Comparative Risk of Serious Infection With Vedolizumab vs Anti-Tumor Necrosis Factor in Inflammatory Bowel Disease: Results From Nationwide Swedish Registers

Abstract

INTRODUCTION:

We aimed to assess the risk of serious infection in patients with inflammatory bowel disease (IBD) treated with vedolizumab compared with those treated with anti-tumor necrosis factors (TNF) and the general population.

METHODS:

In this Swedish cohort study, treatment episodes were identified from nationwide health registers. We used Cox regression with propensity score-matched cohorts to estimate hazard ratios (HRs) for incident serious infections, defined as infections requiring hospital admission.

RESULTS:

During 1,376 treatment episodes in Crohn's disease, the rate of serious infections per 100 person-years (PY) was 5.18 (95% CI = 3.98–6.63) with vedolizumab vs 3.54 (95% CI = 2.50–4.85) with anti-TNF; HR = 1.72 (95% CI = 1.12–2.65), partly explained by more gastrointestinal infections. Compared with the rate of 0.75/100 PY (95% CI = 0.59–0.92) in a matched general population cohort, vedolizumab demonstrated higher risk (HR = 7.00; 95% CI = 5.04–9.72). During 1,294 treatment episodes in ulcerative colitis, the corresponding rates were 3.74/100 PY (95% CI = 2.66–5.11) with vedolizumab vs 3.42/100 PY (95% CI = 2.31–4.89) with anti-TNF; HR = 0.80 (95% CI = 0.47–1.36) during the initial 1.1 years and HR = 2.03 (95% CI = 0.65–6.32) after 1.1 years (truncated due to nonproportional hazards). Pneumonia accounted for 40% of all infections among anti-TNF, whereas no case was observed among vedolizumab episodes. Compared with the rate of 0.69/100 PYs (95% CI = 0.53–0.87) in a matched general population cohort, vedolizumab showed an HR of 5.45 (95% CI = 3.67–8.11).

DISCUSSION:

Vedolizumab was associated with increased risks of serious infections compared with anti-TNF in Crohn's disease but not in ulcerative colitis. Nonetheless, the panorama of serious infections seemed to differ between the drugs. Our findings underscore the importance of clinical awareness of infections and the safety profile of the 2 therapies.

INTRODUCTION

The use of biologics in patients with inflammatory bowel disease (IBD) has been linked to adverse effects including infections; this may especially be so for anti-tumor necrosis factor (TNF) treatment (1–5). Vedolizumab, which have a gut-specific mechanism of action by inhibiting migration of lymphocytes to the intestinal mucosa (6–8), is perceived to be a safer treatment alternative. This perception has been supported by a recent randomized controlled trial (RCT) comparing vedolizumab with adalimumab in ulcerative colitis (UC), where the rate of serious infection was lower in vedolizumab users (1.6 vs 2.2 per 100 person-years [PY]) (5).

However, clinical trials often involve selected patient populations and suffer from limited follow-up periods. Therefore, evaluating safety outcomes in real-world cohorts is also essential, but evidence from real-world data is conflicting (see Supplementary Table 1, Supplementary Digital Content 1, http://links.lww.com/AJG/D344). According to a recent meta-analysis, vedolizumab may offer a net benefit over anti-TNF in UC but not in Crohn's disease (CD) (9). However, the majority of earlier studies have suffered from various methodological limitations, including small sample size, short follow-up period, restriction to single centers, different inclusion criteria, varied data sources, imprecise and varied definitions of IBD and serious infections, and inadequate confounding control (10–18). Particularly, the variability in clinical characteristics and disease severity among different treatment groups in previous research impedes the interpretation of the effect estimates. Various study designs address different questions. Previous RCTs and observational studies have aimed to determine whether different advanced treatments (e.g. vedolizumab or anti-TNF agents) differ with regard to risk of serious infections. However, it is imperative to also contextualize the absolute and relative risk of serious infections in patients with IBD requiring immunosuppressive drugs compared with individuals without IBD not undergoing such treatments.

To address several of these limitations, we conducted a nationwide cohort study of vedolizumab-treated patients with IBD, comparing both overall and organ-specific risks of serious infection, to risks in propensity score-matched anti-TNF-treated patients and also among matched individuals in the general population without IBD.

