Introduction and brief review

            Neuromuscular diseases are relatively common, with a prevalence ranging from 33 to 62.6 in 100000 people. (1,2) Acute respiratory failure with a respiratory infection is the most frequent cause of unplanned hospital admission. The incidence of myasthenia gravis (MG) and Guillain Barre syndrome (GBS) is around 0.25-2 per million people and 0.8-1.9 per 100000 people per year, respectively. (3) Around 15-20 % of patients with acute neuromuscular disease admissions suffer from significant respiratory weakness.

Airway secretion clearance is critical for maintaining respiratory health and managing acute respiratory illness. The diaphragm is the most powerful inspiratory muscle responsible for 70% of the ventilation. (4,5) The mechanical cough reflex consists of three phases: an inspiratory, a contraction, and an expiratory. During the inspiratory phase, the patient inhales up to 60%-90% of total lung capacity. In the contraction phase, the glottis closes for about 0.2 seconds and the expiratory muscles of the rib cage and abdomen contract. This results in an increase in intrathoracic pressure to 300 cmH2O. In the expiratory phase, the glottis opens rapidly during 20-40 ms generating a forceful airflow (up to 360-1000 L/min) to move secretions and trapped foreign materials upwards within the airways. (6) This pathway is impaired at multiple levels in patients with neuromuscular diseases. (1,7)

The involvement of the respiratory muscles in NMD prevents a patient from taking a deep breath, and the resultant decreased inspiratory capacity limits the pre-cough inspired volume. Decreased expiratory muscle strength and impaired glottic closure restrict the effectiveness of forced expired volumes resulting in ineffective cough. (7) Secretions accumulate in the mouth and get aspirated to the lungs due to impaired function of the vocal cords. Acute respiratory failure in NMD patients requires mechanical ventilatory support and is often precipitated by respiratory infections. (8) It is clinically believed that the ineffective cough and severe respiratory weakness increase the risk of developing LRTI. (9,10) But prospective studies to confirm the association are lacking. (7) Acute respiratory failure carries substantial risks to the patients, including prolonged invasive ventilation, worsening of muscle weakness, and death. (9)

POCUS has been used for diaphragm assessment (diaphragm excursions (DE)) to determine the ability to wean in GBS and MG patients. (4,5) Diaphragm thickening fraction (DTF) has shown a significant correlation with FVC. (9) In healthy volunteers, cough peak velocity (CPV) during the expiratory phase of cough predicts PEFR. (12) In neurocritical care patients, CPV is low in patients who failed extubation. A combined inspiratory and expiratory assessment of the respiratory muscles is shown to be feasible. (13) Laryngeal ultrasound (LUS) is a well-tolerated procedure to assess glottis movements in thyroid surgeries. (14) Lung ultrasound aeration score (LUAS) has been correlated with respiratory system compliance and acts as a "densitometer" in acute respiratory failure patients. (15)  

Statement of the problem:

In acute NMD, respiratory muscle weakness affects the cough mechanism at multiple levels. A single assessment using spirometry cannot provide information on which components of the cough mechanism are affected. 

The accepted standard for assessing respiratory muscle strength includes trans-diaphragmatic and esophageal pressure monitoring. Both measurements are invasive, labor-intensive, and difficult to measure when patients are in the wards and not on ventilatory support. Hence, inspiratory pressures, spirometry, oximetry, and capnometry are used as surrogates for respiratory muscle strength assessment. (5) The main drawbacks of these measurements include late indicators of muscle weakness, tests being more subjective are prone to errors, and do not measure all pathways leading to respiratory infections and failure.

To summarize, acute NMD patients are at significant risk of developing LRTI due to respiratory muscle weakness. There is a lack of objective bedside tools to assess the respiratory system comprehensively. The current respiratory measures cannot account for all the aspects of respiratory weakness. The results of this study will be relevant and applicable to other NMD populations and central nervous system disorders. (11)

Here, we propose a comprehensive POCUS assessment of the respiratory system - inspiratory phase by DE and DTF, expiratory phase by CPV, glottis closure by LUS, and lung compliance by LUAS. This study will help us develop a composite score using the above parameters to predict which patients will likely develop LRTI.

