MCV and serum vitB12 levels higher significantly in patients with end stage of ALD
Among the 139 patients, 60(43.1%) patients had serum vitB12 levels>894pg/ml. The serum levels of vitB12 and MCV in the patients that with higher scores of MDF(≥32), MELD(≥20) and ABIC(≥9) were significantly higher than that with lower scores of these models. In contrast to vitB12 and MCV, the serum levels of folic acid were much lower in the patients with higher scores. In addition, serum levels of GGT, and HGB were significantly lower in the patients with higher scores of MDF (P<0.05) and HGB also was significantly lower in that with higher scores of MELD (P<0.05). (Shown in Table 1).
We further analyzed the relationship of vitB12 and MCV with parameters of liver function in ALD patients. These results suggested that the serum levels of vitB12 and MCV were significantly correlated with the main indicators of liver function, and there was a positive correlation between the serum levels of vitB12 and ALT, AST, TBil, PT and INR(r=0.181,r=0.237,r=0.626,r=0.685,r=0.615, respectively. P<0.05).
Serum levels of vitB12 and MCV were not associated with the complications in retrospective ALD-ACLF cohort
We selected the patients diagnosed with ALD-ACLF (all the patients were Chinese) to analyzed whether the clinical parameters including the serum levels of vitB12 and MCV have relationship with the complications of end stage liver disease. The ages of all 79 patients (Male/Female: 77/2) range from 24y to 71y, median age 46.66±8.45 years. The most common complication of ALD-ACLF was bacterial infection(49/79,62%), followed by AKI(30/79,38%) and HE(25/79,31.6%)，UGIB(6/79,7.9%). A univariate Logistic regression analysis was performed and there was no statistical significance in Age, gender, BMI, ALB, ALT, AST, GGT, WBC, HGB, PLT, Serum folic acid, vitB12, Fe and Cre (P>0.05) in those patients with various complications of ascites, HE, UGIB and bacterial infection (P>0.05).
Serum levels of vitB12 were associated with liver failure but not liver inflammation or cirrhosis
We collected and analyzed the clinical data of the patients with AH, AC and ALD-ACLF to show whether there were statistically significant differences on the levels of serum vitB12. The results showed there were statistically significant differences on the levels of vitB12 between AH and ALD-ACLF, AC and ALD-ACLF (P<0.001). No statistically significant differences on the levels of vitB12 between AH and AC. There was no statistically significant difference on MCV levels among the three groups (F=2.67, P=0.072). There was a statistically significant difference on vitB12 levels and MCV between non-ACLF and ACLF (F=12.86, P<0.001, F=11.51, P=0.02, respectively). (Table 2)
MCV, vitB12 and MDF were independent prognostic predictors in ALD-ACLF
According to the results of Univariate Logistic regression analysis and clinical significance, we selected duration of alcohol consumption, Amount of alcohol consumption, MCV, vitB12, MDF score and MELD score as independent variables for Multivariate Logistic regression analysis. The analysis revealed that MCV, vitB12and MDF score were independent prognostic predictors in ALD-ACLF patients (P<0.05, Table 3). MCV and MDF score were risk factors for adverse outcomes of ALD-ACLF.
A new model of MDF combined with vitB12 and MCV improved the predictive value of MDF on 28-day outcome in ALD-ACLF patients
According to the results of multi-factor analysis, finally, MDF, vitB12 and MCV were selected to jointly predict the prognosis of patients with ALD-ACLF, and then performed binary logistic regression analysis again to establish a new short-term prognosis model for patients with ALD-ACLF. The calculation formula is as follows: Logit(P)=-9.273 + 0.078 * MCV + 0.036 * MDF +(-1.786*1) +(-0.709*2) +(-0.98*3). [0=vitB12(≥2000), 1=VitB12(<1175), 2=vitB12(1175-1709), 3=vitB12(≥1709-<2000)]. The new model was proved to be valuable and promising to predict 28-day outcome (liver function index improved significantly or not) in the patients diagnosed ALD-ACLF (0.11±1.28 vs. -0.57±1.46, P<0.03; Figure 2 panel A). The ROC curve analysis shows that the area under the ROC curve (AUC) and Youden index of the new model are obviously better than MDF solely (Figure 2 panel B).