Vedolizumab is effective and safe in elderly inflammatory bowel disease patients: a binational, multicenter, retrospective cohort study

Abstract

Background

Immune modulating therapies are associated with an increased risk of infections and malignancies. This is of particular concern in elderly inflammatory bowel disease patients. This study aims to compare the safety and efficacy of vedolizumab between young and elderly inflammatory bowel disease patients.

Methods

A binational, multicentre, retrospective, cohort study was performed from 2015 to 2019. Patients who underwent treatment with vedolizumab and were followed for at least 14 weeks were studied. They were divided according to age into groups: 40 years or less or 60 years or older. Clinical and endoscopic responses at weeks 14 and 52 and infection development were compared between young and elderly inflammatory bowel disease patient groups.

Results

There were 144 patients (82 Crohn’s disease and 62 ulcerative colitis) in the elderly cohort and 140 patients (83 Crohn’s disease and 57 ulcerative colitis) in the young cohort. The average age was 70.2 ± 7.3 years and 29.6 ± 5.7 years, respectively. Clinical and endoscopic responses were comparable between the groups (week 52 remission of Crohn’s disease: 40% vs. 35%, P = 0.7; week 52 remission of ulcerative colitis: 48% vs. 51%, P = 0.84). Previous anti-tumour necrosis factor biological therapy was independently associated with poor clinical remission rates at week 52 (Crohn’s disease: odds ratio 0.23, 95% confidence interval 0.06–0.79; P = 0.02 and ulcerative colitis: odds ratio 0.10 95% confidence interval 0.01–0.74; P = 0.024). There were significantly more infections in the elderly cohort (2% vs. 12%, P = 0.002), none of which were fatal.

Conclusions

Vedolizumab is equally effective in elderly and young inflammatory bowel disease patients. The findings of this study demonstrate an increased risk of infections among the elderly treated with vedolizumab, which may be related to their age and underlying diseases.

Key summary

Established knowledge on this subject:

• Biological therapies are associated with increased infection and malignancy risk.

• Advanced age and significant comorbidities may increase this risk.

• Vedolizumab appears to be safe and effective in elderly populations.

New findings in this study:

• Vedolizumab appears to have increased infection rates in the elderly patient population.

• Vedolizumab is equally effective in elderly and young patients.

Introduction

Inflammatory bowel diseases (IBDs) are composed of Crohn’s disease (CD) and ulcerative colitis (UC), which are chronic, immune-mediated inflammatory diseases of the gastrointestinal tract. While immune-modulating drugs are effective treatments for IBD, they are associated with serious adverse events, such as infections and treatment-related malignancies.¹,²

Most patients with IBD are diagnosed in their youth, but up to 15% are over the age of 60 years at diagnosis, and they form a group that is increasing in size due to the ageing of the world’s population.³ Concerns over the safety, primarily infections and malignancies, in this population have been raised because of their advanced age and significant comorbidity burden. In addition, the elderly are frequently excluded from clinical trials and observational cohorts.⁴

Vedolizumab is a selective monoclonal antibody which inhibits the interactions of α4β7 integrin (found on the lymphocyte surface) with the cell adhesion molecule MAdCAM-1 (found on the intestinal mucosa endothelium), thus inhibiting the recruitment of lymphocytes into the inflamed tissue. There are few data on the use of vedolizumab in the elderly population. In a small case series of 29 patients over the age of 60 years, efficacy and safety were similar to real-world studies.⁵ Another study reviewed data from the GEMINI studies and stratified the results according to age.⁶ The latter study included 56 IBD patients over the age of 65 years and the results showed no difference in terms of safety or efficacy compared to younger patients. A comparison of the safety and efficacy of anti-tumour necrosis factor (TNF) agents to vedolizumab in an elderly population also revealed no difference in safety and efficacy between them.⁷ In a recent study, a comparison of 25 elderly (age >65 years) IBD patients to disease-matched younger controls (age  < 65 years) again indicated no difference in safety and efficacy.⁸ In addition, a recent multicentre study from the UK also found that vedolizumab is safe and effective in elderly patients and is comparable to age-unstratified vedolizumab-treated cohorts.⁹

The aim of this retrospective study was to determine whether there is any difference in the safety and efficacy of vedolizumab in a large cohort of elderly (>60 years) IBD patients compared to young ( < 40 years) controls.