Conventional tests used to monitor respiratory weakness in NMD patients include FVC and PEFR, Electrical impedance tomography to describe regional ventilation and perfusion, and vocal cord dysfunction with FNL. (8,14,16,17) The comparison of ultrasound findings with these measures adds to the reliability of the former.

Hypothesis/ Research question:

Research question: Can ultrasound-based assessment (POCUS) of the respiratory system as a screening tool objectively identify acute NMD patients at risk of developing LRTI?

Hypothesis: Assessment of the individual components, namely, diaphragm thickness and excursion, cough flow velocity, laryngeal ultrasound, and lung ultrasound aeration score, can be utilized to develop a comprehensive scoring system to identify patients at risk of developing LRTI and predict the disease course in the hospital.

There are only studies of single measures to assess respiratory muscle weakness. A multiparadigm assessment of this complex process of respiration is essential for prediction and planning therapy for LRTI.

Objectives:

Primary:

-       To develop and validate a composite scoring system using POCUS to predict lower respiratory tract infections in acute NMD patients.

Secondary:

-       To estimate the correlation between spirometry and POCUS findings on respiratory muscle strength in acute NMD patients.

-       To estimate the correlation between lung ultrasound aeration score and electrical impedance tomography on regional ventilation in acute NMD patients.

-       To determine the agreement between vocal cord assessment with flexible nasolaryngoscopy and laryngeal ultrasound in acute NMD patients.

Methodology

Study design: Prospective observational study.

Study population: Acute neuromuscular patients (Guillain Barre syndrome, Myasthenia Gravis) admitted to hospital.

Study site: NIMHANS, Bengaluru

Study duration: 3 years

Inclusion criteria:

1.     Admitted patients aged 12 to 70 years with a diagnosis of acute neuromuscular disease with anticipated respiratory muscle involvement (Diagnosed cases of GBS and MG)

Exclusion criteria:

1.     Refusal of consent

2.     Patients with cardiac, renal, or liver failure

3.     Radiological uni- or bi-lateral diaphragmatic paralysis

4.     Poor ultrasound images    

5.     Severe sepsis at admission (Septic shock, multiorgan dysfunction syndrome)

Sample size: Due to the absence of precedent studies, the repeated measures nature of data, and the outcome parameter being dichotomous, simple closed-form formulations of sample size calculations are not applicable. Although simulations can be conducted to calculate power and sample size based on assumed proportions, the method falls short due to the absence of minimal data required to generate assumptions and excessive variability with a small change in assumed patient variance. We decided to keep 20 samples per parameter of the model for model development (doubling the minimal 10 samples per parameter thumb rule). Thus, 100 samples for model development and validation in separate 50 samples.

Study procedure: 

Patients satisfying the inclusion criteria will be recruited to the study after obtaining informed consent. All patients will be managed using standard existing clinical practice guidelines for the respective diagnosis. The treating physician, not part of the study, takes the treatment decisions, which will be recorded. In our institute, NMD patients undergo predominantly plasmapheresis than intravenous immunoglobulin therapy. The patients are cared for on the ward and managed in the ICU in case of respiratory deterioration. The patient’s clinical details: Age, gender, comorbid illness, clinical presentation, physical findings, radiological and laboratory reports will be collected. The course in the hospital includes clinical deterioration, need for mechanical ventilation, duration of ICU stay, and hospital stay will be collected.

Measurements:

In all the patients, the following assessments will be conducted first during recruitment (baseline) and on alternate days till discharge or 14 days.

Point-of-care ultrasound (POCUS): (Figures in supplementary material)

Intensivists (study investigators) with at least three years of experience in critical care ultrasonography will perform all the evaluations. All measurements will be taken with the patient in the semi-recumbent position at 20–40 degrees. The measurements are taken of the right hemidiaphragm. An average of three measurements will be recorded. All images and measurements will be archived for later offline analysis.