METHODS

Setting

This was a nationwide register study based on data from Swedish routine clinical practice. We used the unique personal identity number to link data from Swedish health care registers (19). The International Classification of Diseases (ICD) codes were obtained from the National Patient Register (NPR) (20). Additional data on medical treatment were collected from the Prescribed Drug Register and the Swedish National Quality Register for Inflammatory Bowel Disease (SWIBREG) (21,22). SWIBREG currently includes >60,000 patients with IBD, corresponding to >85% of all patients with IBD receiving advanced therapies in Sweden. The register contains clinical variables such as date of diagnosis, disease phenotype, and medications, including infusions. Four hospitals have 100% coverage and 12 have >90% coverage, using the NPR as the gold standard (23). We used the Total Population Register to identify matched controls in the general population without IBD (24).

Participants

Patients with CD and UC were identified from the NPR and SWIBREG, only including centers with >70% coverage in SWIBREG (25). The diagnosis was defined as having at least one relevant ICD code in the NPR along with at least one diagnosis of IBD in SWIBREG (see Supplementary Table 2, Supplementary Digital Content 1, http://links.lww.com/AJG/D344). A previous patient chart validation study found that >99% of patients fulfilling this definition had true IBD (26,27). Patients with diagnoses who shifted between CD and UC but only had one of the diagnoses in the 5 years preceding the start of follow-up were classified according to their most recent diagnosis. Otherwise, we classified patients with a mix of codes as inflammatory bowel disease unclassified. However, patients with a diagnostic or procedure code typical of CD (see Supplementary Table 3, Supplementary Digital Content 1, http://links.lww.com/AJG/D344) were classified as having CD (28).

Data from SWIBREG and the NPR were used to categorize the IBD disease phenotype according to the Montreal classification (see Supplementary Table 4, Supplementary Digital Content 1, http://links.lww.com/AJG/D344) (29). Medical treatment was recorded based on Anatomical Therapeutic Chemical (ATC) codes (see Supplementary Table 5, Supplementary Digital Content 1, http://links.lww.com/AJG/D344). Patients were included if they were exposed to vedolizumab or an anti-TNF (infliximab, adalimumab, or golimumab) during January 1, 2014, to December 31, 2020. Patients were excluded if they were younger than 18 years at treatment initiation, exposed to biologics before the date of diagnosis of IBD, migrated ≤5 years before baseline, or diagnosed at index date. Patients with a history of any of the following: human immunodeficiency virus infection, tissue or organ transplantation, cancer (except nonmelanoma skin cancer), immunodeficiency, hepatitis B, hepatitis C, or tuberculosis (see Supplementary Table 6, Supplementary Digital Content 1, http://links.lww.com/AJG/D344) were also excluded.

The Charlson comorbidity index (CCI) was used as a measure of comorbidity burden. We used a previously reported list of comorbidities to calculate CCI (see Supplementary Table 7, Supplementary Digital Content 1, http://links.lww.com/AJG/D344) (30). In brief, the index is derived from 17 comorbidities. Each comorbidity is weighted based on its association with mortality risk (range = 1–6), and the total score is calculated by adding the weighted subscores across all comorbidities. We categorized individuals based on their CCI score as no comorbidity (score = 0), mild comorbidity (score = 1–2), and severe comorbidity (score ≥3).

Exposure

From the cohort of individual patients, we identified episodes of treatment. The primary exposure of interest was vedolizumab. Patients could contribute to more than one treatment episode. They were considered exposed from the initiation of treatment (baseline) until 90 days after discontinuation of treatment. The exposure period was defined as days from baseline until the first infection, 90 days after termination of treatment, end of follow-up (i.e. December 31, 2020), emigration, or death, whichever occurred first. Treatment was classified as terminated if a date of discontinuation was recorded in SWIBREG, if more than 90 days passed since the last infusion was recorded in NPR, or if more than 180 days passed since the last prescription was recorded in the Prescribed Drug Register. For individuals captured in multiple registers, the earliest available start date and latest available stop date were used. Switching from a reference product to a biosimilar was not considered discontinuation.

