Spinal cord injury (SCI) is characterised by profound respiratory compromise secondary to the level of loss of motor, sensory and autonomic control associated with the injury. This review aims to detail these anatomical and physiological changes after SCI, and outline their impact on respiratory function. Injury-related impairments in strength substantially alter pulmonary mechanics, which in turn affect respiratory management and care. Options for treatments must therefore be considered in light of these limitations.
Respiratory impairment following spinal cord injury (SCI) is more severe in high cervical injuries, and is characterised by low lung volumes and a weak cough secondary to respiratory muscle weakness.
Autonomic dysfunction and early-onset sleep disordered breathing compound this respiratory compromise.
The mainstays of management following acute high cervical SCI are tracheostomy and ventilation, with noninvasive ventilation and assisted coughing techniques being important in lower cervical and thoracic level injuries.
Prompt investigation to ascertain the extent of the SCI and associated injuries, and appropriate subsequent management are important to improve outcomes.
To describe the anatomical and physiological changes after SCI and their impact on respiratory function.
To describe the changes in respiratory mechanics seen in cervical SCI and how these changes affect treatments.
To discuss the relationship between injury level and respiratory compromise following SCI, and describe those at increased risk of respiratory complications.
To present the current treatment options available and their supporting evidence.
Spinal cord injury impairs respiratory function. The associated impairments can be well managed. http://ow.ly/1Gcy305oQIY
Spinal cord injury (SCI) is thankfully a relatively rare injury, but one with profound personal and community impacts. The personal disability is typically substantial, with significant reductions in community participation, employment and quality of life (QOL) [1–3]. Access Economics estimated the cost of SCI in Australia in 2009 to be $2.0 billion: $5 million per case of paraplegia (lifetime) and almost double that ($9.5 million) for quadriplegia . In Australia, Western Europe and North America, SCI most commonly occurs secondary to motor vehicle accidents, whereas falls from trees and rooftops are the most frequent cause in South-East Asia and Oceania . Worldwide, the average person with traumatic SCI is most likely to be male, in his early 30s, with a 70% chance of being paraplegic, and a 50% likelihood of having a motor and sensory complete lesion . However, alongside increases in general population age, the average age of injury is also increasing . Traumatic SCI incidence varies from 8 to 246 cases per million inhabitants per year with a global prevalence from 236 to 1298 per million [8, 9]. A SCI may also be secondary to nontraumatic causes such as vascular abnormalities, tumours or infection , and far less is known regarding incidence and prevalence in nontraumatic SCI.
Respiratory complications remain the most common cause of mortality following SCI [11, 12]. Patients are most vulnerable to respiratory illness in the first year after injury but continue to suffer from respiratory complications throughout life. The number of respiratory complications suffered during initial admission is more important than level of injury in determining length of stay and hospital costs .
The most important determinants of the extent of respiratory compromise after SCI are the patterns and level of motor, sensory and autonomic neurological impairment. Classification of the individual SCI is made on the basis of clinical examination and comprises a two-stage process of neurological level of injury (NLI) determination and then motor and sensory function. The NLI is defined as the most caudal spinal segmental level with intact motor and sensory function. Motor and sensory impairment are scored and severity of SCI classified using the International Standards for Neurological Classification of Spinal Cord Injury (ISNCSCI) developed by the American Spinal Injury Association (ASIA) and the International Spinal Cord Society (www.iscos.org.uk/international-standards-and-spinal-cord-independence-measure-scim) . The clinical examination to determine ISNCSCI score involves the assessment of 10 muscles (motor) and 28 dermatomes (sensory) bilaterally [14, 15]. Once the NLI is determined, the injury is classified according to the ASIA Impairment Scale (AIS) (table 1).
The effects of SCI on sensory and motor functions are directly related to the local site of the cord injury and any subsequent recovery. However, the effects of SCI on the autonomic nervous system are often more global because of the anatomical organisation of the autonomic nervous system, particularly the sympathetic nervous system. The ISNCSCI taxonomy does not include any assessment of the autonomic nervous system and, while it is likely that complete paralysis (i.e. ISNCSCI A) is associated with a complete sympathetic lesion, this may not be the case in all individuals .
