INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The assessment of diaphragmatic contractility by measuring twitch transdiaphragmatic pressure (Pdi_tw) in response to phrenic nerve stimulation has the advantage of being independent of patient effort and cooperation (1). Recordings of Pdi_tw, however, require placement of esophageal and gastric balloon catheters, which has limited the clinical application of this approach. In healthy subjects (1) and in patients with respiratory muscle weakness but free of lung disease (2), the less invasive measurement of twitch airway pressure (Paw_tw) can reliably predict twitch esophageal pressure (Pes_tw) and Pdi_tw values.

In patients with chronic obstructive pulmonary disease (COPD), Similowski and colleagues (4) reported good correlations between Paw_tw and Pes_tw, and between Paw_tw and Pdi_tw, when electrical stimulation of the phrenic nerves was superimposed on graded voluntary inspiratory efforts. Unfortunately, this investigation did not clarify a number of important issues such as the identification of a threshold value of Paw_tw below which diaphragmatic weakness can be excluded; satisfactory control of lung volume during graded inspiratory efforts; limitations in the use of electrical stimulation of the phrenic nerves; and the limits of agreement (7) between Paw_tw and Pes_tw and between Paw_tw and Pdi_tw. Moreover, electrical stimulation of the phrenic nerves is painful, and it can be difficult to locate the phrenic nerves and avoid twitch potentiation during repeated stimulations (8). Magnetic stimulation of the phrenic nerves has advantages over electrical stimulation, in that it is easy to employ and is well tolerated and reproducible; this makes it an attractive method for the assessment of diaphragmatic contractility in both healthy subjects (8) and patients (2, 5).

Because of concerns with equilibration between esophageal pressure (Pes) and airway pressure (Paw) in patients with severe COPD and the limitations of electrical stimulation, we investigated whether it is possible to develop a simple method whereby measurements of Paw_tw could be used to predict Pes_tw and Pdi_tw elicited by magnetic stimulation of the phrenic nerves in patients with severe COPD.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Twelve men 53 to 76 yr of age, with stable, severe COPD were studied (Table 1). The study was approved by the local ethics committee. Informed consent was obtained from all patients.

Pes and gastric pressures (Pga) were separately measured with two thin-walled, latex-balloon-tipped catheters coupled to pressure transducers. Transdiaphragmatic pressure (Pdi) was obtained by subtraction of Pes from Pga. A flanged mouthpiece was connected to a cylinder (8 cm in length and 15 mm in internal diameter) that could be occluded by a valve located at its distal end. Proximal to the valve, a steel tube (4 cm in length and 1 mm in internal diameter) was passed through and anchored to the wall of the cylinder; this tube had the purpose of preventing glottic closure. Paw was sensed at the cylinder with a pressure transducer. A pneumotachograph was attached to the cylinder by tubing (20 cm in length and 5 mm in internal diameter), and the flow signal was displayed to the patient.

Compound diaphragmatic action potentials (CDAPs) were recorded bilaterally with two sets of surface electrodes at the level of the 7th and 8th intercostal space and the anterior axillary line. Bilateral phrenic nerve stimulation was performed using two magnetic stimulators (Magstim 200; Jali Medical Inc., Newton, MA) with two sets of double 40-mm coils (D40-1183.00) that generated a magnetic field of 3.2 Tesla at maximal output. The area of stimulation associated with the CDAP of greatest amplitude was located by moving the stimulating probe around the posterior border of the sternomastoid muscle at the level of the cricoid cartilage.

Study Protocol

Experiment 1. The purpose of this experiment was to determine the relationship between Paw_tw and Pes_tw, and between Paw_tw and Pdi_tw, when patients relaxed at FRC with a mouthpiece and noseclip in place. To avoid twitch potentiation (8), a rest period of 20 min preceded the first stimulation. Eight to 10 stimulations were delivered while the valve connecting to the mouthpiece was occluded.

Experiment 2. The purpose of this experiment was to determine the relationship between Paw_tw and Pes_tw, and between Paw_tw and Pdi_tw, during a gentle exhalation starting from FRC. At each point that Paw had reached + 5 cm H₂O, the phrenic nerves were stimulated (2), and this process was repeated five to six times.

