Relapse rates and predictors for relapse in a real-life cohort of IBD patients after discontinuation of anti-TNF therapy

Figures & data

Table 1. Patient characteristics in CD and UC patients.

Characteristic
CD patients	n = 77
Male	27 (35%)
Age; median years (IQR)	43 (34–53)
Montreal classification
A1 (<16 years)	11 (14%)
A2 (17–40 years)	64 (83%)
A3 (>40 years)	2 (3%)
L1 (ileum)	24 (31%)
L2 (colon)	20 (26%)
L3 (ileocolonic)	33 (43%)
+L4 (upper GI)	9 (12%)
B1 (non stricturing, non-penetrating)	48 (62%)
B2 (stricturing)	20 (26%)
B3 (penetrating)	9 (12%)
P (Perianal disease)	29 (38%)
Median months anti-TNF use (IQR)	60 (29–91)
Endoscopy available <1year	44 (57%)
Fecal calprotectin available <1 year	49 (64%)
Radiology (MRI, US) available <1 year	3 (4%)
Discontinued anti-TNF type and schedules
Infliximab	37 (48%)
Normal	33 (89%)
Intensified	4 (11%)
Adalimumab	40 (52%)
Normal	31 (78%)
Intensified	9 (12%)
Previous surgical resection
1 resection	19 (25%)
2 resections	11 (14%)
>2 resections	3 (4%)
Previous medication use
Infliximab	28 (36%)
Adalimumab	9 (12%)
Thiopurines	72 (94%)
Methotrexate	27 (35%)
Corticosteroids	63 (82%)
Budesonide	36 (47%)
Disease duration median years (IQR)	19 (13–27)
UC patients	n = 24
Male	8 (33%)
Age; median years (IQR)	48 (33–57)
Montreal classification
E1 (proctitis)	1 (4%)
E2 (left-sided)	10 (42%)
E3 (pancolitis)	13 (54%)
S1 (mild)	0 (0%)
S2 (moderate)	6 (25%)
S3 (severe)	18 (75%)
Median months anti-TNF use (IQR)	29 (16–73)
Endoscopy available <1year	16 (67%)
Fecal calprotectin available <1 year	17 (71%)
Discontinued anti-TNF type and schedules
Infliximab	20 (83%)
Normal	16 (80%)
Intensified	4 (20%)
Adalimumab	4 (17%)
Normal	4 (100%)
Intensified	0 (0%)
Previous surgical resection
1 resection	1 (4%)
Previous medication use
Infliximab	4 (17%)
Adalimumab	1 (4%)
Thiopurines	21 (88%)
Methotrexate	4 (17%)
Corticosteroids	23 (96%)
Budesonide	3 (13%)
Disease duration median years (IQR)	14 (7–31)

Table 2. Continued treatment at time of anti-TNF discontinuation in CD and UC patients.

CD patients	n = 77
None	49 (64%)
Immunomodulators	26 (34%)
Thiopurines	22 (29%)
Methotrexate	4 (5%)
5-ASA	2 (3%)
UC patients	n = 24
None	2 (8%)
Immunomodulators	11 (46%)
Thiopurines	11 (46%)
Methotrexate	0 (0%)
5-ASA	19 (79%)

Table 3. Reasons for discontinuing anti-TNF treatment.

Reasons for discontinuation	n (%)
Side effects	24 (24%)
Infections	7 (7%)
Low trough levels	2 (2%)
Anti-drug antibodies	6 (6%)
Patient preference	14 (14%)
Malignancy	2 (2%)
Pregnancy	2 (2%)

All patients (n = 101) were in remission at time of anti-TNF discontinuation. This table shows additional reasons that contributed to the decision to stop anti-TNF treatment.

Figure 1. Flowchart of included patients.

Figure 2. Time to relapse in all patients (a), CD patients (b) and UC patients (c).

Figures 3. Time to relapse requiring treatment with biologicals, surgery or experimental medication in all patients (a), CD patients (b) and UC patients (c).

Table 4. Treatment started at relapse in CD and UC patients.

	CD patients	UC patients
Patients with relapse	n = 38	n = 18
Infliximab	12 (32%)	8 (47%)
Adalimumab	14 (18%)	1 (6%)
Golimumab	–	2 (8%)
Vedolizumab	3 (4%)	–
Experimental medication	4 (5%)	–
Thiopurines	3 (8%)	2 (11%)
Methotrexate	3 (8%)	2 (11%)
Corticosteroids (oral or topical)	4 (11%)	4 (17%)
Budesonide (oral or topical)	5 (13%)	1 (4%)
5-ASA (oral or topical)	–	8 (33%)
Cyclosporine	–	1 (4%)

All treatments started at relapse in CD and UC patients. Patients could start a combination of these treatments.

