Tweetable abstract
Interpretation of pulmonary function testing in patients with amyotrophic lateral sclerosis must account for coexisting lung diseases, when making patient care decisions. https://bit.ly/3Co2yR0
Case history
A 71-year-old male, with a 30-pack-year history of smoking and COPD, was referred to our clinic for pulmonary evaluation after being recently diagnosed with amyotrophic lateral sclerosis (ALS). His original presenting neurological symptom was bilateral hand weakness.
The patient reported occasional drooling suggestive of early ALS-associated bulbar dysfunction, yet he denied dysphagia or coughing while eating. He denied breathlessness while lying in the supine position (orthopnoea). He also denied dyspnoea on exertion; however, his ambulation was limited to few steps mainly due to ALS-associated lack of balance and mild leg weakness. He reported sleeping well, with no morning headaches or daytime sleepiness. His physical examination was significant for decreased hand grip strength and mild tongue fasciculations. Chest movements were equal bilaterally, there was no abdominal paradox, inspiratory neck muscle recruitment, or staccato voice. Lungs were clear on auscultation. Serum bicarbonate concentration and partial pressure of carbon dioxide were normal. Peak cough flow was 220 L·min−1.
2 years before the onset of ALS symptoms, spirometry in the sitting position revealed a forced expiratory volume in 1 s (FEV1) of 1.86 L (66% predicted), a forced vital capacity (FVC) of 3.08 L (80% predicted) and a FEV1/FVC ratio of 0.60, all suggestive of COPD with moderate airflow obstruction. Upon diagnosis of ALS, FEV1 while sitting was 1.61 L (59% predicted), FVC was 2.66 L (71% predicted) and the FEV1/FVC ratio was 0.61. Supine FVC was 2.02 L, signifying a decrease in FVC from the sitting to supine position of 24%. We also obtained maximal inspiratory pressures (MIP), sniff nasal pressures (SNIP) and maximal expiratory pressures (MEP), which were −62 cmH2O (80% predicted), −68 cmH2O (70% predicted) and +58 cmH2O (35% predicted), respectively. Finally, on ultrasound imaging, diaphragm thickness at functional residual capacity (FRC) was 1.8 mm and the muscle's thickening on inhaling from FRC to total lung capacity (TLC) was 19%.
Should we initiate noninvasive ventilation?
Timing for the initiation of noninvasive ventilation (NIV) in patients with ALS is still a matter of debate [1–3]. According to the European Federation of Neurological Societies, FVC <80% of predicted, with symptoms, is one of the proposed criteria to initiate NIV (table 1) [1]. Both FVC and slow vital capacity (SVC) can be assessed, with the best of the two values being taken into account [3]. NIV increases survival, improves patients’ quality of life and is the preferred therapy to alleviate symptoms of respiratory insufficiency, although this has not been confirmed in patients with bulbar dysfunction [4]. Treatment is usually initiated at night to alleviate nocturnal hypoventilation (figure 1 and table 1). Patients with bulbar dysfunction are less compliant with NIV, due, in part, to their difficulties with oral mask fit, laryngospasm, and inability to control secretions driven posteriorly [4].
Before recommending NIV to our patient, it is important to scrutinise the spirometry procedure used by the pulmonary function laboratory, as patients with bulbar dysfunction can have difficulty in achieving a good mouth seal with a standard, cylindrical mouthpiece. This can lead to spuriously low measurements. In one study the median (interquartile range) SVC with the standard cylindrical mouthpiece was 8.4% (3.9–15.5%) smaller than with the flanged mouthpiece (p<0.001) [5].
Cognisant that the pulmonary function laboratory had used a cylindrical mouthpiece, we repeated spirometry with a flanged mouthpiece. With it, FVC in our patient increased from 71% to 82% predicted. Thus NIV, on account of FVC alone, is not indicated.
Does a near normal FVC signify preserved respiratory muscle strength?
