INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The supersyringe method is generally considered to be the reference technique for measuring the static pressure-volume (P-V) curve of the total respiratory system in patients with acute respiratory failure (ARF) (1, 2). One of the drawbacks of this technique is the length of time necessary to perform the entire P-V curve: the inflation maneuver requires 45 to 60 s and is associated with volume loss related to oxygen uptake by the lungs (3, 4). Another drawback of the technique is that it is mandatory to disconnect the patient from the ventilator. These are the reasons why Levy and colleagues (5) developed a static inspiratory occlusions method that can be performed without disconnecting the patient from the ventilator (6). However, this technique requires around 15 min to perform and is cumbersome at the bedside.

An alternative method named "constant-flow method" or "pulse method" has been proposed to determine the respiratory compliance (7). This "dynamic" method is based on the principle that when a constant flow is blown into the lungs, the rate of change of pressure is inversely proportional to the compliance of the respiratory system. During the procedure, two segments can be identified on the airway pressure curve: the first portion with a step shift in the pressure is related to the resistive properties of the respiratory system; the second portion, characterized by a linear increase in airway pressure at a rate inversely proportional to the compliance, is related to the elastic properties of the respiratory system. Suratt and Owens (8) compared the constant-flow method with the static method and demonstrated that compliance values measured with both methods were closely correlated. It has also been shown that dynamic P-V curves obtained from the constant flow method can be used for detecting PEEP-induced overinflation and alveolar recruitment (9). In these studies where a high constant flow varying from 20 to 60 L/min was administered, lower and upper inflection points could not be accurately determined. Recently, Servillo and colleagues (10) used a lower flow of 15 L/min delivered by a Servo ventilator specially equipped with a computerized prototype and compared this modified constant-flow method with the reference inspiratory occlusions method. They demonstrated that the obtained P-V curves were shifted to the right because of the resistive properties of the respiratory system, therefore resulting in an overestimation of lower and upper inflection points.

The aim of this study was to validate a simple, inexpensive and reliable technique for measuring the inflation P-V curve at the bedside without disconnecting the patient from the ventilator. Thoracopulmonary, pulmonary, and thoracic P-V curves obtained after the administration of two constant flows delivered by a conventional ventilator (3 and 9 L/min) were compared with those obtained with the supersyringe and inspiratory occlusion methods in a series of patients with ARF.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Patients

During a 6-mo period, 14 consecutive patients in whom ARF was diagnosed on admission to the Surgical Intensive Care Unit of La Pitié Hospital in Paris (Department of Anesthesiology) were prospectively studied. The protocol was considered as a part of routine clinical practice and no informed consent was obtained from the patients' next of kin. Inclusion criteria were the presence of unilateral or bilateral infiltrates on a bedside chest radiograph associated with a Pa_O2 < 300 mm Hg using an FI_O2 of 1.0 and zero end-expiratory pressure (ZEEP). Exclusion criteria were a previous history of chronic obstructive pulmonary disease (COPD) and the presence of a chest tube with a persistent air leak.

All patients were orotracheally intubated with endotracheal tube no. 8 and ventilated in volume-controlled mode using a constant inspiratory flow (César Ventilator; Taema, France). The César ventilator is equipped with an additional pressure transducer allowing the measurement of "external pressures." In four patients, a no. 8 Hi-Lo Jet Mallinckrodt tube (Mallinckrodt Inc., Argyle, NY) incorporating one side port ending at the distal tip of the endotracheal tube was used. Between the P-V curve measurements, the following ventilatory settings were applied throughout the study (control mechanical ventilation): tidal volume, 10 ml/kg; respiratory frequency, 18 breaths/min; inspiratory:expiratory (I/E) ratio, 33%; PEEP, 0; FI_O2, 1.0. All patients were sedated with a continuous intravenous infusion of fentanyl 250 µg/h, flunitrazepam 1 mg/h and paralyzed with vecuronium 4 mg/h. A fiberoptic thermodilution pulmonary artery catheter (CCO/SvO₂/ VIPTD catheter; Baxter Healthcare Corporation, Irvine, CA) and a radial or femoral arterial catheter already in place for cardiorespiratory monitoring were used for hemodynamic measurements.

