INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Dynamic pulmonary hyperinflation and intrinsic positive end-expiratory pressure (PEEPi) are well recognized events in patients with severe chronic obstructive pulmonary disease (COPD) during acute respiratory failure as well as in stable condition (1). PEEPi provides an inspiratory threshold load which may be substantial (1), and must be counterbalanced by the patients' inspiratory muscles before starting inspiration (4). Recent work has shown that in ventilator-dependent COPD patients, PEEPi accounts for about 40%, on average, of the total ventilatory workload (5, 6). In these patients accurate measurement of PEEPi is important for the correct setting of the ventilator, and particularly for the level of PEEP and continuous positive airway pressure (CPAP) to be applied (5, 7). In fact, low levels of PEEP/CPAP, compared with the actual PEEPi, will not provide sufficient assistance, whereas excessive levels of PEEP/CPAP can worsen pulmonary hyperinflation, thus reducing further the inspiratory muscle pressure generation capacity (8). Therefore, measurement of PEEPi is important in these patients. Unfortunately, the end-expiratory occlusion method commonly used to quantify PEEPi (4) seldom works in actively breathing patients during assisted ventilation. The esophageal balloon technique estimates the value of dynamic PEEPi (PEEPi,dyn) (5, 9). However, in patients with advanced COPD, PEEPi,dyn may be significantly lower than the mean end-expiratory static recoil pressure of the total respiratory system (PEEPi,st) measured with the end-respiratory occlusion technique (10). In this study we suggest a new method for measuring PEEPi,st in actively breathing patients. The rationale of this method is provided by the well-established relationship between the elastic recoil of the respiratory system and the pressure-generating capacity of the inspiratory muscles (13), such that the difference between the decrease in mouth pressure (Pao) and esophageal pressure (Ppl) during a sustained maximal or submaximal inspiratory effort against occluded airway at end-expiration should provide the value of PEEPi,st. The aim of our study was to assess the reliability of this method for measuring PEEPi,st during spontaneous breathing in ventilator-dependent COPD patients (5) in whom correct measurement of PEEPi is important.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Patients

Eight ventilator-dependent COPD patients, admitted to the intensive care unit (ICU) of the Division of Pulmonary Disease, Medical Center for Rehabilitation, S. Maugeri Foundation, Veruno (NO), Italy, were recruited for this study. The diagnosis of COPD was confirmed by clinical history and pulmonary function tests (Table 1). Patients were all tracheostomized (cuffed tracheostomy cannula, 8 mm internal diameter and 11.5 cm in length; Shiley Inc., Irvine, CA) and had been mechanically ventilated (Evita ventilator, Dräger, Germany) for 45 ± 20 d before the study. All patients had undergone at least two weaning trials during the days preceding the study and were able to sustain a period of at least 20 min of unsupported breathing.

The study was conducted along the guidelines of the Declaration of Helsinki and the protocol was approved by the institution's ethics committee. Written informed consent was obtained from the patients or their next of kin.

Measurements

Flow, volume, and pleural and gastric pressures were measured as previously described (5, 9). Flow () was measured with a heated Lilly pediatric-type pneumotachometer connected to a differential pressure transducer (Screenmate 701240-015009; Jaeger, Wurzburg, Germany), and inserted between the proximal tip of the tracheal cannula and the Y connector of the ventilator (Figure 1). Volume (V) was calculated through numerical integration of the flow signal. Changes in pleural (Ppl) and abdominal (Pab) pressures were estimated from changes in esophageal and gastric pressures, respectively. Both Ppl and Pab were measured using two balloon-tipped catheter systems connected to two differential pressure transducers (143PC03D; Micro Switch, Honeywell, Freeport, IL). The catheters were 80 cm in length and 1.7 mm in internal diameter; the balloons were 10 cm in length and 2.4 cm in circumference. The correct positioning of the esophageal balloon was assessed by means of the occlusion test (14). Another catheter, similar in length and internal diameter, and another pressure transducer (143PC03D; Micro Switch, Honeywell) were used to sample the pressure at the airway opening (Pao) via a side port inserted between the tracheostomy tube and the pneumotachometer. Transpulmonary (PL) and transdiaphragmatic (Pdi) pressures were obtained by subtracting Ppl from Pao and Pab, respectively. Minute ventilation (E), tidal volume (VT), inspiratory (TI) and expiratory time (TE), total respiratory cycle duration (Ttot) breathing frequency (f), mean inspiratory flow (VT/TI), and the duty cycle (TI/Ttot were calculated as average values from one-minute continuous recordings of flow and volume.