Outcome

The primary outcome was any serious infections, defined as the first occurring hospital admission with a primary discharge diagnosis (or contributory, if the primary diagnosis was IBD) listing infection (see Supplementary Table 8, Supplementary Digital Content 1, http://links.lww.com/AJG/D344). The secondary outcomes were 3 site-specific serious infections: respiratory tract infections, gastrointestinal tract infections, and other infections. All outcomes were stratified by IBD subtype (CD and UC).

Matching

To control for potential confounders in comparisons between vedolizumab and anti-TNF, we performed 1:1 nearest-neighbor propensity matching without replacement. The propensity score for vedolizumab exposure was estimated using a logistic regression model including the following covariates recorded at baseline: age group, disease duration, Charlson comorbidity index category, sex, concomitant treatment with immunomodulators and corticosteroids, line of biological therapy, previous IBD-related surgery (see Supplementary Table 9, Supplementary Digital Content 1, http://links.lww.com/AJG/D344), and previous serious infection. The balance of confounder variables was assessed on the basis of standardized differences and plotting histograms of propensity scores before and after matching. We considered a covariate to be well balanced if the absolute standardized difference was <10% (31,32).

Furthermore, treatment episodes with vedolizumab were matched (using exact matching on sex, age in years, baseline date, and place of residency) with 10 general population reference individuals without IBD.

Statistical analysis

Incidence rates (IR) were calculated for each outcome. Crude IR are presented as events/100 person-years with 95% confidence interval (CI). Absolute rate differences were estimated using Poisson regression, and hazard ratios (HRs) with 95% CI were estimated with Cox proportional hazards regression. We used robust sandwich estimators to account for the clustering within matched sets. The proportional hazards assumption was tested with visual assessments of Kaplan-Meier curves, log(−log) plots, and testing of scaled Schoenfeld residuals. The model was corrected with time-axis division when applicable. Kaplan-Meier curves were plotted to visualize cumulative survival without serious infection. All tests were two-tailed, and P-values of <0.05 or 95% CIs not crossing 1 were considered statistically significant. Data processing and statistical analyses were performed in SAS version 9.4 software (SAS Institute Inc., Cary, NC) and Stata version 16.1 software (StataCorp. 2019. Stata Statistical Software: Release 16. College Station, TX: StataCorp LLC).

Prespecified subgroup analyses of the primary outcome were by sex, age group (18–60 and >60 years), previous anti-TNF exposure, and previous serious infection. As a sensitivity analysis, we evaluated the exposure window by using an on-treatment approach to follow-up where patients were only considered exposed between therapy start and therapy stop. An additional sensitivity analysis was performed to minimize misclassification, restricting the primary outcome to hospital admissions with infection as a primary discharge diagnosis only.

Ethical considerations

Ethical approval for this study was granted by the Regional Ethics Committee, Karolinska Institute, Stockholm, Sweden (2007/785-31/5; 2011/1509-32; 2012/601-32; 2015/0004-31; 2015/615-32).

RESULTS

Patient characteristics

Among 44,012 patients with IBD identified in the registers, 1,335 treatment episodes with vedolizumab (1,221 unique patients) and 1,335 treatment episodes with anti-TNF (1,180 unique patients) were included in the propensity-matched cohorts (Flowchart shown in Supplementary Figure 1, Supplementary Digital Content 1, http://links.lww.com/AJG/D344). Vedolizumab episodes were also matched to 11,818 equivalent reference periods in the general population. Patients treated with vedolizumab and anti-TNF were not materially different with regard to most demographic, disease, and treatment characteristics (Tables 1 and 2). Histograms of the propensity score before and after matching are available in the supplementary material (see Supplementary Figure 2, Supplementary Digital Content 1, http://links.lww.com/AJG/D344).

Table 1.

Baseline demographics and clinical characteristics of propensity score-matched cohorts of treatment episodes in Crohn's disease

Table 2.

