Comparative Effectiveness of Ustekinumab and Anti-TNF Agent as First-Line Biological Therapy in Luminal Crohn’s Disease: A Retrospective Study From 2 Referral Centers

Abstract

Background

Real-life data on the efficacy of ustekinumab as first-line therapy for the treatment of luminal Crohn’s disease (CD) compared with anti-tumor necrosis factor (anti-TNF) agents are lacking. We compared the clinical response rates at 3 months in 2 cohorts of biologic-naïve patients treated by ustekinumab and anti-TNF agents.

Methods

Biologic-naïve patients starting either ustekinumab or an anti-TNF agent for luminal CD between 2016 and 2019 in 2 tertiary centers were retrospectively included. The primary endpoint was clinical response at 3 months, defined as a Harvey-Bradshaw Index <4 or a 3-point drop in the score without steroids, need for CD-related surgery, or treatment discontinuation owing to failure or intolerance. Patients treated with ustekinumab were matched to patients receiving anti-TNF agents by a propensity score algorithm.

Results

We included 156 patients starting anti-TNF agents (95 adalimumab and 61 infliximab) and 50 ustekinumab. After matching, clinical response rates at 3 months were 64% and 86% in the ustekinumab and anti-TNF groups, respectively (P = .01). At 12 months, in multivariate analysis adjusted for disease duration, location, concomitant immunosuppressant and steroids, and symptoms, clinical remission was independently associated with the biological therapy received (odds ratio, 2.6 for anti-TNF agent vs ustekinumab; P = .02). With a median follow-up duration of 40 (interquartile range, 23-52) months, no difference was observed in terms of time to drug withdrawal (P = .29) or safety.

Conclusions

This retrospective real-world data suggest that an anti-TNF agent as a first-line biological therapy is associated with higher rates of response at 3 months than ustekinumab in patients with CD.

Lay Summary

We conducted a retrospective real-world study to compare the efficacy of biologics in Crohn’s disease. Our data suggest that an anti-tumor necrosis factor agent as a first-line biological therapy is associated with higher rates of response at 3 months than ustekinumab in Crohn’s disease.

Crohn’s disease, ustekinumab, anti, TNF, infliximab, adalimumab

Key Messages

What is already known?

Ustekinumab and anti-tumor necrosis factor (TNF) agents are effective for patients with luminal Crohn’s disease. Data on their compared effectiveness as first-line therapy are lacking.

What is new here?

We found that anti-TNF agent use was associated with higher rates of clinical response at 3 months and remission at 12 months than ustekinumab in biologic-naïve patients with luminal Crohn’s disease. In the anti-TNF group, 2% of patients experienced severe adverse events compared with none in the ustekinumab group.

How can this study help patient care?

With an increasing number of available biologic agents, comparing the efficacy and safety of biologicals with a different mechanism of action is warranted to guide gastroenterologists’ prescription.

Introduction

In the past decades, anti-tumor necrosis factor (anti-TNF) agents have dramatically changed the therapeutic strategy in patients with Crohn’s disease (CD). However, in the pivotal randomized controlled trials that demonstrated the safety and efficacy of anti-TNF agents, the clinical remission rates at 1 year did not exceed 40%.¹ In addition, a subgroup of patients may lose response after having responded initially.² Recently, ustekinumab, a monoclonal antibody that blocks the p40 subunit shared between interleukin (IL)-12 and IL-23, was approved for the treatment of moderate to severe CD. With an increasing number of available biologic agents, comparing the efficacy and safety of biologicals with a different mechanism of action is warranted to guide gastroenterologists’ prescription. Nowadays in Europe and North America, anti-TNF agents are widely started as a first-line therapy based on physicians’ experience and treatment guidelines.³

To date, comparative data guiding physicians in the optimal sequencing of biological agents in CD are scarce. In the pivotal placebo-controlled randomized trials demonstrating the efficacy of ustekinumab, clinical remission rates were significantly higher in biologic-naïve patients as compared with those previously exposed to biologics.^4,5 In addition, most published real-world data on ustekinumab in CD involved patients who previously failed to achieve remission with another biologic.^6-8 To date, only 1 trial, published as an abstract, has directly compared ustekinumab with anti-TNF agents in a biologic-naïve cohort; the SEAVUE (Safety and Efficacy of Adalimumab Versus Ustekinumab for One Year) trial included biologic-naïve patients with moderate-to-severe luminal CD in a multicenter randomized controlled trial comparing the efficacy and safety of ustekinumab and adalimumab. No difference was observed in terms of clinical remission rates at 12 months between the 2 groups.⁹

Real-world comparative data on the efficacy of ustekinumab as first-line therapy for the treatment of moderate-to-severe CD are lacking. The aim of the present study was to compare the efficacy and safety of ustekinumab and anti-TNF agents as first-line therapy in luminal CD.

