Functional Anatomy of the Diaphragm and Other Respiratory Muscles

• ▪
  

  Primary inspiratory muscles are those respiratory muscles that are phasically recruited with each ventilatory effort, under resting conditions. In humans, the diaphragm, parasternal intercostals, and scalene muscles are considered primary.

  
  
• ▪
  

  The diaphragm is responsible for expanding the lower chest cage. The costal portion is opposed to the lower six ribs, producing a zone of apposition occupying about 30% of the total surface area of the rib cage.

  
  
• ▪
  

  Diaphragm action on the lower rib cage is mediated by two mechanisms:

  
  
  
  
  • ○
    

    Appositional action : with contraction, the diaphragm descends meeting abdominal resistance to its excursion. Abdominal pressure rises, which is transmitted through the zone of apposition to expand the lower rib cage ( Fig. 51-1 ).

    
    

    
    

    
    

    

    
    

    
    

    Figure demonstrating appositional action of the diaphragm. Note: the zone of apposition of the costal diaphragm with the lower six ribs is well depicted. With diaphragm contraction, the muscle moves down like a piston and meets resistance from the relatively incompressible abdominal contents. Intra-abdominal pressure rises, which is transmitted through the zone of apposition, to expand the lower rib cage.

    
    

    (From De Troyer Estenne M. Functional anatomy of the respiratory muscles. Clin Chest Med 1988;9:175-193, with permission.)

    
    
    
    
  • ○
    

    Insertional action : the alignment of costal diaphragm fibers are such that with abdominal resistance to excursion, the vector of the forces generated by contraction is oriented cranially, such that the ribs are pulled up and out ( Fig. 51-2 ).

    
    

    
    

    
    

    

    
    

    
    

    Figure demonstrating insertional action of the diaphragm . Note: the orientation of diaphragm muscle fibers are such that with contraction, the main vector of the forces generated, pulls the lower ribs up and out, to assist with expansion of the lower rib cage.

    
    

    (From De Troyer Estenne M. Functional anatomy of the respiratory muscles. Clin Chest Med 1988;9:175-193, with permission.)

    
    

  
  
  
  
• ▪
  

  Abnormal geometric configuration of the chest cage affects diaphragm function. For example, with hyperinflation, the diaphragm becomes flatter and shorter, reducing its force, generating potential by reducing or eliminating its appositional action and altering its insertional action such that fiber orientation and direction of force vectors are altered to produce in extreme cases, ribcage paradox (i.e., inward movement of the rib cage).

  
  
• ▪
  

  The diaphragm is innervated by the right and left phrenic nerves, which arise from cervical roots 3, 4, and 5.

  
  
• ▪
  

  Other primary inspiratory muscles : the parasternal intercostals are the interchondral portion of the internal intercostals and produce cranial motion of the ribs and an increased anteroposterior dimension of the upper rib cage. They are innervated by the intercostal nerves. The scalene neck muscles increase the anteroposterior and transverse diameters of the upper rib cage and elevate the sternum. They are innervated by cervical nerves C4–C8.

  
  
• ▪
  

  Accessory respiratory muscles are those recruited under conditions of increased demand (e.g., load, chemical drive, and exercise). These include the external intercostals and levator costae, which generally function as reserve inspiratory muscle groups, the sternocleidomastoids (inspiratory, innervated by the X1 cranial nerves), and the internal interosseous intercostals, which generally subserve an expiratory function.

  
  
• ▪
  

  Expiratory muscles: the abdominal muscles are silent at rest but exert phasic expiratory influences on the chest cage under conditions of increased ventilatory demand. In addition, they subserve an important functional role in the act of coughing. They are innervated by branches of T7 to T12.

Introduction to Diaphragm Paralysis

• ▪
  

  Paralysis or severe weakness of the diaphragm can be either unilateral or bilateral.

  
  
• ▪
  

  Paralysis can occur as an isolated phenomenon or as part of a more generalized neuromuscular disorder.

  
  
• ▪
  

  The impact on ventilatory function depends on the degree of weakness, the extent of muscle involvement (bilateral more than unilateral), the rapidity of onset, the extent of other inspiratory muscle weakness, and comorbidities (particularly cardiac, pulmonary, or chest wall abnormalities).