Materials and methods

Study population and period

This is a binational, multicentre, retrospective, cohort study which enrolled patients from centres in Israel and Edinburgh, UK between 2015 and 2019. Consecutive elderly (>60 years of age) and young ( < 40 years of age) patients who were started on vedolizumab and completed at least 14 weeks of therapy were included. An additional cohort of 40 to 60-year-old patients was recruited in order to perform a secondary safety analysis. All consecutive patients meeting these inclusion criteria were included. All of the data were recorded as part of routine clinical care. A standardised electronic data collection pro-forma was used at all centres, and all demographic and relevant clinical and endoscopic data were extracted from the patients’ electronic medical files. The clinical data included age of disease onset, disease duration, disease phenotype, smoking history, previous medications, surgical history, extra intestinal manifestations and disease activity indices. Details regarding adverse events, detailed by the treating physician, were retrieved from the regular clinic follow-up notes. The endoscopic data included relevant endoscopic scores when detailed or the endoscopist’s impression of disease activity.

Definitions

The Harvey–Bradshaw index (HBI) was used to determine clinical activity in the CD patients at weeks 14 and 52. Response was defined as a decrease of three or more points and remission as a HBI less than 5. Endoscopic disease was assessed using the simple endoscopic score for Crohn’s disease (SES-CD) or assessment of the treating physician, and a decrease of 50% was considered an endoscopic response and a score of 2 of less as remission.

The partial Mayo score was used to determine clinical disease activity in the UC patients at weeks 14 and 52. Response was defined as a decrease of 2 or more points from baseline and remission as a score less than 2. Endoscopic disease was assessed using the Mayo endoscopic score or graded by the treating physician as either ‘no response’, ‘improvement’, or ‘remission’. A decrease of 1 or more in the Mayo endoscopic score was considered an endoscopic response, and a score of 0 or 1 was regarded as endoscopic remission.

Corticosteroid-free remission was defined as patients in clinical remission according to HBI (CD patients) or partial Mayo score (UC patients) who were not receiving any corticosteroid therapy at week 52 of treatment.

Ethical considerations

This study received ethics approval from all the participating medical centres’ institutional review boards, and informed consent was waived because the data were de-identified and individual patient data were not published.

Statistical analysis

Continuous variables presented as means ± SD were compared with an independent samples t-test. Categorical variables are presented as proportions and were compared with Fisher’s exact test. Multivariate logistic regression analysis was used to test the association between patient characteristics and treatment outcome in the total patient cohort as well as in each of the patient groups, controlling for universal confounders (age group, sex, disease duration, location and behaviour) as well as potential confounders which were distributed differently between the groups and were associated with treatment response, namely: smoking history and previous treatment history. A P value of less than 0.05 was considered significant in all analyses. The SPSS software package (version 25.0) was used for all analyses.

