Respiratory muscle function in patients with nemaline myopathy

In this cross-sectional study, we comprehensively assessed respiratory muscle function in various clinical forms of nemaline myopathy (NM) including non-volitional tests for diaphragm function. Forty-two patients with NM were included (10 males (25-74 y/o); 32 females (11-76 y/o)). The NM forms were typical (n = 11), mild (n = 7), or childhood-onset with slowness of movements (n = 24). Forced vital capacity (FVC) and maximal inspiratory pressure were decreased in typical NM in comparison with childhood-onset NM with slowness (32.0 [29.0-58.5] vs 81.0 [75.0-87.0]%, p < 0.01, and 35.0 [24.0-55.0] vs 81.0 [65.0-102.5] cmH 2 O, p < 0.01). Eight patients with childhood-onset NM with slowness had respiratory muscle weakness. There was a low correlation between FVC and Motor Function Measure scores (r = 0.48, p < 0.01). End-inspiratory diaphragm thickness and twitch mouth pressure were decreased in patients requiring home mechanical ventilation compared to non-ventilated patients with normal lung function (1.8 [1.5-2.4] vs 3.1 [2.0-4.6] mm, p = 0.049, and -7.9 [-10.9--4.0] vs -14.9 [-17.3--12.6], p = 0.04). Our results show that respiratory muscle weakness is present in all NM forms, including childhood-onset NM with slowness, and may be present irrespective of the degree of general motor function impairment. These ﬁndings highlight the importance for screening of respiratory function in patients with NM to guide respiratory management.

Respiratory muscle weakness is a common feature in patients with NM and can lead to respiratory failure [ 22 , 23 ].In previous clinical studies, the physiological impact of respiratory muscle weakness was mostly assessed by spirometry or based on the presence of clinically evident respiratory insufficiency [ 9 , 24-28 ].A recent NM cohort study also assessed respiratory muscle strength [29] .These studies did not assess respiratory involvement in relation to the clinical NM forms, the relationship between respiratory muscle function and motor function, or specific measures of diaphragm function and structure.Furthermore, respiratory muscle weakness in childhood-onset NM with slowness has not been reported in the cohort studies [ 20 , 30 , 31 ], although it was tested in only one study [31] .Detailed knowledge of respiratory muscle involvement in the different forms of NM is important for respiratory management and for the development of outcome measures for clinical trials.
This cross-sectional study comprehensively assessed the degree of respiratory muscle involvement in different NM forms and genotypes using global measures of respiratory muscle function (spirometry, respiratory muscle strength testing).Moreover, compared to previous work we used non-volitional measures of respiratory muscle function, i.e. ultrasound imaging of diaphragm and intercostal muscles and twitch mouth pressure ( P mo,tw ) elicited by magnetic stimulation of the phrenic nerves.Furthermore, we explored the relationship between respiratory muscle function and motor function, and included patients with childhood-onset NM with slowness of movements.

Study design
This cross-sectional study was performed between October 2018 and September 2020 at the Radboud University Medical Centre, Nijmegen, the Netherlands.Ethical approval (NL65214.091.18) was obtained from the local ethical committee and written informed consent was obtained from all patients or legal representatives.

Study population
Recruitment was performed by contacting patients with NM who had previously visited the outpatient clinic.Moreover, we reached out to clinicians in neuromuscular centres in the Netherlands and to patients via the Dutch patient organisation.The inclusion criteria were genetically confirmed NM or the combination of an NM clinical phenotype, a biopsy confirming NM, and a first degree family member with genetically confirmed NM.The age range included was six to 80 years to increase the likelihood for patients to be cognitively and physically able to carry out the tests.Patients with other conditions affecting pulmonary or respiratory muscle function were excluded.Patient demographics (age, sex, BMI) and genotype were systematically collected.

Form of NM and motor function
The form of NM, ambulation, requirement for home mechanical ventilation (HMV), and spinal deformity were systematically collected.We used the recently proposed revised classification, which distinguishes five forms of NM based on the age of onset and severity [ 1 , 2 ]: 1) severe NM (contractures, fractures, no respiratory effort or no movements at birth); 2) typical NM (perinatal onset and motor milestones delayed but reached); 3) mild NM (childhood or juvenile onset); 4) distal NM (distal weakness); 5) childhoodonset NM with slowness (slowness of movements and corerod histology, caused by the KBTBD13 mutation [18][19][20] ); and 6) recessive TNNT1 (former Amish) NM .Ambulation was classified as unaided, using walking aids including intermittent wheelchair use, or wheelchair-dependent. HMV was categorised as none, nocturnal non-invasive, nocturnal and daytime non-invasive, and invasive ventilation.The presence of scoliosis was assessed by physical examination and by the Adam's forward bend test [32] , and dichotomized as present or absent.The 32-item Motor Function Measure (MFM) was used to examine motor function.The total score is expressed as percentage of the maximum score.
In case the patients were not able or willing to visit the Radboud University Medical Centre, a limited protocol consisting of MFM assessment and respiratory muscle strength testing was performed during home visits.

