Respiratory complications in the adult patient with chronic spinal cord injury

INTRODUCTION — The number of people in the United States living with spinal cord injury (SCI) was estimated to be about 294,000 (range of 250,000 to 368,000) in 2020 [1]. Rates of new injury are estimated to be 54 cases per million [2]. Causes of SCI include trauma (eg, vehicular crashes, falls, gunshot wounds) and non-traumatic causes that include tumors, vascular diseases, and spinal stenosis [1,3]. Respiratory failure is common after acute SCI and in persons with chronic SCI who develop respiratory illnesses. Respiratory complications and related diseases are the most common cause of death [4-9]. The spectrum of pulmonary complications following SCI includes respiratory failure, pneumonia, atelectasis, pulmonary thromboembolism, sleep disorders, dyspnea, and dysphonia.

This topic reviews the prevention and management of respiratory complications of SCI. The diagnosis and management of acute SCI, including the acute management of respiratory failure, the management of other common complications of SCI, and the changes in respiratory physiology that occur following SCI are discussed separately. (See "Acute traumatic spinal cord injury" and "Chronic complications of spinal cord injury and disease" and "Respiratory physiologic changes following spinal cord injury".)

EPIDEMIOLOGY — Pulmonary complications following SCI are common and contribute to morbidity, including rehospitalization. In a Canadian study assessing hospitalization in the year following discharge, 27.5 percent of persons with SCI were rehospitalized at least once, and 11.5 percent of the hospitalizations were attributed to a respiratory cause [10]. In the US, overall rates of hospitalization 5 to 20 years after injury were approximately 20 to 25 percent, with 8.1 percent attributable to diseases of the respiratory system [11].

After circulatory complications, pulmonary complications are the first or second most common underlying cause of death in the years following SCI, accounting for up to 20 to 30 percent of the mortality that occurs beyond the first year after SCI [12-16]. Pulmonary complications are most common immediately following acute SCI and decrease in the months to years after SCI.

Most studies find that the likelihood of pulmonary complications depends upon the level and completeness of SCI, with a greater risk in persons with higher and neurologically more complete SCI [17,18]. However, in a prospective cohort study of patients with SCI who had survived at least one year (mean injury duration of 17 years) and were not receiving chronic mechanical ventilation, neither the level nor completeness of SCI was not associated with an increased risk of future respiratory illness, hospitalization for cardiac or pulmonary causes, or decline in pulmonary function [19]. A potential explanation for these findings which are in contrast with other studies is that this study included mainly long-term survivors and thus excluded persons with the highest level and severity of injury who were more likely to have died earlier.

Additional risk factors for respiratory illness following SCI include a lower percent predicted forced expiratory volume in one second (FEV₁), smoking history, history of chronic obstructive pulmonary disease, and history of pneumonia or bronchitis since acute SCI (approximately two-fold) [12,19-22].

RESPIRATORY INSUFFICIENCY — At the time of acute traumatic SCI, acute transverse myelitis, or spinal stroke, flaccid paralysis occurs immediately, affecting all muscles caudal to the level of injury, potentially including respiratory muscles. The extent of ventilatory muscle impairment depends upon the level and severity of SCI; ventilatory impairment is most common and severe in complete cervical cord lesions, but also prevalent in thoracic cord lesions. (See "Respiratory physiologic changes following spinal cord injury", section on 'Impairment of ventilatory muscle function' and "Initial evaluation and management of blunt thoracic trauma in adults", section on 'Pulmonary injury'.)

Acute management — Respiratory failure due to phrenic and/or intercostal nerve injury with consequent respiratory muscle dysfunction is common in the minutes to days following acute SCI. In addition, injuries associated with blunt chest trauma and acute aspiration may contribute to respiratory insufficiency. Thus, many patients require intubation. The assessment of the level of SCI, associated ventilatory deficits, injuries associated with blunt chest trauma, the decision to intubate, and the timing of tracheostomy are discussed separately. (See "Acute traumatic spinal cord injury", section on 'Clinical presentation' and "Initial evaluation and management of chest wall trauma in adults" and "Acute traumatic spinal cord injury", section on 'In the emergency department' and "Acute traumatic spinal cord injury", section on 'Respiratory complications' and "Tracheostomy: Rationale, indications, and contraindications".)

Ventilator settings — The optimal ventilator settings for patients with chronic respiratory failure due to cervical or high thoracic SCI are not known. In the absence of data to the contrary, we follow the general guidelines for mechanical ventilation and adjust as needed based on underlying lung disease and degree of acute illness, dyspnea, and arterial blood gas results. When the tracheal tube cuff is deflated to facilitate speech, the tidal volume is typically increased to compensate for air leakage around the cuff. (See "Overview of initiating invasive mechanical ventilation in adults in the intensive care unit".)

A controversial area is whether to set the ventilator to deliver high or low tidal volumes to facilitate weaning after an episode of respiratory failure. One hypothesis is that the profound respiratory muscle weakness of cervical SCI leads to atelectasis even during mechanical ventilation and that this propensity is offset by use of high tidal volumes [23]. Thus, it has been suggested practice to ventilate persons with cervical or high thoracic SCI at tidal volumes of 10 to 15 mL/kg predicted body weight (PBW) or higher (table 1 and table 2) [7,23-25]. However, among case series that reported fewer respiratory complications with tidal volumes ≤15 to 20 mL/kg, an uncuffed tracheostomy tube was typically in place which would have resulted in a substantial air leak and lower effective tidal volumes [23,24,26]. Further study is needed to reconcile these observations with the large body of evidence in ventilated patients without SCI suggesting that lower tidal volumes (eg, 6 to 10 mL/kg PBW) reduce the risk of ventilator-associated lung injury. (See "Ventilator-induced lung injury".)

