Comparative Prognostic Performance of Inflammatory and Immunonutritional Indices in Acute Pancreatitis: A Head-to-Head Cohort Study

Background/Objectives: Simple admission laboratory indices have been proposed for early risk stratification in acute pancreatitis (AP), but their relative performance and their value beyond established severity assessment are rarely examined directly. We performed an exploratory head-to-head comparison of eight admission inflammatory and immunonutritional indices, including the Global Immune-Nutrition-Inflammation Index (GINI), for a composite adverse in-hospital outcome. Methods: This single-center retrospective study (2020 to 2026) analyzed the index admission of each adult hospitalized with AP (one observation per patient). Eight indices were evaluated (NLR, CAR, PIV, PNI, HALP, CALLY, CCR, GINI). The primary endpoint was a composite of in-hospital death, intensive care unit (ICU) admission, or prolonged hospitalization (a length of stay of at least 7 days). We assessed discrimination (area under the curve, AUC, with DeLong intervals and Holm-corrected comparison), age-adjusted logistic regression, calibration, bootstrap validation, and incremental value over the Bedside Index for Severity in Acute Pancreatitis (BISAP), with a patient-clustered robustness analysis. Results: Of 238 patients, 73 (30.7%) met the composite outcome (prolonged stay 63, ICU 15, death 2). After age adjustment, all eight indices were associated with the outcome (all p < 0.001). The three numerically highest were NLR (age-adjusted AUC 0.737, 95% CI 0.667 to 0.808), CALLY (0.735), and GINI (0.734), with no significant pairwise difference after correction. Each added discrimination beyond BISAP (BISAP alone AUC 0.631 vs. BISAP plus CALLY 0.752; likelihood ratio p < 0.001). GINI was numerically higher but not significantly better than its components (CAR 0.710, PIV 0.708). Models were well calibrated (calibration slope 0.97 to 1.00). Conclusions: In AP, several admission indices, led by NLR, CALLY, and GINI, gave modest, broadly comparable discrimination for a composite adverse in-hospital outcome and added information beyond BISAP. No index was superior, and the composite was driven largely by prolonged hospitalization. These exploratory findings require prospective external validation before any clinical use.