The ISNCSCI and separate assessments of the integrity of the autonomic nervous system describe the impairment caused by a particular injury. However, the subsequent disability, handicap and QOL are determined by the relationship between life satisfaction, perceived health status and any comorbidities that arise from the injury.
Acute SCI and respiration
Effects of SCI
The ability to breathe deeply and cough forcefully is impaired to varying degrees depending on the level and completeness of the SCI, with greater dysfunction seen at higher injury levels [17–19]. Although a number of authors have documented the changes in pulmonary function following SCI, care is required when these data are reviewed. Because pre-morbid spirometric values are generally not available in SCI, the magnitude of the drop immediately after SCI remains unknown. Similarly, the significance of any measurements obtained after injury is inferred from population values. A recent publication addressed this limitation through the development of SCI-specific prediction equations  and the model is currently being prospectively validated.
Immediately following a traumatic SCI, there is a period of spinal shock resulting in flaccid paralysis of the muscles below the level of injury, which lasts for a period of weeks to months . Flaccid paralysis of the intercostal muscles creates an unstable chest wall such that during inspiration, the negative intrathoracic pressure causes paradoxical inward depression of the ribs [22, 23]. This mechanical imbalance and disadvantage result in less efficient ventilation, increased work of breathing, and a tendency towards distal airway collapse and microatelectasis . Airway secretions may accumulate in the lungs either through increased production  or decreased clearance secondary to impaired cough. It is during this period that intubation and ventilation for respiratory support is highly likely .
With time and a reduction in active chest wall movement secondary to weakness, the tendons, ligaments and joints of the rib cage stiffen. Together with spasticity of the intercostals, this will stabilise the rib cage at a lower absolute lung volume, such that paradoxical breathing lessens [27, 28]. The resolution of spinal shock may also improve lung volumes as the thoracic and abdominal muscles begin to develop spasticity.
The risk of respiratory failure is directly associated with injury level. Patients may lose up to one AIS level within the first few days of injury as a result of cord swelling or bleeding, making this an especially high-risk period . A patient with a complete injury above C5 will typically have impaired diaphragm function and is likely to require a period of endotracheal intubation and mechanical ventilation . A C5 injury level may also involve diaphragm weakness but is more likely to be associated with the ability to breathe independently. Impaired inspiration, lack of cough strength, and no movement of the hands, trunk and lower limbs are seen. A patient with a complete T12 classification will have no observable inspiratory or expiratory impairment, full upper body strength, good trunk strength and balance but no movement of the lower limbs. An understanding of this classification allows the clinician to predict the likely needs and respiratory management of their patient (table 2).
Particularly in traumatic SCI, a number of factors may modify the degree of respiratory impairment. Patients with incomplete injuries have some feeling and, often, movement preserved below the level of injury; as such, they will relatively improve the remaining function of the muscles and systems below the lesion. Traumatic brain injury occurs in many patients with a primary diagnosis of cervical SCI , potentially modifying airway protection and motor control. Injuries that result in traumatic SCI are frequently high velocity, and may also result in thoracic cage trauma, rib fractures and internal organ damage.
Cardiovascular and autonomic impact on respiration
In the acute post-injury period of spinal shock, not only is there flaccid paralysis of the muscles below the injury, in lesions above T6, there is also a reduction in sympathetic nervous system activity  and an unopposed expression of parasympathetic activity via the vagus nerve [32–35]. Impaired control of the autonomic nervous system seen in individuals with high thoracic and cervical SCI can lead to hypotension and cardiac arrhythmias. The most common arrhythmia seen is bradycardia. Higher, more complete injuries will result in more significant arrhythmias and these are most common in the first 14 days . Patients with recurrent bradycardia may benefit from early insertion of a cardiac pacemaker .
