Table 4.
Predictors of the time to bacterial infection or death n (%)
| Bacterial infection or death within 1 yr | Univariate hazard ratio (95%CI) | P value | Adjusted hazard ratio3 (95%CI) | P value | |
| MELD score1 | |||||
| ≤ 13 | 5/20 (25) | 1.00 (reference) | 0.003 | 1.00 (reference) | 0.02 |
| > 13 | 18/26 (69) | 4.45 (1.64-12.07) | 3.23 (1.18-8.83) | ||
| Child-Pugh | |||||
| Score1 | |||||
| ≤ 7 | 1/13 (8) | 1.00 (reference) | 0.010 | Not included | |
| > 7 | 22/33 (67) | 13.63 (1.83-101.73) | (not significant) | ||
| Ascites1 | |||||
| No | 2/15 (13) | 1.00 (reference) | 0.003 | 1.00 (reference) | 0.01 |
| Yes | 21/31 (68) | 9.01 (2.10-38.70) | 6.95 (1.59-30.28) | ||
| IL-61 | |||||
| ≤ 35 pg/mL | 4/17 (24) | 1.00 (reference) | 0.010 | Not included | |
| > 35 pg/mL | 19/29 (66) | 3.87 (1.31-11.47) | (not significant) | ||
| I-FABP12 | |||||
| ≤ 1300 pg/mL | 1/13 (8) | 1.00 (reference) | 0.010 | Not included | |
| > 1300 pg/mL | 16/23 (70) | 13.85 (1.83-105.04) | (missing data2) |
1
Parameters were dichotomized according to the optimum cut-off from ROC analysis using maximum Youden index;
2
Serum IFABP levels were available from 36 patients;
3
Dichotomized parameters from univariate analysis were included in a forward-stepwise Cox regression model (P for inclusion < 0.05).