52 weeks)	At week 52, there were no differences between vedolizumab and ustekinumab for hospitalization (9.0% vs 12.1%; OR = 1.44; 95%CI: 0.48-4.37; P = 0.52)
Adimadhyam et al[27]	Vedolizumab vs tofacitinib	Patients with UC (n = 956); anti-TNF experienced	All-cause hospitalizations	All-cause hospitalization was similar between tofacitinib and vedolizumab cohorts (adjusted HR = 1.23; 95%CI: 0.83-1.84)
Greener et al[22]	Adalimumab	Patients with CD (n = 79); anti-TNF experienced	IBD-related hospitalizations	Of 79 patients included, patients receiving higher maintenance doses of adalimumab (40 mg weekly, ≥ 80 mg every other week; n = 39) had a numerically lower number of IBD-related hospitalizations (21% vs 38%, P = 0.08) than those receiving a standard maintenance regimen of adalimumab (40 mg every other week; n = 40)
Narula et al[23]	Adalimumab	Patients with CD (n = 117); anti-TNF experienced	IBD-related hospitalizations	Patients receiving higher maintenance doses of adalimumab (40 mg weekly, ≥ 80 mg every other week; n = 40) had a numerically lower number of IBD-related hospitalizations [8/40 (20%) vs 28/77 (36.3%), P = 0.09] than those receiving a standard maintenance regimen of adalimumab (40 mg every other week; n = 77)
Kappelman et al[29]	Vedolizumab vs ustekinumab	Patients with CD (n = 1375); anti-TNF experienced	IBD-related hospitalizations with surgery and hospitalizations without surgery (52 weeks)	885 new users of ustekinumab had fewer nonsurgical CD hospitalizations than 490 new users of vedolizumab [adjusted HR = 0.58 (0.40-0.83)]. There was a numerical trend toward fewer surgical CD hospitalizations [adjusted HR = 0.83 (0.57-1.22)]
Dulai et al[34]	Vedolizumab, anti-TNFs	Patients with CD (n = 1803); biologic naive and anti-TNF experienced	CD-related hospitalizations and surgery (12 months)	Patients were stratified into high, intermediate, and low probabilities of response in the Optum dataset (n = 358) and high and low probabilities of response in the Truven dataset (n = 1445). In the Optum dataset, rates of hospitalization were 2-4 times lower for vedolizumab-treated patients with a high probability of response vs those with a low probability of response (hospitalization, 19.0% vs 48.1%, P < 0.001). In the Truven dataset, a significantly lower proportion of the high probability of response group experienced a CD-related hospitalization (14.4% vs 19.6%, P = 0.011) during the 12 months after vedolizumab initiation
Narula et al[24]	Adalimumab	Patients with CD (n = 61); biologic naive	CD-related hospitalizations	At week 48, 11/61 (18.0%) patients had a CD-related hospitalization. Of the 11 patients with CD-related hospitalizations, 8 patients had 1 CD-related hospitalization and 3 patients had 2 CD-related hospitalizations
Lenti et al[28]	Vedolizumab vs ustekinumab	Patients with CD (n = 400); anti-TNF experienced	IBD-related hospitalizations (52 weeks)	There was a significant risk difference between ustekinumab (n = 282) and vedolizumab (n = 118) in the unmatched cohort (risk difference, 12; 95%CI: 3-20; P = 0.01), which was no longer significant in the matched population (risk difference, 9; 95%CI: -4 to 22; P = 0.20)
Singh et al[31]	Vedolizumab, ustekinumab, or TNF-α antagonists	Patients with CD (n = 2965); anti-TNF experienced	All-cause hospitalizations	No difference in the risk of hospitalization between ustekinumab vs TNF-α antagonists (HR = 0.99; 95%CI: 0.89-1.21), ustekinumab vs vedolizumab (HR = 0.76; 95%CI: 0.54-1.07), or vedolizumab vs TNF-α antagonists (HR = 1.32; 95%CI: 0.98-1.77)
Abraham et al[25]	Infliximab-dyyb	Patients with IBD (n = 115); biologic naive (n = 39) and biologic experienced (n = 76)	IBD-related hospitalizations (12 months)	In the overall population, IBD-related hospitalizations were recorded in 9.6% (n = 11/115) of patients at baseline and 1.2% (n = 1/84) within the 12-month observation period. Of biologic-naive patients, this was 6 (15.4%) and 0 (0%) at baseline and 12 months, respectively. No patients who switched from infliximab to the infliximab biosimilar had an IBD-related hospitalization
Alharbi et al[26]	Infliximab	Patients with IBD (n = 86; UC, n = 24; CD, n = 62); received ≤ 2 biologic agents	Hospitalizations	Over 2 years, the mean number of hospitalization days was 1.1 days for patients with CD and was 0.7 days for patients with UC
Rundquist et al[32]	Vedolizumab vs anti-TNF-α	Patients with IBD (n = 400; CD, n = 198; UC, n = 202); anti-TNF experienced (second-line biological)	IBD-related hospitalizations (12 months)	After propensity score matching, vedolizumab-treated patients with UC had significantly lower drug survival without IBD-related hospitalization after a 12-month follow-up compared with anti-TNF-treated patients (82% vs 93%, P = 0.02). There were no significant differences between treatments for drug survival without IBD hospitalization in patients with CD (88% vs 82%, P = 0.24)
Surgery
D’Amico et al[17]	Not restricted (included infliximab, adalimumab, vedolizumab)	All patients with UC (n = 156); LOT not mentioned	IBD-related surgery (median follow-up, 30.5 months)	Surgery, n = 16; hospitalized, n = 44; mean (SD) hospitalization duration of 8.97 (8.55) days. 16 patients with histological disease activity at baseline underwent surgery compared with no patients in the histological remission group (14% vs 0%)
Panés et al[19]	Infliximab, conventional therapy	Patients with UC (n = 2239); LOT not mentioned	Surgery (colectomy) (5-year follow-up)	During 5 years of follow-up, 271 patients had colectomy: 174/1059 (16.4%) in the infliximab group and 48/1180 (4.1%) in the conventional therapy group
Meyer et al[75]	Vedolizumab vs ustekinumab	Patients with UC (n = 218); anti-TNF experienced	IBD-related surgery (52 weeks)	At week 52, there were no differences between vedolizumab and ustekinumab UC-related surgery (8.8% vs 12.3%; OR = 2.42; 95%CI: 0.77-7.59; P = 0.13) outcomes
Adimadhyam et al[27]	Vedolizumab vs tofacitinib	Patients with UC (n = 956); anti-TNF experienced	Surgery (colectomy)	There were no significant differences between the secondary outcome of colectomy between new users of tofacitinib (n = 234) and new users of vedolizumab (n = 722) for the secondary outcome of colectomy (incidence rate per 1000 person-years: Tofacitinib, 97.17; Vedolizumab, 59.25; Adjusted HR = 1.79; 95%CI: 0.93-3.44)
Hirai et al[20]	Adalimumab, infliximab, or ustekinumab	Patients with CD (n = 1346); biologic naive (first-line biologic treatment)	Surgery any time after biologic introduction; time to first surgery	292/1346 (21.7%) patients had a surgery any time after biologic initiation. The mean (SD) time from biologic introduction to first surgery was 21.8 (20.8) months
Greener et al[22]	Adalimumab	Patients with CD (n = 79); anti-TNF experienced	IBD-related surgery	Patients receiving the standard maintenance doses (40 mg every other week; n = 40) had a higher risk of surgery than those receiving higher maintenance doses (40 mg weekly, ≥ 80 mg every other week; n = 39) of adalimumab (18% vs 3%, P = 0.03)
Narula et al[23]	Adalimumab	Patients with CD (n = 117); anti-TNF experienced	Surgery	There was no difference between patients receiving higher maintenance doses of adalimumab (40 mg weekly, ≥ 80 mg every other week; n = 40) and those receiving a standard maintenance regimen of adalimumab (40 mg every other week; n = 77) in the number of surgeries [4/77 (5.1%) vs 1/40 (2.5%), P = 0.65]
Vu et al[30]	Vedolizumab or ustekinumab	Patients with CD (n = 1122); biologic naive	IBD-related surgery	After 1-year of follow-up, there was a significantly lower rate of CD-related surgeries in vedolizumab-treated (n = 578) vs ustekinumab-treated (n = 544) patients (7.7% vs 11.6%; HR = 0.67; 95%CI: 0.44-0.99; P = 0.047)