Unopposed parasympathetic activation of the airways has been demonstrated to increase bronchial reactivity [38–40] and the effect may be reversed with ipratropium bromide, an anticholinergic agent. The same authors also observed that anticholinergic pre-treatment completely abolished airway reactivity following methacholine challenge . Notwithstanding the parasympathetic influence on the airways, the effect is minimal compared with the restrictive pulmonary deficit secondary to muscle paralysis, and appears to affect the forced expiratory volume in 1 s (FEV1) and forced vital capacity (FVC) equally; hence, there is no effect on the forced expiratory ratio. It has been demonstrated that short-term β-agonist administration in acute tetraplegia may confer some benefit ; however, there is no clear evidence of longer-term benefit from bronchodilation in cervical SCI.
Pulmonary oedema and pulmonary embolism
Pulmonary oedema can affect as many as 50% of individuals with acute tetraplegia . The causes are multifactorial and include excessive fluid resuscitation in the presence of hypotension in the acute post-injury setting. The risk of pulmonary embolism is increased following acute SCI, with an incidence of 4.5% and a mortality rate of 3.5% in the first 3 months following injury . In the first month post-injury, SCI patients have 500 times the risk of death from pulmonary embolism compared to age- and sex-matched controls without SCI . In a retrospective audit of 222 patients with SCI from 2010 to 2013, Clements et al.  found 33 had at least one pulmonary embolism and eight of these also had a deep venous thrombosis. They found occurrence of venous thromboembolism to be associated with increased weight, male sex, completeness of motor paralysis, length of stay, associated pelvic or lower limb fracture and delayed admission to the specialist SCI centre.
Lung function and respiratory muscle pressure-generating capacity change over time, and are correlated with lesion level [46, 47]. The ventilatory pattern alters significantly during the first year following injury. Immediately following injury, there is flaccid paralysis of the intercostal and abdominal muscles, with marked paradoxical abdominal and thoracic movement, and a reduction in vital capacity to approximately 20–60% of the predicted value in tetraplegia and 80–90% in paraplegia [22, 48, 49]. The paradoxical ventilation is due to the diaphragm contracting against an unstable rib cage and is more marked following cervical injuries . Truncal and intercostal tone increase with time , stabilising the rib cage and returning the vital capacity to approximately 60% of the pre-injury level . A progressive reduction in the functional residual capacity also occurs during this time, associated with atelectasis and basal pulmonary fibrosis . The residual volume remains elevated compared with normal population values [51–54], especially in the sitting position with the abdomen unbound, although the magnitude of the elevation reduces with time [51, 52]. The total lung capacity does not appear to recover significantly during this time [51, 52].
Moving from upright to supine affects the respiratory function of the tetraplegic and high paraplegic individual differently to the able-bodied person. Frisbie and Brown  compared the change in neck and abdominal circumference in 13 male subjects with tetraplegia with 11 male, age-matched controls. The waist size increased in both groups when sitting but increased by a significantly (p<0.001) greater amount in the subjects with tetraplegia (4.1±1.1 inches versus 1.6±0.8). This increase in abdominal girth when sitting in tetraplegia is secondary to decreased abdominal muscle strength  and the associated increased abdominal wall compliance [56–58]. In the seated position, the abdominal contents are less supported by the decreased abdominal wall muscle tone and fall forward, increasing the waist size and lowering the diaphragm. In able-bodied subjects, the FVC is reduced in the supine position [52, 57, 59], whereas in tetraplegia it is increased [49, 52, 53, 57, 59]. Postural changes are associated with symptoms; patients with an acute, high SCI report less breathlessness when supine compared to sitting . In the supine position, the weight of the abdominal contents forces the diaphragm to a higher resting level so that contraction produces greater absolute excursion of the diaphragm; an effect that can be increased when the person with tetraplegia is tipped 15° head down from supine such that the vital capacity rises by a further 6% .
The ability to produce an effective cough is severely impaired in patients with cervical or high thoracic SCI [62, 63]. Patients who have loss of innervation to the abdominal muscles and the internal intercostals lose the ability to produce a forced expiration . De Troyer and Estenne  have shown that patients with injuries at C5–C8 can utilise the clavicular portion of pectoralis major to generate an expulsive force, although the extent to which this is clinically important is unclear. Linn et al.  found that in a group of patients with high tetraplegia, peak expiratory flow rate was <50% of predicted normal values. The most effective peak cough flows can be achieved with a positive pressure-supported inspiration followed by expiration augmented by negative pressure or an assisted cough [66, 67].