A limitation of assessing the efficacy of a particular approach has been a lack of sufficiently powered trials of weaning following acute SCI and in a chronic SCI setting. For example, in one trial, 118 subjects with SCI within the prior six months were screened and only 34 were randomized to tidal volumes of 10 mL/kg or 20 mL/kg in a two week weaning protocol [27]. There was no difference in median days to wean, ventilator associated pneumonia, or atelectasis. However, in two recent retrospective analyses of ventilation following acute SCI (in 181 patients [28] and 84 patients [29]), higher tidal volumes (defined as >10 mL/kg and 15 mL/kg, respectively) were associated with a greater risk of pneumonia. Therefore, in order to balance the goals of avoiding atelectasis and ventilator-associated lung injury in patients with SCI-associated respiratory failure but otherwise normal lungs, we typically aim for a tidal volume of 8 to 10 mL/kg PBW.

Some evidence suggests that tetraplegic patients may experience more dyspnea when ventilated at lower tidal volumes (eg, <10 mL/kg PBW), although this has only been assessed in a small number of patients [30]. Other factors such as inspiratory flow rate, mucus plugging, and gas exchange should be assessed and corrected as needed.

Positive end-expiratory pressure (PEEP) is a standard method to prevent atelectasis in patients with acute respiratory distress syndrome (ARDS), but the optimal PEEP in persons with respiratory failure due to SCI is not known [25,26]. We typically use 5 cm H₂O or less of PEEP, unless the patient has complicating ARDS. Excessive PEEP may cause flattening of the diaphragm and lower its position in the chest, resulting in a mechanical disadvantage during inspiration [6,7]. (See "Positive end-expiratory pressure (PEEP)", section on 'Indications'.)

Weaning from the ventilator — Improvements in pulmonary muscle function over the weeks following SCI are due to several factors. These include functional descent of the neurologic injury level as spinal cord inflammation resolves, enhanced recruitment of accessory ventilatory muscles, retraining of deconditioned muscles, and the evolution from flaccid paralysis (known as spinal shock) to spastic paralysis. As the intercostal and abdominal muscles become less flaccid, rib cage stabilization improves. The mean duration of mechanical ventilation in 186 patients with acute cervical SCI was three to four weeks in one series [31]. In another study of 222 patients with traumatic cervical injury requiring mechanical ventilation, 63 percent did not require mechanical ventilation at discharge from an acute care setting [32]. Thus, ongoing attempts at weaning from mechanical ventilation are appropriate and often successful in the weeks to months after SCI, except for patients with high cervical cord lesions. (See "Initial weaning strategy in mechanically ventilated adults" and "Weaning from mechanical ventilation: Readiness testing" and "Management of the difficult-to-wean adult patient in the intensive care unit".)

Protocols describing an approach to weaning the patient with cervical and high thoracic SCI (and stable respiratory status) suggest an approach of gradually increasing periods of ventilator-free breathing (VFB) [7,26,33]. The accuracy of weaning predictors (eg, rapid shallow breathing index) in predicting weaning success has not been specifically addressed in this patient population, and we typically follow standard clinical criteria for weaning, as described in the table (table 3). During the VFB trials, agitation, respiratory rate, heart rate, and pulse oxygen saturation are monitored; an increase in respiratory rate and a fall in tidal volume (eg, one-half to two-thirds of the starting value) are indicators of poor tolerance of VFB [24,26]. (See "Initial weaning strategy in mechanically ventilated adults" and "Weaning from mechanical ventilation: Readiness testing".)

Prevention of atelectasis is a key component of successful weaning. In addition to adjusting tidal volume as noted above, meticulous attention to chest physiotherapy is essential. In addition to routine times of chest physiotherapy (eg, manually assisted cough, mechanical insufflation-exsufflation, endotracheal suctioning), chest physiotherapy is used to clear the airways of secretions in preparation for each weaning effort. (See 'Chest physiotherapy' below.)

Patient position during weaning trials is also important. In motor complete cervical SCI, weaning tends to be more successful with the patient in a more recumbent position as the diaphragm is higher in the chest, due to pressure from the abdominal contents [34]. The higher the diaphragm is in the chest, the more effective its contraction. Forced vital capacity and forced expiratory volume in one second (FEV₁) are higher in persons with cervical SCI when supine [35].

Respiratory muscle training — Respiratory muscle training typically involves inspiration or expiration through a device with a narrow orifice or against resistance once or twice a day, five to six days a week. Inspiratory muscle training (IMT) has been evaluated as an adjunct to weaning from mechanical ventilation [6,7,36,37]. Systematic reviews of IMT in mechanically ventilated patients without SCI have found improved respiratory muscle strength but insufficient evidence to determine whether the improvement in muscle strength leads to a shorter duration of mechanical ventilation [38]. In patients with SCI, systematic reviews have similarly found an improvement in respiratory muscle strength, as assessed by improved peak maximal inspiratory and expiratory pressures and vital capacity, but data are insufficient to conclude a benefit in weaning [36,37,39]. A meta-analysis of six trials reported an increase in maximal inspiratory pressure but no effect on measures of quality of life, maximal expiratory pressure, or FEV₁ [40].

Given the lack of adverse effects and the theoretical advantages of improved strength, one center describes using IMT once patients are able to tolerate a few minutes without ventilatory support. IMT may be initiated with an inspiratory resistance of 7 to 10 cm H₂O for up to one minute initially, twice daily [6,7]. The level of resistance, duration, and frequency of treatment are increased as strength improves. To maintain any improvement in strength, it is necessary to continue training, as with any muscle strengthening program.