Dulai et al[34]	Vedolizumab, anti-TNFs	Patients with CD (n = 1803); biologic naive and anti-TNF experienced	CD-related surgery (12 months)	Patients were stratified into high, intermediate, and low probabilities of response in the Optum dataset (n = 358) and high and low probabilities of response in the Truven dataset (n = 1445). In the Optum dataset, rates of surgery were 2-4 times lower for vedolizumab-treated patients with a high probability of response vs those with a low probability of response (8.4% vs 44.4%, P < 0.001). In the Truven dataset, a significantly lower proportion of the high probability of response group experienced a CD-related surgery (13.5% vs 23.3%, P < 0.001) during the 12 months after vedolizumab initiation
Narula et al[24]	Adalimumab	Patients with CD (n = 61); biologic naive	CD-related surgery	At week 48, 7/61 (11.5%) had a CD-related surgery
Singh et al[31]	Vedolizumab, ustekinumab, or TNF-α antagonists	Patients with CD (n = 2965); anti-TNF experienced	IBD-related surgery	There were no differences in the 1-year risk of IBD-related surgery for the comparisons of ustekinumab and vedolizumab [propensity score-matched cohort HR = 1.42 (0.54-3.72)], vedolizumab and anti-TNF antagonists [propensity score matched cohort HR = 0.63 (0.27-1.47)], or ustekinumab and anti-TNF antagonists [propensity score matched cohort HR = 1.08 (0.69-1.70)]
Bressler et al[33]	Vedolizumab and anti-TNF-α	Patients with IBD (n = 1095; CD, n = 491; UC, n = 604); biologic naive	IBD-related surgery	There was no difference in disease-related surgeries between vedolizumab and anti-TNF cohorts in patients with CD (incidence rate, 0.7 vs 1.5 per 100 years). Only 1 patient with UC had a colectomy (anti-TNF group)
Rundquist et al[32]	Vedolizumab vs anti-TNF-α	Patients with IBD (n = 400; CD, n = 198; UC, n = 202); anti-TNF experienced (second-line biological)	IBD-related surgery (12 months)	After propensity score matching, there were no significant differences between vedolizumab and anti-TNF treatments for drug survival without IBD-related surgery in patients with UC (93% vs 97%, P = 0.21) or CD (89% vs 82%, P = 0.17)
Clinical relapse
Park et al[21]	Ustekinumab	Patients with CD (n = 99); anti-TNF experienced	Clinical relapse assessed via treatment failure, defined as the need for additional corticosteroid use, withdrawal of ustekinumab, need for CD-related surgery, or CD-related hospitalization	11/99 (11.1%) patients needed a CD-related surgery and 22/99 (22.2%) patients needed a CD-related hospitalization
EIMs
Eder et al[41]	Vedolizumab	Patients with UC (n = 100); biologic naive (n = 55) and biologic experienced (n = 45)	EIMs	EIMs were reported in 13/100 (13.0%) patients with UC at baseline, and 12/91 (13.2%), 3/62 (4.8%), and 3/61 (4.9%) at weeks 14, 54, and 80, respectively. The most frequently reported extraintestinal symptom was arthralgia (n = 11 at week 14 vs n = 3 at week 54)
Straatmijer et al[39]	Ustekinumab	Patients with CD (n = 252); LOT not mentioned	EIMs (2 years)	59/252 patients with CD who were treated with ustekinumab had EIMs at baseline. All patients with uveitis, aphthous, stomatitis, and pyoderma gangrenosum achieved EIM remission and 1 of the 2 patients with erythema nodosum achieved EIM remission during the 2-year follow-up period
Ferretti et al[42]	Vedolizumab vs non-gut-selective biologics	Patients with IBD (n = 1182); LOT not mentioned	EIMs (3-4 years)	224 patients with new or preexisting EIMs were included, who received ongoing biologic for a median of 3 (vedolizumab) and 4 (non-gut-selective agents) years. Patients treated with vedolizumab experienced a slightly higher rate of clinical worsening than those treated with non-gut-selective therapies (15.5% vs 7.3%, P = 0.08)
Livne-Margolin et al[43]	Vedolizumab vs ustekinumab	Patients with IBD (n = 111); LOT not mentioned	EIMs	Clinical response of EIM at week 52 was achieved in 36% (18/50) of ustekinumab-treated patients and 34% (19/54) of vedolizumab-treated patients (P = 0.9)
Kopylov et al[44]	Vedolizumab	Patients with IBD (n = 99; CD, n = 55; UC, n = 44); LOT not mentioned	EIMs (12 months)	In patients with moderately to severely active IBD, within 12 months of vedolizumab initiation 25.3% of patients reported resolution of all EIMs, and 49.5% reported improvement of EIMs
Long-term complications
Death
D’Amico et al[17]	Various (included infliximab, adalimumab, vedolizumab)	Patients with UC (n = 156); LOT not mentioned	Deaths (median follow-up, 30.5 months)	No deaths occurred
Macaluso et al[47]	Adalimumab biosimilar	All patients with IBD (n = 559); biologic naive and biologic experienced	Deaths	No deaths occurred
Cancer
Singh et al[45]	Vedolizumab vs TNF-α antagonists	Patients with IBD (n = 5566); biologic naive	Cancer	There were no differences in the risk of incident malignancy between patients treated with vedolizumab (n = 759) vs those treated with TNF-α antagonists (n = 4807) (incidence rate, 9.0 vs 11.6 per 1000 patient-years; incidence rate ratio, 1.28; 95%CI: 0.61-2.45)
Macaluso et al[47]	Adalimumab biosimilar	All patients with IBD (n = 559); biologic naive and biologic experienced	Neoplasia	No neoplasia occurred
Vedamurthy et al[46]	Vedolizumab and anti-TNF-α	Patients with IBD (n = 463); LOT not mentioned	New or recurrent cancer (median follow-up, 6.2 years)	Incidence rates of new or recurrent cancer following an index cancer diagnosis were 22, 4.2, and 5.6 per 1000 person-years in the vedolizumab, anti-TNF, and no immunosuppression groups, respectively. Neither vedolizumab [HR = 0.72 (0.38-1.36)] nor anti-TNF therapy [HR = 1.03 (0.65-1.64)] had an increased risk of new or recurrent cancer compared with no immunosuppression
Impact on patient’s life
IBDQ
Gatopoulou et al[49]	Golimumab	Patients with UC (n = 81); anti-TNF naive	IBDQ-32 remission (12 months), IBDQ response (6 and 12 months)	IBDQ-32 remission was achieved by 76.9% of patients at 12 months of treatment. The proportion of patients achieving IBDQ-32 response at 6 months and 12 months was 73.0% (46/63) and 75.0% (39/52), respectively
Karami et al[51]	Adalimumab vs infliximab	Patients with UC (n = 238); LOT not mentioned	IBDQ-9	Mean (SD) IBDQ-9 scores were similar between the adalimumab [37.42 (10.72); n = 160] and infliximab [36.61 (11.12); n = 78] treatment groups
Mendonça et al[53]	Various	Patients with IBD (n = 56); LOT not mentioned	IBDQ-32 (cross-sectional)	Total IBDQ-32 scores were significantly lower in females than in males. The difference in IBDQ-32 scores between males and females was larger in patients with CD than in those with UC
Alharbi et al[26]	Infliximab	Patients with IBD (n = 86; CD, n = 62; UC, n = 24); received ≤ 2 biologic agents	IBDQ	This study was terminated early due to low patient numbers. The mean (SD) change from baseline to 3 months in IBDQ score was 11.3 (39.6) points
IBD-DI
Ruiz-Casas et al[56]	Various	Patients with UC (n = 2966); LOT not mentioned	IBD-DI (12 months); IBD-Control Questionnaire	The IBD-DI score was 30.6 in patients with moderate or severe UC (n = 1000) and 22.3 in patients with moderate or severe UC who achieved mild UC or remission status 12 months before the index date (n = 647). The IBD-Control score (United Kingdom only) was a mean 7.2 overall and was similar between cohorts (7.1 vs 7.3)
SIBDQ
Bamias et al[48]	Vedolizumab	Patients with UC (n = 96); LOT not mentioned	SIBDQ (weeks 14 and 54)	Vedolizumab-treated patients with UC reported significant improvements in SIBDQ scores from baseline (mean score, 44.9) to weeks 14 (mean score, 55.2; P < 0.001) and 54 (mean score, 56.5; P < 0.001)
Matsuoka et al[52]	Vedolizumab	Patients with UC (safety, n = 268); anti-TNF experienced	SIBDQ (54 weeks)	There was an increase in mean (SD) SIBDQ score from 44.2 (11.89) at baseline (n = 223) to 53.3 (11.05) after vedolizumab induction (n = 210), with a mean (SD) change (n = 204) of 8.9 (12.53)
Abraham et al[25]	Infliximab-dyyb	Patients with IBD (n = 115; CD, n = 67; UC, n = 48); biologic naive (n = 39) and biologic experienced (n = 76)	SIBDQ (12 months)	There were significant improvements in SIBDQ scores from baseline to 12 months in patients with CD and patients with UC who were treated with infliximab-dyyb (all P < 0.05)
SF-36
Bellone et al[54]	Vedolizumab and infliximab	Patients with IBD (n = 50); LOT not mentioned	SF-36 (14 weeks)	There was a significant improvement from baseline to 14 weeks of treatment in all domains of the SF-36

