INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Breathing with a constant load at a maintained breathing cycle increases respiratory effort sensation (RES) over time (1, 2). The reason for the increased RES when a given diaphragmatic contraction pattern is maintained is unclear. Some studies have suggested that increased RES occurs secondary to increases in central respiratory drive and increased activation of the respiratory muscles required to compensate for developing weakness and/or fatigue (1, 2). However, no controlled evaluation of this hypothesis exists. Breathing with a constant pattern and increased load also decreases the power spectrum center frequency of the diaphragm (CFdi) electrical activation (EAdi) (3, 4).

Therefore, the aims of the present study were, first, to determine if breathing with a fixed pattern at a constant level of diaphragm activation increases RES over time. Second, we wanted to evaluate whether breathing with constant activation at different levels of maintained pressure generation produced corresponding changes in RES and CFdi over time.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Six healthy subjects (five male and one female) aged 36 ± 7 yr were tested after giving their informed consent. Five subjects had previous experience with diaphragm function testing, whereas one subject had no previous experience. Five of the six subjects were naive to the purpose of the study. The experimental protocol was approved by the scientific and ethical committees of the hospital.

Experimental Set-up

Electrical activity of the crural diaphragm was obtained via a nasogastric catheter with nine electrodes (2 mm wide and 2 mm in diameter), placed 10 mm apart on silicone tubing (2 mm diameter), creating an array of eight sequential electrode pairs. Two teflon tubes (diameter = 0.75 mm), placed inside the silicone tubing, connected two differential pressure transducers (Sensym Inc., Milpitas, CA; ± 350 cm H₂O) to each of two 5-cm-long, 1.5-cm-diameter latex balloons mounted 5 cm below the most distal electrode and 2 cm above the most cephalad ring to allow for measurement of gastric (Pga) and esophageal pressures (Pes), respectively. Pdi was obtained in the conventional manner (Pga  Pes). Inspiratory flow was measured with a heated pneumotach (#2; Hewlett Packard, Palo Alto, CA). Rib cage and abdominal displacements (Respitrace; Ambulatory Monitoring, Ardsly, NY) were plotted as Konno and Mead (5) plots. Oxygen saturation (Sa_O2, Ohmeda, Tewksbury, MA) and end-tidal CO₂ pressure (PET_CO2, infrared CO₂ analyzer; Puritan Bennet, Carlsbad, CA) were also measured.

Data Acquisition

Diaphragm electrical signals were amplified (INA102; Burr-Brown, Tucson, AZ), high-pass filtered at 10 Hz (6 dB/octave), and low-pass filtered at 1 kHz (80 dB/octave; Frequency Devices, Haverhill, MA). All signals were A/D converted (Data translation 2839; Marlboro, MA) with 12-bit resolution and stored on hard disk for off-line analysis. EAdi signals were sampled at 2 kHz, whereas all other parameters were sampled at 125 Hz.

Signal Processing

The segments of EAdi used in the analysis were automatically selected between 50% and 75% of the electrocardiogram (ECG) R-R interval (6). DC levels and trends were removed by linear regression analysis; the tails of the raw signals were zero padded from the first and last zero crossings to fit the segments for a fast Fourier transform of 1,024 points. The relative position of the center of the electrically active region of the contracting crural diaphragm (EARdi center) with respect to the electrode array was determined (7, 8). The "double subtraction technique" was then applied to minimize the effects of electrode filtering and electrode-to-muscle distance, and to enhance the signal-to-noise ratio (8). The double subtracted signal was then converted into the frequency domain by fast Fourier transform. Signal quality was evaluated from the power spectrum according to Sinderby and coworkers (6), and the following inclusion criteria were used: signal-to-noise ratio  15 dB, signal-to-motion artifact ratio  12 dB, drop in power density of the spectrum ratio  30 dB, and spectral deformation index  1.4. For those signals deemed to be of acceptable quality, we calculated EAdi (as the root mean square [RMS]) and CFdi (6). Note that in the current study, CFdi was analyzed only when EAdi was voluntarily increased to and maintained at a high steady level during both the volume and pressure maneuvers such that quality indexes were always satisfied. To avoid influence of power spectral shifts on RMS (9), we calculated RMS on the spectral moment of order 1 (M₁), which is insensitive to conduction velocity changes (10).

