Surgical treatment of phrenic nerve injury

INTRODUCTION — Injury to one phrenic nerve leads to paralysis of the ipsilateral diaphragm, often leading to symptoms of dyspnea, which may improve with time. If both phrenic nerves are injured, both diaphragms are affected. If the injury is complete, this usually results in ventilator dependency.

The incidence of injury to the neuromuscular pathways that control the diaphragm is difficult to determine, but it can be estimated based on some of the more common iatrogenic and traumatic etiologies of phrenic nerve injury. As examples, phrenic nerve injury is reported to occur in up to 10 percent of those undergoing cardiac procedures [1-3]. Furthermore, among the roughly 6000 new cervical spinal cord injuries per year in the US, up to 75 percent have diaphragmatic paralysis leading to chronic ventilatory support, which increases morbidity and mortality [4,5]. The number of patients affected who do not have a clearly identifiable etiology may far surpass those with known traumatic or iatrogenic injuries, many of whom likely have either a viral insult (eg, Parsonage-Turner syndrome or neuralgic amyotrophy) or, more commonly, a chronic cervical radiculopathy [5,6].

The surgical management of phrenic nerve injury, including anatomy of the neuromuscular pathways supporting diaphragm function, etiologies leading to injury, and phrenic nerve reconstruction, is reviewed here. An overview of unilateral and bilateral diaphragmatic paralysis and nonsurgical treatments is reviewed separately. (See "Diagnosis and management of nontraumatic unilateral diaphragmatic paralysis (complete or partial) in adults" and "Diagnostic evaluation of adults with bilateral diaphragm paralysis" and "Treatment of bilateral diaphragmatic paralysis in adults".)

ANATOMY OF RESPIRATORY FUNCTION

Neuromuscular pathways — The neuromuscular pathways that control respiratory function transmit signals that originate from voluntary or involuntary motor control centers in the brain. A baseline level of involuntary control, which originates from brainstem nuclei, is necessary to activate breathing and initiate an inspiratory effort during sleep. Respiratory centers in the cerebral cortex may also provide a conscious override to stimulate a respiratory effort. There are established connections between both sides of the brainstem, including a described "crossed phrenic phenomenon" whereby a spinal cord hemisection that disrupts ipsilateral respiratory activity will be restored through a rerouting of impulses from the contralateral, uninjured side [7-10].

The impulses from the central nervous system descend through the anterior horn cells in the spinal cord, emerging into the peripheral nervous system through the upper cervical roots (C3-5) and contributing to the formation of phrenic nerves (figure 1). The majority of axonal tracts to the phrenic nerves are provided by the fourth cervical root (C4) with additional contributions from the third and fifth cervical roots [11]. There is some anatomic variability in the relationship between the cervical roots and phrenic nerves, including one such variation described in a pediatric patient with Klippel-Feil syndrome (characterized by cervical vertebral fusion) who presented with diaphragmatic paralysis from neck trauma; the phrenic nerve arose solely from the fifth cervical root [12]. Approximately 60 percent of individuals have an accessory phrenic nerve that arises from adjacent cervical roots and/or the cervical plexus and merges with the phrenic nerve low in the cervical region or upper thorax [13,14].

Once formed, the phrenic nerves pass through the neck and chest to innervate the diaphragm muscle. The phrenic nerves originate deep in the prevertebral space and course through the supraclavicular region between the anterior scalene muscles (deep) and the prevertebral fascia (superficial) (figure 1). The upper roots of the brachial plexus descend in close approximation to the phrenic nerves but in a slightly more oblique orientation when compared to their more vertical trajectory. The nearly 100 percent temporary diaphragmatic paralysis that occurs in association with interscalene nerve blocks performed for proximal upper extremity analgesia during shoulder surgery is explained by the relative proximity of phrenic nerves and the fifth and sixth cervical roots of the brachial plexus [15].

The right and left phrenic nerves descend immediately adjacent to the thymus and pericardium (figure 2 and figure 3). The phrenic nerves are primarily motor nerves to the diaphragm, but they also provide sensory feedback from the pericardium and gastrointestinal structures (eg, hiccups involve a continuous reflex arc emanating from inputs through the phrenic nerve, vagus nerve, and sympathetic chain; transmitting to a central mediator; and sending motor output to the diaphragm, glottis, and intercostals) [16].

Diaphragm — The diaphragm muscle is skeletal in nature and is divided into two hemidiaphragms by its midline central tendon (figure 4 and figure 5). The lateral attachments to each chest wall and its position in the center of the trunk account for its importance in body posture and stability. Upon entering the medial portion of each diaphragm at the nerve-muscle interface, the phrenic nerves then arborize laterally throughout the extent of the muscle (figure 4). Anatomic studies have documented terminal subdiaphragmatic extensions of the nerve [17].

The diaphragm is the primary inspiratory muscle, working in conjunction with several accessory respiratory muscles (trapezius, sternocleidomastoid, pectoralis major/minor, intercostal muscles, and the hyoid musculature) to expand the thoracic cavity in a vertical dimension; the intercostal muscles are primarily responsible for horizontal expansion of the ribcage. In response to the increased thoracic domain, the lung expands passively. The expiratory phase of breathing reverses the dimensions of the thoracic cavity back to its resting state.

