Capnodynamic assessment of mixed venous oxygen saturation in a porcine experimental endotoxemic model

Abstract

Sepsis continues to be a major cause of death and illness globally, posing significant challenges for healthcare professionals. In the pursuit of more accurate and timely monitoring tools, the concept of capnodynamically derived mixed venous oxygen saturation (Capno-SvO₂) has emerged as a promising method. Capno-SvO₂ provides a non-invasive way to assess and track SvO₂ and could serve as an additional tool alongside more invasive methods like the pulmonary artery catheter. This could potentially be of great value in the care of critically ill patients with sepsis, where alternative minimal invasive monitoring methods may vary in reliability. The aim of the current study was to compare capno-SvO₂ against values obtained through pulmonary artery blood sample CO-oximetry and continuous fiberoptic SvO₂ monitoring, using a well-established porcine experimental sepsis model. Anesthetized pigs were exposed to a standardized endotoxin infusion sepsis protocol, followed by a series of maneuvers typically applied in sepsis care. Simultaneous recordings were done throughout the experiment for all three monitoring methods. Bland-Altman analysis corrected for repeated measurements was used to assess the agreement of absolute values between the paired recording of CO-oximetry and Capno-SvO₂ as well as between CO-oximetry and fiberoptic SvO₂. The ability of Capno-SvO₂ and fiberoptic SvO2 to track changes was assessed by concordance rate. A total of 10 animals and 275 paired datapoints were included in the study. The majority of the animals displayed pronounced hemodynamical instability in response to endotoxin exposure and subsequent treatment interventions. Analysis of all paired data points showed a bias between Capno-SvO₂ and CO-oximetry SvO₂ of + 1% with 95% limits of agreement of -14% to + 17%. The corresponding numbers for fiberoptic SvO₂ and CO-oximetry SvO₂ were -4% and -15% to + 8%. The concordance rate as compared to CO-oximetry, were 97% and 93% for Capno-SvO₂ and fiberoptic SvO₂, respectively. In this experimental sepsis model, continuous, non-invasive Capno-SvO₂ generates average absolute values comparable to the gold standard CO-oximetry albeit with relatively wide limits of agreement. Capno-SvO₂ displayed a concordance rate of 97% against CO-oximetry and exhibits better trending ability compared to invasive fiberoptic SvO₂.

Keywords: Hemodynamic monitoring; Mixed venous saturation; Non-invasive; Sepsis; Septic shock; SvO2.

Fig. 1

Event plot showing SvO₂ responses to the various hemodynamic challenges for all three SvO₂ monitoring methods. Values are mean (95% CI). N = 10 (refers to number of animals). Vertical black arrows indicate recalibration points for fibreoptic SvO₂. BL1-5: Precision Baseline, Induction BL: Baseline before induction sepsis, FiO₂: Fraction of inspired oxygen; PEEP: Positive End Expiratory Pressure (cmH2O), Norepi; Norepinephrine, HemBL; Baseline before start hemorrhage.

Fig. 2

Bland–Altman plots for pooled recordings for Capno-SvO₂ vs. CO-oximetry (A) and fiberoptic SvO₂ vs. CO-oximetry (B). Dotted black line represents bias and black broken lines represent upper and lower limits of agreement. N = 10. A: 275 paired data points, B: 275 paired data points.

Fig. 3

Four quadrant plots presenting the concordance between Capno-SvO₂ and CO-oximetry SvO₂ (A) and between fiberoptic SvO₂ and CO-oximetry SvO₂ (B). A: 70 paired delta values, B: 70 paired delta values. The central boxed area illustrates the 10% exclusion zone. Dotted line is line of identity. N = 10 (refers to number of animals).

Fig. 4

Continuous data from animal nr 5 and 7. Blue line represents unfiltered Capno-SvO2 (updated breath by breath), orange line represents Capno-SvO2 filtered as a moving mean over 50 s, green line represents Fiberoptic SvO2 and black crosses represents blood gases from the PAC.