Protocol

Respitrace bands were positioned on the subjects and secured in place. The esophageal catheter was passed through the nose, swallowed, and the electrodes were positioned at the level of the gastroesophageal junction with feedback from an on-line display. This position of the catheter placed the Pes and Pga balloons in the lower third of the esophagus and in the stomach, respectively. Correct positioning of the balloons was confirmed by an occlusion test (11) and Valsalva maneuver.

While seated in an upright chair, facing a computer monitor with visual feedback, subjects performed the following maneuvers: (a) a combined expulsive and Mueller maneuver (12) to obtain a maximum voluntary Pdi (Pdi_max) and (b) a maximal inspiration to total lung capacity (TLC) to obtain a voluntary maximal EAdi value (13). At least three reproducible maneuvers were required. Pdi_max was obtained at either end-expiratory lung volume for the pressure maneuver protocol (see below) or at the lung volume attained during the targeted level of EAdi in the volume protocol (see below). The Pdi_max values obtained at the given lung volume where each maneuver was performed were subsequently referred to as Pdi_{max @LGT}.

Subjects were asked to perform 10 min periods of intermittent contractions consisting of 10 s of near-isometric contractions followed by a 7-s relaxation period where free breathing was allowed. The duty cycle was imposed by a sound signal. Subjects did not breathe during the near-isometric contraction period, but the relaxation period allowed enough time of unloaded breathing to avoid hypoxia and hypercapnia. To obtain two different Pdi levels for the same target EAdi, subjects were instructed to perform two different types of maneuvers: the pressure maneuver and the volume maneuver.

Pressure maneuver protocol. Subjects were requested to perform a pressure maneuver consisting of either an expulsive maneuver (i.e., Pdi generated mainly by gastric pressure) as depicted in the left panel of Figure 1, or by a Mueller manever (i.e., Pdi generated mainly by esophageal pressure), which is demonstrated in the middle panel of Figure 1, in order to reach a target level of diaphragm activation. Both maneuvers were performed at end-expiratory lung volume. Subjects sustained their targeted level of EAdi for 10 s, followed by 7 s of a few unloaded breaths (Figure 1, left and middle panel).

View larger version (98K): [in this window] [in a new window]   Figure 1.   Tracings of tidal volume (VT), diaphragm electrical activity (EAdi), transdiaphragmatic pressure (Pdi), esophageal pressure (Pes), and gastric pressure (Pga) during the "pressure" and "volume" maneuvers performed in the study. Pressure maneuvers consisted of either an expulsive or a Mueller maneuver.

Volume maneuver protocol. Subjects were instructed to inspire to a lung volume that produced the target level of EAdi (Figure 1, right panel). With open glottis, the inspiratory effort was then maintained for 10 s at this lung volume. Free breathing was allowed for 7 s between the volume maneuvers.

All maneuvers were performed with online feedback of EAdi and mechanical parameters. The targeted level of EAdi was 40% of the maximum EAdi. Subjects were allowed to rest 15 min before the main test started. Both the pressure and volume maneuvers were tested on the same day with 30 min of rest allowed between the two. Half of the subjects started with the pressure maneuver, and the remainder with the volume maneuver. A schematic description of the protocol is presented in Figure 2. Four subjects performed the pressure maneuver with the expulsive maneuver and two subjects with the Mueller maneuver.

View larger version (29K): [in this window] [in a new window]   Figure 2.   Schematic description of the protocol. Pdimax @LGT = maximal Pdi obtained during maximal activation at the corresponding lung volume. EAdi = diaphragm electrical activation.

During both periods in which pressure and volume maneuvers were performed, RES was evaluated using a modified Borg scale (14). Subjects were asked to rate RES by pointing to the Borg scale at 0, 3, 6, and 9 min into the runs.

Statistical Analysis

Comparisons across conditions or between parameters were performed using the repeated measures ANOVA. Pairwise post-hoc comparisons were made using the Tukey test. Association between parameters was tested using linear regression analysis. Data are presented as mean values ± standard deviations (SD).