Clinical Features

• ▪
  

  Orthopnea is a classic symptom of bilateral diaphragm paralysis and occurs rapidly with supine posture. It may be seen with acute presentations of unilateral diaphragm paralysis.

  
  
• ▪
  

  Exertional dyspnea may be seen in patients with both bilateral and unilateral paralysis, culminating in reduced exercise tolerance.

  
  
• ▪
  

  Immersion dyspnea in water is another symptom of bilateral diaphragm paralysis. This occurs when the water is at the level of the lower ribcage.

  
  
• ▪
  

  Abdominal paradox (inward inspiratory motion of the abdomen) while supine is a key sign of bilateral diaphragm paralysis. The sign is less evident with lesser degrees of weakness and may be absent with maximum transdiaphragmatic pressure (Pdi _max ) greater than 30 cmH ₂ O. Abdominal paradox may also be noted with unilateral diaphragm paralysis.

  
  
• ▪
  

  In general, inspiratory muscle strength of patients with unilateral diaphragm paralysis is most decreased in those with associated cardiopulmonary disease.

  
  
• ▪
  

  Sleep-disordered breathing and nocturnal hypoventilation may occur with diaphragm paralysis. This condition is most commonly noted in patients in whom weakness of other inspiratory muscles is also present.

  
  
• ▪
  

  Arm elevation may result in increased metabolic and ventilatory cost, producing dyspnea.

Diagnostic Tests

• ▪
  

  Imaging:

  
  
  
  
  • ○
    

    Chest x-ray study shows elevated hemidiaphragm with unilateral diaphragm paralysis or the appearance of small lung volumes and basilar subsegmental atelectasis in the case of bilateral diaphragm paralysis, which can be misinterpreted as poor inspiratory effort. Note: An elevated hemidiaphragm lacks specificity for unilateral diaphragm paralysis (44%) with sensitivity and negative predictive values of 90% and 93%, respectively.

    
    
  • ○
    

    Fluoroscopy , demonstrating paradoxical motion of a hemidiaphragm when supine following a vigorous sniff, is useful with unilateral paralysis. However, with bilateral paralysis, the patient is unable to lie supine and the test lacks sensitivity. In the upright posture, expiratory muscle recruitment elevates the flaccid diaphragm, which then descends passively at the onset of inspiration as abdominal muscles relax.

    
    
  • ○
    

    Ultrasound can assess both diaphragm motion as well as diaphragm muscle morphometry. Ultrasound can determine diaphragm atrophy with paralysis (thickness at zone of opposition with chest wall at a functional residual capacity [FRC] <2 mm and <20% increase in thickness during inspiration). M-mode ultrasound was also useful in assessing diaphragm motion in a large cohort of pediatric patients and can be employed serially at the bedside.

  
  
  
  
• ▪
  

  Pulmonary function studies

  
  
  
  
  • ○
    

    Table 51-1 depicts pulmonary function in patients with bilateral and recent unilateral diaphragm weakness or paralysis.

    
    
    
    

    TABLE 51-1 ▪

    
    

    Pulmonary Function with Diaphragm Paralysis

    
    

    
    
    
    
    
    
    
    
    
    
    
    
    
    
    
    
    
    
    
    
    
    

    	VC (L) (% pred)	Supine fall in VC (%)	TLC (% pred)	RV (% pred)
    Bilateral	2.1 + 0.7 (48)	37 + 9	67 + 11	104 + 23
    Unilateral	2.9 + 0.9 (76)	11.8 + 8.1	88 + 17	108 + 23

     
    

    Values are means ± SD.

    
    

    RV, residual volume; TLC, total lung capacity; VC, vital capacity.
    
    
    
    
  • ○
    

    Although the vital capacity (VC) may be influenced by inspiratory and muscle strength, other factors such as compliance of the lungs or chest wall and airway closure may also contribute.

    
    
  • ○
    

    Thus, the VC exhibits poor specificity in the diagnosis of respiratory muscle weakness and is considered to be less sensitive than maximum mouth pressures in the presence of mild respiratory muscle weakness. Serial measurements in the intensive care unit (ICU) offer more reliable data and trends.

    
    
  • ○
    

    An exaggerated fall in VC when supine is a useful test.

  
  
  
  
• ▪
  

  Arterial blood gases

  
  
  
  
  • ○
    

    Awake arterial blood gases are relatively insensitive indices of respiratory muscle weakness until late and severe impairment of respiratory muscles is present.

