In vivo calibration of esophageal pressure in the mechanically ventilated patient makes measurements reliable

Esophageal manometry has been proposed as a clinical surrogate for pleural pressure direct measurement, to guide mechanical ventilation in acute respiratory failure [1, 2]. Esophageal pressure (Pes) is currently measured by an esophageal catheter with an air-filled balloon placed in the mid-lower portion of the esophagus. The prerequisite for using this technique is that the pressure surrounding the esophagus, i.e., the intrathoracic pressure, is correctly transmitted to the balloon. Several factors may affect the results [3, 4]. Large-size balloons, only partially inflated, are used to avoid any additional pressure due to balloon wall stretching [5, 6]. Nonetheless, due to the elastance of the esophagus wall, the filling volume was found to affect measured Pes even when the balloon is only partially inflated [5, 7]. Milic-Emili and coworkers [6] proposed two strategies to eliminate this artifact: using very low filling volumes (≤0.5–1.0 ml) or correcting Pes values by extrapolation to zero balloon volume. The first option, simpler, was thereafter adopted by researchers and clinicians, and discussed no further. Only recently, an in vitro study clearly demonstrated that, under positive-pressure conditions like during mechanical ventilation, very low filling volumes may be insufficient to accurately transmit both absolute values and tidal swings of Pes [8]. On the other hand, higher than traditional filling volumes result in disproportionately high Pes values, as recently reported by Chiumello and coworkers [9]. These findings support the concept that only respiratory swings of Pes are reliable, and not absolute values, thus firing up the controversy on alternative methods to compute transpulmonary pressure [10].

With the present study, we propose a new Pes calibration procedure designed to solve both the issue of low Pes transmission due to insufficient balloon filling, and the opposite issue of Pes overestimation due to significant pressure generated by the esophageal wall as a reaction to balloon filling. This procedure was tested on patients with acute respiratory failure (ARF), to evaluate its feasibility and possible advantages over traditional, not calibrated, approaches.

We enrolled sedated and paralyzed patients with ARF under pressure-controlled mechanical ventilation, in whom an esophageal balloon catheter (Nutrivent, Sidam, Mirandola, Italy) had been inserted in the mid-lower third of the thoracic esophagus for clinical purposes. The esophageal balloon has a length of 10 cm, a nominal volume of 10 ml and a factory-recommended inflating volume of 4 ml of air. Mechanical characteristics of this catheter and other commercially available ones were previously studied in vitro [8]. Briefly, the devices were exposed to different surrounding pressures and progressively filled with air to obtain esophageal balloon pressure-volume curves. The appropriate range of filling volumes was defined by a null balloon transmural pressure and corresponded, for all the devices, to the intermediate linear section of the curve. The appropriate range was found to be catheter-specific and related to the surrounding pressure, being 0.5–8 ml for the NutriVent catheter when the surrounding pressure was in the 0–30 cmH₂O range. Appropriate catheter position was confirmed by visualization of cardiac artifacts on Pes and radiopaque markers on chest X-ray.

In each patient, we recorded the static Pes at end-expiration (Pes_EE) and end-inspiration (Pes_EI) while the esophageal balloon was inflated with increasing volumes from 0 to 8 ml, by 0.5-ml steps from 0 to 3 ml and 1-ml steps from 3 to 8 ml. At each volume step, the balloon was completely deflated by applying a negative pressure, then fully inflated with 10 ml of air and finally deflated to the target volume. Static pressures were obtained by airway occlusion maneuvers of 5 s. From those data, we obtained two curves expressing the individual pressure-volume (PV) relationship between balloon filling volume and Pes, respectively at end-expiration and end-inspiration (Fig. 1).

On the end-expiratory PV curve we graphically identified the intermediate linear section and its lower and upper limits, expressed as minimum and maximum filling volumes (V_MIN and V_MAX). According to in vitro results [8], V_MIN is the smaller filling volume to be injected into the esophageal catheter to pressurize it at the same level of the pressure surrounding the esophageal balloon, i.e., to reach a balloon transmural pressure of zero. Further volume injection induces progressive inflation of the esophageal balloon, V_MAX being the larger filling volume that does not induce overstretch of the balloon wall [8]. Therefore, the range between V_MIN and V_MAX was considered to correspond to appropriate balloon filling, with volumes below V_MIN denoting underfilling and volumes above V_MAX denoting overfilling of the balloon. Within the appropriate volume range, we identified the volume providing the maximum difference between Pes_EI and Pes_EE (V_BEST).

The slope of the intermediate linear section of the end-expiratory PV curve was obtained by least square fitting and it was considered to express the elastance of the esophagus (Ees) [7]. For any filling volume V_X above V_MIN, the pressure generated by the esophageal wall (Pew) was calculated as: Pew = (V_X – V_MIN) * Ees (Figure S1 in Additional file 1). Pew was considered null at filling volumes lower than V_MIN, the balloon transmural pressure being negative and therefore the esophagus reaction to balloon filling negligible in this case.

