Programming assisted ventilation parameters in daily practice is a challenge, since excessive assist results in diaphragmatic dysfunction due to muscle rest and facilitates trigger failure. On the other hand, excessive respiratory effort causes dyspnea,1 asynchrony and even muscle fatigue and damage,2 with the risk of prolonging the duration of mechanical ventilation (MV).3 It therefore seems advisable to determine muscle effort, and recent reviews address this issue as well as the difficulty of incorporating it to routine monitoring practice in the Intensive Care Unit (ICU).4 At the patient bedside or point of care, respiratory effort is assessed through clinical examination and monitoring of the respiratory signals generated by the respirator. The clinical assessment of respiratory difficulty is based on a series of clinical signs, though observation and quantification are characterized by important inter-observer variability.5 To the best of our knowledge, no clinical studies have compared clinical evaluation with objective measures of respiratory muscle effort. Interpretation of the curves generated onscreen by the respirator depends on the experience of the physician, though even experienced professionals can have difficulties in interpreting such curves.6 The changes which respiratory muscle effort produce in the airway pressure curve in volume assist control ventilation (ACV) and in inspiratory flow in ventilation with pressure support (PS) allow us to deduce the intensity of respiratory effort.7,8 In these cases we are assuming that patient effort modifies the pressure in ACV or PS, and therefore the morphology of the curve with respect to passive or controlled ventilation represented by the calculated respiratory muscle pressure (Pmus).9,10 The incorporation of this signal to the monitoring of patients subjected to mechanical ventilation allows for improved interaction with the respirator,8,11 but little information is available on the capacity to estimate pleural pressure amplitude in assist mode ventilation.12 In this regard, it has been used in ventilation with PS to optimize the assist level, showing good correlation with esophageal pressure (Pes),8 but the concordance between the two signals has not been established, and it is therefore not clear whether they are equivalent in measuring respiratory effort.13

The present study was carried out to evaluate the concordance between invasive measurement of respiratory effort with the Pes probe and noninvasive measurement based on the curves generated by the respirator.

Ten patients were studied. Their characteristics are summarized in Table 1. A total of 50 recordings were obtained, of which 41 proved suitable for analysis. The rest were discarded due to artifacts in the Pes signal. The applied CPAP was 5cmH2O, except in one case that required 18cmH2O because of PEEPi 20cmH2O. In this mode the peak inspiratory flow (V′max) was 0.7±0.1l/s and the minute volume (VE) was 9±2l/m. In ACV1, the tidal volume was 0.42±0.04l, VE was 9±2l/m and V′ 1±0.2l/s. The inspiratory flow decreased to 0.49±0.1l/s for ACV5, with VE 9±3. In PS the assist for maximum muscle rest (PS1) was 15±5cmH2O, and decreased to 9±4cmH2O in PS5, without significant changes in V′ 0.8±0.1 versus 0.8±0.04l/s; VE was higher than with the other modes, 10±3l/m (Fig. 1).

Figure 1.

Representative example, recorded (Paw, Pes, IF) and calculated signals (Pmus, vol) in CPAP, ACV and PS. With low assist, the respiratory variation of calculated esophageal and muscle pressure is evident and of greater amplitude than with high assist. The amplitude of both is different, however. In this case measurement is difficult: in the case of esophageal pressure due to the oscillation caused by the heart beat, and in the case of calculated pressure muscular due to oscillations or “spikes” caused by changes in inspiratory flow. ACV1: assist control ventilation with flow 0.95l/s; ACV5: assist control ventilation with flow 0.75l/s; CPAP: continuous positive airway pressure 5cmH2O; F: flow (l/s); Pes: esophageal pressure; Pmus: calculated muscle pressure (cmH2O); PS5: pressure support of 8; PS1: 16cmH2O with PEEP 4cmH2O; Paw: airway pressure; Vol: volume (l).

(0.44MB).

The variables measured in Pes and Pmus are shown in Table 2. The increase in respiratory assist reduced the measured and calculated effort to a significant degree (p<0.001). On comparing these measurements in each ventilation mode, the mean differences did not prove significant, but the concordance analysis reflected important dispersion (Table 3). Accordingly, the variation between the measurement and calculation of effort could be 16cmH2O for respiratory variation and 264cmH2O·s/m in PTPm (Figs. 2 and 3). These differences were not influenced by the amplitude of effort, but the percentage error between the measurements and calculations was smaller with low assist levels (CPAP, ACV5, PS5).

Figure 2.

Linear regression analysis and 95% confidence interval with Bland–Altman diagram: mean difference between the two signals and limits of agreement (confidence interval).

Delta Pe-Pmus.

(0.28MB).

Figure 3.

Linear regression analysis and 95% confidence interval with Bland–Altman diagram: mean difference between the two signals and limits of agreement (confidence interval).

PTPPe-Pmus.

(0.26MB).

The results of our study indicate that the calculated respiratory effort does not allow us to infer the measured value in Pes. These differences may be due to a number of factors, particularly the parameters that intervene in the calculation of Pmus, the accuracy of the respiratory mechanics measurements, and their validity in the presence of muscle effort. The measurement of Pes and the transmission of effort to the airway may also exert an influence.

The calculated Pmus signal depends on the respiratory mechanics parameters (E and R) that intervene in the equation of movement.19,20 Any measurement method has limitations, since these parameters can change from one respiratory cycle to another, even with deep sedation.21 In practice, the most widely used method is inspiratory and expiratory occlusion, based on a mathematical model that assumes the absence of effort.22 We used a previously validated model under spontaneous breathing,23 assisted ventilation13,24 and controlled ventilation.25 The differences observed between measured and calculated effort were similar to those published elsewhere using this approach based on a simple mechanical method.26 More complicated methods have failed to improve the results. In ventilation with PS,27,28 respiratory effort calculated with Pmus and other ventilation variables may differ with respect to the measured value by up to 7j/m (normal 5–10j/m). Experimental studies using complex mathematical algorithms have reported differences of 0.3j/l.29 In patients with proportional assisted ventilation, the difference in measured and calculated respiratory effort may be 5cmH2O/s,11 which is high on considering the acceptable limit in mechanical ventilation (4cmH2O/s17). In this ventilation mode greater similarity between the two signals was to be expected. By definition, in proportional assisted ventilation, pressure and flow are proportional to the calculated respiratory effort,30 and the respiratory mechanics measurements are periodic.18

The conditions of the respiratory mechanics measurements may exert and influence. These parameters are estimated in passive ventilation and may change under conditions of spontaneous respiration,31 limiting their usefulness for calculating muscle effort.32,33 Thus, after failed respirator weaning, a decrease in compliance has been observed, with an increase in resistances and PEEPi.34,35 However, these results have not been confirmed in other studies in patients requiring mechanical ventilation due to chronic obstructive pulmonary disease (COPD) and acute respiratory distress syndrome (ARDS).36 Differences in the respiratory pattern may modify the respiratory mechanics in assisted ventilation, due to the dependency of elastance and resistance upon volume and flow respectively.37 The high inspiratory flow used for passive ventilation may overestimate the inspiratory resistances, and the calculated Pmus therefore will be greater than Pes.

Other factors can influence the differences between calculated and measured respiratory muscle effort. The fact that Pes does not adequately represent pleural pressure and/or the respiratory effort does not modify Paw. The Pes is a point pressure value and might not reflect the global lung surface. Thus, in the presence of lung zones with different mechanical characteristics, pleural pressure will not only exhibit a gravitation distribution and will prove more heterogeneous, as has been evidenced in individuals with COPD38 and recently in patients with ARDS.39 In these cases, respiratory effort produces a redistribution of the air at the start of inspiration, without changes in volume, and Pes does not represent pleural pressure, particularly in dependent zones. On the other hand, respiratory effort might not generate a proportional pressure change in the airway, and consequently calculated Pmus will not represent the change in pleural or esophageal pressure. In healthy individuals, the change in alveolar pressure during inspiration is similar to the change in Paw, though there may be inter-individual differences of up to 12%.40 However, when airway resistance is increased (COPD), respiratory effort does not adequately transmit to the airway, as has been demonstrated following electrical stimulation of the phrenic nerve,41 and respiratory effort is underestimated.19,38 Likewise, in the presence of low lung compliance, respiratory effort causes compression of the inhaled gas, with few changes in lung volume and a disproportion between Pes and the airway.42 Our results show that in the presence of high assist and weak effort, the measured Pes is greater than the calculated value. In other words, in the case of high assist, minor effort does little to modify Paw in ACV or inspiratory flow in PS. In contrast, in the case of low assist, with high muscle effort, the calculated pressure exceeds the measured value, due to the aforementioned causes.

The main limitation of our study is the heterogeneity of the group of patients involved. Although they were undergoing weaning from the respirator, the respiratory mechanics and days of mechanical ventilation differed greatly. A larger and/or stratified sample could improve the results obtained in this regard.

In conclusion, the pressure and flow curves of the respirator allow us to assess the changes in respiratory muscle effort on modifying the ventilation parameters, but not to quantify their amplitude.

Francisco Ruiz-Ferrón: Preparation of the protocol, registries in 5 patients, analysis of the data and preparation of the manuscript.

Jose Manuel Serrano-Simón: Preparation of the protocol, registries in 5 patients, analysis of the data and preparation of the manuscript.

None declared.

None declared.