Background
Esophageal pressure (Pes), which is commonly measured by a catheter with an air-filled balloon, has been used to estimate pleural pressure [
1,
2]. Recently, this technique has regained attention to guide lung-protective ventilation in patients with acute respiratory distress syndrome [
3‐
7]. An accurate measurement of Pes depends on the proper filling of the balloon [
8‐
13]. Under in vitro conditions at atmospheric pressure, during progressive inflation of the balloon, the balloon pressure-volume (P-V) curve exhibits a nearly plain intermediate section, indicating a volume range with negligible balloon recoil pressure [
8‐
11]. Nevertheless, when the balloon is placed in the esophagus, inflation of the balloon yields an inclined linear intermediate section on the P-V curve [
8,
12,
13]. This phenomenon has been thought to be due to the response of the esophageal wall recoil to the balloon inflation, and the slope of the intermediate linear section is considered to be equal to the esophageal wall elastance (Ees) [
8,
12,
13]. Although release-derived and elastance-derived strategies have been proposed to compute relative transpulmonary pressure [
14], eliminating the influence of balloon surrounding structures on the absolute measurement of Pes may also be required [
3,
4,
7]. Based on the estimation of Ees and pressure generated by the esophageal wall during balloon inflation (Pew), Mojoli and coworkers introduced an in vivo calibration procedure to make the Pes measurement more reliable [
13]. However, to the best of our knowledge, no study has been performed to testify the certainty of this calibration method. Moreover, only one type of balloon with a relatively large geometric volume has been investigated [
13].
In the present study, we established a bench model to simulate different levels of balloon-surrounding elastance and performed balloon volume test using a small-volume-balloon catheter. The balloon-surrounding elastance was estimated by the slope of the intermediate linear section on the balloon P-V curve. We primarily aimed to examine the agreement between the estimated and measured elastance. The balloon volume test was also performed in passive patients under controlled ventilation. Balloon P-V curves obtained in vitro and (in vivo conditions were compared. Based on previously introduced methods [
12,
13], we developed a simplified procedure for the Ees estimation and Pes calibration. The secondary aims included the assessment of the agreement between the standard and simple methods, and the comparison of calibrated Pes values among different filling volumes.
Discussion
For the small-volume-balloon catheter investigated in the present study, we found that: 1) the slope of the intermediate linear section on the balloon P-V curve agreed with simulated balloon-surrounding elastance in our bench model; and 2) Ees estimated by the simple method agreed closely with the standard linear regression method.
It is well known that reliable Pes measurement depends on proper filling of the balloon [
9,
10]. However, even within the optimal balloon volume range, esophageal wall reaction to balloon filling may elevate the absolute value of Pes [
12,
13]. Milic-Emili et al. first proposed correction of raw Pes values by extrapolating to the zero balloon volume in healthy volunteers during spontaneous breathing [
12]. Mojoli and colleagues further introduced a Pes calibration procedure by estimating the Ees and Pew in passive patients under controlled ventilation [
13]. In the present bench experiment, glass chambers with different inner volumes were used to simulate different balloon-surrounding elastance (Fig.
1). The chamber elastance was also estimated using the slope of the intermediate linear section on the balloon P-V curve when the balloon was inflated in the chambers (Fig.
2a). Agreement analysis showed that the slope of the intermediate section on the balloon P-V curve overestimated chamber elastance by a mean bias of 0.5 cmH
2O/ml (Fig.
3a), which approximated the slope of the linear section on the balloon P-V curve at atmospheric pressure (0.4 cmH
2O/ml). While no surrounding elastance existed when the balloon was tested at atmosphere, we reasoned that a slight inclination of the linear section on the balloon P-V curve might have resulted from the tendency of the balloon returning to its original shape. We also found that intermediate sections on in vitro and in vivo balloon P-V curves were nearly overlaid (Additional file
7: Figure S2), additionally supporting the rationality of the methodology for Ees estimation.
