Evaluation of Optimal Esophageal Catheter Balloon Inflation Volume in Mechanically Ventilated Children

Abstract

BACKGROUND:

Accuracy of esophageal pressure measured by an air-filled esophageal balloon catheter is dependent on balloon filling volume. However, this has been understudied in mechanically ventilated children. We sought to study the optimal filling volume in children receiving ventilation by using previously reported calibration methods. Secondary objectives included to examine the difference in pressure measurements at individualized optimal filling volume versus a standardized inflation volume and to study if a static hold during calibration is required to identify the optimal filling volume.

METHODS:

An incremental inflation calibration procedure was performed in children receiving ventilation, <18 y, instrumented with commercially available catheters (6 or 8 French) who were not breathing spontaneously. The balloon was manually inflated by 0.2 to 1.6 mL (6 French) or 2.6 mL (8 French). Esophageal pressure (P_es) and airway pressure tracings were recorded during the procedure. Data were analyzed offline by using 2 methods: visual determination of filling range with the calculation of the highest difference between expiratory and inspiratory P_es and determination of a correctly filled balloon by calculating the esophageal elastance.

RESULTS:

We enrolled 40 subjects with median (interquartile range [IQR]) age 6.8 (2–25) months. The optimal filling volume ranged from 0.2 to 1.2 mL (median [IQR] 0.6 [0.2–1.0] mL) in the subjects with a 6 French catheter and 0.2–2.0 mL (median [IQR] 0.7 [0.5−1.2] mL) for 8 French catheters. Inflating the balloon with 0.6 mL (median computed from the whole cohort) gave an absolute difference in transpulmonary pressure that ranged from −4 to 7 cm H₂O compared with the personalized volume. P_es calculated over 5 consecutives breaths differed with a maximum of 1 cm H₂O compared to P_es calculated during a single inspiratory hold. The esophageal elastance was correlated with weight, age, and sex.

CONCLUSIONS:

The optimal balloon inflation volume was highly variable, which indicated the need for an individual calibration procedure. P_es was not overestimated when an inspiratory hold was not applied.

Background

The understanding of protective measures to decrease the risk of ventilator-induced lung injury has evolved extensively over the past two decades.¹ This includes, among others, limiting plateau pressure to reduce lung stress and strain.² Lung stress is defined as the pressure developed within the lung structures onto which the distending force is applied and is poorly reflected by airway pressures (P_aw).^3,4 The transpulmonary pressure (P_tp) represents the stress applied to the lung parenchyma and is computed by the difference between P_aw and pleural pressure. Measuring transpulmonary pressure may aid in a more tailored selection of tidal volumes and pressures to keep the lung open as part of a lung-protective ventilation strategy.

At the bedside, pleural pressure cannot be measured. Alternatively, esophageal pressure (P_es) manometry allows estimating pleural pressure to compute the end-inspiratory and end-expiratory transpulmonary pressure.^5,6 Furthermore, patient effort can be quantified by using esophageal manometry by calculating work of breathing according to the Campbell diagram or its surrogates, pressure-time and pressure-rate product. However, esophageal pressure manometry is mostly limited to research settings. This is because many factors influence P_es measurements, including body position, intrathoracic pressure, lung volume, heart weight, and the esophageal elastance, which makes interpretation at the bedside challenging.⁷ Esophageal balloon elastance is also of critical importance because inadequate balloon inflation will lead to erroneous P_es readings. Both under- and overinflation of the balloon returns under- or overestimated P_es values.^8,9 This also holds true for pediatric patients. It was observed in a bench test that tested commercially available catheters inserted into a simulated pediatric esophagus that P_es was overestimated when inflation volumes within the manufacturers’ recommended ranges were used.¹⁰ This signals the need for individualized titration of the balloon volume to obtain accurate measurements. Mojoli et al¹¹ proposed an in vivo calibration technique that included recording the static P_es at end-expiration and at end-inspiration with the esophageal balloon being stepwise inflated from 0 to 8 mL in adults. This allowed mapping of the individual end-expiratory and end-inspiratory pressure-volume relationship between balloon filling volume and P_es. The optimal filling volume was then identified by the maximum difference between P_es at end-inspiration and P_es at end-expiration in the mid-linear section.