Sleep disordered breathing and sleep apnoea syndrome
People with SCI have marked impairment of both respiratory function and strength after injury, and as such, the normal ventilatory instability associated with sleep may further compromise their already marginal ventilator system. In the able-bodied, the loss of wakefulness coincides with a decrease in alveolar ventilation and an associated 2–7-mmHg increase in arterial carbon dioxide tension . Sleep hypoventilation is secondary to a number of separate physiological processes and there is some evidence that it may be exaggerated at sleep onset in SCI .
The most significant sleep-related respiratory problem post SCI is obstructive sleep apnoea (OSA) [70, 71]. The OSA appears as a direct consequence of cervical injury  and is up to 83% prevalent in the first year. Chronically, OSA has a prevalence of 40–91% [73–78]. The prevalence in paraplegia does not appear significantly different to that of the able-bodied population; however, the literature is far less comprehensive in this group. In the general population, untreated sleep disorders are associated with cardiovascular disease and impaired cognition . Neurocognitive impairments including decreased memory and attention have also been linked to nocturnal hypoxia in tetraplegic subjects with untreated OSA . In chronic SCI, most authors have observed associations between increasing age, body mass index (BMI) and neck circumference and OSA prevalence [73, 74, 77, 81], but these relationships appear weaker acutely . It may be that OSA in tetraplegia is a biphasic disorder, acutely caused by the cervical SCI that partially resolves only to increase in prevalence again as people with SCI age and gain weight [72, 73]. An individual with acute tetraplegia and undiagnosed or untreated OSA may struggle to participate in the demanding process of rehabilitation , and any ongoing cognitive impairment  will also likely limit future employment opportunities that are skewed towards computer and desk based tasks due to the physical SCI disability.
As with all clinical practice, background anatomical, pathophysiological and disease knowledge inform the clinical assessment of patients. In SCI, much can be gleaned by careful assessment of the patient’s breathing pattern to identify paradoxical or unequal chest wall movement. Qualitative diaphragm movement can be assessed when viewing the patient in the supine position from the foot end of the bed to assess symmetry on deep inspiration. Simple spirometry, particularly serial measures, provides clinically relevant information regarding the vital capacity and any airflow obstruction, and can be undertaken regularly at the bedside. Measures of muscle strength, such as the maximal inspiratory pressure (MIP) and the most sensitive, noninvasive measure of functional respiratory muscle strength, the sniff nasal inspiratory pressure, can also be readily performed as the bedside . The ability to cough has been routinely assessed by measuring peak cough flow using an “asthma” peak flow meter; however, the reliability of these devices is poor .
Early surgical stabilisation of the spine is typical practice in the majority of spinal units globally [85, 86]. Patients are typically intubated for surgery, admitted to intensive care post-operatively and once surgically stabilised, able to be sat out of bed. However, this upright position whilst in spinal shock may lead to increased work of breathing and hypotension, which can be countered to a degree through the use of an abdominal binder and antihypotensive medication.
Lung volume maintenance
Restoration of lung volume is a mainstay of treatment for the person with acute tetraplegia or high paraplegia. Intermittent positive pressure breathing (IPPB) via a mouthpiece or facemask to support inspiration prior to manually assist coughing can augment lung volume and increase exhalation flow. Similarly, insufflation using a portable, noninvasive ventilator or an in/exsufflation device will also boost inspiratory volumes prior to assisted coughing or exsufflation [66, 67]. Both IPPB and in/exsufflation can be delivered via a variety of interfaces including mouthpiece, facemask and tracheostomy connector. The introduction of many new ventilators for noninvasive ventilation (NIV) has overcome some of the limitations of IPPB machines such as a lack of choice of interface and the requirement for pressurised gas .
Breath stacking is a technique where a resuscitation bag is used with a mouth piece or facemask to deliver two or more breaths prior to exhalation in order to augment lung volume and aid in secretion clearance. This low-cost treatment can be provided at home or in the subacute environment [88–90] and while it has been shown to slow the decline in FVC when introduced for those with neuromuscular disorders , controlled trial evidence for efficacy is lacking.