Spirometry and respiratory muscle strength testing
Spirometry was carried out with a handheld electronic spirometer (SpiroUSB, Vyaire Medical, Mettawa, IL, USA connected to PC Spirometry software, Spida CareFusion 2.3.0.10 for Windows 7).Forced vital capacity (FVC) and vital capacity (VC) upright and supine, and forced expiratory volume in the first second (FEV1) were measured and compared with reference values [33] .Peak cough flow (PCF) was also measured.Respiratory muscle strength testing, including maximal inspiratory pressure (MIP), maximal expiratory pressure (MEP), and sniff nasal inspiratory pressure (SNIP), was performed with a handheld electronic manometer (Micro RPM, Micro Medical, Rhymney, UK connected to Puma software version 1.4.2) and compared with reference values [ 34 , 35 ].Both spirometry and respiratory muscle strength tests were performed in accordance with the standards of the American Thoracic Society and the European Respiratory Society [36][37][38] .

Respiratory muscle ultrasound
Right hemidiaphragm and intercostal muscle thickness and echogenicity was measured in supine position using an Esaote MyLab Twice ultrasound machine (Esaote SpA, Genoa, Italy) equipped with a 3-13 MHz LA533 linear transducer (Esaote SpA, Genoa, Italy) according to previously described standards [ 39 , 40 ].The ultrasound probe was placed near the zone of apposition of the diaphragm, typically in the 8 th or 9 th intercostal space, on the antero-axillary line.Thickness was measured at resting end-expiration (DT EE ), i.e. at functional residual capacity, and at maximal end-inspiration (DT EI ), i.e. at total lung capacity, from the inner part of the peritoneal layer to the inner part of the pleural layer.An average value was calculated from three separate measurements.Thickening ratio (DTR) was calculated as DT EI / DT EE [39] and absolute change in thickness was calculated as DT EI -DT EE .Thickness of the intercostal muscles was measured at resting end-expiration (IT EE ).Muscle echogenicity was measured by manually selecting a region of interest, using the trace method [41] , in the diaphragm or intercostal muscular tissue in the images at end-expiration using custom developed software in Matlab (R2018a, Mathworks, Natick, MA, USA).Echogenicity cannot be reliably measured in a diaphragm thickness of less than 1 mm, leading to exclusion of echogenicity analysis.Three measurements were averaged to calculate the mean echogenicity in grey values.Echogenicity evaluates the muscle structure and increased echogenicity may result from fibrosis, inflammation or fatty degeneration of the muscle [42] .As the reference values of echogenicity are dependent of the ultrasound machine and its settings, standardised scores (z-scores) were calculated as the number of standard deviations from the predicted value.These predicted values were calculated using a linear regression equation including age.This linear regression was obtained from diaphragm ultrasound images from 83 healthy subjects (2-80 y/o) measured in our lab with an even distribution of males and females.
2.6.Twitch mouth pressure P mo,tw was measured to obtain a nonvolitional measurement of diaphragm function.To measure P mo,tw we used an automated inspiratory-pressure trigger device, as described previously [43] .The patient wore a nose clip in a seated position and breathed through a flanged mouthpiece connected to a two-way nonrebreathing valve with an inflatable balloon connected to the inspiratory port (Series 9300, Hans Rudolph Inc, Kansas City, MO, USA).Mouth pressure was recorded from a small port near the mouthpiece that was connected to a pressure transducer (MP160 equipped with DA100C and TSD160E, connected to AcqKnowledge software version 5.0.2,Biopac systems Inc.Goleta, CA, USA).Two small figure-of-eight coils (D40 Alpha Branding Iron Coil 4179-00, Magstim, Whitland, UK) were placed bilaterally on the phrenic nerves at the posterior border of the sternomastoid muscle at the level of the cricoid cartilage.The pulses were simultaneously generated by two magnetic stimulators (Magstim 200 2 , Magstim, Whitland, UK).Optimal coil position and intensity were determined by varying the position of the coils and increasing the intensity.The stimulations were automatically delivered at a predetermined inspiratory pressure trigger of -5 cmH 2 O to ensure an open glottis.We performed five measurements with 30 seconds in between to avoid twitch potentiation [44] .The highest value was chosen.The compound muscle action potential of the diaphragm was recorded using electrodes (MP160 equipped with BIO100C and MEC110C, connected to AcqKnowledge software version 5.0.2,Biopac systems Inc.Goleta, CA, USA) positioned at the costal margin to check for correct stimulation of the diaphragm.P mo,tw values below -11 cmH 2 O were considered abnormal, indicating diaphragm dysfunction [45] .

