Comparison of two porcine acute lung injury models: a post-hoc analysis – Intensive Care Medicine Experimental

In the present post-hoc analysis of two different animal ARDS models, we investigated similarities and/or differences of macroscopic findings to identify a pre-clinical model with the most suitable clinical features seen in relation to human ARDS. Three major findings are:

1. (1)

   both ARDS models are feasible and reproducible and contribute to an impairment of the gas exchange, whereas

2. (2)

   the LPS-induced ARDS caused the most severe cardiovascular and metabolic insufficiency and

3. (3)

   the double-hit model impressed with higher mechanical ventilator settings and altered pulmonary mechanics.

Many different animal models exist to mimic human ARDS and its pathophysiological features to help better understand this syndrome [3]. The chosen animal model should accurately reproduce the various patterns of the disease. Pigs used as a model for lung injury provide numerous advantages compared to rodents: the anatomy, genetics, and physiology are remarkably similar to humans [16]. There is no doubt, when chosen the right model, that acute lung injury models with pigs will offer new crucial findings in the complex pathophysiology of ARDS in the future with innovative renewals from bench to (human) bedside. In the following, the findings of our post-hoc analysis are classified in the context of alterations seen in human ARDS.

Focus hemodynamic system

Hemodynamic alterations are often seen in ARDS. Up to 60% of the patients experience hemodynamic failure due to (1) pulmonary hypertension and acute cor pulmonale, (2) vasoplegia and (3) arterial remodeling due to sepsis-induced vascular dysfunction [17, 18]. To improve perfusion, cardiac output and to measure EVLWI and balance the infusion therapy, it is recommended to use transpulmonary thermodilution systems as done in our studies [19]. The hemodynamic changes mentioned were observed more frequently in the LPS group in the presented analysis. Heart rate, mean arterial pulmonary pressure, cardiac output and wedge pressure were elevated similar to human septic ARDS. Mean arterial blood pressure was kept stable in both groups. However, this only happened due to significantly higher norepinephrine doses in the LPS group. This observation may suggest a septic-induced vascular dysfunction in these animals. The role of different catecholamines in (bovine) ARDS has not been investigated so far. Thereby, the LPS induced ARDS model could offer a suitable option. Furthermore, EVLWI was elevated in both groups. One key finding in the pathophysiology of ARDS is the development of lung oedema. Elevated EVLWI reflects the persistence of pulmonary edema. Similar results were found in a septic model in minipigs. In this study, EVLWI was elevated over time after fecal peritonitis [20]. As seen here, both models show a significant damage to the alveolar capillary unit which causes the characteristic oedema. The measurement of the PCWP, as a surrogate marker for a cardiogenic pulmonary oedema, supports these findings. Neither the ARDS induced with LPS nor the double-hit ARDS lead to relevant changes in the PCWP values. It is well-known, that elevated PCWP values are associated with poor outcome in ARDS due to right ventricular failure [21, 22]. In our study, it remains unclear why the PCWP do not raise in the double-hit model. It is reported that especially oleic acid infusion elevates significantly the PCWP in an ARDS model in dogs [23]. To summarize, different fluid, catecholamines and transfusion therapy regimes could be addressed and studied with both models to reduce the impact of an edema and ameliorate the hemodynamics in ARDS.