The structural makeup of the diaphragm consists of an approximately equal makeup of slow twitch (type I) and fast twitch (type II) muscle fibers [18,19]. The resting thickness of the diaphragm muscle, when measured at the "zone of apposition" (ie, area of the diaphragm that directly opposes the rib cage) is estimated to be approximately 1.5 mm, expanding by 2 mm with functional activation [20]. Rapid deterioration in diaphragm structure is seen with phrenic nerve injury and with positive pressure ventilation. In a study comparing structural aspects of the diaphragm from individuals who had or had not been on positive pressure ventilation, there was marked atrophy of the diaphragm and a 57 percent decrease in type I slow twitch muscle fibers after only 18 hours of mechanical ventilation [21].

ETIOLOGIES

Central nervous system — Disruption of the central nervous system (CNS) pathways may occur from vascular insults to the brain or brainstem (ie, stroke, aneurysm), tumor, cervical spinal cord injury or compression [22-24], or systemic neurodegenerative conditions (ie, amyotrophic lateral sclerosis [25]). In addition, central hypoventilation syndrome, or central sleep apnea, prevents a normal response to elevated carbon dioxide levels detected in the brainstem. The symptomatic consequence of many of these conditions (table 1), which often cause bilateral dysfunction, is moderate-to-severe dyspnea requiring partial or complete dependency on oxygen supplementation, positive pressure support, and/or mechanical ventilation. (See "Diagnostic evaluation of adults with bilateral diaphragm paralysis".)

Ventilator dependency in high cervical spinal cord injury is directly responsible for increased morbidity and mortality, even when compared with cervical tetraplegics who are not ventilator dependent [5]. Cervical spinal cord injury leads to a high incidence of ventilator dependency. Of the roughly 6000 new cervical spinal cord injuries per year in the United States, up to 75 percent may require mechanical ventilation due to diaphragmatic paralysis [4].

It is estimated that in at least 20 percent of cervical spinal cord injuries requiring mechanical ventilation, the neural degeneration is not limited to the spinal cord [6]. The loss of anterior horn cells in the spinal tracts will also result in Wallerian degeneration within the phrenic nerves complicating efforts to achieve ventilator weaning. (See "Respiratory physiologic changes following spinal cord injury", section on 'Impairment of ventilatory muscle function'.)

Peripheral nervous system — Diaphragmatic paralysis has been associated with a cervical radiculopathy (C5), either as an isolated finding or in conjunction with other muscular deficits [26,27]. Cervical radiculopathy is characterized by pain, paresthesias, and/or paralysis resulting from irritation of a nerve emanating from the cervical spine (table 2) [28]. Upper extremity muscle weakness due to a C5, C6, or C7 radiculopathy is often readily detectable on clinical examination during strength testing of the shoulder, biceps, or triceps, respectively. However, an isolated C4 radiculopathy may be more difficult to identify, presenting with nonspecific axial neck pain; however, it can also be associated with diaphragmatic paralysis [26,27]. (See "Clinical features and diagnosis of cervical radiculopathy" and "Treatment and prognosis of cervical radiculopathy".)

Idiopathic paralysis and viral neuritis (ie, Parsonage-Turner syndrome) are other etiologies for diaphragmatic paralysis reported in the literature [5,6]. Although viral neuritis has very specific presenting signs and symptoms (eg, fever, malaise, arm weakness, nausea/vomiting) that may be correctly diagnosed when exhibited in close temporal relation to the onset of dyspnea, idiopathic paralysis is truly a diagnosis of exclusion. Clinicians should suspect that a subset of patients initially diagnosed with "idiopathic" or "virally mediated" diaphragmatic paralysis may have a C4 or C5 radiculopathy. (See "Brachial plexus syndromes" and "Diagnosis and management of nontraumatic unilateral diaphragmatic paralysis (complete or partial) in adults", section on 'Etiology'.)

Phrenic nerve injury occurring in the neck is commonly from iatrogenic or traumatic injury. In the chest, iatrogenic mechanisms predominate. Injury to the phrenic nerve due to iatrogenic or traumatic mechanisms are discussed in the next sections. Such injuries are typically unilateral. (See 'Iatrogenic' below and 'Traumatic' below.)

Iatrogenic — Phrenic nerve injury has been reported as a complication of surgical procedures and other interventions that involve structures in the vicinity of the neuromuscular pathways in the neck or mediastinum.

Phrenic nerve injury occurs in up to 10 percent of the estimated 250,000 cardiac procedures performed annually in the United States [29]. Phrenic nerve injury resulting from cardiac bypass surgery is most often due to either hypothermic damage from use of heart cooling or direct injury during isolation and transfer of the internal mammary artery pedicle. (See "Neurologic complications of cardiac surgery", section on 'Phrenic nerve'.)

