Colitis outcomes

	Overall	Microscopic colitis	Non-microscopic colitis	p-value
Interventions
Budesonide	15/38 (39.5%)	12/13 (92.3%)	3/25 (12.0%)	< 0.001*
Any systemic glucocorticoids	25/38 (65.8%)	3/13 (23.1%)	22/25 (88.0%)	< 0.001*
Systemic glucocorticoids < 1 mg/kg/d	14/38 (36.8%)	2/13 (15.4%)	12/25 (48.0%)	0.077
Systemic glucocorticoids ≥1 mg/kg/d	12/38 (31.2%)	1/13 (7.7%)	11/25 (44.0%)	0.030*
Time from symptom onset to initiation of glucocorticoids (days)
Mean +/− SD	32.4 +/− 37.1	33.2 +/− 26.0	32.0 +/− 42.7	0.928
Median	21.0	28.0	18.0	0.928
Proportion continuing CPI course	14/38 (36.8%)	10/13 (76.9%)	4/25 (16.0%)	< 0.001*
Proportion eventually discontinuing immunotherapy	29/38 (76.3%)	8/13 (61.5%)	21/25 (84.0%)	0.226
Due to toxicity alone	22/38 (57.9%)	6/13 (46.2%)	16/25 (64.0%)	1.000
Due to any progressive disease	7/38 (18.4%)	2/13 (15.4%)	5/25 (25.0%)	1.000
Average number of additional infusions#	3.0 +/− 5.7	5.8 +/− 6.8	1.6 +/− 4.5	0.030*
Additional irAEs after colitis	8/38 (21.1%)	3/13 (23.1%)	5/25 (20.0%)	1.000
Response to first line treatment	22/38 (67.9%)	9/13 (69.2%)	14/25 (56.0%)	0.429
Proportion of patients requiring second-line immunosuppression	23/38 (60.5%)	5/13 (38.5%)	18/25 (72.0%)	Not calculated
Proportion requiring TNFα inhibitor	16/38 (42.1%)	4/13 (30.8%)	11/25 (44.0%)	0.429

The p-value was calculated by ANOVA for numerical covariates and chi-square test or Fisher’s exact for categorical covariates, where appropriate. TNFα: tumor necrosis factor α. CPI: immune checkpoint inhibitor. SD: standard deviation. irAE: immune-related adverse event. ^#calculated among patients who continued CPIs. *denotes statistical significance at p = 0.05. Second-line immune suppression includes exposure to different glucocorticoids in the microscopic colitis cohort

All of the boldfaced numbers should be statistically signficant