Measurements

All measurements were recorded on a Macintosh personal computer (Apple Computer Inc., Cupertino, CA), using a commercially available data acquisition system MP 100 WS (Biopac System, Inc., Goleta, CA) and simultaneously on a strip-chart recorder (Gould ES 1000; Gould Instruments, Cleveland, OH). All variables were sampled on-line by an analog/digital-converter at a rate of 100 Hz except for the constant flow method for which data were sampled at a rate of 10 Hz. All measurements were performed in ZEEP.

Pressure and Cardiorespiratory Measurements

Inspiratory and expiratory flow were measured using a heated and calibrated pneumotachograph (Hans Rudolph Inc., Kansas City, MO) inserted between the Y-piece and the endotracheal tube. Tidal volume was obtained by integration of the flow signal. Airway pressure (Paw) was measured at the proximal tip of the endotracheal tube by means of a connecting piece related to a pressure transducer (PX-1X2; Baxter SA, Maurepas, France). In the four patients intubated with the Hi-Lo Jet Mallinckrodt tube, distal tracheal pressure was also measured by connecting an additional pressure transducer to the side port ending at the distal tip of the endotracheal tube. Esophageal pressure was measured using a water-filled catheter (Argyl; Sherwood Medical, Belgium) connected to a calibrated pressure transducer (PX-1X2; Baxter SA) and positioned at the level of the lower third of the esophagus according to a previously described technique (11). The patients were kept in a half-sitting position in order to minimize the effect of the weight of the mediastinum in the supine position, and the pressure transducer was zeroed at 2 cm above the posterior axillary line at the level of the ninth intercostal space. In the first seven patients, an "occlusion test" was performed before administering muscle relaxants, as recommended (12). In a given patient, three consecutive 5-s occlusions of the tubing connecting the Y-piece and the inspiratory limb were performed while Pes and Paw were simultaneously sampled at a rate of 10 Hz. Individual ratios between changes in Pes and Paw (Pes/Paw), computed over 50 data points per patient, were, respectively, 1.21, 0.95, 1.04, 0.95, 1.16, 0.94, and 1.00, suggesting that the water-filled catheter method was accurate for measuring pleural pressure. Quasi-static respiratory compliance (Cqs) was calculated during mechanical ventilation by dividing the tidal volume by end-inspiratory pressure minus intrinsic PEEP (PEEPi) (13). Total respiratory resistance (Rrs) was measured using the end-inspiratory occlusion technique (14). Rrs was computed by dividing peak inspiratory pressure minus plateau airway pressure by the constant inspiratory flow immediately preceding the end-inspiratory pause. Systemic and pulmonary arterial pressures were simultaneously measured using arterial and fiberoptic pulmonary artery catheters connected to two calibrated pressure transducers (PX-1X2; Baxter SA) positioned at the midaxillary line. All pressures were measured at end-expiration. Arterial pH, Pa_O2, and P _O2 were measured using an IL BGE blood gas analyzer. Hemoglobin and arterial and mixed venous oxygen saturation (Sa_O2 and S _O2) were measured using a calibrated OSM3 hemoximeter. True pulmonary shunt (S/T) was calculated according to standard formulas.

Measurement of Pressure-Volume Curves

Thoracopulmonary, pulmonary, and thoracic P-V curves were obtained in each patient using three different methods: (1) supersyringe method, (2) inspiratory occlusions method, and (3) constant-flow method using two different flows: 3 and 9 L/min.

In order to standardize the lung volume history, each method was preceded by the same sequence of control mechanical ventilation.

The supersyringe method. The patient was disconnected from the ventilator and connected to a specifically designed supersyringe at the end of a 3-s expiration as previously described (15). A 3-L syringe (Model Series 5540; Hans Rudolph) was used for insufflating the lungs with pure O₂ in 100-ml steps until a volume corresponding to a plateau pressure of 30 cm H₂O was reached. Intervals between two steps lasted 3 s.