View larger version (15K): [in this window] [in a new window]   Figure 1.   Schematic representation of the apparatus used to perform the occluded inspiratory effort maneuver. A selective inspiratory occlusion was performed during expiration by means of the inspiratory pneumatic shutter. Consequently, patients were allowed to exhale freely, after which they were encouraged to perform a maximum inspiration against the occluded airway. The expiratory shutter was kept open during the maneuver.

Dynamic pulmonary compliance (Cdyn_L) and pulmonary resistance at midinspiratory volume (RL) were computed from PL, , and V records as previously described (5). RL included the endotracheal tube resistance. PEEPi,dyn was measured as the decrease in Ppl preceding the start of inspiratory flow (5, 9) in patients who did not show expiratory muscle activity (Figure 2). Five patients exhibited expiratory muscle activity, as indicated by a rise in Pab during expiration followed by a fall at the beginning of the inspiratory effort (Figure 3). In these patients PEEPi,dyn measurement was corrected for the overestimation induced by expiratory muscle activity by subtracting the decrease in Pab from the drop in Ppl when both are present in the interval between the onset of inspiratory effort and the point of zero flow (5, 9). Expiratory flow limitation during tidal breathing was qualitatively assessed by means of the technique of external positive end- expiratory pressure (PEEP) withdrawal (15). Maximum inspiratory pressure (MIP) and maximum pleural (Ppl max) pressure swings were measured in each subject by means of the Mueller maneuver, i.e., a sustained, maximum inspiratory effort against occluded airway starting at EELV (5, 9, 16).

View larger version (17K): [in this window] [in a new window]   Figure 2.   Mueller maneuver in a representative patient without expiratory muscle activity. = flow; Pao = airway opening pressure; Ppl = esophageal pressure; PL = transpulmonary pressure; Pab = gastric pressure; Pdi = transdiaphragmatic pressure; first vertical line = start of the inspiratory effort; second vertical line = zero-flow point; horizontal dashed line = zero-flow. Expiratory muscle activity was negligible, as indicated by Pab tracing that, after the initial decay during expiration because of the postinspiratory activity of the diaphragm, remains stable throughout most of expiration and does not change in the time interval between the beginning of Ppl and Pdi inspiratory swings and the point of zero-flow (time interval between the two vertical lines).

View larger version (16K): [in this window] [in a new window]   Figure 3.   Mueller maneuver in a representative patient with expiratory muscle activity. First vertical line = start of the inspiratory effort; second vertical line = zero-flow point; horizontal dashed line = zero-flow. Expiratory muscle activity was observed, as indicated by the increase in Pab during expiration followed by its decrease at the beginning of the inspiratory effort. In this case, to compute PEEPi,dyn and Ppl max, the end-expiratory level of Ppl was corrected by the negative Pab preceding the zero-flow point. For definition of abbreviations, see Figure 2.