Respiratory symptoms
We explicitly asked for the presence of respiratory symptoms and symptoms of sleep-related hypoventilation as part of the medical history.We systematically asked about exertional dyspnoea (dyspnoea when washing, dressing and/or walking), orthopnoea, and dyspnoea when bending forward.Symptoms of sleep-related hypoventilation were waking up more often in comparison with five years ago, restlessness during sleep, nightmares, morning headache, and daytime sleepiness (all at multiple days a week).

Data analysis
The data was stored in Castor EDC (Castor clinical data management platform, Amsterdam, The Netherlands) and the statistical analysis was performed using GraphPad Prism software version 5.03 (GraphPad Software, San Diego, CA, USA).Results are presented in two different categorisations: form of NM and degree of respiratory involvement.The latter is based on the presence of a restrictive lung pattern, i.e.FVC < 80% of predicted, and use of HMV.
Results are expressed as median [interquartile range (IQR)].Kruskal-Wallis tests with Dunns post-test analysis were used to compare the groups.Chi-squared tests were used to compare categorical data.Spearman tests assessed correlation analyses.A pvalue < 0.05 was considered to be statistically significant.

Patient characteristics
Forty-three female patients and 21 male patients with NM were identified, of whom five patients did not fulfil the inclusion criteria because their diagnosis had not been genetically verified.Seventeen patients declined to participate because of various reasons, including the visit considered to be too strenuous or timeconsuming.Forty-two patients with NM participated in this study, including ten male patients (24%; aged 25 to 74 years) and 32 female patients (76%; aged 11 to 76 years).

Form of NM and motor function
The forms of NM were either typical (n = 11), mild (n = 7), or childhood-onset with slowness (n = 24).Walking aids were used by 12 patients and six patients were wheelchair-dependent.A total of 11 patients required HMV.Within this group, seven patients required only nocturnal non-invasive HMV, two NEB patients required daytime and nocturnal non-invasive HMV, and two ACTA1 patients required continuous invasive HMV via tracheostomy.The initial indication for HMV in these 11 patients was sleep-related    1 .

Spirometry and respiratory muscle strength testing
Results of spirometry and respiratory muscle strength testing are presented in Table 2 and Fig. 1 , subdivided based on NM forms.At group level, patients with a typical form showed a lower FVC, VC, PCF, MIP, MEP and SNIP in comparison with patients with childhood-onset NM with slowness.However, a large variation in the degree of respiratory muscle involvement was observed within the three groups.All NEB (n = 4) and ACTA1 (n = 3) patients and eight of the 19 KBTBD13 patients had an FVC < 80% of predicted.FEV1/FVC was above 70% in all patients, thus no obstructive lung pattern was present.MIP was correlated with FVC ( Fig. 2 A).There was a low correlation between MFM scores and FVC ( Fig. 2 B).
To further characterise the patients, they were subdivided based on the presence of a restrictive lung pattern, i.e. an FVC of less than 80% predicted, and requirement for HMV.One ACTA1 patient was excluded from this subdivision as spirometry could not be performed due to severe respiratory muscle weakness.MIP and MEP were decreased in patients requiring HMV compared to nonventilated patients with FVC ≥ 80% of predicted ( Table 3 , Fig. 3 A and 3 B).MEP was decreased in comparison with patients with FVC < 80% of predicted without HMV too.The spirometry and diaphragm and intercostal muscle ultrasound results are obtained from patients who visited the hospital, i.e. 10 patients with a typical, 5 patient with a mild, and 19 patients with a childhood-onset NM with slowness form.The PCF result of one patient with a typical form is missing.Three patients were excluded from diaphragm echogenicity measurement due to a diaphragm thickness of less than 1 mm.Pmo,tw was measured in 5,

