Background: Acute liver failure is a potentially fatal disease of various etiologies for which liver transplantation is the only known curative treatment. Although the decision-making on transplantation is largely dependent on the severity of liver injury (based on predicting a fatal outcome), a statistical analysis to predict "survival" has not been extensively conducted. In this study, we investigate the medical history of patients in two distinct areas of Japan with the aim of identifying the predictors of survival in patients with acute liver injury (ALI).
Methods: Datasets of 301 patients with ALI in two distinct areas (93 in southern Kyushu and 208 in northern Tohoku) of Japan, who were treated from 2004 to 2014, were included in the analysis.
Results: Among the enrolled 301 cases, 263 patients survived without transplantation. A PT-INR of ≥ 1.3 during the clinical course was found to be adequate for predicting a poor prognosis, because all of the fatal cases emerged from this population (hazard ratios: southern Kyushu, 0.2827; northern Tohoku, 0.1862). All surviving patients showed a reduction in their PT-INR during treatment, whereas the PT-INR did not decrease in the patients with a poor prognosis. A PT-INR of < 1.3 on days 7 and 8 efficiently predicted transplant-free survival (log-rank test: southern Kyushu, P = 0.0030; northern Tohoku, P = 0.0022).
Conclusions: A PT-INR of ≥ 1.3 during the clinical course might identify cases with a poor prognosis, while the recovery of the PT-INR to < 1.3 predicts transplant-free survival.
Keywords: ALF; ALI; HGF; PT-INR; Survival.