Results

Demographic characteristics

A total of 284 IBD patients who received at least 14 weeks of vedolizumab between 2015 and 2019 were included in the study. Of these, 144 patients (82 with CD and 62 with UC) were older than 60 years of age and 140 patients (83 with CD and 57 with UC) were younger than 40 years of age at study inclusion. Of these 284 patients, 222 patients reached week 52. At the time of the analysis 44 patients had not yet reached week 52 of treatment, 16 patients were lost to follow-up (10 patients (7%) in the elderly group and 6 (4%) in the young group, P = 0.44) and two patients decided to stop treatment after becoming pregnant. Patients who stopped therapy due to poor treatment response, disease complications or adverse events were included in the final week 52 analysis. The average age of the elderly group was 70.2 ± 7.3 years, and IBD was diagnosed at an average age of 60.1 ± 14 years. The average age of the young group was 29.6 ± 5.7 years, and the average age at IBD diagnosis was 20.2 ± 7.3 years. Disease duration was similar for the two groups (9.4 vs. 10.2 years, respectively, P = 0.5). The elderly patients in both the CD and UC groups had significantly more cardiovascular (CD: 38% vs. 5%, respectively, P < 0.001; UC: 47% vs. 7%, respectively, P < 0.001), metabolic (CD: 38% vs. 4%, respectively, P < 0.001; UC: 38% vs. 14%, respectively, P = 0.003) and respiratory (CD: 17% vs. 2%, respectively, P = 0.001; UC: 15% vs. 0%, respectively, P = 0.003) comorbidities. The mean duration of follow-up was similar between the elderly and young cohorts (35.9 ± 21 vs. 39.3 ± 16.6 weeks, P = 0.13). Significantly more young patients were previously treated with an anti-TNF agent (CD: 88% vs. 51%, P < 0.001; UC: 70% vs. 47%, P = 0.03). The reasons for stopping biological therapy were similar between the groups. The baseline characteristics of the two study groups are detailed in Tables 1 and 2.

Table 1.

Demographics of the Crohn’s disease cohort.

Variable	Young (n = 83)	Elderly (n = 82)	P value
Age ± SD	29 ± 5.47	70 ± 7.44	<0.001
Age at onset ± SD	19 ± 6.42	60 ± 14.66	<0.001
Duration of disease  ± SD	10 ± 5.5	10 ± 11.75	0.9
Sex
Male (%)	36 (43)	33 (40)	0.75
Disease location
L1 (%)	21 (25)	34 (41)	0.3
L2 (%)	15 (18)	18 (22)	0.56
L3 (%)	45 (54.2)	29 (35)	0.2
L4 (%)	3 (4)	2 (2)	n/s
Disease behaviour
B1 (%)	44 (53)	41 (50)	0.64
B2 (%)	21 (26)	29 (35)	0.24
B3 (%)	17 (20)	12 (15)	0.3
Active perianal disease (%)	12 (14)	11 (13)	0.67
Harvey–Bradshaw index at initiation  ± SD	8.4 ± 4.78	7.3 ± 3.12	0.09
Comorbidities
Cardiovascular (%)	5 (6)	31 (38)	<0.001
Metabolic (%)	3 (4)	31 (38)	<0.001
Solid tumour (%)	1 (1)	12 (15)	0.001
Haematological disease/malignancy (%)	1 (1)	4 (5)	0.2
Inflammatory disease (%)	1 (1)	8 (10)	0.18
Dermatological disease (%)	1 (1)	3 (4)	0.4
Respiratory disease (%)	2 (2)	14 (17)	0.001
Other (%)	3 (4)	4 (5)	n/s
Smoking history
Current (%)	14 (17)	21 (26)	0.19
Past (%)	6 (7)	22 (28)	0.001
Never (%)	63 (76)	29 (40)	<0.001
Previous surgery	28 (35)	31 (38)	0.5
Previous treatment
5-ASA (%)	23 (28)	30 (37)	0.3
Thiopurine (%)	70 (84)	43 (52)	<0.001
Steroids (%)	28 (34)	15 (18)	0.02
MTX (%)	26 (31)	16 (20)	0.07
Nutritional (%)	0 (0)	1 (1)	n/s
Other (%)	8 (10)	7 (9)	n/s
Previous biological therapy
Any (%)	73 (88)	42 (51)	<0.001
IFX (%)	60 (72)	33 (40)	0.002
ADA (%)	63 (76)	36 (44)	0.002
Goli (%)	0 (0)	1 (1)	0.45
UST (%)	3 (4)	2 (2)	n/s
Two or more agents (%)	51 (61)	28 (34)	0.005
Reason for stopping previous biological agent
Primary LOR (%)	20 (24)	12 (15)	0.53
Secondary LOR (%)	41 (49)	22 (27)	0.5
Adverse event (%)	13 (16)	12 (15)	0.65
Other (%)	2 (2)	0 (0)	n/s

