INTRODUCTION

The introduction of biologic therapies, including antitumor necrosis factor therapies (anti-TNFs), revolutionized the treatment of moderate-to-severe inflammatory bowel disease (IBD), including Crohn's disease (CD) and ulcerative colitis (UC). Anti-TNFs, namely infliximab, have been shown in large randomized controlled trials (RCTs) to be associated with greater rates of clinical remission when used as so-called “combination therapy” with immunomodulators, including methotrexate, azathioprine, or 6-mercaptopurine (1,2). Use of combination therapy has been associated with lower rates of antidrug antibody (ADA) development and increased anti-TNF drug levels, a commonly proposed mechanism for the increased efficacy observed with this practice (3).

Novel biologics, including vedolizumab, an anti-integrin, and ustekinumab, an interleukin 12/23 inhibitor, represent unique treatment mechanisms from anti-TNFs. Similar to anti-TNFs, higher levels of vedolizumab and ustekinumab are associated with increased rates of clinical remission; however, rates of ADA development seem to be lower for vedolizumab and ustekinumab, and ADAs do not always seem to be closely correlated with clinical outcomes (4–10).

The risks and benefits of adding an immunomodulator to vedolizumab or ustekinumab remain unclear. A growing body of data including a recent meta-analysis of 16 studies and 1 prospective study did not show clinical benefit with this practice (11–13). Conversely, a recent, large 2-nation population-based study did show benefit of thiopurines in combination with vedolizumab, although statistically significant only in patients with CD treated with vedolizumab (14). In contrast to anti-TNFs, vedolizumab and ustekinumab carry a uniquely low or no increased risk of serious and opportunistic infections (15,16). Immunomodulators have been shown to be associated with numerous potential complications, including lymphoma, nonmelanoma skin cancers, and serious infections, with serious infections being of particular concern in those aged 65 years or older (17–19). The favorable safety profile of vedolizumab and ustekinumab could therefore be compromised by the addition of an immunomodulator.

While more large-scale prospective data are needed, the known risk of adding an immunomodulator with ustekinumab or vedolizumab without clear benefit warrants further investigation of current prescribing practices. In this study, we aimed to assess the prevalence and predictors of immunomodulator use as combination therapy with vedolizumab or ustekinumab in a real-world cohort of patients with IBD. In assessing these real-world practice patterns, we hypothesized that combination therapy would be used commonly, especially in those with prior exposure to anti-TNFs, younger age, and certain disease characteristics associated with more difficult-to-treat disease, including perianal involvement. In addition, we hypothesized that the use of combination therapy with these novel biologics would decrease with time from the US Food and Drug Administration (FDA) approval as more data emerged regarding rates of ADA formation and immunogenicity.

METHODS Data source

TARGET-IBD is a prospective longitudinal observational cohort of consented patients with IBD receiving usual care at 34 centers (13 community, 21 academic) across the United States. The structure of TARGET-IBD has been described elsewhere and is similar to that of other TARGET observational cohorts (TARGET-NASH, TARGET-HCC, TARGET-DERM, TARGET-PBC, TARGET-HBV), which aim to describe real-world diagnosis, management, and disease course (20–23). Adults (aged 18 years or older) and children with a diagnosis of CD, UC, or indeterminate colitis as determined by a treating physician and who are receiving any prescription therapy for IBD are eligible for inclusion in TARGET-IBD. Patients unable to provide written informed consent/assent, enrolled in any interventional study for IBD therapy, those with a history of prior total abdominal colectomy for UC, or those not meeting the inclusion criteria are excluded. Central institutional review board approvals are obtained prior to enrollment. Patients provide informed consent for up to 3 years of retrospective redacted medical information and up to 5 years of prospective data. Redacted medical records from participating sites are uploaded every 6 months into a secured data repository. There are no predetermined interventions or assessments.

For the purposes of this study, clinical information for adults enrolled in TARGET-IBD between July 2017 and November 2020 was ascertained from electronic medical records by trained abstractors and stored in a secured repository. Extraintestinal manifestations (EIMs) were identified using a text search that included joint EIMs (ankylosing spondylitis, enteropathic arthritis, arthropathy, joint ankylosis, peripheral arthritis, sacroiliitis, and spondyloarthropathy), ocular EIMs (anterior uveitis, episcleritis, iritis, scleritis, and uveitis), and cutaneous EIMs (erythema nodosum, pyoderma, and pyoderma gangrenosum).