Further research is needed to clarify whether inspiratory muscle training improves weaning outcomes.

Long-term assisted ventilation — Certain categories of patients with SCI will require ongoing assistance with ventilation.

Cervical cord injury with bulbar muscle weakness — Some patients with cervical SCI are unable to wean from assisted ventilation, particularly those with high cervical cord injury that affects the lower cranial nerves at the craniocervical junction. These high cervical cord injuries impair bulbar function (causing weakness of pharyngeal and palatal muscles) and cause recurrent aspiration of saliva and inability to protect the airway. In addition, these patients are unable to maintain a pulse oxygen saturation of 95 percent or greater despite optimal use of noninvasive ventilation and mechanically assisted cough because of ongoing saliva aspiration [41]. Thus, these patients require long-term tracheostomy and mechanical ventilation. (See "Management and prognosis of patients requiring prolonged mechanical ventilation".)

Cervical cord injury with intact bulbar function — Among patients with cervical SCI, who have good bulbar function, but are unable to breathe independently, most are able to wean to a combination of noninvasive ventilation and mechanically assisted cough. We generally start with continuous noninvasive positive pressure ventilation (NPPV) and then use the other techniques, such as glossopharyngeal breathing, abdominal respirators, and diaphragmatic pacing, to allow patients more time independent from the NPPV machine [41,42]:

●Noninvasive positive pressure ventilation – NPPV can be used to provide full and continuous ventilatory support [41]. At night, NPPV can be delivered via nasal, oronasal, or lip-seal interfaces with a conventional ventilator. During the day, some patients are able to breathe independently. Otherwise, a 15-mm angled mouth piece may be used as an interface with the ventilator to better enable speech. The angled mouthpiece can be kept in the mouth or fixed near the mouth for frequent, but intermittent use. Alternatively, patients can use a nasal interface during the day and prevent excessive oral leakage by closing their mouth or pressing their tongue against the soft palate. (See "Noninvasive ventilation in adults with chronic respiratory failure from neuromuscular and chest wall diseases: Patient selection and alternative modes of ventilatory support" and "Noninvasive ventilation in adults with chronic respiratory failure from neuromuscular and chest wall diseases: Practical aspects of initiation" and "Noninvasive ventilatory support and mechanical insufflation-exsufflation for patients with respiratory muscle dysfunction".)

●Glossopharyngeal breathing – Intermittent glossopharyngeal breathing in high cervical level patients utilizes oral, pharyngeal, and laryngeal muscles to enhance ventilation. The technique involves the use of the glottis to add to an inspiratory effort by gulping boluses of air into the lung. The technique can be used by patients with almost no spontaneous vital capacity to take a deep inhalation for coughing or supporting voice volume, or it can be used for VFB time by patients who otherwise cannot tolerate disconnection [43,44].

●Abdominal respirators – Abdominal respirators, such as the intermittent abdominal pressure ventilator (IAPV; exsufflation belt) depend on the motion of abdominal viscera to augment diaphragmatic excursion in patients with diaphragmatic paralysis [41]. The IAPV requires that the patient sit at an angle of at least 30° from horizontal and may need to be augmented by glossopharyngeal breathing [41]. Some patients prefer IAPV to NPPV for daytime support [45]. (See "Noninvasive ventilation in adults with chronic respiratory failure from neuromuscular and chest wall diseases: Patient selection and alternative modes of ventilatory support", section on 'Abdominal respirators'.)

Other negative pressure ventilatory devices, such as the tank, cuirass, or poncho ventilators are cumbersome and may result in upper airway obstruction during sleep [42,46]. (See "Noninvasive ventilation in adults with chronic respiratory failure from neuromuscular and chest wall diseases: Patient selection and alternative modes of ventilatory support" and "Noninvasive ventilation in adults with chronic respiratory failure from neuromuscular and chest wall diseases: Practical aspects of initiation" and "Noninvasive ventilatory support and mechanical insufflation-exsufflation for patients with respiratory muscle dysfunction".)

●Role of an abdominal binder – In tetraplegic SCI, abdominal binders are sometimes used while the patient is in the seated position in an attempt to improve respiratory mechanics by mimicking the non-functioning abdominal muscles. By compressing the abdominal contents to increase intra-abdominal pressure, the diaphragm is elevated into a more optimal position for inspiration. A meta-analysis indicated that abdominal binders improved vital capacity by an average of 0.32 L (95% CI = 0.09, 0.55), but decreased functional residual capacity by 0.41 L (0.14, 0.67), thereby potentially promoting atelectasis [47]. The use of binders is individualized, as there is insufficient evidence to either support or discourage their use in this patient population. The binder should not overlap the rib cage or pelvis [7].

●Diaphragmatic pacing – Patients with complete SCI rostral to C3 (ie, above the takeoff level of the phrenic nerve) often have diaphragmatic paralysis in addition to intercostal and abdominal wall muscle paralysis [48,49]. For patients with SCI-associated diaphragmatic paralysis, diaphragmatic pacing may allow weaning from mechanical ventilation [50-54]. (See "Pacing the diaphragm: Patient selection, evaluation, implantation, and complications".)

Most patients prefer phrenic pacemaker systems to chronic positive pressure ventilation because pacing systems are small and can be kept out of view; breathing through the mouth feels more natural; and taste and smell are enhanced by breathing through the mouth and nose [50,55,56]. Furthermore, patients with diaphragmatic pacers have speech that is more continuous and contains fewer and shorter pauses compared with patients on positive-pressure ventilation [57]. (See 'Phonation' below.)