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The inspiratory volume during the volume maneuver averaged 71.2 ± 14.1% of the IC. Qualitative analysis of the chest wall configuration (by observation of Konno and Mead plots) indicated outward abdominal and rib cage displacements in all subjects during the volume maneuver, with respect to the FRC configuration. During the pressure maneuver, the chest wall configuration did not change in Subjects 1, 3, 5, and 6; abdominal circumference decreased slightly in Subject 2 and increased modestly in Subject 4, with respect to the FRC configuration.

All subjects were able to keep their imposed duty cycle (0.59 ± 0.05 and 0.59 ± 0.03, p = 0.872) and targeted EAdi (39.0 ± 3.5 and 38.8 ± 4.7% of maximum, p = 0.950) during the volume and pressure maneuvers, respectively. The PET_CO2 decreased by 0.6 ± 0.5% (p = 0.035) and 0.8 ± 0.8 (p = 0.068, one-way ANOVA) from the beginning to the end of the volume and pressure maneuvers periods, respectively. Sa_O2 was normal (> 96%) during all runs.

As depicted in the left panel of Figure 3, the Pdi_{max @LGT} obtained at FRC was 185.0 ± 47.9 cm H₂O and was reduced to 72.7 ± 18.5 (p = 0.010, one-way ANOVA) at the lung volume corresponding to that attained during the volume maneuver (i.e., mean of 71% of IC). During the pressure maneuver, Pdi was 55.0 ± 22.7 cm H₂O whereas it was lower (21.4 ± 5.2, p = 0.010, one-way ANOVA) during the volume maneuver (left panel of Figure 3). As shown in the right panel of Figure 3, the relative difference in the Pdi observed between the pressure and volume maneuver was similar to that observed for Pdi_{max @LGT} at the two lung volumes, suggesting an activity-independent length-tension relationship.

View larger version (14K): [in this window] [in a new window]   Figure 3.   Effect of inspiratory volume on transdiaphragmatic pressure (Pdi) during maximal and submaximal voluntary activation. The left panel shows the target submaximal Pdi (in cm H2O) obtained at the target level of diaphragm electrical activation (EAdi) as well as the maximal Pdi (obtained during maximal activation at the two corresponding lung volumes, i.e., Pdimax @LGT) plotted versus inspiratory volume. In the right panel the Pdi values obtained during maximal and targeted submaximal voluntary activation levels are expressed as percentage of their respective FRC values. Note that although the reductions in Pdi at maximal or submaximal diaphragm activation levels were different in absolute terms (left panel), when expressed as a fraction of the Pdi obtained at functional residual capacity (FRC), the reductions were similar.

Pdi values did not decrease significantly from the beginning to end of both the volume and pressure maneuvers, the observed changes being 3.4 ± 9.3 cm H₂O (p = 0.410 one-way ANOVA) and 5.9 ± 8.7 cm H₂O (p = 0.158 one-way ANOVA), respectively.

As depicted in Figure 4, whereas the mean CFdi decreased significantly with time during both pressure (circles) and volume (triangles) maneuvers (p < 0.001 for both), the observed decrease was larger (p = 0.003, two-way ANOVA) during the pressure maneuver than during the volume maneuver. The RES increased with time during both the pressure and volume maneuvers (p < 0.001, two-way ANOVA) and attained higher values during the pressure maneuver compared with the volume maneuver (p = 0.008, two-way ANOVA). The mean values for the CFdi and RES during both the volume and pressure maneuvers are shown in the right panel of Figure 4. The superposition of the two relationships indicates that the changes in CFdi were related to the changes in RES independent of the length and Pdi at which the contractions were performed. Figure 5 demonstrates the relationship of CFdi and RES obtained during both the volume and pressure maneuvers in each subject studied.

View larger version (18K): [in this window] [in a new window]   Figure 4.   Left panels show group mean values with SD for respiratory effort sensation (RES) and diaphragm EAdi power spectrum center frequency (CFdi) at 0, 3, 6, and 9 min during the pressure (circles and solid line) and volume maneuvers (triangles and dashed lines). The right panel demonstrates the relationship of changes in RES to changes in CFdi over time during both maneuvers.