SD: standard deviation; n/s: non-significant; disease location: L1 – ileal, L2 – colonic, L3 – ileocolonic, L4 – proximal; disease behaviour: B1 – inflammatory, B2 – fibrostenotic, B3 – penetrating; MTX: methotrexate; IFX: infliximab; ADA: adalimumab; Goli: golimumab; UST: ustekinumab; LOR: loss of response

Table 2.

Demographics of the ulcerative colitis cohort.

Variable	Young (n = 57)	Elderly (n = 62)	P value
Age  ± SD	30.6 ± 6.02	70 ± 7.09	<0.001
Age at onset  ± SD	22.5 ± 8	59.9 ± 13.2	<0.001
Duration of disease  ± SD	8.1 ± 6.4	10.12 ± 12.2	0.25
Sex
Male (%)	31 (54)	34 (55)	n/s
Disease extent
E1 (%)	7 (12)	8 (13)	n/s
E2 (%)	16 (28)	24 (39)	0.25
E3 (%)	31 (54)	37 (40)	0.14
Partial Mayo score at initiation  ± SD	4.84 ± 2.7	5.22 ± 1.75	0.41
Comorbidities
Cardiovascular (%)	4 (7)	28 (46.7)	<0.001
Metabolic (%)	8 (14)	23 (38.3)	0.003
Solid tumour (%)	2 (3.5)	7 (11.9)	0.17
Haematological disease/malignancy (%)	1 (1.8)	2 (3.3)	n/s
Inflammatory disease (%)	8 (14)	3 (5)	0.12
Dermatological disease (%)	1 (1.8)	2 (3.3)	n/s
Respiratory disease (%)	0 (0)	9 (15)	0.003
Other (%)	1 (1.8)	1 (2)	n/s
Smoking history
Current (%)	0 (0)	1 (2)	N/S
Past (%)	2 (3.5)	17 (27)	<0.001
Never (%)	53 (96)	36 (58)	<0.001
Previous surgery	0 (0)	4 (6.5)	0.12
Previous treatment
5-ASA (%)	47 (87)	58 (95)	0.19
Thiopurine (%)	45 (83)	30 (49)	<0.001
Steroids (%)	18 (33)	23 (38)	0.7
MTX (%)	1 (1.9)	10 (16.4)	0.01
Nutritional (%)	0 (0 )	0 (0)
Other (%)	0 (0)	0(0)	n/s
Previous biological therapy
Any (%)	37 (70)	25 (47)	0.03
IFX (%)	33 (62)	20 (38)	0.02
ADA (%)	13 (25)	12 (23)	n/s
Goli (%)	2 (3.8)	1 (2)	n/s
UST (%)	2 (3.8)	0 (0)	0.5
Two or more agents (%)	11 (20.8)	7 (13.2)	0.45
Reason for stopping previous biological agent
Primary LOR (%)	15 (40)	10 (31)	0.6
Secondary LOR (%)	19 (51)	14 (44)	0.63
Adverse event (%)	2 (5.4)	5 (16)	0.2
Other (%)	0 (0)	0 (0)	n/s

SD: standard deviation; n/s: non-significant; disease extent: E1 – proctitis, E2 – left sided, E3 – extensive; MTX: methotrexate; IFX: infliximab; ADA: adalimumab; Goli: golimumab; UST: ustekinumab; LOR: loss of response.