To test the hypothesis that the proposed calibration method is also feasible in children receiving mechanical ventilation, we evaluated this calibration technique in a cohort of children who were critically ill and examined if an inspiratory and expiratory hold as part of this calibration technique was required to identify the optimal filling volume in children. We also wanted to examine the difference between the optimal filling volume identified with the calibration technique and a standardized inflation volume based on the aggregate measurements of the whole cohort (to reflect the situation in which a pre-set volume is delivered by a ventilator). During our study, a different calibration technique was proposed that made use of an alternative approach to determining over- and underinflation and the optimal filling volume on the pressure-volume loop.¹⁰ We modified our study protocol accordingly so that we could also evaluate feasibility of this alternative approach.

Methods

Study Design and Setting

This study was performed between March 2018 and May 2020 in the medical-surgical pediatric ICU of the Beatrix Children’s Hospital/University Medical Center Groningen. Included were subjects receiving invasive mechanical ventilation <18 y who were not breathing spontaneously (achieved through deep sedation and/or infusion of continuous neuromuscular blocking agents) and were instrumented with a 6 or 8 French SmartCath esophageal catheter (Vyaire, Mettawa, Illinois). The catheter size was based on weight (6 French, < 15 kg; 8 French, >15 kg). Unit policy is to use esophageal catheters as standard of care in patients with an expected duration of ventilation of >24 h. The correct position of the catheter was verified by the presence of cardiac oscillations on the P_es – time scalar. Chest radiographs were only performed when clinically indicated and not solely to confirm the correct position of the esophageal catheter. Subjects were excluded if malposition of the esophageal balloon was suspected based on an abnormal P_es – time scalar readings (either before the actual measurements or during offline visual inspection before data analysis) or after review of the chest radiograph or when spontaneous breathing efforts were registered during an expiratory hold of 3 s. The institutional review board of the University Medical Center Groningen (2017/307; date July 10, 2017) approved the study and waived the need for informed consent.

Data Acquisition and Study Protocol

The subjects were studied while they were in the supine position. Ventilator settings were managed by the medical team and not changed for this study. For data acquisition, the esophageal catheter was connected to a BiCore-II pressure transducer (Vyaire). P_es was sampled at 200 Hz. The analog signal of the mechanical ventilator (AVEA, Vyaire) was connected to a computer through an analog-to-digital converter to obtain P_es. A customized P_es cable with a 3-way valve was used to manually inflate the Pes and bypass the inflation procedure of the mechanical ventilator (Supplementary Fig. 1 [see the supplementary materials at http://www.rcjournal.com]). P_es was zeroed to the atmosphere before each measurement. Subsequently, we incrementally inflated the balloon with 0.2 mL up to 1.6 mL (6 French catheter) or 2.6 mL (8 French catheter); after each increment, we performed a 3-s inspiratory and expiratory hold. By following a protocol modification after 32 measurements, we first measured P_es and P_aw during an inspiratory and expiratory hold with the balloon fully deflated and then calibrated the pressure to the outside atmosphere for at least 15 s.

At the beginning of the data acquisition, we recorded subject-specific ventilator settings (PEEP, peak inspiratory pressure, mean P_aw, F_IO2, set breathing frequency) and ventilator data (expiratory tidal volume normalized to body weight). To characterize the cohort, we collected patient demographics from the patient record (age, weight, sex, reason for intubation, 24-h Pediatric Risk of Mortality Score III, Pediatric Incidence of Mortality II score, endotracheal tube size, and, if applicable, the correct position of the esophageal catheter on chest radiograph (defined as the catheter positioned in the lower third part of the esophagus) and clinical data, including the use of continuous neuromuscular blocking agents and the severity of oxygenation disorder classified by the oxygenation index (OI) or, if lacking, by the oxygen saturation index (none, OI < 4 or oxygen saturation index < 5; mild, 4 ≤ OI < 8 or 5 ≤ oxygen saturation index < 7.5; moderate, 8 ≤ OI < 16 or 7.5 ≤ oxygen saturation index < 12.3; severe, OI > 16 or oxygen saturation index ≥ 12.3).

Data Analysis

By using a custom-build MatLab script (MATLAB 2018a, The MathWorks, Natick, Massachusetts), cardiac oscillations and noise were filtered from the P_es signal by using a band-stop filter set at a heart rate of ±20 beats/min and a 5-Hz low-pass filter. The inspiratory P_aw was measured as the mean of 5 samples 0.2 s after the flow was < 0.5 L/min. The expiratory P_aw was defined as the mean over the last 200 samples (1 s) during the expiratory hold. The static P_es was calculated as the mean during 200 samples (1 s), 0.2 s after the flow was < 0.5 L/min, and the last second of the expiratory hold. The dynamic P_es was calculated from the mean of 5 consecutive breaths, in which the expiratory P_es per breath was calculated as a mean over the 40 samples just before inspiration and the inspiratory P_es over 5 samples just before the start of the expiration phase. The dynamic P_es was obtained on the unfiltered signal to simulate bedside measurement.

