Electrical Impedance Tomography-based Ventilation Patterns for Evaluating Proper Ventilator Settings and to Classify Lung Morphofunction

To the Editor:

We read with great interest the study by Iwata and colleagues, published in a recent issue of the Journal (1). Utilizing electrical impedance tomography (EIT) in postoperative patients, the authors categorized three EIT ventilation patterns that are based on the dorsal fraction of ventilation: <50%, between 50 and 59%, and ⩾60%. This classification has potential applications in assessing the adequacy of positive end-expiratory pressure (PEEP) levels to prevent lung both overdistension and collapse. Notably, the authors observed an association between an uneven distribution of ventral-to-dorsal ventilation and a higher incidence of postoperative complications. We extend our congratulations to the authors for this important work.

In this letter, we aim to address the application of EIT phenotyping on the basis of ventilation distribution patterns to distinguish impairment in lung morphofunction as focal or nonfocal. The proposed classification in the study by Iwata and colleagues (1) is likely appropriate for patients with previously healthy lungs in which the use of the phenotypes is focused on the effect of defining an optimal PEEP. Conversely, this classification can be limited when both focal lung disease and PEEP settings affect ventilation distribution. For instance, elevated PEEP levels in the supine position may induce lung overdistension in the ventral region, leading to a dominance of dorsal ventilation. If this dorsal shift in ventilation occurs in patients with focal dorsal lung impairment, such as bilateral aspiration pneumonia, the ventilation distribution might, indeed, become more homogeneous, thus misclassifying the site of the disease.

Recently, we introduced the concept of morphofunctional phenotypes in patients with acute hypoxemic respiratory failure considering EIT-ventilation distribution among lung quadrants (i.e., ventral right, ventral left, dorsal right, and dorsal left) (2). A nonfocal phenotype was defined when the ventilation distribution was preserved among the lung quadrants (Figure 1A). Focal lung dysfunction was discriminated into focal-bilateral (indicative of reduced ventilation bilaterally in the dorsal region) or focal-unilateral (indicative of reduced ventilation in one of the dorsal quadrants or in a single lung) phenotypes (Figure 1A). Here, we incorporate the phenotype with bilateral ventral hypoventilation, a pattern suggestive of lung overdistension proposed by Iwata and colleagues (1).

Figure 1.

Classification of lung morphofunction. (A) Phenotypes of lung morphofunction according to the distribution of ventilation assessed by electrical impedance tomography (EIT) and (B) the change in the phenotype classification induced by different positive end-expiratory pressure (PEEP) levels. Overdistension is defined by a bilateral ventral hypoventilation (i.e., ventilation distribution <20% for ventral right and ventral left regions). Nonfocal phenotype is defined by a homogeneous ventilation distribution among lung quadrants (i.e., ventilation distribution ⩾20% for ventral right, ventral left, dorsal right, and dorsal left regions). Nonfocal phenotype is further divided into high compliance (bold H; >0.4 ml/cm H₂O/kg) and low compliance (bold L; ⩽0.4 ml/cm H₂O/kg) on the basis of respiratory system compliance. This combination of EIT pattern and respiratory system compliance can help assess the severity of nonfocal diseases, such as in patients with minimal or severe diffuse ground-glass opacities. The focal-unilateral phenotype is defined by reduced ventilation in only one of the dorsal quadrants (i.e., ventilation distribution <20% for dorsal right or dorsal left) or in a single lung (i.e., ventilation distribution <20% for ventral right and dorsal right or ventral left and dorsal left). The focal-bilateral phenotype consists of reduced ventilation bilaterally in the dorsal region (i.e., ventilation distribution <20% for dorsal right and dorsal left). The graph in (B) was produced using available data from previous studies on intubated patients with acute hypoxemic respiratory failure (Pa_O2/Fi_O2 < 300 mm Hg) assessed with EIT from four centers in three countries (Brazil, Chile, and the United States). All patients were monitored with EIT during a decremental PEEP trial in the supine position under pressure or volume-controlled modes without inspiratory effort. cm H₂O = centimeters of water; D = dorsal; L = left; PBW = predicted body weight; R = right; V = ventral.

We contend that accurately classifying morphofunctional phenotypes necessitates the exclusion of confounding effects because of insufficient or excessive PEEP levels. The effect of PEEP on these EIT-morphofunctional phenotypes is evident in Figure 1B. This figure utilizes data from previous studies conducted on intubated patients with acute hypoxemic respiratory failure (n = 80), with a ratio of Pa_O2/Fi_O2 < 300 mm Hg, assessed with EIT from four centers in three countries. These studies were approved by their respective ethical committees: Massachusetts General Hospital in the United States (approval number: 2019P001995), Heart Institute (or, InCor) of the University of São Paulo in Brazil (approval number: 4001231), Federal University of Pernambuco in Brazil (approval number: 4362977), and Hospital Clínico Universidad de Chile (approval number: 027/2016).

Note the significant reduction, exceeding 50%, in the frequency of patients classified with focal-bilateral lung dysfunction when PEEP levels were increased from 5–8 cm H₂O to 15–18 cm H₂O. This phenotype was almost nonexistent at higher PEEP levels. Conversely, lowering PEEP levels markedly diminished the incidence of overdistension, nearly eliminating it. This reduction in PEEP revealed that bilateral ventral hypoventilation is seldom a focal morphofunctional phenomenon but is predominantly affected by ventilatory settings. In addition, it is unlikely to have acute hypoxemic respiratory failure secondary to bilateral ventral lung disease, except for bilateral pneumothorax.

Although reducing PEEP relieves lung overdistension, it is important to recognize that low levels of PEEP may induce dorsal lung derecruitment, which is expected in patients under deep sedation and controlled ventilation, thus indicating a reversible focal lung hypoventilation in the posterior region.

Therefore, to properly assess lung morphofunction with EIT, PEEP should be protocolized to minimize the impact of ventilatory settings. A feasible alternative is using the PEEP value at the crossing point of the overdistension and recruitable collapse curves measured by EIT during a decremental PEEP trial (3), thus concurrently minimizing both phenomena. Physiological studies indicate that this crossing point corresponds to an end-expiratory transpulmonary pressure ranging from 0 to +2 cm H₂O (4, 5), a known threshold associated with reduced lung collapse and overdistension.

Last, both the evaluation of proper PEEP settings and morphofunctional lung disease are influenced by the presence of spontaneous effort. Strong inspiratory efforts can increase ventilation in dependent lung regions (6) and confound accurate morphofunction interpretation, which should be performed under strictly controlled ventilation.

In conclusion, EIT holds promise as a tool in understanding lung pathophysiology and optimizing respiratory support delivery. However, the interpretation of EIT images depends on various factors, including respiratory disease characteristics and ventilator settings. Awareness of these factors is crucial for avoiding misinterpretation of EIT data.