Diaphragm pacing is not synchronous with activation of the oropharyngeal muscles that prevent airway collapse during breathing while asleep. Therefore, we generally maintain a tracheostomy in patients with diaphragmatic pacers in order to avoid upper airway obstruction during sleep, facilitate suctioning, and assist in management in case of pacemaker failure.

In addition, concurrent use of intercostal nerve pacing may allow the use of diaphragmatic pacing in a greater number of candidates, although this is still experimental [48,58,59]. (See "Pacing the diaphragm: Patient selection, evaluation, implantation, and complications".)

VENOUS THROMBOEMBOLISM

Risk — Patients who have suffered an acute SCI are among the highest risk groups for developing venous thromboembolism (VTE) with a risk that remains above that of the general population for at least one year following injury [60-63]. The greatest risk for VTE is between 72 hours and two weeks following the injury; however, the risk remains high for the first three months and thereafter drops substantially:

●In 222 patients with acute SCI admitted to an acute SCI and rehabilitation service, 47 (21.2 percent) developed VTE; 90 percent of the deep vein thrombosis (DVT) cases and all of the pulmonary embolism cases occurred the first three months [64].

●In 94 patients followed prospectively for a median of three years, the risk was 34.4 VTE events/100 patient-years in the first three months and 0.3 VTE events/100 patient-years thereafter [60].

●In a retrospective assessment of 16,240 acute SCI cases from 1991 through 2001, the cumulative incidence of VTE was 4.8 percent during the initial hospital admission, with an additional 0.6 percent through 91 days, and 0.6 percent through one year after SCI [61].

●In a comparison of 47,916 SCI patients and 191,664 controls, the VTE risk in acute SCI patients was greatest in the first three months following injury (hazard ratio [HR] 17) and subsequently fell thereafter (HR<2 for up to eight years) [62].

●In a study of 424 patients with SCI for one year or more who underwent plastic or reconstructive surgery lasting 60 minutes or more between 2004 and 2014, the 90-day rate of VTE was 0.2 percent compared with a rate of 1.7 percent in 777 general surgical patients who underwent a variety of procedures and a mix of pharmacologic and mechanical VTE prophylaxis. Only 7.8 percent of the SCI group had pharmacologic prophylaxis (9 percent mechanical). This suggests that the risk of VTE is generally low in chronic SCI patients undergoing plastic surgery compared with general surgical patients [65].

●Using a discharge database (1995 through 2010), the incidence of VTE was highest during the first three months (10 percent); rates subsequently fell to 1 percent at six months, 0.64 percent at nine months, and 0.35 percent at one year [66]. Between one to three years after SCI, rates varied but were about 0.31 percent, which compared favorably with estimated VTE rates of the general population (0.1 to 0.2 percent). These data indicate that the long-term risk of VTE during the chronic phase of SCI remains ill-defined, but suggests that risk may remain marginally elevated even after one year compared with general population rates.

Prophylaxis — The approach to VTE prophylaxis in patients with SCI varies among centers and is dependent upon time elapsed from the original SCI event as well as the risk of bleeding, degree of mobility, and hospital policy. The precise time periods that demarcate the acute, post-acute, and chronic phases of SCI are not rigorously defined. However, for the purposes of VTE prophylaxis, it is useful to consider the following phases and an approach similar to that outlined by the 2016 Consortium for Spinal Cord Medicine guidelines as well as the Neurocritical Care Society and the Society of Critical Care Medicine [67,68]:

●Acute hospitalization for SCI – Patients with acute SCI (eg, following injury and/or through hospital discharge) are patients with the highest risk of VTE such that pharmacologic prophylaxis with subcutaneous low molecular weight (LMW) heparin should be administered provided there is no contraindication [67,69-71]. Subcutaneous unfractionated heparin (UFH) is an alternative if LMW heparin cannot be used. Full anticoagulation with oral anticoagulants (warfarin or direct oral anticoagulants [DOACs]) is not suggested during the acute phase due to bleeding risk [67,69]. Mechanical methods of prophylaxis including intermittent pneumatic compression (IPC) and graduated compression stockings (GPS) are often used by many experts as an adjunct to pharmacologic prophylaxis [67,69,70].

For those with a contraindication to anticoagulant use, mechanical methods of prophylaxis (IPC, GPS) are indicated; pharmacologic prophylaxis should be promptly administered once the increased bleeding risk is resolved. Observational data suggest that LMW heparin may be given safely within 48 hours of surgical spine fixation [72]. VTE prophylaxis in acutely ill hospitalized patients is discussed separately. (See "Prevention of venous thromboembolic disease in adult nonorthopedic surgical patients" and "Prevention of venous thromboembolic disease in acutely ill hospitalized medical adults".)

●Post-acute hospitalization SCI (ie, typically during inpatient rehabilitation) – In this population, the optimal duration and method of VTE prophylaxis is unknown and depends upon factors including mobility, presence of other VTE risk factors, intercurrent acute illness, and bleeding risk. VTE prophylaxis should be individualized with a low threshold to continue VTE prevention measures for a minimum of eight weeks; longer durations are appropriate, (eg, 12 weeks) in patients with persistent immobility or with other risk factors [67,69]. We typically use pharmacologic prophylaxis with LMW heparin. An alternative suggested regimen includes therapeutic dose warfarin (Internationalized Normalized Ratio 2 to 3). Although the DOACs are also suggested alternatives to warfarin post-acute hospitalization [67,73], there have been no clinical trials in patients with SCI using these agents.

●Post-rehabilitation SCI – Prophylaxis is typically not continued beyond 8 to 12 weeks in most patients, regardless of their ambulatory status. If a patient requires rehospitalization, VTE prophylaxis is managed according to the risk of VTE in acutely hospitalized patients [67]. (See "Acute traumatic spinal cord injury", section on 'Venous thromboembolism and pulmonary embolism' and "Prevention of venous thromboembolic disease in adult nonorthopedic surgical patients".)