TNF: Tumor necrosis factor; IBD: Inflammatory bowel disease; EIM: Extraintestinal manifestation; IBD-DI: Inflammatory Bowel Disease Disability Index; IBDQ: Inflammatory Bowel Disease Questionnaire; SIBDQ: Short Inflammatory Bowel Disease Questionnaire; SF-36: 36-Item Short-Form Health Survey; CD: Crohn’s disease; UC: Ulcerative colitis; LOT: Line of therapy; IBDQ-32: 32-Item Inflammatory Bowel Disease Questionnaire; IBDQ-9: 9-Item Inflammatory Bowel Disease Questionnaire; CI: Confidence interval; HR: Hazard ratio; OR: Odds ratio.

Table 3 Summary of differences between standard protocol items: Recommendations for interventional trials disease-modification endpoints and endpoints used in real-world studies.

SPIRIT endpoints		Measure	Timeframe	Number of publications using SPIRIT definition	Number of publications using alternative definition	Difference from SPIRIT
Midterm	Bowel damage in patients with CD	Lemann Index	12-24 months	0	0
	IBD-related surgery	Any colectomy (UC), any CD-related surgery, endoscopic balloon dilation, perianal surgery (CD)	24-36 months	8	5	Different timeframes used
	IBD-related hospitalizations	Composite of number of hospitalizations and cumulative length of hospital stay	12-24 months	0	14	No length of stay reported
	Disease extension in UC1	Macroscopic proximal disease extension	2-5 years after diagnosis	0	0
	Extraintestinal manifestations		12-36 months	5	0
	Permanent stoma			0	0
	Short-bowel syndrome	Small-bowel length < 200 cm		0	0
Long-term	Dysplasia or cancer		5 years	0	3	No 5 year follow up
	Mortality	IBD and non IBD related	5 years	0	2	No 5 year follow up
Impact of patients’ life	Health-related quality of life	Composite of IBDQ-36 and SF-36	Every 6-12 months	0	8	Shorter forms and no composite use
	Disability	IBD Disability Index	Every 6-12 months	1	0
	Fecal incontinence2	Cleveland score or Jorge-Wexner score	Every 6-12 months	0	0

¹Excluding patients with pancolitis.