Generation of a FVC manoeuvre requires maximal voluntary recruitment of the inspiratory muscles, followed by maximal voluntary recruitment of the expiratory muscles [6]. Although FVC is affected by impairments of the inspiratory and expiratory muscles, FVC remains normal, or only minimally reduced, if respiratory muscle strength is above 50% predicted [6]. This finding results from the sigmoid shape of the pressure–volume relationship of the respiratory system.
Patients with diaphragmatic paralysis experience a substantial fall in FVC on lying down [6]. On this basis, Allen et al. [7] measured FVC in the standing and supine positions in healthy subjects, in patients with restrictive lung disease, and in patients with obstructive lung disease. No study participant had a history of respiratory muscle weakness. The investigators concluded that a decrease in FVC (ΔFVC) >25% in heathy and restricted subjects from the standing to the supine position, or >35% in patients with obstructive lung disease, should be an indication for a further study of diaphragm strength. In 2019, the panel of experts for the European Respiratory Society (ERS) statement on respiratory muscle testing noted that they would suspect bilateral diaphragm weakness if, among other findings, ΔFVC was >30% [8]. They stressed, however, that this threshold was an “expert opinion [and it was] …not intended as a recommendation for clinical practice” [8]. More recently, Brault et al. [9] assessed 84 patients, 17 of whom had bilateral diaphragm dysfunction (where diaphragm dysfunction was defined as a diaphragm thickening fraction ≤30% during maximal inspiration); they concluded that a ΔSVC value ≤15% was strongly associated with the presence of bilateral diaphragm dysfunction. Only two patients among those with bilateral diaphragm dysfunction had COPD (the mean±sd FEV1/FVC for the group with bilateral diaphragm dysfunction was 79±9%). In an accompanying editorial, Dres and Laveneziana [10] concurred with the investigators that a limitation of the study was the lack of a diagnostic gold standard for diaphragmatic dysfunction (i.e. bilateral anterior phrenic nerve stimulation). Other strategies used to assess respiratory muscle strength, and in particular inspiratory muscle strength, include measurements of MIP, SNIP and transdiaphragmatic pressure (Pdi) elicited by phrenic nerve stimulation [6, 11].
Clinical pearl
A decrease in FVC or SVC on switching from the upright to the horizontal position of >25% in patients without obstructive lung disease and of >35% in patients with obstructive lung disease should alert a clinician to the possibility of inspiratory muscle weakness [7], and MIP should be checked. A MIP more negative than −80 cmH2O in men and −70 cmH2O in women excludes clinically relevant muscle weakness [12]. A false-positive result can arise from poor technique. The addition of SNIP to MIP helps reduce false-positive diagnosis of inspiratory muscle weakness by 20% [13, 14]. SNIP more negative than −70 cmH2O in men and more negative than −60 cmH2O in women excludes clinically significant weakness [12]. With less negative pressures, clinicians should consider measuring Pdi, using voluntary manoeuvres and/or stimulation of the phrenic nerves. Sniff Pdi is useful when SNIP yields suspiciously low values, such as in cases with upper airway obstruction (e.g. hypertrophy of the adenoids, rhinitis, polyps) or lower airway obstruction [8]. Hyperinflation, a common finding in COPD, can decrease MIP and SNIP in the absence of respiratory muscle weakness. However, SNIP less negative than −40 to −50 cmH2O is unlikely to arise though hyperinflation alone. Nocturnal oximetry is a simple tool to screen for nocturnal desaturations. Prolonged desaturations may reflect ventilation/perfusion mismatch or nocturnal hypoventilation, which indicate the need for NIV [1–3]. Nocturnal transcutaneous carbon dioxide [2, 3] can distinguish between desaturations due to nocturnal hypoventilation and desaturations due to ventilation/perfusion mismatch [3]. Unfortunately, cost, limited availability, frequent technical failures (drift), and a lack of a clear definition of nocturnal hypoventilation limit the use of this technique [2, 3].
How do we reconcile a moderately reduced FVC with the contrasting findings seen on diaphragm ultrasound?