Measurement of PEEPi,st. PEEPi,st during spontaneous breathing was measured as the difference between MIP and Ppl max at the plateau. Patients enrolled for the present study were severe COPD patients. In this condition, it has been shown that an accurate evaluation of inspiratory muscle strength (Pmus) is obtained only by adding the total respiratory system recoil pressure (Prs) to maximum airway opening pressure, as indicated by the equation: Pmus = Prs + MIP (Equation 1) (13, 17). Because Pmus is equivalent to the change in Ppl between the start and the end of an occluded inspiratory effort (Ppl) (18), and because the total respiratory system recoil pressure is PEEPi,st, an alternative way to write Equation 1 is: Ppl max = PEEPi,st + MIP (Equation 2). A rearrangement of Equation 2 produces a third equation: PEEPi,st = MIP  Ppl max (Equation 3), given that MIP and Ppl max are both negative values. In summary, from a theoretical point of view, it is possible to assess PEEPi,st during spontaneous breathing by the difference between MIP and Ppl max. In the five patients who had expiratory muscle activity, PEEPi,st measured as MIP  Ppl max was corrected with the same method used in the measurement of PEEPi,dyn (9), this correction amounting to 2.0 ± 1.6 cm H₂O on average.

Experimental Procedure and Study Design

Patients were studied in a semirecumbent position throughout the procedure. Following topical anesthesia (10% xylocaine spray), the two balloon-tipped catheters were positioned through the nose into the stomach and the esophagus as previously described (5, 9). The occlusion test was satisfactory in every instance, the average Ppl/Pao being 1.015 ± 0.045 (range, 0.938 to 1.078).

Once the patient appeared to be accustomed to the experimental setting and measuring equipment, the ventilatory mode and setting prescribed by the caring physician was discontinued. The measuring equipment was composed of a connector for airway pressure sampling, the pneumotachometer, and a T-tube circuit with the same resistive characteristics as the ventilator tubings. The inspiratory and expiratory lines of the T-tube were separated by a Hans-Rudolph one-way valve to allow either inspiratory or expiratory line occlusions by means of two pneumatic shutters (Figure 1). Both shutters were used to perform the occlusion test. The assessment of maximum inspiratory pressures was performed by occluding only the inspiratory shutter.

After the spontaneous breathing pattern was recorded, the inspiratory limb of the circuit was occluded during a free expiration through the expiratory limb of the circuit. Therefore, the inspiratory occluded effort was performed at end-expiration. At this point, patients were encouraged to perform maximum inspiratory muscle contraction against the occluded airway (Mueller maneuver) (16). The occlusion was released after at least 2 s, and the maneuver was repeated after 30 s of spontaneous breathing until two values of MIP were within 5% of each other. A maximum of 5 maneuvers were sufficient in all patients.

Validation of PEEPi,st measurement. At the end of the above described procedure, the patients were sedated (propofol 2.5 mg/kg, intravenously) and ventilated in the control mode with the settings prescribed by the primary physician. On average, the settings of the ventilator were as follows: VT = 0.54 ± 0.09 L; f = 16.0 ± 3.7 breaths/ min; inspiratory flow = 0.83 ± 0.21 L/s; TI/Ttot = 0.23 ± 0.05. After a few minutes of mechanical ventilation to ensure for respiratory muscle relaxation, an end-inspiratory occlusion was performed using the appropriate "hold" button of the ventilator. After a plateau was reached in Pao, the occlusion was released, and a series of brief interruptions was performed throughout the following relaxed expiration by manual occlusion of the expiratory part of the ventilator, until expiratory flow became nil, according to the technique described in detail elsewhere (19, 20). After each airway occlusion, airflow fell to zero, lung volume did not change, and Pao became positive, exhibiting in all instances a plateau reflecting the elastic recoil pressure of the respiratory system at various lung volumes during lung deflation. We checked that the PL value at the plateau during the end-inspiratory occlusion was higher than that at the plateau recorded during the Mueller maneuver performed in the first part of the procedure. The quasi-static pressure-volume (P/V) relationship of the respiratory system, as well as of its lung and chest wall compartments during lung deflation were determined by plotting the plateau in postinterruption Pao, PL, and Ppl pressures, respectively, against the corresponding volume above relaxation volume (Vr, i.e., the end-expiratory position observed when expiratory flow and Pao became nil) (19, 21).