Methods

Study Design and Population

We conducted a 2-center retrospective study at the Inflammatory Bowel Disease Centres of the McGill University Health Centre (Montreal, Quebec, Canada) and Bordeaux University Hospital (Bordeaux, France) including consecutive biologic-naïve patients starting either ustekinumab or an anti-TNF agent (infliximab or adalimumab) for active luminal CD between January 2016 and December 2019. Patients older than 18 years of age with a confirmed diagnosis of active luminal CD based on clinical and endoscopic or radiographic data having received ustekinumab or an anti-TNF agent as first-line biological therapy were included if they had at least 1 clinical follow-up after initiation of therapy. We excluded patients treated with ustekinumab subcutaneous induction or for ulcerative colitis, indeterminate colitis, pouchitis or CD of the pouch, immediate postoperative prophylaxis for asymptomatic CD, or isolated perianal CD without luminal activity. Inclusion date corresponded to the day of first administration of the biological therapy. The follow-up period lasted from inclusion until death, CD-related surgery, treatment discontinuation, or last medical record before end of the follow-up period in September 2021.

Biological therapies were used according to licensed regimens. Ustekinumab was approved as first-line biological therapy in Canada in 2017. Therapeutic choice between an anti-TNF agent and ustekinumab was made by Canadian physicians after discussion with the patient. In France, ustekinumab use as first-line therapy is restricted to patients with a contraindication to anti-TNF agents. In both countries, gastroenterologists selected infliximab or adalimumab based on personal experience and patient’s preference. At induction, infliximab was started at 5 mg/kg intravenously at weeks 0, 2, and 6; adalimumab at 80 and 40 or 160 and 80 mg subcutaneously at weeks 0 to 2; and ustekinumab at approximately 6 mg/kg intravenously. During the maintenance phase, drugs were optimized based on clinical need and experienced physicians’ opinion to a maximum of 10 mg/kg every 4 weeks for infliximab, 80 mg every week for adalimumab, and 90 mg every 4 weeks or intravenous reinduction at 6 mg/kg for ustekinumab.

Data Collection

Data on variables of interest were collected in medical charts: age at diagnosis, age at inclusion, weight, sex, smoking status, Montreal classification, prior luminal CD-related surgery, extraintestinal manifestations, concomitant treatments (steroids or immunomodulators), ustekinumab and anti-TNF dose for induction, therapy optimization, adverse events, baseline disease activity (Harvey-Bradshaw Index [HBI], C-reactive protein [CRP]), and follow-up assessments (HBI, endoscopic, CRP, steroid and immunomodulator use) at 3, 6, 9, and 12 months ± 2 months when available. Fecal calprotectin measurements were only available for some patients included in Canada and were not routinely performed in France. No prespecified clinical or endoscopic assessment was planned during follow-up. Need for luminal CD-related surgery, treatment discontinuation, and occurrence of adverse events were captured.

Outcome Measures

The primary endpoint was clinical response at 3 months after therapy initiation. Secondary endpoints were clinical remission at 3 and 12 months, clinical response at 12 months, biochemical remission at 3 and 12 months, time to drug withdrawal, endoscopic healing at 6 to 18 months, adverse events, and severe adverse events. Factors associated with clinical response and remission at 3 and 12 months were determined.

Clinical response was defined as a 3-point drop in the HBI score compared with inclusion or HBI score <4 without use of steroids or need for CD-related surgery or treatment discontinuation due to failure or intolerance.¹⁰ Clinical remission was defined as an HBI score <4 without use of steroids or need for CD-related surgery or treatment discontinuation due to failure or intolerance. Biochemical remission was defined as clinical remission associated with a CRP ≤5 mg/L, corresponding to the upper limit of the normal range in our laboratories.¹¹ Endoscopic healing was defined as a Simple Endoscopic Score for CD score <4. Severe adverse events were defined as adverse events leading to hospital admission or prolonged stay. Therapy optimization was defined as an escalation in frequency or dose of biological therapy administration. Use of steroids included only oral or intravenous steroids, excluding topical steroids.

