ObjectiveTo evaluate the efficacy of γ-glutamyl transpeptidase to platelet ratio (GPR) and its component indexes γ-glutamyl transpeptidase (GGT) and platelet (PLT) in predicting the necro-inflammatory and fibrotic levels of liver in HBeAg-positive and -negative patients with chronic hepatitis B (CHB).
MethodsA total of 323 HBeAg-positive and 254 HBeAg-negative CHB patients were enrolled for the study, who had undergone liver biopsy. Liver pathological sates were evaluated using the Scheuer scoring system.
ResultsIn HBeAg-positive patients, the areas under receiver operating characteristic curves (AUROCs) of GPR in predicting pathological grade ≥G2 and ≥G3 (0.795 and 0.831) were significantly larger than those of PLT in predicting pathological grade ≥G2 and ≥G3 (0.695 and 0.742) (P < 0.000 1 and P= 0.004 6, respectively). The AUROCs of GPR in predicting pathological stage ≥S2, ≥S3 and ≥S4 (0.726, 0.818 and 0.850) were significantly larger than those of GGT in predicting pathological stage ≥S2, ≥S3 and ≥4 (0.692, 0.770 and 0.791) (P=0.002 8, P=0.000 1 and P=0.000 1, respectively). According to the Youden indexes, the optimal cutoffs of GPR in predicting pathological grade ≥G2 and ≥G3 were > 0.376 and > 0.662, respectively, and those in predicting pathological stage ≥S2, ≥S3 and ≥S4 were > 0.368, 0.420 and > 1.106, respectively. In HBeAg-negative patients, AUROCs of GPR in predicting pathological grade ≥G2 and ≥G3 (0.853 and 0.908) were significantly larger than those of PLT in predicting pathological grade ≥G2 and ≥G3 (0.701 and 0.718) (P < 0.000 1 and P < 0.000 1, respectively). The AUROCs of GPR in predicting pathological stage ≥S3 and ≥S4 (0.839 and 0.858) were significantly larger than those of GGT in predicting pathological stage ≥S3 and ≥S4 (0.798 and 0.804) (P＝0.002 8 and P＝0.004 6, respectively). According to the Youden indexes, the optimal cutoffs of GPR in predicting pathological grade ≥G2 and ≥G3 were > 0.562 and > 0.943, respectively, and those in predicting pathological stage ≥S2, ≥S3 and ≥S4 were > 0.566, > 0.798 and > 0.963. With reference to the optimal cutoffs, the sensitivity and specificity of GPR in predicting pathological grade ≥G2, ≥G3 and stage ≥S3, ≥S4 were more than 70% regardless of the HBeAg-positive or -negative patients.
ConclusionGPR can effectively predict the different pathological states of the liver whether HBeAg is positive or negative. However, the optimal cutoffs of GPR are not exactly consistent with predicting the same pathological states of HBeAg-positive and -negative patients.
DownLoad: