Bench Assessment of Work of Breathing During a Spontaneous Breathing Trial on Zero Pressure Support and Zero PEEP Compared to T-Piece

Abstract

BACKGROUND:

Analysis of observational data suggests that both a T-piece and zero pressure support ventilation (PSV) and zero PEEP impose work of breathing (WOB) during a spontaneous breathing trial (SBT) similar to what a patient experiences after extubation. The aim of our study was to compare the WOB imposed by the T-piece with zero PSV and zero PEEP. We also compared the difference in WOB when using zero PSV and zero PEEP on 3 different ventilators.

METHODS:

This study was conducted by using a breathing simulator that simulated 3 lung models (ie, normal, moderate ARDS, and COPD). Three ventilators were used and set to zero PSV and zero PEEP. The outcome variable was WOB expressed as mJ/L of tidal volume.

RESULTS:

An analysis of variance showed that WOB was statistically different between the T-piece and zero PSV and zero PEEP on all the ventilators (Servo-i, Servo-u, and Carescape R860). The absolute difference was lowest for the Carescape R860, which increased WOB by 5–6%, whereas the highest for Servo-u, which reduced the WOB by 15–21%.

CONCLUSIONS:

Work may be imposed or reduced during spontaneous breathing on zero PSV and zero PEEP when compared to T-piece. The unpredictable nature of how zero PSV and zero PEEP behaves on different ventilators makes it an imprecise SBT modality in the context of assessing extubation readiness.

Introduction

A patient’s ability to successfully sustain spontaneous ventilation after extubation depends on how well the respiratory muscles cope with the imposed load on the respiratory system. As a screening test, clinicians aim to replicate the conditions that the patient will face after extubation, that is, extubation readiness during a spontaneous breathing trial (SBT). The SBT requires variable effort, depending on whether it is conducted without (ie, T-piece SBT) or with low levels of inspiratory pressure such as pressure support ventilation (PSV) or continuous positive airway pressure (CPAP) SBT. Low levels of pressure support are typically applied during the PSV SBT to compensate for the imposed work load due to the endotracheal tube and ventilator circuit.¹

However, PSV assists both the elastic and resistive portions of the work of breathing (WOB), and the addition of low levels of pressure support can produce large reductions in inspiratory work in patients who are on mechanical ventilation. For instance, addition of pressure support of 5 cm H₂O decreases inspiratory work by 31 to 38%.^2,3 The addition of PEEP during an SBT can also substantially increase cardiac output in patients with left-ventricular dysfunction.^4,5 Therefore, extubation after a weaning trial with PEEP of 5 cm H₂O could unmask an increased cardiorespiratory load that might lead to extubation failure.⁶ Although the T-piece SBT can avoid these problems, setting up the T-piece circuit could be time consuming and further stretch respiratory care resources.⁷ Performing an SBT without disconnecting the patient from the ventilator allows for additional patient monitoring, which adds another layer of safety to the weaning trial.

The purpose of this simulation-based study was to compare the imposed WOB during an SBT with a T-piece versus a connection to a ventilator with zero PSV and zero PEEP. We hypothesized that WOB by using zero PSV and zero PEEP as an SBT modality could be comparable with the T-piece SBT and thus closely simulate the respiratory load that a patient might face after extubation.³ Furthermore, we aimed to compare the imposed WOB during zero PSV and zero PEEP SBT using 3 different ventilators.

Quick Look.

Current Knowledge

Performing a spontaneous breathing trial (SBT)without disconnecting the patient from the ventilator has many merits and safety advantages for the patient. Although the optimal condition in which an SBT should be performed is still unknown, there is interest in doing an SBT with a T-piece because it imposes the work of breathing (WOB) that a patient might experience after extubation, especially in an obese population. Simulating this SBT condition without disconnection from the ventilator by using zero PSV and zero PEEP is of clinical interest.

What This Paper Contributes to Our Knowledge

The main results were that (1) the WOB imposed during zero PSV and zero PEEP is different from the WOB imposed with a T-piece; (2) the WOB imposed during zero PSV and zero PEEP varies among ventilators. Variable ventilator performance appears to complicate the utility of an SBT.