The inspiratory occlusions method. According to Levy and colleagues (5) and Amato and coworkers (16), P-V curves were obtained by performing "study breath" occlusions at different tidal volumes. Briefly, measurements were made using the end-inspiratory and expiratory pause hold functions of the César ventilator. During control mechanical ventilation and before administering the tidal volume corresponding to a given "study breath," PEEPi was determined by activating the end-expiratory pause knob. Then, the expiratory pause knob was released and control mechanical ventilation was resumed for five cycles. At the end of the expiration of the fifth cycle, the tidal volume corresponding to the "study breath" was set on the ventilator and the inspiratory pause knob was pressed for obtaining a 3-s postinspiratory pause. The same sequence was repeated for each "study breath" corresponding to different tidal volumes. The smallest tidal volume was 100 ml and the highest tidal volume was the tidal volume corresponding to a plateau airway pressure of 30 cm H₂O. Between these two extremes, tidal volumes in 100-ml increments were administered at random. Respiratory frequency was kept constant independently of the studied tidal volume.

The constant-flow method. The new method was performed during volume-controlled mechanical ventilation using two different constant flows: 3 L/min and 9 L/min. The César ventilator is not capable of delivering an inspiratory flow for a period longer than 9.6 s because the minimum respiratory frequency and the maximum I/E ratio that can be set are 5 breaths/min and 80%, respectively. With those minimum ventilatory settings, delivering a constant flow of 3 or 9 L/min over 9.6 s requires the administration of a tidal volume of 500 or 1,500 ml. Therefore, after the administration of five cycles of control mechanical ventilation, the following ventilatory settings were set on the ventilator: I/E ratio of 80%, respiratory frequency of 5 bpm, tidal volume of 500 ml for obtaining a constant flow of 3 L/min or 1,500 ml for obtaining a constant flow of 9 L/min. Simultaneously, the P-V curve diagram was displayed on the screen of the ventilator. With these ventilatory settings, a constant inspiratory flow was administered to the patient for 9.6 s (flow rate: 50 ml/s for 3 L/min, 150 ml/s for 9 L/min), generating a P-V curve on the screen of the ventilator. As shown in Figure 1, the obtained P-V curve was frozen on the screen and control mechanical ventilation was resumed. Two cursors present on the screen were used to determine the lower and the upper inflection points and the slope of the linear portion of the P-V curve (thoracopulmonary compliance). The parameters were calculated automatically by the ventilator after positioning the cursors. The whole maneuver took 2 min at the bedside without requiring any special equipment. Simultaneously, pressures, flows, and volumes were recorded on the Biopac system and the Gould ES 1000 recorder.

View larger version (24K): [in this window] [in a new window]   Figure 1.   Trace recordings obtained from the screen of the César ventilator during a P-V curve maneuver using the 9-L/min constant-flow method. On the right side, it is indicated that the patient is in volume-controlled mode (VAC) using a tidal volume (VT) of 1,480 ml, a respiratory frequency (f ) of 4.9 breaths/min, and an FIO2 of 97%. This is equivalent to administering a constant flow (débit) of 9 L/min for 9.6 s (middle part of the figure). On the left side, the P-V curve is represented allowing the determination of its slope by means of two cursors (+).

Standardization of lung volume history. Lung volume history was standardized between the three methods and for each P-V curve as follows: five tidal volumes of 10 ml/kg were administered at a respiratory frequency of 18 breaths/min using an I/E ratio of 33% and ZEEP (control mechanical ventilation). At the end of the expiratory phase of the fifth tidal volume, the P-V curve was measured by one of the three methods. For the constant-flow method, the airways were occluded at end-expiration of the fifth tidal volume by activating the expiratory pause knob of the ventilator and, simultaneously, the respiratory frequency was turned down to 5 breaths/min, I/E ratio up to 80% and VT to 500 or 1,500 ml. The P-V curve was immediately displayed on the screen of the ventilator when the expiratory pause knob was released. This maneuver allowed a rigid standardization of lung volume history between the three methods, particularly regarding the apnea time preceding the P-V curve.