For the in vivo calibration technique constructed by Mojoli et al¹¹ (from here on referred to as calibration A), we first plotted the inspiratory and expiratory P_es against balloon inflation volume (Fig. 1). The plot was then visually inspected to identify the linearity, and the lower limit (V_min) and upper limit (V_max) of this linear part. The optimal filling volume was determined by the highest Δ P_es (static P_es expiratory – static P_es inspiratory) within V_min and V_max. The esophageal elastance was calculated by least squares fitting on the linear part (between V_min and V_max) of the expiratory P_es. Pressure of the esophageal wall at a certain filling volume (V_x) was calculated by the formula: pressure of the esophageal wall = (V_x − V_min) × esophageal elastance. The pressure of the esophageal wall was considered zero in volumes below V_min. End-expiratory transpulmonary pressure (Ptp_PEEP) and end-inspiratory transpulmonary pressure (Ptp_plat) values measured at the optimal filling volume measured on static conditions as part of method A were also compared with the median filling volume computed from the aggregate measurements from the whole cohort.

Fig. 1.

An example of a pressure-volume loop with the minimal, maximal, and optimal filling volume marked in a 6 French catheter according to method A. The minimum volume (V_min) and maximum volume (V_max) were visually determined. The optimal filling volume (Vbest) was determined as the biggest difference between the inspiratory and expiratory lines, in the linear section of the curve (between V_min and V_max). From Reference 11, with permission.

We modified our study protocol to test the in vivo calibration technique proposed by Hotz et al¹⁰ (method B) in the measurements that included a respiratory hold with a completely deflated balloon. In agreement with their technique, we calculated the elastance of the system. The elastance of the system was defined as the difference in P_es divided by the difference in volume: elastance of the system = (P_es at V_x − P_es at V_x+1)/(V_x − V_x+1). Volume steps were equal (0.2 mL), therefore, the difference in measured P_es at every volume step was used. Fifty percent of the highest elastance of the system values simulating overdistention were excluded. From the remaining values, the lowest volume was chosen as the best volume.¹⁰ See Table 1 for an overview of definitions used and calculations.

Table 1.

Overview of Definitions and Formulas

Statistical Analysis

For continuous data, descriptive statistics were used in which mean ± SD was used for normally distributed data, which was assessed by the Kolmogorov-Smirnov test, and median (interquartile range [IQR]) for non-normal distributed data. Paired continuous data were analyzed by using the Wilcoxon signed-rank test. Linear regression analysis was used to study the correlation among continuous variables. All statistical analyses were performed with SPSS 24 (IBM, Chicago, Illinois). P < .05 was considered as statistically significant.

Results

In total, 57 subjects were enrolled. Data of 17 subjects (29.8%) were excluded due to technical issues (n = 4) with suspected (n = 8) or confirmed (n = 5) esophageal catheter malposition. This left data from 40 subjects eligible for analysis, of whom 32 had a 6 French catheter. Characteristics of the study cohort are summarized in Table 2. For method A, we observed that, with the 6 French catheter V_min varied between 0.2 and 0.6 mL, and V_max 0.8 and 1.2 mL. The optimal filling volume varied between 0.2 and 1.2 mL, with a median (IQR) of 0.6 (0.2–1.0) mL. Among the subjects with an 8 French catheter, V_min varied between 0.2 and 0.8 mL, and V_max varied between 1.0 and 2.6 mL. The optimal filling volume varied between 0.2 and 2.0 mL with a median (IQR) of 0.7 (0.45–1.15) mL. For method B, the optimal filling volume ranged from 0.2 to 0.6 mL in the 6 French catheter (n = 18) and the median (IQR) was 0.4 (0.4–0.6) mL. In the 4 subjects with an 8 French catheter, the optimal filling volume ranged from 0.4 to 1.2 mL with a median of 0.7 mL.

Table 2.