²Excluding patients with isolated liminal ileal Crohn’s disease.

SPIRIT: Standard protocol items: Recommendations for interventional trials; CD: Crohn’s disease; IBD: Inflammatory bowel disease; UC: Ulcerative colitis. IBDQ: Inflammatory Bowel Disease Questionnaire; SF-36: 36-Item Short-Form Health Survey.

Midterm complications

The midterm complications defined in the SPIRIT consensus include bowel damage in patients with CD, IBD-related surgery, IBD-related hospitalizations, disease extension in UC, extraintestinal manifestations (EIMs), permanent stoma and short-bowel syndrome (Table 1).

In total, 24 studies (UC, n = 6; CD, n = 11; IBD, n = 7) reporting midterm complications were identified: The majority were IBD-related surgery or IBD-related hospitalization rates. No studies reported bowel damage, disease extension in UC, permanent stroma or short-bowel syndrome. Of these 24 studies, eight included biologic-naive patients (UC, n = 2; CD, n = 4; IBD, n = 2), 14 included anti-TNF experienced patients (UC, n = 4; CD, n = 7; IBD, n = 3), and six did not mention the line of treatment (UC, n = 2; CD, n = 1; IBD, n = 3), with four studies including biologic-naive and anti-TNF experienced patients.

IBD-related hospitalizations and IBD-related surgery

IBD-related hospitalizations or IBD-related surgery rates were frequently recorded in the studies, mostly as secondary outcomes; however, none calculated the composite of the number of IBD-related hospitalizations and cumulative length of hospital stay as recommended by the SPIRIT consensus. The three observational studies in patients with UC included an observational, retrospective cohort study of consecutive patients (n = 156) who underwent colonoscopy with biopsies and had histological reports available over a median 30-month follow-up[17]. Patients were monitored over time to assess the onset of negative disease symptoms, including escalation of medical therapy (initiation of systemic corticosteroids, immunomodulators, or biologics; dose or interval optimization; switch to another drug class), UC-related hospitalization, need for UC-related surgery, development of colorectal dysplasia or malignancy, and all-cause death. The most frequently used drugs were oral aminosalicylates (89.7%), systemic steroids (67.3%), rectal aminosalicylates (53.8%), infliximab (53.2%), thiopurines (48.1%), adalimumab (37.2%), and vedolizumab (34.6%). Sixteen patients underwent surgery and 44 were hospitalized, with a mean (SD) hospitalization duration of 9.0 (8.6) days. Sixteen (14%) patients with histological disease activity at baseline underwent surgery compared with none in the histological remission group. A similar trend was seen for hospitalizations, with significantly more patients with histological disease activity being hospitalized than those with histological remission (36.0% vs 7.1%). The second study was a 54-week, real-world European observational study, GO OBSERVE, conducted in patients with moderate to severe UC (n = 109) who were biologic naive or received ≤ 1 advanced therapy[18]. Secondary effectiveness analyses included the proportion of patients with UC-related hospitalizations (through week 54) and UC-related surgery (through week 54). The numbers of both outcomes were low [UC-related hospitalization, 3.7% (4/109); UC-related surgery, 0% (0/109)]; however, there was a substantial amount of missing patient information for each of these endpoints [61.5% (67/109) and 65.1% (71/109), respectively]. In the third study, rates of colectomy were reported as part of the 5-year data from the Observational Postmarketing UC Study (OPUS) registry, which investigated the safety of infliximab in European patients with moderate to severe UC[19]. A prespecified secondary analysis of OPUS was to compare colectomy rates in patients treated with infliximab (n = 1059) or conventional therapy (n = 1180). During 5 years of follow-up, 271 patients had colectomy: 174 (16.4%) in the infliximab group and 48 (4.1%) in the conventional therapy group. The majority of colectomy events occurred within the first 12 months.

The five observational studies in patients with CD included an observational longitudinal study in Japan using claims data to describe outcomes in patients who were prescribed ≥ 1 advanced therapy[20]. A secondary outcome in this study was time from advanced therapy initiation to first surgery, which showed of the 1346 patients included in the study, 292 (21.7%) patients had surgery any time after biologic initiation. The mean (SD) time from biologic introduction to first surgery was 21.8 (20.8) months. The second study in patients with CD was a multicenter observational study enrolling Korean patients with moderate to severe CD (n = 99) who were refractory to conventional therapies and who demonstrated clinical response after induction therapy with ustekinumab[21]. CD-related surgery and CD-related hospitalizations during ustekinumab maintenance therapy were individual components of the composite primary outcome (clinical relapse defined as a therapeutic failure including additional corticosteroid use, withdrawal of ustekinumab, and need for a CD-related surgery or CD-related hospitalization, developing after the initial clinical response). Eleven (11.1%) patients needed CD-related surgery and 22 (22.2%) needed CD-related hospitalization.

Two retrospective observational studies of patients with CD (n = 79[22] and n = 117[23]) reported IBD-related hospitalizations as a secondary outcome. In both, patients receiving higher maintenance doses of adalimumab (40 mg weekly, ≥ 80 mg every other week) had fewer IBD-related hospitalizations than those receiving a standard maintenance regimen of adalimumab (40 mg every other week). Finally, CD-related hospitalizations and surgeries were assessed as individual components of the composite primary endpoint (disease progression was defined as any new internal fistula or abscess, stricture, perianal fistula or abscess, CD-related hospitalization, or CD-related surgery) in the effect of tight control management on CD (CALM) open-label, phase 3, long-term extension study comparing tight control with conventional strategies among biologic-naive patients with moderately to severely active CD who were treated with adalimumab[24]. Of the 121 participants, 61 had baseline and end-of-study ileocolonoscopy videos available. Eleven of 61 (18.0%) patients needed CD-related hospitalization at week 48, eight (72.7%) of whom had one hospitalization and three (27.3%) of whom had two hospitalizations. CD-related surgeries were needed in seven of 61 (11.5%) patients. Even though new internal fistula, internal abscess, intestinal strictures, perianal disease, or new perianal abscesses are not included in the STRIDE consensus, they occurred in five of 61 (8.2%), three of 61 (4.9%), six of 61 (9.8%), and two of 61 (3.3%) patients, respectively, in the CALM long-term extension and could be clinically relevant additions to describe disease progression.

