Comparative analysis of critical care echocardiography and invasive hemodynamic monitoring in a porcine model of hyperdynamic sepsis

Background: Porcine models offer anatomical and physiological parallels to humans, enabling invasive monitoring and echocardiographic assessment. However, agreement between transthoracic echocardiography and invasive cardiac output measurements in porcine sepsis remains unclear. Aim: To evaluate septic hemodynamics in a porcine model using integrated echocardiographic and invasive monitoring, focusing on concordance between methods. A further aim was to examine how antibiotic treatment influenced disease progression. Methods: This prospective study included 21 anesthetized, mechanically ventilated pigs. Five were excluded for inadequate imaging quality; the remaining 16 were randomized to untreated sepsis (n = 8) or sepsis with antibiotic treatment (piperacillin-tazobactam, n = 8). Three animals died before the second measurement, resulting in n = 8 and n = 5, respectively. Sepsis was induced via titrated infusion of Escherichia coli (CPE NDM-1) to achieve moderate pulmonary hypertension. Invasive monitoring used a femoral thermistor-tipped arterial catheter, Swan-Ganz catheter, and PiCCO. Transthoracic echocardiography assessed LVEF (Teichholz), tissue Doppler S', tricuspid annular plane systolic excursion, diastolic indices, and LVOT VTI-derived cardiac output. Measurements were obtained at baseline and 12 hours post-induction. Agreement was evaluated using Pearson correlation and Bland-Altman analysis. Results: Both groups developed hyperdynamic, vasoplegic sepsis characterized by elevated heart rate and cardiac output with reduced systemic vascular resistance; preload, stroke volume, and systolic function were preserved. dPmax and RVSW increased significantly, with the greatest increases in untreated animals. Echocardiographic cardiac output showed a moderate-to-strong correlation with invasive measurements (r = 0.720, p < 0.05). Bland-Altman analysis showed systematic underestimation, with a mean bias of +1.259 L/min (95% LoA −1.332 to +3.850 L/min) and a percentage error of 74.3%, exceeding the 30% threshold for interchangeability. No proportional bias was detected (r = −0.292, p = 0.148). Antibiotic treatment attenuated the hyperdynamic response, reflected by lower cardiac output and heart rate, though interpretation was complicated by higher norepinephrine requirements. Conclusion: Critical care echocardiography correlated moderately to strongly with invasive cardiac output measurements, supporting its utility for tracking hemodynamic trends in porcine sepsis. However, it consistently underestimated cardiac output, underscoring limitations for precise quantification and supporting its role primarily as a qualitative, trend-based tool in porcine sepsis research.