Inferior vena cava filters should not be routinely placed nor should surveillance with compressive ultrasound be performed for the early identification of DVT.

The rationale for this approach is based on the known high risk of VTE in this population and systematic reviews of small observational studies and small randomized trials that suggest LMW heparin is superior to UFH for the prevention of VTE following injury and through the rehabilitation period:

●In one 2008 systematic review of nine studies, compared with UFH, LMW heparin was associated with a reduction in rates of pulmonary embolism (PE; 9 versus 3 percent) [74]. There was no difference in the rate of fatal PE, DVT, or bleeding. However, significant heterogeneity and methodologic flaws prohibit firm conclusions from this review.

●In a randomized study of 107 patients with SCI that compared thrice daily UFH plus IPC to enoxaparin 30 mg twice daily alone, there was no difference in the overall rate of VTE or bleeding [75]. However, there appeared to be a reduction in the rate of symptomatic PE (5 versus 18 percent).

●In a six week extension study of the above randomized trial, LMW heparin was associated with a reduction in the rate of VTE (9 versus 22 percent) without any difference in the bleeding rate [76].

●Other observational studies report no differences in rates of VTE or bleeding between various LMW heparin agents [77,78] and based on a meta-analysis [79].

PULMONARY INFECTION — The risk of pneumonia is highest in the first year following SCI, but patients remain at increased risk over their lifetime [17]. In a prospective series, pneumonia complicated acute SCI in approximately 30 percent of patients during a mean follow-up of 9.5 months [80]. In contrast, among outpatients with chronic SCI, rates of pneumonia were approximately 3 percent per year [81,82]. Interventions, such as chest physiotherapy and vaccination, can help reduce the risk of pneumonia following SCI.

Risk factors — A number of consequences of SCI contribute to the high risk of pneumonia. Whether these factors are present in a given patient depends on the neurologic level and completeness of injury. Risk factors for pneumonia include [83-92]:

●Expiratory muscle weakness, which results in an ineffective cough

●Altered levels of consciousness from concomitant head trauma or sedating medications

●Ileus, which limits diaphragmatic excursion and increases the risk of aspiration of gastric contents

●Failure of spontaneous sighing

●Bronchial mucus hypersecretion, which occurs in approximately 20 percent of acute cervical level SCI patients, possibly from impairment of the peripheral sympathetic nervous system

●Associated rib fractures or thoracoabdominal surgery

●Dysphagia and aspiration

Clinical presentation and diagnosis — The clinical presentation of pneumonia in patients with SCI includes the typical features of cough, increased respiratory secretions, dyspnea, tachypnea, and fever. In addition, patients with SCI at higher neurologic levels and greater completeness of injury, may exhibit a rapid progression to respiratory failure because of an inability to clear secretions adequately or to increase ventilation to compensate for worsened gas exchange and increased carbon dioxide production. Infections in patients with SCI may also be accompanied by autonomic dysfunction. (See "Chronic complications of spinal cord injury and disease", section on 'Autonomic dysreflexia'.)

Chest radiographs are frequently obtained in the supine position in patients with SCI and may be inadequate to define the true extent of parenchymal consolidation, atelectasis, or pleural effusion. Thus, clinicians should have a low threshold for obtaining computed tomography (CT) imaging when pneumonia is suspected [93].

While pneumonia is the most frequent cause of late ventilatory failure, other complications should also be considered including secretion retention, new onset cardiopulmonary disease (chronic obstructive pulmonary disease, heart failure), subdiaphragmatic problems that impair motion of the diaphragm, restrictive pulmonary disease secondary to progressive kyphoscoliosis, post-traumatic syringomyelia, cervical spondylostenosis with progressive myelopathy, and obesity. In this setting, pulmonary function tests, arterial blood gas evaluation, and a chest radiograph can help determine whether the problem is an intrinsic lung problem or deterioration in respiratory muscle strength. (See "Respiratory physiologic changes following spinal cord injury" and "Tests of respiratory muscle strength".)

Management — The potential for resistant organisms being the cause of pneumonia is increased if the patient is currently or recently hospitalized, so attempts to establish a bacteriologic diagnosis with sputum gram stain and culture are warranted. Invasive sampling techniques may be necessary if the patient's cough is weak and does not produce sputum for examination or the patient does not respond promptly to initial empiric antibiotics. (See "Aspiration pneumonia in adults" and "Epidemiology, pathogenesis, microbiology, and diagnosis of hospital-acquired and ventilator-associated pneumonia in adults" and "Community-acquired pneumonia in adults: Assessing severity and determining the appropriate site of care" and "Clinical evaluation and diagnostic testing for community-acquired pneumonia in adults".)

The treatment of aspiration, hospital–acquired, and community-acquired pneumonia is discussed separately. (See "Treatment of hospital-acquired and ventilator-associated pneumonia in adults" and "Treatment of community-acquired pneumonia in adults who require hospitalization" and "Aspiration pneumonia in adults", section on 'Treatment'.)

Clearance of respiratory secretions during pneumonia can be a major problem for patients with SCI. Close attention to pulmonary hygiene is essential. (See 'Chest physiotherapy' below.)

For patients with pneumonia who have lobar atelectasis or impaired gas exchange that does not respond to chest physiotherapy, we perform therapeutic bronchoscopy. We find it helpful to perform intermittent upper abdominal thrusts ("quad coughs") on the patient during the procedure to assist with mobilization of pulmonary secretions. (See "Flexible bronchoscopy in adults: Indications and contraindications", section on 'Therapeutic indications'.)