Propensity Score Matching

A propensity score matching analysis was performed to establish comparable groups of patients treated by ustekinumab and anti-TNF for primary outcome analysis. The following patient- and disease-related baseline variables were determined a priori and included in the propensity score model predicting the probability of the treatment given: age, CD duration shorter or greater than 18 months, smoking status, CD location according to the Montreal classification and history of associated perianal disease. Patients treated with ustekinumab were matched 1:2 to patients receiving an anti-TNF agent. Patients were matched using a nearest-neighbor matching algorithm with no replacement (MatchIt R package version 4.3.2). The maximum accepted difference between propensity scores in each pair was set at 0.1 SD. Patients in the ustekinumab or the anti-TNF groups without matching points in the other group within 0.1 SD were discarded. Comparability of baseline characteristics between matched groups was verified using chi-square test for categorical variables and Mann-Whitney U test for continuous variables. We performed a sensitivity analysis to compare separately patients treated with ustekinumab to patients treated by infliximab or adalimumab and to compare only patients treated by biological therapy alone without concomitant immunomodulator using the same propensity score analysis with a 1:1 match. The use of propensity score weighting method instead of matching would have increased the statistical power by making it possible to analyze the whole population but would have complexified the interpretation of the results.

Statistical Analysis

Statistical analyses were performed using R 3.6.3 (R Foundation for Statistical Computing, Vienna, Austria). Continuous variables were reported as median and interquartile range (IQR) and compared using Mann-Whitney U test. Categorical variables were expressed as frequencies and compared using a chi-square test or Fisher’s exact test when expected numbers were lower than 5. Two-sided statistical tests were used for all analyses. A P value <.05 was considered as significant. Missing data were not imputed. Kaplan-Meier curves were plotted for time from inclusion to drug withdrawal for failure or adverse event for each treatment group and compared using log-rank statistics. Patients discontinuing the treatment for another reason were excluded from the analysis. A logistic regression was conducted for eligible predictive factors of clinical response at 3 months and clinical remission at 3 and 12 months among age, sex, weight, smoking status, disease location, disease phenotype, history of perianal disease, extraintestinal manifestations, history of CD-related luminal surgery, disease duration, HBI at inclusion, concomitant steroids or immunomodulators at inclusion, and biological therapy received. Continuous variables were dichotomized at the median. First, a univariate logistic regression was performed. Results are presented as odds ratio (OR) with 95% confidence intervals (95% CIs). Variables with a P value below .20 and variables of biological interest (ie, disease location, disease duration, concomitant immunosuppressant at inclusion, and steroids at inclusion) were included in a full model of multivariable logistic regression, interactions were tested, and the variables were selected for the final model using a backward elimination.

Ethical Considerations

This study was conducted in accord with the Tri-Council Policy Statement: Ethical Conduct for Research Involving Humans (2014), as well as in respect of the requirements set out in the applicable standard operation procedures of the Research Institute of the McGill University Health Centre Research Institute and of the McGill University Health Centre Research Ethics Board. The study was approved by the McGill University Health Centre institutional review board and ethics committee (REB no. 2020-5825).

Results

Study Population

Among the 206 included patients (103 [50%] women, median age 30 [IQR, 22-47] years), 156 (76%) patients were treated with an anti-TNF agent (95 [46%] with adalimumab and 61 [30%] with infliximab) and 50 (24%) were treated with ustekinumab as first-line therapy. Median CD duration was 21 (IQR, 6-104) months, and 45 (22%) patients had previously undergone a CD-related abdominal surgery. At inclusion, the median HBI was 5 (IQR, 2-8), and 81 (39%) patients were receiving oral corticosteroids. Compared with patients on an anti-TNF agent, patients receiving ustekinumab were less frequently active smokers (P = .02), had less frequently a history of perianal CD (P = .04), and were less frequently treated with concomitant immunomodulators (P = .01). A CRP measurement at inclusion was available for 43 of 50 patients in the ustekinumab group and 155 of 156 patients in the anti-TNF group. The proportion of patients with CRP >5 mg/L at inclusion was similar between the 2 groups: 30 (70%) of 43 vs 108 (70%) of 156 (P > .9), respectively. Compared characteristics of the 2 groups are displayed in Table 1.

Table 1.

