. 2022 Aug 23;42(11):2453–2465. doi: 10.1111/liv.15386

TABLE 4.

Factors associated with hepatic severe adverse events during treatment with obeticholic acid

	Univariate			Multivariate
Variable	RR	95%CI	p	aRR	95%CI	p
Age at OCA start, years	1.01	0.93–1.09	.855
Age at PBC diagnosis, years	1.01	0.93–1.09	.854
Duration of PBC, years	1.00	0.93–1.07	.947
Female sex	0.42	0.06–2.74	.366
Diabetes mellitus	0.92	0.13–6.70	.932
BMI, kg/m²	0.92	0.79–1.08	.311
ANA positivity	1.85	0.49–6.97	.366
AMA positivity	0.72	0.16–3.16	.66
PBC‐AIH overlap	0.77	0.10–5.68	.796
Oesophageal varices	2.48	0.79–9.87	.123
Concomitant fibrate therapy	1.12	0.16–8.10	.907
History of ascites	4.54	2.67–7.72	<.001	3.50	1.85–6.50	<.001
Platelets <150 000 /mm³	0.59	0.16–2.23	.437
INR ^a ^, ^b	2.11	1.25–3.56	.005	1.91	1.10–3.36	.024
Albumin, g/dl ^b	0.13	0.06–0.28	<.001	0.18	0.06–0.51	.001
Creatinine, mg/dl	1.52	0.37–6.18	.561
Child‐Pugh score ^c	2.32	1.75–3.06	<.001	2.43	1.50–4.04	<.001
MELD ^c	1.32	1.15–1.50	<.001	1.23	1.09–1.39	<.001
OCA dose ^d	0.85	0.27–2.67	.775
ALP/ULN at baseline	1.26	0.79–1.99	.331
ALT/ULN at baseline ^b	1.66	0.98–2.83	.06	1.00	0.54–1.99	.918
AST/ULN at baseline ^b	1.82	1.40–2.35	<.001	0.91	0.58–1.43	.680
GGT/ULN at baseline	0.98	0.93–1.04	.574
Total bilirubin at baseline ^b	1.53	1.34–1.74	<.001	1.30	1.05–1.56	.014

Note: Risk ratios with 95% confidence intervals were from Poisson regression models with robust error variance. All variables associated at univariate analysis with a p < .10 entered the multivariate model.

Abbreviations: AIH, autoimmune hepatitis; ALP, alkaline phosphatase; ALT, alanine transferase; AMA, antimitochondrial antibodies; ANA, antinuclear antibodies; aRR, adjusted risk ratio; AST, aspartate transferase; BMI, body mass index; GGT, gamma‐glutamyl transferase; INR, international normalized ratio; MELD, model for end‐stage liver disease; OCA, obeticholic acid; PBC, primary biliary cholangitis; RR, risk ratio; UDCA, ursodeoxycholic acid; ULN, upper limit of normal.

^{^a}

Risk estimates reported for 1 standard deviation increase to provide a more clinically useful result.

^{^b}

Alternatively included in multivariable models to avoid multiple collinearity.

^{^c}

Since included in their computation, Child‐Pugh score and MELD were included in multivariate models after exclusion of history of ascites, albumin, INR and total bilirubin (for Child‐Pugh score) and INR and total bilirubin for MELD.

^{^d}

OCA dose categorized as <5 mg daily, 5 mg daily and >5 mg daily.