CD patients

There were no significant differences in the rates of clinical response or remission between young and elderly CD patients at weeks 14 and 52 of vedolizumab therapy (Figure 1). Corticosteroid-free remission (Figure 2) and endoscopic response and remission rates (Supplementary Figure 1) were also comparable across age groups. Young patients naive to biological therapy showed significantly higher rates of clinical response and remission compared to non-naive young patients at week 14 (100% vs. 59%, P = 0.04 and 71% vs. 29%, P = 0.03, respectively), but the response and remission rates at week 52 were similar. Elderly patients who were naive to biological therapy had a trend towards higher response rates at both weeks 14 and 52 (80% vs. 50%, P = 0.07 and 75% vs. 47%, P = 0.08, respectively), and showed significantly higher remission rates at both weeks 14 and 52 (52% vs. 10%, P < 0.001 and 69% vs. 31%, P = 0.03, respectively) (Supplementary Figure 2). In a multivariate analysis adjusting for age group, sex, disease duration, location, behaviour and smoking, previous use of an anti-TNF was independently negatively associated with clinical response at week 14 (odds ratio (OR) 0.24, 95% confidence interval (CI) 0.07–0.79; P = 0.02) (Table 3). There was a trend towards poorer clinical response among anti-TNF-exposed CD patients at week 52 (OR 0.34, 95% CI 0.1–1.18; P = 0.091) (Table 3). Prior anti-TNF therapy was independently associated with poor clinical remission rates at both weeks 14 and 52 (OR 0.12, 95% CI 0.04–0.39; P < 0.001 and OR 0.23, 95% CI 0.06–0.79; P = 0.02, respectively). Another multivariate analysis was performed in order to determine treatment outcomes in each of the age groups (young ≤40 years and elderly ≥60 years). Following adjustment for gender, smoking, disease duration and disease behaviour, it was found that among elderly CD patients, prior anti-TNF exposure, was independently associated with poor clinical remission rates at both weeks 14 and 52 (OR 0.07, 95% CI 0.014–0.39; P = 0.002 and OR 0.20, 95% CI 0.04–0.85; P = 0.03, respectively) (Table 3). This was not seen in the young CD cohort.

Figure 1.

Clinical response and remission rates in the young and elderly Crohn’s disease patients (a) and (b) and ulcerative colitis patients (c) and (d) cohorts at weeks 14 (a) and (c) and 52 (b) and (d).

Figure 2.

Corticosteroid-free clinical remission at week 52 in Crohn’s disease (a) and ulcerative colitis (b) cohorts.

Table 3.

Adjusted association between past anti-TNF exposure (vs. no exposure) and vedolizumab clinical outcomes.