View larger version (23K): [in this window] [in a new window]   Figure 5.   The relationship of changes in respiratory effort sensation (RES) to changes in diaphragm EAdi power spectrum center frequency (CFdi) over time during the pressure (filled circles and crossed symbols represent expulsive and Mueller maneuvers, respectively) and volume (open circles) maneuvers in the six subjects studied. Symbols are values obtained at 0, 3, 6, and 9 min into the test period of the respective maneuvers.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The present study introduces two new concepts important in the understanding of RES. First, it shows that RES can increase over time, to a level described as "severe," while breathing with a fixed breathing pattern, constant EAdi, and maintained level of Pdi. Second, the progressive increase in RES was closely associated with a progressive decrease in CFdi, independent of the level of diaphragm pressure generation.

It is currently accepted that RES can be affected over time by changes in respiratory drive, breathing pattern, pressure generation, weakness, fatigue, and blood gases (15). Throughout each test period in the present study, breathing pattern and EAdi were successfully maintained by each subject (i.e., respiratory drive was constant) and PET_CO2 and Sa_O2 were constant, which precludes their possible involvement in the increased RES over time. Given that Pdi remained constant throughout each test period, it is unlikely that afferent feedback from mechanoreceptors contributed to the increase in RES observed over time during both the volume and pressure maneuvers.

With respect to the development of fatigue and RES, Laghi and coworkers (16) demonstrated that the time course of RES was not different during the first 2 min of repeated breathing periods with the same inspiratory loading, despite a progressive development of contractile fatigue. This seems to indicate that during early stages of loaded breathing, the RES is not related to the absolute degree of contractile fatigue. With contractile fatigue, the EAdi (i.e., the neural drive) must increase in order to maintain a constant Pdi, when the recruited motor units become weaker. In the present study, diaphragm activation was constant and Pdi did not decrease throughout the test periods, suggesting that diaphragm contractile function was preserved, and by extension the observed changes in RES over time were therefore unlikely to be due to fatigue.

Another explanation for the increased RES over time is that development of neuromuscular transmission failure would require an increased central nervous system output in order to recruit additional motor units necessary to maintain a constant EAdi. That the increased RES perceived over time, under these circumstances, may be influenced by the increased central nervous system output cannot be excluded. Existing evidence suggests that transmission failure does not occur prior to contractile diaphragm fatigue in humans (17, 18) and animals (19) during loaded breathing. Furthermore, the EAdi level of 40% targeted in the present study corresponds to the level that is maintained during resting breathing in patients with chronic respiratory insufficiency (13). Therefore, the likelihood that neural transmission failure contributed significantly to the increased RES over time in the present study is remote. Hence, it is unlikely that factors such as as respiratory drive, breathing pattern, pressure generation, weakness, fatigue, and blood gases caused RES to increase over time in the present study.

A very interesting finding of the present study was that changes in RES over time were strongly correlated with changes in CFdi. The finding that RES was higher (at comparable times) during the pressure maneuver, where a higher Pdi was generated, confirms the previously described involvement of mechanoreceptorssensitive to changes in forceon RES (15). However, the fact that CFdi was also lower (at comparable times) during the pressure maneuver suggests a common factor other than mechanoreceptors likely influenced the former. Diaphragm contractions with a high Pdi can be expected to hinder blood flow relatively more than contractions producing a low Pdi (22, 23). Mortimer and coworkers (24) showed that impaired blood flow (independent of oxygen availability) to the human biceps muscle resulted in an electromyogram (EMG) power spectrum shift to lower frequencies. Later, Körner and coworkers (25) showed that shifts toward lower frequencies in the EMG power spectrum, induced by isometric contractions of the biceps brachii, could be reversed only when intramuscular pressure fell below 20 mm Hg. It could therefore be suggested that Pdi is a common factor that affects both RES and CFdi. Although this may explain why CFdi was reduced and RES was increased with the increased Pdi at comparable times, it cannot explain why CFdi decreased and RES increased proportionally over time.

