Efficacy of Biologic and Small Molecule Agents as Second-line Therapy After Exposure to TNF Inhibitors in Patients With Ulcerative Colitis

A Propensity-matched Cohort Study

Gursimran S. Kochhar; Aakash Desai; Francis A. Farraye; Raymond K. Cross; Sandra El-Hachem; Parambir S. Dulai; Miguel Regueiro

Disclosures

Aliment Pharmacol Ther. 2023;58(3):297-308.

Abstract and Introduction

Abstract

Background: There is limited real-world data on comparative effectiveness of different biologic or small molecule agents as second-line therapies in patients with ulcerative colitis (UC) with prior exposure to a tumour necrosis factor inhibitor (TNFi).

Methods: We conducted a retrospective cohort study using TriNetX, a multi-institutional database to assess the efficacy of tofacitinib, vedolizumab and ustekinumab in patients with ulcerative colitis (UC) with prior exposure to a TNFi. Failure of medical therapy was defined as a composite outcome of intravenous steroids or colectomy within 2 years. One-to-one propensity score matching was performed for demographics, disease extent, mean haemoglobin, C-reactive protein, albumin and calprotectin, prior IBD medications and steroid use between cohorts.

Results: Among 2141 patients with UC and prior exposure to TNFi, 348 (16.2%), 716 (33.4%) and 1077 (50.3%) were switched to tofacitinib, ustekinumab and vedolizumab, respectively. After propensity-score matching, there was no difference in the composite outcome (aOR: 0.77, 95% CI: 0.55–1.07) but higher risk of colectomy (aOR: 2.69, 95% CI: 1.31–5.50) in the tofacitinib cohort than the vedolizumab cohort. There was no difference in the risk of composite outcome (aOR: 1.29, 95% CI: 0.89–1.86) but higher risk of colectomy (aOR: 2.63, 95% CI: 1.24–5.58) in the tofacitinib cohort than the ustekinumab cohort. The vedolizumab cohort had a higher risk of composite outcome (aOR: 1.67, 95% CI: 1.29–2.16) than the ustekinumab cohort.

Conclusion: Ustekinumab might be the preferred second-line therapy over tofacitinib and vedolizumab in patients with UC that were previously exposed to a TNFi.

1 2 3 4 5

Next Section

Abstract and Introduction
Background
Methods
Results
Discussion
References

Table 1. Baseline characteristics of patients in the tofacitinib and vedolizumab cohort before and after propensity-score matching.
	Before propensity score matching			After propensity score matching
	Tofacitinib cohort (n = 348)	Vedolizumab cohort (n = 1077)	p value	Tofacitinib cohort (n = 325)	Vedolizumab cohort (n = 325)	p value
Demographics
Age at index (mean ± SD)	42 ± 16	42.5 ± 18.1	0.66	41.3 ± 15.8	41.6 ± 18.9	0.83
Female gender	173 (49.7%)	596 (55.3%)	0.06	163 (50.1%)	165 (50.7%)	0.87
Race
White	290 (83.3%)	856 (79.4%)	0.11	271 (83.3%)	269 (82.7%)	0.83
African American	16 (4.5%)	58 (5.3%)	0.56	23 (7%)	33 (10.1%)	0.16
Hispanic	24 (6.8%)	62 (5.7%)	0.43	23 (7%)	33 (10.1%)	0.16
Diabetes mellitus	31 (9%)	129 (12%)	0.13	27 (8.2%)	26 (7.9%)	0.88
Nicotine dependence	31 (9%)	102 (9.5%)	0.80	29 (8.8%)	28 (8.5%)	0.88
Obesity	53 (15.4%)	210 (19.5%)	0.09	49 (15%)	45 (13.8%)	0.65
Disease extent^a
Pan-colitis	202 (59%)	584 (54.4%)	0.13	197 (61.1%)	191 (59.3%)	0.62
Left-sided colitis	50 (14.6%)	133 (12.4%)	0.28	46 (14.2%)	38 (11.8%)	0.34
Recto-sigmoiditis	45 (13.1%)	128 (11.9%)	0.54	41 (12.7%)	49 (15.2%)	0.36
Proctitis	33 (9.6%)	136 (12.6%)	0.13	33 (10.2%)	30 (9.3%)	0.69
Labs^b
Haemoglobin (g/dL)	12.7	12.7	0.95	12.7	12.4	0.12
Albumin (g/dL)	3.95	3.93	0.68	3.94	3.85	0.08
C-reactive protein (mg/L)	15.2	14.5	0.79	14.8	20.6	0.14
Calprotectin (μg/g)	751	613	0.11	751	550	0.09
Prior medications^c
Prednisone	211 (60.6%)	601 (55.8%)	0.11	195 (60%)	202 (62.1%)	0.57
Methylprednisolone	112 (32.1%)	334 (31%)	0.68	104 (32%)	116 (35.6%)	0.31
Azathioprine	51 (14.6%)	154 (14.2%)	0.86	50 (15.3%)	46 (14.1%)	0.65
Mercaptopurine	23 (6.6%)	78 (7.2%)	0.68	23 (7%)	30 (9.2%)	0.31
Methotrexate	36 (10.3%)	76 (7%)	0.04	28 (8.6%)	29 (8.9%)	0.88
Type of TNFi^d
Infliximab	154 (45%)	572 (53.3%)	0.003	150 (46.2%)	159 (49%)	0.47
Adalimumab	175 (51.1%)	518 (48.3%)	0.35	164 (50.6%)	146 (45%)	0.15
Golimumab	22 (6.4%)	32 (2.9%)	0.003	17 (5.2%)	16 (4.9%)	0.85