	Crohn’s disease patients			Ulcerative colitis patients
	Total CD cohort OR (95% CI)*, Pv	Young OR (95% CI), Pv	Elderly OR (95% CI), Pv	Total UC cohort OR (95% CI)*, Pv	Young OR (95% CI), Pv	Elderly OR (95% CI), Pv
Clinical response at week 14
	(n=136)	(n=78)	(n=57)	n=97	(n=49)	(n=49)
Any past anti-TNF treatment	0.24 (0.07–0.79) 0.020	ns	0.37 (0.09–1.45) 0.156	0.25 (0.07–0.84) 0.025	0.16 (0.01–1.50) 0.109	0.34 (0.06–1.37) 0.123
One past anti-TNF treatment (vs. none)	0.25 (0.06–0.93) 0.039	ns	0.26 (0.08–1.60) 0.185	0.14 (0.04–0.50) 0.003	0.07 (0.007–0.74) 0.027	0.16 (0.03–0.84) 0.031
Two past anti-TNF treatment (vs. none)	0.20 (0.06–0.70) 0.012	ns	0.38 (0.09–1.61) 0.194	1.41 (0.14–13.79) 0.764	0.43 (0.02–8.34) 0.582	ns
Clinical remission at week 14
Any past anti-TNF treatment	0.12 (0.04–0.39) <0.001	0.16 (0.02–1.05) 0.056	0.07 (0.014–0.39) 0.002	0.18 (0.07–0.48) 0.001	0.27 (0.07–1.10) 0.069	0.14 (0.03–0.56) 0.006
One past anti-TNF treatment (vs. none)	0.23 (0.06–0.81) 0.022	0.32 (0.04–2.59) 0.289	0.17 (0.02–1.15) 0.070	0.17 (0.06–0.50) 0.001	0.25 (0.05–1.09) 0.066	0.14 (0.03–0.69) 0.015
Two past anti-TNF treatment (vs. none)	0.09 (0.02–0.31) <0.001	0.15 (0.02–1.10) 0.062	0.02 (0.002–0.31) 0.004	0.20 (0.05–0.83) 0.027	0.36 (0.05–2.35) 0.286	0.12 (0.01–1.34) 0.086
Clinical response at week 52
	(n=110)	(n=65)	(n=45)	n=76	(n=41)	(n=36)
Any past anti-TNF treatment	0.34 (0.10–1.18) 0.091	0.54 (0.02–10.87) 0.546	0.29 (0.07–1.17) 0.082	0.09 (0.01–0.50) 0.006	ns	0.16 (0.02–1.20) 0.076
One past anti-TNF treatment (vs. none)	0.25 (0.06–1.01) 0.052	0.21 (0.01–3.35) 0.274	0.24 (0.04–1.49) 0.128	0.06 (0.01–0.41) 0.003	ns	0.13 (0.01–1.16) 0.068
Two past anti-TNF treatment (vs. none)	0.30 (0.08–1.05) 0.060	0.34 (0.02–4.42) 0.410	0.26 (0.06–1.17) 0.080	0.12 (0.01–0.81) 0.029	ns	0.23 (0.02–2.68) 0.237
Clinical remission at week 52
Any past anti-TNF treatment	0.23 (0.06–0.79) 0.020	0.38 (0.01–7.95) 0.540	0.20 (0.04–0.85) 0.030	0.12 (0.03–0.43) 0.001	0.07 (0.008–0.70) 0.024	0.10 (0.01–0.74) 0.024
One past anti-TNF treatment (vs. none)	0.32 (0.08–1.28) 0.110	0.50 (0.02–11.66) 0.667	0.38 (0.06–2.34) 0.301	0.85 (0.02–0.33) <0.001	0.04 (0.004–0.55) 0.015	0.05 (0.004–0.61) 0.019
Two past anti-TNF treatment (vs. none)	0.19 (0.05–0.68) 0.011	0.36 (0.017–7.49) 0.511	0.13 (0.02–0.69) 0.016	0.22 (0.05–0.97) 0.046	0.12 (0.01–1.55) 0.106	0.22 (0.02–2.22) 0.199

Odds ratios (ORs) are adjusted for gender, smoking, disease duration (years) and disease phenotype (ulcerative colitis (UC) patients: disease extent, Crohn’s disease (CD) patients: disease behaviour).

CI: confidence interval; TNF: tumour necrosis factor.

*Adjusted for age group.