1.     Diaphragm Thickness (DT) and Diaphragm Thickening fraction (DTF)¹⁸

This measurement will be performed with a linear array (7.5-10 MHz) transducer positioned in a craniocaudal direction, perpendicular to the skin in the zone of apposition, between the mid-axillary and the antero-axillary line. The diaphragm will be identified at a depth of two to four centimeters as a three-layered structure between the peritoneal and pleural membranes. The DT (both in inspiration and expiration) will be measured perpendicular to its muscle fiber direction between the peritoneal and pleural membrane, excluding the membranes. DT will be measured in the same frozen M–mode image at three places, and an average value will be considered. The diaphragm thickens with active shortening and, therefore, DT reflects contractile activity. DTF is calculated in M-mode as

2.     Diaphragmatic excursion¹⁸ 

Using a phased-array probe (2–5 MHz) positioned just below the right costal margin at the midclavicular line, angled as much as possible cranially and perpendicular to the posterior third of the diaphragm. The diaphragm appears as a bright line covering the liver. Initially, B- mode will be applied to obtain a good image and to select the exploration line. The excursion will be quantified in M-mode with the M-line placed perpendicular to the direction of motion; the sweep speed is adjusted to around ten mm/s to obtain a minimum of three respiratory cycles within one image.

3.     Cough peak velocity (CFV) using Tissue Doppler Imaging (TDI):¹²

The TDI measurements will be done with the following adjustments in the ultrasound machine; 1. Adjusting the sample volume to include the full thickness of the diaphragm. 2. Velocity scale adjusted to slow or less than 10 cm/sec. Initially, B mode will be used to obtain a good image, and a tissue Doppler will be applied. The following measurements will be taken from the TDI waveform: The patient will be instructed to cough after deep inspiration, and the peak expiration velocity will be measured.

4.     Vocal cord assessment by laryngeal ultrasound:¹⁴

The linear probe will be placed transversely, in the midline, over the middle portion of the thyroid cartilage. Symmetry, mobility, and position of vocal cords (true or false) during breathing and phonation will be assessed. The diagnosis of vocal cord dysfunction will be based on the vocal fold’s asymmetric abduction and adduction movements. The movement of vocal folds will be assessed while whispering the vowel "e". Suppose the vocal folds are not visible from a midline approach due to acute shape or calcification of the thyroid cartilage. Each vocal fold will be assessed lateral to the cartilage in that case. Normal vocal cord movement will grade as "0", incomplete vocal cord paralysis will be graded as "1", and complete paralysis as "2".

5.     Lung aeration ultrasound score:¹⁵

The linear transducer probe (7.5 MHz) will be used for examination. The thorax will be divided into 12 quadrants (anterior, lateral, and posterior zones separated by the anterior and posterior axillary lines, and each zone divided into upper and lower portions across the mamillary lines). Intercostal spaces of each of these areas shall be scanned. Each of the 12 quadrants will be assigned a score of 0 to 3 according to a simple grading system, the lung ultrasound aeration score. (Score 0 - A-profile; Score 1 - typical B-profile, small subpleural consolidation with a clear pleural line; Score 2 - confluent B-profile, multiple subpleural consolidations, and irregular pleural line; Score 3 -consolidated lung with air bronchograms) The LUAS score (0–36) is calculated by adding up the 12 individual quadrant scores with higher scores indicating more severe aeration loss.

Respiratory system assessment:

1.     Spirometry:^19,20

The patient is assessed in a semi-recumbent position. Forced vital capacity (FVC) and forced expiratory volume in the first second (FEV1): The procedure of spirometry is performed 3 phases: 1) the patient is asked to take maximal inspiration; 2) then instructed to exhale with a maximal effort to produce a "blast" of exhalation; 3) and later continue till complete exhalation to the end of the test. Forced vital capacity (FVC) is the total volume of air exhaled throughout, and FEV1 is the volume exhaled in the first second. Peak expiratory flow rate: The patient breathes out (blows out; lasts out) maximally into the peak flow meter after taking maximum inspiration. At least five efforts must be made, of which three should fall within 10% of one another for PEFR and within 5% for FVC. The mean value is taken from the best of the three efforts. 