The PL/V and Pao/V curves were used to compute the EELV during spontaneous breathing and the reference PEEPi,st (PEEPi,st-Ref) at that volume, respectively. To this purpose, the absolute value of PL at the plateau during the Mueller maneuver (Figures 2 and 3) were considered to be the result of the static lung elastic recoil at the EELV of the breath preceding the occluded inspiratory effort. The mirror image of PL plateau (PL plateau) was superimposed on the mirror image of the passive PL/V curve (PL/V) obtained during mechanical ventilation (Figure 4, left). The intercept between PL plateau and PL/V curve identified the EELV during spontaneous breathing referenced to Vr. The intercept between EELV referenced to Vr and the Pao/V curve gave the value of PEEPi,st-Ref for that particular lung volume (22) (Figure 4, right).

View larger version (11K): [in this window] [in a new window]   Figure 4.   Passive pressure-volume relationship from a representative patient. V = increase in lung volume above the relaxation volume; Pao = airway opening pressure; PL = mirror image of the transpulmonary pressure; PEEPi,st-Ref = reference static intrinsic positive end-expiratory pressure.

Data Analysis

Using a personal computer (Compaq 386 equipped with 80387 math coprocessor; Compaq, Houston, TX, and with DT2801/A A/D board; Data Translation, Marlboro, MA) all signals were analog-to-digital converted, displayed on-line through the procedure, and stored on 3.5-inch floppy diskettes at a sampling rate of 100 Hz. The data were collected and the tracings analyzed using appropriate software (Labdat and Anadat; RHT-InfoDat Inc., Montreal, PQ, Canada). Data are reported as mean ± SD unless otherwise specified. Simple linear regression techniques (23) were employed to calculate the correlation between PEEPi,st and PEEPi,st-Ref, and between PEEPi,dyn and PEEPi,st. Comparisons between MIP and Ppl max and between PEEPi,dyn and PEEPi,st were performed by means of the Wilcoxon signed rank test (24). A value of p < 0.05 was taken as the criterion for accepting statistically significant difference. PEEPi,st and PEEPi,st-Ref were also compared by means of the Bland and Altman analysis (25).

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Table 2 shows mean values of breathing pattern and lung mechanics during spontaneous breathing. In general, VT was small and the frequency of breathing high. Lung mechanics was abnormal because of low dynamic compliance, high pulmonary resistance, and PEEPi (Table 2).

Mean values of Ppl max and MIP are also shown in Table 2. On average Ppl max was significantly greater than MIP (p < 0.05). The difference between Ppl max and MIP, i.e., the change in PL occurring between the start of Ppl decrease and the PL plateau value reflected PEEPi,st (Table 2). The PL occurring during the Mueller maneuver is composed of two parts (Figure 2): the first part is the change in PL between the start of Ppl decrease and the point of zero flow corresponding to PEEPi,dyn; the second is the further increase in PL occurring during the occluded inspiratory effort after the point of zero flow to reach the plateau. The PEEPi,dyn/PEEPi,st ratio averaged 0.55 ± 0.15. Individual values are shown in Figure 5.

View larger version (18K): [in this window] [in a new window]   Figure 5.   Relationship between PEEPi,st and PEEPi,dyn. PEEPi,st = static PEEPi assessed as the difference between MIP and Ppl max; PEEPi,dyn = dynamic PEEPi; solid line = identity line; dashed lines = isophlets for different PEEPi,dyn to PEEPi,st ratios. The simple linear regression equation is provided in the figure.