In our patient, diaphragm thickness at resting FRC was <2 mm, and the change in thickness when inhaling from FRC to TLC was <20%. In one study, Gottesman and McCool [15] enrolled 30 study participants, five of whom had bilateral diaphragm paralysis, and suggested that diaphragm thickness at resting FRC <2 mm together with a change in thickness when inhaling from FRC to TLC of <20% discriminates between a weak or paralysed diaphragm and a normal diaphragm. Unfortunately, diaphragm thickness at resting FRC ranges from 1.2 to 11.8 mm in healthy individuals [16]. Additionally, measurements of diaphragm thickening to TLC are marred by limitations. For example, measurements of diaphragm thickening explain as little as one third of the variability in inspiratory effort, and the within-session reproducibility of diaphragmic thickening is weak [6, 16, 17].
How valid are current ultrasound imaging techniques in determining diaphragm function?
Validity refers to “the degree to which the test or instrument measures what is supposed to measure” [18]. In our case, the “test” in question (diaphragm ultrasonography) is supposed to provide information (a measure) about diaphragm pressure output and spirometry (e.g. vital capacity). Despite initial enthusiasm for this technique, we now know that the correlation between tidal changes in Pdi and diaphragm pressure–time product (PTPdi) and diaphragm thickening have an “r” value of only 0.40 and 0.38, respectively. This signifies that diaphragm thickening (by ultrasound) explains approximately only 20% and 19% of the variability of tidal changes in Pdi and PTPdi. The correlation between Pdi and PTPdi and diaphragm thickening in patients with diaphragm weakness (i.e. mechanically ventilated patients) is even less robust [19]. In healthy individuals, investigators have reported no significant correlation between diaphragm thickening during voluntary contractions and inspiratory pressures [20]. Similarly, the correlation between diaphragm excursions measured by the subcostal approach [10] and diaphragm pressure output is extremely poor [21]. Not surprisingly, the correlation between diaphragm ultrasound measurements and spirometry is very poor. For example, in a cross-sectional study of 33 patients with ALS, Rajula et al. [22] reported that diaphragm thickening was associated with FVC % predicted (r=0.376), signifying that diaphragm thickening explains only about 18.8% of the variability of FVC % predicted. In a study of 47 patients, 29 with ALS and 18 with myotonic dystrophy or other neuromuscular disorders, Carrié et al. [23] reported that diaphragmatic excursion measured by a right subcostal ultrasound approach explained about 50% of the variability in FVC % predicted (r2=0.529). In that study, a FVC % predicted of 60% could be associated with diaphragmatic excursions ranging from about 2 cm to about 7 cm (see figure 2 in Carrié et al. [23]). In a preliminary study of 29 patients with ALS and 13 subjects without ALS, Viccaro et al. [24] reported that diaphragmatic excursion correlated only modestly with FVC % predicted in the orthostatic (r=0.42) and supine (r=0.39) positions, and with the orthostatic to supine ΔFVC (r= −0.25). Thickening fraction correlated only modestly with FVC % predicted in the orthostatic (r=0.30) and supine positions (r=0.26). These results signify that diaphragmatic excursion and thickening explain only about 13% to 21% of the variability of the various FVC measurements in the two positions [24].
In a prospective study of 40 patients with ALS, Pinto et al. [25] reported that maximal and minimal diaphragm ultrasound thickness, during full inspiration and expiration, did not change over 4 months. In contrast, the amplitude of the diaphragm compound action potential, MEP and SNIP declined significantly. Lecci et al. [26] assessed whether the expiratory excursion of the right hemidiaphragm could be used as an indirect estimate of peak cough flow in 21 patients with ALS. The investigators concluded that the coefficient of determination (r2) between peak cough flow and expiratory excursion of the right hemidiaphragm was 0.75. Unfortunately, the relationship between peak cough flow and diaphragmatic expiratory excursion was poor when the peak cough flow was reduced (i.e. the peak cough flow in the range of greatest clinical interest). For example, a peak cough flow of 200 L·min−1 could be associated with a diaphragmatic expiratory excursion ranging from about 2 cm to about 5 cm. In a preliminary study of 24 patients with ALS, Ferrer Espinosa et al. [27] reported that diaphragm echo intensity was greater in patients on NIV than in patients not on NIV. A major limitation of this technique is that within the diaphragm there is often an additional bright layer due to connective tissue and vessels (figure 2) that is bound to increase echo intensity independent of any ALS-associated change in muscle fibres. Finally, to the best of our knowledge, prospective studies supporting the use of ultrasound imaging for clinical decision making in the individual patient with ALS are non-existent. Put simply, current quantitative ultrasound measurements of the diaphragm are not valid methods to provide information about diaphragm function or spirometry [16]. Therefore, we reason that it is premature to recommend ultrasound imaging in the routine clinical evaluation of patients with ALS.
