A novel clinical test of respiratory muscle endurance

Subjects

The protocol was approved by the Ethics Committee of the Royal Brompton Hospital, London, UK, and all participants gave their informed consent.

Study 1 included 30 healthy subjects who underwent RME testing to determine the relationship between t_lim and the load/capacity ratio. Study 2 included six of the 30 healthy subjects who underwent RME testing on three separate occasions in order to determine the intrasubject variability of the test. Finally, study 3 included eight patients with interstitial lung disease due to systemic sclerosis (SS/ILD) who underwent RME testing to assess the validity and acceptability of the technique; these patients were compared with eight controls.

Apparatus

Negative pressure threshold loading was performed using a low resistance inspiratory threshold loading (ITL) device, modified from that described by Chen et al. 9. The device consisted of a cylindrical adjustable pressure chamber, 41×10 cm, which was connected to a two-way nonrebreathing valve (PK Morgan Ltd, Rainham, Kent, UK). The negative pressure (range, −10–−135 cmH₂O) was generated by a vacuum cleaner (Numatic International, Chard, Somerset, UK), and pressure in the chamber was measured using a pressure monitor (Magnehelic, Dwyer Instruments Inc., Michigan City, IN, USA). Expiratory flow was measured with a pneumotachograph head (Fleisch, Lausanne, Switzerland) attached to a tube of dimensions 41×3.5 cm connected to an electrospirometer (GM Instruments Ltd, Kilwinning, UK).

Pressure measurements

An oesophageal balloon catheter (Ackrad Laboratories, Cranford, NJ, USA) was inserted and positioned in the standard manner 10. Oesophageal pressure (P_oes) and mouth pressure (P_mouth) were measured using differential pressure transducers (Validyne, Northridge, CA, USA) and amplified by carrier amplifiers (Validyne). All signals were passed to an analog-digital board and Power PC (Apple Computers, Cupertino, CA, USA) running Labview 4.1 software (National Instruments, Austin, TX, USA).

Pressure/time product

Oesophageal pressure-time product (PTP) per total breath cycle (PTP_oes) and mouth PTP per total breath cycle (PTP_mouth) were calculated from P_oes and P_mouth traces for the whole RME test in each subject. The area subtended by the pressure time curve was determined for each breath using a semi-automated modification of Labview 4.1 software. Average PTP_oes and average PTP_mouth were used as indicators of the load on the respiratory muscles.

Dynamic compliance

Dynamic compliance (C_L,dyn) was calculated during resting breathing by dividing tidal volume by the difference in P_oes at points of zero flow.

Maximum inspiratory pressure

Before each RME test, all participants performed maximum inspiratory manoeuvres using a mouthpiece attached to a metal tube, which had a small air leak to prevent glottic closure 11. Efforts were repeated from functional residual capacity (FRC) at least 10 times until three reproducible efforts were obtained. Patients and subjects who were unfamiliar with the manoeuvre had a preliminary session for instruction in order to ensure that maximum pressures were achieved. Peak (rather than the 1-s average) negative oesophageal pressure (P_oes,max) and peak negative mouth pressure (P_mouth,max) were determined, and these measurements were used as the capacity of the respiratory muscles.

Inspiratory threshold loading protocol

All subjects and patients were exposed to the same relative target negative pressure (70% of P_oes,max). No restriction was placed on the pattern of breathing. During the first minute of the test, the threshold pressure was progressively made more subatmospheric from −10 cmH₂O to 70% of P_oes,max. Subjects had to generate a pressure at the mouth, slightly greater than the chamber pressure, to open the valve in order to initiate and sustain inspiratory flow. The ITL device provided a constant inspiratory pressure when the valve was open 9. All the subjects underwent 2–5 practice endurance runs, up to 70% P_oes,max, to become accustomed to the apparatus and the sensation of dyspnoea. It was important for all subjects to familiarize themselves with the equipment and start with low workloads to ensure that the best strategy for handling higher loads was chosen 12.t_lim was defined as the period from when the negative pressure imposed reached 70% of P_oes,max to task failure (i.e. the incremental loading minute was excluded). Task failure was defined as the inability to generate the target pressure and open the valve. Strong verbal encouragement was given throughout the RME test, and subjects and patients were instructed to stop and come off the mouthpiece when they could no longer open the valve. All subjects continued to exhaustion or until the sensation of dyspnoea was too uncomfortable to tolerate.