Study Population Characteristics

We found that esophageal elastance was significantly higher in the subjects with a 6 French catheter (median [IQR] 7.97 [6.21–9.82] cm H₂O/mL) compared to those with an 8 French catheter (median [IQR] 2.75 [2.33–3.40] cm H₂O/mL) and varied between 0.82 and 13.9 cm H₂O/mL. Esophageal elastance was moderately inversely related to body weight (Pearson r = –0.55, P. < .001) and age (r = –0.57, P < .001) and significantly higher among the female subjects (median [IQR] 8.07 [6.43–10.33] cm H₂O/mL vs 5.79 [3.02–8.41] cm H₂O/mL) (P = .034). We did not observe significant differences in the median (IQR) Ptp_plat (10.69 [6.78–13.07] cm H₂O vs 10.53 [8.52–12.36] cm H₂O) or the median (IQR) Ptp_PEEP (–3.02 [−7.08 to –0.27] cm H₂O vs −3.25 [–4.15 to −1.31] cm H₂O) when the individualized optimal filling volume was compared with the median optimal filling volume (0.6 mL) from the aggregate measurements of the subjects with 6 French catheters (Fig. 2).

Fig. 2.

Differences in Ptp_PEEP and Ptp_plat obtained at a balloon volume of 0.6 mL and personalized optimal filling volume (Vbest). The data represents median (IQR).

Static versus Dynamic Measurements

For the 6 French catheter, P_es values were significantly different at 4 different balloon volumes when measured under zero-flow conditions compared with measurements under dynamic conditions (Fig. 3). The absolute difference in P_es varied between –0.31 and 0.63 cm H₂O. Similar findings were made for the 8 French catheter, with significantly higher P_es up to 0.5 cm H₂O under dynamic conditions. End-expiratory P_es was not significantly different when measured under static versus dynamic conditions (Supplementary Tables S1 and S2 [see the supplementary materials at http://www.rcjournal.com]).

Fig. 3.

Static holds versus dynamic measurement of esophageal pressure. * P < .05

Comparison of Calibration A with Calibration B

In 22 of the 40 subjects (55%), both methods were used to identify the optimal filling volume. We found that, in only 3 subjects, the same optimal filling volume was identified by both methods. Bland-Altman analyses showed a slight overestimation by method A (levels of agreement: –0.7 to 0.7). When using method B in vivo calibration techniques resulted in a higher optimal filling volume in 10 subjects. At optimal filling volume by method B, the inspiratory P_es differed from –4.2 to 3.5 cm H₂O from the inspiratory P_es at the optimal filling volume by method A. Similar observations were made for the expiratory P_es. When P_es was corrected for the pressure of the esophageal wall, the maximum difference was ∼1 cm H₂O.

Quick Look.

Current Knowledge

Esophageal manometry is a valuable tool to individualize mechanical ventilation. Accurate esophageal pressure readings mandate individual balloon filling volume titration. To date, however, this has been uncommonly studied in children.

What This Paper Contributes to Our Knowledge

Our study shows that esophageal pressure readings were inaccurate when a standard balloon inflation volume was used. An individualized titration maneuver is necessary; this can be performed without respiratory holds.

Discussion

To our knowledge, our clinical study was the first and largest to report the variability in optimal filling volume of commercially available esophageal balloon catheters in mechanically ventilated children who were not breathing spontaneously. We observed a wide range of inflation volumes in younger and older children. Inflating the balloon with a pre-set volume resulted in clinically relevant over- and underestimation of the transpulmonary pressure. The use of an algorithm to estimate the esophageal pressure without performing respiratory holds did not lead to significant differences in the obtained esophageal pressures and the optimal filling volume.

Multiple in vitro and in vivo calibration studies have been performed with various catheters in adult populations.^8,11,12 Although it is apparent that the correct balloon volume is necessary for accurate esophageal pressure readings, there is a lack of studies that have examined this in the pediatric context. One group of investigators reported the optimal filling volume to be between 0.2 and 0.5 mL in 3 young children instrumented with a 6 French catheter.¹⁰ Although our findings support these preliminary observations, we observed a wider range in our cohort, and we had performed the titration maneuver slightly different than theirs (we used steps of 0.2 mL instead of 0.1 mL). Our procedure to determine the optimal filling volume was adopted from a calibration technique developed in adults.¹¹ This makes it difficult to compare our findings with the adult study. For example, we filtered the esophageal pressure readings by using a band-stop filter to eliminate cardiac oscillations. Also, in our cohort, we mainly studied subjects instrumented with the small 6 French catheter.

