Table 3.
Univariable and multivariable Cox regression analyses for the overall risk of graft loss after LT in the training set
| Variables | Training set (N = 1262) | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| Univariable analysis | Multivariable analysis | ||||||||
| Beta-coefficient ± SE | OR | 95% CI | p value | Beta-coefficient ± SE | Wald | HR | 95% CI | p value | |
| Waiting time (per day) | 0.00 ± 0.00 | 1.00 | 1.00–1.00 | 0.7 | – | – | – | – | – |
| Patient age at LT (per year) | − 0.00 ± 0.01 | 1.00 | 0.99–1.01 | 0.8 | – | – | – | – | – |
| Patient male gender (yes vs. no) | 0.11 ± 0.12 | 1.11 | 0.89–1.40 | 0.4 | – | – | – | – | – |
| Patient BMI at LT (per unit)a | − 0.02 ± 0.01 | 0.98 | 0.96–1.00 | 0.08 | – | – | – | – | – |
| MELD at LT (per point)b | 0.02 ± 0.01 | 1.02 | 1.01–1.03 | 0.002 | – | – | – | – | – |
| Donor age (per year)a | 0.01 ± 0.00 | 1.01 | 1.00–1.02 | 0.001 | 0.01 ± 0.00 | 9.38 | 1.01 | 1.00–1.01 | 0.002 |
| Donor male gender (yes vs. no) | − 0.06 ± 0.10 | 0.94 | 0.77–1.14 | 0.5 | – | – | – | – | – |
| Donor BMI at LT (per unit) | 0.01 ± 0.01 | 1.01 | 0.98–1.03 | 0.5 | – | – | – | – | – |
| CVA as cause of donor deatha | 0.26 ± 0.10 | 1.29 | 1.05–1.58 | 0.01 | – | – | – | – | – |
| Donor ICU stay (per day) | 0.00 ± 0.01 | 1.00 | 0.98–1.03 | 0.9 | – | – | – | – | – |
| Donor DM2a | 0.26 ± 0.18 | 1.30 | 0.91–1.85 | 0.2 | – | – | – | – | – |
| Donor local procurement | − 0.09 ± 0.10 | 0.91 | 0.75–1.11 | 0.4 | – | – | – | – | – |
| CIT (per min) | 0.00 ± 0.00 | 1.00 | 1.00–1.00 | 0.5 | – | – | – | – | – |
| WIT (per min)a | 0.01 ± 0.00 | 1.01 | 1.00–1.01 | 0.003 | – | – | – | – | – |
| BAR score (per point)a | 0.04 ± 0.01 | 1.04 | 1.01–1.06 | 0.002 | 0.03 ± 0.01 | 6.58 | 1.03 | 1.01–1.05 | 0.01 |
| D-MELD (per 100 points)a,b | 0.03 ± 0.01 | 1.03 | 1.02–1.05 | < 0.001 | – | – | – | – | – |
| EADa | 0.44 ± 0.10 | 1.55 | 1.28–1.89 | < 0.001 | – | – | – | – | – |
| CCIa | 0.04 ± 0.00 | 1.04 | 1.04–1.05 | < 0.001 | 0.04 ± 0.00 | 552.95 | 1.05 | 1.04–1.05 | < 0.001 |
The Backward Wald method was used for selecting the covariates in the multivariable analyses. − 2log likelihood = 4,955.53
SE standard error, HR hazard ratio, CI confidence intervals, LT liver transplantation, BMI body mass index, MELD model for end-stage liver disease, CVA cerebrovascular accident, ICU intensive care unit, DM diabetes mellitus, CIT cold ischemia time, WIT warm ischemia time, BAR balance of risk, D-MELD donor-MELD, EAD early allograft dysfunction, CCI comprehensive complication index
aVariables initially introduced in the multivariable model
bThe multivariable model was constructed introducing only the variable “D-MELD”, with the intent to avoid collinearity phenomena with the variable “MELD”. A similar model with the variable “MELD” instead of “D-MELD” was contextually constructed. In both cases, D-MELD or MELD were deleted during the backward Wald method