INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

In amyotrophic lateral sclerosis (ALS) degeneration of motor neurones results in weakness and wasting of the dependent muscles. Black and Hyatt (1) showed that maximal static expiratory mouth pressures (MEP) were frequently reduced in ALS, indicating involvement of the nerves supplying the inspiratory and expiratory muscles. In patients with ALS, maximal static mouth pressures and other indices of respiratory muscle pump function decline as the disease progresses (2, 3). Moreover, disease involving the inspiratory muscles indicates a poor prognosis in ALS, whether judged by static inspiratory mouth pressure (4) or by surrogate measurements of inspiratory muscle strength (5).

In contrast the effect of ALS on expiratory muscle function, beyond the recognition that static expiratory pressures are commonly reduced (1, 3, 6), is poorly understood. Kreitzer and coworkers (7) studied 32 patients with ALS and found that expiratory muscle weakness was associated with a reduction of peak flow and a blunting of the effort-dependent portion of the maximal expiratory flow volume (MEFV) curve. In their study expiratory muscle weakness seemed to be an isolated finding in that such patients could generate inspiratory and transdiaphragmatic pressures similar to those produced by patients with a normally shaped MEFV curve. Abdominal muscles are the principal muscles of active expiration, and cough is largely (8), though not exclusively (9), dependent on the integrity of this muscle group.

Therefore, the aims of the present study were first to investigate whether abdominal muscle weakness in ALS is an isolated finding or whether inspiratory muscle weakness is also usually found. Second, since one of the principal functions of the abdominal muscles is to generate an effective cough, we examined the relationship between abdominal muscle weakness and cough efficacy as judged by the ability to generate transient supramaximal flows (10). Third, since upper airway problems are increasingly recognized in ALS (11), we made a visual inspection of vocal cord movement in the subgroup with respiratory symptoms. Finally, we sought to evaluate the relative contributions of inspiratory and expiratory muscle weakness to ventilatory failure and the presence of respiratory symptoms.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The protocol was approved by our ethics committee, and all patients gave written informed consent to participate. The 26 patients studied were judged by a neurologist (PNL) to have ALS; by the classification proposed by the El Escorial criteria they had clinically definite, probable, or suspected ALS (12). Sixteen of these (Patients 1 to 16) had respiratory symptoms; 10 patients (Patients 17 to 26) without respiratory symptoms were also studied. For study purposes the following were considered as respiratory symptoms: dyspnea, orthopnea, symptoms suggestive of nocturnal respiratory failure (for example, morning headache or daytime somnolence), or ineffective cough. Only one patient (Patient 15), with a 27-yr history of asthma, admitted the presence of respiratory symptoms prior to the onset of ALS. The duration (in months) of symptoms in general and, when applicable, respiratory symptoms was noted. The extent of ALS was not otherwise assessed except for a crude evaluation of speech and swallowing (see legend to Table 1).

The FEV₁ and the slow VC were measured in accordance with the guidelines of the British Thoracic Society using a bellows spirometer (Vitalograph, Bucks, UK). Blood gas tensions were estimated from arterialized earlobe samples (Radiometer ABL 30; Radiometer A/s, Copenhagen, Denmark) (13).

Abdominal muscle strength was assessed from a balloon catheter 110 cm in length (PK Morgan, Rainham, Kent, UK) swallowed pernasally into the stomach; a second catheter, placed in the esophagus in a conventional manner, allowed measurement of esophageal pressure (Pes) and, by subtraction of Pes from gastric pressure (Pga), transdiaphragmatic pressure (Pdi). The catheters were connected to differential pressure transducers (Validyne MP45-1; Validyne, Northridge, CA), carrier amplifiers (PK Morgan), a 12-bit NB-MIO-16 analogue-digital board (National Instruments, Austin, TX), and a Macintosh Quadra Centris 650 personal computer (Apple Computer Inc., Cupertino, CA) running Labview software (National Instruments). Signals were sampled at 100 Hz.