Diaphragm and intercostal muscle function
At group level, there was no difference in diaphragm and intercostal muscle thickness, and echogenicity between the different forms of NM ( Table 2 ).Due to declination and contraindications for magnetic stimulation, P mo,tw was measured in a selected number of adult patients (n = 25).P mo,tw was decreased in typical NM in comparison with childhood-onset NM with slowness.In Table 3 , results of diaphragm and intercostal muscle measurements are shown in patients with and without restrictive lung pattern and HMV.There was no difference in DT EE between groups, but DT EI , absolute change in thickness, and DTR in patients requiring HMV were lower in comparison with patients with FVC ≥ 80% predicted ( Fig. 3 C).IT EE showed no difference between the groups.Three patients were excluded from diaphragm echogenicity measurement due to a diaphragm thickness of less than 1 mm.There were no differences in echogenicity of the diaphragm and intercostal muscles between patients, but the majority had a higher echogenicity than predicted, indicated by a z-score > 0 ( Fig. 3 D).P mo,tw was decreased in patients requiring HMV in comparison with patients with FVC ≥ 80% predicted.

Respiratory symptoms
Table 3 shows the respiratory symptoms in groups with different degrees of respiratory muscle weakness.Complaints indicative of respiratory muscle weakness were present in all patients.

Discussion
This study showed that respiratory muscle weakness is present in a large proportion (59%) of patients with NM of all clinical forms.In contrast to previous findings, we found that respiratory muscle weakness is also present in patients with childhood-onset NM with slowness of movements.There Table 3 Results from respiratory muscle strength testing, diaphragm and intercostal muscle ultrasound, twitch mouth pressure, and respiratory symptoms in patients with FVC ≥ 80% predicted, FVC < 80% predicted without HMV, and FVC < 80% predicted with HMV.  was a low correlation between motor function and the degree of respiratory muscle weakness in our cohort, indicating that respiratory muscle function can be compromised independently of motor performance.Furthermore, we showed, using diaphragm ultrasound and twitch mouth pressure, that impaired diaphragm function is a key component of respiratory muscle weakness in NM.
Respiratory muscle function was most severely compromised in the typical form of NM.However, within the different forms a large variability was observed.This variability is partly explained by distinct differences in respiratory muscle function in relation to genotypes, but also by variation within genotypes.All patients with NEB and ACTA1 mutations included had respiratory muscle weakness and most of these patients required non-invasive or invasive HMV, which is in accordance with previous studies [ 46 , 47 ].In contrast, respiratory muscle function varied widely in the patients with TPM2, TPM3, RYR1, CFL2, LMOD3, and MYPN mutations.The adult RYR1 patients and the adult LMOD3 patient had no respiratory muscle weakness.These findings stand out from previously described patients with severe or typical NM forms with respiratory insufficiency [ 3 , 5 , 6 ], which adds to the variability of the clinical spectrum of these genotypes.Another contrasting finding is that the included TPM2 patients showed respiratory muscle weakness with a high MFM score, while mildly affected patients with mutations in this gene usually do not have respiratory muscle weakness [14] .Interestingly, we found mild respiratory muscle weakness in eight of the 24 patients with the childhood-onset with slowness form caused by a mutation in the KBTBD13 gene, which has not been observed in previous cohort studies [ 20 , 30 , 31 ].However, respiratory function tests were not systematically performed in these studies.A case report assessing FVC in a patient with a mutation in the KBTBD13 gene did find respiratory muscle weakness [48] .Overall, the variability in respiratory muscle weakness between different NM forms and genotypes in our cohort emphasises the need for regular screening of respiratory muscle function in all patients with NM.
Importantly, we found a low correlation between motor function and spirometry.This may partly be explained by a known ceiling effect of the MFM in mildly affected patients [49] .However, 20 patients with high MFM scores (90% or more) showed a broad range of FVC values (48-93% predicted).Moreover, two NEB patients and one ACTA1 patient with respiratory involvement in our study had a mild form of NM.This confirms the clinical observation that the degree of motor function impairment is not representative for respiratory muscle function in all patients with NM, as shown previously in case reports [50][51][52] .The low correlation could also be caused by the heterogeneity in genotypes.For example, the TPM2 patients stand out with a high MFM score and a decreased FVC.Moreover, the correlation is influenced by a large group of mildly affected KBTBD13 patients.We cannot exclude that there is a stronger correlation between motor function and spirometry in specific genotypes or clinical forms of NM.The monocentric nature of our study and consequently the relatively low number of patients per genotype or clinical form of NM limits these subgroup analyses.
Diaphragm and intercostal muscle thickness and echogenicity did not differ between the different forms of NM.However, echogenicity of these respiratory muscles was higher than their predicted normal value in most patients.This finding is in line with a previous study which found high muscle echogenicity in peripheral muscles in NM caused by fatty degeneration [53] .A possible explanation for the lack of difference between the forms of NM might be that structural changes in the respiratory muscles precede changes in pulmonary function.In patients receiving HMV impaired diaphragm function is a key component of impaired respiratory function, shown by a decreased diaphragm thickness (end-inspiration) and a decreased P mo,tw .In line with these results, previous post-mortem studies in cases with severe respiratory insufficency have found a high concentration of nemaline rods in the diaphragm, indicating an important role of diaphragm involvement [ 54 , 55 ].Whether respiratory muscle ultrasound is suitable as an outcome measure for respiratory function in NM patients should be evaluated in future studies.
All patients reported respiratory symptoms regardless of the degree of respiratory muscle weakness.Thus, in our study respiratory symptoms were not indicative of the severity of respiratory muscle weakness.
This study has several limitations.First, the number of patients for each form of NM, except childhood-onset NM with slowness, was limited.Second, genetically unresolved cases were not included.Third, no patients with severe NM were included.This led to a bias towards less severely affected patients.Fourth, there was a female predominance, which cannot be explained by a selection bias.This predominance of female patients was not reported in other studies [ 24 , 29 ].Hence, because of this coincidence, these data are more representative for female patients.Last, were unable to perform a quantitative analysis on the reported respiratory symptoms, as we did not use a validated questionnaire.
Given that respiratory muscle weakness is an important feature in all forms of NM, future longitudinal studies and clinical trials should evaluate respiratory muscle function as well as general motor function.Longitudinal studies will be able to determine the intervals needed for respiratory monitoring by assessing the progression of respiratory muscle weakness.Relevant outcome measures for a longitudinal study of nemaline myopathy, including respiratory outcomes, have also been discussed [56] .Furthermore, these studies should also focus on which measurements and corresponding cut-off values can predict the need for HMV.Interestingly, diaphragm ultrasound has recently been shown as a novel predictor of nocturnal hypercapnia in slowly progressive myopathies [57] .
In conclusion, respiratory muscle weakness in NM is present in all clinical forms and genotypes, and may be present irrespective of the degree of motor function impairment.Diaphragm dysfunction was identified as an important contributor to respiratory muscle weakness.In contrast to previous findings, we also found respiratory muscles weakness in patients with childhood-onset NM with slowness of movements.Therefore, screening of respiratory function is important to guide respiratory management in all patients with NM.