Statistical analysis

Due to the curvilinear relationship between t_lim and the load/capacity ratio, a single-segment logarithmic conversion of t_lim was performed to produce a linear model to describe the relationship between t_lim and PTP_oes/P_oes,max. Furthermore, this variance-stabilizing transformation of the data permitted the definition of 95% confidence bands as well as the establishment of RME values in healthy subjects.

Linear regression analysis was used to determine the relationships between log t_lim and PTP_oes/P_oes,max; PTP_oes and the duty cycle (duration of inspiration/duration of total breathing cycle (t_I/t_tot)); log t_lim and t_I/t_tot; and log t_lim and PTP_mouth/P_mouth,max. The relationship between P_oes and P_mouth data was assessed using a Bland and Altman plot 13. An analysis of variance (ANOVA) test for repeated measures was used to evaluate differences between three consecutive RME tests. Differences between controls and patients with SS/ILD were made with an unpaired t-test. A value of p<0.05 was considered statistically significant. Data are presented as mean±sd.

Study 1: healthy subjects

Thirty subjects (20 males) were studied with a mean±SD age of 38±10 yrs (range 26–65 yrs). Fifteen of these subjects were naïve to the study.

Relationship between endurance time and respiratory load/capacity ratio

The t_lim showed a curvilinear relationship with PTP_oes/P_oes,max (fig. 1a⇓). A linear relationship was demonstrated between log t_lim and PTP_oes/P_oes,max (r=0.88, p<0.0001; fig. 1b⇓).

• Download figure
• Open in new tab
• Download powerpoint

Fig. 1.—

a) “Best-fit” plot of the curvilinear relationship between endurance time (t_lim) and oesophageal load/capacity ratio (oesophageal pressure-time product per total breath cycle (PTP_oes)/maximum negative oesophageal pressure (P_oes,max)) in normal subjects (n=30). b) Regression analysis with the 95% confidence bands of the linear relationship between log t_lim and PTP_oes/P_oes,max in normal subjects (n=30; r=0.88, p<0.0001).

Breathing strategy

There was a linear relationship between PTP_oes/P_oes,max and t_I/t_tot (fig. 2a⇓; r=0.87, p<0.0001) and also between log t_lim and t_I/t_tot (fig. 2b⇓; r=0.77, p<0.0001).

• Download figure
• Open in new tab
• Download powerpoint

Fig. 2.—

a) Regression analysis of the linear relationship between duty cycle (duration of inspiration/duration of total breathing cycle (t_I/t_tot)) and oesophageal load/capacity ratio (oesophageal pressure-time product per total breath cycle (PTP_oes)/maximum negative oesophageal pressure (P_oes,max)) in normal subjects (n=30; r=0.87, p<0.0001). b) Regression analysis of the linear relationship between the log endurance time (log t_lim) and duty cycle (t_I/t_tot) in normal subjects (n=30; r=0.77, p<0.0001).

Comparison of oesophageal and mouth load/capacity ratio

Bland and Altman plot 13 demonstrated agreement between PTP_oes/P_oes,max and PTP_mouth/P_mouth,max (fig. 3a⇓). The r-value for the linear relationship between log t_lim and PTP_mouth/P_mouth,max was 0.82 (p⩽0.0001; fig. 3b⇓).

• Download figure
• Open in new tab
• Download powerpoint

Fig. 3.—

a) Bland and Altman plot 13 of the difference in oesophageal pressure-time product per total breath cycle (PTP_oes)/maximum negative oesophageal pressure (P_oes,max) and mouth pressure-time product per total breath cycle (PTP_mouth)/maximum negative mouth pressure (P_mouth,max) against the mean of PTP_oes/P_oes,max and PTP_mouth/P_mouth,max in normal subjects (n=29). b) Regression analysis with 95% confidence bands of the linear relationship between log endurance time (log t_lim) and mouth load/capacity ratio (PTP_mouth/P_mouth,max) in normal subjects (n=29; r=0.82, p<0.0001).