Two observational studies included patients with CD and UC. A prospective observational study reported IBD-related hospitalizations as a secondary outcome in patients treated with the infliximab biosimilar infliximab-dyyb[25]. Thirty-nine patients were biologic naive, 57 switched from infliximab to infliximab-dyyb, and 19 switched from other biologics. IBD-related hospitalizations were recorded in 9.6% (n = 11/115) of patients at baseline, 5.5% (n = 6/109) at 3 months, 1.0% (n = 1/99) at 6 months, and 1.2% (n = 1/84) at 12 months. In biologic-naive patients, this was six (15.4%), two (5.4%), one (3.0%), and zero (0%) at baseline and 3 months, 6 months, and 12 months, respectively. In patients who switched from infliximab to infliximab-dyyb, 5.5% (3/109) of patients had an IBD-related hospitalization at 3 months, whereas none had an IBD-related hospitalization at baseline or at 6 months or 12 months. A second prospective observational study conducted in patients who received two or fewer biologic agents aimed to assess the effectiveness of infliximab and golimumab in clinical practice in the Middle East and Northern Africa[26]. This was terminated early due to the low number of participating patients and high patient withdrawal; at study termination, all patients had received infliximab, with data available only up to 3 months. The mean (SD) numbers of hospitalization days for patients with CD and UC were 1.1 (7.8) and 0.7 (1.8), respectively[26].

Although studies have used IBD-related hospitalizations and surgery as endpoints, they have been used in an inconsistent way. Also, when using IBD-related hospitalizations and surgery as endpoints, it remains complex to delineate clinically relevant information due to the diversity of patient populations, settings, timepoints, and drugs used, which all lead to variable surgery and hospitalization rates.

Comparative effectiveness of advanced therapies for midterm complications

An observational cohort study in patients with UC using insurance data compared the effectiveness and safety of tofacitinib vs vedolizumab in patients who previously used anti-TNF inhibitors and were newly initiating study treatment. Follow-up began the day after cohort entry and ended at the earliest occurrence of outcome, 365 days post cohort entry, disenrollment, death, or end of data. There were no significant differences between new users of tofacitinib (n = 234) and vedolizumab (n = 722) for the secondary outcome of colectomy [incidence rate (IR) per 1000 person-years: Tofacitinib, 97.17; Vedolizumab, 59.25; Adjusted hazard ratio (HR) = 1.79; 95% confidence interval (CI): 0.93-3.44][27].

Three of four identified studies in patients with CD compared vedolizumab with ustekinumab. A propensity-score matched study from the United Kingdom compared patients treated with ustekinumab and those treated with vedolizumab who failed prior anti-TNF agents[28]. An exploratory outcome at 52 weeks was hospitalization rates before and after matching. The results of the exploratory analysis showed that there was a significant risk difference between ustekinumab and vedolizumab in the unmatched cohort (risk difference, 12%; 95%CI: 3%-20%; P = 0.01), which was no longer significant in the matched population (risk difference, 9%; 95%CI: -4 to 22; P = 0.20). A second study was a retrospective cohort study using longitudinal claims data in patients who used anti-TNF treatment in the previous 6 months[29]. Secondary outcomes included hospitalization for CD with and without surgery. A total of 1217 new users of ustekinumab had fewer nonsurgical CD hospitalizations than 667 new users of vedolizumab [IR per 1000 person-years: Ustekinumab, 76.25; Vedolizumab, 136.19; Adjusted HR = 0.58 (0.40-0.83)][29]. There was a numerical trend toward fewer surgical hospitalizations for CD among new users of ustekinumab vs new users of vedolizumab [IR per 1000 person-years: Ustekinumab, 87.96; Vedolizumab, 97.78; Adjusted HR = 0.83 (0.57-1.22)]. A retrospective cohort study using United States administrative claims data to compare CD-related surgery during the follow-up period in biologic-naive patients was also identified[30]. After 1-year of follow-up, there was a significantly lower rate of CD-related surgeries in vedolizumab-treated vs ustekinumab-treated patients (7.7% vs 11.6%; HR = 0.66; 95%CI: 0.44-0.99; P = 0.047).

These studies in patients with CD comparing vedolizumab with ustekinumab had varying results for CD-related hospitalizations and surgeries, with two studies reporting similar rates between treatments, one study favoring vedolizumab, and one study favoring ustekinumab. Interestingly, the similar rates of surgery between treatments were shown in patients with prior anti-TNF treatments, whereas vedolizumab was favored in the study including biologic-naive patients. It should be noted that endpoints were often secondary or exploratory outcomes and therefore were not sufficiently powered to detect differences between outcomes. The endpoint definitions vary between trials, making it challenging to compare results.

The fourth study in patients with CD made comparisons between more than two biologics, with outcomes assessed within 1 year of biologic initiation[31]. The risk of IBD-related surgery was investigated as a secondary outcome in 2965 patients. Results showed no differences in the 1-year risk of IBD-related surgery for the comparisons of ustekinumab and vedolizumab [propensity-score matched cohort HR = 1.42 (0.54-3.72)], vedolizumab and anti-TNF antagonists [propensity-score matched cohort HR = 0.63 (0.27-1.47)], or ustekinumab and anti-TNF antagonists [propensity-score matched cohort HR = 1.08 (0.69-1.70)][31].

Three studies in patients with IBD compared vedolizumab with anti-TNF agents, again with varying results. A population-based cohort study was conducted that included 1363 patients with UC or CD who received first-line anti-TNF treatment and initiated treatment with a second-line anti-TNF agent or vedolizumab[32]. After propensity-score matching, vedolizumab-treated patients with UC had significantly lower drug survival without IBD-related hospitalization after a 12-month follow-up compared with anti-TNF-treated patients. There were no significant differences between treatments for drug survival without IBD-related surgery in patients with UC or CD, and drug survival without IBD-related hospitalization in patients with CD[32]. In addition, a retrospective medical chart review of 1095 biologic-naive patients with IBD (UC, n = 604; CD, n = 491) that included disease-related surgeries after a 24-month follow-up as a secondary outcome found no difference between vedolizumab and anti-TNF cohorts in patients with CD (IR: 0.7 vs 1.5 per 100 years)[33]. Only one patient with UC had a colectomy (anti-TNF group). These two studies concluded that IBD-related surgeries are similar between vedolizumab and anti-TNF agents in biologic-naive patients and those treated with a second-line biologic. The differences in endpoint definitions make the results of these studies difficult to compare. Greater adherence to the SPIRIT consensus when conducting real-world studies would help standardize endpoints across studies and make comparisons of endpoints more meaningful.