Patient characteristics at initiation of biological therapy

Variable	Overall Population (N = 206)	Anti-TNF Group (n = 156)	Ustekinumab Group (n = 50)	P Value
Center				<.01a
Bordeaux	118 (57)	116 (74)	2 (4)
Montreal	88 (43)	40 (26)	48 (96)
Age, y	30 (22-47)	29 (22-45)	37 (25-56)	.05
Female	103 (50)	77 (49)	26 (52)	.70
Weight, kg	62 (55-77)	61 (55-76)	64 (56-80)	.20
Active smoker	51 (25)	45 (29)	6 (12)	.02a
Disease locationa				.08
Ileal	55 (27)	44 (28)	11 (22)
Ileocolonic	94 (46)	75 (48)	19 (38)
Colonic	57 (28)	37 (24)	20 (40)
Disease phenotypea				.09
Inflammatory	126 (61)	89 (57)	37 (74)
Stricturing	48 (23)	41 (26)	7 (14)
Penetrating	32 (16)	26 (17)	6 (12)
History of perianal disease	61 (30)	52 (33)	9 (18)	.04a
Extraintestinal manifestations	45 (22)	31 (20)	14 (28)	.20
History of CD-related surgery	45 (22)	37 (24)	8 (16)	.30
Disease duration <18 mo	102 (50)	81 (52)	21 (42)	.20
HBI at inclusion	5.0 (2.0-8.0)	5.0 (2.0-8.0)	5.0 (2.0-8.2)	>.90
CRP at inclusion, mg/L	13.9 (3.4-32.3)	12.5 (3.4-33.4)	16.6 (3.7-26.8)	.80
Concomitant steroids	81 (39)	57 (37)	24 (48)	.11
Concomitant immunomodulators	78 (38)	68 (44)	10 (20)	.01a
Follow-up duration, mo	40 (23-52)	46 (32-56)	21 (15-27)	<.01a

Values are n (%) or median (interquartile range).

Abbreviations: CD, Crohn’s disease; CRP, C-reactive protein; HBI, Harvey-Bradshaw Index; TNF, tumor necrosis factor.

^aAccording to Montreal classification.

Regarding drug regimens for ustekinumab, intravenous induction was administered at a median dose of 5.6 (IQR, 5.0-6.3) mg/kg; maintenance regimens were 90 mg every 8 weeks in 49 (98%) of 50 patients and 90 mg every 12 weeks in 1 patient.

During a median follow-up duration of 40 (IQR, 23-52) months, the biological therapy was intensified in 26 (52%) of 50 patients in the ustekinumab group vs 73 (47%) of 156 in the anti-TNF group (P = .50) with a median time to escalation of 7 (IQR, 4-11) months and 7 (IQR, 4-13) months, respectively. In the ustekinumab group, median interval after intensification was 90 mg every 4 (IQR, 4-4) weeks including 2 (8%) of 26 patients receiving intravenous reinduction at 6 mg/kg. In the anti-TNF group, 21 (34%) of 61 patients underwent dose or interval intensification for infliximab and 52 (55%) of 95 for adalimumab. In patients undergoing intensified anti-TNF maintenance regimens, median dose and intervals were 7.5 (IQR, 5.0-10.0) mg/kg and 6 (IQR, 6-7) weeks for infliximab and 40 (IQR, 40-40) mg and 1 (IQR, 1-1) week for adalimumab, including 1 (2%) of 52 patients receiving reinduction at 160 mg.

Unmatched outcomes at 3 months

At 3 months, clinical assessment was available for 44 (88%) of 50 and 140 (90%) of 156 patients in the ustekinumab and the anti-TNF groups and CRP values for 33 (66%) of 50 and 131 (83%) of 156 patients, respectively. The remaining patients were excluded from the analysis because the HBI score or CRP measurement was missing. Clinical response and clinical remission were observed in 29 (66%) of 44 and 114 (81%) of 140 patients and in 25 (57%) of 44 and 106 (76%) of 140 patients in the ustekinumab and the anti-TNF groups, respectively (P = .03 and P = .02). Biochemical remission was observed in 11 (33%) of 33 and 78 (60%) of 131 in the ustekinumab and the anti-TNF groups, respectively (P = .01).

Matched and Adjusted Outcomes at 3 Months

Each ustekinumab patient was matched to the 2 closest anti-TNF patients regarding the variables included in the propensity score when available. Overall, 39 patients treated with ustekinumab and 69 patients treated with anti-TNF were used for the matched analysis. Comparison of baseline characteristics of the matched populations is displayed in Supplementary Table 1. Clinical response rates at 3 months were 25 (64%) of 39 and 59 (86%) of 69 in the ustekinumab and anti-TNF groups, respectively (OR, 0.30; 95% CI, 0.12-0.77; P = .01) (Figure 1). In the univariate analysis, clinical response at 3 months was associated with the biological therapy received (OR, 1.17; 95% CI, 1.02-1.34; P = .03 for anti-TNF agent vs ustekinumab) and absence of steroids at inclusion (OR, 1.18; 95% CI, 1.04-1.33; P = .01). In the multivariate analysis, absence of steroids at inclusion (OR, 2.20; 95% CI, 1.02-4.82; P = .05) was independently associated with clinical response at 3 months (Table 2).

Table 2.