Validation of PEEPi,st. The lung volume at which the end-inspiratory occlusion was performed during mechanical ventilation averaged 1.045 ± 0.342 L above Vr. The EELV of the breath preceding the Mueller maneuver averaged 0.772 ± 0.289 L above Vr. On average, PEEPi,st-Ref computed on the passive Pao/V curve for that lung volume amounted to 13.1 ± 3.0 cm H₂O. The relationship between PEEPi,st measured as MIP  Ppl max and Peepi,st-Ref computed from the quasi-static expiratory Pao/V curve is shown in Figure 6. All the individual data essentially lie on the identity line. The strong agreement between individual values of PEEPi,st and PEEPi,st-Ref is confirmed by the Bland and Altman analysis shown in Figure 7.

View larger version (17K): [in this window] [in a new window]   Figure 6.   Relationship between PEEPi,st and PEEPi,st-Ref. PEEPi,st = static PEEPi assessed as the difference between MIP and Ppl max; PEEPi,st-Ref = reference static PEEPi computed as described in Figure 4 (see also text); dashed line = identity line; solid line = regression line. The simple linear regression equation is provided in the figure.

View larger version (11K): [in this window] [in a new window]   Figure 7.   Comparison between PEEPi,st-Ref and PEEPi,st with the Bland and Altman analysis. Average PEEPi = average value of each pair of measurements (PEEPi,st-Ref + PEEPi,st/2); PEEPi = difference between the two values; d = mean bias; s = standard deviation of the individual differences; d + 2.306 = upper limit of agreement; d  2.306 = lower limit of agreement between PEEPi,st-Ref and PEEPi,st; PEEPi,st = 13.0 ± 2.9 cm H2O; PEEPi,st-Ref = 13.1 ± 3.0 cm H2O; d = 0.1 ± 0.4 cm H2O; p = not significant. A strong agreement is observed between PEEPi,st and PEEPi,st-Ref.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The data of this study show that PEEPi,st can be measured in actively breathing patients from the difference between MIP and Ppl max during the maximum occluded inspiratory effort, i.e., the Mueller maneuver commonly used to assess the inspiratory muscle strength. Clearly, an esophageal balloon is needed to measure changes in Ppl and the occlusion test must be satisfactory, which is the case when an equal decrease occurs in Pao and Ppl during respiratory efforts against occluded airway (14) because the lung volume does not change, and hence, PL (i.e., Pao  Ppl) remains constant throughout the occlusion (14). Similarly, PL remains constant during the Mueller maneuver (Figures 2 and 3) (14). In normal subjects in whom the EELV is coincident with Vr, a simultaneous decrease in Pao and Ppl indicates the onset of the inspiratory effort against occluded airway. By contrast, in patients with severe COPD and dynamic pulmonary hyperinflation, the decrease in Pao at the beginning of the occluded effort is preceded by the decrease in Ppl (Figures 2 and 3). The value of Ppl in that interval has been interpreted as the amount of pressure needed to counterbalance PEEPi,dyn (4). Because Pao does not change in that interval, the change in PL also provides measurement of PEEPi,dyn. This interpretation is valid provided that expiratory muscles are relaxed and airway opening is not occluded at end-expiration. This was the case in three of our eight patients in the present study (Figure 2), whereas in the other five patients expiratory muscles were active during expiration such that correction for expiratory muscle relaxation at the beginning of the inspiratory effort was needed (5, 9, 26). In all eight patients, after the initial change due to PEEPi,dyn, PL exhibited an additional modification before reaching the plateau which lasted throughout the occlusion. The fact that this additional change was present in patients regardless of expiratory muscle activity strongly supports that it was not the result of abdominal muscle relaxation. We excluded an artifact due to bad positioning of the esophageal balloon not only because the occlusion test was satisfactory in all the patients, but also because, after the slight initial change following the offset of PEEPi,dyn, PL exhibited a plateau throughout the occluded inspiratory effort in all patients though both Pao and Ppl continued to decrease to the point of minimum value, namely MIP and Ppl max, respectively. According to Equation 3, the difference between MIP and Ppl max, and hence the PL from the start of the inspiratory effort to PL plateau represented PEEPi,st. This was confirmed by the identity of PEEPi,st and the reference PEEPi,st measured on the quasi-static P/V curve. Therefore we interpreted the difference in PL between the end of PEEPi,dyn period and the plateau as due to stress adaptation phenomena of the lung tissues plus equilibration among lung units with different time constants () during the occlusion. Our interpretation is in agreement with the results of Petrof and colleagues (10) and Maltais and colleagues (11) who have shown that in COPD patients PEEPi,dyn is smaller than the true end-expiratory elastic recoil, i.e., PEEPi,st.