The thoracic nerve roots were bilaterally stimulated over the tenth thoracic intervertebral space using a 19-pin, 90-mm circular coil powered by a Magstim DEM stimulator (Magstim Co., Whitland, Dyfed, UK) (14). A minimum of five stimulations were given at 100% of maximal stimulator output. Before stimulation the patients rested for 20 min to minimize twitch potentiation (14, 15). Stimulation was performed from relaxed end-expiration (as judged from Pes) in the seated position with the patient wearing a noseclip.

We also measured Pga during a maximal voluntary cough effort (16, 17). This maneuver was performed with the patient also seated but not wearing a noseclip. Specific instructions were not given regarding lung volume; thus, invariably the patients inspired prior to coughing. Repeated efforts were performed until no further increase was obtained. A screen displaying Pga was deliberately made visible to the patient while performing this, and other, tests (18).

The pressure measured at the mouth during a maximal static expiratory effort against a closed shutter (MEP) was also obtained (1). We use a flanged mouthpiece and, in this study, the mouthpiece was held by an investigator rather than by the patient (19). This maneuver was performed from TLC; repeated efforts were performed until no further increase was obtained.

Inspiratory muscle strength was assessed by measurement of Pdi and Pes during a maximal voluntary sniff (20), and by measurement of Pdi after a single bilateral cervical magnetic stimulation (21). The methods used were those described in a previous study (22) and are not elaborated in greater detail here.

To investigate the presence or absence of transient supramaximal flow during coughing (cough spikes), a constant-volume whole-body plethysmograph was used (PK Morgan). Use of the plethysmograph was not necessary in itself, but it enabled the use of a pneumotachograph and software that could superimpose repeated MEFV loops. The patients were seated comfortably and wearing noseclips. Patients were asked to breathe quietly and then perform a MEFV maneuver. Without coming off the mouthpiece they were then asked to repeat the maneuver but to cough vigorously during the expiratory phase. A minimum of three maneuvers were attempted for each patient, but this was increased if necessary for patients who had difficulty performing the maneuver.

Patients with respiratory symptoms (11 of the 16) also received a fiberoptic examination of the vocal cords. This examination was conducted using a bronchoscope (Olympus Optical Co., Tokyo, Japan) passed pernasally using topical anesthesia. Once a good view of the vocal cords was obtained the patients were asked to vocalize and also to cough and make forced expirations.

Data Handling and Conventions

Both the twitch gastric pressure (Tw Pga) and the cough gastric pressure (Cough Pga) were defined as the difference between baseline Pga at resting end expiration and the subsequent peak (typical traces are shown in Figure 1). Likewise, sniff Pdi (Sn Pdi), sniff Pes (Sn Pes), and Tw Pdi are defined as the difference between resting end expiration and the subsequent peak; for simplicity, subatmospheric deflections of Pes are given a positive value. Cough spikes were defined as being present or absent (Figure 2) according to the classification used by Szeinberg and coworkers (10).

View larger version (11K): [in this window] [in a new window]   Figure 1.   Example of gastric pressure (Pga) tracings obtained with T10 magnetic stimulation (upper panel ) and during a maximal voluntary cough (lower panel ) in a patient without respiratory symptoms.

View larger version (12K): [in this window] [in a new window]   Figure 2.   Examples of maximal flow-volume loops with cough efforts superimposed in a patient without cough spikes (left panel ) and in a patient with spikes (right panel ). The scales on the axes are: vertical, each division 1 L/s; horizontal, each division 1 L.

Statistics were computed using unpaired t tests, simple/multiple regression or Fisher's exact test, as appropriate, using Statview 4.02 (Abacus Concepts, Berkeley, CA). A level of p < 0.05 was taken as significant.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Clinical spirometric and symptomatic data are shown in Table 1. The patients with respiratory symptoms had a lower vital capacity (by 1.6 L; 95% confidence intervals [CI], 2.0 to 1.2) (p < 0.0001) than did the asymptomatic patients. There was no difference in the duration of general symptoms, sex, age, or disease type between the group with respiratory symptoms and the group without. Typical traces for Cough Pga and Tw Pga are shown in Figure 1; strength data are shown in Table 2. Compared with patients without respiratory symptoms, symptomatic patients had a lower Sn Pdi, Sn Pes, and Tw Pdi (p < 0.0001 for all three). Of the expiratory tests only Tw Pga was significantly lower (p = 0.0009) in the symptomatic patients. Among symptomatic patients those volunteering the symptom of poor cough did not have lower expiratory muscle strength, nor did measures of inspiratory muscle strength relate to specific symptoms.