Factors associated with clinical response at 3 months in univariate and multivariate logistic regression (n = 184)

Variable	Univariate Analysis		Multivariate Analysis
	OR (95% CI)	P Value	OR (95% CI)	P Value
Age <30 y		.90
Male		.90
Weight <60 kg		.50
Active smoker		>.90
Disease location		.60		>.90
Disease phenotype		>.90
History of perianal disease		.10		—
Extraintestinal manifestations		.30
History of CD-related luminal surgery		.20
Disease duration <18 mo		.40		.40
HBI at inclusion		.09		.40
Absence of concomitant steroids	1.18 (1.04-1.33)	.01	2.20 (1.02-4.82)	.05
Concomitant immunomodulators		.20		.50
Biological therapy received (anti-TNF agent vs ustekinumab)	1.17 (1.02-1.34)	.03		.08

Abbreviations: CD, Crohn’s disease; CI, confidence interval; HBI, Harvey-Bradshaw Index; OR, odds ratio; TNF, tumor necrosis factor.

Figure 1.

Clinical response rates at 3 months according to the biological therapy received. The numbers in each colored bar indicate the frequency for each group. The top and the bottom of the black line indicate the lower and upper bounds of the 95% confidence interval. For this analysis, 39 patients treated by ustekinumab were matched to 69 patients treated by an anti-tumor necrosis factor agent based on a propensity score, and rates were compared using a chi-square test. The corresponding P value is displayed on top of the colored bars.

In the univariate analysis, clinical remission at 3 months was associated with the biological therapy received (OR, 1.21; 95% CI, 1.04-1.41; P = .02 for anti-TNF agent vs ustekinumab), a lower HBI at inclusion (P < .01), concomitant immunomodulator treatment at inclusion (OR, 1.15; 95% CI, 1.00-1.31; P = .05), and absence of steroids at inclusion (OR, 1.21; 95% CI, 1.06-1.38; P = .01). In the multivariate analysis, the biological therapy received (OR, 2.26; 95% CI, 1.01-5.08; P = .05 for anti-TNF agent vs ustekinumab) and a lower HBI at inclusion (P = .01) were independently associated with clinical remission at 3 months (Supplementary Table 2).

Unmatched outcomes at 12 months

At 12 months, clinical assessment was available for 49 (98%) of 50 and 142 (91%) of 156 patients and CRP values were available for 43 (86%) of 50 and 133 (85%) of 156 in the ustekinumab and anti-TNF groups, respectively. Clinical and biochemical remission were observed in 31 (63%) of 49 and 116 (82%) of 142 patients and in 18 (42%) of 43 and 83 (62%) of 133 patients in the ustekinumab and anti-TNF groups, respectively (P = .01 and P = .02).

Adjusted Outcomes at 12 Months

In univariate analysis, clinical remission at 12 months was associated with the biological therapy received (OR, 1.17; 95% CI, 1.02-1.34; P = .03 for anti-TNF agent vs ustekinumab), a lower HBI at inclusion (P < .01), and history of perianal CD (OR, 1.16; 95% CI, 1.02, 1.32; P = .02). In multivariate analysis, the biological therapy received (OR, 2.62; 95% CI, 1.18, 5.79; P = .02 for anti-TNF agent vs ustekinumab) and a lower HBI at inclusion (P = .01) were independently associated with clinical remission at 12 months (Table 3 and Figure 2).

Table 3.

Factors associated with clinical remission at 12 months in univariate and multivariate logistic regression (n = 191)

Variable	Univariate Analysis		Multivariate Analysis
	OR (95% CI)	P Value	OR (95% CI)	P Value
Age <30 y		.80
Male		.05		—
Weight <60 kg		>.90
Active smoker		.70
Disease location		.40		.60
Disease phenotype		.80
History of perianal disease	1.16 (1.02-1.32)	.02		—
Extraintestinal manifestations		.06		—
History of CD-related luminal surgery		.90
Disease duration <18 mo		.70		.60
HBI at inclusion	0.98 (0.97-0.99)	<.01	0.87 (0.80-0.94)	<.01
Absence of concomitant steroids		>.90		.30
Concomitant immunomodulators		.50		.90
Biological therapy received (anti-TNF agent vs ustekinumab)	1.17 (1.02-1.34)	.03	2.62 (1.18-5.79)	.02

Abbreviations: CD, Crohn’s disease; CI, confidence interval; HBI, Harvey-Bradshaw Index; OR, odds ratio; TNF, tumor necrosis factor.

Figure 2.