Prevention

Chest physiotherapy — Based on limited data, and not on well-designed randomized trials, chest physiotherapy appears to benefit patients with cervical or thoracic SCI by decreasing the risk of mucus retention and atelectasis [94]. A decrease in the frequency with which ventilatory support is required and improvement in survival following acute SCI have also been noted [84]. Based on the patient's ability to cough and clear secretions, requirement for chest physiotherapy may be life long, particularly during respiratory infections. We treat all patients with either cervical or thoracic SCI with chest physiotherapy, including [6,17,86,94,95]:

●Encouragement of deep breathing, frequent changes of position, postural drainage of secretions [7].

●Manually assisted coughing, also called a "quad cough," uses forceful upper abdominal thrusts in a posterior and cephalad direction timed to coincide with exsufflation or a voluntary cough effort. Preceding the cough with deep inspiration or glossopharyngeal breathing (repetitive air gulping) improves the success of this maneuver. Contraindications to performing quad coughs are the presence of an inferior vena cava filter or abdominal aortic aneurysm. (See "Respiratory muscle weakness due to neuromuscular disease: Management", section on 'Manual-assisted cough (abdominal thrust)'.)

●We use mechanical insufflation-exsufflation as an effective adjunct to chest physiotherapy. The mechanical insufflator-exsufflator (MIE) is a device that delivers a deep breath via a mask or while directly attached to an endotracheal or tracheostomy tube and assists with exhalation by "sucking" the air out and generating a negative pressure [96-98]. Insufflations at pressures of 40 to 60 cm H₂O are followed immediately by exsufflations to -40 to -60 cm H₂O [41,96]. This may be accompanied by manually assisted coughing. The use of this device is common in SCI units and is frequently provided for home use [97]. We also use this device in our intensive care unit for patients with SCI to assist with mucus clearance as required clinically, particularly during weaning after acute pulmonary infections. Although potential complications of insufflation-exsufflation include abdominal distention, adverse effects are rarely reported. The measurement of cough velocity (<5 L/sec) has been suggested to screen for persons who may benefit from the MIE for chronic use, but usually the need is determined based on clinical behavior [41]. In our experience, frequent use of mechanical insufflation-exsufflation has resulted in avoiding bronchoscopy to remove secretions, and if needed, hourly use is described as an adjunct during weaning [99]. We also provide devices for patients to use at home during periods of increased secretions. (See "Respiratory muscle weakness due to neuromuscular disease: Management", section on 'Mechanical insufflation-exsufflation'.)

●For patients who are intubated or have undergone tracheostomy, thorough suctioning via the tracheal tube is appropriate. Nasotracheal or orotracheal suctioning is sometimes used in non-intubated patients to facilitate secretion removal, but is uncomfortable for the patient, may cause laryngeal trauma, and misses the left mainstem bronchus 90 percent of the time [100].

Other chest physiotherapy modalities may also be of benefit, but there are fewer data to support their use. Many clinicians use beta-agonist bronchodilator therapy to enhance mucociliary clearance, although there are no data demonstrating clinical efficacy. Expiratory muscles innervated by segments C5 through C7 can be trained in SCI patients [36,101,102], but it is not known if such training improves cough sufficiently to reduce mucus retention and prevent complications. Intermittent positive pressure breathing does not appear to be beneficial [103,104].

Electrical stimulation devices are under investigation as a means to improve the cough of SCI patients. The safety and clinical utility of these electrical stimulation techniques require further investigation, but the techniques have the potential to decrease the incidence and severity of pneumonia in patients with SCI. In a clinical trial, an electrical stimulation system was surgically implanted in nine subjects with complete cervical SCI to activate the expiratory muscles by stimulation at the T9, T11, and L1 spinal levels [105]. Subjects were instructed to use the cough system every 30 seconds for 5 to 10 minutes, two to three times a day on a chronic basis, and also as needed for secretion management. The incidence of acute respiratory tract infections fell from 2.0 +/- 0.5 to 0.7+/-0.4 events/subject year [105]. Augmentation of cough is also possible using external posterolateral electrodes [106] and has been described following functional electrical stimulation over the abdominal muscles [107]. Implantation of an electrical cough stimulation system in conjunction with a diaphragm pacing system improved cough and expectoration in one case series [108].

Vaccination — Vaccination against influenza and pneumococcus may help reduce the risk of pneumonia following SCI. The ability to respond to the pneumococcal and influenza vaccines is not impaired [109-111].

●Published guidelines for influenza vaccination note that persons with SCI are a high risk group and vaccination is recommended [112]. (See "Seasonal influenza vaccination in adults".)

●Although the pneumococcal vaccine guidelines do not include specific reference to individuals with SCI, as SCI is considered a chronic medical condition, it is recommended that all such patients be vaccinated [113,114]. (See "Pneumococcal vaccination in adults", section on 'Approach to vaccination'.)

PULMONARY EDEMA — Neurogenic pulmonary edema (NPE) is a rare complication of SCI. In acute SCI, it is generally associated with complete cervical cord injuries [115-118]. NPE can also develop in patients with chronic SCI above T6 as a complication of autonomic dysreflexia [115,118], an episodic syndrome characterized by profound systemic hypertension, tachycardia or bradycardia, headaches, flushing, diaphoresis above the SCI level, and pupillary changes [119,120]. (See "Chronic complications of spinal cord injury and disease", section on 'Autonomic dysreflexia'.)

NPE typically presents with dyspnea, tachypnea, tachycardia, and basilar crackles, although radiographic evidence of NPE may be present in the absence of clinical findings. The chest radiograph typically shows a normal heart size with septal thickening and patchy or perihilar ground glass opacities. The diagnosis and management of NPE are discussed separately. (See "Neurogenic pulmonary edema".)