A third study assessed the ability of either a five-variable (Optum dataset) or three-variable (Truven dataset) clinical decision support tool to predict differences in CD-related hospitalizations and surgery in patients treated with vedolizumab or an anti-TNF[34]. Patients treated with vedolizumab were stratified into high, intermediate, and low probabilities of response in the Optum dataset (n = 358) and high or low probabilities of response in the Truven dataset (n = 1445). Patients treated with an anti-TNF (n = 814) were stratified into high (n = 650), intermediate (n = 158), and low (n = 6) probabilities of response in the Optum dataset, although patients with a low probability of response were not included in the analysis due to the low number of patients. In the Optum dataset, rates of hospitalization and surgery were two to four times lower for vedolizumab-treated patients with a high probability of response vs those with a low probability (hospitalization, 19.0% vs 48.1%, P < 0.001; surgery, 8.4% vs 44.4%, P < 0.001). In the Truven dataset, a significantly lower proportion of the high probability of response group experienced a CD-related hospitalization (14.4% vs 19.6%, P = 0.011) or CD-related surgery (13.5% vs 23.3%, P < 0.001) during the 12 months after vedolizumab initiation. In patients from the Optum dataset who were treated with an anti-TNF, there was also a significant association between the intermediate or high probability of response and surgery (P = 0.0041) or hospitalization (P = 0.0463), although the strength of association was lower than that for vedolizumab. This clinical prediction tool, which included IBD-related hospitalization and IBD-related surgery, was better at predicting response to vedolizumab than anti-TNF agents.

EIMs

EIMs are common in UC and CD[35], occurring when the disease affects other parts of the body; thus, they should be differentiated from events occurring secondarily due to chronic disease and malnutrition or treatment side effects[36]. EIMs were included as an outcome measure in the SPIRIT consensus because they are indicative of the systemic nature of the disease and may impact therapeutic decisions; conversely, some treatments (including surgery and biologics) are risk factors for EIMs, with a genome-wide study implicating pathways and genes that are targets for existing drugs (e.g., anti-TNFs, cell adhesion molecules, Janus kinase inhibitors, and IL-6)[16,37,38]. We identified five studies that captured EIMs within 12-36 months of follow-up as recommended by the SPIRIT consensus[16].

In a real-world prospective cohort study of 252 patients with CD who were treated with ustekinumab and followed up for ≥ 2 years, the absence of EIMs was included as a secondary outcome. Results showed that at baseline, there were 67 EIMs reported in 59 patients. During follow-up, 22 of 67 (32.8%) arthralgia events achieved EIM remission. In addition, all events of uveitis, aphthous stomatitis, and pyoderma gangrenosum achieved EIM remission and one of two patients with erythema nodosum achieved EIM remission during the 2-year follow-up[39]. Thirty-nine patients developed new arthralgias during follow-up, of whom 23 (66.7%) achieved EIM remission before week 104. Other EIMs that developed during follow-up were transient aphthous stomatitis (n = 3), uveitis (n = 2), erythema nodosum (n = 2), and pyoderma gangrenosum (n = 1). A second study was the noninterventional, prospective, POLONEZ study that included patients with UC eligible for treatment with vedolizumab in Poland (biologic naive, n = 55; biologic experienced, n = 45)[40,41]. The results showed > 2-fold reduction in the presence of EIMs throughout vedolizumab treatment, from 13% at baseline and week 14 to approximately 5% at weeks 54 and 80. The most frequently reported EIM was arthralgia, occurring in 12% of patients at baseline and week 14, and approximately 5% of patients at weeks 54 and 80[41].

A multicenter observational study of patients with IBD receiving biologic treatment assessed the clinical course of new-onset and preexisting EIMs as a secondary outcome[42]. A total of 224 patients with new or preexisting EIMs were included in the analysis, who received a median 3 (vedolizumab) and 4 (non-gut-selective agents) years of ongoing advanced therapy. Patients treated with vedolizumab experienced a numerically higher rate of clinical worsening than those treated with non-gut-selective therapies (15.5% vs 7.3%, P = 0.08). Both groups had similar rates of modification of the therapeutic regimen to control clinical worsening. An observational study of patients with IBD who were treated with vedolizumab or ustekinumab for up to 52 weeks[43] showed no difference between the two therapies regarding their effect on EIMs during the follow-up [clinical response of EIMs at week 52: Ustekinumab, 18/50 (36%); Vedolizumab 19/54 (34%);P = 0.9; Clinical response of arthralgia at week 52: Ustekinumab, 18/46 (36%); Vedolizumab 19/55 (34%); P = 0.3][43]. A further noninterventional, multicenter, retrospective medical chart review study was conducted in 99 patients with moderately to severely active IBD and concurrent active EIMs at vedolizumab initiation[44]. The primary endpoint was resolution of all EIMs at 6 months of vedolizumab initiation. The most frequent EIMs were arthralgia (69.7%), peripheral spondyloarthritis (21.2%), and axial spondyloarthritis (10.1%). The primary endpoint of resolution of all EIMs at 6 months was achieved by 19.2% of patients; within 12 months of vedolizumab initiation, this increased to 25.3%[44]. Improvement (resolution or partial response) of EIMs was reported in 36.5% and 49.5% of patients at 6 months and 12 months, respectively. Individual EIMs that resolved at 6 months and 12 months were arthralgia [14/69 (20.3%); 18/69 (26.1%)], peripheral spondyloarthritis [6/21 (28.6%); 7/21 (33.3%)], erythema nodosum [3/7 (42.9%); 3/7 (42.9%)], primary sclerosing cholangitis [1/6 (16.7%); 2/6 (33.3%)], and oral aphthous ulcers [0/3 (0%); 1/3 (33.3%)]. No resolution was observed in patients with axial spondyloarthritis and uveitis. During the study, new-onset EIMs of arthralgia, peripheral spondyloarthritis, primary sclerosing cholangitis, and uveitis were reported in 1% of patients.

Although included in the SPIRIT guidelines, few studies have reported EIMs, which may reflect that many studies were started before the guidelines were published. Of the studies that report EIMs, there is considerable variation in baseline EIM prevalence, although treatment has shown improvements in EIMs over time. More evidence regarding the prevalence and treatment of EIMs is required, ideally those verified by board-certified rheumatologists or dermatologists.

Long-term complications

The SPIRIT consensus identified two long-term complications, dysplasia/cancer (gastrointestinal and extraintestinal neoplastic lesions) and mortality (IBD related and non-IBD related), both over a 5-year follow-up period. Four studies reporting these long-term complications were identified, of which two reported cancer outcomes and two reported deaths. Of these four studies, one included biologic-naive patients with IBD, one included biologic-naive and biologic-experienced patients with IBD, and two did not mention the line of therapy.

