NON-INVASIVE SERUM BIOMARKERS TO PREDICT ADVANCED FIBROSIS IN AUTOIMMUNE HEPATITIS IN MEXICAN POPULATION: RPR AND HALP-SCORE

Introduction and Objectives

Looking for non-invasive and cost-effective models to predict fibrosis in autoimmune hepatitis (HAI), modern scores such as RDW/Platelet ratio (RPR) and the hemoglobin, albumin, lymphocyte, and platelet (HALP) have been studied. These have demonstrated greater accuracy than traditional markers such APRI or FIB-4. These require further study and validation in mexican population.

Materials and Methods

Eighty-one patients histopathological diagnosis of HAI from Hospital de Especialidades IMSS Monterrey, Mexico, were included from January 2022 to December 2024. Fibrosis was assessed using METAVIR system, categorized as F0: absence of fibrosis; F1: portal fibrosis without septa; F2: portal fibrosis with few septa; F3: multiple bridging; and F4: cirrhosis. Grades F3-F4 were considered advanced fibrosis. Biochemical values ​​were obtained 1–20 days prior biopsy. The RPR was calculated RDW [%]/Platelets [10⁹/Liter] and HALP-Score was obtained (Hemoglobin [g/dL] by albumin [g/dL] by lymphocytes [k/uL]/Platelets [10⁹/Liter]. Statistical analyses were performed using SPSS 25.0 (IBM SPSS Statics). Type of study: Analytical, retrospective and observational.

Results

For RPR, an area under the curve (AUC) of 0.628 was obtained with confidence intervals (95%) of 0.461–0.796. The cutoff point, calculated using Youden index, was 0.0617. The sensitivity of the RPR was 87.9%, with a specificity of 26%. The positive likelihood ratio was 1.2 and the negative likelihood ratio was 0.45. The false negative rate was 9.8%. For HALP-Score, an AUC of 0.560 was obtained with confidence intervals (95%) of 0.636–0.600. The cutoff point, calculated using Youden index, was 0.497. The sensitivity of HALP-Score is 62.1%, with a specificity of 38%. The positive likelihood ratio is 1.04 and the negative likelihood ratio is 0.95. The false negative rate is 30.86%.

Conclusions

These models didn't demonstrate statistical significance, and therefore warrant further study for validation as an auxiliary tool in our population.

Conflict of interest: None

ROC Curve