Analysis of the P-V Curves

For the three methods, thoracopulmonary P-V curves were traced using either absolute values of Paw or Paw minus PEEPi. For the supersyringe method, PEEPi was considered as the positive pressure at zero volume when present; for the constant-flow technique, as the positive pressure measured at zero-flow when present (Figure 2); and for the inspiratory occlusions method, as the PEEPi measured during control mechanical ventilation and before each tidal volume characterizing a given "study breath." Pulmonary and thoracic P-V curves were also computed for the three methods.

View larger version (25K): [in this window] [in a new window]   Figure 2.   Determination of intrinsic PEEP (PEEPi), resistive pressure (Pres), and lower inflection point (Pflex) on the thoracopulmonary P-V curve obtained from the 9-L/min constant-flow method in Patient 7. On the left side of the figure, the 9-L/min constant flow (upper panel ) and corresponding to changes in lung volume (lower panel ) are represented according to increases in airway pressure (Paw) during the P-V curve maneuver. On the right side of the figure, change in flow (upper panel ) and airway pressure (Paw) according to the time are represented. The continuous line corresponds to PEEPi and zero flow and the dashed line corresponds to Pres and the time from which the continuous flow is constant at 9 L/min. PEEPi (Arrow 1) is defined as the increase in Paw corresponding to the change in expiratory flow from its end- expiratory value to zero (Arrow 2); Pres (Arrow 3) related to the constant flow is measured as the airway pressure at which the inspiratory flow becomes constant (Arrow 4) minus PEEPi. Pflex is computed as the pressure corresponding to the intersection between the starting compliance and the slope of the P-V curve (Arrow 5).

For each method and patient, the following P-V curves were constructed: (1) the thoracopulmonary P-V curve by plotting lung volumes against airway pressures; (2) the pulmonary P-V curve by plotting lung volumes against pressure differences between airway and esophageal pressures; (3) the thoracic P-V curve by plotting lung volumes against esophageal pressures. Each P-V curve was transformed by means of a graphic software (Microcal Software, Inc., Northampton, MA) to determine lung volumes corresponding to standardized increments of airway pressure. This transformation gave the possibility of averaging the P-V curves and comparing the three methods according to the same airway pressure scale. For thoracopulmonary and pulmonary P-V curves, lung volumes corresponding to 2.5-cm H₂O increments of airway pressure were computed, whereas for thoracic P-V curves, lung volumes corresponding to 0.5-cm H₂O increments of airway pressure were computed.

The slopes of thoracopulmonary P-V curves corresponding to total respiratory system were determined by linear regression analysis taking into account only values above the lower inflection point. The determination of the value of the lower inflection point was performed on the P-V curves using absolute value of Paw as described by Gattinoni and colleagues (17): the starting compliance was computed as the ratio between the first inflation of 100 ml and its corresponding airway pressure, and, as shown in Figure 2, the lower inflection point (Pflex) was computed as the pressure corresponding to the intersection between the starting compliance and the slope of the P-V curve.

Analysis of the Resistive Component Related to the Constant-Flow Method

The resistive pressure related to the constant flow was analyzed by calculating the initial increase in airway pressure until the inspiratory flow became constant minus PEEPi, defined as the increase in airway pressure corresponding to the decrease in expiratory flow from its end-expiratory value to zero (18) (Figure 2).

In four patients in whom proximal and distal tracheal pressures were recorded simultaneously, the resistive pressure related to the endotracheal tube was calculated as the difference between proximal and distal pressures, the latter being measured at the distal tips of the endotracheal tube at a lung volume corresponding to a pressure of 30 cm H₂O. The flow resistance of the endotracheal tube was calculated by dividing this difference by the constant inspiratory flow (19).