Baseline demographics and clinical characteristics of propensity score-matched cohorts of treatment episodes in ulcerative colitis

Outcome in Crohn's disease

The median follow-up time was 1.22 (0.56–2.38) years among vedolizumab episodes (n = 688), 1.01 (0.49–2.10) years among anti-TNF episodes (n = 688), and 1.26 (0.61–2.55) years in the general population (n = 6,038). The main anti-TNF drug was adalimumab, 451 (66%) of the anti-TNF episodes used this treatment. The remaining episodes were represented by infliximab (n = 189; 27%) and golimumab use (n = 48; 7%). Serious infection occurred during follow-up among 63 (9%) vedolizumab episodes, 38 (6%) anti-TNF episodes, and 84 (1%) general population episodes (Tables 3 and 4). This resulted in incidence rates of 5.18 (3.98–6.63) serious infections per 100 PY with vedolizumab and 3.54 (2.50–4.85) per 100 PY with anti-TNF, corresponding to an HR of 1.72 (1.12–2.65; Figure 1a). Male sex, aged younger than 60 years, and previous anti-TNF treatment were associated with higher HRs of serious infection in vedolizumab users compared with anti-TNF users (Table 5).

Table 3.

Association between use of vedolizumab and risk of serious infections compared with anti-TNF

Table 4.

Association between use of vedolizumab and risk of serious infections compared with the general population

Figure 1.

Cumulative incidence of serious infection in propensity score-matched cohorts of vedolizumab and anti-TNF episodes with (a) Crohn's disease and (b) ulcerative colitis. anti-TNF, anti-tumor necrosis factor.

Table 5.

Additional analyses of association between vedolizumab use and risk of serious infection in propensity score-matched cohorts of treatment episodes in Crohn's disease

As an exploratory analysis, we separately compared vedolizumab vs infliximab and vedolizumab vs subcutaneous anti-TNF (golimumab and adalimumab). The IR for serious infection among infliximab episodes was 2.77 (1.33–5.10), and the corresponding rate for subcutaneous anti-TNF was 3.92 (2.61–5.67). The CIs overlapped with the IR in vedolizumab episodes (5.18 [3.98–6.63]). When examining site-specific infections, an incidence rate difference of 0.87 (0.09–1.66) serious gastrointestinal infections per 100 PY with vedolizumab vs anti-TNF was observed, corresponding to an HR of 2.47 (0.96–6.39, Table 3). The most common specific serious infection among vedolizumab and anti-TNF episodes was pneumonia (Figure 2a). Infection with Clostridioides difficile occurred in 3 vedolizumab episodes and 1 anti-TNF episode. All subtypes of serious gastrointestinal infections are listed in Supplementary Table 10 (Supplementary Digital Content 1, http://links.lww.com/AJG/D344).

Figure 2.

Specific infectious diagnoses observed in propensity score-matched cohorts of vedolizumab and anti-TNF episodes with (a) Crohn's disease and (b) ulcerative colitis. anti-TNF, anti-tumor necrosis factor.

Compared with the general population (IR per 100 PY: 0.75; 0.59–0.92), vedolizumab users had a 7.00-fold higher hazard of serious infection (HR = 7.00; 5.04–9.72; Table 4). There were statistically significant associations between vedolizumab treatment and all 3 site-specific serious infections, with the HR of gastrointestinal tract infections being of highest magnitude (HR: 31.92; 11.85–85.94; Table 4).

Sensitivity analyses

Restricting the exposure window by using an on-treatment approach to follow-up and ending exposure at the date of stopping treatment or defining the primary outcome as hospital admissions with infection as a primary discharge diagnosis only did not influence our HRs notably (Table 5).

Outcome in ulcerative colitis

The median follow-up time was 1.05 (0.51–2.31) years among vedolizumab episodes (n = 647), 0.79 (0.45–1.69) years among anti-TNF episodes (n = 647), and 1.06 (0.51–2.34) years in the general population (n = 5,780). The main anti-TNF drug was adalimumab, 347 (53%) of the anti-TNF episodes used this treatment. The remaining episodes were represented by infliximab (n = 239; 37%) and golimumab use (n = 61; 9%). Serious infection occurred during follow-up among 39 (6%) vedolizumab episodes, 30 (5%) anti-TNF episodes, and 66 (1%) general population episodes (Tables 3 and 4). This resulted in an incidence rate of 3.74 (2.66–5.11) serious infections per 100 PY with vedolizumab and 3.42 (2.31–4.89) per 100 PY with anti-TNF (Table 3). The HR until 1.1 years was 0.80 (0.47–1.36) and 2.03 (0.65–6.32) after 1.1 years (Figure 1b). No significant association of serious infection was observed when analyses were stratified by prespecified subgroups of UC (Table 3).

As an exploratory analysis, we separately compared vedolizumab vs infliximab and vedolizumab vs subcutaneous anti-TNF (golimumab and adalimumab). The IR for serious infection among infliximab episodes was 3.70 (1.97–6.32), and the corresponding rate for subcutaneous anti-TNF was 3.24 (1.19–5.19). The CIs overlapped with the IR in vedolizumab episodes 3.74 (2.66–5.11). Compared with anti-TNF episodes, there was no significant association between vedolizumab treatment and any of the 3 site-specific serious infections (Table 3). Gastroenteritis (33%) was the most frequent specific infection in vedolizumab episodes, whereas pneumonia (40%) was the most common infection in anti-TNF episodes. In addition, no case of pneumonia was observed among vedolizumab episodes (Figure 2b). Infection with C. difficile occurred in 5 vedolizumab episodes and 2 anti-TNF episodes. All subtypes of serious gastrointestinal infections are listed in Supplementary Table 10 (Supplementary Digital Content 1, http://links.lww.com/AJG/D344).

Compared with the general population (IR per 100 PY: 0.69; 0.53–0.87), vedolizumab users had 5.45-fold higher hazard of serious infection (HR = 5.45; 3.67–8.11; Table 4). Vedolizumab episodes carried a statistically significant association with all 3 site-specific infections, and the highest HR was observed for serious gastrointestinal infections (Table 4).

Sensitivity analyses

Sensitivity analyses confirmed our main findings, and there was no statistically significant difference in risk of serious infections between vedolizumab and anti-TNF use (Table 6).

Table 6.

Additional analyses of association between vedolizumab use and risk of serious infection in propensity score-matched cohorts of treatment episodes in ulcerative colitis

DISCUSSION

In this propensity score-matched nationwide cohort study, based on data from routine clinical practice, vedolizumab use was linked to an increased risk of serious infections in patients with CD compared with anti-TNF treatment. This was in part driven by an increased risk of gastrointestinal infections. However, no significant association was observed in UC. Importantly, vedolizumab episodes were only associated with serious gastrointestinal infections and upper respiratory infections, whereas pneumonia was common during anti-TNF episodes. Compared with the general population, vedolizumab users were at increased risk of serious infections in both CD and UC.

The results from existing data on the risk of serious infections in patients with IBD treated with these biologics are mixed (5,10–18,33–39). A recent meta-analysis of vedolizumab vs anti-TNF reported odds ratios for serious infections of 1.03 (95% CI: 0.78–1.35) in CD and of 0.68 (95% CI: 0.56–0.83) in UC (9). However, the reported odds for serious infections in CD was driven by a pronounced decreased risk with vedolizumab in 2 studies. Using an American claims database, Kochar et al (15) identified an adjusted HR of 0.33 (95% CI: 0.13–0.85) in a retrospective cohort study of elderly patients (aged ≥65 years). Similarly, Bressler et al (37). reported an adjusted HR of 0.26 (95% CI: 0.08–0.87) in a 24-month retrospective medical chart study of adult patients. Remaining studies, including a large combined French and American propensity score-matched cohort study of 86,349 patients with IBD (33), reported numerically increased risk estimates in vedolizumab-treated patients or comparable risks. Some previous observational studies in UC have found associations between vedolizumab treatment and risk of serious infections (15,33), reporting HR down to 0.54 (40), even though differences have not been statistically significant in most reports (16,37).

The discrepancy between our results and the meta-analysis could be attributed to several factors. The risk of drug-related serious infections is determined by the drugs, immunosuppressive effect, and its effectiveness to control inflammation. Thus, the risk-benefit ratio may vary across cohorts due to differences in clinical characteristics, follow-up periods, disease severity, and healthcare systems (41). Intriguingly, we observed a nonproportional relative risk of serious infections in patients with UC treated with vedolizumab vs anti-TNF, with a numerically decreased HR during the first 1.1 years that shifted to a numerically elevated HR beyond this period. The early period finding aligns with the results from the VARSITY study (5). However, our findings raise the question of whether the ability of vedolizumab and anti-TNF therapy to control inflammation differs over time in UC and between the 2 subtypes of IBD. The risk of vedolizumab-associated serious infection seems to differ between CD and UC that may potentially be explained by different patient characteristics. Unfortunately, we had limited data on several factors linked to infections during treatment episodes, such as changes in corticosteroid treatment, frailty, endoscopic activity, therapeutic drug monitoring, and inflammatory markers, such as C-reactive protein and calprotectin. Especially, the lack of measures of disease activity limits our conclusions.

