Abstract
<title>Abstract</title> <p>Background Polytrauma is a global killer, yet the standard Shock Index (SI) frequently misclassifies elderly patients and neurotrauma victims due to blunted chronotropic drives or severe traumatic brain injury (TBI) autonomic dysregulation. Composite indices integrating age (ISA), neurology (SI/G), or both (ISA/G) offer theoretical benefits but lack validation in complex Latin American referral settings. Objective To evaluate and compare four SI variants (SI, ISA, SI/G, ISA/G) for predicting intrahospital mortality in adult polytrauma patients, identifying independent death predictors through multivariable regression. Methods Retrospective cohort study (2022) at a Peruvian Level IV center (N = 296). Upon emergency admission, four indices were calculated using triage vital signs: SI (HR/SBP), ISA (SI×Age), SI/G (SI/GCS), and the combined ISA/G [(SI×Age)/GCS]. After Kolmogorov-Smirnov normality testing, categorical variables were compared using Chi-square and continuous variables via Student's t or Mann-Whitney U tests. Area under the ROC curve (AUC) measured discriminative capacity for hospital death. Independent predictors were isolated using backward conditional logistic regression (Stata v.17) Results The cohort exhibited extreme clinical severity (73.99% overall mortality; 81.76% Priority I). Standard SI performed at random chance level (AUC = 0.496; 95%CI: 0.419–0.573). Conversely, adding sequential physiological adjustments progressively restored the prognostic signal, with the doubly-adjusted ISA/G index achieving peak discrimination (AUC = 0.660; 95%CI: 0.592–0.728). Multivariable modeling confirmed that an entry ISA/G ≥ 1.602 increased adjusted odds of death fivefold (OR = 4.989; 95%CI: 1.898–13.113), functioning independently of severe cranial trauma (OR = 10.112) and immediate invasive airway requirement (OR = 5.428). Conclusion Standard SI is insufficient in highly critical trauma cohorts. The composite ISA/G index bridges crucial demographic and clinical gaps, providing an immediate, resource-independent bedside triage tool that seamlessly integrates into regional risk-stratification frameworks supports integration into triage protocols during the Golden Hour in complex Latin American trauma centers.</p>