Aortic or mitral valve repair/replacement can also lead to phrenic nerve injury. It is not yet known whether minimally invasive methods of valve surgery will alter incidences of nerve injury. Procedures performed to alleviate atrial fibrillation, such as the Maze procedure and cardiac ablation, have both been reported to result in diaphragmatic paralysis [30,31].

Mediastinal procedures such as thymectomy, particularly when performed for malignancy, are associated with a 1 to 2 percent incidence of phrenic nerve injury [32,33]. Nerve resection may be required as part of the resection to achieve negative margins. As an example, carcinoma of the lung may require intentional sacrifice of the phrenic nerve if there is direct invasion [34-36]. Alternatively, diaphragmatic paralysis may be an unintended consequence due to dissection in a scarred operative field filled with adhesions. Patients undergoing lung transplantation may also suffer the effects of phrenic nerve injury due to the extensive restructuring of the thoracic cavity and the formation of dense adhesions that can cause a chronic compressive neuropathy [37]. Phrenic nerve injury has also been associated with liver transplantation, possibly from irritation or compression of the intramuscular portion of the phrenic nerve within the diaphragm [38].

Transient and sometimes permanent diaphragmatic paralysis can be a consequence of interscalene nerve blocks [39]. The incidence is estimated at <1 percent. These are likely due to either mechanical or pharmacological insult to the phrenic nerve, or a combination, that leads to segmental nerve ischemia and loss of conduction [40,41]. Transient diaphragmatic paralysis with interscalene block is nearly uniform with an onset of 5 to 10 minutes after the block is placed, with resolution typically four to six hours later depending on dosing [15]. Modified techniques and dosing have reduced the incidence [42,43]. Cervical spine disease may increase the risk for persistent paresis following interscalene nerve block [44]. (See "Upper extremity nerve blocks: Techniques", section on 'Interscalene block'.)

Other procedures that can lead to phrenic nerve injury include cardiac ablation procedures, cardiac pacemaker implantation, cervical lymphadenectomy, carotid-subclavian bypass, thoracic outlet release, and chiropractic manipulation [2,29,40,41,45,46]. Cardiac ablation procedures may result in phrenic nerve injury from transmission of thermal energy between the pulmonary vein and the phrenic nerve in the region where they intersect in the upper chest. Alternatively, stimulation of cardiac pacemaker wires in proximity to the phrenic nerve as it courses near the pericardium may conduct repetitive electrical activity resulting in nerve dysfunction. During thoracic outlet surgery, dissection in and around the anterior scalene muscle during rib resection may result in dense adhesions forming around the phrenic nerve reducing or eliminating nerve transmission. Lastly, repetitive chiropractic manipulation in the cervical region has been associated with phrenic nerve injury and is likely due to recurrent inflammation in the muscles and fascia surrounding the cervical roots and phrenic nerve, resulting in chronic compression.

Some patients have restricted diaphragmatic movement due to noncompliance within the lung or chest wall, rather than from neuromuscular deficits. This occurs most commonly due to extensive intrathoracic scarring from prior surgery, radiation therapy, or chronic infection. (See "Chest wall diseases and restrictive physiology".)

The phrenic nerve has been described as a nerve transfer donor for brachial plexus reconstruction [47]. Donor site morbidity has not been assessed adequately. The authors have personal experience in evaluating and treating two patients who presented with exertional dyspnea one year following this procedure performed elsewhere. Based on diagnostic criteria in our treatment algorithm, one patient was offered phrenic nerve reconstruction while plication was recommended for the second patient.

Phrenic nerve resection may be required during oncologic procedures, such as tumors of the phrenic nerve or malignant thymomas [48], which results in the expected respiratory symptomatology as described above.

Traumatic — Blunt or penetrating injury occurring anywhere along the neuromuscular pathways, including injury of the diaphragm, can impair respiratory function.

●Neck/shoulder – Traumatic phrenic nerve injury is most often from traction or whiplash injuries impacting the neck and shoulder. There is also a reported 15 percent incidence of phrenic nerve injury in severe brachial plexus injuries [49-51]. While cervical radiculopathy is more often due to degenerative disc disease, it may be related to antecedent trauma in 15 percent [52]. (See "Brachial plexus syndromes".)

●Chest – Penetrating or blunt trauma to the lower thoracic region is more likely to result in direct injury to the diaphragm muscle and, if severe enough, can restrict or eliminate functional muscle activity. Traumatic phrenic nerve injury within the thoracic cavity is uncommon as the nerve pathway is deeply interposed between the medial portion of the lung and the pericardium but is possible with rupture of the pericardium [53,54]. Any direct injury that occurs is likely to be associated with significant other potentially lethal injuries. Care is taken during the conduct of resuscitative thoracotomy to avoid injury to the left phrenic nerve. (See "Recognition and management of diaphragmatic injury in adults" and "Resuscitative thoracotomy: Technique".)

●Abdomen – Penetrating or blunt trauma to the abdomen is more likely to directly injure the diaphragm muscle, rather than the phrenic nerve. (See "Recognition and management of diaphragmatic injury in adults".)