Methods

A breathing simulator was attached to a T-piece setup and a ventilator set at zero PSV and zero PEEP; an endotracheal tube was not attached to the system. The breathing simulator was programmed to achieve a target minute ventilation. The simulator software calculated WOB per breath (mJ).

Breathing Simulator

This study was conducted using the Active Servo Lung 5000 (ASL 5000 sw3.6, IngMar Medical, Pittsburgh, Pennsylvania), programmed with linear airway resistance and linear respiratory system compliance to simulate adult patients with moderate ARDS, COPD, and normal lung.⁸ The simulator ran with a closed loop control of a target tidal volume (V_T) so that simulated inspiratory effort would change automatically according to disease state and the imposed impedance of the attached airway system (ie, T-piece or ventilator). In this way, any difference in work would be due to the airway system for any given disease state. The simulation parameters for each disease state, as defined by Arnal et al⁸ in their study that described respiratory system’s mechanical properties, are displayed in Table 1. The mean ± SD for inspiratory work (mJ) was based on 10 consecutive breaths after steady state and calculated by the simulator and collected from the post-run analysis screen. Data were collected for an initial WOB calculation to be done without any respiratory system attached (ie, T-piece or ventilator) and described as baseline patient WOB.

Table 1.

Patient Simulation Parameters Programmed in the ASL 5000

Ventilators

The Servo-u, Servo-i ventilator (Getinge, Gothenburg, Sweden), and Carescape R860 (GE HeathCare, Boston, Massachusetts) were used for all experiments. The ventilator pre-check was performed before each experiment run. The experiment was performed without a heated humidifier or heat and moisture exchanger in the ventilator circuit. All the ventilators were set in pressure support mode, with zero pressure support and zero PEEP, ${\text{F}}_{{\text{IO}}_{\text{2}}}$ = 0.21, inspiratory rise time = 0.15 s, and inspiratory cycling = 30% of peak inspiratory flow. Inspiratory flow trigger threshold was set at 1.4 L/min in all the ventilators. The bias flow was set to 2 L/min in the Carescape R860 because the bias flow for the Servo ventilators defaults to this value for adults.

Procedure

The ASL 5000 simulator was set to collect data with body temperature and pressure corrections. The pre-use ventilator checks performed included an internal leakage test, ventilator circuit test with circuit compliance, flow, and pressure transducer calibration on the Servo-u, Servo-i, and Carescape ventilator.

Statistical Analysis

A sample of 10 breaths was measured, after steady state was achieved, for each SBT modality. The V_T and work done per breath (expressed as mJ) were recorded. The outcome variable was patient inspiratory work per breath expressed in mJ/L of V_T. This was obtained by dividing the work done per breath by the V_T recorded for each individual breath. Normal distributions were represented as mean ± SD. Means were compared by using one-way analysis of variance, whereas proportions were compared using the Fisher exact test. The mean ± SD was calculated from a sample of 10 resulting observations. The mean values for work (mJ/L) during baseline measurement, zero PSV and zero PEEP and the T-piece were compared by using one-way analysis of variance. The mean WOB differences among baseline, T-piece, Servo-u, Servo-i, and Carescape R860 were compared using the Tukey honest significant difference test for multiple comparisons. P values < .05 indicated significance. For these analyses, JMP (JMP Pro, version 16.0, SAS Institute) was used.

