Use of esophageal balloon pressure-volume curve analysis to determine esophageal wall elastance and calibrate raw esophageal pressure: a bench experiment and clinical study

A commercially available esophageal balloon catheter (Cooper: LOT 177405, Cooper Surgical, USA) was used in the present study [9‐11]. During the study, pressure waveforms were measured and recorded by a dedicated data acquisition system [11].

Five glass chambers with an inner volume of 1000, 500, 250, 175 and 125 ml were used to obtain five levels of chamber elastance approximately from 1 to 8 cmH₂O/ml (Fig. 1a) [9, 15]. In each chamber, six levels of baseline chamber pressure (5, 10, 15, 20, 25 and 30 cmH₂O) were also established by injecting different amounts of air into the chamber. Thus, 30 balloon-surrounding conditions with five levels of chamber elastance and six levels of baseline chamber pressure were simulated (Fig. 1b).

Five Cooper catheters were selected and randomly introduced into the five different chambers. Before each balloon volume test, the residual volume of the balloon was standardized as the method described in previous studies [16‐18]. Briefly, the chamber pressure was adjusted to 5 cmH₂O by air injection and the balloon lumen was opened to the atmosphere for 3 min. We arbitrarily defined the balloon volume under this condition as the zero filling volume. Then the balloon was intermittently inflated in 0.2-ml increments up to 2.4 ml. A balloon volume test was first performed at atmospheric pressure and then in each chamber with certain elastance at different levels of baseline chamber pressure (Fig. 1b).

Balloon pressure and chamber pressure were simultaneously measured. Balloon transmural pressure was defined as the difference between the balloon pressure and the chamber pressure (balloon pressure - chamber pressure) [9, 10]. The balloon volume with transmural pressure within ±1.0 cmH₂O was defined as the minimal and maximal balloon volumes (V_MIN and V_MAX), which represented the optimal filling volume with the least influence of balloon recoil pressure [9, 10]. Balloon working volume (V_WORK) was calculated as the difference between V_MIN and V_MAX. The balloon volume with the closest to zero transmural pressure was defined as the best filling volume (V_BEST).

Balloon pressure and chamber pressure were plotted against the balloon volume (Fig. 2a). The chamber pressure increased linearly as a function of balloon volume, and the slope obtained by least square fitting was defined as the measured chamber elastance. The balloon P-V curve exhibited a sigmoid shape with an intermediate linear section corresponding to the V_MIN to V_MAX range. We used the slope of this linear section to estimate chamber elastance. Baseline chamber pressure was estimated as the difference between the measured balloon pressure at V_BEST and the product of estimated chamber elastance and V_BEST.

The clinical study was conducted in the intensive care unit, Beijing Tiantan Hospital, Capital Medical University, Beijing, China. The study protocol was reviewed and approved by the local Institutional Review Board (KY-2016-11-22) and the study was registered at ClinicalTrials.​gov (NCT02976844). Written informed consent was obtained from the patient or appropriate substitute decision makers.