Intramuscular and others — Penetrating trauma to the lower thoracic or abdominal region is more likely to result in direct injury to the diaphragm muscle. Traumatic or iatrogenic injury that disrupts the integrity of the diaphragm muscle or its attachments, and subsequent muscular repair, can lead to respiratory dysfunction.

There are systemic myopathies, such as muscular dystrophy, that can impact the diaphragm muscle directly, leading to dysfunction. (See "Diagnosis and management of nontraumatic unilateral diaphragmatic paralysis (complete or partial) in adults" and "Diagnostic evaluation of adults with bilateral diaphragm paralysis".)

SURGICAL REFERRAL — Immediate referral is appropriate when phrenic nerve injury is a known transection due to iatrogenic or traumatic injury to determine the optimal timing for surgery [55]. Early referral and repair may improve outcomes. In other circumstances, patients should be referred for surgical evaluation if symptoms persist in spite of six to eight months of optimal medical management [56], which may have included efforts to improve diaphragm function using diaphragm physical therapy. (See 'Timing' below.)

However, many patients do not receive evaluation for possible intervention for a variety of reasons. A subset of these patients may not come to medical attention due to favorable compensatory respiratory mechanisms that have reduced, but have not eliminated, symptoms, or symptoms may not be recognized as being related to diaphragm dysfunction. Others with a known diagnosis may simply not be referred due to a lack of awareness of treatment options and outcomes.

SURGICAL EVALUATION — Clinical and diagnostic evaluation by the surgeon performing surgery is necessary to determine feasibility of surgical correction, the most appropriate treatment approach, and the timing of surgery. Clinical findings are helpful to confirm the diagnosis, identify the location of injury, judge the severity of symptoms, and, in the recovery period, help monitor improvement.

Clinical evaluation — The goals of surgical treatment are to reverse the symptoms (partially or completely) associated with unilateral or bilateral paralysis, which are often quite disabling; reduce complications (infection related to atelectasis, sleep-disordered breathing, exacerbation of underlying pulmonary disease, fatigue, depression); and improve quality of life [57-60].

Symptoms associated with unilateral and bilateral diaphragm dysfunction include the following:

●Unilateral (see "Diagnosis and management of nontraumatic unilateral diaphragmatic paralysis (complete or partial) in adults", section on 'Clinical manifestations'):

•Exertional dyspnea, particularly when bending (eg, tying shoes), and orthopnea [61]. Certain activities, such as swimming and ascending stairs, tend to result in the greatest degree of respiratory difficulties. The well-conditioned individual with a normal body mass index will sometimes be well compensated and not exhibit any obvious symptoms other than a reduction in maximal physical activity.

•Exacerbation of gastrointestinal reflux and bloating can occur with left-sided diaphragmatic paralysis due to the altered position of the stomach with elevation of the diaphragm [30].

●Bilateral: Bilateral diaphragmatic dysfunction often leads to severe respiratory disturbances and often requires supplemental oxygen or mechanical ventilation. (See "Diagnostic evaluation of adults with bilateral diaphragm paralysis", section on 'Clinical manifestations'.)

Secondary symptoms reported by patients with unilateral or bilateral diaphragmatic dysfunction include easy fatigability, lethargy, chronic cough, breathlessness, sexual incompetence, and depression.

Comorbid conditions, body mass index, and patient age must be taken into consideration in surgical planning. Sleep-disordered breathing from diaphragmatic paralysis needs to be clearly distinguished from obstructive sleep apnea. These patients are susceptible to respiratory infection due to incomplete lung expansion as a consequence of chronic atelectasis [58]. It is unknown whether chronic diaphragmatic paralysis is also a risk factor for the development or exacerbation of parenchymal lung disorders such as asthma or chronic obstructive pulmonary disease (COPD). However, it has been clearly demonstrated that patients with COPD exhibit diminished diaphragmatic function secondary to reduced neural drive [59,60].

On physical examination, the most obvious finding is diminished breath sounds at the base on the involved side when auscultating the lung fields. Occasionally, there will be elicitation of tingling or a "pins and needles" sensation in the supraclavicular region in response to light tapping (ie, Tinel's sign), supporting the diagnosis of a phrenic neuropathy in the cervical region. Unless the diagnosis is due to a major insult to the cervical roots and/or brachial plexus, examination of the upper extremities will be unremarkable. The breathing patterns of those with bilateral dysfunction reveal significant dependence on accessory respiratory muscles for adequate ventilatory exchange.

Preoperative therapies — Diaphragmatic dysfunction resulting from disease or trauma results in both respiratory and postural impairments. For those who are exhibiting spontaneous recovery, physical therapy for diaphragm retraining can be extremely effective for symptom relief. Therapy is continued as long as necessary [62-67]. In patients with a diaphragm devoid of innervation, therapy can only improve compensatory mechanisms by strengthening accessory respiratory muscles. For patients without evidence of spontaneous improvement on diagnostic evaluation and subjective reporting, diaphragm physical therapy will not provide a benefit and is reserved for the postsurgical period. (See 'Diaphragm physical therapy' below.)