UC patients

Clinical response and remission rates were similar among the young and elderly UC patients at week 14 (75% vs. 77%, P = 0.8 and 31% vs. 24%, respectively, P = 0.85) and at week 52 (68% vs. 70%, P = 1 and 51% vs. 48%, P = 0.84) of vedolizumab therapy (Figure 1). Corticosteroid-free remission (Figure 2) and endoscopic response and remission rates (Supplementary Figure 1) were also comparable across age groups. The patients in both groups who were naive to biological therapy showed significantly higher response rates compared to patients who were exposed to biological therapy at weeks 14 and 52 (Supplementary Figure 2). In a multivariate analysis adjusting for age group, sex, disease duration, disease extent and smoking, previous use of an anti-TNF was independently negatively associated with clinical response at weeks 14 and 52 (OR 0.25, 95% CI 0.07–0.84; P = 0.025, OR 0.09, 95% CI 0.01–0.5; P = 0.006, respectively) (Table 3). Prior anti-TNF therapy was independently associated with poor clinical remission rates at both weeks 14 and 52 (OR 0.18, 95% CI 0.07–0.48; P = 0.001 and OR 0.12, 95% CI 0.03–0.43; P = 0.001, respectively). In the next multivariate analysis, performed in order to determine treatment outcomes in each of the age groups (young ≤40 years and elderly ≥60 years), following adjustment for gender, smoking, disease duration and disease behaviour, it was found that among elderly UC patients, prior anti-TNF exposure was independently associated with poor clinical remission rates at both weeks 14 and 52 (OR 0.14, 95% CI 0.03–0.56; P = 0.006 and OR 0.10, 95% CI 0.01–0.74; P = 0.024, respectively) (Table 3). This was not seen in the young UC cohort.

Adverse events among the entire patient population (Table 4 and Supplementary Table 1)

There was a total of 49 adverse events, of which 23 (16%) were in the young group and 26 (18%) were in the elderly group (P = 0.5) (Table 4). There were significantly more infections involving the nasopharynx, urinary tract, skin and vulva, as well as Clostridioides difficile infections in the elderly patients (17 (12%) vs. three (2%); P = 0.002). Treatment was halted due to urinary sepsis in only one elderly patient. One elderly patient died from small bowel obstruction, which was not thought to be related to vedolizumab. There were no differences in the rates of dermatological, rheumatological, or allergic adverse events across the age groups. There were significantly more non-specific adverse events, such as headache and myalgia, in the young cohort (11 (8%) vs. three (2%); P = 0.03). Vedolizumab was discontinued due to adverse events in six of the young patients (4%) events compared to four elderly patients (3%) (P = 0.5). The reasons for drug discontinuation were headache, worsening arthralgia, liver enzyme abnormalities and acute dyspnoea following administration, and they were similar for both groups. There were no differences in adverse event rates when the elderly population was stratified according to age in decades.

Table 4.

Adverse events in the entire patient population.

Adverse event	Young (n = 140)	Elderly (n = 144)	P value
Any event (%)	23 (16)	26 (18)	0.52
Required discontinuation (%)	6 (4)	4 (3)	0.54
Dermatological (%)	3 (2)	2 (1)	0.68
Allergic reactions (%)	3 (2)	2 (1)	0.68
Musculocutaneous (%)	3 (2)	2 (1)	0.68
Infections (%)	3 (2)	17 (12)	0.002
Nasopharyngeal (%)	3 (2)	9 (6)	0.13
Gastrointestinal (%)	0 (0)	1 (0.6)	1
Clostridium difficile (%)	0 (0)	2 (1)	0.5
Urinary tract (%)	0 (0)	2 (1)	0.5
Cellulitis (%)	0 (0)	2 (1)	0.5
Vulvar abscess (%)	0 (0)	1 (0.6)	1
Other (%)	11 (8)	3 (2)	0.03

In order to ascertain whether these differences were due to the exclusion of patients between the ages of 40 and 60 years, we obtained data from an additional cohort of 37 patients in this age group. Elderly patients (≥60 years) still had significantly increased rates of infections (4% vs. 12%, P = 0.008) with a non-significant trend towards more nasopharyngeal infections (2% vs. 6%, P = 0.07) when compared to young ( < 60 years) patients. There were no other significant differences, in terms of adverse events, between the groups.

Discussion

The objective of this study was to compare selected parameters of the safety and efficacy of vedolizumab between elderly (≥60 years) and young (≤40 years) patients. To the best of our knowledge, it has the largest reported number of elderly IBD patients who were being treated with vedolizumab and the first to include a control group of young patients under 40 years of age.^5–8 The elderly patients had a significantly higher incidence of cardiovascular, metabolic and respiratory comorbidities, although both age groups had similar IBD duration, disease phenotype and activity characteristics. The young patients were more biologically experienced than the elderly patients, which may indicate the prescribing physician’s inclination to be more conservative in treating IBD in older patients due to the overall higher risk of both infections and malignancies in elderly patients who have an increased comorbidity burden.