We enrolled postoperative patients with delayed emergence from general anesthesia admitted to the unit for mechanical ventilation. For patients after elective intracranial operations, the reasons for delayed extubation and mechanical ventilation mainly included large brain stem tumor resection, length of operation longer than six hours, and major intra-operative bleeding or brain swelling [19]. Mechanical ventilation was also continued in severe traumatic brain-injured patients after emergent operations. For patients after orthopedic and vascular surgery, delayed extubation was performed because of long duration of operation or major bleeding. Prolonged mechanical ventilation was usually required in patients with brain stem lesions and severe traumatic brain injury. In our unit, Pes monitoring was usually used in mechanically ventilated patients to guide ventilator settings and weaning. Exclusion criteria were as follows: 1) age under 18 years; 2) contraindications for esophageal balloon catheter insertion, including evidence of severe coagulopathy, diagnosed or suspected esophageal varices, and history of esophageal, gastric or lung surgery; and 3) evidence of active air leak from the lung, including bronchopleural fistula, pneumothorax, pneumomediastinum, and an existing chest tube. During the study, most of the patients did not recover from anesthesia and neuromuscular paralysis. In the case of recovered spontaneous breathing, continuous intravenous infusion of midazolam 0.05–0.2 mg/kg/h and fentanyl 0.1 mg/h were given, and an intravenous bolus of vecuronium 0.1 mg/kg was used as needed. The absence of spontaneous inspiratory effort was confirmed by the absence of a negative airway pressure (Paw) swing during a 3-s end-expiratory occlusion. The patients were ventilated under a volume-controlled mode with constant flow, set as the tidal volume of 6–8 ml/kg predicted body weight and clinical positive end-expiratory pressure and fraction of inspired oxygen. Hypoxemia was documented as the ratio of partial pressure of arterial oxygen to fraction of inspired oxygen less than 300. Acute respiratory distress syndrome (ARDS) was diagnosed according to the Berlin definition [20]. The absence of the chest X-ray criterion was defined as acute hypoxemic respiratory failure (AHRF).

The Cooper balloon was placed in the lower two thirds of the esophagus, which was confirmed by cardiac artifacts on Pes tracing and bedside X-ray examination.

During the study, the patients were remained in supine position with the head of the bed elevated to 30°. Before each balloon volume test, the balloon was inflated to 3.0 ml and then was deflated by generating a negative pressure followed by opening to the atmosphere for 3 min. The balloon was intermittently inflated in 0.2-ml increments up to 2.4 ml. At each tested volume, after 3-min equilibration, the airway was occluded at end-expiration and end-inspiration, each for 3 s. Positive pressure occlusion test was performed at end-expiratory occlusion, and the ratio of changes in Pes to Paw (∆Pes/∆Paw) during the compression of the chest wall was calculated [21].

Esophageal balloon pressure was plotted against the balloon volume (Fig. 2b). The method introduced by Mojoli et al. was used to determine the Ees and optimal balloon volume [13]. An intermediate linear section was visually inspected on end-expiratory balloon P-V curves, and the lower and upper limits were defined as the clinical V_MIN and V_MAX. The range between V_MIN and V_MAX was defined as V_WORK and considered as the optimal balloon filling volume for clinical Pes measurement. The clinical V_BEST was defined as the balloon volume with the largest difference between end-expiratory and end-inspiratory Pes. The slope of the intermediate linear section on the end-expiratory balloon P-V curve was defined as the Ees [13]. Esophageal wall recoil pressure reacting to balloon filling; i.e., Pew, was estimated by the product of Ees and the balloon volume within V_MIN to V_MAX, using the method previously introduced by Milic-Emili et al. [12]. The raw Pes values were calibrated by extrapolating to the zero balloon volume, which could also be expressed as: calibrated Pes = raw Pes - Pew.

In each of the patient’s balloon volume test, intermediate linear section on the end-expiratory balloon P-V curve (i.e. range of clinical V_MIN to V_MAX) enclosed filling volume of 0.6 to 1.4 ml. Therefore, we simplified the estimation of Ees only using parameters at these two filling volumes, as the difference of end-expiratory balloon pressure between 0.6 ml and 1.4 ml divided by 0.8 ml (1.4–0.6 ml).

Categorical variables are reported as numbers and percentages. Continuous data are presented as the median and interquartile range (IQR) and were compared using a Kruskal-Wallis test with post hoc comparison by Bonferroni’s correction.

Correlation of optimal balloon volume with balloon pressure was analyzed. Spearman’s correlation coefficient (rho) was calculated.

Bland-Altman’s analysis was used to examine the agreement for chamber elastance and baseline chamber pressure between the estimated and the measured value in the bench experiment, and for Ees between values calculated by the standard and the simple method in the clinical study. Bias and standard deviation (SD) of the mean bias were calculated. Upper and lower limits of agreement were defined as bias ±1.96 SD of the mean bias. The sample size in the bench experiment was based on the conditions’ setup (n = 30). The sample size in the clinical study (n = 40) was chosen on the basis of previous studies [13, 21]. The respective sample size gave a 95% confidence interval of the agreement of limit as ±0.32 × SD and ±0.27 × SD [22].