Study 2: reproducibility in healthy subjects

Six subjects (four male) were studied on three separate occasions. Mean time between each test was 78±82 days.

Relationship between endurance time and respiratory load/capacity ratio

Log t_lim was plotted against PTP_oes/P_oes,max for three RME tests in each subject. These serial data were compared to log t_lim and PTP_oes/P_oes,max data obtained from the 30 healthy subjects (fig. 4⇓). Although the results of t_lim and PTP_oes/P_oes,max for each subject changed between each successive test (table 1⇑), figure 4⇓ shows that log t_lim and PTP_oes/P_oes,max data remain within, or close to, the normal values determined by study 1.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 4.—

Plot of log endurance time (log t_lim) and oesophageal load/capacity ratio (oesophageal pressure-time product per total breath cycle (PTP_oes)/maximum negative oesophageal pressure (P_oes,max)) for three separate respiratory muscle endurance (RME) runs in six normal subjects. The arrows indicate 1st (^#), 2nd (^¶) and 3rd (⁺) RME test. The shaded area represents the simple regression analysis with 95% confidence bands from figure 1b⇑. □: subject 1; ▴: subject 2; ○: subject 3; ▪: subject 4; ▵: subject 5; •: subject 6.

Study 3: patients with interstitial lung disease due to systemic sclerosis

Characteristics of the patients are summarized in table 2⇓. Eight patients underwent RME testing and were compared with eight controls (table 3⇓). Log t_lim, less than predicted from the PTP_oes/P_oes,max when compared with normal RME values in study 1 (fig. 1b⇑), was taken as a reduction in RME. All patients had normal cardiac function on echocardiogram and evidence of interstitial lung disease on computed tomography.

Respiratory muscle strength and endurance

There was no difference in mean P_oes,max between the SS/ILD group and the controls. Log t_lim and PTP_oes/P_oes,max were calculated for the patients and compared to the PTP_oes/P_oes,max and log t_lim values obtained from the 30 normal subjects (fig. 5⇓). The log t_lim value was less than predicted from the PTP_oes/P_oes,max value in patients 1, 2, 4, 6–8, indicating a reduction in RME. Patients 3 and 5 had a value for log t_lim that would be predicted from the PTP_oes/P_oes,max, indicating normal RME.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 5.—

Plot of log endurance time (log t_lim) and oesophageal load/capacity ratio (oesophageal pressure-time product per total breath cycle (PTP_oes)/maximum negative oesophageal pressure (P_oes,max)) for eight patients with systemic sclerosis and interstitial lung disease, compared directly with that of the 30 healthy normals in figure 1b⇑. □: patient 1; ▴: patient 2; ○: patient 3; ▪: patient 4; ▵: patient 5; •: patient 6; ⋄: patient 7; ♦: patient 8.

Critique of method

Respiratory load/capacity ratio to assess respiratory muscle endurance

Previous investigators have evaluated the endurance properties of the respiratory muscles and the effect of pressure and timing on diaphragm 5 and ribcage muscle 6 fatigue. These studies used pressure time index (PTI=(mean pressure generated/maximum pressure generated)×(inspiratory time/duty cycle)) to determine when fatigue is likely to occur. The present authors found the calculation of PTI to be problematic, even with the modern software available. To calculate PTI, according to the method of Bellemare and Grassino 5, a square-wave breathing pattern must be maintained throughout inspiration, and the plateau pressure is recorded as the mean pressure generated. However, untrained healthy subjects and patients find it difficult to breathe with a square-wave breathing pattern and, in fact, breathe with much more “smoothed-out” patterns, so that it is difficult to accurately determine the plateau pressure. Furthermore, this type of breathing pattern is associated with a more gradual fall and rise in the negative intrathoracic pressure, which is not accommodated by PTI. Even if a nonsquare wave breathing pattern is used, the computation would require an algorithm such that, where P is the pressure generated per breath and n is the number of P points sampled throughout the inspiration 27. A much simpler method uses PTP. Appendini and Donner 28 recently re-evaluated PTI and suggested that PTI is the PTP developed over the total breath cycle (PTP/t_tot) expressed as a percentage of the maximum pressure that can be generated. This modified version of PTI, or load/capacity ratio, was used in the present study to determine the relationship between (P_oes,tp·total breath cycle⁻¹/P_oes,max) and t_lim in subjects breathing with an unrestricted pattern. Interestingly, in the healthy normals, the mean PTP_oes·cycle⁻¹/P_oes,max was 0.29, which is close to the 0.30 threshold value for PTI reported by Zocchi et al. 6, above which ribcage muscle fatigue occurred in four highly-trained subjects performing square-wave breathing.