    
    
  • ○
    

    In patients with chronic myopathies and no lung disease per se, awake hypercapnia was unlikely, until respiratory muscle strength was reduced to less than 40% of predicted and the VC to less than 50%.

    
    
  • ○
    

    Borderline hypercapnia may rapidly change to overt respiratory failure when even mild changes in load or ventilatory drive ensue.

  
  
  
  
• ▪
  

  Measurement of respiratory muscle pressures in the workup of diaphragm dysfunction

  
  
  
  
  • ○
    

    Maximum inspiratory mouth pressure (PI _max )

    
    
    
    
    • ▪
      

      Global test of inspiratory muscle strength . Reflects the pressure generated by the inspiratory muscles (P mus) plus the passive elastic recoil pressure of the respiratory system (lung and chest wall; Prs). At FRC, Prs = 0, whereas at residual volume (RV), Prs can be considerable (e.g., −30 cmH ₂ O). Thus, measure PI _max at FRC rather than at RV.

      
      
    • ▪
      

      Normal values are depicted in Table 51-2 .

      
      
      
      

      TABLE 51-2 ▪

      
      

      NORMAL VALUES FOR TESTS OF RESPIRATORY MUSCLE PERFORMANCE

      
      

      
      
      
      
      
      
      
      
      
      
      
      
      
      
      
      
      
      
      
      
      
      
      
      
      
      
      
      
      
      
      
      
      
      
      
      
      
      
      
      
      
      
      
      
      
      
      
      
      
      
      
      
      
      
      
      
      

      Test		Male	Female	Reference
      * PI max (cmH 2 O)	(tube MP) (flanged MP)	−123 ± 21 −106 ± 30	−87 ± 15−73 ± 21
      * PE max (cmH 2 O)	(tube MP) (flanged MP)	232 ± 42 148 ± 34	152 ± 5693 ± 16
      Sniff Pes (cmH 2 O)		−105 ± 25	−89 ± 21
      Sniff Pdi (cmH 2 O)		145 ± 24	121 ± 25
      Maximum Static Pdi (Mueller) (cmH 2 O)		108 ± 30	65 ± 31
      Pdi (combined Mueller/expulsive technique) (cmH 2 O)		180 ± 14
      UMS-Tw-Pdi (cmH 2 O)		16 ± 3 (left; <10 abnormal) 12 ± 4 (right; <6 abnormal)
      BAMPS-Tw-Pdi (cmH 2 O)		27.6 ± 1.6 (male and female; < 20 abnormal)

       
      

      BAMPS, bilateral anterior magnetic phrenic stimulation; UMS-Tw-Pdi, unilateral magnetic stimulation, twitch Pdi.

      
      

      * Normal values in the elderly and in children are available.

      
      
      
      
    • ▪
      

      Major disadvantage : test is effort dependent and requires patient motivation, cooperation, and coordination. In the presence of intrinsic positive end-expiratory pressure (PEEP), the pressure generated by the inspiratory muscles will be underestimated.

    
    
    
    
  • ○
    

    Maximal esophageal (Pes) and nasal (Pnas, sn) pressures:

    
    
    
    
    • ▪
      

      Also global inspiratory measures but less variances. Requires insertion of an esophageal balloon catheter or nasal catheter.

      
      
    • ▪
      

      In nonintubated patients, Pes following a maximum sniff may be easier to perform.

      
      
    • ▪
      

      Pnas, sn, using a catheter wedged in one nostril, may represent a relatively noninvasive measure mirroring Pes measurements. Pnas, sn tends to underestimate Pes, sn in patients with lung disease (e.g., chronic obstructive pulmonary disease [COPD]).

    
    
    
    
  • ○
    

    Pdi

    
    
    
    
    • ▪
      

      Measured as the difference between gastric pressure (Pga) and Pes (i.e., Pdi = Pga − Pes). Is a measure of diaphragm strength .

      
      
    • ▪
      

      Maximal Pdi (Pdi _max ) may be achieved using different techniques: a maximal static inspiratory effort (Mueller maneuver), a Mueller maneuver combined with an abdominal expulsive effort, and a maximal sniff (sniff Pdi).

      
      
    • ▪
      

      Such efforts require patient motivation, cooperation, and coordination, which may be difficult to achieve in critically ill patients. (See Table 51-2 for normal values.)

