Transpulmonary Pressure to Guide Mechanical Ventilation: Art or Science?

The pressure required to move the respiratory system and generate a breath in spontaneously breathing subjects is negative and is produced by the contraction of the respiratory muscles; in patients on controlled mechanical ventilation, it is positive and is generated by the ventilator (1). Considering inertia nihil, the pressure required to generate a breath 1) comprises the pressure to overcome resistance and elastance of the respiratory system; 2) is partitioned among the different components of the respiratory system (i.e., lung and chest wall); and 3) is generated by the gradient developed across the different components of the respiratory system (“transmural pressure”) (Figure 1) (1).

Figure 1.

Schematic representation of the relevant positive pressures required during controlled mechanical ventilation to generate a breath and generated by the gradient developed across the different components of the respiratory system (“transmural pressure”).

We learned that ventilatory strategies minimizing the pressure applied to the lung (i.e., transpulmonary pressure [Pl]) can decrease mortality (2). On the basis of the assumption that the pressure across the respiratory system (Prs) closely approaches Pl, protective ventilatory settings are obtained, limiting Prs to a maximal value of 30 cm H2O (2). However, Talmor and colleagues found that only 24% of the variance in Pl was explained by Prs, whereas 52% was due to variation in chest wall pressure (Pw) (3). Several studies confirmed these observations (4, 5), leading to the suggestion that assessment of Pl instead of Prs should be considered to optimize protective ventilatory strategies (2).

The most used method to provide an indirect determination of Pw uses an air-containing latex balloon sealed over a catheter placed in the esophagus and transmitting balloon pressure to a transducer (esophageal pressure [Pes]) (1). Recent studies highlight the potential clinical relevance of using measurements of Pes to estimate Pl. Grasso and colleagues suggested that positive end-expiratory pressure (PEEP) should be titrated to reach a maximal value of 20 cm H2O of Pl instead of a maximal value of 30 cm H2O of Prs (4). Recent clinical trials tested the hypothesis that alveolar collapse and hypoxemia may be determined by negative values of Pl caused by high Pw and that, therefore, setting PEEP to maintain a positive Pl might optimize alveolar recruitment, minimizing the risk of hyperinflation and of hemodynamic impairment (6, 7). Central in this hypothesis is the assumption that the absolute value of Pes would provide an estimate of Pw accurate enough to allow a clinically relevant assessment of Pl. The first single-center trial enrolled 61 patients and found that the ratio of the partial pressure of arterial oxygen to the FiO2 at 72 hours was 88 mm Hg higher in the Pes-guided group than in the control group (95% confidence interval, 78.1–98.3; P = 0.002) (6). The second multicenter trial enrolled 200 patients. The primary outcome, a ranked composite score incorporating death and ventilator-free days among survivors through Day 28, was not different between groups (7).

In this issue of the Journal, Sarge and colleagues (pp. 1153–1163) present a post hoc analysis of the latter trial (8). The authors found that Pes-guided PEEP was associated with lower mortality among patients with lower Acute Physiology and Chronic Health Evaluation-II (APACHE-II), having the opposite effect in patients with higher APACHE-II. Moreover, in patients without shock or hypotension at baseline, Pes-guided PEEP was associated with more days free of shock and of ventilator support. By contrast, in patients with vasopressor-dependent shock at baseline, Pes-guided PEEP was associated with fewer shock- and ventilator-free days. The relevance of the observation that, independent of treatment group or multiorgan dysfunction severity, mortality was lowest when PEEP titration achieved end-expiratory Pl near 0 cm H2O should not escape our notice (8). Despite the intrinsic limitations of a retrospective, post hoc, not prespecified analysis, and the fact that the APACHE score does include variables not directly related to organ failure, these data seem to suggest that measurements of Pes may allow replacing Prs with Pl as a tool to set ventilator settings and that targeting an end-expiratory Pl nearer 0 cm H2O might enhance lung protection.

Years ago, the clinical use of Pes was proposed by clinical physiologists (9) that Berwick would define as “visionary innovators” (10). However, because clinicians tend to adopt innovations very slowly, to introduce Pes in the clinical practice, the scientific community still needs “to move to sort through ambiguous evidence, proceed to collaboration within a community of observers, and finally move to consensus for action” (11). According to the German philosopher Immanuel Kant (12), “actions are acceptable only if everyone could do or could receive it.” Under these circumstances, several steps need to be taken to make measurement of Pes acceptable according to Immanuel Kant’s view. First, Pes measures a local value of pressure, while Pw is not uniform (13), and therefore, relating overall lung volume and Pl to a local value of Pw may be misleading (13). Moreover, whereas Pes in upright subjects closely reflects absolute values of Pw, in supine or semirecurrent postures, absolute values of Pes do not accurately assess absolute values of Pw and may be largely different, depending on gravitational forces and the lung–chest shape (13). Terragni and colleagues performed direct measurements of Pl outside the thorax (during ex vivo lung perfusion “before” lung transplantation) and indirect measurement of Pl using values of Pes of the same lung inside the thorax (after lung transplantation) (14) and suggested that a correction factor of 5 cm H2O should be subtracted from the measured values of Pes to have a correct estimate of absolute Pes (14). Rigorously conducted studies are therefore needed to assess the validity of absolute or corrected values of Pes that should be used to calculate Pl. Second, future clinical trials should use enrichment strategies to optimize selection of patients who may benefit the most from the use of a Pes-guided ventilatory strategy and eventually randomize only patients with a severe impairment of the elastic properties of the chest wall as assessed by prerandomization measurements of chest wall elastance. Third, large, multicenter international observational studies collecting Pes values might be key, but this implies that the physiological meaning and interpretation of Pes should be uniform for all clinicians, and that we must expand the teaching of all technical skills. In case clinical monitors or ventilators implement “automatic” measurement of Pes, we should avoid what is described with the reading and interpretation of “automatic” measurement of pulmonary arterial pressure (15).

In the late 1980s, Joseph Milic-Emili entitled an editorial: “Is weaning an art or a science?” (16). We respectfully allude to his title in this manuscript, knowing that, as far as weaning is concerned, for Pes in clinical practice there is still (as John Lennon and Paul McCartney would phrase it) a “long and winding road” ahead of us (17).