Abbreviations: SD, standard deviation; TNFi, tumour necrosis factor inhibitor.
^aDisease extent within 1 year prior to the initiation of second-line therapy based on ICD 10 codes.
^bLabs within 1 year prior to the initiation of second-line therapy.
^cMedications within 1 year prior to initiation of second-line therapy.
^dPatients could have been on more than one medication.

Table 2. Baseline characteristics of patients in the tofacitinib and ustekinumab cohort before and after propensity-score matching.
	Before propensity score matching			After propensity score matching
	Tofacitinib cohort (n = 348)	Ustekinumab cohort (n = 716)	p value	Tofacitinib cohort (n = 303)	Ustekinumab cohort (n = 303)	p value
Demographics
Age at index (mean ± SD)	42 ± 16	40.4 ± 15.8	0.12	41.3 ± 15.7	41.3 ± 16.3	0.88
Female gender	173 (49.7%)	393 (54.8%)	0.11	152 (50.1%)	157 (51.8%)	0.68
Race
White	290 (83.3%)	565 (78.9%)	0.08	251 (82.8%)	258 (85.1%)	0.43
African American	16 (4.5%)	64 (8.9%)	0.01	14 (4.6%)	11 (3.6%)	0.54
Hispanic	24 (6.8%)	34 (4.7%)	0.14	22 (7.2%)	23 (7.5%)	0.87
Diabetes mellitus	31 (9%)	56 (7.9%)	0.54	26 (8.2%)	24 (7.6%)	0.76
Nicotine dependence	31 (9%)	101 (14.3%)	0.01	31 (9.8%)	27 (8.5%)	0.58
Obesity	53 (15.4%)	147 (20.9%)	0.03	51 (16.1%)	47 (14.9%)	0.66
Disease extent^a
Pan-colitis	206 (59.1%)	229 (31.9%)	<0.0001	170 (56.1%)	178 (58.7%)	0.51
Left-sided colitis	51 (14.6%)	38 (5.3%)	<0.0001	33 (10.8%)	29 (9.5%)	0.59
Recto-sigmoiditis	45 (12.9%)	52 (7.2%)	0.002	30 (9.9%)	37 (12.2%)	0.36
Proctitis	36 (10.3%)	58 (8.1%)	0.22	30 (9.9%)	32 (10.5%)	0.78
Labs^b
Haemoglobin (g/dL)	12.7	12.9	0.26	12.7	12.8	0.58
Albumin (g/dL)	3.95	3.99	0.38	3.94	3.98	0.53
C-reactive protein (mg/L)	15.2	16.8	0.58	15.5	16.9	0.70
Calprotectin (μg/g)	903	762	0.26	921	900	0.89
Prior medications^c
Prednisone	211 (60.6%)	342 (47.7%)	<0.0001	177 (58.4%)	182 (60%)	0.67
Methylprednisolone	112 (32.1%)	196 (27.3%)	0.10	93 (30.6%)	92 (30.3%)	0.92
Azathioprine	51 (14.6%)	107 (14.9%)	0.90	48 (15.8%)	41 (13.5%)	0.42
Mercaptopurine	23 (6.6%)	38 (5.3%)	0.39	17 (5.6%)	16 (5.2%)	0.85
Methotrexate	36 (10.3%)	83 (11.5%)	0.54	30 (9.9%)	31 (10.2%)	0.89
Type of TNFi^d
Infliximab	154 (45%)	275 (39.1%)	0.06	138 (44.9%)	144 (46.9%)	0.62
Adalimumab	175 (51.1%)	441 (62.7%)	0.0004	162 (52.7%)	161 (52.4%)	0.93
Golimumab	22 (6.4%)	25 (3.5%)	0.03	18 (5.8%)	13 (4.2%)	0.35