Patients were included in the combination therapy group if they were on vedolizumab or ustekinumab with an immunomodulator (methotrexate, azathioprine, or 6-mercaptopurine) at any point in the retrospective or prospective period, with the exception of the analysis of prescription practices by year, where patients were considered to be on combination therapy only if an immunomodulator was used with vedolizumab or ustekinumab within the specified period.

Outcomes

The primary outcome of this study was the prevalence of combination therapy with ustekinumab and vedolizumab in our real-world IBD cohort. We compared this with the prevalence of combination therapy with anti-TNFs in the same cohort.

Secondary outcomes included predictors of combination therapy with vedolizumab and ustekinumab, stratified by disease type (UC and CD). Examined predictors included type of practice (academic vs community), patient age, sex, race, ethnicity, body mass index (BMI), tobacco use, disease duration, previous biologic exposure(s), and previous steroid use, and for CD, history of IBD surgery, disease location, and history of perianal disease. Additional secondary outcomes included changes in practices over time (within 2 years of the US FDA drug approval and after 2 years of the US FDA drug approval) and relationship of EIMs to combination therapy use with vedolizumab and ustekinumab.

Statistical analysis

Rates of anti-TNF, vedolizumab, and ustekinumab monotherapy and combination therapy at any time during the study period and during enrollment were estimated. Continuous variables were summarized using the mean and SD. Categorical variables were presented as counts and proportions. Rates of combination therapy were stratified by disease type (UC vs CD) and time from drug approval. Patient characteristics were compared using the Cochran-Mantel Haenszel general association test for categorical variables and the Kruskal-Wallis test for continuous variables. All statistical analyses were performed using SAS version 9.4 statical software (SAS Institute, Cary, NC).

RESULTS

We identified 4,039 patients with IBD, of whom 60.9% had CD. In total, 18.8% of patients were treated with vedolizumab, and 13.0% were treated with ustekinumab. Most of the patients treated with vedolizumab or ustekinumab had prior anti-TNF exposure (73.1% vedolizumab and 89.2% ustekinumab, Table 1), and many patients on ustekinumab had both anti-TNF and vedolizumab exposure (27.5%). The proportion of patients on combination therapy with vedolizumab and ustekinumab exceeded 30% (30.7% combination therapy with vedolizumab, 36.2% combination therapy with ustekinumab), whereas 41.1% of patients treated with anti-TNFs were on combination therapy.

Table 1. - Demographics and clinical characteristics of patients in TARGET-IBD treated with biologic therapy Treatmenta Anti-TNF (N = 2075) Vedolizumab (N = 759) Ustekinumab (N = 527) Combination therapy (all patients), n (%) 853 (41.1) 233 (30.7) 191 (36.2)  Combination therapy (patients with CD), n (%) 604 (41.9) 146 (33.1) 187 (36.7)  Combination therapy (patients with UC), n (%) 249 (39.3) 87 (27.4) 4 (23.5) Crohn's disease, n (%) 1,441 (69.4) 441 (58.1) 510 (96.8) Age at therapy start (yr)b, n (%)  <65 1931 (93.3) 659 (86.8) 494 (93.7)  ≥65 138 (6.7) 100 (13.2) 33 (6.3)  Female, n (%) 1,115 (53.7) 429 (56.5) 312 (59.2) Race, n (%)  White 1736 (83.7) 654 (86.2) 438 (83.1)  Other 265 (12.8) 82 (10.8) 68 (12.9)  Not reported 74 (3.6) 23 (3.0) 21 (4.0) Academic site, n (%) 1,485 (71.6) 544 (71.7) 424 (80.5) Prior anti-TNF exposurec, n (%) 555 (73.1) 470 (89.2) Prior anti-TNF + prior vedolizumab exposurec, n (%) 145 (27.5) No. of previous biologic exposuresc, n (%)  0 1,452 (70.0) 193 (25.4) 50 (9.5)  1 419 (20.2) 262 (34.5) 135 (25.6)  2 142 (6.8) 195 (25.7) 158 (30.0)  3 or more 62 (3.0) 109 (14.4) 184 (34.9) Any previous steroid usec, n (%) 1,127 (54.3) 551 (72.6) 408 (77.4) Steroid use ongoing at enrollment, n (%) 527 (25.4) 250 (32.9) 172 (32.6) History of IBD surgery (among patients CD), n (%) 606 (42.1) 214 (48.5) 246 (48.2) Crohn's disease location, n (%)  n 1,441 441 510  Colon 238 (16.5) 86 (19.5) 72 (14.1)  Ileocolon 718 (49.8) 202 (45.8) 281 (55.1)  Ileum 347 (24.1) 108 (24.5) 121 (23.7)  Not reported 138 (9.6) 45 (10.2) 36 (7.1) History of perianal disease (among patients with CD), n (%) 561 (38.9) 168 (38.1) 194 (38.0)