Factors associated with clinical remission at 12 months in multivariate analysis. Dots correspond to the odds ratio and lines correspond to the 95% confidence intervals. A logistic regression was conducted including factors with P value <.2 in univariate logistic regression and variables of biological interest. HBI, Harvey-Bradshaw Index; TNF, tumor necrosis factor.

Drug Withdrawal and Endoscopic Healing

The biological therapy was discontinued for failure or adverse event in 5 (10%) of 50 patients in the ustekinumab group and 34 (22%) of 156 patients in the anti-TNF group (P = .03). Median time to treatment discontinuation was 9 (IQR, 7-11) months in the ustekinumab group vs 12 (IQR, 4-29) months in the anti-TNF group. Detailed reasons for all treatment discontinuation during follow-up are described in Table 4. No difference was observed in terms of time to drug withdrawal (P = .29) (Figure 3). Rates of drug persistence at 12 months were 88% in patients treated with ustekinumab and 87% in those treated with an anti-TNF agent, respectively.

Table 4.

Reasons for all treatment discontinuation according to the biological therapy received

Discontinuation Reason	Overall Population (N = 206)	Anti-TNF Group (n = 156)	Ustekinumab Group (n = 50)
Overall number of discontinuations	54 (26)	48 (31)	6 (12)
Failure	30 (15)	25 (16)	5 (10)
Adverse event	9 (4)	9 (6)	0 (0)
Patient’s decision	1 (0.5)	1 (0.5)	0 (0)
Pregnancy	4 (2)	3 (2)	1 (2)
Cancer	1 (0.5)	1 (0.5)	0 (0)
Remission	9 (4)	9 (6)	0 (0)

Values are n (%).

Abbreviation: TNF, tumor necrosis factor.

Figure 3.

Drug persistence during follow-up according to the biological therapy received. Kaplan-Meier curves plotted from biological therapy initiation to drug discontinuation for failure or intolerance. The STROBE statement is a checklist of items that should be included in reports of cohort studies. TNF, tumor necrosis factor.

Between 6 and 18 months after therapy initiation, 28 (56%) of 50 patients in the ustekinumab group and 78 (50%) of 156 patients in the anti-TNF group were evaluated by an ileocolonoscopy. Endoscopic healing was observed in 17 (61%) of 28 patients in the ustekinumab group vs 45 (58%) of 78 in the anti-TNF group (P = .78).

Safety

Adverse events were observed in 7 (14%) of 50 patients in the ustekinumab group vs 20 (13%) of 156 patients in the anti-TNF group (P = .80). In the anti-TNF group, 3 (2%) of 156 patients had severe adverse events observed compared with 0 (0%) of 50 in the ustekinumab group (P > .90). These events occurred in 3 patients treated by adalimumab and consisted of 1 patient admitted for cerebral aspergillosis, 1 for postinfectious macrophage activation syndrome, and 1 for severe folliculitis needing abscess drainage. None of the severe adverse events led to death. Detailed characteristics of adverse events are displayed in Table 5.

Table 5.

Adverse events according to biological therapy received

Variable	Overall Population (N = 206)	Anti-TNF Group (n = 156)	Ustekinumab Group (n = 50)	P Value
All adverse events	27 (13)	20 (13)	7 (14)	.80
Infection	8 (4)	7 (5)	1 (2)	.70
Abdominal pain	2 (1)	2 (1)	0 (0)	>.90
Arthralgia	9 (4)	5 (3)	4 (8)	.20
Infusion reaction	6 (3)	4 (3)	2 (4)	.60
Headache	4 (2)	2 (1)	2 (4)	.20
Dermatological lesions	4 (2)	4 (3)	0 (0)	.60
Severe adverse events	3 (2)	3 (2)	0 (0)	>.90

Values are n (%).

Sensitivity Analyses

Clinical response rates at 3 months were compared separately between the ustekinumab group and the 2 anti-TNF groups. After matching 1:1 to patients treated by infliximab, they were 17 (59%) of 29 in the ustekinumab group vs 25 (86%) of 29 in the infliximab group (P = .02). After matching 1:1 to patients treated by adalimumab, they were 24 (69%) of 35 and 26 (74%) of 35 in the ustekinumab and the adalimumab groups, respectively (P = .60).

In patients not receiving a concomitant immunomodulator, after matching 1:1, clinical response at 3 months was observed in 22 (67%) of 33 patients in the ustekinumab monotherapy group and 24 (72%) of 33 in the anti-TNF monotherapy group (P = .59).