Experiment 3. The purpose of this experiment was to determine the relationship between Paw_tw and Pes_tw, and between Paw_tw and Pdi_tw, during a gentle inhalation starting from FRC. Patients were instructed to initially exhale to FRC, and, after closure of the valve of the mouthpiece, to inhale gently. At each point that an inspiratory flow of 40 ml/s was reached, the phrenic nerves were stimulated (3), and this process was repeated five to six times.

Signal Processing and Data Analysis

Twitch pressures were measured as the difference between the maximal pressure displacement secondary to phrenic nerve stimulation and the value of each pressure signal immediately after stimulus delivery. When pressure tracings were biphasic, twitch pressure was calculated as the difference between the value immediately after stimulus delivery and the maximal excursion (either positive or negative) after stimulation. Individual twitch responses were rejected from analysis according to previously described criteria (see Table 2).

All data are reported as the mean and standard deviation (mean ± SD). Linear regression analysis and Bland-Altman plots (7) were performed when appropriate.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Phrenic Nerve Stimulation at Relaxed FRC (Experiment 1)

In almost all instances Paw_tw and Pes_tw were poorly matched, probably because of variable glottic closure (Table 3 and Figures 1 and 2), and Paw_tw was smaller than Pes_tw (Figure 1, upper left panel ).

View larger version (11K): [in this window] [in a new window]   Figure 1.   Twitch airway pressure (Pawtw) (dotted lines) and twitch esophageal pressure (Pestw) (solid lines) recordings after magnetic stimulation of the phrenic nerves in four representative patients with severe COPD. (Upper left panel ) At relaxed FRC, Patient A demonstrated a fall in Pestw that was accompanied by a small decrease in Pawtw. (Upper right panel ) During a gentle inspiratory effort (inspiratory flow rate ~ 40 ml/s) from FRC, the Pawtw signal in Patient B closely followed the Pestw profile. (Bottom panels) During gentle exhalations (Paw + 5 cm H2O) from FRC, the Pawtw signal closely followed the Pestw profile in Patient C, but the two signals displayed a considerable discrepancy in Patient D.

View larger version (14K): [in this window] [in a new window]   Figure 2.   Twitch airway pressure (Pawtw) versus twitch esophageal pressure (Pestw) (left panel ), and Pawtw versus twitch transdiaphragmatic pressure (Pditw) (right panel ) in response to phrenic nerve stimulation at relaxed FRC in 12 patients (106 data points). The correlations between Pawtw and Pestw (r = 0.44, p < 0.001) and between Pawtw and Pditw (r = 0.52, p < 0.001) were significant but weak, probably because of varying degrees of glottic closure, increased upper airway compliance, and increased airway resistance.

Phrenic Nerve Stimulation during Gentle Exhalation from FRC (Experiment 2)

Phrenic nerve stimulation elicited positive Paw_tw values in two patients (Patients 1 and 6 in Table 3). Because assessment of inspiratory activity of the diaphragm was the aim of the study, the data in these patients were excluded in the calculation of mean values and correlation coefficients. The relationship between Paw_tw and Pes_tw (Figure 1, bottom panels; and Figure 3) and between Paw_tw and Pdi_tw was variable (Figure 3).

View larger version (12K): [in this window] [in a new window]   Figure 3.   Twitch airway pressure (Pawtw) versus twitch esophageal pressure (Pestw) (left panel ), and Pawtw versus transdiaphragmatic pressure (Pditw) (right panel ) in response to phrenic nerve stimulation during gentle exhalation (Paw + 5 cm H2O) from FRC in 10 patients (57 data points). The correlations between Pawtw and Pestw (r = 0.62, p < 0.001) and between Pawtw and Pditw (r = 0.33, p = 0.01) were significant but weak, probably because of increased upper airway compliance and expiratory flow limitation.

Phrenic Nerve Stimulation during Gentle Inhalation from FRC (Experiment 3)

Phrenic nerve stimulation elicited positive Paw_tw values in two patients (Patients 1 and 12 in Table 3), which were excluded from analysis. The relationship between Paw_tw and Pes_tw was good in most instances (Figure 1, upper right panel; and Figure 4), but not between Paw_tw and Pdi_tw (Figure 4). The bias between Paw_tw and Pes_tw (the mean of the difference between Paw_tw and Pes_tw) was 1.2 cm H₂O, and the limits of agreement (bias ± 2 SD) ranged from 3.7 to 6.0 cm H₂O (Figure 4, bottom panel ).