CD, Crohn's disease; IBD, inflammatory bowel disease; TNFs, tumor necrosis factor therapies; UC, ulcerative colitis.

aIncludes treatment use at any time during the retrospective or prospective periods. Participants may be included in more than 1 column. All patients included.

bBased on earliest known date of use of treatment (anti-TNF, vedolizumab, or ustekinumab) from corresponding column.

cUse prior to the earliest known date of use of treatment (anti-TNF, vedolizumab, or ustekinumab) from corresponding column.

Predictors of combination therapy

When examining predictors of combination therapy with vedolizumab or ustekinumab, we found that patients who had any previous biologic exposure were significantly more likely to use combination therapy (P < 0.001 for vedolizumab in CD, P = 0.045 for vedolizumab in UC, P < 0.001 for ustekinumab in CD, Table 2). Furthermore, additional biologic exposures were associated with an increased likelihood of combination therapy (P < 0.001). Combination therapy was also more likely in patients with perianal disease (P = 0.02 for vedolizumab, P = 0.03 for ustekinumab). Older patients were significantly less likely to be on combination therapy with ustekinumab in CD (P = 0.01); however, this relationship was not seen for patients with UC or CD on vedolizumab. No significant difference in combination therapy use was seen for sex, race, BMI, tobacco use, duration of disease, or steroid use with either vedolizumab or ustekinumab. In addition, in patients with CD, a history of IBD surgery and location of CD were not associated with the use of combination therapy. Last, the presence of EIMs was only associated with increased rates of combination therapy for those patients with cutaneous EIMs on ustekinumab (Table 3). Patients with UC treated with ustekinumab were excluded from these analyses because of small numbers (n = 17) in the setting of more recent approval for this indication.