Discussion

In this retrospective cohort, we found that anti-TNF agents as a first-line biological therapy led to higher rates of clinical response at 3 months in matched analysis and clinical remission at 12 months in multivariate analysis than ustekinumab in biologic-naïve patients with luminal CD. Drug persistence and endoscopic healing rates were similar between the 2 groups. No difference was observed in terms of overall adverse events between the 2 therapies. Three severe adverse events (2%) occurred in the anti-TNF group compared with none in the ustekinumab group.

Although regulatory agencies mainly rely on the randomized controlled trials to approve medication, it has been shown that only 30% of patients seen in a regular practice would fulfill the criteria to be included in those studies.¹² Real-world studies are useful to guide clinical practice and determine the efficacy of a given therapy. We present one of the first real-world studies comparing the efficacy of ustekinumab and anti-TNF agents in biologic-naïve patients with luminal CD. In the recent SEAVUE trial, biologic-naïve patients with active moderate-to-severe CD were randomized to receive ustekinumab or adalimumab for 1 year. Clinical remission, defined as Crohn’s Disease Activity Index <150 at 12 months, was achieved by 65% of patients in the ustekinumab group and 61% in the adalimumab group (P = .42).⁹ In the present cohort study, we observed a similar efficacy rate for ustekinumab than the SEAVUE trial but a higher efficacy rate for anti-TNF group (63% and 82%, respectively, in our cohort). This anti-TNF efficacy is consistent with what was described in the Leuven cohort—94% infliximab failure–free survival at 1 year in a population of recently diagnosed CD.² There are differences between our study population and the SEAVUE trial that may explain this discrepancy. First, we included patients receiving either infliximab or adalimumab. Indeed, in our sensitivity analysis, a significant difference in terms of efficacy was only found between infliximab and ustekinumab. Second, in the SEAVUE trial, therapy intensification was not allowed while half of our patients received intensified anti-TNF treatment. However, therapy intensification was also prescribed to half of the patients receiving ustekinumab in our study. Finally, median disease duration was shorter in our cohort (21 months compared with 30 months in the SEAVUE trial). A shorter disease duration has been strongly associated with anti-TNF efficacy in CD.^2,13,14 Only 1 real-world study comparing ustekinumab and anti-TNF agents in CD has been published to date. In an Australian national population-based analysis, Ko et al⁸ found superior treatment persistence with ustekinumab (12-month rates: 80% for ustekinumab, 68% for infliximab, and 64% for adalimumab; P = .01). However, owing to the study design, the authors could not compare rates of remission and response between ustekinumab and anti-TNF as we did. We found higher drug persistence rates at 12 months than Ko et al both in patients treated by ustekinumab and in patients treated by anti-TNF (88% and 87%, respectively). This could be explained by the fact that one-third of the patients had been previously exposed to a biologic in the Australian registry. Our data are also consistent with a network meta-analysis showing that infliximab was the treatment with higher efficacy to induce clinical remission after induction in biologic-naïve patients followed by adalimumab and ustekinumab using pooled data from randomized controlled trials.¹⁵

As clinical remission rate at 12 months was lower in the ustekinumab group, one could have expected that drug discontinuation rates would be higher in this group. However, in Canada, where the patients from the ustekinumab group were mostly recruited, ustekinumab intensification is only allowed after 4 months. Moreover, the first year of treatment overlapped with the start of the COVID-19 (coronavirus disease 2019) pandemic for a proportion of patients. As a consequence, they were less frequently evaluated than usual in outpatient clinics. Finally, gastroenterologists are less experienced with ustekinumab use than anti-TNF agents. These factors could have led gastroenterologists to continue ustekinumab even in case of incomplete clinical remission to avoid therapeutic switch in uncertain times and wait longer for the efficacy of therapy intensification.

The choice between currently available biologic therapies cannot be based only on efficacy. Safety, need for a concomitant immunomodulator, availability, route of delivery, patient preference, and cost should also be considered.³ In our cohort, 2% of patients exposed to an anti-TNF agent experienced a severe adverse event compared with none in the ustekinumab group. In trial SEAVUE trial, excluding those related to worsening CD, severe adverse events were observed in 18 (9%) of 195 adalimumab patients compared with 20 (10%) of 191 ustekinumab patients. The PSOLAR (Psoriasis Longitudinal Assessment and Registry) study compared cumulative rates of adverse events including serious infection, major adverse cardiovascular events, and malignancy in more than 12 000 patients treated by ustekinumab or infliximab for psoriasis.¹⁶ Higher rates of serious infections were observed in the infliximab group compared with ustekinumab group. No difference was described regarding the other events. Anti-TNF agents are frequently used in association with an immunomodulator to prevent immunogenicity.³ This factor can influence the safety profile of anti-TNF agents because combination therapy has been shown to be associated with higher risks of serious infection and lymphoma.^17,18