Abbreviations: SD, standard deviation; TNFi, tumour necrosis factor inhibitor.
^aDisease extent within 1 year prior to the initiation of second line therapy based on ICD 10 codes.
^bLabs within 1 year prior to the initiation of second line therapy.
^cMedications within 1 year prior to initiation of a second line therapy.
^dPatients could have been on more than one medication.

Table 3. Baseline characteristics of patients in the vedolizumab and ustekinumab cohort before and after propensity-score matching.
	Before propensity score matching			After propensity score matching
	Vedolizumab cohort (n = 1077)	Ustekinumab cohort (n = 716)	p value	Vedolizumab cohort (n = 566)	Ustekinumab cohort (n = 566)	p value
Demographics
Age at index (mean ± SD)	42.5 ± 18.1	40.4 ± 15.8	0.01	39.9 ± 17.5	40.6 ± 16.2	0.48
Female gender	596 (55.3%)	393 (54.8%)	0.85	311 (54.9%)	323 (57%)	0.47
Race
White	856 (79.4%)	565 (78.9%)	0.77	447 (78.9%)	453 (80%)	0.65
African American	58 (5.3%)	64 (8.9%)	0.003	39 (6.8%)	42 (7.4%)	0.72
Hispanic	62 (5.7%)	34 (4.7%)	0.35	34 (6%)	31 (5.4%)	0.70
Diabetes mellitus	129 (12%)	56 (7.9%)	0.006	40 (6.9%)	52 (9%)	0.19
Nicotine dependence	102 (9.5%)	101 (14.3%)	0.001	68 (11.7%)	70 (12.1%)	0.85
Obesity	210 (19.5%)	147 (20.9%)	0.49	114 (19.7%)	121 (20.9%)	0.60
Disease extent^a
Pan-colitis	675 (62.9%)	320 (45.5%)	<0.0001	297 (51.4%)	298 (51.6%)	0.95
Left-sided colitis	196 (18.2%)	89 (12.6%)	0.0001	91 (15.7%)	81 (14%)	0.40
Recto-sigmoiditis	189 (17.6%)	90 (12.8%)	0.006	89 (15.4%)	76 (13.1%)	0.27
Proctitis	236 (22%)	110 (15.6%)	0.0009	106 (18.3%)	98 (16.8%)	0.53
Labs^b
Haemoglobin (g/dL)	12.7	12.9	0.13	12.8	12.8	0.86
Albumin (g/dL)	3.93	3.99	0.08	3.97	3.97	0.90
C-reactive protein (mg/L)	14.5	16.8	0.25	13.4	16.7	0.16
Calprotectin (μg/g)	614	762	0.07	697	797	0.35
Prior medication^c
Prednisone	601 (55.8%)	342 (47.7%)	0.0008	287 (50.7%)	291 (51.4%)	0.81
Methylprednisolone	334 (31%)	196 (27.3%)	0.09	173 (30.5%)	163 (28.7%)	0.51
Azathioprine	154 (14.2%)	107 (14.9%)	0.70	88 (15.5%)	83 (14.6%)	0.67
Mercaptopurine	78 (7.2%)	38 (5.3%)	0.10	31 (5.4%)	33 (5.8%)	0.79
Methotrexate	76 (7%)	83 (11.5%)	0.0009	51 (9%)	47 (8.3%)	0.67
Type of TNFi^d
Infliximab	572 (53.3%)	275 (39.1%)	<0.0001	239 (42.1%)	249 (43.9%)	0.54
Adalimumab	518 (48.3%)	441 (62.7%)	<0.0002	337 (59.4%)	323 (56.9%)	0.39
Golimumab	32 (2.9%)	25 (3.5%)	0.50	18 (3.1%)	17 (2.9%)	0.86