      
      
    • ▪
      

      Phrenic nerve stimulation at FRC provides an objective, reproducible nonvolitional test.

      
      
    • ▪
      

      This can be achieved using either electrical stimulation (surface or needle electrodes) or magnetic stimulation in which rapidly changing magnetic fields produce brief electric fields within conducting tissues.

      
      
    • ▪
      

      Although cervical magnetic stimulation produces bilateral phrenic stimulation, it requires the stimulating coil to be positioned behind the neck, making this technically challenging in ICU patients, who are commonly supine. Unilateral or bilateral anterior magnetic stimulation approaches to stimulate the phrenic nerves are better suited to the ICU environment ( Fig. 51-3 ).

      
      

      
      

      
      

      

      
      

      
      

      Bilateral anterior magnetic stimulation of phrenic nerves.

      
      

      (From Polkey MI, Moxham J. Clinical aspects of respiratory muscle dysfunction in the critically ill. Chest 2001;119:926-939, with permission.)

      
      
      
      
    • ▪
      

      Electrical or magnetic phrenic nerve stimulation uses single supramaximal stimuli to produce a twitch (Tw) Pdi. In general, bilateral phrenic nerve Tw-Pdis are about a quarter of the Pdi _max achieved by maximum voluntary effort (see Table 51-2 ).

    
    
    
    
  • ○
    

    Data with diaphragm paralysis:

    
    
    
    
    • ▪
      

      Table 51-3 summarizes mean respiratory pressures generated in patients with bilateral severe diaphragm weakness/paralysis or hemidiaphragm paralysis.

      
      
      
      

      TABLE 51-3 ▪

      
      

      RESPIRATORY MUSCLE PRESSURES WITH DIAPHRAGM PARALYSIS

      
      

      
      
      
      
      
      
      
      
      
      
      
      
      
      
      
      
      
      
      
      
      
      
      
      

      	PI max (% pred)	PE max (% pred)	Pdi max (Mueller)	Pdi max (Sniff)	TwPdi
      Bilateral	4.0 ± 18 (43)	135 ± 51 (98)	11 ± 8	13 ± 6	0.8 ± 2
      Unilateral	49 ± 9 (62)	112 ± 53 (95)	*	62 ± 13: 1.2 ± 1.6 (affected) 11.1 ± 3.5 (normal)

       
      

      Values are means ± SD.

      
      

      Pdi, transdiaphragmatic pressure; PI _max , maximum inspiratory mouth pressure; PE _max , maximum expiratory mouth pressure; Tw, twitch.

      
      

      Data from Laroche CM, Carroll N, Moxham J, Green M. Clinical significance of severe isolated diaphragm weakness. Am Rev Respir Dis 1988;138: 862-866; and Laroche CM, Mier AK, Moxham J, Green M. Diaphragm strength in patients with recent hemidiaphragm paralysis. Thorax 1988;43:170-174.

      
      

      * Maximum static values variable and depend on gender.

      
      
      
      
    • ▪
      

      Typical tracings of pressure deflections for Pdi and its components (gastric and esophageal pressures) following a sniff maneuver in a patient with bilateral diaphragm weakness and a control subject are depicted in Figure 51-4 .

      
      

      
      

      
      

      

      
      

      
      

      Serial measurement of maximum transdiaphragmatic pressures (Pdi) in a patient with severe diaphragm weakness compared to a normal subject.

      
      

      (From Spiteri MA, Mier AK, Pantin CF, Green M. Bilateral diaphragm weakness. Thorax 1985;40:631-632, with permission.)

      
      
      
      
    • ▪
      

      The ratio of ▵ Pga to ▵ Pes is normally less than −1. With diaphragm paralysis, other chest and neck inspiratory muscles are recruited such that the ratio approaches +1. This index, which correlates with Pdi _max , may, thus, be a useful index of diaphragm weakness.

    
    
    
    
  • ○
    

    Nerve conduction and electrophysiology:

    
    
    
    
    • ▪
      

      Phrenic nerve conduction time ( PNCT) may also be measured following unilateral phrenic nerve stimulation (normal in adults: 6–8 ms). In general, marked prolongation of PNCT suggests a demyelinating process, whereas preserved PNCT coupled with reduced amplitude of the compound action potential suggests axonal damage.