Then we compared five different balloon filling strategies: 0.5 ml (V_0.5), as per traditional recommendations; 4.0 ml (V_4.0), as per manufacturer’s recommendations; 8.0 ml (V_8.0), i.e., approximately at full balloon inflation; filling volume equal to individual V_MIN; and filling volume equal to individual V_BEST. At each of these filling volumes, we measured again the static Pes_EE and Pes_EI and performed a validation occlusion test [3, 11]. Our patients being sedated and paralyzed, we recorded simultaneous changes of Pes (ΔPes) and airway pressure (Paw) (ΔPaw) while applying gentle compressions on the patient’s chest during an end-expiratory occlusion maneuver (Fig. 2). The hypothesis that V_BEST was associated with the ΔPes/ΔPaw ratio closest to 1 among the different balloon filling strategies was evaluated by repeated measures ANOVA and Cochran’s Q test.

From the Pes measurements obtained at a filling volume equal to individual V_BEST and the corresponding Pew, we computed the calibrated values of Pes (Pes_CAL), as previously suggested [6, 7]: Pes_CAL = Pes – Pew. The hypothesis that Pes_CAL significantly differed from raw measurements of Pes at V_0.5 (Pes_V0.5), V_4.0 (Pes_V4.0) and V_BEST (Pes_VBEST) was evaluated by Bland-Altman analysis and repeated measures ANOVA. The study was approved by the committee on research ethics at the institution in which the enrollment was conducted (Fondazione IRCCS Policlinico S. Matteo, Pavia, Italy) and informed consent was obtained as required.

Fifty series of measurements were performed in 36 patients (20 males, 57 ± 17 years) undergoing pressure controlled mechanical ventilation. In nine patients, multiple measurements at different body positions and/or positive end-expiratory pressure (PEEP) settings were performed: six patients were studied twice, one patient three times and two patients four times. Body position was supine with the bed at 20–30 degrees head up, lateral and prone in 45, 3 and 2 cases respectively. Settings of pressure-controlled mechanical ventilation were: inspiratory oxygen fraction (FiO₂) 0.71 ± 0.22, PEEP 11.8 ± 5.0 cmH₂O, peak pressure 28.6 ± 5.5 cmH₂O, respiratory rate 15.7 ± 5.2 bpm. Tidal volume/ideal body weight was 7.9 ± 1.8 ml/kg, plateau pressure 27.6 ± 5.0 cmH₂O and partial pressure of oxygen in arterial blood (PaO₂)/FiO₂ ratio 165 ± 76 mmHg. Causes of acute respiratory failure were: community-acquired pneumonia (7), heart failure (2), chronic obstructive pulmonary disease (COPD) decompensation (1), acute respiratory distress syndrome (ARDS) (7), abdominal compartment syndrome (6), pulmonary alveolar proteinosis (8), and postoperative respiratory complications (5). Pulmonary alveolar proteinosis patients were studied during the whole lung lavage procedure previously described [12]. In five patients, body mass index was higher than 50 kg/m².

We evaluated esophageal manometry in patients with ARF under positive-pressure mechanical ventilation. Our main findings are:

Esophageal pressure is used as surrogate for pleural pressure (Ppl) in order to compute transpulmonary (airways minus pleural) pressure (P_L), i.e., the effective pressure distending the lung parenchyma. Ideally, in mechanically ventilated patients we should maintain positive values of P_L at end-expiration while avoiding excessive P_L at end-inspiration, to limit both lung derecruitment and overdistention. These theoretical assumptions were recently translated into two different practical approaches to set PEEP, with potential advantages in ARF [1, 2]. The two approaches are based on different methods to compute transpulmonary pressure: Talmor and colleagues assumed the absolute value of esophageal pressure to be equal to pleural pressure (direct method), whereas Grasso and colleagues computed pleural pressure by partitioning airway pressure according to respiratory swings of esophageal pressure (elastance-derived method).

It was recently demonstrated that, under positive-pressure conditions, the use of traditional small filling volumes is associated with underfilling and thus malfunctioning of the esophageal balloon catheter in most cases [8]. Underfilling involves under-transmission of both absolute values and respiratory swings of esophageal pressure, and therefore may affect both the direct and the elastance-derived method.