2.     Electrical impedance tomography:^17,21

A 16-electrode belt will be attached to the patient’s chest circumference at the level of the 4th−6th intercostal space, measured at the parasternal line. EIT measurements will be conducted at a scan rate of 50 images per second using the PulmoVista 500 EIT device (Dräger, Lübeck, Germany). EIT data will be analyzed offline using the EIT Data Analysis Tool 6.1 and PulmoVista PC Software 1.2 (Dräger, Lübeck, Germany). The reconstructed EIT images will be used to assess regional and temporal inhomogeneity of lung function. The typical workflow for EIT data analysis is the following: EIT data files will be loaded; EIT sections of 4 min for analysis are defined; EIT data are reconstructed. For EIT data reconstruction, a low-pass filter with a cutoff frequency of 50 min^−1 is applied to exclude cardiac-related variations. Within the generated tidal images, four horizontal layers for each side are defined as regions of interest and labeled from ventral to dorsal: ventral, mid-ventral, mid-dorsal, and dorsal. The global inhomogeneity (GI) index is a derived parameter estimating the extent of ventilation distribution inhomogeneity will be recorded. The images will be stored for review.

3.     Flexible nasolaryngoscopy (FNL):¹⁴

FNL will be done in all patients at admission to record baseline and later repeated when the patient has features of bulbar dysfunction or at discharge. After applying local anesthesia to the nostrils and upper airway, bedside FNL with a 2.2 mm flexible fiberoptic nasolaryngoscope will be performed. The vocal cord movement will be assessed for symmetry, mobility, and movement during breathing and phonation. The movement of vocal folds will be assessed while whispering the vowel "e". Normal vocal cord movement will grade as "0", incomplete vocal cord paralysis will be graded as "1", and complete paralysis as "2". The images will be stored for review.

Outcomes:

Lower respiratory tract infection (LRTI): (CDC criteria) The patient will be assessed for LRTI on alternate days after recruitment with the following clinical criteria.

Clinical criteria:

At least one of the following: 

1. Fever (>38.0^oC or > 100.4^oF) 

2. WBC <4000 or >12000/mm³

3. Adults ≥ 70years, altered mental status with no other recognized cause

And at least two of the following:

1. New onset purulent sputum/ increased respiratory secretions

2. New onset or worsening of cough, or dyspnea, or tachypnea

3. Rales or bronchial breath sounds

4. Worsening gas exchange (O_2­­ desaturations, increased O₂ requirements, or increased ventilator demand)

If the clinical criteria are positive, then an imaging modality as part of routine care will be performed.

Radiological criteria: two or more serial chest imaging (X-ray or CT scan) with at least one of the following:

1. New and persistent (infiltrates or consolidation or cavitation)

2. Progressive and persistent (infiltrates or consolidation or cavitation)

If both clinical and radiological criteria are positive, then LRTI is considered positive. The isolation of organisms will be performed as per the following criteria.

Microbiological criteria: Presence of at least one of the following: 

1. Positive quantitative culture or corresponding semi-quantitative culture results from minimally contaminated LRT specimens (specifically, Sputum, BAL, protected specimen brushing, or endotracheal aspirate).

2. Intracellular bacteria on a direct microscopic exam (for example, Gram’s stain).

3. Organism identified from blood.

4. Organism identified from pleural fluid (If indicated)

Statistical methods: The collected data will be entered into a Microsoft Excel sheet. The analysis will be conducted using R software. Appropriate packages used for each analysis will be represented in statistical methods. The data will be represented as mean and standard deviation or median and interquartile range or frequency and percentage as appropriate. 