Focus respiratory system

In the past, the adverse effects of mechanical ventilation in patients with ARDS are discussed [24]. The inhomogeneity of gasless regions up to hyperinflated areas are present in the lung and contribute to the ventilator-induced lung injury (VILI) [25]. Furthermore, lung oedema, anatomic variations and the reduced ventilatable lung tissue make “non-adverse” ventilation difficult. Not higher tidal volumes itself damage the lung, it is the mechanical power delivered by the mechanical ventilation that affects the lung and the development of VILI [25]. Respiratory rate, mean and peak airway pressures as well as PEEP are other main determinants of mechanical power. Optimal PEEP in patients is still discussed. For example, some authors recommend higher PEEP in patients with the highest recruitable lung parenchyma and the most hypoxemic patients [26]. As seen in our analysis, higher peak and mean airway pressures were generated directly after ARDS induction and maintained over time in the double-hit model. Furthermore, respiratory rate and minute ventilation was increased in the double-hit model compared to LPS. The impact of different animal models to mechanical ventilation parameters has not been investigated so far, whereas different types of ventilation modes and their impact of VILI and inflammation were well-analyzed. In an ARDS model with piglets, the influence of spontaneous breathing or mechanical ventilation on abdominal oedema and inflammation was investigated [27]. Decreased lung compliance is another pathophysiological finding in human ARDS. Since COVID-19, it is known that different compliance phenotypes in human ARDS exist and the research focus on this parameter became more popular [28]. In both groups, the compliance decreased immediately after ARDS induction and remained lower over time. This effect was more worsen in the double-hit group. Yet, changes in the lung compliance have been investigated in pigs during pronation in ARDS [27]. Due to the reduced compliance, hypoxemia was more severe in the double-hit model immediately after ARDS induction. The PEEP levels also increased more in this group. Determining the optimal PEEP settings in clinical routines is challenging, especially when protective ventilation strategies must be followed. Optimizing PEEP research is often combined with the use of the EIT. In a bovine double-hit ARDS model, using lavage and high tidal volumes to induce lung injury, similar results were found as reported in our study [29]. In conclusion, the double-hit model offers advantages in respiratory research when focusing on different mechanical ventilator strategies (a.e. influence of different PEEP and airway pressures on lung inflammation).

Focus inflammation

LPS-induced and sepsis-associated ARDS is characterized by systemic inflammatory changes: imbalance of inflammatory response, immune dysfunction and mitochondrial damage [30]. The pathology of sepsis-induced ARDS is extremely complex. IL-6 plays a key role in promoting pulmonary vascular dysfunction, microthrombi and failure of hypoxic pulmonary vasoconstriction (HPV) with in consequence elevated pulmonal arterial pressures [31]. In our study, elevated expression of IL-6 was observed in the lung tissue from LPS-animals probably contributing to significant higher mPAP overtime in this group. Furthermore, elevated lactate levels, decreased base excess and lower pH values may reflect tissue hypoxia and changes in lactate metabolism [32]. Especially elevated lactate levels over time are associated with higher mortality in critical ill patients [32]. In addition, the drops in the white blood cell count and thrombocytes are also common in sepsis-induced ARDS. These “sepsis-like changes” were seen only in the LPS group in our comparison. Gram-negative and LPS-associated sepsis is one of the most causes in human ARDS and has clinical relevance [33]. Nevertheless, a major disadvantage of this model is that the response to endotoxin has significant interspecies variation, with dogs being more tolerant to endotoxin exposure than pigs, sheep or humans [33]. In conclusion, in the double-hit model no remarkable inflammatory changes were observed. The LPS-induced ARDS model offers clear advantages when focusing on studies with inflammatory changes.

Our study has some limitations. (1) The short follow-up period of 8 h addresses only the acute phase of an ARDS. Long-term effects will not be shown. (2) To reduce confounding variables all pigs were of the same gender, a situation not seen in clinical daily praxis. (3) Concerning the double-hit model, it remains unclear in which part the oleic acid infusion or the bronchoalveolar lavage cause the observed lung injury. (4) Despite all the similarities between humans and pigs, results of animal studies need to be translated to clinical practice.

In conclusion, the LPS-induced ARDS caused the most severe cardiovascular and metabolic insufficiency and has clinical relevance due the gram-negative nature of LPS. The double-hit model impressed with higher mechanical ventilator settings. The results are in conclusion with findings in sheep and humans that support the value of different animal models [34]. We can conclude that the different causes of ARDS resulted in the same clinical starting point with severe gas exchange problems. However, in the short time of 8 h in our experiments the underlying causes of ARDS affected the clinical properties of this models in the further course. The choice of which ARDS animal model to use must be carefully considered based upon the focus of the study. Acid aspiration, hyperoxia and bleomycin models also exist and needed to be addressed in further comparison studies to identify an ARDS animal model with the most clinical features and accordance of ARDS in the future.