Detection of occult hemorrhage using multivariate non-invasive technologies: a porcine study.

Abstract

Objective Occult hemorrhage (OH) can develop subtly after trauma, often without immediate changes in traditional vital signs like heart rate and blood pressure. This creates a critical diagnostic challenge, as clinically significant OH may remain undetected until it progresses to hemodynamic instability or shock. Early identification is crucial since delays in diagnosis are linked to poor outcomes. We evaluated poly-anatomic multivariate technologies-electrical impedance tomography (EIT), near-infrared spectroscopy (NIRS), electrical impedance spectroscopy (EIS), Plethysmography (Pleth), and ECG-in a porcine model to detect OH without needing prior baseline measurements. Approach Forty female swine were bled at controlled slow rates to induce subclinical hemorrhage. During this phase, traditional vital signs remained stable. Continuous data from EIT, NIRS, EIS, Pleth, and ECG were recorded. A supervised voting classification approach was employed to detect OH, using minimally transformed measurements from each technology to train classifiers. A soft majority voting technique combined outputs from multiple technologies to improve accuracy. Main Results The multivariate approach consistently outperformed the best univariate technology (EIT) in predicting OH. Receiver operating characteristic (ROC) analysis revealed that after 21 minutes of observation, the multivariate approach achieved an area-under-the-curve (AUC) of 0.98, compared to 0.92 for EIT alone. The most effective technologies were EIT on the thorax, NIRS on the abdomen, and EIS on the thorax. Significance In this clinically relevant porcine model, multivariate non-invasive measurements were superior in detecting OH compared to univariate methods or standard vital signs. This approach eliminates the need for patient-specific baseline data and enables early detection, potentially improving outcomes by allowing timely intervention before hemodynamic deterioration. Advanced technologies like EIT, NIRS, and EIS hold significant promise for clinical applications in trauma care.