Previous in vitro experiments found that V
WORK, i.e., the balloon volume range between V
MIN and V
MAX, was narrower in small-volume-balloons [
9‐
11]. Among the tested esophageal catheter types, the geometric volume of the Cooper balloon was the smallest (2.8 ml). Thus, we questioned whether in vivo determination of Ees and calibration of Pes introduced by Mojoli et al. could be used in this type of balloon. Although our clinical V
WORK (median of 1.2 ml, ranging from 0.8 to 2.0 ml) was smaller than that in large-volume balloon (respective mean V
MIN and V
MAX of 1.5 and 5.4 ml) [
13], an intermediate linear section on the end-expiratory P-V curve could be observed in each of the clinical balloon volume tests with at least five P-V data points for linear fitting (Fig.
2 and Additional file
7: Figure S2). This intermediate section is adequate for the calculation of Ees. In accordance with previously reported in vitro and in vivo studies [
10,
13], V
MIN directly correlated with the balloon-surrounding pressure in our bench experiment and with Pes in our clinical study. However, we did not find a positive relationship between V
BEST with balloon pressure in either in vitro or in vivo condition. Additionally, Mojoli et al. found that the V
BEST for large-volume-balloon displayed significant variability in different bench conditions [
10] and among patients [
13], while in the present study the V
BEST was quiet similar among patients (Table
1). These results might be explained by the low variation in balloon with small geometric volume.
Our clinical results suggested that inflating the small-volume balloon also induced esophageal wall reactions. In our group of patients, the median Ees was 3.3 (2.5–4.1) cmH
2O/ml, corresponding to a median Pew at V
BEST of 2.8 (2.5–3.5) cmH
2O with the maximum value of 5.2 cmH
2O. This level of Pew might be large enough to significantly overestimate the Pes and underestimate the transpulmonary pressure. Thus, we recommend that the balloon volume test and Pes calibration should be performed even when a small-volume balloon is used. Moreover, it could be noticed that in vivo V
BEST was markedly larger than in vitro value (Table
1), which might suggest the contact reaction of the balloon to the esophageal wall. However, this phenomenon requires further investigation.
Because the range of V
MIN to V
MAX was 0.6 to 1.4 ml in all clinical tests, we simplified the estimation of Ees by only using the esophageal balloon pressures measured at these two balloon volumes. The results obtained by the simple method agreed closely with those by the conventional linear regression method (Additional file
8: Figure S3). For the small-volume balloon used in the present study, all clinical V
BEST were also located between 0.6 and 1.4 ml, and no significant difference was found in either the ∆Pes/∆Paw ratio during the occlusion test or calibrated Pes among balloon volumes within this range (Table
3). Therefore, we further suggest a simple procedure for balloon volume test and Pes calibration for the Cooper balloon catheter. The balloon can only be inflated to three volumes of 0.6, 1.0 and 1.4 ml, and the volume with largest tidal swing in balloon pressure could be selected as the V
BEST. Ees can be simply estimated only using pressures at balloon volumes of 0.6 and 1.4 ml, and the raw Pes at the V
BEST can be calibrated. Whether this procedure is suitable for other types of balloons, especially for large volume balloons, needs further investigation.
The major strength of our study is the combined reporting of bench and clinical results. There are limitations in our study. First, although different levels of balloon-surrounding elastance were successfully produced in our bench experiment, we definitely recognized that our model did not simulate the real scenario within the esophagus because the esophageal contact reaction to the balloon inflation was omitted. This limitation can be illustrated in the merged in vitro and in vivo balloon P-V curves (Additional file
7: Figure S2). At higher balloon volumes, in vivo curves elevated earlier and more markedly, which might be due to the esophageal contact reaction. However, our simulated levels of balloon-surrounding elastance and pressure covered clinical reported ranges of Ees and Pes [
3,
4,
6,
7,
13,
21,
23]. When combining the results in the bench experiment and the clinical study, we reasonably thought that this simplified bench model was enough to demonstrate the relationship between the slope of the intermediate balloon P-V curve and the Ees. Second, we only investigated one specific type of balloon with a relatively small geometric volume. Third, only postoperative patients with normal oxygenation and airway driving pressure were studied. Fourth, balloon volume test was only performed in passive patients in controlled ventilation. Therefore, our findings cannot be directly generalized to other balloon catheters and other patient populations.