Although using a fixed volume of 0.6 mL did not lead to statistically significant differences in calculated transpulmonary pressures in this study, the absolute differences in transpulmonary pressures for 6 French catheters ranged from –4.1 to 7.3 cm H₂O (Ptp_plat) and –4.4 to 5.8 cm H₂O for the Ptp_PEEP compared with the optimal filling volume that we think is clinically very relevant when setting mechanical ventilation in children. This difference highlights the need for an individualized balloon titration volume (Fig. 4).

Fig. 4.

Ptp_PEEP and Ptp_plat 0.6 mL vs optimal filling volume (Vbest).

Esophageal pressure monitoring in the clinical setting is still seldom.¹³ The esophageal pressure and optimal filling is affected by multiple factors, including position. Therefore, determining P_es in the clinical setting may become very time-consuming. However, we observed no difference in obtained P_es when the optimal filling volume was calculated on the raw signals instead of performing respiratory holds. The balloon filling titration procedure included 18 (6 French) or 28 (8 French) respiratory holds of 3 s with an interruption of the ventilation in children who were critically ill. We would therefore suggest eliminating the respiratory holds and to calculate the P_es to improve patient comfort and labor intensity of the filling volume titration procedure.

Multiple calibration procedures are described; they differ mostly in the definition of the optimal filling volume. In this study, we used two different methods and found absolute differences of ±4 cm H₂O in P_es at the optimal filling volume.^10,11 We consider these differences as clinically important. However, it remains unclear which method represents the true esophageal pressure, necessitating further studies to determine the optimal titration procedure in children. Furthermore, P_es measurements estimate pleural pressure and are subject to inaccuracies. Our findings were limited to subjects in the supine position, and the heart’s weight has been shown to affect P_es measurements and accuracy of transpulmonary pressure. This could affect balloon compliance and required filling volumes in patients in prone or semi Fowler’s positions. Also, the subjects were passive (not actively breathing) on the ventilator. In patients who are triggering the ventilator, the pleural pressure becomes negative, which could affect the optimal balloon compliance and esophageal elastance. This relationship needs to be explored further.

It is attractive to study optimal filling volume in an in vitro pediatric model as has been done by others,¹⁰ although it is challenging to adequately simulate the tone and motion of the esophagus.^8,10,12 In fact, esophageal elastance and esophageal wall pressure are important contributors and need to be accounted.^11,12 In adults with larger catheters, a pressure of the esophageal wall of 2.0 (±1.9) cm H₂O was found at the optimal filling volume, and, in adults with small volume catheters, a median pressure of the esophageal wall of 2.8 cm H₂O was found.¹¹ These values are comparable with the median pressure of the esophageal wall of 2.5 cm H₂O that we observed in our cohort. We also found that the esophageal elastance had a moderate correlation with sex, age, and body weight, which makes adequate in vitro models even more challenging.¹¹ Also, other factors need to be considered, such as the elastance of the air-filled esophageal balloon material and patient-specific factors, such as the amount and consistency of saliva.¹⁴ In our study, all esophageal catheters used were tested by using the mechanical ventilator (AVEA) before they were inserted into the patient. This procedure included a full expansion of the balloon and tests for leakage. Hereafter, the balloon was recalibrated and inflated with a pre-set inflation volume every 30 min as long as it was connected to the mechanical ventilator. The data in this study were obtained at different time points after insertion; this could have influenced the technical specifics and elastance of the balloon. However, in clinical scenarios, an esophageal catheter is used continuously and is only replaced for a new one in the same frequency as other nasogastric feeding tubes or in the case of technical failure. Therefore, we think this variation should be included and gives a realistic representation of the clinical situation.

There are some limitations to our study that need to be discussed. First, our study was designed as a single-center study, which potentially limited the generalizability of our findings. Second, we observed a fairly high failure rate (∼30%) of catheter placements. Although we do not know any studies that reported failure rates, we think it reflects the challenges so far in the implementation of routine esophageal pressure manometry in children who are ventilated. Third, we only measured the esophageal pressure with the balloon fully deflated (ie, 0 mL) in half the subjects, and we arbitrarily decided to use inflation steps of 0.2 mL. It cannot be ruled out that these increments might be too large in the pediatric patient; however, a comparative study has not yet been performed.

Conclusions

Analysis of our data supports the implementation of an individualized balloon inflation volume calibration in children who are receiving ventilation. This procedure can be performed without holds, but the optimal volume per step needs to be further studied. Pre-set volume should not be used because this leads to erroneous esophageal pressure readings.