Analysis for the primary objective: Analysis will be conducted using R software. Initial testing for components of the POCUS score against LRTI will be conducted using univariate binary logistic regression. One hundred samples will be selected randomly for the sample frame of 150 for the analysis. Suppose the parameters significantly predict LRTI (coefficient P<0.05). In that case, they will be entered into a generalized linear mixed effects model with a logit link, with random intercept by subject for developing the model (package "nlme" in R). 

After the development of the final model, the coefficients will be used to predict the probability of LRTI in the remaining 50 samples. Probabilities will be used to predict LRTI occurrence based on a cutoff probability of 0.5. The predicted LRTI occurrence will be compared with observed LRTI occurrence using a confusion matrix, and sensitivity, specificity, and positive and negative likelihood ratios will be calculated.

If clinically indicated, interactions between variables will be tested and excluded if found to not contribute (p>0.05). If significant collinearity is observed between model parameters (coefficient>0.5), the variable will be excluded based on the clinical importance of the said variables.

Two blinded study investigators will remeasure all the archived ultrasound images to assess for inter-observer agreement.

Analysis for the first secondary objective: Correlations between spirometry and POCUS will be conducted separately for expiratory and inspiratory variables. For expiratory, PEFR and CFV will be correlated; for inspiratory, FVC and FEV1 will be correlated with DTF and DE. Since the variables are collected at multiple time points, the correlation will be conducted using repeated measures correlation which is done by standardizing coefficient derived by constructing a linear mixed effects model using one of the variables and time as predictors with random intercept by subject. The same can be conducted using package "rmcorr" in R. P < 0.05 is considered as statistically significant.

Analysis for the second secondary objective: Correlations between the EIT and POCUS will be conducted separately for regional and overall ventilation. For regional, specific corresponding areas will be correlated, and for overall, LUAS will be correlated with the GI index. The variables are collected at multiple time points; correlation will be conducted using repeated measures correlation by standardizing coefficient derived by constructing a linear mixed effects model using one of the variables and time as predictors with random intercept by subject. The same can be conducted easily using package "rmcorr" in R. P < 0.05 is considered as statistically significant.

Analysis for the third secondary objective: Agreement will be conducted using kappa statistic (package "epitools" in R software). P < 0.05 is considered as statistically significant.

Reference:

1.     Benditt JO. Respiratory Care of Patients With Neuromuscular Disease. Respir Care. 2019 Jun;64(6):679-688.

2.     Hull J, Aniapravan R, Chan E, Chatwin M, Forton J, Gallagher J, Gibson N, Gordon J, Hughes I, McCulloch R, Russell RR, Simonds A. British Thoracic Society guideline for respiratory management of children with neuromuscular weakness. Thorax. 2012 Jul;67 Suppl 1:i1-40. 

3.     Rajagopalan S, Baghshomali S, Navarro JC, Kalanuria AA. Intensive Care Management of the Neuromuscular Patient. Textbook of Neuroanesthesia and Neurocritical Care. 2018 Dec 8:63–96. doi: 10.1007/978-981-13-3390-3_5. PMCID: PMC7120052.

4.     Watson K, Egerton T, Sheers N, Retica S, McGaw R, Clohessy T, Webster P, Berlowitz DJ. Respiratory muscle training in neuromuscular disease: a systematic review and meta-analysis. Eur Respir Rev. 2022 Nov 29;31(166):220065. 

5.     Sahni AS, Wolfe L. Respiratory Care in Neuromuscular Diseases. Respir Care. 2018 May;63(5):601-608. 

6.     Camela F, Gallucci M, Ricci G. Cough and airway clearance in Duchenne muscular dystrophy. Paediatr Respir Rev. 2019 Aug;31:35-39. 

7.     Sheers N, Howard ME, Berlowitz DJ. Respiratory adjuncts to NIV in neuromuscular disease. Respirology. 2019 Jun;24(6):512-520. 

8.     Hutchinson D, Whyte K. Neuromuscular disease and respiratory failure. Pract Neurol. 2008 Aug;8(4):229-37. 

9.     Boentert M, Wenninger S, Sansone VA. Respiratory involvement in neuromuscular disorders. Curr Opin Neurol. 2017 Oct;30(5):529-537. 

