Physiological effects of lung-protective ventilation in patients with lung fibrosis and usual interstitial pneumonia pattern versus primary ARDS: a matched-control study

BACKGROUND

Although patients with interstitial pneumonia pattern (ILD-UIP) and acute exacerbation (AE) leading to severe acute respiratory failure may require invasive mechanical ventilation (MV), physiological data on lung mechanics during MV are lacking. We aimed at describing the physiological effect of lung-protective ventilation in patients with AE-ILD-UIP compared with primary ARDS.

METHODS

Partitioned lung and chest wall mechanics were assessed in a series of AE-ILD-UIP patients matched 1:1 with primary ARDS as controls (based on BMI and PaO2/FiO2 ratio). Three PEEP levels (zero = ZEEP, 4-8 cmH2O = PEEPLOW, and titrated to achieve positive end-expiratory transpulmonary pressure PL,EE = PEEPTITRATED) were used for measurements.

RESULTS

Ten AE-ILD-UIP patients and 10 matched ARDS were included. In AE-ILD-UIP median PL,EE at ZEEP was - 4.3 [- 7.6- - 2.3] cmH2O and lung elastance (EL) 44 [40-51] cmH2O/L. At PEEPLOW, PL,EE remained negative and EL did not change (p = 0.995) versus ZEEP. At PEEPTITRATED, PL,EE increased to 0.8 [0.3-1.5] cmH2O and EL to 49 [43-59] (p = 0.004 and p < 0.001 compared to ZEEP and PEEPLOW, respectively). ΔPL decreased at PEEPLOW (p = 0.018) and increased at PEEPTITRATED (p = 0.003). In matched ARDS control PEEP titration to obtain a positive PL,EE did not result in significant changes in EL and ΔPL.

CONCLUSIONS

In mechanically ventilated AE-ILD-UIP patients, differently than in patients with primary ARDS, PEEP titrated to obtain a positive PL,EE significantly worsened lung mechanics.

Numerical Assessment of Damage Parameters for a Hard Interface Model

Adhesive interfaces are suitable modelling tools to describe very thin elastic layers and the related occurring phenomena (such as damage, viscosity, friction, etc.), without using a volumetric description, which is often computationally prohibitive in a large-scale numerical simulation. A major drawback of these kinds of models is the identification of free parameters, because of the smallness of a direct observation scale. This paper proposes a numerical assessment of two model parameters, a damage energy threshold and a damage viscosity, of a hard interface model previously formulated by authors. The proposed assessment protocol uses macroscopic experimental data, available in the literature, on structural adhesives under standard characterization tests. The numerical results obtained give insights into the physical interpretation of these parameters.