As noted by Barnes and Simon [28], establishing the validity of diaphragm ultrasound in assessing respiratory muscle weakness would require a reference standard beyond those currently available, for example, simultaneous diaphragmatic electromyogram and oesophageal manometry, which would be both invasive and time-consuming.
Would the patient benefit from a mechanical in-exsufflator?
The most widely available and user-friendly tool to measure cough effectiveness is the peak cough expiratory flow meter device. A peak cough flow value >160 L·min−1 is needed to clear secretions in stable conditions [29] and a value >255 L·min−1 is needed during acute lower respiratory tract infections [3]. Clinicians recommend using cough assist devices when peak cough flow values are <270 L·min−1 [29]. Among patients with ALS, those with a peak cough flow >337 L·min−1 have a greater chance of being alive at 18 months than those with a smaller peak cough flow [29].
Our patient's peak cough flow was 220 L·min−1. The interpretation of our patient's peak cough flow value is complicated by two pre-existing conditions. First, he has COPD, a condition which can lead to expiratory flow limitation. This by itself decreases peak cough flow values independently of respiratory muscle weakness. Second, he has bulbar dysfunction. This can impair a good seal with the mouthpiece, or cause difficulties due to upper airway collapse, laryngospasm, and inability to control secretions. Therefore, it is impossible to determine whether the reduced peak cough flow value is a true positive or a false positive.
Given the low risk of harm of a cough assist device, we offered one to the patient, but he was very reluctant. He stated: “If you can tell me for sure I need it, I'll take it.”
What other options do we have to assess expiratory muscle function?
One reasonable option is to repeat the peak cough flow manoeuvre using a flanged mouthpiece [5] or an oronasal mask [3]. To the best of our knowledge, however, no investigators have compared peak cough flow recorded with a cylindrical or with a flanged mouthpieces or with an oronasal mask in patients with bulbar ALS. A second option is to record MEP. Interpretation of a low MEP can be difficult because patients with facial muscle weakness (bulbar ALS) have difficulty in performing this manoeuvre [30]. Additionally, MEP is subject to false positives if the effort is insufficient.
If we are not confident that our patient is able to achieve a good seal even with a flanged mouthpiece, or we suspect a volitional error, we could record gastric pressure during a cough manoeuvre (cough Pga) [6, 11]. Complementing MEP with cough Pga recordings decreases false-positive diagnosis of expiratory muscle weakness by 30–42% [6]. But, like any voluntary effort, cough Pga can be subject to volitional errors.
To eliminate volitional errors, a non-volitional evaluation of the abdominal muscles can be achieved by measuring changes in Pga elicited by stimulation of thoracic nerve roots that innervate the abdominal muscles (or, twitch gastric pressure, Pgatw) (figure 3) [31].
Clinical pearl
A MEP less than +80 cmH2O in men and less than +60 cmH2O in women arouses suspicion of expiratory muscle weakness [14], and cough Pga should be checked. A cough Pga greater than +130 cmH2O in men and +95 cmH2O in women excludes clinically important muscle weakness [14]. As cough Pga is a volitional test, a smaller value can result from poor technique. The addition of Pgatw to MEP and cough Pga reduces false-positive diagnosis of expiratory muscle weakness by 56% [14]. Pgatw greater than +16 cmH2O in men and women excludes clinically important expiratory muscle weakness [14]. In patients with ALS, an effective cough is unlikely when Pgatw is less than +7 cmH2O [32]. Caution must be taken when interpreting low Pgatw values as they may result from submaximal stimulation of the thoracic nerve roots and not because of muscle weakness [31]. Unfortunately, recordings of volitional cough Pga and Pgatw require placement of a nasogastric catheter, a procedure that can be difficult, particularly in patents with bulbar dysfunction.
