PEEP guided by esophageal pressure--any added value?

Treatment of the acute respiratory distress syndrome (ARDS) with positive end-expiratory pressure (PEEP) was introduced more than 40 years ago.1,2 Few practitioners question the effects of PEEP on hemoglobin saturation, but the optimal dose titration is controversial. Usually PEEP is simply adjusted to a suitable increase in oxygenation efficiency (i.e., the ratio of the partial pressure of arterial oxygen [PaO2] to the fraction of inspired oxygen [FiO2]) or according to an algorithm such as the Acute Respiratory Distress Syndrome Network (ARDSNet) protocol.3 Because PEEP can impair cardiac output, some clinicians advocate titration to maximal oxygen delivery, a method that requires measurement of cardiac output. More complex techniques such as titration with the use of the pressure–volume relationship in the respiratory system have been reported.4 In this issue of the Journal, Talmor and colleagues5 describe a randomized trial of a mechanical-ventilation strategy in which PEEP was adjusted according to end-expiratory transpulmonary pressures (ClinicalTrials.gov number, NCT00127491). Transpulmonary pressure was measured as the difference between the airway opening pressure and the pleural pressure; pleural pressure was estimated from esophageal pressure. A small catheter with a balloon covering the holes at its end was used to measure pneumatic pressure in the esophagus. PEEP was then adjusted to produce an estimated transpulmonary pressure at end expiration of 0 to 10 cm of water, according to PaO2:FiO2. Originally designed to enroll 200 subjects, the trial was stopped by the safety board after only 61 subjects were enrolled. Termination was based on an a priori stopping rule requiring a difference in PaO2:FiO2 between groups that was of a critical significance level (P<0.02). Adjusting PEEP on the basis of transpulmonary pressure is a reasonable physiological premise, since this pressure is derived by calculating the difference between airway pressure and the highly variable pleural pressure. Pleural pressure can be unpredictable in cases of critical illness due to pleural effusion, elevated abdominal pressure, and variations in the elastance of the chest wall. For example, a large pleural effusion may raise pleural pressure, thereby reducing transpulmonary pressure at any given PEEP. However, accurate measurements of pleural pressure in the intensive care unit (ICU) are rarely possible. Because of this drawback, pleural pressure can be estimated only from esophageal pressure. The methods and limitations of estimating esophageal pressure have been published previously.6,7 In brief, many assumptions must be made in order to accept that pressure in the esophagus dynamically and accurately reflects pleural pressure. For instance, we must assume that the balloon pressure reflects the esophageal pressure, that the transmural pressure in the esophagus is 0 cm of water, that the esophagus is not compressed by intrathoracic structures such as the heart, that the pressures in the periesophageal area are the same as the pleural pressure, and that pleural pressure is relatively uniform throughout the thorax. In fully one third of the patients in the study by Talmor et al., the balloon placement was inadequate, according to monitoring pressure recordings, and alternative placement techniques were used; this requirement calls into question the consistency of positioning and thus the reliability and reproducibility of the measurements. In addition, a correction factor of 5 cm of water was subtracted from the esophageal pressure in an attempt to compensate for the known artifacts of mediasti-