Heart failure with subsequent circulatory shock is usually ascribed to the left ventricle and the systemic circulation. However, in recent years much attention has been paid to acute right ventricular failure (ARVF) and its contribution to circulatory shock, especially in intensive care medicine [1]. ARVF is most often caused by increased afterload of the right ventricle (RV) by chronic or acute pulmonary hypertension (PH) combined with regionally or globally reduced contractile capacity of the RV, although ARVF in the absence of PH is possible [2].

The pulmonary circulation is a low-pressure, low-resistance system. An increase in pulmonary resistance instantaneously raises the RV end-systolic pressure and volume [2–4]. This prolongs the isovolumic contraction, shortens the ejection phase and results in a pumping mechanism similar to that of the left ventricle. The result is both higher oxygen consumption and reduced oxygen supply, as coronary perfusion is diminished by elevated intramyocardial pressure. A normal RV can adapt to both chronic and, to a certain degree, acute afterload increases by hypertrophy. However, when the RV fails to compensate for the increased afterload, devastating progressive ARVF and total circulatory breakdown eventually develops.

There is certainly a need for more research on the mechanisms leading to ARVF as well as how to treat the condition [2]. With its complex nature, and because it is caused by relatively small pressure changes, it is important to use good intact animal models to study the phenomenon. Several previous large animal models have used open chest approaches [5–7]. We believe it is vital to keep the chest closed and leave the pericardium intact as these surgical manipulations alter the pressure-flow relations of the RV and the pulmonary circulation, and hence, disturb the ARVF model.

We therefore aimed to construct a new and feasible large animal model where the afterload of the RV could be changed—preferably reversibly—with minimal invasiveness and closed chest. In such a model the hemodynamic effect of increased afterload on RV can be studied in detail, and it would be possible to include this method in different models of RV failure.

In the pulmonary vasculature, vasoconstriction occurs in response to hypoxia, termed hypoxic pulmonary vasoconstriction (HPV). The mechanism acts in response to low alveolar pO₂ and is a vital physiological function to direct blood to the ventilated alveoli, and thereby substantially reduce pulmonary shunting. With a decrease in alveolar pO₂, HPV onsets within seconds [8], increases rapidly and gradually plateaus. The constriction is mainly localized in small resistance pulmonary arteries (PAs, <200 µm) [9] and is reversible with restoration of alveolar normoxia. Global hypoxia will induce HPV in all lung segments, resulting in a rise in pulmonary vascular resistance (PVR), with estimates ranging from 50 to 300% [10]. HPV has been used as a means to increase both acute and chronic PVR in animal experiments [11], but there is no systematic description in the literature on how to use HPV to induce acute RV afterload increase in pigs.

In this study we explored the use of HPV as a means to increase the pressure load on the RV aiming to determine the fraction of inspired oxygen (FiO₂) level yielding the desired effect without leading to serious dysoxia. We used the standard definition of PH as a mean pulmonary artery pressure (mPAP) of ≥25 mmHg at rest as the target level [12].

The individual animals had substantially different baseline mPAP values despite similar age, weight and handling, 18.3 (4.2) mmHg [mean (sd)] (range 14–27 mmHg). The FiO₂ level needed to increase mPAP above 25 mmHg also varied considerably, 0.15 (0.026) [mean (sd)], range 0.13–0.21. (In the one animal with a baseline mPAP above 25 mmHg, the FiO₂ was lowered to 0.15.) This yielded a highly significant overall increase in mPAP from 18.3 (4.2) to 28.4 (4.6) mmHg (p ≪ 0.0001) (Table 1). Since CO did not change, the change in PVR was also consistent in all animals: from 254(76) dyns/cm⁵ to 504(191) dyns/cm⁵, yielding a p-value of 0.001 (Fig. 1).

Pig	Baseline (BL)					Pulmonary hypertension (PH)
	FiO2	mPAP (mmHg)	PWP (mmHg)	CO (l/min)	PVR (dyn s/cm5)	FiO2	mPAP (mmHg)	PWP (mmHg)	CO (l/min)	PVR (dyn s/cm5)
1	0.30	22	11	3.5	251.4	0.15	34	11	3.9	471.8
2	0.30	17	7	3.2	250.0	0.21	31	7	4.0	480.0
3	0.30	17	7	2.4	333.3	0.13	25	7	3.0	480.0
4	0.30	14	7	2.2	254.5	0.13	25	5	3.9	410.3
5	0.30	17	7	4.5	177.8	0.14	26	8	4.5	320.0
6	0.30	15	8	3.4	164.7	0.14	25	8	2.9	469.0
7	0.30	27	6	4.3	390.7	0.15	36	6	2.5	960.0
8	0.30	17	9	3.0	213.3	0.15	25	9	2.9	441.4

Both mPAP and PVR increase with reduction of FiO₂ and globally induced hypoxic pulmonary vasoconstriction

FiO ₂ fraction of inspired oxygen, mPAP mean pulmonary arterial pressure, PWP pulmonary wedged pressure, CO cardiac output, PVR pulmonary vascular resistance

Blood gases taken at baseline and after reduction of FiO₂ displayed a slight reduction in oxygen delivery (DO₂) from 409 (92) ml/min to 354 (52) ml/min. This was not a statistically significant reduction with a p value of 0.14. Nor did it affect the oxygen consumption (VO₂), which remained unchanged, from 195 (35) ml/min to 200 (41) ml/min (Fig. 2). This was caused by increased oxygen extraction after induction of PH, with a highly significant reduction in mixed venous oxygen saturation (SvO₂) from 0.51 (0.076) to 0.36 (0.070), yielding a p - value of ≪0.0001 (Fig. 3). Additionally, there were no changes in lactate, pH or base excess from baseline to after FiO₂ reduction. For details, see Additional file 1.