Diaphragmatic paralysis is associated with sleep-disordered breathing from reduced inspiratory muscle force, and patients with diaphragmatic paralysis benefit from nocturnal positive pressure support regardless of whether there is an established co-diagnosis of obstructive sleep apnea [57]. Positive pressure ventilation using either continuous positive airway pressure (CPAP) or bilevel positive airway pressure (BiPAP) can be extremely helpful to improve respiratory activity during sleep, improving sleep quality. However, it is not a mechanism for diaphragm retraining or strengthening since the diaphragm is passively mobilized during treatment activation. There is no evidence that positive pressure ventilation improves spontaneous diaphragmatic activity or lessens primary symptoms, especially exertional dyspnea.

Whether the patient benefits more from BiPAP compared with CPAP may help to distinguish sleep disorders due to isolated diaphragmatic paralysis from obstructive sleep apnea. The ability to maintain higher pressures during inspiration and then provide a lower level during the expiratory phase would seem to favor BiPAP for an inspiratory muscle disorder, whereas upper airway obstruction may benefit more from higher pressures during both phases of breathing (ie, CPAP). In a review of 66 patients with unilateral or bilateral diaphragmatic paralysis, all of whom exhibited abnormal sleep studies consistent with sleep-disordered breathing, patients exhibited demonstrable improvements using positive airway pressure supplementation [68]. Not surprisingly, less than 40 percent tolerated CPAP, with the rest requiring BiPAP.

Preoperative studies — The manner in which patients with diaphragm dysfunction present to the surgical specialist is variable. Sometimes patients present with complete diagnostic evaluation; at other times there are no studies to work from. Diagnostic studies (table 3) are repeated as necessary depending on how long ago they were performed and whether the patient has reported any improvement or worsening of symptoms. The diagnostic evaluation confirms the presence of disease and gauges the severity. The location of the lesion is determined based upon history or suggested by the location of any organic pathology.

Chest fluoroscopy — Chest fluoroscopy or sniff testing evaluates the resting (baseline) position of the diaphragm and the extent of downward excursion during resting and forceful inspiration. The performance of the sniff test is described separately. (See "Diagnosis and management of nontraumatic unilateral diaphragmatic paralysis (complete or partial) in adults", section on 'Imaging' and "Diagnostic evaluation of adults with bilateral diaphragm paralysis", section on 'Imaging studies'.)

Sniff testing identifies paresis (ie, limited downward movement), paralysis (ie, complete absence of movement), or paradoxical movement of the diaphragm (ie, upward movement during rapid inspiratory efforts) to confirm the diagnosis. With unilateral paralysis, diaphragm dysfunction is easy to identify by comparing movements on the video with the normal contralateral diaphragm; however, with bilateral diaphragm dysfunction, subtle abnormalities may be difficult to appreciate. For most patients, performing the diagnostic study in supine and upright positions can reveal differences that may assist in qualifying the severity of the dysfunction. The sniff maneuver can also be performed using ultrasonography, which provides more accurate measurements of diaphragm thickness.

Pulmonary function testing — Pulmonary function tests (PFTs) are performed as part of a standard diagnostic evaluation in patients with diaphragmatic dysfunction. While PFTs are not used to directly stratify which patients are offered surgery, preoperative tests are used for comparison as an objective measure of functional recovery.

The reduction in pulmonary compliance as a result of the compromised inspiratory muscle activity results in a mild-to-moderate restrictive ventilatory defect. Reductions in expiratory volumes may be more pronounced in the supine position. Well-conditioned individuals with unilateral paralysis may often have percent predicted values within a normal range for their age. Alternatively, there are other patients with diaphragmatic paralysis who develop secondary pulmonary disorders, such as asthma or sleep-disordered breathing, and demonstrate mixed restrictive-obstructive deficits on spirometry testing. When bilateral diaphragmatic dysfunction is present (eg, cervical stenosis), the results of spirometry testing will usually indicate much more severe restrictive ventilatory deficiency. The spirometry parameters recorded during an expiratory effort (ie, FEV1 [forced expiratory volume in one second]) may underestimate the overall impact of an inspiratory deficiency due to diaphragmatic dysfunction. (See "Diagnosis and management of nontraumatic unilateral diaphragmatic paralysis (complete or partial) in adults", section on 'Pulmonary function testing' and "Diagnostic evaluation of adults with bilateral diaphragm paralysis", section on 'Pulmonary function tests'.)

Cross-sectional imaging — Computed tomography (CT) or magnetic resonance (MR) imaging is almost always appropriate to rule out organic pathology, such as cervical disease or tumor, and should be recommended based on the particulars of patient history. As an example, individuals with a history of neck or back pain, especially with concomitant upper extremity weakness or paresthesias, require cervical MR imaging to look for cord compression, disc disease, or radiculopathy. Alternatively, patients with diaphragmatic paralysis whose history is significant for benign or malignant tumors of the thyroid, thymus, breast, or lung require imaging to eliminate tumor pathology causing neural injury. Current MR imaging standards do not reliably permit identification of phrenic nerve entrapment or chronic nerve inflammation, as opposed to similar pathology involving the larger brachial plexus neural structures.

