Liver transplantation in the critically ill: a multicenter Canadian retrospective cohort study

Table 4 Multivariable analysis: predictors of receipt of liver transplant in 221 critically ill cirrhosis patients on transplant list in ICU

Covariate	Unadjusted (n= 221)	Model 1 (n= 126)	Model 2 (n= 174)	Model 3 (n= 125)
Age	0.99 (0.96-1.01)	0.95 (0.91-0.99) ^a	0.98 (0.94-1.01)	0.95 (0.91-0.99) ^a
Gender (female)	1.06 (0.59-1.89)	1.51 (0.64-3.57)	1.02 (0.51-2.05)	1.58 (0.66-3.80)
Etiology (HCV versus non-HCV)	1.28 (0.72-2.28)	1.78 (0.76-4.14)	1.63 (0.81-3.28)	1.72 (0.73-4.03)
SOFA (admission)^b	0.94 (0.88-0.99) ^a
Lactate (admission, natural log)^c	0.52 (0.34-0.79) ^a
SOFA (48 hours)	0.88 (0.82-0.94) ^a		0.89 (0.82-0.97) ^a	0.92 (0.82-1.04)
48-hour lactate (natural log)	0.30 (0.17-0.53) ^a	0.29 (0.15-0.53) ^a		0.32 (0.17-0.61) ^a
χ² (degrees of freedom)		31.03 (6)	23.75 (6)	33.23 (7)

All multivariable models were adjusted for site (two sites) and decade of transplant (2000 through 2009 versus 1990 through1990). Lactate (on admission and at 48 hours) was converted to a natural logarithm (normal distribution). Hosmer-Lemeshow goodness-of-fit performed well (P > 0.3) for all models. All models included age, gender (female), etiology (HCV versus non-HCV). Model 1 (n = 126, 95 missing) included lactate (natural log) at 48 hours after admission. Model 2 (n = 174, 37 missing) included SOFA score at 48 hours. Model 3 (n = 125, 96 missing) included SOFA score and lactate (natural log) at 48 hours. Significant results (P values) included Age: Model 1 (P = 0.036), Model 3 (P = 0.043); SOFA (at 48 hours): Unadjusted (P < 0.001), Model 2 (P = 0.006); Lactate 48 hours after admission (natural logarithm): Unadjusted (P < 0.001), Model 1 (P < 0.001), Model 3 (P = 0.001). See Additional file 2 for full derivation of all models and P values.

ISSN: 1364-8535