Glossopharyngeal breathing (GPB) can be used to increase lung volumes and assist secretion clearance in the person with high tetraplegia . Vital capacity may be increased by as much as 1000 mL  and in ventilator users, GPB can provide security in case of ventilator failure or for brief periods of ventilator-free breathing .
An effective cough
A large inspiratory effort followed by a quick and forceful expiration is required in order to achieve a successful cough. A compressive force directed inwards and upwards under the diaphragm, and compression of the rib cage may provide manual, expiratory assistance, and replace the work of the abdominal and internal intercostal muscles. Frownfelter and Massery  describe various methods of achieving assisted cough. The therapist must synchronise the applied compressive force with the expiratory effort of the patient. Mechanical in/exsufflation devices to assist coughing have been documented to assist with cough in patients with neuromuscular disorders and respiratory muscle weakness [66, 67, 95–97]; however, there are essentially no well-controlled, longer-term comparison data.
Respiratory muscle training and the position of the diaphragm
A recent Cochrane and associated review and meta-analysis have demonstrated that respiratory muscle training (RMT) can significantly improve respiratory muscle strength, function and endurance during the period of training in tetraplegia [98, 99]. A significant benefit of RMT was demonstrated for vital capacity (mean difference (95% CI) 0.41 (0.17–0.64) L) MIP (10.66 (3.59–17.72) cmH2O), maximal expiratory pressure 10.31 (2.80–17.82) cmH2O), maximum voluntary ventilation (17.51 (5.20–29.81) L⋅min−1) and inspiratory capacity (0.35 (0.05–0.65) L). Unfortunately, no high-quality, long-term data are available and as with all muscle training, the effect on strength appears to decay or be lost with cessation of training.
Elasticated abdominal binders are used to minimise the effect of postural hypotension and aid respiration [100–103]. Abdominal binding decreases abdominal compliance, and restores pressure transference across the thorax and abdomen, allowing the diaphragm to assume a more normal resting position in the upright posture . Wadsworth et al.  found that not only does an abdominal binder improve vital capacity, FEV1, peak expiratory flow and MIP but also increases the time the patient can sustain voice.
Secretion and ventilation management of individuals with acute cervical SCI differs from that required by individuals with pulmonary dysfunction due to non-neurological injuries . Secretions accumulate secondary to increased production , poor cough and, in some instances, aspiration of saliva. Atelectasis leads to impaired aeration, infection and pneumonia . A literature review of respiratory management during the first 6 weeks following cervical SCI showed a protocol using a combination of techniques, which may include IPPB, manually assisted coughing, respiratory muscle resistance training, NIV and/or a clinical pathway is most likely to provide positive patient outcomes .
Readiness for weaning from mechanical ventilation is indicated by a vital capacity of 15 mL⋅kg−1 or greater [107, 108], a decreasing sputum load, ability to co-operate, a patent upper airway, a relatively clear chest radiograph with no new changes, and reduction in the requirement for ventilator assistance [35, 109, 110]. Ideally, extubation of the patient with SCI is a co-ordinated decision involving medical, surgical, nursing and physiotherapy staff. Extubation should occur early in the day to allow for a period of intensive observation and chest physiotherapy in order to minimise the risk of respiratory deterioration and re-intubation. Adequate secretion clearance and techniques aimed at volume maintenance/augmentation are recommended [105, 106, 110, 111]. Extubation failure is a poor patient outcome, associated with increased mortality, tracheostomy rate and length of hospital stay [112, 113].
High-level tetraplegic patients may be safely managed without invasive ventilation or tracheostomy in units with significant SCI and NIV expertise . Whilst Berney et al.  showed early extubation and intensive physiotherapy may decrease the intensive care length of stay in suitable patients, this and the previous study by Bach  reflect the experience of specialist SCI and NIV services and, as such, may not be generalisable. Failure to intubate in a timely way in the presence of tetraplegia can lead to the need for emergency airway intervention [29, 109] or catastrophic airway loss and death . In a generalist unit not familiar with SCI-specific management, intubation and invasive ventilation followed by the use of a tracheostomy to facilitate weaning from ventilation may provide a safer option for the patient.