We acknowledge some limitations in our study. Patients in the 2 groups were not comparable in terms of smoking status, history of perianal disease, and use of concomitant immunomodulator. Active smoking was less frequent in the ustekinumab group, which should have favored this therapy. Regarding the 2 other factors, they are related to prescription habits of gastroenterologists: a concomitant immunomodulator treatment is recommended when infliximab is prescribed, and very few data have been published so far on the efficacy of ustekinumab on perianal disease. We used a propensity score matching to overcome these baseline differences to analyze the primary outcomes and multivariate analysis adjusted on disease characteristics for the other outcomes. The propensity score matching lowered the sample size because only patients with close propensity scores in both groups could be compared. Owing to a low proportion of patients treated with a concomitant immunomodulator in the ustekinumab group, this variable could not be added in the propensity score, and matched groups were still unbalanced. The effect of the concomitant immunomodulator could explain partly the superiority of anti-TNF agents over ustekinumab that we observed. When adjusting on concomitant immunomodulator use in the multivariate logistic regression model, the biological therapy received was not associated with clinical response at 3 months. However, it was still associated with clinical remission at 12 months. Treatment groups were not balanced between the 2 centers, as ustekinumab is approved as a first-line biological therapy only in Canada. Owing to its retrospective design, endoscopic evaluation was available only in a selected subset of patients, and fecal calprotectin was not routinely measured in the French cohort. Clinical outcomes are not the most reliable markers of treatment efficacy, especially in the subgroup of patients with previous CD-related surgery. At each time point, steroid use was taken into account at the day of evaluation without possibility to exclude with certainty short course of steroids between evaluations. A small group of patients were not seen for clinical evaluation during the year 2020 because of the COVID-19 pandemic. Finally, the small size of the ustekinumab group limited the statistical power for the sensitivity analysis and the reporting of adverse events.

Conclusions

This real-world study suggests that anti-TNF therapy, especially when used in combotherapy, is associated with higher rates of clinical response at 3 months and remission at 12 months than ustekinumab in biologic-naïve patients with active luminal CD. Registry studies, retrospective experiences, and randomized controlled trials are needed to investigate further the compared safety and efficacy of these 2 classes of biological therapies with the goal of determining a sequence of therapy adapted to a specific patient profile.

Author Contribution

Guarantor of the article: T.B.. Study concept and design; acquisition of data; analysis and interpretation of data; drafting of the manuscript; critical revision of the manuscript for important intellectual content; statistical analysis; technical or material support; study supervision: P.R., T.B., D.L. Acquisition of data and critical revision of the manuscript: C.K., A.N., G.P., M.H., N.A., F.P., A.B., F.Z., P.L.L., and W.A. All authors approved the final version of the manuscript.

Funding

None received.

Conflicts of Interest

P.R. has received consultancy and speaker fees from AbbVie, Amgen, and Janssen. C.K., G.P., A.N., M.H., and N.A. have no conflicts of interest to declare. F.P. has received consultancy and speaker fees from AbbVie, MSD, Ferring, Pfizer, Takeda, Biogen, and Janssen. A.B. has participated in advisory boards with AbbVie, Janssen, Pfizer, Takeda, Hoffman-LaRoche, and Amgen; received research support from AbbVie; and received educational support from Fresenius Kabi and Takeda. F.Z. has received speaker fees from Janssen. P.L.L. has served as a speaker and/or advisory board member for AbbVie, Amgen, Arena Pharmaceuticals, Fresenius Kabi, Genentech, Gilead, Janssen, Merck, Mylan, Organon, Pharmacosmos, Pfizer, Roche, Takeda, Tillotts, and Viatris; and received unrestricted research grant support from AbbVie, Gilead, Takeda, and Pfizer. WA has served on advisory boards for or received honoraria from AbbVie, Amgen, Arena Pharmaceuticals, Dynacare, Fresenius Kabi, Janssen, Merck, Novartis, Pfizer, Sandoz, and Takeda. D.L. has provided counseling for, served on boards for, or received transports or fees from AbbVie, Biogaran, Biogen, Ferring, HAC-pharma, Janssen, MSD, Novartis, Pfizer, Prometheus, Roche, Takeda, Theradiag, and Tillotts. T.B. has served as a speaker or advisor for AbbVie, Alimentiv (formerly Robarts Inc.), Amgen, Bristol-Myers-Squibb, Ferring, Fresenius Kabi, Gilead, Janssen, Merck, Pentax, Pfizer, Roche, Sandoz, Takeda, and Viatris.

Data Availability

The data underlying this article will be shared on reasonable request to the corresponding author.