Table 4. Comparison of efficacy between different biologic or small molecule agents as a second-line therapy in patients with UC after exposure to a TNFi.
Outcome	Cohorts	N (%)	aOR	95% CI
Tofacitinib vs. vedolizumab
Composite outcome	Tofacitinib	92 (28.3%)	0.77	0.55–1.07
Composite outcome	Vedolizumab	110 (33.8%)	0.77	0.55–1.07
IV steroids	Tofacitinib	80 (25%)	0.69	0.49–0.98
IV steroids	Vedolizumab	104 (32%)	0.69	0.49–0.98
Colectomy	Tofacitinib	28 (8.6%)	2.69	1.31–5.50
Colectomy	Vedolizumab	11 (3.3%)	2.69	1.31–5.50
Oral steroids	Tofacitinib	149 (45.8%)	1.32	0.96–1.80
Oral steroids	Vedolizumab	127 (39%)	1.32	0.96–1.80
Change in therapy	Tofacitinib	67 (20.6%)	1.49	0.99–2.25
Change in therapy	Vedolizumab	48 (14.7%)	1.49	0.99–2.25
Tofacitinib vs. ustekinumab
Composite outcome	Tofacitinib	86 (28.3%)	1.29	0.89–1.86
Composite outcome	Ustekinumab	71 (23.4%)	1.29	0.89–1.86
IV steroids	Tofacitinib	74 (24.4%)	1.07	0.74–1.56
IV steroids	Ustekinumab	70 (23.1%)	1.07	0.74–1.56
Colectomy	Tofacitinib	25 (8.2%)	2.63	1.24–5.58
Colectomy	Ustekinumab	10 (3.3%)	2.63	1.24–5.58
Oral steroids	Tofacitinib	136 (44.8%)	1.11	0.80–1.53
Oral steroids	Ustekinumab	128 (42.2%)	1.11	0.80–1.53
Change in therapy	Tofacitinib	61 (20.1%)	2.38	1.48–3.82
Change in therapy	Vedolizumab	29 (9.5%)	2.38	1.48–3.82
Vedolizumab vs. ustekinumab
Composite outcome	Vedolizumab	201 (35.5%)	1.67	1.29–2.16
Composite outcome	Ustekinumab	140 (24.7%)	1.67	1.29–2.16
IV steroids	Vedolizumab	193 (34%)	1.60	1.29–2.16
IV steroids	Ustekinumab	138 (24%)	1.60	1.29–2.16
Colectomy	Vedolizumab	17 (3%)	1.72	0.78–3.79
Colectomy	Ustekinumab	10 (1.7%)	1.72	0.78–3.79
Oral steroids	Vedolizumab	237 (41.8%)	1.14	0.90–1.44
Oral steroids	Ustekinumab	219 (38.6%)	1.14	0.90–1.44
Change in therapy	Vedolizumab	59 (10.4%)	1.45	0.95–2.19
Change in therapy	Ustekinumab	42 (7.4%)	1.45	0.95–2.19

Note: TriNetX obfuscates patient counts ≥ 1 and <10 to maintain patient confidentiality.
Abbreviations: aOR, adjusted odds ratio; CI, confidence interval.

Table 5. Comparison of safety outcomes between different biologic or small molecule agents as a second-line therapy in patients with UC after exposure to a TNFi.
Outcome	Cohorts	N (%)	aOR	95% CI
MACE	Tofacitinib	10 (3.1%)^a	0.73	0.31–1.70
MACE	Vedolizumab	13 (4.2%)	0.73	0.31–1.70
Malignancy	Tofacitinib	10 (3.5%)^a	0.96	0.39–2.34
Malignancy	Vedolizumab	10 (3.5%)^a	0.96	0.39–2.34
VTE	Tofacitinib	10 (3.1%)^a	0.95	0.39–2.32
VTE	Vedolizumab	10 (3.2%)^a	0.95	0.39–2.32
Opportunistic infections	Tofacitinib	19 (7.9%)	0.73	0.39–1.39
Opportunistic infections	Vedolizumab	23 (10.5%)	0.73	0.39–1.39
MACE	Tofacitinib	10 (3.2%)^a	0.98	0.40–2.39
MACE	Ustekinumab	10 (3.2%)^a	0.98	0.40–2.39
Malignancy	Tofacitinib	10 (3.5%)^a	0.77	0.32–1.81
Malignancy	Ustekinumab	12 (4.5%)	0.77	0.32–1.81
VTE	Tofacitinib	10 (3.2%)	0.92	0.37–2.25
VTE	Ustekinumab	10 (3.4%)	0.92	0.37–2.25
Opportunistic infections	Tofacitinib	19 (8.2%)	1.21	0.60–2.46
Opportunistic infections	Ustekinumab	15 (6.8%)	1.21	0.60–2.46
MACE	Vedolizumab	13 (2.2%)	1.31	0.57–3.01
MACE	Ustekinumab	10 (1.7%)	1.31	0.57–3.01
Malignancy	Vedolizumab	17 (3.3%)	1.46	0.69–3.09
Malignancy	Ustekinumab	12 (2.3%)	1.46	0.69–3.09
VTE	Vedolizumab	10 (1.7%)	0.96	0.39–2.33
VTE	Ustekinumab	10 (1.8%)	0.96	0.39–2.33
Opportunistic infections	Vedolizumab	41 (9.4%)	1.61	0.96–2.70
Opportunistic infections	Ustekinumab	25 (6%)	1.61	0.96–2.70