Detailed experimental results are provided in Additional file 2. Example balloon P-V curves at atmospheric pressure and different levels of chamber elastance are shown in Additional file 3: Figure S1. The slope of the intermediate linear section on the balloon P-V curve at atmospheric pressure was 0.4 (0.3–0.4) cmH₂O/ml. In accordance with previous investigations, gas-tight chambers with different inner volumes produced different levels of chamber elastance (Additional file 4: Table S1).

Bias (lower and upper limits of agreement) was 0.5 (0.2 to 0.8) cmH₂O/ml between the slope of the linear section on the balloon P-V curve (estimated chamber elastance) and measured chamber elastance, and − 0.6 (− 1.2 to 0.0) cmH₂O between the estimated and measured baseline chamber pressure (Fig. 3).

Optimal balloon volume parameters in the bench experiment are shown in Table 1. Bench V_MIN and V_MAX correlated linearly with balloon pressure (rho = 0.715 and 0.674, p = 0.010), but no correlation was found between V_BEST and balloon pressure (rho = 0.284, p = 0.128).

Forty patients were enrolled. Baseline characteristics are shown in Table 2. Hypoxemia was presented in 13 (32.5%) patients, with ARDS and AHRF diagnosis in 4 (10%) and 9 (22.5%) patients, respectively. Detailed results in the clinical study are presented in Additional file 5.

An intermediate linear section was visually identified on the end-expiratory esophageal balloon P-V curve in each of the clinical tests. Optimal balloon volume parameters obtained in the clinical study are also shown in Table 1 and Additional file 6: Table S2. V_MIN and V_MAX correlated linearly with either end-expiratory or end-inspiratory esophageal balloon pressures (rho = 0.320 to 0.352, p = 0.026 to 0.044). V_BEST was 1.0 (0.8–1.2) ml, ranging from 0.6 to 1.4 ml. No significant correlation was found between the V_BEST and esophageal balloon pressure (rho = 0.096 and 0.176, p = 0.557 and 0.279). Ees was 3.3 (2.5–4.1) cmH₂O/ml, ranging from 1.2 to 5.8 cmH₂O/ml. Estimated Pew at V_BEST was 2.8 (2.5–3.5) cmH₂O, ranging from 1.4 to 5.2 cmH₂O.

In vivo and in vitro curves closely resembled each other in the intermediate linear section, but in most cases, in vivo P-V curve elevated earlier and more significantly than the paired in vitro curve. Examples of in vivo esophageal balloon P-V curves during end-expiratory occlusion in combination with in vitro curves in the chambers with comparable elastance and baseline pressure are shown in Additional file 7: Figure S2.

In each of the patient’s tests, the V_MIN to V_MAX range enclosed 0.6 to 1.4 ml, and V_BEST was also located within 0.6 to 1.4 ml (Table 1). ∆Pes/∆Paw ratios during occlusion test were 0.725 (0.251–0.848) at balloon volumes below 0.6 ml, with 35.0% (42/120) within 0.8 to 1.2. Ratios at balloon volumes above 1.4 ml were 0.884 (0.734–1.041), and 59.4% (98/165) was within 0.8 to 1.2. At balloon volumes between 0.6 and 1.4 ml, there were 92.5% (185/200) ∆Pes/∆Paw ratios within 0.8 to 1.2, and no significant difference was found among different balloon volumes within this range (p = 0.067, Table 3). There was also no significant difference in the calibrated Pes (p = 0.997) among balloon volumes within 0.6 to 1.4 ml (Table 3).

Compared with the standard linear regression method, bias (lower and upper limits of agreement) of Ees estimated by the simple method was − 0.1 (− 0.7 to 0.6) cmH₂O/ml (Additional file 8: Figure S3).