10.  Harlaar L, Ciet P, van der Ploeg AT, Brusse E, van der Beek NAME, Wielopolski PA, de Bruijne M, Tiddens HAWM, van Doorn PA. Imaging of respiratory muscles in neuromuscular disease: A review. Neuromuscul Disord. 2018 Mar;28(3):246-256. 

11.  Gipsman AI, Lapinel NC, Mayer OH. Airway clearance in patients with neuromuscular disease. Paediatr Respir Rev. 2023 Feb 23:S1526-0542(23)00005-2. 

12.  Norisue Y, Santanda T, Homma Y, Tomita S, Saito S, Kataoka J, Fujimoto Y, Nabeshima T, Tokuda Y, Fujitani S. Ultrasonographic Assessment of Passive Cephalic Excursion of Diaphragm During Cough Expiration Predicts Cough Peak Flow in Healthy Adults. Respir Care. 2019 Nov;64(11):1371-1376. 

13.  Hirolli D, Srinivasaiah B, Muthuchellappan R, Chakrabarti D. Clinical Scoring and Ultrasound-Based Diaphragm Assessment in Predicting Extubation Failure in Neurointensive Care Unit: A Single-Center Observational Study. Neurocrit Care. 2023 Mar 1. doi: 10.1007/s12028-023-01695-4. Epub ahead of print. 

14.  Sciancalepore PI, Anzivino R, Petrone P, Petrone D, Quaranta N. Clinical usefulness of transcutaneous laryngeal ultrasonography in otolaryngology practice during COVID-19 pandemic: a literature review. J Ultrasound. 2023 Mar;26(1):1-12. 

15.  Biasucci DG, Loi B, Centorrino R, Raschetti R, Piastra M, Pisapia L, Consalvo LM, Caricato A, Grieco DL, Conti G, Antonelli M, De Luca D. Ultrasound-assessed lung aeration correlates with respiratory system compliance in adults and neonates with acute hypoxemic restrictive respiratory failure: an observational prospective study. Respir Res. 2022 Dec 18;23(1):360.

16.  Brennan M, McDonnell MJ, Duignan N, Gargoum F, Rutherford RM. The use of cough peak flow in the assessment of respiratory function in clinical practice- A narrative literature review. Respir Med. 2022 Mar;193:106740. doi: 10.1016/j.rmed.2022.106740. Epub 2022 Jan 15. 

17.  Sella N, Pettenuzzo T, Zarantonello F, Andreatta G, De Cassai A, Schiavolin C, Simoni C, Pasin L, Boscolo A, Navalesi P. Electrical impedance tomography: A compass for the safe route to optimal PEEP. Respir Med. 2021 Oct;187:106555. 

18.  Fayssoil A, Behin A, Ogna A, Mompoint D, Amthor H, Clair B, Laforet P, Mansart A, Prigent H, Orlikowski D, Stojkovic T. Diaphragm: pathophysiology and ultrasound imaging in neuromuscular disorders. Journal of neuromuscular diseases. 2018 Jan 1;5(1):1-0.

19.  Ponce MC, Sankari A, Sharma S. Pulmonary Function Tests. [Updated 2022 Sep 3]. In: StatPearls [Internet]. Treasure Island (FL): StatPearls Publishing; 2023 Jan-. Available from: https://www.ncbi.nlm.nih.gov/books/NBK482339/ 

20.  Dikshit MB, Raje S, Agrawal MJ. Lung functions with spirometry: an Indian perspective--II: on the vital capacity of Indians. Indian J Physiol Pharmacol. 2005 Jul-Sep;49(3):257-70. 

21.  Dargvainis M, Ohnesorge H, Schädler D, Alkatout I, Frerichs I, Becher T. Recruitable alveolar collapse and overdistension during laparoscopic gynecological surgery and mechanical ventilation: a prospective clinical study. BMC Anesthesiol. 2022 Aug 6;22(1):251.