Statistical Analysis

The results were expressed as mean ± SD in the text and in the tables, and as mean ± SEM in the figures. P-V curves among the different methods were compared by calculating the areas under the curves created by the projection of pressures over the volume axis. A one-way analysis of variance for repeated measures was used to compare the areas, the slopes of the P-V curves, and the lower inflection points between the three methods. The level of significance was considered to be 5%.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Among the 11 men and three women (mean age, 61 ± 13 yr) enrolled in the study, two were admitted to the Surgical Intensive Care Unit after multiple trauma, eight after postoperative complications following major surgery, and four after an acute medical disease. At admission, simplified acute physiologic score (SAPS II) was 44 ± 17 and lung injury severity score (LISS) was 2.3 ± 0.8. The main cause of ARF was pneumonia. Mortality rate was 50%. Hemodynamic and respiratory measurements demonstrated arterial hypoxemia (Pa_O2, 171 ± 92 mm Hg), increased true pulmonary shunt (S/T, 39 ± 11%), pulmonary artery hypertension (, 27 ± 11 mm Hg), and reduced respiratory compliance (Cqs, 48 ± 13 ml/cm H₂O). The total respiratory resistance was 4.3 ± 1.1 cm H₂O/L/s. The maximal value of Rrs was 6.2 cm H₂O/L/s. PEEPi was present in the majority of patients. Although different from one method to another in a given patient, PEEPi was not different as a mean between the three methods and was 2.7 ± 0.4 cm H₂O.

Pressure-Volume Curves

As shown in Figure 3, pulmonary and thoracopulmonary P-V curves obtained from the supersyringe, the inspiratory occlusions and the constant-flow method using a 3 L/min flow were superimposable. Thoracopulmonary P-V curves traced by plotting lung volumes either against absolute values of Paw or against the difference between Paw and PEEPi were identical. However, the thoracopulmonary P-V curves obtained from the 9-L/min constant-flow method were slightly shifted to the right with areas under the curves significantly different from the other methods (p < 0.01). The thoracic P-V curves were identical among all methods. In the four patients in whom proximal and distal tracheal pressures were recorded simultaneously, P-V curves traced by plotting lung volume either against proximal or distal Paw were superimposable during the 9-L/min constant flow method (Figure 4).

View larger version (27K): [in this window] [in a new window]   Figure 3.   Thoracopulmonary P-V curves before and after subtraction of PEEPi (upper panel ), pulmonary P-V curves, and thoracic P-V curves (lower panel ) obtained from the supersyringe (open circles), the inspiratory occlusions (open squares), the 3-L/min (closed circles), and the 9-L/min constant-flow methods (open triangles). The P-V curves obtained by the 9-L/min constant-flow method are slightly shifted to the right. The other three curves obtained from the three methods are not significantly different. Paw = airway pressure; Paw  PEEPi = airway pressure minus intrinsic PEEP; PL = change in transpulmonary pressure; Pes = change in esophageal pressure.

View larger version (16K): [in this window] [in a new window]   Figure 4.   Thoracopulmonary P-V curves obtained from the 9-L/ min constant-flow method computed using the proximal (open circles) and the distal airway pressures (open squares). The two curves are not significantly different.

Slopes of the P-V Curves

The individual values of compliance of the respiratory system using the supersyringe, the inspiratory occlusions, and the constant-flow methods are shown in Table 1. The values of compliance of the respiratory system by the three methods and computed by using two constant flows were not significantly different.

Lower Inflection Points

The individual values of the lower inflection point of the respiratory system obtained from the thoracopulmonary P-V curves with the supersyringe, the inspiratory occlusions, and the constant-flow methods are shown in Table 2. The values of Pflex obtained by the supersyringe and the inspiratory occlusions methods were similar. Pflex identified on P-V curves obtained using the 9-L/min constant-flow method were on average 1 cm H₂O higher than those obtained by other methods, but the difference was not significant. In one patient (Patient 8), Pflex was overestimated by 3 cm H₂O when using the 9-L/ min constant-flow method. Pflex identified on P-V curves obtained using the 3-L/min constant-flow method were slightly lower than those obtained by static methods, the difference being statistically insignificant. In one patient (Patient 12), Pflex identified on the thoracopulmonary P-V curve using the 3-L/min constant-flow method was not observed on the P-V curves obtained by the other methods.

Resistive Component Related to the Constant-Flow Method

For the 3-L/min constant flow method, the time required by the ventilator to achieve the constant flow was of 0.11 ± 0.03 s, corresponding to a resistive pressure of 1.0 ± 1.0 cm H₂O. For the 9-L/min constant flow method, the time required by the ventilator to achieve the constant flow was of 0.14 ± 0.05 s, corresponding to a resistive pressure of 1.8 ± 2.1 cm H₂O.

