Development and validation of a multivariable predictive model using immunological and physiological biomarkers for disease severity in amoebic liver abscess

Abstract Early detection of patients at risk of developing severe amoebic liver abscess (ALA) is crucial for timely interference and improved clinical results. The current study aimed to develop and validate a multivariable prediction model combining between routine and inflammatory indicators for clinical severity of ALA. The study was conducted on 70 patients diagnosed with amoebic liver abscess and classified into mild and severe groups. The study markers C-reactive protein (CRP), alanine aminotransferase (ALT), aspartate aminotransferase (AST), alkaline phosphatase (ALP), albumin, total bilirubin, lactate dehydrogenase (LDH), interleukin-6 (IL-6) and interleukin-10 (IL-10) were measured in serum. Independent predictors were recognized by multivariable logistic regression analysis and integrated into a nomogram. Model discrimination was estimated by receiver operating characteristic (ROC) curve analysis and area under the curve (AUC). The results indicated that a significant increase was observed in biochemical markers (CRP, ALT, AST, ALP, Total Bilirubin, LDH, IL-6, and IL-10) in the severe group (52.7 ± 6.3, 92.8 ± 8.4, 85.3 ± 5.6, 277.2 ± 5.2, 3.5 ± 0.82, 421.6 ± 9.7, 96.3 ± 7.6, and 88.2 ± 6.1) compared to mild (18.4 ± 2.2, 42.5 ± 4.1, 38.7 ± 3.5, 134.8 ± 6.2, 1.3 ± 0.61, 234.5 ± 11.3, 64.87 ± 7.8, and 58.77 ± 6.3) respectively and control groups (6.3 ± 1.3, 24.2 ± 3.26, 26.4 ± 3.97, 127.4 ± 4.5, 0.92 ± 0.22, 94.2 ± 3.3, 10.6 ± 2.2,and 8.3 ± 1.74) at ( p < 0.001). These results indicate that differences in hepatic and inflammatory predictors between mild and severe cases suggest their potential in estimating disease severity. A multivariable model provides an accurate and clinically beneficial tool for prediction of severe ALA over individual marker and supports its potential implementation in patient management.