View larger version (21K): [in this window] [in a new window]   Figure 4.   Twitch airway pressure (Pawtw) versus twitch esophageal pressure (Pestw) (upper left panel ) and Pawtw versus transdiaphragmatic pressure (Pditw) (upper right panel ) in response to phrenic nerve stimulation during gentle inhalation (inspiratory flow rate ~ 40 ml/s) from FRC in eight patients (46 data points). The correlation between Pawtw and Pestw was good (r = 0.92, p < 0.001), probably because of maintenance of glottic patency, a decrease in upper airway compliance, and more uniform pleural pressure distribution during the inspiratory maneuver. The correlation between Pawtw and Pditw was relatively weaker (r = 0.61, p < 0.001). (Bottom panel ) Bland-Altman plot of the difference between Pawtw and Pestw versus the mean of Pawtw and Pestw. The bias, i.e., the mean of the difference between Pawtw and Pestw (solid line), was 1.2 cm H2O; the limits of agreement, i.e., bias 2 SD to bias +2 SD (dashed lines), were 3.7 to 6 cm H2O.

During each protocol, many patients displayed initial positive deflections in both Paw and Pes (Figure 1), probably because of magnetic stimulation of the expiratory rib cage muscles (8). Some patients displayed a positive value of Paw_tw accompanied by a positive value of Pes_tw of lesser magnitude (Figure 1, upper left panel, and Table 3), probably in part because of magnetic coactivation of the upper airway musculature with consequent displacement of a column of air towards the mouth (9).

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

While patients with severe COPD performed a gentle inhalation starting from FRC, stimulation of the phrenic nerves resulted in a strong correlation between Paw_tw and Pes_tw. When stimulations were delivered at relaxed FRC or during a gentle exhalation from FRC, the correlation between Paw_tw and Pes_tw was weak. During all maneuvers, the correlation between Paw_tw and Pdi_tw was weak.

The finding that phrenic nerve stimulations delivered at relaxed FRC resulted in only a weak correlation between Paw_tw and Pes_tw is not surprising (3) and is thought to result from glottic closure during phrenic nerve stimulation. Such glottic closure can be avoided by instructing the subjects to inhale to total lung capacity and then exhale passively through an expiratory resistance to FRC (1). With this maneuver, however, lung volume is changing continuously and twitch potentiation is unavoidable (8), limiting its applicability in clinical practice. In healthy subjects and in patients suspected of respiratory muscle weakness (but free of lung disease), Hamnegård and colleagues (2) found it possible to predict Pes_tw from Paw_tw obtained from stimulations delivered during a gentle expiratory effort without inducing twitch potentiation. Unlike Hamnegård and colleagues, we did not observe a satisfactory correlation between Paw_tw and Pes_tw during gentle exhalation (Figure 3). The different experiences in the two studies probably relate to our selection of patients with severe COPD, who invariably have varying degrees of expiratory flow limitation. Consequently, the transmission of intrathoracic pressure to the upper airway was likely impaired in our patients.

Similar to our findings in healthy subjects (3), the patients with severe COPD demonstrated a good correlation between Paw_tw and Pes_tw during gentle inhalation (Figure 4). We speculate that this result was partly due to the avoidance of glottic closure during the stimulation, and partly because of the decrease in upper airway compliance during the gentle inspiratory effort. As expected, the values of Pdi_tw during gentle inhalation were less than the values obtained at FRC (15.4 ± 9.5 versus 20.1 ± 6.9 cm H₂O), reflecting the increase in lung volume during the inhalation maneuver (3).