In four patients in whom proximal and distal tracheal pressures were recorded simultaneously, the resistance related to the endotracheal tube appeared flow-dependent. Using a 3-L/ min constant flow, the resistive pressure related to the endotracheal tube was of 0.3 ± 0.2 cm H₂O, corresponding to an added resistance of 5.8 ± 3.8 cm H₂O/L/s, whereas using a 9-L/ min constant flow, the resistive pressure related to the endotracheal tube was of 1.0 ± 0.7 cm H₂O, corresponding to an added resistance of 8.5 ± 6.0 cm H₂O/L/s.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The present study, performed in patients with ARF, has demonstrated that the thoracopulmonary, the pulmonary, and the thoracic P-V curves obtained using the 3-L/min constant-flow method are similar to those obtained with two reference methods, the supersyringe and the inspiratory occlusions techniques. The constant-flow method using a flow of 9 L/min is associated with a slight rightward shift of the P-V curves because of the resistive component. However, the values of the lower inflection point and of the slopes of the P-V curves are not significantly different between the methods. This study has shown that the constant-flow method can correctly assess the elastic properties of the lung and the chest wall without being artifacted by the resistive properties of the endotracheal tube and the respiratory system if a constant flow of 3 L/min is used.

The constant-flow method is a dynamic method aimed at measuring not only static respiratory compliance but also the lower inflection point, which is generally considered as a determinant for optimizing the PEEP level (1, 20). Although the resistive properties of the respiratory system do not influence the slope of the P-V curve, they can markedly interfere with the determination of the lower infection point (10). The ventilator cannot provide a constant flow from the very onset of inspiration (14). The time required to achieve the constant flow corresponds to an immediate step change in Paw because of the resistive component of the respiratory system. During volume-controlled ventilation, the total resistance of the respiratory system is related to respiratory circuits, endotracheal tube, conducting airways, and lung parenchyma. The resistive component related to the endotracheal tube can be easily bypassed by measuring the intratracheal pressure directly. However, the conducting airways, the inhomogeneity of regional time constants within the lung parenchyma, and the viscoelastic behavior of the lung tissue are a permanent source of resistance. As a consequence, the first part of the P-V curve obtained by the constant-flow method can be influenced by the resistive properties of the respiratory system, leading to an overestimation of the inflection point (Pflex) (Figure 2). This phenomenon occurred in Patient 8 in whom an elevated respiratory resistance was observed, resulting in an overestimation of Pflex by the 9-L/min constant-flow method. The resistance of the respiratory system is also related to the level of inspiratory flow (19, 21). Servillo and colleagues (10) demonstrated that the P-V curves were shifted to the right when a constant flow of 15 L/min was delivered by the ventilator. In contrast, Mankikian and colleagues (22) have demonstrated that P-V curves obtained with a constant flow of 1.7 L/min are superimposable with the P-V curves obtained from the supersyringe method. Our results, obtained in a series of patients with normal or slightly elevated respiratory resistance, demonstrate that P-V curves obtained with a constant flow of 3 L/min are similar to those obtained with static methods, whereas those obtained by a constant flow of 9 L/min show a slight rightward shift. These results outline the flow dependence of the resistance of the respiratory system (23). In this study, patients with COPD and those who were bronchospastic were excluded, and further studies are required in order to define the optimal flow rate in patients with elevated airway resistance.

In all patients, the pressures were measured at the proximal tip of the endotracheal tube, therefore eliminating increases in airway pressure related to the resistive properties of the ventilatory circuits. In the four patients in whom pressures were simultaneously measured at the proximal and the distal tips of the endotracheal tube, P-V curves were superimposable when plotting lung volume against either proximal or distal Paw, suggesting that the resistance opposed by an endotracheal tube no. 8 does not result in a significant increase in airway pressure when a constant flow of 9 L/min is administered.