      
      
    • ▪
      

      Although surface and needle electrodes have been used, the recent use of esophageal electrodes offers precise and reproducible measurement of PNCT and amplitude of the diaphragm compound action potential.

    
    

  
  

Causes of Diaphragm Paralysis

A comprehensive classification of diaphragm paralysis is provided in Table 51-4 .

Specific Conditions

• ▪
  

  Phrenic nerve injury after cardiac surgery:

  
  
  
  
  • ○
    

    Phrenic nerve injury occurs because of phrenic nerve cooling when ice slush is placed around the heart for cooling. Complete phrenic nerve conduction block has been well documented with cooling in dogs, and more recently, changes in the evoked electromyographic (EMG) response of the diaphragm have been demonstrated in patients undergoing cardiac surgery using bilateral magnetic phrenic stimulation. Note: The latter occurred with mild degrees of hypothermia (31°C).

    
    
  • ○
    

    Internal mammary harvesting may also increase the risk of phrenic injury.

    
    
  • ○
    

    Traction and vascular compromise of the phrenic nerve may contribute to injury.

    
    
  • ○
    

    The incidence of phrenic injury following cardiac surgery has decreased with the introduction of preventive measures, including insulation. Prior incidence has been reported as 10% to 36%.

    
    
  • ○
    

    Diehl and coworkers reported the prevalence of clinically significant diaphragm dysfunction to be 2.1% (ice slush) and 0.5% (with insulation). In general, bilateral diaphragm paralysis is rare (<5% of cases with phrenic injury).

    
    
  • ○
    

    A low incidence has recently been described in children (0.28% bilateral diaphragm paralysis).

    
    
  • ○
    

    Phrenic dysfunction has also been described after heart-lung and lung transplantation.

    
    
  • ○
    

    The clinical presentation of diaphragm paralysis may be acute or subacute and may contribute to significant morbidity and/or mortality (see clinical features discussed earlier).

    
    
  • ○
    

    The majority of affected patients recover function. With injury limited to the myelin sheath, recovery by 12 weeks is expected. With accompanying axonal damage, recovery is further delayed and full recovery can take 1 to 2 years. About 30% of patients fail to fully recover diaphragm function.

    
    
  • ○
    

    Supportive ventilatory measures may be required in the acute or symptomatic chronic dysfunction. In the latter scenario, diaphragm plication (either unilateral or bilateral) is a therapeutic option (see discussion of treatment later).

  
  
  
  
• ▪
  

  Neuralgic amyotrophy:

  
  
  
  
  • ○
    

    Neuralgic amyotrophy is an idiopathic inflammatory condition affecting the bronchial plexus, which may lead to unilateral or bilateral diaphragm paralysis.

    
    
  • ○
    

    Patients may present with dyspnea (particularly with exercise or with immersion) after a prodromal flulike illness in which neck and shoulder pain are prominent.

    
    
  • ○
    

    Nocturnal desaturation and obstructive sleep apnea have been described, possibly related to involvement of upper airway muscles.

    
    
  • ○
    

    Upper limb weakness/paralysis resolves within 3 years in most cases.

    
    
  • ○
    

    Long-term recovery of diaphragm strength was delayed and generally took greater than 3 years.

    
    
  • ○
    

    No specific therapy is available. In patients with delayed recovery (>2–3 years), diaphragm plication can be considered.

  
  
  
  
• ▪
  

  Systemic lupus erythematosis (SLE)

  
  
  
  
  • ○
    

    The mechanisms responsible for producing the shrinking lung syndrome in patients with SLE are not well understood.

    
    
  • ○
    

    Pathogenic mechanisms include ( ) demyelinating phrenic neuropathy, which may be responsive to corticosteroid therapy, ( ) diaphragm myopathy, and ( ) chest wall restriction rather than diaphragm weakness per se. Diaphragm fiber atrophy and fibrosis have been described in an autopsy study.

  
  
  
  
• ▪
  

  Spinal cord trauma

  
  
  
  
  • ○
    

    Complete cord injury at levels C1–C3: produce bilateral diaphragm paralysis as well as paralysis of other primary inspiratory and expiratory muscles.

    
    
  • ○
    

    Complete cord injury at levels C3–C5: produce variable loss of diaphragm force–generating ability, as well as paralysis of other primary inspiratory and expiratory muscles.