As off-label treatment is commonly used in Sweden, we included golimumab episodes in Crohn's disease. This indication is not approved, but our research group has previously reported on clinical outcomes of golimumab treatment in Crohn's disease (42). Differences in mechanisms of action can result in varying risk estimates for different subtypes of infections. Consistent with some previous findings, we observed an HR of 2.47 (95% CI: 0.96–6.39) for gastrointestinal infections among vedolizumab-treated patients with CD. Sing et al observed a 2.9-fold increased risk of serious gastrointestinal infection in patients with CD treated with vedolizumab compared with anti-TNF (40). In a RCT, C. difficile infections were more frequently reported in patients with UC randomized to vedolizumab compared with adalimumab (5). This is congruent with the observed association between vedolizumab and hospital admissions due to gastroenteritis in patients with UC in our cohort, while pneumonia was more common in anti-TNF-treated patients. However, it remains unknown whether the gut-specific mechanisms of vedolizumab predispose for gastrointestinal infections or if residual disease activity contributes to this risk. Interestingly, pneumonia accounted for 40% of the serious infections among anti-TNF episodes in ulcerative colitis, while it was 0% in vedolizumab episodes, where all respiratory infections affected the upper respiratory tract aside from coronavirus disease 2019. The distribution of specific infections was more similar between vedolizumab and anti-TNF episodes in Crohn's disease except for other viral infections that only occurred in anti-TNF episodes (8%). This ratio of viral infection was conversely in ulcerative colitis. However, the number of other viral infections was low (3 cases), so it is difficult to draw any conclusions from this finding.

To enhance the clinical interpretation of our results, we also contrasted vedolizumab users with matched reference individuals from the general population. High HRs for serious infections were consistently noted across all comparisons, especially for serious gastrointestinal infections.

Magnitudes of associations within subgroups of patients with CD seemed to differ. Higher HRs of vedolizumab vs anti-TNF were observed in patients previously exposed to anti-TNF agents, male patients, and in those aged 18–60 years. The association with prior anti-TNF therapy may suggest vedolizumab being less effective in controlling inflammation in treatment-refractory CD. While differences between male and female sex is puzzling, it is important to recognize that infections pose a significant concern in elderly patients. Consistent with prior literature (9,10,12,18,43), we found no differential risk of serious infection between vedolizumab and anti-TNF use in patients older than 60 years.

As a clinician, it is important to be aware of the increased risk of serious infections in patients with IBD, linked to both immunosuppressive therapies and disease activity, which this study highlights. Overall, patients with IBD on biological therapies represent a fragile population that is at increased risk of infections compared with persons without IBD. However, associations of serious infections may differ between various types of biological therapies, even in patients with similar disease severity. In our cohort, the increased risk of serious infections with vedolizumab in Crohn's disease seemed partly explained by more frequent gastrointestinal infections such as C. difficile, which are generally less severe than lower respiratory tract infections. This association should be considered when starting treatment in patients with previous C. difficile infections. Regarding respiratory tract infections in ulcerative colitis, pneumonia only occurred in anti-TNF episodes, whereas only upper respiratory tract infections and coronavirus disease 2019 were observed in vedolizumab episodes. These findings should be taken into account when initiating biological therapy in frail patients with IBD. Strengths of our study include the use of prospectively recorded real-world data from nationwide registers and the analyses of a period where both treatments were available. The large sample size and comprehensive information about covariates enabled rigorous confounding control. However, like in all observational studies, analyses are still susceptible to confounding by indication. We cannot control for potential unmeasured confounders including smoking that may have influenced the risk of, e.g., respiratory infections. Confounding of the association of serious infections may also have occurred due to potential differences in concomitant use of corticosteroids or immunomodulators during follow-up since combination therapy is known to increase the risk of infections (44,45). The use of data on prescribed medication from the Swedish Prescribed Drug Register and on infusion from centers with >70% coverage in SWIBREG reduce the risk of exposure misclassification. To specifically evaluate the impact of treatment on risk of infection, we also excluded patients with comorbidities strongly associated with infections. Misclassification of the outcome, i.e., serious infections, cannot be entirely excluded, but high positive predictive values for defining outcomes from ICD codes in the Swedish national patient register have been reported (20). Finally, the number of events was relatively low in some subsets of data, resulting in some imprecise subgroup and site-specific analyses.

