Supplementary Table 4.

Univariate and Multivariate Analysis for ALP Drop by 20% or More From Baseline to Last Follow-up Among Patients Not on Ursodeoxycholic Acid (n = 41)

Variable	Univariate		Multivariate
	Odds ratio (95% CI)	P value	Odds ratio	P value
Cirrhosis	7.80 (1.60–38.11)	.011	5.72 (0.68–47.79)	.107
Baseline ALP >ULNa	2.83 (0.62–13.04)	.181	2.12 (0.27–16.34)	.472
Ulcerative colitisb	0.24 (0.05–1.10)	.067	0.41 (0.06–2.92)	.370
Male gender	0.42 (0.10–1.81)	.245	0.42 (0.06–3.03)	.392
Age at diagnosis of PSCc	1.06 (1.00–1.12)	.053	1.03 (0.95–1.11)	.468
Small-duct PSCd	1.75 (0.14–21.88)	.664	—
PSC-AIH overlapd	3.50 (0.20–62.42)	.394	—
Duration vedolizumabe	1.06 (0.98–1.14)	.157	—
IBD improvedf	0.75 (0.14–4.13)	.741	—
Previous anti-TNF use	0.83 (0.19–3.75)	.812	—

AIH, autoimmune hepatitis; ALP, alkaline phosphatase; CI, confidence interval; IBD, inflammatory bowel disease; PSC, primary sclerosing cholangitis; TNF, tumor necrosis factor; ULN, upper limit of normal.

^aBaseline indicating last ALP taken before vedolizumab commenced.

^bVersus IBD-unspecified or Crohn’s disease.

^cPer 1-year increase.

^dVersus large-duct PSC.

^ePer 1-month increase in vedolizumab duration.

^fEndoscopic improvement versus unchanged/worsened.