It therefore follows that mechanical ventilation with or without tracheostomy is usual for most patients with acute traumatic cervical SCI. Patients are more likely to progress to tracheostomy if they experience respiratory complications, or have a complete motor SCI or an admission AIS motor score of ≤22. Lesion level, as well as completeness, is important. Changes on magnetic resonance imaging at the C3 level are predictive of the need for tracheostomy  and Como et al.  found a tracheostomy rate of 81–83% in patients with a complete SCI above the level of C5. Overall tracheostomy incidence in tetraplegia is reported to be between 10% and 60% [116, 117] with FVC being an important predictor of the need for tracheostomy, with cut-points of <830 and ≤500 mL. Berney et al.  also found volume of pulmonary secretions and gas exchange to be predictive of airway management. Tracheostomy may be performed as early as day 4 following anterior cervical surgery with no increase in the incidence of wound or implant infection , and has been associated with fewer days on the ventilator and a shorter hospital stay in selected patients .
The best method of weaning from ventilation in SCI is unclear. Reports suggest it is best performed with the patient supine [28, 59, 110] or with a head down tip [120, 121], to optimise diaphragm function. Cohn  suggests that weaning should be thought of as a conditioning process for the diaphragm and warns that fatigue of the muscle should be avoided. Wallbom et al.  suggest a regime of twice-daily trials with a 3–4-h rest period in between trials and gradual increases for each session until weaning is achieved. An evidence-based protocol to guide weaning of ventilator-dependent cervical SCI patients has been shown to increase mean maximal inspiratory and expiratory pressure, mean vital capacity, and mean off-ventilator breathing times .
Speech options with a tracheostomy
Communication options for patients with tracheostomy in the intensive care setting should be explored as soon as practicable. Patients with intact hand function can write or type whilst ventilated, but this option is either impossible or very difficult for the person with tetraplegia. Lip reading, or use of eye gaze boards or eye blinks are also useful communication methods; however, all of these methods are slow, often frustrating for the patient and carer, and give rise to frequent miscommunication. Full or partial deflation of the tracheostomy cuff can allow for leakage of air across the vocal cords and therefore speech. “Leak speech” can be optimised by taking a multidisciplinary approach consulting with the speech pathologist, physiotherapist or respiratory therapist and physician [123, 124].
Ventilator adjustments to compensate for loss of air through the mouth and nose during cuff deflation will ensure the patient remains adequately ventilated . Increasing tidal volume, positive end-expiratory pressure and inspiratory time will facilitate speech by compensating for leak and allowing the patient to remain stable on ventilation . It is also possible to place a one-way speaking valve in line with the patient’s breathing circuit to redirect all exhaled air across the vocal folds . This requires the circuit to be reconfigured in order to facilitate speech in addition to cuff deflation and ventilator adjustment. Cuff deflation may also allow for use of a one-way speaking valve in the nonventilated, tracheostomised patient . Speaking valve use may increase work of breathing and monitoring for fatigue is required.
Removal of the tracheostomy tube in the patient with SCI requires assessment of cough effectiveness, airway patency and adequacy of the swallow [129, 130]. Other considerations such as co-operation of the patient, oxygen requirements, medical stability, presence of respiratory infection or pending surgery should be taken into account [131, 132]. As the incidence of aspiration is moderately high in acute tetraplegia (16–41%), speech pathology assessment is required [133, 134]. Speaking valve use during oral intake has been shown to decrease aspiration risk in nonventilated patients  and persistent aspiration following cervical SCI may not be an obstacle to decannulation if a risk management approach is taken . Management of the SCI patient by a specialist multidisciplinary tracheostomy team can decrease total cannulation time and acute length of stay with associated cost savings . While most patients will be able to clear secretions post-tracheostomy removal by use of mechanical in/exsufflation or manually assisted coughing [110, 111], insertion of a minitracheostomy may be considered. Minitracheostomy is also useful where manually assisted coughing is contraindicated such as in the presence of intra-abdominal trauma or fractured ribs.