Weakness of the expiratory and inspiratory muscles frequently coexisted; maximal Sn Pes correlated significantly with maximal Cough Pga (r = 0.57, p = 0.002) (Figure 3), MEP (r = 0.47, p = 0.02) and Tw Pga (r = 0.56, p = 0.003). Tw Pdi correlated significantly with Tw Pga (r = 0.67, p = 0.0003). There was no relationship between MEP and Cough Pga (r = 0.12, p = 0.58) or Tw Pga (r = 0.27, p = 0.21) (Figure 4); however, Tw Pga correlated with Cough Pga (r = 0.4, p = 0.04) (Figure 5). As shown in Figure 4, high values for both the Cough Pga and the Tw Pga were frequently found in association with a low MEP. Blood gas data are shown in Table 3; two patients (Patients 4 and 21) had negative alveolar-arterial oxygen gradients. Pa_CO2 showed a significant association with Tw Pga and all measures of inspiratory muscle strength. Multivariate regression analysis was therefore performed using all six indices of respiratory muscle strength; only Sn Pes emerged as a significant independent predictor of PCO₂ (p = 0.04) (Figure 6). Compared with asymptomatic patients, patients in the symptomatic group had a lower pH (p = 0.0004), a higher Pa_CO2 (p < 0.0001), and a higher bicarbonate (p < 0.0001).

View larger version (11K): [in this window] [in a new window]   Figure 3.   Relationship between maximal Cough Pga and maximal Sn Pes showing that inspiratory and expiratory muscle weakness frequently coexisted. Closed circles are patients with respiratory symptoms; open circles are those without.

View larger version (11K): [in this window] [in a new window]   Figure 4.   Relationship between MEP and Cough Pga (upper panel ) and Tw Pga (lower panel ).

View larger version (10K): [in this window] [in a new window]   Figure 5.   Relationship between Tw Pga and the maximal Cough Pga.

View larger version (10K): [in this window] [in a new window]   Figure 6.   Relationship between maximal Sn Pes and arterialized capillary CO2 tension. Closed circles are patients with respiratory symptoms; open circles are those without.

Typical cough spike traces are shown in Figure 2. The presence or absence of spikes was related to the strength of the cough as judged by maximal Cough Pga (p = 0.02). Illustrating these data graphically (Figure 7), it appears that a threshold effect operates so that below a certain level of abdominal muscle strength cough spikes were usually absent. These values were, approximately, Cough Pga < 50 cm H₂O, Tw Pga < 7 cm H₂O, and MEP < 30 cm H₂O. Categorizing the patients (irrespective of symptom category) with respect to these threshold values showed significant differences (using Fisher's exact test) in the ability to generate cough spikes for a Cough Pga > 50 cm H₂O (p = 0.009) or a Tw Pga > 7 cm H₂O (p = 0.006). The threshold for MEP (> 30 cm H₂O) failed to reach significance (p = 0.08). Two of the eleven patients examined fiberoptically had abnormal vocal cord motion. These abnormalities were unilateral abduction failure (one case) and unwanted bilateral cord closure during rapid expiration (one case).

View larger version (8K): [in this window] [in a new window]   Figure 7.   Abdominal muscle strength assessed using maximal Cough Pga (upper panel ), MEP (middle panel ), and Tw Pga (lower panel ) in relation to the presence or absence of cough spikes.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Our data show the following in ALS. First, expiratory muscle weakness is often associated with inspiratory muscle weakness. Second, the ability to achieve transient supramaximal flows during a cough (indicating dynamic airway compression) is related to abdominal muscle strength; however, because this operates with a threshold effect this ability is not lost until substantial levels of weakness are present. Third, expiratory muscle weakness, in contrast to inspiratory muscle weakness, is not an independent predictor of hypercapnia. Finally, two of 11 vocal cord examinations in patients with respiratory symptoms were abnormal, confirming that vocal cord dysfunction is a feature of established ALS.