Table 5. Univariable Cox proportional hazard analysis of potential predictors for relapse.

All patients (n = 101)			CD patients (n = 77)		UC patients (n = 24)
Factor	HR (95%CI)	p-value	HR (95%CI)	p-value	HR (95%CI)	p-value
Predictors for relapse in patients requiring any kind of IBD treatment (n = 56)
Diagnosis	1.51 (0.86–2.66)	.15	–	–	–	–
Gender	0.90 (0.52–1.58)	.722	0.98 (0.50–1.91)	.94	0.68 (0.25–1.89)	.464
Previous anti-TNF use	1.22 (0.70–2.10)	.484	1.40 (0.74–2.65)	.305	1.26 (0.35–4.48)	.72
Intensified anti-TNF	1.37 (0.72–2.60)	.340	1.96 (0.92–4.16)	.081	0.43 (0.12–1.59)	.206
IM continuation at stop	0.69 (0.39–1.21)	.195	0.79 (0.40–1.57)	.504	0.37 (0.13–1.09)	.071
Calprotectin <50*	0.68 (0.38–1.24)	.208	0.53 (0.25–1.14)	.114	1.05 (0.37–3.00)	.93
Calprotectin <25*	0.49 (0.24–1.03)	.059	0.34 (0.12–0.96)	.041	0.79 (0.25–2.54)	.697
Trough level >2*	2.27 (1.13–4.53)	.021	2.89 (1.20–6.98)	.018	1.14 (0.35–3.65)	.829
Disease duration (years)	1.003 (0.98–1.03)	.844	1.008 (0.97–1.04)	.665	1.002 (0.97–1.04)	.917
Anti-TNF duration (months)	0.997 (0.989–1.005)	.411	0.998 (0.988–1.008)	.708	1.00 (0.99–1.02)	.968
Surgical resection CD	–	–	1.35 (0.72–2.56)	.353	–	–
Age diagnosis <17 CD	–	–	2.29 (1.04–5.07)	.040	–	–
Colonic involvement CD	–	–	0.82 (0.42–1.62)	.566	–	–
Complicated disease CD	–	–	0.75 (0.39–1.42)	.370	–	–
Perianal disease CD	–	–	0.91 (0.47–1.76)	.777	–	–
Predictors for relapse in patients requiring treatment with biologicals, surgery or experimental medication (n = 45)
Diagnosis	1.06 (0.54–2.04)	.875	–	–	–	–
Gender	0.94 (0.51–1.73)		0.98 (0.50–1.91)	.943	0.68 (0.25–1.89)	.464
Previous anti-TNF use	1.45 (0.80–2.64)	.220	1.38 (0.70–2.74)	.356	2.30 (0.62–8.57)	.216
Intensified anti-TNF	1.37 (0.67–2.84)	.405	2.00 (0.90–4.45)	.089	0.32 (0.04–2.50)	.278
IM continuation at stop	0.62 (0.33–1.16)	.132	0.83 (0.40–1.71)	.60	0.26 (0.07–0.95)	.042
Calprotectin <50	0.66 (0.34–1.27)	.207	0.68 (0.31–1.47)	.323	0.61 (0.17–2.17)	.609
Calprotectin <25	0.45 (0.18–1.07)	.069	0.43 (0.15–1.22)	.110	0.54 (0.12–2.57)	.411
Trough level >2	3.31 (1.37–8.00)	.008	4.18 (1.42–12.30)	.009	1.59 (0.34–7.53)	.561
Disease duration (years)	1.01 (0.99–1.04)	.363	1.01 (0.98–1.05)	.585	1.02 (0.98–1.06)	.426
Anti-TNF durations (months)	1.00 (0.99–1.01)	.58	1.00 (0.99–1.01)	.716	1.01 (0.99–1.02)	.551
Surgical resection CD	–	–	0.99 (0.49–1.97)	.968	–	–
Age diagnosis <17 CD	–	–	2.68 (1.19–6.01)	.017	–	–
Colonic involvement CD	–	–	1.08 (0.52–2.28)	.832	–	–
Complicated CD	–	–	0.60 (0.30–1.18)	.139	–	–
Perianal disease CD	–	–	0.88 (0.43–1.78)	.716	–	–

Hazard ratios (HR), 95% confidence intervals (CI) and p values for all variables tested with univariable Cox proportional hazard analysis. Significant associations or trends toward a significant association are shown in bold.