Table 2. - Demographics and clinical characteristics of patients in TARGET-IBD treated with vedolizumab and ustekinumab Summary Vedolizumab-Crohn's disease Ustekinumab-Crohn's disease Vedolizumab-ulcerative colitis Combination therapya (N = 146) Monotherapy (N = 295) P valueb Combination therapy (N = 187) Monotherapy (N = 323) P value Combination therapy (N = 87) Monotherapy (N = 231) P value Age at treatment start, n (%)c  <65 131 (89.7) 253 (85.8) 0.2435 182 (97.3) 297 (92.0) 0.0144 76 (87.4) 198 (85.7) 0.7058  ≥65 15 (10.3) 42 (14.2) 5 (2.7) 26 (8.0) 11 (12.6) 33 (14.3) Female, n (%) 81 (55.5) 183 (62.0) 0.1868 117 (62.6) 186 (57.6) 0.2701 40 (46.0) 125 (54.1) 0.1962 Race, n (%)  White 128 (87.7) 256 (86.8) 0.5677 159 (85.0) 264 (81.7) 0.2278 77 (88.5) 193 (83.5) 0.5439  Other 12 (8.2) 31 (10.5) 24 (12.8) 42 (13.0) 8 (9.2) 31 (13.4)  Not reported 6 (4.1) 8 (2.7) 4 (2.1) 17 (5.3) 2 (2.3) 7 (3.0) Academic site, n (%) 106 (72.6) 222 (75.3) 0.5488 147 (78.6) 265 (82.0) 0.3434 62 (71.3) 154 (66.7) 0.4343 BMI at treatment start (kg/m2)  Mean (SD) 25.9 (5.9) 26.5 (7.5) 0.4364 27.0 (6.2) 26.9 (6.9) 0.4856 30.0 (1.4) 25.6 (4.4) 0.4389 Tobacco use at enrollment, n (%)  Current smoker 14 (9.6) 37 (12.5) 0.7926 20 (10.7) 41 (12.7) 0.0545 2 (2.3) 5 (2.2) 0.9444  Former smoker 43 (29.5) 79 (26.8) 34 (18.2) 89 (27.6) 21 (24.1) 59 (25.5)  Never smoker 86 (58.9) 172 (58.3) 126 (67.4) 186 (57.6) 62 (71.3) 159 (68.8)  Unknown 3 (2.1) 7 (2.4) 7 (3.7) 7 (2.2) 2 (2.3) 8 (3.5) Duration of disease at treatment start (yr)  Mean (SD) 15.6 (11.0) 15.4 (12.6) 0.4120 14.2 (9.5) 14.0 (10.7) 0.4486 8.1 (7.5) 8.7 (8.8) 0.9965 Any previous biologic exposure, n (%) 133 (91.1) 230 (78.0) 0.0007 181 (96.8) 283 (87.6) 0.0005 65 (74.7) 145 (62.8) 0.0453 No. of previous biologic exposures, n (%)  0 13 (8.9) 65 (22.0) <0.0001 6 (3.2) 40 (12.4) <0.0001 22 (25.3) 86 (37.2) <0.0001  1 24 (16.4) 87 (29.5) 26 (13.9) 89 (27.6) 21 (24.1) 104 (45.0)  2 44 (30.1) 86 (29.2) 55 (29.4) 95 (29.4) 32 (36.8) 36 (15.6)  3 or more 65 (44.5) 57 (19.3) 100 (53.5) 99 (30.7) 12 (13.8) 5 (2.2) Any history of steroid use at treatment start, n (%) 97 (66.4) 203 (68.8) 0.6151 145 (77.5) 252 (78.0) 0.9003 66 (75.9) 192 (83.1) 0.1411 Steroid use ongoing at treatment start, n (%) 47 (32.2) 95 (32.2) 0.9980 81 (43.3) 112 (34.7) 0.0528 45 (51.7) 116 (50.2) 0.8108 History of IBD surgery at treatment start, n (%)d 65 (44.5) 128 (43.4) 0.8220 91 (48.7) 146 (45.2) 0.4505 Location of Crohn's disease at treatment start, n (%)  Colon 26 (17.8) 48 (16.3) 0.7122 27 (14.4) 44 (13.6) 0.4517  Ileocolon 60 (41.1) 118 (40.0) 102 (54.5) 159 (49.2)  Ileum 36 (24.7) 67 (22.7) 36 (19.3) 82 (25.4)  Not reported 24 (16.4) 62 (21.0) 22 (11.8) 38 (11.8) History of perianal disease at treatment start, n (%) 56 (38.4) 81 (27.5) 0.0201 77 (41.2) 103 (31.9) 0.0346

Bold entries are statistically significant P values.

BMI, body mass index; IBD, inflammatory bowel disease.

aPatients are classified as being on combination therapy if they were on combination therapy at any point during the retrospective or prospective periods and classified as being on monotherapy if they were not on combination therapy. Combination therapy includes the use of vedolizumab or ustekinumab concurrently with methotrexate, azathioprine, or 6-mercaptopurine.

bP values for categorical variables are from the Cochran-Mantel-Haenszel general association test, and P values for continuous variables are from the Kruskal-Wallis test.

cTreatment start is defined as the start date of combination therapy for patients on combination therapy and defined as the start date of vedolizumab or ustekinumab for patients on monotherapy.

dHistory of IBD surgery includes the following: small intestinal resection, Ileocecal resection, Colectomy, Colectomy total, Ileocolectomy, Proctocolectomy, Ileostomy, Jejunostomy, Colorectostomy, Colostomy, Enterostomy, Ileocolostomy, Rectal fistula repair, Anal fistula excision, Anal fistula, Anal fistula repair, Colon fistula repair, Gastrointestinal fistula repair, Intestinal fistula repair, Abscess drainage, Abscess management, and Perirectal abscess.