Results

The baseline WOB measured for models with normal, ARDS, and COPD lung models was 1,153 (1,151.9–1,154.1) mJ/L, 1,230 (1,228–1,232) mJ/L, and 1,820 (1,819–1,821) mJ/L, respectively. Measurements for the different SBT modalities showed the T-piece circuit imposed an inspiratory WOB of 1,167 (1,166.3–1,167.7) mJ/L, 1,243 (1,242.2–1,243.8) mJ/L, and 1,830 (1,828.9–1,831.1) mJ/L for normal, ARDS, and COPD models, respectively. Ventilators set at zero PSV and zero PEEP for normal, ARDS, and COPD models WOB imposed by Servo-u was 909 (95% CI 908.7– 909.3) mJ/L, 1,008 (1,007.4–1,009.6) mJ/L, and 1,550 (1,546.9–1,553.3) mJ/L; by Servo-i was 998 (997.2–998.8) mJ/L, 1,083 (1,082–1,084) mJ/L, and 1,673 (1,672–1,673) mJ/L; and by Carescape R860 was 1,222 (1,221–1,223) mJ/L, 1,288 (1,286.6–1,289.4) mJ/L, and 1,933 (1,931–1,935) mJ/L, respectively (Table 2). An analysis of variance on the WOB yielded significant variation among the different SBT modalities P < .001. A post hoc Tukey test showed that WOB for the T-piece and zero PSV and zero PEEP (on all 3 ventilators) differed significantly, at P < .001 for all 3 lung models (see the supplementary materials at http://www.rcjournal.com).

Table 2.

Summary Data for WOB for Each Type of SBT

To add clinical context, the WOB for each SBT modality was divided by the baseline WOB and the difference in terms of percentage was calculated for each of the 3 lung models. As shown in Figure 1, WOB was significantly different among the T-piece, Servo-u, Servo-i, and Carescape R860. The T-piece imposed a 1% additional WOB for all the lung models compared with baseline. Zero PSV and zero PEEP on Servo-u and Servo-i ventilators decreased WOB, whereas the Carescape R860 ventilator increased it for all 3 lung models (Table 3). The percentage difference in WOB in the 4 SBT modalities differed significantly, at P < .001 for the 3 lung models, simulating normal, ARDS, and COPD mechanics. Analysis of the pressure-time waveform curves (Fig. 2) showed that, despite the pressure support was set to 0, Servo-u added 4.5 cm H₂O pressure above PEEP; similarly, Servo-i added 3.5 cm H₂O, and Carescape R860 added 2.3 cm H₂O.

Fig. 1.

WOB was significantly different among the T-piece, Servo-u, Servo-i, and Carescape R860 (P < .001, Tukey honest significant difference test for multiple comparisons) (supplementary tables [see the supplementary materials at http://www.rcjournal.com]). For the T-piece, WOB increased for all lung models compared with baseline. Zero PSV and zero PEEP on the Servo ventilators decreased WOB, whereas the Carescape R860 ventilator increased WOB for all 3 lung models. The baseline WOB was calculated by measuring baseline WOB without any ventilator modality attached (ventilator or T-piece) using the equation ∫ (pairway − PEEP + P_mus) dVolume from [start of inspiration] to [start of expiration]. WOB = work of breathing; PSV = pressure support ventilation; P_mus = muscle pressure; dVolume = change in volume.

Table 3.

Comparison of the Change in WOB to Baseline

Fig. 2.

The pressure time curve analysis that shows hidden pressure support provided by all 3 ventilators when set on zero PSV and zero PEEP. The trigger work imposed by Carescape R860 was significantly different from the Servo ventilators despite the same trigger settings. PSV = pressure support ventilation.

Discussion

To our knowledge, this is the first bench study that measured and compared WOB imposed during an SBT on the T-piece and zero PSV and zero PEEP (zero CPAP), which simulated different lung mechanics. We found that there were significant differences in imposed WOB among the different setups and ventilators. Furthermore, when set to zero PSV and zero PEEP, ventilators performed in a variety of ways and the work they eventually impose depends on the proprietary software of the ventilator. A recent multi-center trial in France demonstrated that, when compared with the T-piece, SBTs done with zero PSV and zero PEEP had similar re-intubation rates.⁹ Our study showed that zero PSV and zero PEEP may not demand the respiratory effort that a patient would experience after extubation, the work imposed may be lower (Servo-u, Servo-i) or slightly higher (Carescape R860), depending on the ventilator used.

