Study Ethics

This research was conducted as part of a protocol that was approved by the Institutional Animal Care and Use Committee of the University of Minnesota and was compliant with the National Research Council's 2011 Guidelines for the Care and Use of Laboratory Animals (Protocol No. 2305-41090A). Clinical trial number: not applicable.

Study Preparation

This study was an exploratory pilot with no prior data to guide a formal sample size calculation. We selected 11 animals based on alignment with similar hemodynamic studies. Healthy male Yorkshire swine (73.4 ± 5.9 kg) with an average age of 6 months that had previously acclimated to the animal facility setting for 2 −6 weeks and fasted for 12 h, were sedated with Telazol (≤ 500 mg/kg) intramuscularly, followed by Methohexital (≤ 50 mg/kg) after establishing intravenous access. After intubation, anesthesia was maintained with a 1‐to‐1.5% minimum alveolar concentration (MAC) of isoflurane during continuous ventilation. All animals received periodic intravenous boluses of 5,000 IU of Heparin with a total administration of 30,000 IU.

In each animal, the left femoral artery was cannulated with an 18 G 1–3/4″ Jelco IV catheter (Smiths Medical, Minneapolis, MN, USA) for continuous intraarterial pressure (IAP) monitoring. For venous access, a 12.5 French, 13 cm long SafeSheath® II Hemostatic Peel-away Introducer System for Vascular Access (Medtronic, Minneapolis, USA) was used in the right femoral vein and right internal jugular vein. The position of the catheter in the inferior vena cava was confirmed with fluoroscopy. Intraosseous access was gained using a 25 mm,15G EZ-IO® device (Teleflex®, Inc., PA, U.S.A.), drilling it into the proximal tibia of the same side as the femoral arterial access, unless the intraosseous access was not successful (Fig. 1). The lines were flushed with 10 ml heparinized saline (10 IU/ml); no continuous infusions were attached to the lines.

Fig. 1

Micro-computed tomography reconstruction of a 25 mm intraosseous access within the proximal porcine tibia of one of the animals

Intraosseous cannulation was deemed successful when there was a palpable loss of resistance, the needle remained upright, blood could be withdrawn, and saline flush was possible [10]. A successful intraosseous pressure reading was defined as a pulsatile pressure signal under resting conditions.

All measurements were done with the animals in the supine position. The intraosseous, intraarterial, and intravenous lines (in case of central venous pressure) were zeroed at the approximate mid-axillary level. Pressure calibration was calibrated at 0 and 200 mmHg using a Utah Medical Delta-Cal 650–950 (Utah-Medical West Midvale, Utah, USA) transducer simulator and tester.

Simultaneous data were recorded using a physiological data monitoring system (EMKA Technologies, Paris, France), with data points continuously recorded every 2 ms. These pressure measurements underwent denoising with a 10 Hz filter.

Comparative Analyses of the Relationships between Arterial and Intraosseous Pressures

To describe and test the temporal and relative pressure relationships between recorded intraarterial and intraosseous pressures, continuous recordings were shortened to 40-s measurements. Before analyzing a given data set, time differences in pulse peaks (phase shift due to different lengths of the fluid-filled line to the transducer) were corrected (Temporal phase correction by 30–60 ms).

Arterial pressures were compared to intraosseous pressure at baseline (normotension n = 6), during: 1) anesthesia-induced hypotension induced by (systolic blood pressure < 90 mmHg occurring during prolonged anesthesia, n = 2), 2) right jugular transvenous ventricular cardiac pacing at rates of 120/min (right ventricular screw-in lead via the right jugular vein n = 1), 130 (n = 1), 140 (n = 2, Supplementary Fig. 1), 3) pacing-induced ventricular arrhythmias (n = 1, Supplementary Fig. 2), 4) ventilator-induced breath hold (breath hold was performed by locking the ventilator in an inspiratory hold for up to 30 s n = 2); 5) temporary balloon occlusion of the distal inferior vena cava to increase femoral vein pressures (IVC occlusion with a TORAY PTMC-30 Inoue-Balloon fluoroscopically positioned, n = 1, Supplementary Fig. 3), and 6) administration of an epinephrine bolus during spontaneous circulation with stable hemodynamics (200 µg i.v., n = 1).

Intraosseous Pressure Changes Associated with Loss of Spontaneous Circulation

All studies ended in asystole, confirmed by loss of intraarterial pressures. Simultaneous intraarterial and intraosseous pressure recordings were continued during the initiation of cardiac arrest via infusion of St Thomas’ Hospital cardioplegic solution in 6 cases [11]. This led to reproducible patterns of normotension transitioning to hypotension with values < 60 mmHg, analogous to what is clinically seen in pseudo-PEA with the final transition to complete pulselessness and asystole (Figure 4).

Finally, one case of cardiac arrest with subsequently administered low-quality cardiopulmonary resuscitation (CPR) with an automatic piston chest compression device programmed at 110 compressions per minute and a depth of 5 cm, resulting in sustained arterial diastolic pressure < 25 mmHg, was recorded [12].

Statistical Analyses

All pressure values were presented as means ± standard deviations. Statistical analyses were done with R version 3.6.1. Bayesian vector autoregression estimates were calculated using EViews 13.

We calculated Pearson's correlation coefficients for arterial and intraosseous pressure and modeled the dynamic links between intraosseous and intraarterial pressures using Bayesian VAR estimations.

To evaluate abilities to detect cardiac arrest on both arterial and intraosseous pressure tracings, peak pulse pressures (pmax – pmin) before and after arrest were calculated and compared using Wilcoxon rank-sum tests.

A p-value of less than 0.05 was considered statistically significant.

A post-hoc power analysis based on our observed baseline correlation (r ≈ 0.80) confirmed that the sample size was sufficient to detect a strong correlation with 80% power. Subgroup analyses remain exploratory, given the sample size.