Table 3. - Extraintestinal manifestations of patients enrolled in TARGET-IBD and treated with vedolizumab and ustekinumab Extraintestinal manifestationsa Vedolizumab-Crohn's disease Ustekinumab-Crohn's disease Vedolizumab-ulcerative colitis Combination therapy (N = 146) Monotherapy (N = 295) P value Combination therapy (N = 187) Monotherapy (N = 323) P value Combination therapy (N = 87) Monotherapy (N = 231) P value Jointb 33 (22.6%) 69 (23.4%) 0.8538 40 (21.4%) 70 (21.7%) 0.9407 11 (12.6%) 45 (19.5%) 0.1543 Ocularc 24 (16.4%) 47 (15.9%) 0.8919 33 (17.6%) 45 (13.9%) 0.2618 13 (14.9%) 31 (13.4%) 0.7263 Cutaneousd 31 (21.2%) 59 (20.0%) 0.7627 50 (26.7%) 57 (17.6%) 0.0152 9 (10.3%) 25 (10.8%) 0.9023

Bold entries are statistically significant P values.

EIM, extraintestinal manifestation; IBD, inflammatory bowel disease.

P values for categorical variables are from the Cochran-Mantel-Haenszel general association test.

Vedolizumab and ustekinumab combination therapy and monotherapy are based on treatments ongoing at any time during the retrospective or prospective periods. Combination therapy includes the use of vedolizumab or ustekinumab concurrently with methotrexate, azathioprine, or mercaptopurine.

aIncludes EIMs noted at any time.

bJoint EIMs include ankylosing spondylitis, arthritis enteropathic, arthropathy, joint ankylosis, peripheral arthritis, reactive arthritis, sacroiliitis, and spondyloarthropathy.

cOcular EIMs include anterior uveitis, episcleritis, iritis, scleritis, and uveitis.

dCutaneous EIMs include erythema nodosum, pyoderma, and pyoderma gangrenosum.

Rates of combination therapy initiation

Rates of combination therapy initiation decreased over time. Rates of vedolizumab combination therapy initiation decreased from 27.6% when started in the first 2 years after the US FDA drug approval to 21.9% when started in subsequent years (25.0%–19.8% in UC, 28.7%–23.8% in CD). Rates of ustekinumab combination therapy initiation in CD decreased from 30.7% when started within 2 years of drug approval as opposed to 22.7% when started in subsequent years (Table 4).

Table 4. - Patterns of monotherapy and combination therapy use, analyzed by date of therapy initiation Summary Crohn's disease Crohn's disease Ulcerative colitis Vedo started 2014–2016 (N = 181) Vedo started 2017–2020 (N = 260) Ustek started 2016–2018 (N = 361) Ustek started 2019–2020 (N = 132) Vedo started 2014–2016 (N = 76) Vedo started 2017–2020 (N = 242) Treatment type, n (%)a  Monotherapy 129 (71.2) 198 (76.1) 250 (69.2) 98 (74.2) 57 (75.0) 194 (80.2)  Combination therapy with methotrexate 15 (8.3) 19 (7.3) 39 (10.8) 16 (12.1) 2 (2.6) 10 (4.1)  Combination therapy with thiopurine 37 (20.4) 43 (16.5) 72 (19.9) 18 (13.6) 17 (22.4) 38 (15.7)

Vedolizumab/ustekinumab start year is based on earliest known date of use of treatment.

aCombination therapy includes the use of vedolizumab or ustekinumab concurrently with methotrexate, azathioprine, or mercaptopurine, at any point during the retrospective or prospective periods.

DISCUSSION

Analysis of current prescribing practices revealed that combination therapy with vedolizumab and ustekinumab exceeded 30% in this cohort. For comparison, combination therapy with anti-TNFs, where supporting data are more robust, was used in just more than 40% of participants. The high prevalence of combination therapy with ustekinumab and vedolizumab in this population is notable, given the unclear benefit of this practice, the known low immunogenici