    
    
  • ○
    

    Complete cord injury at levels C6–C8: diaphragm and neck accessory muscles intact.

    
    
  • ○
    

    A recent study revealed the following on hospital discharge or transfer : 100% ventilator dependence with complete cord injury of C4 and higher; 56% with C5 injury; 15% with C6 injury.

    
    
  • ○
    

    Need for long-term mechanical support depends on complete versus partial spinal cord injury, the level of the injury, resolution of edema and inflammation with apparent descent of the injury level, onset of muscle spasticity and rib cage stiffening, age, comorbidities and complications.

  
  

Treatment of Diaphragm Paralysis

• ▪
  

  Diaphragm plication:

  
  
  
  
  • ○
    

    Surgical plication of the diaphragm is a technique in which sutures are placed in the paralyzed hemidiaphragm in order to render it taut, thus preventing it from being sucked up into the thorax with inspiration. On chest x-ray study, the diaphragm is usually elevated to halfway up the chest cavity. Plication puts the diaphragm back in its normal position.

    
    
  • ○
    

    This procedure is mostly performed in patients with symptomatic unilateral diaphragm paralysis. This procedure should not be performed until enough time has passed to allow a phrenic nerve to recover function (usually 1 year). Note: bilateral diaphragm plication has been used in selected cases with good long-term success.

    
    
  • ○
    

    The procedure can be performed using a video-assisted thorascopic approach, conventional posterolateral thoracotomy, or via a laparoscopic approach.

    
    
  • ○
    

    Diaphragm plication has been associated with improved gas exchange, chest mechanics (including supine VC), respiratory muscle function (including Pdi _max ), exercise performance, symptoms, and social well-being.

    
    
  • ○
    

    In children, diaphragm plication may also facilitate weaning from mechanical ventilation.

    
    
  • ○
    

    Sustained long-term effects (up to 10 years) have been reported.

    
    
  • ○
    

    Abdominal compartment syndrome has been reported as a rare complication of diaphragm placation.

  
  
  
  
• ▪
  

  Diaphragm pacing

  
  
  
  
  • ○
    

    This technique may be considered in patients with high cervical spinal cord injuries (above C3), that is, high quadriplegia. Advantages: independence from mechanical ventilation, the ability to speak, normalized gas exchange, reduced infections, and improved quality of life.

    
    
  • ○
    

    Major prerequisites: intact lower phrenic motor neurons, normal cognitive function and absence of recovery of diaphragm function after a period of at least 4 months.

    
    
  • ○
    

    If the phrenic nerve is not functional, an intercostal nerve is grafted to the phrenic nerve. The pacing wires are attached to the intercostal nerve. This procedure is successful in 50% of cases. Re-enervation of the phrenic nerve takes 6 months, so there is a considerable delay before success can be determined.

    
    
  • ○
    

    Diaphragm pacing electrodes can be inserted around phrenic nerve trunks in the neck (cervical approach) or in the chest (standard thoracotomy or thoracoscopic approach).

    
    
  • ○
    

    There is also increasing experience with intramuscular implantation of pacing electrodes into the diaphragm laparoscopically.

    
    
  • ○
    

    Once electrodes are inserted, conditioning regimens of gradually increasing periods of diaphragm pacing are required for up to 9 months before full uninterrupted pacing.

    
    
  • ○
    

    Long-term favorable results have been reported in patients with quadriplegia, who have been fully paced for a mean period of 14.8 years.

    
    
  • ○
    

    Combined intercostal and diaphragm pacing has recently been described in ventilator-dependent patients due to spinal cord injury.

  
  
  
  
• ▪
  

  Ventilatory support

  
  
  
  
  • ○
    

    Acute bilateral diaphragm paralysis with or without associated pulmonary complications or other inspiratory muscle involvement may require intubation and full mechanical ventilation.

    
    
  • ○
    

    Prolonged continuous or intermittent mechanical ventilatory requirements necessitate tracheostomy and positive pressure ventilation.

    
    
  • ○
    

    Symptomatic patients with isolated bilateral diaphragm paralysis may be managed with noninvasive positive pressure ventilatory support (nocturnal + intermittent daytime).

    
    
  • ○
    

    Other noninvasive devices used include the pneumobelt, rocking bed, and negative pressure devices (cuirass, pneumo-wrap).