In this nationwide cohort study, we found that vedolizumab was associated with increased risk of any serious infection compared with anti-TNF in CD, but not in UC. The observed association in patients with CD seemed to be partly explained by more gastrointestinal infections, which are generally less severe infections than lower respiratory tract infections. In ulcerative colitis, pneumonia accounted for 40% of the serious infections among anti-TNF episodes, whereas it did not occur in vedolizumab episodes, where serious respiratory infections only included upper respiratory tract infections and COVID-19. Compared with the general population, vedolizumab users in both subtypes of IBD were at higher risk of serious infection. The results from this real-world comparative study underscore the importance of enforcing clinical awareness of infections in vedolizumab-treated patients and may help clinicians understanding of the optimal positioning of vedolizumab in the treatment of IBD.

CONFLICTS OF INTEREST

Guarantor of the article: Jonas Halfvarson, MD, PhD.

Specific author contributions: O.O.: Acquisition of data. S.K. and Y.C.: Data analysis, creating figures and tables. S.K. and J.H.: Drafting the article. Substantial contributions to the study design, method, interpretation of data, and revising manuscript critically for important intellectual all authors. All authors approved the final version of the article, including the authorship list.

Financial support: S.K. received ALF funding from Region Örebro County (OLL-972338 and OLL-986027). J.H. received grants from the Swedish Research Council (grant number 2020-02021).

Potential competing interests: S.K. received lecture fee from Takeda. C.E. received grant support lecture fee advisory board from BMS, Takeda, Janssen Cilag, Pfizer, Abbvie. J.F.L. has coordinated an unrelated study on behalf of the Swedish IBD quality register (SWIBREG). That study received funding from Janssen corporation. J.F.L. has also received financial support from MSD developing a paper reviewing national health care registers in China. J.F.L. has an ongoing research collaboration on celiac disease with Takeda. O.O. has been PI on projects at Karolinska Institutet financed by grants from Janssen, Pfizer, AbbVie, Bristol Myers Squibb, Takeda, and Ferring, and Karolinska Institutet has received fees for lectures and participation on advisory boards from Janssen, Ferring, Galapagos, Bristol Myer Squibb, Takeda, and Pfizer. O.O. also reports grants from Pfizer, Janssen, Galapagos, and AbbVie in the context of a national safety monitoring programs. J.H. served as speaker andor advisory board member for AbbVie, Aqilion, BMS, Celgene, Celltrion, Dr. Falk Pharma and the Falk Foundation, Eli Lilly, Ferring, Galapagos, Gilead, Hospira, Index Pharma, Janssen, MEDA, Medivir, Medtronic, MSD, Novartis, Pfizer, Prometheus Laboratories Inc., Sandoz, Shire, Takeda, Thermo Fisher Scientific, Tillotts Pharma, Vifor Pharma, UCB. J.H. also received grant support from Janssen, MSD, and Takeda. M.S., Y.C., and S.M. have no conflicts to declare.

Study Highlights.

WHAT IS KNOWN

• ✓ Use of biological treatments is associated with adverse effects including infections.

• ✓ Real-world data on risks of serious infections comparing vedolizumab with anti-tumor necrosis factor are inconclusive.

WHAT IS NEW HERE

• ✓ Vedolizumab was associated with a higher hazard ratio of serious infections compared with anti-tumor necrosis factor in Crohn's disease.

• ✓ Compared with the general population, vedolizumab users were much more likely to develop serious infections.

Supplementary Material

acg-119-2480-s001.docx ^{(215.1KB, docx)}