Critique of the Method

Validity of tests of expiratory muscle strength. In contrast to the range of tests available for assessment of inspiratory muscle strength (23), relatively few tests are available to assess expiratory muscle strength, and only one, the MEP, has established normal values. The conclusions that would be drawn from examination of the MEP measurements in isolation from the Cough Pga and the Tw Pga would be substantially similar to those drawn from examination of all three. However, although Tw Pga and Cough Pga have limitations as tests of expiratory strength (discussed below), the MEP also has limitations, even though it is an established test. There are direct data to suggest that technique influences the value obtained for MEP (19), and, in addition, experience obtained from other maximal isometric voluntary force maneuvers suggest that many patients may not always make a maximal effort (24), particularly if, as is the case with an MEP maneuver, the muscle length is long (25).

Normal values for maximal Cough Pga in our laboratory are > 175 cm H₂O for men and > 100 cm H₂O for women (16); using this criterion abdominal muscle weakness was a frequent finding in our patients. The maximal voluntary Cough Pga has not been formally described as a test of expiratory muscle strength, although this maneuver, with measurement of Pes, was considered by Byrd and Hyatt (17) to be superior to the MEP, at least for use in patients with lung disease. Cough Pes is similar to Cough Pga since, in general, diaphragm pressure generation during a voluntary cough is small in magnitude (for example, see Figure 1) and therefore Cough Pga and Cough Pes are usually numerically similar.

For the Tw Pga, as with any twitch technique (15, 26), potentiation can falsely increase the measured pressure; we therefore took great care to avoid twitch potentiation. A further problem is that thoracic nerve root stimulation is recognized not to be supramaximal (14, 27); thus, relatively minor variations in adiposity or skeletal deformities might result in variation in field penetration and hence results. Nevertheless, many of the values reported in the present study are substantially lower than we observed in normal subjects (14). It is thus probable that Tw Pga does detect abdominal muscle weakness and thus, given that Tw Pga is reproducible for individuals (14), this technique might be of value in patients with ALS unable to perform voluntary tests.

Generation of cough spikes. Our conclusions rest on the capability of our apparatus to detect cough spikes and on the patient to generate them. With our method we have been able to reliably detect cough spikes in normal subjects and also, as shown, in a proportion of the patients with ALS. There is no way of ensuring that our patients made a maximal effort beyond asking them to perform repeated maneuvers. We strongly encouraged patients to make maximal efforts and believe that they were well motivated. Despite this, Patients 14 and 21 failed to generate spikes despite performing relatively strongly on tests of expiratory muscle strength. This could indicate that they failed to make a maximal cough effort, although this seems unlikely.

However, it is also of interest that these patients had marked bulbar disease. Supramaximal flows are believed to result from air displaced from collapsing airways. Although supramaximal flow is observed when the glottis is open (as during a "huff" [28]), the sudden opening of the previously closed glottis in normal cough amplifies this. The action of the glottis during cough is evident from the figures presented by Szeinberg and coworkers (10) where interruption to expiratory flow before the cough effort is observed whether or not a spike follows. This pattern was also observed in our subjects, although this is not clearly demonstrated in Figure 2 because of the superimposition (by the plethysmograph software) of the resting flow-volume loop. Thus, the possibility that glottic dysfunction could increase the threshold strength beyond which cough spikes can be detected deserves consideration. An additional mechanism that could operate in patients with bulbar disease would be that inappropriate glottic narrowing in the expulsive phase of cough could dampen or prevent transmission of a transient supramaximal flow to the pneumotachograph at the mouth.

Significance of the Findings

Impaired expiratory muscle strength is confirmed as a recognized finding in established ALS. Our data extend this by showing that once weakness extends beyond a critical level, there is an associated inability to generate supramaximal flow transients during cough. This in turn suggests that when weakness is severe, dynamic airway compression does not occur and therefore that cough efficacy is compromised. This has not been previously investigated in patients with ALS, but these data are similar to those reported by Szeinberg and coworkers (10) in their study of patients with muscular dystrophy. They concluded that the lowest level of MEP consistent with production of flow transients was 60 cm H₂O; we were not able to obtain data of this clarity for the MEP, but the threshold obtained from our data for Cough Pga (50 cm H₂O) supports this observation. Many patients, with a variety of neuromuscular disorders, have sufficiently severe expiratory muscle weakness to place them close to or below this threshold level (29, 30). Moreover, in neuromuscular disease expiratory muscle strength can fall acutely during upper respiratory tract infections (30). Our data predict that, because of the threshold effect, some patients who can generate flow transients when well would be unable to do so during acute infection. These patients might particularly benefit from therapies to assist or augment cough and could be identified by measurement of expiratory muscle strength using any of the three techniques used in the present study.