The measurement of PEEPi,st as MIP  Ppl max could be affected by substantial errors if correspondence between Ppl and Pao, quantified by the Ppl/Pao ratio obtained during the occlusion test (14), is lacking. Of course, the error would increase with increasing Ppl max. It follows that Ppl max should be corrected for Ppl/Pao ratio before subtracting Ppl max from MIP to obtain PEEPi,st. The correction can be accomplished by dividing the measured Ppl max by Ppl/Pao found during the occlusion test.

A recent study by Yan and colleagues in normal subjects in whom dynamic hyperinflation has been induced by breathing through a Starling resistor (27) suggested that PEEPi,st could be underestimated because of persistent postinspiratory muscle activity. This could be the case in asthmatics (28, 29), but it is unlikely to be the case in patients with COPD in whom several studies have shown that the postinspiratory activity decays early in expiration (30, 31). In our study, the close agreement found between PEEPi,st measured during the Mueller maneuver and PEEPi,st-Ref measured from the quasi-static P/V curve (Figures 6 and 7) supports our conclusion. The Mueller maneuver to measure PEEPi does still require the invasive procedure of the esophageal and gastric catheters, but it helps to overcome the methodological problems of the end-expiratory occlusion maneuver in actively breathing patients. In fact, active patients react to the end-expiratory airway occlusion such that the continuous inspiratory and expiratory muscle activity may prevent the plateau in the pressure at the airway opening, thus negating noninvasive measurement of PEEPi,st without sedation. Moreover, severe COPD patients often exhibit an increased respiratory frequency (16) and breath-by-breath variability of the respiratory cycle duration that can make it impossible for the operator to select the correct time for airway occlusion. In contrast, during the Mueller maneuver with the technique used in this study, only the inspiratory side of the circuit is occluded, such that the patient can perform a complete expiration before the occluded inspiratory effort.

As previously mentioned, a possible limitation of the method we propose in the present study is the need to insert both gastric and esophageal balloons, a procedure that can be uncomfortable for the patient (32). However, this limitation is common to the method of end-expiratory occlusion to exclude or correct for expiratory muscle activity. On the other hand, during the Mueller maneuver, all the following measurements can be obtained: an estimation of the pressure-generating capacity of the inspiratory muscles; the evaluation of PEEPi,st; both PEEPi,dyn and PEEPi,st measurements within the same breath, i.e., at the same end-expiratory volume, provided that appropriate corrections are done for effects of Ppl/Pao ratio and expiratory muscle activity on the measurement.

The possibility of assessing PEEPi,st during spontaneous breathing provides some clinical advantages. First, PEEPi,st represents a more accurate estimate of the true elastic lung recoil than PEEPi,dyn, and some definition of lung inhomogeneities through the PEEPi,dyn/PEEPi,st ratio (Figure 5) (11). Second, and more important, PEEPi,st could provide a better guideline to set the optimal level of PEEP and CPAP in ventilator-dependent patients with flow limitation (11). Hence, our technique could prove very useful in clinical practice.

In summary, our study has shown that the occluded inspiratory effort maneuver (Mueller maneuver) is a reliable procedure for measuring PEEPi,st in patients with severe COPD during spontaneous breathing. This method provides advantages with respect to the traditional end-expiratory occlusion maneuver, and may prove suitable for both diagnostic and therapeutic purposes.