There are strong reasons to assume that changes in CFdi, observed in the present study, represented changes in muscle fiber action potential conduction velocity (APCV). The muscle fiber APCV depends on the cable properties of the fiber, which remain relatively stable during muscle contractions (26), and the membrane excitability (27, 28). The latter depends on ion gradients across the membrane generating the driving electric force and the properties of the proteins making up the gating ion channels. If the electrolyte imbalance across the sarcolemma leads up to a slight depolarization of the membrane (e.g., with increased activation and/or reduced blood flow), inactivation of the voltage-dependent sodium channels occurs, reducing membrane excitability (27, 28) and slowing propagation of the action potentials (29). Proportionality between APCV and the power spectrum frequency shift has been theoretically described in detail (7, 9) and the CFdi has been demonstrated to reflect the mean APCV in the canine diaphragm (26). In addition to APCV, the EAdi power spectrum is also influenced by motor unit territory, number of fibers in the motor unit, dispersion in arrival times of the single contributions in the motor unit signal, and dispersion in action potential conduction velocities between motor units. However, these influences are minor in healthy muscles (9). It is therefore possible that once the capacity of the Na⁺/K⁺ pump, which regulates the transmembrane electrolyte gradient, is exceeded (30), and/ or when the blood flow is reduced, a progressively altered chemical milieu of the diaphragm fiber can be expected over time, which should be associated with a progressive decline in CFdi, although activation and Pdi remain constant.

With respect to RES, it is interesting to note that blood electrolyte changes observed during exercise can promote changes in ventilation (31) by stimulating K⁺ sensitive cells located in the carotid body (32). Changes in blood electrolyte concentration were not measured in the present study; however, considering the muscle mass and the activation of the diaphragm, it is not likely that systemic plasma (K⁺) would increase to a level that would be induced by exercise. Maybe more relevant, Supinski and coworkers (33) demonstrated that injection of potassium (K⁺) directly into the diaphragm circulation in dogs had an excitatory effect on diaphragm activation, which was mediated by phrenic nerve pathways. In the present study, the changes in CFdi suggest that an electrolyte imbalance may have occurred in the diaphragm. Although Supinski and coworkers (33) interpreted their findings as evidence for a phrenic-to-phrenic reflex loop, it does not exclude the possibility that the same afferent pathways also can modulate RES in conscious subjects.

Another possible link between RES and diaphragm electrolyte imbalance is obtained by the fact that dyspnea is improved in patients with COPD following inhalation of ₂-agonists despite poor evidence of an improved FEV₁ (34, 35). The ₂-agonist salbutamol has been demonstrated to be very potent in stimulating the Na⁺/K⁺ pump (36), restoring ionic gradients across the sarcolemma, and thereby improving fiber membrane excitability and muscle contractility (30). During inspiratory threshold loading in healthy subjects, the ₂-agonist fenoterol increased twitch Pdi, increased duration to task failure, reduced the inspiratory effort sensation, and reduced the decrease in CFdi (37). No change in FEV₁, VC, or FRC was observed after inhalation of fenoterol (37). Similarly, when patients with COPD performed resistive breathing after inhalation of the ₂-agonist albuterol, they also experienced a reduction in inspiratory sense of effort, despite a clinically insignificant increase in FEV₁. Based on findings of related increases in IC and twitch Pdi, the authors (35) attributed their results to an albuterol-induced restoration of end-expiratory lung volume; however, the possibility that IC was affected by an increase in TLC was not evaluated. Although a common impact of the diaphragm fiber electrolyte gradient on both CFdi and RES could explain the findings of the present study, future studies are required to validate this hypothesis. However, it is possible that the discomfort associated with RES may act to protect the excitability of the diaphragm sarcolemma.

Considering methodological and technical aspects of the present study, it is important to mention that for an accurate physiological measurement of CFdi it is necessary to control for (a) changes in muscle-to-electrode distance, (b) electrode positioning with respect to the muscle fiber direction and location, (c) electrode configuration, (d) signal-to-noise ratio, (e) influence of cross-talk from other muscles (including the heart and the esophagus), and (f) electrode movement-induced artifacts (6, 7, 8, 26, 38, 39). The technology used to measure the EAdi power spectrum in the present study used methods and devices to minimize all these influences (6, 8). Changes in chest wall configuration/lung volume do not affect CFdi obtained during voluntary contractions in healthy subjects (10, 38, 40) nor do changes in its length (28, 39).

In conclusion, when breathing with a fixed pattern, constant diaphragm activation and maintained pressure generation (a) RES increases over time and (b) the changes in RES over time are associated with changes in CFdi, independent of Pdi level. This suggests that common intrinsic properties of the diaphragm may affect, directly or indirectly, the time course of both RES and CFdi.