Abbreviations: aOR, adjusted odds ratio; CI, confidence interval; MACE, major adverse cardiovascular event; VTE, venous thromboembolism.
^aTriNetX obfuscates patient counts ≥1 and <10 to maintain patient confidentiality.

Authors and Disclosures

Gursimran S. Kochhar¹, Aakash Desai², Francis A. Farraye³, Raymond K. Cross⁴, Sandra El-Hachem¹, Parambir S. Dulai⁵ and Miguel Regueiro⁶

¹Division of Gastroenterology, Hepatology & Nutrition, Allegheny Health Network, Pittsburgh, Pennsylvania, USA
²Division of Gastroenterology & Hepatology, MetroHealth Medical Center, Case Western Reserve University, Cleveland, Ohio, USA
³Division of Gastroenterology, Hepatology & Nutrition, Mayo Clinic, Jacksonville, Florida, USA
⁴Division of Gastroenterology & Hepatology, University of Maryland School of Medicine, Baltimore, Maryland, USA
⁵Division of Gastroenterology, Hepatology & Nutrition, Northwestern University, Chicago, Illinois, USA
⁶Division of Gastroenterology, Hepatology & Nutrition, Cleveland Clinic, Cleveland, Ohio, USA

Correspondence
Gursimran S. Kochhar, Division of Gastroenterology, Hepatology & Nutrition, Allegheny Health Network, Pittsburgh, PA 15212, USA. Email: drkochhar@gmail.com
Gursimran S. Kochhar and Aakash Desai are co-first authors due to equal contribution.

Author Contributions
Gursimran Kochhar: Conceptualization (equal); methodology (equal); supervision (equal); writing – original draft (equal). Aakash Desai: Methodology (equal); writing – original draft (equal); data collection (lead); data analysis (lead). Francis A Farraye: Writing – review and editing (supporting). Raymond K. Cross: Writing – review and editing (supporting). Sandra El Hachem: Writing – review and editing (supporting). Parambir S Dulai: Writing – review and editing (supporting). Miguel Regueiro: Conceptualization (equal); supervision (equal); writing – review and editing (supporting).

Conflict of Interest Statement
GS: Speaker: Lilly Pharma. Advisor Board: GIE Medical, Lilly Pharma, CorEvitas Research foundation. Consultant: Pentax endoscopy. Stock Options: Digbi Health. AD: None. RC: Dr. Cross has served on advisory boards for Abbvie, BMS, CorEvitas, Fresenius Kabi and Janssen; has served as a consultant for Fzata, Magellan Health, Pfizer, Samsung Bioepis and Sebela; has served on the Data Safety Monitoring Board for Adiso and has a research grant with Janssen. FF: Dr. Farraye has served on advisory boards for BMS, Braintree Labs, Fresenius Kabi, GI Reviewers, Glaxo Smith Kline, Iterative Health, Janssen, Pfizer and Sebela. Dr. Farraye serves on a Data Safety Monitoring Board for Adiso Therapeutics and Lilly. SE: Dr. El-Hachem has served on the speakers' board for Abbvie, Janssen, Pfizer, Bristol Myers Squibb, Takeda, Ferring, UCB. PD: Research support and/or consulting Abivax, Addiso, Roivant, Takeda, Pfizer, Abbvie, Janssen, Lilly, Gilead, BMS and Precidiag. MR: Dr. Regueiro has served on advisory boards and as a consultant (both) for Abbvie, Janssen, UCB, Takeda, Pfizer, BMS, Organon, Amgen, Genentech, Gilead, Salix, Prometheus, Lilly, Celgene, TARGET Pharma Solutions, Trellis and Boehringer Ingelheim Pharmaceuticals, Inc. (BIPI).