While a mechanical in-exsufflator device can increase peak cough flow [33], there is very little outcome data [34]. Therefore, we decided to not perform any further expiratory muscle testing and invited the patient to use a mechanical in-exsufflator device. Unfortunately, due to the bulbar involvement [35], the patient derived no benefit and returned the device to the ALS clinic.
6 months later, the patient has worsened significantly. He notes orthopnoea and a weak cough. He is unable to close his mouth, and he can no longer phonate. Serum bicarbonate is now 33 mEq·dL−1. We ordered repeat spirometry and peak cough flows, but the technician was unable to obtain usable data, even with a flanged mouthpiece or an oronasal mask.
What options remain to quantify respiratory function?
Since spirometry is no longer possible, one option is SNIP. If SNIP is normal, it would be reassuring. If it is low, the interpretation is less clear. SNIP does not correlate well with either arterial carbon dioxide or FVC in patients with bulbar dysfunction [36]. This is thought to stem from either upper airway collapse during the manoeuvre, difficulty in achieving complete mouth closure, or the combination of the two. Measurement of Pdi could yield better data. Such a measurement, however, requires placement of oesophageal and gastric catheters, and it seems unlikely our patient with severe bulbar dysfunction would be able to tolerate this procedure.
Measurement of respiratory function in this patient would ultimately be used to make two decisions: risk stratification of percutaneous endoscopic gastrostomy and whether to initiate NIV [1]. For the first decision, the patient opted to forgo percutaneous endoscopic gastrostomy and thus, from this standpoint, any measurement of respiratory function is unnecessary. As for the second decision, we already have enough data to initiate NIV, i.e. the patient complains of orthopnoea and his serum bicarbonate is high, which is likely a compensatory metabolic alkalosis arising from chronic hypercapnia.
Knowing full well that patients with bulbar dysfunction do not benefit from NIV as much as non-bulbar patients [4], we still offered NIV to him. Patients should have an opportunity to demonstrate failure. Otherwise, they are denied a chance at success (albeit a low one).
We initiated NIV, but the patient did not tolerate it. He declined tracheostomy and enrolled in a hospice. He died at home, peacefully, 2 months later.
Key points
Use caution when interpreting spirometry and peak cough flow data in patients with neuromuscular disease and co-existing obstructive lung disease. Considering the usual slow progression of obstructive lung disease, comparing with spirometry performed not long before the onset of ALS symptoms may assist you in estimating the extent of respiratory muscle weakness.
Use caution when interpreting spirometry and peak cough flow data in patients with neuromuscular disease and bulbar dysfunction. Use a flanged mouthpiece and/or sniff nasal pressures. In selected patients, consider recording gastric pressure during a voluntary cough or during stimulation of the thoracic nerve roots.
Quantitative diaphragm ultrasound imaging as currently performed is not a valid tool to estimate the strength of the diaphragm.
Decisions regarding respiratory management of patients with ALS can be made using indirect assessments of respiratory muscle strength, such as clinical signs and symptoms, overnight oximetry, and serum bicarbonate.
Footnotes
Conflict of interest: F. Laghi reports grants or contracts from Merit Review Award, Veterans Administration Research (1 I01 RX002803-01A1), outside the submitted work. S.W. Littleton reports no conflicts of interest.
Support statement: Supported by grants from the Veterans Administration Research Service.
- Received February 11, 2023.
- Accepted June 6, 2023.
- Copyright ©ERS 2023
Breathe articles are open access and distributed under the terms of the Creative Commons Attribution Non-Commercial Licence 4.0.