Reduction of FiO₂ from 0.3 to 0.15 (0.024) is an efficient method for increasing the afterload of the RV, yielding an overall doubling of the PVR. The effect took place in all animals and should therefore be considered a reliable general physiological feature. As such, this method can be used in future studies of RV adaption to increased afterload; the combination with some method of regional or global RV contractility reduction should in theory induce ARVF (Additional file 1: Table S1).

There was considerable variation in individual animals’ baseline mPAPs and the FiO₂ reduction needed for a sufficient mPAP increase. One could therefore ask whether this kind of model should aim at increasing the mPAP (or PVR) above a clinical definition of PH, as was done in this study, or whether one should aim for a relative change in afterload in each respective animal, say, a 100% increase in mPAP (or PVR) from baseline. Either way, the FiO₂ reduction has to be adapted to the individual animal.

Some animals experienced a relatively large reduction of arterial oxygen saturation (SaO₂) as a consequence of the FiO₂ reduction, leading to decreased DO₂ overall. This did not lead to a decreased VO₂ because of an increased oxygen extraction. It should be noted, however, that this occurred in the stable situation 30 min after establishing the lower FiO₂ value (Fig. 2). We noticed during the initial phase of FiO₂ reduction a transient change in skin colour and a decline in plethysmographic measured SaO₂. After some minutes with their new oxygen supply, the animals regained their previous colour and SaO₂ level. This immediate physiological instability after serious system perturbations followed by stabilization after a few minutes is an interesting physiological feature that we have also experienced in other animal models.

Another word of caution is that in some of the pilot animals prior to this study, we lowered the FiO₂ too rapidly to levels that were too low, yielding an extremely low SvO₂. This is a very strong stimulus to the sympathetic nervous system, leading to tachycardia and a high CO to compensate for the hypoxia. The pulmonary circulation is constructed to be able to receive large volumes of blood without increasing the pulmonary arterial pressure. This is mainly caused by an increase of the total cross-sectional area of the pulmonary vasculature, resulting from passive distension and recruitment of collapsed vessels [13]. In this situation the animals ended up with an increased mPAP from baseline, but with the same or even a reduced PVR. It is therefore important to be very careful when using our method, because it approaches the limit of the animal’s ability to compensate.

In theory, this method of increasing mPAP and PVR by reducing FiO₂ should be fully reversible. Unfortunately, we did not study whether the mPAP returned to its baseline value after re-establishing normal FiO₂, or the time frame needed for this to occur. In the literature, HPV is described as biphasic with one rapid and one late phase. The time aspect is inconclusively described with variations in onset and duration for both phases [10, 13–15]. The long phase is said to last from 45 min to several hours, but there are no conclusive descriptions of how long the hypoxia must be maintained for this phase to kick in, nor how fast the vascular tone is returned to normal with restoration of normoxia. This emphasizes the need for further investigation of the method’s reversibility.

This study was not designed to investigate RV changes as a consequence of increased afterload. In accordance with the theory as described in the background, increased afterload itself should not be enough to cause failure in a healthy RV. This was also our impression during the study, as CVP remained unchanged and CO didn’t show any systematically decrease. However, a full assessment of RV’s adaption to changing loading conditions requires assessment with conductance catheters or similar in order to extract pressure–volume loops [6, 11].

Resistance as defined in practical clinical medicine and in this manuscript as a measure of afterload is an application of Ohm’s law where resistance = Δ pressure/flow. In reality, this is only valid for a system of linear flow. In the pulmonary (and systemic) circulation, flow is oscillatory, in which impedance is a better calculation of the overall afterload. The continuous flow and pressure signals are assessed by Fourier analyzes and then the pressure/flow relations are compared for the respective harmonics [16]. Thus, in order to examine impedance, high-resolution continuous pressure and flow signals are mandatory, which we did not have in this study.

We showed in this study that the mPAP and the PVR could be raised by approximately 100% in the study animals by lowering the FiO₂ from 0.30 to 0.15 (0.024). The responses were individual, as were the baseline mPAPs. This FiO₂ reduction led to reduced SaO₂ and hence DO₂ in some animals, but the VO₂ was not reduced after a stabilization period. We note that in the initial phase, the animals seemed to be unstable, with clinical signs of dysoxia, but they regained normal oxygenation after some minutes. Furthermore, this method should be used with some caution, as reducing FiO₂ too rapidly might lead to sympathetic nervous activation with tachycardia and high CO, resulting in a secondary normalization of PVR despite high mPAP. The method should in principle be reversible, although we did not study this in this work. We therefore present a novel method for minimal-invasive (closed chest) RV afterload manipulations intended for future studies of ARVF.