Identification of regional overdistension, recruitment and cyclic alveolar collapse with electrical impedance tomography in an experimental ARDS model

Abstract

Background: Information on regional ventilation distribution in mechanically ventilated patients is important to develop lung protective ventilation strategies. In the present prospective animal study, we introduce an electrical impedance tomography (EIT)-based method to classify lungs into normally ventilated, overinflated, tidally recruited/derecruited and recruited regions.

Methods: Acute respiratory distress syndrome (ARDS) was introduced with repeated bronchoalveolar lavage in ten healthy male pigs until the ratio of arterial partial pressure of oxygen and fraction of inspired oxygen (PaO2/FiO2) decreased to less than 100 mmHg and remained stable for 30 minutes. Stepwise positive end-expiratory pressure (PEEP) increments were performed from 0 cmH2O to 30 cmH2O with 3 cmH2O increase every 5 minutes. Respiratory system compliance (Crs), blood gases and hemodynamics were measured at the same time. Lung regions at end-expiration and during tidal breathing were identified in EIT images.

Results: Overinflated regions contain air at end-expiration but they are not or are only minimally ventilated. Recruited regions compared to reference PEEP level contain air at end-expiration of arbitrary PEEP level but not at that of reference PEEP level. Tidally recruited/derecruited regions are not represented in lung regions at end-expiration but are ventilated during tidal breathing. The results coincided with measurements of blood gases. The coefficient for correlation between the number of recruited pixels and PaO2/FiO2 was 0.89 ± 0.12 (p = 0.02).

Conclusion: The proposed novel EIT-based method provides information on overinflation, recruitment and cyclic alveolar collapse at the bedside, which may improve the ventilation strategies used.

Keywords: Acute respiratory distress syndrome; Electrical impedance tomography; Mechanical ventilation; Overinflated; Recruitment/derecruitment.

Figures

Fig. 1

Illustration of the proposed analysis method. First row dynamic images at end-inspiration (a) and end-expiration (b) are identified (averaging of five corresponding images is used to minimize noise if necessary). Status image (c) of tidal variation equals pixel impedance values at end-inspiration minus that at end-expiration. Second row lung regions containing air are identified at reference positive end-expiratory pressure (PEEP) level and arbitrary PEEP level (d, white regions end-expiration at reference PEEP; e, light gray end-expiration at arbitrary PEEP; f, gray lung regions for tidal breathing). Third row overdistended regions (i, white regions) are defined as regions that contain air at end-expiration but no air comes in or out during tidal breathing. Fourth row regions subjected to repeated alveolar collapse and reopening (l, red regions) are defined as regions that are not in the lung regions at end-expiration but are ventilated during tidal breathing. Fifth row recruited regions compared to reference PEEP level (o, blue regions) are defined as regions that contain air at end-expiration of the arbitrary PEEP level but not at that of the reference PEEP level. For the combination of the three types of regions please refer to, e.g., Fig. 2, PEEP = 9 cmH2O

Fig. 2

Regions that were identified as overdistended (white), recruited (blue) and tidally recruited/derecruited (red) during the incremental positive end-expiratory pressure (PEEP) trial in one pig. PEEP increased from 9 to 30 cmH2O. Recruited regions were calculated compared to a PEEP of 6 cmH2O. Lung regions are outlined in yellow. Percentage values at the bottom of each sub-figure were calculated by dividing theabsolute number of pixels of the different conditions by the total number of lung pixels

Fig. 3

Summary of results obtained in all 10 pigs during the incremental positive end-expiratory pressure (PEEP) trial. Left number of pixels in overinflated (black solid line), recruited (blue broken line) and tidally recruited/derecruited (red broken line) regions divided by total number of lung pixels in percentage (medians and half interquartile ranges). Recruited regions were calculated compared to a PEEP of 6 cmH2O. The partial pressure of O2 in arterial blood/fraction of inspired oxygen (PaO2/FiO2) ratio (blue stars) and respiratory system compliance (Crs, red circles) are presented as median and half-interquartile ranges