Symptomatic patients with abnormalities on cervical MR imaging should be evaluated by an orthopedic spine surgeon or neurosurgeon. If there is no indication for surgical repair of cervical spine abnormalities, there may still be an option for performing peripheral nerve surgery to correct phrenic nerve impingement to reinnervate the diaphragm, and appropriate referral should be considered. (See "Diagnosis and management of nontraumatic unilateral diaphragmatic paralysis (complete or partial) in adults", section on 'Imaging' and "Diagnostic evaluation of adults with bilateral diaphragm paralysis", section on 'Imaging studies'.)

Electrodiagnostic testing — Electrodiagnostic evaluation is an effective way to quantify the nerve injury and muscular dysfunction and includes both phrenic nerve conduction studies and electromyography of the diaphragm. Although not typically able to identify the location of the lesion, the values on the dysfunctional side can be compared to the contralateral normal side (for unilateral injury) or to standard reference values (for bilateral injury) [69]. Identification of intact voluntary motor units (VMUs) in the diaphragm indicates the potential for functional diaphragmatic recovery after phrenic nerve reconstruction and forms the basis for surgical treatment algorithms [30]. The limiting factor with this diagnostic modality is technical proficiency and skill for precise needle placement to record accurate neuromuscular activity. (See "Overview of nerve conduction studies", section on 'Motor unit number estimates (MUNE)'.)

Transdiaphragmatic pressure measurement — Measurement of pressure differentials across the diaphragm is performed by simultaneously recording from transducers in the esophagus and stomach, thus obtaining values for thoracic and abdominal pressures, respectively (waveform 1). The force developed by the diaphragm is measured indirectly as a pressure (ie, the force developed divided by the surface area over which the force acts). Although there are numerous reports of the utility of this technique in the clinical setting to diagnose (and quantify) diaphragmatic dysfunction, this diagnostic modality is more widely performed outside of the United States [70].

PHRENIC NERVE RECONSTRUCTION — Phrenic nerve reconstruction was first reported in 2011 as an application of the nerve repair techniques commonly and successfully used with other peripheral nerve injuries (eg, brachial plexus [71]) [31]. Phrenic nerve reconstruction has since become well established as a method of restoring functional activity in patients with diaphragmatic paralysis.

Timing — The timing for surgical intervention varies based upon the mechanism of injury and the severity of symptoms [72].

When phrenic nerve injury is a known transection or sacrifice of the nerve during surgery in proximity to its course, immediate repair or early referral for repair is most likely to result in recovery of function [34,55]. It may be optimal to repair the phrenic nerve in the same setting; however, only a few institutions have this capability. When this option is not available, early referral (days to weeks) for phrenic nerve repair or diaphragm plication during the index surgery (or later) may be appropriate depending upon the age and medical comorbidities of the patient.

For other injury mechanisms (eg, compression, hypo/hyperthermia, traction, blunt head/neck trauma), where nerve continuity would be expected, watchful waiting is appropriate since some patients with unilateral injury will have improvements in respiratory function over time [73]. Although no conclusive studies exist on the perfect timing of phrenic nerve repair, based on data for other peripheral nerve injuries classified as those with a possibility of spontaneous recovery [56], a waiting period of at least six months is recommended. For most cases of phrenic nerve injury, waiting 8 to 12 months identifies those whose symptoms are not improving with optimal medical management (eg, diaphragm physical therapy) and who may benefit from phenic nerve repair. Earlier intervention may be warranted, particularly if there is bilateral diaphragm dysfunction with oxygen or ventilator dependency, or if the patient's individual circumstances require a more aggressive approach to prevent detriment to his/her career or functional status (eg, active duty in military).

During the period of watchful waiting, the patient is offered ongoing diaphragm physical therapy. Quarterly or biannual reevaluation should be obtained to document any incremental improvements in function. Observation is continued if there is unequivocal partial or complete return of functional diaphragmatic activity. In the absence of progressive improvements, referral should be considered. Patients who report mild-to-moderate subjective improvement without objective diagnostic evidence of diaphragmatic recovery may be compensating well using accessory respiratory muscles but can still be considered for surgical intervention.

Candidates — Candidates for phrenic nerve reconstruction are those who have moderate-to-severe respiratory impairment from unilateral phrenic nerve injury and who demonstrate intact voluntary motor units (VMUs) in the diaphragm on electrodiagnostic testing (algorithm 1). Whether the location of injury is in the cervical region, mediastinum, or thoracic cavity, the phrenic nerve can be approached and reconstructed successfully. Immediate intervention is offered to patients who experience a known or inadvertent phrenic nerve transection, most often during thymectomy or pulmonary lobectomy, and occasionally from penetrating neck trauma. In these settings, it would be optimal to repair the phrenic nerve when the injury is identified, but if expertise is not available, the young, otherwise healthy patient can be referred for repair as soon as possible whereas the sedentary, older patient or those with a poor prognosis should undergo diaphragm plication at the time of the index surgery. (See 'Etiologies' above and 'Timing' above and 'Procedures' below.)