Learning effect with endurance testing

Previous studies have demonstrated learning effects with the different tests of RME 26. This has contributed to the difficulty in determining a set of normal values for RME, as a healthy subject would have to undergo a number of endurance trials to achieve a reliable result. In study 2, there was wide intra- and intersubject variability in t_lim (table 1⇑).

The results from these repeated tests raises two relevant points: 1) repeated tests of RME may not always result in an increased t_lim; individuals can actually adopt breathing strategies that reduce t_lim (subjects 1 and 2); and 2) variations in t_lim between RME tests in the same individual can be explained if t_lim is related to the respiratory load/capacity ratio such that a decrease in PTP_oes/P_oes,max correlates with an increase in log t_lim (fig. 4⇑). This is advantageous when assessing sequential clinical or experimental data.

Load/capacity ratio and endurance time

The curvilinear relationship between t_lim and the respiratory load/capacity ratio (fig. 1a⇑) is similar to that described in previous studies 5, 6. The present approach recognizes and utilizes this relationship without the need for a fixed breathing strategy and expresses RME as a function of the relationship between t_lim and PTP_oes/P_oes,max. To have a more clinically-useful measure of RME and to establish normal values, logarithmic conversion of t_lim was performed as a variance-stabilizing transformation (fig. 1b⇑). This allows a simple, direct comparison of patient data with normal values (fig. 5⇑).

It could be argued that t_I/t_tot is sufficient to explain the variance in log t_lim, but comparison of figures 1b and 2b⇑⇑ shows that the correlation between log t_lim is better for the load/capacity ratio (r=0.88) than for t_I/t_tot (r=0.77). Thus, t_I/t_tot has a good, but not invariable, relationship with the load/capacity ratio (fig. 2a⇑).

Assessment of respiratory muscle endurance from mouth pressures

A Bland and Altman plot 13 confirmed the close agreement between the P_mouth and P_oes measurements during ITL (fig. 3a⇑), with log t_lim being linearly related to PTP_mouth/P_mouth,max (r=0.82; fig. 3b⇑). Thus, measurements based on P_mouth would permit a noninvasive assessment of RME.

Assessment of respiratory muscle endurance in patients with systemic sclerosis and interstitial lung disease

RMS in patients with ILD is reported as normal 29, but these patients suffer from severe breathlessness and impaired exercise performance. The authors hypothesized that reduction in exercise performance and increased breathlessness is due to increased intrinsic elastic load, resulting in an imbalance in the load/capacity ratio and a reduction in RME. Indeed, all the patients with SS/ILD had normal RMS and six of the patients had a reduction in RME (fig. 5⇑). The increased load in the patients with SS/ILD imposed a different respiratory pattern, with an increased t_I/t_tot, which was reflected in a higher load/capacity ratio than for the controls (table 3⇑). However, this load/capacity ratio could not be sustained, and t_lim was reduced. Interestingly, in the patient group, those with reduced RME had a trend for the C_L,dyn to be lower (0.05 L·cmH₂O⁻¹) than in those with normal RME (0.09 L·cmH₂O⁻¹; p=0.1). Thus, the reduction in RME could result from the combination of the intrinsic elastic load in series with the imposed extrinsic threshold load.

In summary, endurance time can be expressed as a function of the respiratory load/capacity ratio over a range of breathing strategies, using constant submaximal negative pressure threshold loading. The use of this relationship can detect abnormal respiratory muscle endurance in patients with cardiorespiratory disease. Furthermore, mouth pressure accurately reflects oesophageal data, and the noninvasive assessment of respiratory muscle endurance will allow normal values to be established in a large population, against which patients can be evaluated.