With increased survival after high SCI, long-term ventilation is now 6–8% prevalent in the developed world [119, 138–140]. It is a common assumption amongst the able-bodied that life with a high SCI would not be worth living ; however, research into perceived QOL in the ventilator-associated SCI group indicates that this is not the case. The majority of those living in the community report good or excellent QOL and, in some cases, higher QOL than their non-ventilator associated counterparts . There is however, a discrepancy in QOL in the first year following discharge from hospital in ventilator users and non-ventilator users with SCI . This may indicate a period of increased stress for those with SCI and their carers as they adapt to life in the community with high care needs. Discharging a person with full time ventilation into the community requires a coordinated multidisciplinary team and liaison with the person’s local community. Comprehensive training, back-up equipment and planning for unusual events such as power failures need to be considered to ensure the safety of the ventilated person. Peer support can be useful to allay fear and assist with practical planning. Annual medical evaluations of the patient’s ability to breathe may identify delayed diaphragm recovery .
Readmissions to hospital with a chest infection
As survival of individuals following SCI increases, the impact of ageing on respiratory health and wellbeing has become more evident [144, 145]. Postma et al.  found that pulmonary function tests were a stronger predictor of respiratory infection in the first year following injury than lesion level or completeness and similarly, Raab et al.  observed a linear relationship between inspiratory muscle strength and pneumonia. Patients with a vital capacity of <2 L are at greatest risk of developing late-onset respiratory failure , and an association between decline in vital capacity over time with increased BMI and an age at injury >30 years has been described .
Smoking cessation is a clear priority for those with SCI, particularly for those with higher lesions and impaired cough effectiveness. Saunders et al.  found a higher incidence of smokers compared to the rest of the population in their study of 833 American adults with SCI. They also found that those with lower socioeconomic status were more likely to smoke, as were those with poor access to care.
Treatment of OSA
The gold standard for treatment of OSA remains continuous positive pressure therapy (CPAP), but only 20–50% of people with SCI and OSA are able to adhere with CPAP [74, 76, 77, 150]. People with tetraplegia are far more likely to need CPAP for OSA, yet upper-limb motor dysfunction, reduced independence for donning and doffing of masks, and the associated need for increased carer support time challenge adherence. The less claustrophobic nature of nasal pillows may mean that they are well suited to tetraplegic users and the lower pressures required by many tetraplegic CPAP users allows for greater flexibility in interface choice as leakage is less likely at lower pressures . These practical challenges to the use of CPAP for OSA in tetraplegia and the uncertainty as to the real benefit of the therapy in SCI limit the ability of clinicians to provide evidence-based advice to their patients. The results of a current study investigating the effect of CPAP for OSA in acute tetraplegia will assist with informing future practice recommendations in this area .
Electrical and magnetic stimulation of breathing
A paralysed diaphragm can be stimulated via a phrenic nerve pacer or by direct motor point pacing of the diaphragm if the phrenic nerve is intact, and the cell bodies of C3, C4 and C5 at the spinal cord are viable [152, 153]. Full pre-operative assessment is required to ascertain a patient’s suitability for stimulation . Extensive post-operative training is necessary to increase diaphragmatic endurance, and to teach the patient, family and carers the necessary skills and understanding of the device. The conditioning period may be complicated by initial feelings of dyspnoea, particularly if the patient has previously been ventilated with large tidal volumes and associated low bicarbonate . For some patients, phrenic nerve pacing will provide a full-time alternative to the ventilator and the tracheostomy will no longer be required, but for others, the ventilator and tracheostomy will be required for additional hours during the day or night .
Benefits include greater wheelchair mobility, elimination of the fear of accidental ventilator disconnection, loss of social stigma associated with being attached to a ventilator, improved speech, no noise from the ventilator, reduced need for carer input, and improved well-being and overall health [155, 154, 156, 157]. Disadvantages include the need for major surgery involving thoracotomy, risk of surgical damage to the phrenic nerve and failure of the implanted device [154, 158].
Conflict of interest Disclosures can be found alongside this article at breathe.ersjournals.com
- ©ERS 2016
Breathe articles are open access and distributed under the terms of the Creative Commons Attribution Non-Commercial Licence 4.0.