Fig. 1 .
Fig. 1.FVC (A), PCF (B), MIP (C) and MEP (D) in patients with a typical, mild or childhood-onset NM with slowness form.The genotypes are represented by different colours and the horizontal bars represent the median.Significant differences are indicated by p-values of the Dunns post-test.FVC, forced vital capacity.PCF, peak cough flow.MIP, maximal inspiratory pressure.MEP, maximal expiratory pressure.NM, nemaline myopathy hypoventilation in seven patients and daytime hypercapnia in three patients.In one patient it was not specified other than respiratory insufficiency.Scoliosis was present in 11 patients.Seven patients were visited at home.The patient demographics and characteristics subdivided based on NM form are displayed in Table1.

Fig. 2 .
Fig. 2. Correlation between MIP and FVC (A), and MFM and FVC (B).The genotypes are represented by different colours.The dotted vertical line represents the definition of restrictive lung pattern (less than 80% of predicted FVC).FVC, forced vital capacity.MIP, maximal inspiratory pressure.MFM, Motor Function Measure

Table 1
Demographics and characteristics of included patients with nemaline myopathy.

Table 2
Spirometry, respiratory muscle strength tests, diaphragm and intercostal muscle ultrasound, and twitch mouth pressure results in patients with a typical, mild, or a childhood-onset NM with slowness form.
Three patients were excluded from diaphragm echogenicity measurement due to a diaphragm thickness of less than 1 mm.P mo,tw was measured in 11 patients with FVC ≥ 80% predicted, 10 with FVC < 80% predicted without HMV, and 4 with FVC < 80% predicted with HMV.Data are presented as median[IQR].Significant in comparison to patients with FVC < 80% without HMV.HMV, home mechanical ventilation.FVC, forced vital capacity.NM, nemaline myopathy.BMI, body mass index.MFM, Motor Function Measure.MIP, maximal inspiratory pressure.MEP, maximal expiratory pressure.SNIP, sniff nasal inspiratory pressure.DT EE , endexpiratory diaphragm thickness.DT EI , maximal end-inspiratory diaphragm thickness.DT EI -DT EE, absolute change in diaphragm thickness.DTR, thickening ratio diaphragm.IT EE , end-expiratory intercostal muscle thickness.P mo,tw , twitch mouth pressure.IQR, interquartile range.