Contrary to the expectation that Paw_tw should be of equal magnitude to Pes_tw when glottic patency is maintained or less negative than Pes_tw when some degree of glottic closure occurs, Paw_tw was more negative than Pes_tw during at least two maneuvers in two patients: in Patient 12 at FRC and during gentle exhalation, and to a lesser extent in patient 5 during all three maneuvers (Table 3). This finding contrasts with the experience of Yan and colleagues (1) who noted that Paw_tw is virtually always less negative than Pes_tw. The discrepancy between the studies may be due to the employment of different stimulation techniques. Electrical stimulation of the phrenic nerves, employed by Yan and colleagues induces isolated contraction of the diaphragm with consequent outward motion of the lower rib cage and inward motion of the upper rib cage. The deformation of the upper rib cage causes dissipation of the pressure swing, and, thus, the swings in the Pes are greater in the region of the diaphragm than in the upper thorax. In contrast, magnetic stimulation causes contraction of both the diaphragm and the rib cage muscles (8). Upper rib cage paradox is not observed, and, indeed, this region tends to move outward (8). As a result, the cranio-caudal gradient in Pes excursions observed during isolated diaphragmatic contraction (1, 10) is likely diminished during magnetic stimulation. Moreover, contraction of the upper rib cage muscles contributes to the generation of intrathoracic pressure (6). Because Paw reflects the global change in intrathoracic pressure generated by contraction of both the diaphragm and upper rib cage muscles (1), the value of Paw_tw elicited by magnetic stimulation can be more negative than Pes_tw. Finally, hyperinflation can displace the diaphragm to such an extent that it functions as an expiratory, rather than as an inspiratory, muscle. Interestingly, the patients in whom Paw_tw was consistently more negative than Pes_tw (Patients 5 and 12) were two of the three patients in whom the ratio of swings in Pga to Pes (Pga/Pes) was positive during resting breathing. The Pga/Pes ratio reflects the diaphragmatic contribution to tidal breathing, and a positive value indicates an inefficient inspiratory action of the diaphragm (5).

Similowski and colleagues (4) have reported good correlations between Paw_tw and Pes_tw performed during graded voluntary inspiratory maneuvers in patients with COPD. A correlation coefficient, however, measures the strength of a relationship between two variables, not the agreement between them (8). We also observed a good correlation between Paw_tw and Pes_tw during gentle inhalation, but the limits of agreement were wide, making prediction of Pes_tw from Paw_tw measurements unreliable (Figure 4).

The weak correlation between Paw_tw and Pdi_tw contrasts with previous reports (2, 3), and especially with the experience during a gentle inhalation maneuver (3). This discrepancy can be explained by the greater scatter in the Pes_tw/Pdi_tw ratios in the patients with severe COPD than in healthy subjects: the respective intersubject coefficients of variation were 64% (current investigation) and 11% (3). The wide intersubject variability in Pes_tw/Pdi_tw ratios may be due to differences in rib cage distortability between patients with severe COPD and healthy subjects (11). This difference could be the result of variable recruitment of rib cage muscles by magnetic stimulation (8) and/or to differences in the intrinsic characteristics of the rib cage of patients as compared with healthy subjects. A wide intersubject variability in Pes_tw/Pdi_tw ratios implies that a given value of Pes_tw (the only component of Pdi_tw that has potential for transmission to the upper airway) is associated with a wide range of Pdi_tw values among different patients. Accordingly, reliable prediction of Pdi_tw from measurement of Paw_tw is impossible in patients with severe COPD.

Areas of clinical practice in which Paw_tw might prove useful include the screening of patients suspected to have neuromuscular weakness and the monitoring of changes in diaphragmatic contractility (e.g., development of fatigue in patients experiencing an acute illness, or improvement in contractility after a therapeutic intervention). Contrary to the experience in healthy subjects (2, 3), we could not identify a breakpoint in the values of Paw_tw during gentle inhalation (which ranged from 15 to 1 cm H₂O) below which respiratory muscle weakness could be comfortably excluded. This should not necessarily lead us to discard Paw_tw as a screening tool. On post-hoc analysis, Paw_tw values more negative than 7 cm H₂O were always associated with more negative values of Pes_tw (Figure 4). This suggests that whenever a patient with severe COPD displays a Paw_tw more negative than 7 cm H₂O during gentle inhalation, significant neuromuscular compromise of the diaphragm may be excluded. A Paw_tw value less negative than 7 cm H₂O suggests that the patient's diaphragm is weak or fatigued, although it could be a normal variant or caused by excessive "noise" in the Paw_tw-Pes_tw relationship (Figure 4, bottom panel ). Accordingly, when Paw_tw is less negative than 7 cm H₂O, esophageal and gastric balloons should be employed to determine whether the patient truly has diaphragmatic dysfunction. For the purpose of monitoring change in diaphragmatic contractility, Paw_tw might prove useful in patients with a baseline Paw_tw of 7 cm H₂O or less. For both of these clinical applications, it is essential that Paw_tw measurements exhibit excellent reproducibility over time in a given patient with severe COPD.

In conclusion, although twitch airway pressures elicited during a gentle inspiratory effort in patients with severe COPD achieved a close correlation with simultaneous measurements of twitch esophageal pressures, the precision of the measurement was not sufficient to allow reliable prediction of twitch esophageal pressure in all patients.