Urinary BA Indices as Prognostic Biomarkers for Complications Associated with Liver Diseases

Table 5

Multivariate logistic regression analyses for ascites in the entire liver-patient population.

(a) BALDC model

BA parameters	value	Standard error	value	Odds ratio (OR): Exp ()
				1-unit	10%	20%
Intercept	-3.463	—	0.001	0.031	—	—
% MDCA	-2.452	1.112%	0.027	0.086	0.909	0.826
% Primary BA	0.045	0.008%	0.001	1.046	1.234	1.524

Using the regression coefficients () from this table, the estimated (OR) of developing ascites by the BALDC model is .

(b) Non-BA model

Non-BA parameters	value	Standard error	value	Odds ratio (OR): Exp ()
				1-unit	10%	20%
Intercept	0.947	—	0.560	2.577	—	—
MELD	0.189	0.050	0.001	1.208	1.185	1.404
Albumin level	-1.205	0.387	0.002	0.300	0.640	0.410

Using the regression coefficients () from this table, the estimated (OR) of developing ascites by the non-BA model is .

(c) Mixed BA and Non-BA model

Mixed BA and non-BA parameters	value	Standard error	value	Odds ratio (OR): Exp ()
				1-unit	10%	20%
Intercept	-0.275	1.768	0.894	0.79	—	—
% CDCA	0.029	0.012%	0.014	1.029	1.104	1.218
Primary BA/secondary BA	-0.077	0.032	0.015	0.926	0.983	0.967
Albumin level	-1.143	0.407	0.004	0.319	0.655	0.429
MELD	0.189	0.053	0.001	1.208	1.185	1.404

Using the regression coefficients () from this table, the estimated (OR) of developing ascites by the mixed BA and non-BA model is .

(d) Original MELD model

MELD parameters	value	Standard error	value	Odds ratio (OR): Exp ()
				1-unit	10%	20%
Intercept	-4.049	0.554	0.001	1.317	—	—
MELD	0.276	0.045	0.001	0.017	0.026	0.001

Using the regression coefficients () from this table, the estimated (OR) of developing ascites by the original MELD model is .