DYSPNEA WITH DAILY ACTIVITIES — Among persons with chronic SCI, breathlessness is a common complaint and is largely related to the level of SCI [121].

Patients who develop acute worsening of dyspnea should be evaluated for acute pneumonia, thromboembolic disease, secretion retention, obstructive airways disease, and potential cardiovascular causes of dyspnea. (See 'Pulmonary infection' above and 'Clinical presentation and diagnosis' above.)

Contributing factors — Breathlessness among persons with chronic SCI is largely related to the level of SCI [121]. In one study, a respiratory health questionnaire was used to assess the relationship of degree of injury and dyspnea [122]. Among 130 adult males with neurologically complete motor SCI who used a hand-propelled wheelchair greater than 50 percent of the time, those with tetraplegia reported breathlessness more frequently (21 to 33 percent) than those with high thoracic (9 to 15 percent) or lower injury levels (2 to 11 percent). The report of dyspnea was independent of obesity, smoking, age, and years since SCI.

However, the prevalence of dyspnea does not always correlate with the level of SCI, suggesting that the perception of dyspnea may be abnormal in tetraplegia [6]. This is illustrated by the observation that similar percentages (10 to 15 percent) of subjects with cervical SCI, high paraplegia, and low paraplegia reported the need to "stop to catch their breath" [121].

Dyspnea can limit activities of daily life to a variable extent. In a study of 441 motorized wheelchair users with chronic SCI, dyspnea was associated with talking (18 percent), eating (5 percent), dressing/undressing (2 percent), or leaving the house (4 percent) [123]. In this group, the risk of breathlessness was not attributable to lower pulmonary function (eg, forced vital capacity, forced expiratory volume in one second [FEV₁]). It is thought that the high rate of breathlessness during talking may be related to difficulty interrupting breathing to manipulate phrasing, volume, rate, and intensity of speech. In a separate study, dressing was associated with dyspnea in 12 percent of those with SCI [121]. Dyspnea while talking, eating, or dressing was also associated with lower health related quality of life [124].

Management — Physical activity and exercise may ameliorate dyspnea [125]. Among self-reported wheelchair athletes in the VA Boston study, the prevalence of breathlessness was 8 of 49 (16 percent) compared with 48 of 134 (36 percent) non-athletes with SCI [125]. Adjusting for smoking, neurologic level, and history of obstructive lung disease, non-athletes were 2.3 times more likely to report breathlessness than athletes, although this finding was of borderline significance. This relationship persisted when adjusted for percent predicted FEV₁ and maximal expiratory and inspiratory pressures. In a study of 347 individuals able to walk or use a manual wheelchair who were recruited from five SCI centers in the United States, adjusting for chronic obstructive pulmonary disease/asthma history and mobility mode, participation in a planned exercise program was also associated with less dyspnea [126]. These findings suggest an effect of exercise on reducing report of dyspnea that is not attributable to be improvement in respiratory muscle performance, pulmonary function, or underlying pulmonary disease.

A systematic review of inspiratory and expiratory respiratory muscle training in cervical SCI found improved pulmonary function, but no improvement in dyspnea [37]. Further research is needed to determine the role of respiratory muscle training in SCI. (See 'Respiratory muscle training' above.)

SLEEP-DISORDERED BREATHING — A number of reports indicate that patients with SCI are at increased risk of sleep apnea, predominantly of the obstructive or mixed types [127-135]. A prevalence of approximately 40 to 80 percent has been reported in patients with cervical SCI, with a lower prevalence (15 percent) in non-obese subjects [131]. These rates are greater than reported in men and women from the general population [136]. However the severity of injury likely does not explain all the impact on risk. (See "Clinical presentation and diagnosis of obstructive sleep apnea in adults".)

Possible mechanisms for the increased prevalence of sleep apnea include obstruction produced by hypertrophy of the neck musculature, ventilatory muscle spasticity, use of sedative antispasmodic medications, obesity, or an effect of SCI upon an undefined spinal cord pathway involved with control of sleep [137]. Patients with SCI also spend more time supine while sleeping than the general population.

Similar to patients with neuromuscular disorders, patients with severe ventilatory muscle weakness due to SCI are at risk for hypoventilation during rapid eye movement (REM) sleep and may develop hypoxemia [132,138,139]. This may be due to a combination of worsened ventilation-perfusion matching, upper airway narrowing, and REM-related inhibition of inspiratory muscle contraction. (See "The effect of sleep in patients with neuromuscular and chest wall disorders".)

Symptoms of sleep disordered breathing include excessive daytime sleepiness; disrupted nocturnal sleep; morning heading; changes in personality, concentration, or memory; and snoring. These, along with other laboratory findings, are indications for polysomnography in patients with SCI (table 4). Patients with SCI diagnosed with sleep apnea may benefit from specific interventions, depending on the severity. (See "Evaluation of sleep-disordered breathing in adult patients with neuromuscular and chest wall disorders" and "Obstructive sleep apnea: Overview of management in adults".)

PHONATION — Spontaneously breathing patients with cervical cord injury and expiratory muscle weakness can suffer speech abnormalities, such as reduced vocal amplitude and a lower number of syllables per breath compared with able-bodied individuals [140]. Many patients compensate for expiratory muscle weakness by inhaling maximally (sometimes with use of glossopharyngeal breaths) to initiate speech at the highest possible lung volume. This pattern of ventilation is probably chosen to optimize the recoil pressure of the respiratory system and produce maximal flow across the vocal cords.