In this study in patients with ARF under controlled mechanical ventilation, the minimum filling volume (V_MIN) that was able to accurately transmit the end-expiratory esophageal pressure was substantially larger than 0.5 ml in most cases, being on average 1.5 ml and reaching values as high as 3 ml. As previously observed in vitro [8], the in vivo V_MIN was proportional to the surrounding pressure: the higher the Pes, the larger the volume to be injected into the catheter. Moreover, the filling volume providing optimal transmission of Pes respiratory swings (V_BEST) was substantially larger than V_MIN, being on average 3.5 ml and directly related to the Pes value. The positive relationship between V_BEST and Pes is explained by the fact that part of the volume injected into the esophageal catheter must be spent to pressurize the system at the pressure level surrounding the balloon. The dynamic occlusion test is the classic method to validate a Pes measurement [3, 11]: similar Pes and Paw swings should be observed during spontaneous respiratory efforts (or manual chest compressions in the case of a passive patient) against the closed airway opening. A major finding of our study is that in almost all cases the balloon filling volume required to pass the validation occlusion test was much larger than 0.5 ml. Therefore we suggest abandoning the systematic use of the “traditional” small filling volumes. A simple option could be to use a fixed higher filling volume. However, V_BEST was highly variable among different patients and conditions, ranging from 0.5 to 6 ml in our series. This suggests that the esophageal catheter filling volume should be adapted to the intrathoracic pressure condition of any given patient. Therefore, in order to obtain reliable esophageal pressure waveforms, we suggest filling the esophageal balloon catheter progressively step-by-step and selecting, within the catheter-specific range of filling volumes [8], the volume associated with the largest esophageal pressure tidal swing. Since dynamic and static tidal variation of Pes are very similar [13], for clinical purposes the procedure can be performed without occlusion maneuvers at each filling step. Many factors, like mechanical ventilation setting, body position, fluid balance and intra-abdominal pressure, can influence intrathoracic pressure in critically ill patients. At any significant change of one or more of these factors, for better results the optimal filling volume of the esophageal balloon catheter should be rechecked.

The use of larger filling volumes stimulated significant esophageal reaction in our patients. The pressure generated by the esophageal wall, when progressively distended, has both a passive (collagen and elastic fibers) and an active (smooth muscle tone) component, as well described by Orvar and colleagues [14]. When a balloon was progressively inflated in the esophagus of awake healthy volunteers, the pressure generated by the esophageal wall (Pew) was linearly related to the esophageal cross-sectional area (CSA): for each 10 mm² increase in CSA above 50 mm², Pew increased by 1 cmH₂O. Thresholds for peristaltic contractions and for pain or pressure sensation were 150 and 300 mm² respectively. The esophageal catheter we used is a nasogastric tube (20 mm² CSA) provided with a 10-cm-long esophageal balloon that reaches 180 mm² CSA when fully inflated with approximately 10 ml. Therefore, a detectable Pew was expected at intermediate balloon filling volumes (but not in case of deflated or minimally inflated balloon), whereas peristaltic contractions were expected only for higher filling volumes, approximating full balloon inflation. On average we observed a Pew of about 1 cmH₂O for each milliliter of injected volume above V_MIN. When V_BEST was injected into the catheter, Pew was 2 cmH₂O on average and reached values as high as 6 cmH₂O. With a balloon near fully inflated, Pew was about 8 cmH₂O on average with single values as high as 20 cmH₂O. This esophageal reaction to balloon inflation is a well-known artifact affecting absolute values of esophageal pressure, which may therefore become substantially more positive than pleural pressure [5, 7]. Such an artifact can result in significant and unpredictable overestimation of Pes and underestimation of P_L, possibly leading to inappropriately high PEEP levels and/or end-inspiratory lung overdistention when a Pes-guided mechanical ventilation strategy is adopted. Our data suggest that the risk overestimation of Pes is higher in case of female subjects and/or moderate-low body mass indexes. These findings are consistent with gender differences in the responses to esophageal balloon distention [15] and high prevalence of hypomotility of the esophageal body in obese patients [16]. Milic Emili and colleagues [6] suggested using very small filling volumes for the esophageal balloon in order to easily minimize the esophageal artifact and their recommendation survived for about 50 years, being only recently rediscussed [3, 8]. Our findings prove that, in mechanically ventilated ARF patients, it is not possible to avoid balloon underfilling and esophageal artifact at the same time. This technical dilemma may explain why different research groups obtained low [17], moderate [18, 19] or high [9] absolute Pes values in the same condition of supine subjects at functional residual capacity.

To solve this technical problem, we reconsidered and adapted a procedure originally proposed by Milic-Emili who suggested – as an alternative to small filling volumes – to extrapolate to zero balloon volume the Pes measured at higher filling volumes [6]. Our calibration procedure significantly improved the quality of Pes measurement. Tidal swings of Pes were maximized, thus allowing passing of the validation test in almost all cases. Moreover, it was possible to recognize and subtract from Pes the pressure generated by the esophageal wall, which would have significantly raised Pes above the pleural value. In our opinion, calibrated Pes represents the best surrogate for absolute pleural pressure in critical care patients and may significantly improve the direct method proposed by Talmor and colleagues [1].

The present study demonstrates that an approach based on a calibrated esophageal pressure is feasible and provides significant improvement of the technique over traditional approaches. Clinical advantages of using a calibrated esophageal pressure to guide mechanical ventilation should be demonstrated by future trials.