The established test for expiratory muscle strength, the MEP, did not correlate with either Cough Pga or Tw Pga. This could be either because Tw Pga and Cough Pga were not valid measures of abdominal muscle strength or, alternatively, because the MEP is not a consistently useful test in patients with ALS. If the first hypothesis were correct then when the MEP gives low values, given that the abdominal muscles are the primary muscles of active expiration (31), the Cough Pga and the Tw Pga must also be low. Examination of Figure 4 shows that this is not so; in fact, high values of both Cough Pga and Tw Pga were commonly obtained in the presence of a low MEP. Further support for the alternative proposition, that the MEP is not always reliable in ALS, is provided by the relationship between these data and the presence or absence of cough spikes that are an independent physiologic measurement (Figure 7). These data show that although patients judged to be strong by any test usually had spikes, among weaker patients there was more overlap in values between spike-present and spike-absent groups with the MEP than either the Cough Pga or the Tw Pga. These observations suggest that Cough Pga and Tw Pga could have a role as part of a range of tests for evaluating abdominal muscle strength. Specifically, as with the maximal static inspiratory pressure, a high MEP value excludes muscle weakness; for this purpose the MEP has clear advantages of patient tolerance and well-defined normal ranges. However, for patients with a low MEP value the Cough Pga or the Tw Pga might be useful to distinguish weakness from normality. Further studies are therefore warranted in normal subjects to establish normal ranges.

The importance of generating a supramaximal flow has been recently questioned by the work of Bennett and Zeman (28) showing that the generation of enhanced supramaximal flow does not, by itself, increase the excretion of radiolabeled particles from the lung. However, their study was addressing the lung clearance capacity of neurologically normal subjects performing a cough enhanced (or not) by a valve device against a huff. Whichever of these three maneuvers was performed supramaximal flow (compared with the MEFV loop) occurred; dynamic airway compression would therefore be expected with all three maneuvers. Thus, their study compared enhanced supramaximal flow against supramaximal flow rather than, as would be the case with patients with severe expiratory muscle weakness, supramaximal flow against maximal flow. Our data show that some patients with ALS cannot generate supramaximal flow and therefore by implication dynamic airway compression (32). Further studies are therefore required to examine whether losing the expiratory pressure-generating strength required to develop supramaximal flows correlates with functional ability to clear the lung.

The observation that, in multivariate analysis, only Sn Pes predicts hypercapnia in ALS is of interest because sniff pressures in the esophagus are closely related to those in the nasopharynx (33), at least in patients without lung disease. Sniff nasal pressure can now be easily and cheaply measured using a portable handheld meter; it is arguable that this test should form part of the clinical assessment of all patients with ALS.

Vocal cord abnormalities have not been previously investigated in ALS, although epiglottic obstruction of the airway is recognized (34), and upper airway dysfunction has been inferred from oscillations of the flow-volume loop (11). We have seen two further patients with vocal cord abnormalities; one of these had unilateral vocal cord paralysis and the other had unwanted bilateral cord closure during both rapid inspiratory and expiratory maneuvers. Thus, vocal cord dysfunction seems to be a feature of some cases of established ALS.

In summary, expiratory muscle strength and cough function were investigated in ALS. Abdominal muscle weakness was seldom isolated from inspiratory muscle weakness. In multivariate analysis, expiratory muscle strength, unlike inspiratory muscle strength, was not a predictor for the development of hypercapnia. In patients with ALS with severe abdominal muscle weakness, we observed an inability to generate transient supramaximal flows, suggesting that such patients cannot achieve dynamic airway compression.