A good agreement was found between the values of the lower Pflex obtained by the three methods as well as by the two levels of constant flow, indicating that the constant-flow method using a flow equal to or less than 9 L/min is not associated with a major flow-dependent increase in resistive pressure. The resistive pressure resulting from the administration of the 9-L/min constant flow method was 1.8 ± 2.1 cm H₂O. If the resistive pressure related to the endotracheal tube is subtracted, the resistive pressure resulting from the administration of the 9-L/min constant flow through the tracheobronchial tree towards the lung parenchyma is less than 1 cm H₂O, a value without clinical relevance.

In order to standardize the lung volume history between the three methods, all P-V curves were performed after five cycles of control mechanical ventilation with a fixed expiratory time of 2.2 s. As a consequence, in the majority of patients, PEEPi was present at the beginning of each procedure. As recommended, PEEPi has to be subtracted from the airway pressure when the compliance is measured by the end-inspiratory occlusion technique (24). With the constant-flow method, the PEEPi was included in the initial portion of the airway pressure and was identified as the increase in airway pressure corresponding to the decrease in expiratory flow from its end-expiratory value to zero (Figure 2). In other words, PEEPi had to be subtracted from the initial increase in airway pressure (18) in order to accurately estimate the resistive pressure related to the constant flow (Figure 2). In clinical practice at the bedside, we recommend a slightly different procedure to get rid of PEEPi: we suggest to reduce first the respiratory frequency to 5 breaths/min without changing the tidal volume, and to wait until the end of the following expiration before increasing inspiratory time and tidal volume. Such a procedure allows a complete emptying of the lungs and suppresses any PEEPi.

The constant-flow method has many advantages over the supersyringe and the inspiratory occlusions methods: (1) there is no need for disconnecting the patient from the ventilator; (2) the construction of the P-V curve on the screen requires 10 s, whereas the whole maneuver, including determinations of the characteristics of the P-V curve, takes less than 2 min; (3) there is no lung volume loss secondary to pulmonary oxygen uptake; (4) the method is easy to implement, inexpensive, and does not require special equipment; (5) the analysis of the P-V curve can be performed at the bedside if the ventilator is equipped with software that allows to display the P-V curve on the screen and to measure the slope and the lower Pflex of the curve using mobile cursors. Moreover, if the ventilator is equipped with an additional pressure transducer that can be connected to the distal trachea, a P-V curve that is not influenced by the resistive properties of the ventilatory circuits and the endotracheal tube can be obtained at the bedside. Although this technique for measuring P-V curve during mechanical ventilation was tested with the César ventilator, it can be implemented with many ICU ventilators equipped with a screen and software that allows to display to analyze the P-V curve.

This new method also has some limitations. First, the inspiratory flow is indirectly preset by setting the I/E ratio, the respiratory frequency, and the tidal volume. Second, obtaining an inspiratory constant flow for a period longer than 10 s is not possible, because the minimum respiratory frequency that can be used in modern ventilators is equal or superior to 5 breaths/ min. As a consequence, the maximal tidal volume that can be administered for tracing the P-V curve is limited to 500 or 1,500 ml for administering a 3-L/min or 9-L/min constant flow, and the upper Pflex cannot be determined with this technique in patients with mild ARF. Further studies enrolling patients with severe ARDS are required to assess the accuracy of the constant-flow method for measuring the upper Pflex of the P-V curve.

Ideally, forthcoming ICU ventilators should be equipped with a flow generator providing constant flows of 3, 6, and 9 L/min that could be administered for 10, 20, and 30 s. They should also have a hold knob that allows the performance of a prolonged expiration before the maneuver, a screen, and software for displaying and analyzing the P-V curve. Rather than multiplying the number of ventilatory modes available on a ventilator, manufacturers should rather focus on respiratory monitoring at the bedside. Rendering the P-V curve easily accessible to the intensivist without disconnecting the patient from the ventilator would certainly be a significant advance in the field of mechanical ventilation. While waiting for these technical developments, the constant-flow method obtained by using nonconventional ventilatory settings for obtaining a low constant flow equal or less than 9-L/min during volume-controlled ventilation is an easy-to-implement, inexpensive, safe, and reliable method for performing P-V curves at the bedside in critically ill patients receiving mechanical ventilation.