Patients with nerve injuries previously deemed to be idiopathic or virally induced are often candidates for phrenic nerve reconstruction as long as systemic disease is excluded and electrodiagnostic testing supports the potential for neural regeneration. Many presumed "idiopathic" cases are often determined intraoperatively to be from localized or diffuse C4 and/or C5 radiculopathies and benefit from these same nerve reconstructive techniques to achieve restoration of phrenic nerve integrity and diaphragmatic function.

Patient selection is critical to successful outcomes after phrenic nerve reconstruction. Patients with major comorbid conditions and/or an inability to participate in a program of diaphragm rehabilitation will not be suitable candidates for this surgical treatment. Optimal candidates for phrenic nerve reconstruction are the same as for any peripheral nerve injury repair: those of younger age, those with recent injury who can undergo early intervention, and those with no adverse comorbid conditions (eg, obesity, diabetes, peripheral neuropathy, autoimmune disease).

Inability to participate in an aggressive program of diaphragm rehabilitation will reduce the likelihood of significant long-term recovery. Patients who are unable to participate in diaphragm physical therapy due to comorbid conditions (ie, obesity, knee arthritis) have lower rates of functional recovery. (See 'Diaphragm physical therapy' below.)

Phrenic nerve reconstruction can sometimes be offered simultaneously with diaphragm pacemaker implantation so that electrical stimulation may be used to recondition the diaphragm in lieu of physical therapy. But, for long-standing diaphragmatic paralysis, particularly in older adults or in patients with diabetes, diaphragm plication may be a better choice. (See 'Role of diaphragm plication' below.)

Procedures — Specific details regarding techniques for phrenic neurolysis, interposition nerve grafting, and neurotization vary depending on the neural pathology [31,34].

Neurolysis — Any peripheral nerve in the body, particularly those in an anatomically compact area, may be subject to chronic impingement or compression. Carpal tunnel syndrome is an obvious example of this. The fourth and fifth cervical root contributions to the phrenic nerve and the phrenic nerve proper are interposed snugly between the prevertebral fascia and the anterior scalene muscle. Any swelling in the scalene, thickening of the fascia, or adhesions with adjacent blood vessels (ie, transverse cervical artery and vein, or branches of the prevertebral artery) can result in nerve impingement and a deficiency in nerve transmission.

The basic approach to phrenic nerve reconstruction includes first accessing the phrenic nerve at or near the location of injury and performing a nerve decompression (ie, neurolysis). A neurolysis consists of removal of external sources of nerve impingement or compression from scar and fibrosis, usually due to thickened or adherent muscle, fascia, and/or vasculature.

Neurotization/nerve transfer — To further improve axonal recovery and restore nerve conductivity, nerve transfer (or neurotization) or an interposition graft can be performed to recruit new sources of axons. These nerve reconstruction techniques facilitate nerve regeneration over a duration of time that is proportional to the distance between the location of repair and the target muscle.

Neurotization is a technique whereby an intact donor nerve is divided distally and brought into proximity of the recipient (damaged) nerve and coapted to bypass the injured segment in an end-to-side fashion. For patients with suspected cervical radiculopathy, neurotization is more likely to be performed as a means to recruit axons from nerves arising at different cervical levels, thereby "bypassing" the degenerated cervical root. In doing so, care is taken not to create a secondary deficit by using donor nerves with overlapping innervation distributions, such as the spinal accessory nerve.

Diaphragmatic paralysis resulting from a segmental, localized iatrogenic injury in the cervical region will more likely be repaired with a sural nerve interposition graft coapted to the phrenic nerve above and below the location of injury. It is possible to return to an operative field and repair an iatrogenic injury; however, as would be expected, this can be a technically demanding reconstruction.

Adjunctive procedures — Patients with complex unilateral or bilateral diaphragmatic dysfunction are considered for multimodality treatment to enhance recovery (algorithm 1). We define complex as remote injuries (>5 years), hostile surgical fields from prior surgery or radiation, or comorbid conditions that would limit or prevent participating in physical therapy.

Diaphragm pacing — Diaphragm pacemakers are approved by the US Food and Drug Administration (FDA) as a surgical treatment option for ventilator dependency in spinal cord injury (and other central nervous system disorders) and can be extremely effective for partial or complete ventilator weaning in high cervical tetraplegia [74]. In these patients, diaphragm pacing aims to provide natural negative pressure ventilation to replace positive pressure mechanical ventilation, to improve quality of life by making it easier to eat and speak, to improve airway clearance to reduce pulmonary morbidity, and to increase mobility [75]. However, if there is complete loss of phrenic nerve conductivity, which can occur with nerve transection, or long-standing injuries with complete neuromuscular degeneration, diaphragm pacemakers will not be effective. (See "Pacing the diaphragm: Patient selection, evaluation, implantation, and complications".)