In mechanically ventilated patients, a fenestrated, cuffed tracheostomy tube should be used whenever possible to allow voice communication. While the use of a fenestrated tube and/or deflation of the cuff can result in a significant loss of tidal volume through the mouth, the inspired tidal volume can be increased in order to compensate. Patients can minimize the tidal volume leak by controlling glottal patency and by using accessory muscles of ventilation. However, because these compensatory mechanisms are lost during sleep, the inner cannula should be inserted (thereby occluding the fenestration) and the cuff inflated during sleep. We prefer to make adjustments in ventilator delivered breaths, rather than using a "speaking valve", a one-way inspiratory valve inserted between the tracheostomy tube and the ventilator. Speaking valves carry a high risk of asphyxiation if the tracheal tube cuff is not deflated prior to insertion of the valve.

Patients can phonate during most of ventilator-driven inhalation and the early part of exhalation when a fenestrated tracheostomy tube is in place [57]. The change to speaking during inhalation rather than during exhalation takes some practice. The time available for speech during each breath is often short and accompanied by a long intervening period of silence, which makes it difficult to partake in activities such as group conversation or using a telephone. The fraction of each breath available for phonation can be increased by decreasing the inspiratory flow rate and adding positive end-expiratory pressure; these interventions prolong the period during which tracheal pressures exceed supraglottic pressures and produce usable airflow over the vocal cords [6,141,142]. Such minor ventilator manipulations improve speech timing, stabilize volume, and result in a more pleasant speaking voice.

Use of an abdominal binder and glossopharyngeal breathing may also improve speech quality [6]. (See 'Cervical cord injury with intact bulbar function' above and 'Management' above.)

AIR TRAVEL — Care should be taken when transporting patients with SCI by air. Patients with respiratory compromise at sea level may require supplemental oxygen to avoid the hypoxemia that can develop at altitude. (See "Evaluation of patients for supplemental oxygen during air travel".)

Mucus plugging can be particularly problematic if the patient breathes dry air cycled from outside of the aircraft; inspired air should be humidified when possible [143]. Suctioning equipment should be readily accessible during transport.

SUMMARY AND RECOMMENDATIONS

●Epidemiology – Pulmonary complications of spinal cord injury (SCI) are common, particularly during the period immediately following acute SCI, and are the most common cause of death. (See 'Epidemiology' above.)

●Respiratory insufficiency – Our approach is the following:

•Patients with high cervical cord tetraplegia often require long-term ventilatory support depending on the neurologic level and completeness of injury. Mechanical ventilation is usually continued via a tracheostomy for patients with impaired bulbar function.

•Patients with lower cervical cord tetraplegia may need ventilatory support for a few days to weeks but can eventually wean from mechanical ventilation to independent breathing or noninvasive positive pressure ventilation, as long as bulbar function is intact. (See 'Respiratory insufficiency' above.)

•For patients who are unable to wean from ventilatory support, potential alternatives include non-invasive ventilation and phrenic nerve or diaphragmatic pacing. (See 'Respiratory insufficiency' above.)

●General respiratory care – Patients with SCI are at increased risk for developing pneumonia, particularly in the first year following SCI. We believe that chest physiotherapy is standard care for the prevention of atelectasis and pneumonia in patients with cervical or thoracic SCI. Chest physiotherapy typically includes encouragement of deep breathing, manually assisted coughing ("quad cough"), and frequent use of a mechanical insufflator-exsufflator, including at home. We recommend vaccination against influenza and pneumococcus for all patients with respiratory impairment due to SCI (Grade 1B). (See 'Pulmonary infection' above and "Seasonal influenza vaccination in adults" and "Pneumococcal vaccination in adults".)

●Venous thromboembolism – For patients with SCI following acute injury and through inpatient rehabilitation who are not at increased risk of bleeding, we recommend pharmacologic thromboprophylaxis rather than mechanical methods or no prophylaxis (Grade 1B). We prefer use of low molecular weight heparin. For those in whom there is a contraindication to anticoagulant prophylaxis, we suggest mechanical methods including intermittent pneumatic devices and/or graduated compression stockings with prompt transition to pharmacologic prophylaxis once the contraindication has resolved. We individualize the duration of therapy with a view to longer durations (8 to 12 weeks) particularly in those whose mobility is limited or in whom additional risk factors are present. Following rehabilitation, most patients do not need venous thromboembolism prophylaxis unless they are rehospitalized for an acute illness. (See 'Venous thromboembolism' above and "Prevention of venous thromboembolic disease in acutely ill hospitalized medical adults".)

●Dyspnea – Among patients with chronic SCI, those with tetraplegia due to high cervical SCI report breathlessness with activity more frequently than do those with low paraplegia (45 versus 25 percent). (See 'Dyspnea with daily activities' above.)

●Sleep-disordered breathing – Patients with SCI are at increased risk of sleep apnea, predominantly of the obstructive or mixed types. Sleep apnea is more common in patients who are obese and those with a higher level of cord injury, but other risk factors may be more important. Patients with severe ventilatory muscle weakness are at risk to develop hypoxemia due to hypoventilation during rapid eye movement (REM) sleep even in the absence of obstructive or central causes of sleep apnea. (See 'Sleep-disordered breathing' above.)

●Phonation – Spontaneously breathing patients with cervical cord injury and expiratory muscle weakness can suffer speech abnormalities, such as reduced vocal amplitude or a lower number of syllables per breath. Strategies to improve speech include inhaling maximally and then terminating speech at higher than usual lung volumes. For patients who are mechanically ventilated, management includes use of a fenestrated tracheostomy tube during the day, decreasing the inspiratory flow rate to prolong inhalation, and adding positive end-expiratory pressure. (See 'Phonation' above.)