Optimizing PEEP levels during injury development is essential to achieve clinically relevant ARDS in animal models

To the Editor

Preclinical animal models of acute respiratory distress syndrome (ARDS) are a critical component of knowledge generation in the quest to advance the clinical management of human ARDS. However, there is a paucity of positive preclinical findings translated into clinical practice that subsequently impacts patient outcomes (i.e., mortality) (Pham and Rubenfeld 2017). One possible contributor to this issue is a lack of positive end expiratory pressure (PEEP) optimization during the injury development before reaching the ARDS criteria triggering observation phase, which may impact the translatability of animal findings to human clinical practice. By addressing this and other similar issues, the preclinical knowledge may be more efficiently and effectively translated into clinical practice.

Although PEEP optimisation is essential to improve oxygenation and achieve an appropriate level of lung injury, our systematic review of PEEP in ARDS showed that this ventilatory strategy has not been appropriately applied in preclinical settings (Millar et al. 2019). PEEP optimization must be prioritized, especially in circumstances where adjunctive therapies, such as extracorporeal membrane oxygenation (ECMO), are required. However, past animal models of ARDS and ECMO consistently applied insufficient levels of PEEP during injury development (Millar et al. 2019), with ECMO initiated immediately after achieving the target PaO₂/FiO₂ ratio, as a surrogate endpoint of severity, rather than achieving clinically relevant ventilator settings (Battaglini et al. 2024). Since insufficient PEEP levels could often introduce severe hypoxaemia in a short period, this approach might result in subjects with mild lung injury (i.e., not ARDS) being included, leading to erroneous conclusions and difficulty applying preclinical findings to clinical human scenarios. Thus, we call for PEEP optimization, adjusting PEEP levels according to the injury progression (e.g., increase in PEEP as PaO₂/FIO₂ ratio decreases), to be incorporated in all preclinical ARDS model development to achieve consistent and clinically relevant lung injury.

In an ovine model of ARDS instigated by the administration of oleic acid and lipopolysaccharide, we investigated the effects of PEEP optimisation during the injury phase and before achieving the target criteria (Fig. 1A) using data from two randomised controlled preclinical studies [Wildi et al. 2023, Unpublished study]. These studies used the same injury protocol and hemodynamic management, but a different PEEP strategy (Table 1). Upon achieving PaO₂/FiO₂ < 150 cmH₂O and using similar tidal volumes (6 ml/kg) and respiratory rate, two different levels of PEEP were applied during the second injury with lipopolysaccharide: (1) high-PEEP group (n = 15, PEEP 10–14cmH₂O) [Unpublished study], and (2) low-PEEP group (n = 9, PEEP 5–8cmH₂O) (Wildi et al. 2023) (Fig. 1B).

Fig. 1

A Injury Protocol, B Mechanical Ventilation Strategy, C PaO₂/FiO₂ ratio based on the results of arterial blood gas. ECMO extracorporeal membrane oxygenation, FiO₂ fraction of inspiratory oxygen, LPS lipopolysaccharides, PaO₂ partial pressure of arterial oxygen, PEEP positive end-expiratory pressure, RR respiratory rate A 0.03 ml/kg of oleic acid injection was repeated every 30 min until the PaO₂/FiO₂ ratio met Criteria 1 (PaO₂/FiO₂ ratio < 150 cmH₂O). After the initiation of the PEEP adjustment, 0.5 µg/kg of LPS in 50 ml of saline was continuously infused for one hour. Arterial blood gas was measured every 15 min and when necessary, since the first oleic acid injection over the injuries

Table 1 Differences in protocol between the two groups

Despite similar oleic acid doses, we found a notable discrepancy in PaO₂/FiO₂ ratios between the groups within just one hour from PEEP adjustment (Fig. 1C). In the high-PEEP group, the PaO₂/FiO₂ ratio improved to 218 ± 21 mmHg, exceeding the typical clinical indication threshold for ECMO (i.e. PaO₂/FiO₂: 80–120). Consequently, ECMO would likely be not clinically indicated in such a scenario. Conversely, in the low-PEEP group, the PaO₂/FiO₂ ratio declined to 102 ± 11 mmHg, supporting the clinical initiation of ECMO.

Our findings suggest that optimizing PEEP during injury development could substantially alter injury levels in preclinical ARDS models, and more accurately align ARDS characteristics with those seen in clinical practice. Preclinical models must reproduce clinical settings and adhere to clinical guidelines as closely as possible during the development of ARDS (Grasselli et al. 2023) in order to ascertain relevant, clinically applicable results in this area. Additional research is imperative to explore how to effectively refine the application of PEEP in preclinical ARDS models, aiming to provide guidance for its application and establish reliable models capable of replicating the complexity of ARDS, thereby facilitating translatable discoveries. Finally, future preclinical studies should incorporate clinically relevant adjunctive therapies, such as neuromuscular blockade and prone positioning, to better replicate the complexity of ARDS in real-world scenarios.