The first of the two studies reporting dysplasia or cancer was a retrospective cohort study using administrative claims data to identify patients with IBD without prior malignancy who were new users of vedolizumab (n = 759) or anti-TNF antagonists (n = 4807) between 2014 and 2018[45]. The results showed that there were no differences in the risk of incident malignancy between patients treated with vedolizumab vs those treated with anti-TNF (IR = 9.0 vs 11.6 per 1000 patient-years)[45]. The second study reporting dysplasia or cancer was a retrospective cohort study of patients with IBD with a history of current or prior cancer and who initiated vedolizumab, anti-TNF, or had no immunosuppressive therapy after the cancer diagnosis. The follow-up period was a median of 6.2 person-years. The results showed that neither vedolizumab nor anti-TNF therapy had an increased risk of new or recurrent cancer compared with no immunosuppression[46]. It should be noted that neither of these two studies used a 5-year follow-up period as recommended by the SPIRIT consensus. No patients died in the two studies that included mortality as a long-term outcome[17,47]. Studies reporting long-term consequences of IBD were scarce, with no long-term complication measured over a period of 5 years as recommended by the SPIRIT consensus.

Impact on patients’ lives

The SPIRIT consensus identified three outcomes that impact on patients’ lives: HRQoL [combination of 36-Item IBD Questionnaire (IBDQ-36) and 36-Item Short-Form Health Survey (SF-36)], disability [IBD Disability Index (IBD-DI)], and fecal incontinence (Jorge-Wexner, Cleveland score). We identified nine studies reporting on impacts to patients’ lives; the majority assessed HRQoL and one assessed disability. No studies used composite measures of HRQoL.

Regarding HRQoL, the IBDQ was used in eight studies[25,26,48-53] and the SF-36 in one study[54]. No studies used the SPIRIT-recommended composite of the IBDQ-36 and SF-36, and all of the studies included used shorter versions of the IBDQ rather than the 36-item tool.

A noninterventional, prospective, multicenter study of 81 anti-TNF-naive patients with UC who were treated with golimumab was the only study to use IBDQ-32 remission (total IBDQ-32 score ≥ 170 points) over 12 months as a primary endpoint. Secondary endpoints included 6- and 12-month IBDQ-32 responses (total score increase of ≥ 16 points from baseline). The results showed that IBDQ-32 remission was achieved by 76.9% of patients at 12 months of treatment[49]. The proportion of patients achieving IBDQ-32 response at 6 and 12 months was 73.0% (46/63) and 75.0% (39/52), respectively.

Another study that used the IBDQ-32 was a prospective cross-sectional study of patients with IBD at one hospital in Brazil. In the study, patients with UC were mainly (78.6%) treated with aminosalicylates, whereas patients with CD were mainly (78.6%) treated with biologics. They found that total IBDQ-32 scores were significantly lower in females than males, inferring that they had greater impairment in HRQoL. The difference in IBDQ-32 scores between males and females was larger in patients with CD than those with UC[53].

A further study used the IBDQ to assess HRQoL but did not specify which version was used. This study aimed to assess the effectiveness of infliximab and golimumab in clinical practice in the Middle East and Northern Africa, but was terminated early due to the low number of participating patients and high patient withdrawal; therefore, data were available only up to 3 months. At study termination, all patients had received infliximab. The mean (SD) change from baseline to 3 months in IBDQ score was 11.3 (39.6) points[26].

Three studies used the Short IBDQ (SIBDQ), a 10-item questionnaire giving a total score of 10-70 points, with lower scores indicating more severe impairment[55]. The first was a prospective observational study reporting significant SIBDQ score improvements from baseline to 12 months in patients with CD (n = 67) and UC (n = 48) treated with the infliximab biosimilar infliximab-dyyb[25]. The second, a prospective observational study of 96 patients with UC who were treated with vedolizumab, reported significant SIBDQ score improvements from baseline to weeks 14 and 54[48]. Finally, results from an interim analysis of the Japanese postmarketing surveillance study of vedolizumab in patients with moderate to severe UC reported improvements in HRQoL after vedolizumab induction, with a mean SIBDQ score increase of 8.9 (12.53) points[52].

In addition, a 1-year prospective observational study in patients with moderate to severe UC from Iran who were treated with adalimumab (n = 160) or infliximab (n = 78) assessed HRQoL using a 9-item short-form IBDQ (total score 9-63 points). The results showed that the mean (SD) 9-item IBDQ scores were similar between the adalimumab [37.42 (10.72)] and infliximab [36.61 (11.12)] treatment groups[51].

The SF-36 is a self-report questionnaire exploring eight health-related domains (mental health, emotional role, social functioning, vitality, general health, body pain, physical role, and physical functioning), with scores ranging from 0 to 100 and higher scores representing better HRQoL. The SF-36 was used in a prospective observational study in patients with IBD prescribed infliximab or vedolizumab. The primary outcome was the correlation between markers of inflammation and HRQoL, measured using the SF-36, Functional Assessment of Chronic Illness Therapy, and the Work Productivity and Activity Impairment: General Health Questionnaire tools at baseline and 14 weeks after the start of biological treatment. The results showed a significant improvement in all domains of the SF-36 from baseline to 14 weeks of treatment[54].

Disability was assessed using the IBD-DI in a retrospective, cross-sectional, observational, pan-European study[56]. Patients with moderate or severe UC (n = 1000) in the 12 months before the index date were compared with those with moderate or severe UC who achieved mild UC or remission status 12 months before the index date (n = 647) to explore the effect of UC on HRQoL and productivity as a secondary endpoint. The IBD-DI score was 30.6 in patients with moderate or severe UC and 22.3 in patients with moderate or severe UC who achieved mild UC or remission status 12 months before the index date[56]. IBD-DI scores varied by country, with the highest scores in Italy and Romania and the lowest score in Denmark. This study also reported the IBD-Control Score from patients in the United Kingdom, which was a mean 7.2 overall and was similar between cohorts (7.1 vs 7.3).

No studies that used the SPIRIT-recommended IBDQ-36 and SF-36 together were identified, with the majority using a shorter version of the IBDQ-36 over a period of ≤ 12 months. Together, these studies provide evidence that HRQoL is lower in females than males and in CD vs UC. In addition, biologic treatment has been shown to improve HRQoL. The only study comparing two different treatments suggested that these improvements were similar between adalimumab- and infliximab-treated patients. Interestingly, no studies used the Cleveland score or Jorge-Wexner score for fecal incontinence.