The ideal SBT modality has been extensively debated but should ideally simulate the physiologic stress that the patient might experience after extubation. Sassoon et al¹⁰ studied the impact of CPAP and pressure support on pressure-time product in patients recovering from acute respiratory failure. Pressure-time product was used as an estimate of the metabolic work (O₂ consumption) of respiratory muscles. They observed that CPAP during weaning reduced pressure-time product by decreasing resistance of the respiratory system and reducing the intrinsic PEEP relative to the airway pressure, whereas pressure support offloaded the respiratory muscles by assuming part of the work. Sklar et al,³ in a meta-analysis of 16 studies, which included 239 subjects, similarly found that CPAP and pressure support provided during SBT reduced WOB compared with conditions after extubation. All respiratory effort parameters during SBT were similar between the T-piece and zero CPAP and approximated patient WOB after extubation. Furthermore, Mahul et al¹¹ in their study of subjects who were morbidly obese (body mass index ≥ 35 kg/m²) and undergoing weaning compared low levels of pressure support and PEEP with the T-piece. They observed that the T-piece or zero PSV and zero PEEP trials better predicted postextubation WOB in this population. Our findings in this bench study challenge these findings and indicate that such WOB may differ in magnitude, depending on the SBT setup.

Conducting an SBT without disconnecting the patient from the ventilator has several merits. Such a setup allows for quick and immediate resumption of mechanical ventilatory support in the case of weaning failure. Moreover, alarms inherent to ventilators provide an additional layer of safety in the early detection of patients who are struggling. Finally, the transition to an SBT requires less respiratory therapist time. In our study, the WOB on zero PSV and zero PEEP was 10–20% lower than baseline, irrespective of the disease state for the Servo ventilators. The Servo-u ventilator consistently offloaded the respiratory muscles 5–7% more than did the Servo-i. This offloading of respiratory muscles was due to the inherent small end-inspiratory positive airway pressure (mean 3.5 cm H₂O for Servo-i, 4.5 cm H₂O for Servo-u, and 2.3 cm H₂O for Carescape R860) induced by the CPAP engineering algorithm that supplies rapid fresh gas delivery, which seems to function as low-level pressure support. In addition, although the Servo-u ventilator was set at zero PEEP, the ventilator still provided a small CPAP level of 1 cm H₂O by design. This assistance probably accounts for the reduction in WOB observed for the servo ventilators.^10,12

In contrast, the Carescape R860 ventilator imposed a higher WOB while on zero CPAP compared with the Servo ventilators and 4–6% higher than the WOB at baseline. On further analysis of the pressure-time waveform, the higher WOB by the Carescape R860 was due to the apparently higher trigger work and lower pressure support level (Fig. 2), despite all ventilators having the same setting for the trigger threshold (ie, sensitivity) and pressure support setting (ie, pressure support was set to 0 cm H₂O). Importantly, all 3 ventilators by design provided a small pressure support level while set at zero CPAP that served to reduce WOB (Fig. 2).¹³

From a clinical standpoint, a difference of 10–20% in WOB might not seem like a drastic difference and perhaps deemed insignificant in the assessment of extubation readiness.^14,15 However, we showed that how a ventilator behaves on zero PSV and zero PEEP depends on the proprietary software and that the specific ventilator design might end up imposing higher or lower WOB compared with baseline and the T-piece. This might create a lack of precision in assessing patients when zero PSV and zero PEEP is used as an SBT modality with different ventilators. Furthermore, the intrinsic pressure support provided by ventilators during a zero PSV and zero PEEP trial might affect predictors for readiness of a weaning trial (eg, rapid shallow breathing index) by 10%.³ Our study was limited because it was performed by simulating lung mechanics in a single compartment model, whereas lungs behave as a heterogenous unit in clinical practice. We used the 3 ventilators to which we had access, which might limit the generalizability of our findings for ventilators from other manufacturers.

Conclusions

The inspiratory WOB imposed during a SBT using a T-piece was not identical numerically when using zero PSV and zero PEEP on ventilators (Servo-i, Servo-u, and Carescape R860). The WOB imposed when using zero PSV and zero PEEP depended on the proprietary behavior of the ventilator. The variability of WOB imposed by the different ventilator models when operating at zero PSV and zero PEEP might confound its universal application as an SBT modality.