Patients in both groups had similar clinical responses and remission rates at both weeks 14 and 52 regardless of previous biological experience. There was no difference in endoscopic response or remission rates at week 52. These findings are in keeping with those of previous studies which demonstrated similar efficacy for young and elderly patients. Clinical response and remission rates for both groups were similar to those of previous real-world studies.¹⁰,¹¹

The multivariate analysis in this study demonstrates that previous anti-TNF exposure in patients treated with vedolizumab is an independent predictor of poor treatment response and remission rates, particularly in the elderly population. Our findings indicate that the early, first-line introduction of vedolizumab, before the implementation of anti-TNF therapy, may have more favourable clinical outcomes although this statement is limited by the low number of anti-TNF-naive patients in the young cohort.

Our study is the first to report that elderly IBD patients receiving vedolizumab have significantly increased rates of infections compared to young IBD patients receiving vedolizumab. These were predominantly upper respiratory infections and did not result in therapy discontinuation. One elderly patient, however, did have to discontinue vedolizumab consumption due to urinary sepsis. It remains unclear whether this increased risk is related to the greater comorbidity burden of elderly individuals in general or specifically to advanced age. This aspect should be further investigated, as it may have implications for the younger patients who already have a significant comorbidity burden. Our young study patients had significantly increased rates of non-specific adverse events, such as headache and myalgia. Overall, however, vedolizumab was well tolerated by patients in both age groups, with only 10 patients (3.5%) requiring discontinuation of therapy due to adverse events. The small bowel obstruction complications that were the cause of death in one elderly patient were not judged to be related to vedolizumab use.

The major limitation of this study is the retrospective nature of the data collection, with the inherent problems of accuracy and potential biases including recall bias and reporting bias of side effects which patients may not have reported or may not have been recorded. In particular, the reporting of side effects such as headache or myalgia, that can represent ’flu-like symptoms, may have been miss-classified and may have been underreported particularly among the younger patients. In addition, this study was initiated when vedolizumab was first introduced into the market, possibly leading to the inclusion of more treatment-refractory patients particularly among young patients who were significantly more exposed to previous biological therapy. This is in keeping with a previous study which reported lower use of immunomodulation and biological therapy in patients over than 60 years of age not necessarily related to their having a milder disease course. On the other hand, it may indicate that the young population had more severe disease than the included elderly population. Further limitations include the incomplete endoscopic data and the inconsistent use of accepted scoring methods and that we could not correct for various biochemical indicators of severity, such as albumin and inflammatory markers, because these data were not uniformly collected. The strengths of this study include the large size of the patient population as well as the collection of data from two countries, both of which may increase the generalisability of the results.

In conclusion, this study supports previous studies by demonstrating that vedolizumab has equal overall safety and efficacy in both young and elderly IBD patients. In addition, the use of vedolizumab in anti-TNF-naive patients was associated with improved clinical response and remission rates. Prior use of anti-TNF agents was associated with decreased clinical response and remission rates in all the patients who were treated with vedolizumab, but especially the elderly patients. The heretofore unreported higher infection rate in the elderly population is of special concern, and it may call for close follow-up, close adherence to vaccination protocols and heightened awareness of its occurrence in elderly IBD patients.

Summary

This large binational, multicentre retrospective cohort study shows that the use of vedolizumab in elderly patients compared to young patients has similar clinical and endoscopic efficacy but increased rates of infections.

ORCID iDs

Nathaniel Aviv Cohen https://orcid.org/0000-0001-9252-9208

Nikolas Plevris https://orcid.org/0000-0002-3229-8759

Uri Kopylov https://orcid.org/0000-0002-7156-0588