DISCUSSION

Considerable gaps in SPIRIT-recommended endpoint data were identified in this review of recently published real-world studies in patients with IBD treated with advanced therapies. Specifically, for midterm complications, no studies were identified that used the Cleveland or Jorge-Wexner scores for fecal incontinence, the Lemann Index for bowel damage in patients with CD, macroscopic proximal disease extension in UC, permanent stoma, or short-bowel syndrome. In addition, no studies calculated the composite of the number of IBD-related hospitalizations and cumulative length of hospital stay. Regarding long-term outcomes, none of the identified studies investigating cancer or mortality were conducted over the 5-year period recommended by the SPIRIT consensus. A potential reason for some of these endpoint data gaps is the timing of when the SPIRIT guidelines were published: Some of the studies included in our review would have been initiated before publication in 2021, which would explain the inconsistencies in disease progression endpoint reporting. In addition, data for other biologics, especially those published before the SPIRIT guidelines, may not have been included. A possible reason why no studies used some of the midterm complications defined by the SPIRIT consensus may be due to the data not being available in databases/registries. These outcomes may not be routinely collected for all patients and therefore there may be a high proportion of missing data. Some endpoints, such as the combination of the IBDQ-36 and SF-36, may be used less frequently due to patients being less willing to complete longer and/or multiple forms and more likely to be compliant with shorter forms. Finally, physicians may be waiting for the results of clinical trials to determine whether evaluating disease-modification endpoints should become routine in clinical practice.

In our review, no studies used the composite measures of the IBDQ-36 and SF-36 recommended by the SPIRIT consensus, while the majority of studies investigating HRQoL used a shorter version of the IBDQ, rather than the 36-item tool. Both the IBDQ-32 and the SIBDQ have been shown to have good measurement properties, and to have proven validity, reliability, and responsiveness[57]. An advantage of the IBDQ-32 is that it has been translated into 93 languages and therefore can be applied more easily to global studies, allowing more generalizable results. A possible reason as to why the SIBDQ is used more frequently in real-world practice is that it may be considered more practical, with significant time savings for patients and clinicians compared to using the IBDQ-36 or both the IBDQ-36 and SF-36 as recommended by the SPIRIT guidelines.

One study used the IBD-Control Questionnaire as recommended by the Innovative Medicines Initiative endorsed Health Outcomes Observatories (IMI-H₂O) consensus. This consensus recommends IBD-specific questions be used at a minimum of every 12 months, including use of the full Patient-Reported Outcome 2 CD[58]/UC[59] and IBD-control questionnaires[60]. In addition, it recommends self-standing questions be used to measure bowel incontinence, bowel urgency, feeling informed, IBD medication adherence, perianal disease symptoms, and ostomy symptoms. It also includes generic questions that are not IBD-specific, including the patient-reported outcomes measurement information system (PROMIS) global health[61], PROMIS self-efficacy[62], and subjective health experience model[63] questionnaires. The IMI-H₂O consensus aligns with the SPIRIT consensus in recommending the inclusion of hospitalization, EIMs, bowel incontinence, colorectal cancer, and mortality; however, they do not include bowel damage in patients with CD, macroscopic proximal disease extension in UC, permanent stoma, or short-bowel syndrome. The differences between the two consensuses may cause confusion for clinicians; however, the primary difference lies in their objectives and the types of outcomes they define. While the SPIRIT consensus focuses on defining endpoints for clinical trials aimed at preventing disease progression, the IMI-H₂O consensus focuses on a core set of outcomes that can be used in routine clinical practice.

Unlike other disease areas such as rheumatology, for which disease-modification outcomes have been clearly defined and validated[13,64,65], the SPIRIT consensus recommendations require further validation in clinical studies before implementation in disease-modification trials[16]. Therefore, it is not surprising that the majority of data retrieved in our search were derived from secondary outcomes and were not the primary focus of the studies. Overall, there was considerable heterogenicity in the study designs, from which robust inferences on disease-modification endpoints remain challenging. In addition, the limited observation periods of ≤ 12 months in the majority of studies may be too short to make conclusions on disease-modification outcomes. In neurodegenerative diseases like Parkinson’s and Alzheimer’s, disease-modification efforts require a paradigm shift from lumping to splitting, meaning that patients with the same disease may not share the same biological drivers; therefore, splitting the disease into small molecular/biological subtypes to match individuals to the therapies most likely to benefit them is warranted[66]. Whether this principle of precision medicine would also apply to IBD has yet to be fully investigated, although outcomes differ between patients with UC and CD. Newer, more stringent endpoints for disease modification, such as histologic remission or disease clearance, have been proposed for patients with UC and should also be validated in clinical trials[17,67-69].

Although the SPIRIT consensus recommends that the ultimate therapeutic goal in IBD should be the prevention of mid- and long-term complications and minimizing the impact of the disease on patients’ lives[16], disease clearance is a new method of comprehensive disease control that may be the first step toward disease modification. Disease clearance is measured through a combination of clinical, endoscopic, and histological remission; many physicians believe achieving these decreases the long-term progression of UC and prevents associated complications[68]. The prognostic value of histologic remission has been reported[70-72]. It is particularly useful in patients who still have histologic disease activity, but show endoscopic improvement or remission[73]. One ongoing trial, VERDICT (NCT04259138), is validating this outcome in a population of patients with UC treated with vedolizumab[69]: Patients are being randomized according to a treatment target algorithm in which the targets are: (1) Corticosteroid-free symptomatic remission (defined as Mayo rectal bleeding sub-score 0); (2) Corticosteroid-free endoscopic remission (Mayo endoscopic score ≤ 1) plus symptomatic remission; or (3) Corticosteroid-free histologic remission (Geboes Score < 2.B0) plus endoscopic and symptomatic remission. An interim analysis (March 2023) reported that of 432 patients enrolled and randomized, 51% (95%CI: 40%-61%), 37% (95%CI: 29%-46%), and 33% (95%CI: 27%-40%) met these treatment targets, respectively[69]. These results suggest that these targets are feasible in patients with moderate to severe UC, with the lower rates of histological remission reflecting that it is a harder treatment target for patients to achieve. A separate analysis (August 18, 2023) showed that more biologic-naive than biologic-experienced patients achieved the disease clearance target (corticosteroid-free histologic remission plus endoscopic and symptomatic remission) at week 16 [36% (n = 86/253) vs 24% (n = 9/37)][74]. In line with the SPIRIT guidelines, HRQoL will also be assessed in VERDICT.

CONCLUSION

In conclusion, there remains an unmet need for SPIRIT-recommended endpoints to be fully incorporated into prospective disease-modification trials. To meet this unmet need, the SPIRIT consensus recommendations require further validation in clinical studies. In order to enable comparisons between real-world studies, registry protocols should be updated to include these specific endpoints, with data recorded over long-term periods of observation.

ACKNOWLEDGEMENTS

Medical writing support was provided by Claire Line, PhD, and Jon Waldron, PhD, of Envision Catalyst, an Envision Medical Communications agency, a part of Envision Pharma Group, and was funded by Takeda.