Indications for diaphragm pacemakers have expanded to include unilateral or bilateral diaphragmatic dysfunction. Electrical stimulation may enhance nerve regeneration and lessens muscle atrophy. In one small study, the use of diaphragm pacemakers in 21 patients with diaphragm dysfunction (and at least partial preservation of phrenic nerve activity) improved clinically relevant respiratory parameters in 62 percent [76]. A study evaluating simultaneous bilateral pacemaker implantation and staged, unilateral phrenic nerve reconstruction in 14 patients with bilateral phrenic nerve injuries demonstrated nerve reconstruction plus pacemaker implantation resulted in greater functional muscle recovery compared with only pacemaker placement [77,78]. All 14 patients reported improvements in respiratory function.

It is the authors' practice to consider simultaneous phrenic nerve reconstruction and diaphragm pacemaker implantation in complex unilateral cases to improve overall outcomes (algorithm 1) [79]. (See 'Timing' above.)

Candidates for diaphragm pacing have intact VMUs on electrodiagnostic testing. (See 'Preoperative studies' above and "Pacing the diaphragm: Patient selection, evaluation, implantation, and complications", section on 'Selection of potential candidates'.)

In addition:

●Phrenic nerve integrity must be present. If there is no native phrenic nerve function, then nerve reconstruction should be considered simultaneous with, or following, pacemaker implantation.

●There cannot be any active respiratory infection (eg, pneumonia, bronchitis).

●Cognitive function should be intact to allow for patient motivation and retraining.

●Family and caregiver support must be adequate.

Role of diaphragm plication — Plication of the diaphragm has been a long-standing surgical treatment option for reversing symptomatology in chronic diaphragmatic paralysis. But, once a plication has been performed, it is unlikely that the diaphragm will retain the capacity for functional movement, and failed plication surgery due to relapse or recalcitrant symptoms cannot be successfully salvaged with phrenic nerve reconstruction or implantation of a pacemaker. Thus, for patients with the potential for reinnervation (algorithm 1), we reserve diaphragm plication for primary treatment failures. Surgical plication can be offered as first-line treatment for treatment of symptomatic dyspnea when other methods of reinnervation or pacing are not feasible (ie, no potential for reinnervation), or for those who are not good candidates for early nerve repair following injury or planned nerve resection. (See 'Candidates' above.)

Although no conclusive studies exist on the optimal timing of plication for permanent injury, a 12-month waiting period typically allows for sufficient recovery to judge that the damage is permanent, unless it is known that the nerve was sacrificed or transected.

The plication technique is performed with a permanent multifilament polyester suture (eg, 0-Ethibond or 0-Mersilene) as a way to minimize relapse of the elevated diaphragm position. In doing so, more scar is created within the muscle, essentially rendering it nonfunctional and not likely to be amenable to functional correction using nerve reconstruction methods or a diaphragm pacemaker.

Plication is more useful for improving symptoms in patients with chronic diaphragmatic paralysis, rather than as a means to wean from mechanical ventilation. Freeman reported that only one of four ventilator-dependent patients was weaned from the ventilator after plication, whereas 17 of 19 patients who were unable to work secondary to dyspnea were able to return to work by six months following plication [78]. Patients who are severely obese and those with long-standing paralysis are less likely to benefit from plication [80].

The aim of plication is to minimize the loss of thoracic space and prevent paradoxical motion (figure 6). The increased thoracic domain increases lung volumes (eg, total lung capacity, vital capacity, expiratory reserve volume, functional residual capacity), decreases atelectasis, improves ventilation perfusion mismatch, and increases arterial PaO₂ [78,81,82]. Spirometry testing is improved in both sitting and supine positions [83].

The traditional approach for plication is through standard posterolateral thoracotomy [80,83-89]. Minimally invasive approaches (eg, video-assisted thoracoscopic surgery [VATS], laparoscopy) have reduced complication rates. A VATS approach has progressively replaced open thoracotomy for this procedure [80,82,90]. In a review comparing the VATS approach with thoracotomy, VATS achieved similar results based on pulmonary function tests (PFTs), dyspnea scores, and functional assessment but with shorter length of stay, lower complication rates, and lower mortality rate [86]. Several authors have also supported a laparoscopic approach [87,91]. Robotic technology can be applied to either VATS or laparoscopic approaches [92,93]. Laparoscopic approaches on the right are limited by the liver, and prior abdominal operations may make the laparoscopic approach more difficult compared with VATS. Prior chest approaches may make VATS more difficult, but experienced surgeons frequently perform redo surgeries and are able to overcome these issues.

Complications occur in approximately 15 percent, more commonly when using open techniques, and have included pneumonia, pleural effusion, and pulmonary embolus. Recurrence of respiratory symptoms due to weakening or loosening of the plication is not commonly reported but would be expected to occur in a certain subset of patients.

Outcomes — There are no randomized trials comparing phrenic nerve reconstruction with or without adjunctive diaphragm pacer placement with other therapies. Case reports and case series have shown excellent outcomes for immediate repair of anticipated nerve resection or known iatrogenic injury [34]. For chronic diaphragmatic paralysis, case reports [94], case series [22,31,39,79,95], and observational studies [30,96] have encouraging results. Respiratory parameters are consistently improved following successful phrenic nerve reconstruction, and between 75 and 85 percent of patients do not require any form of respiratory care.

In the initial experience of phrenic nerve reconstruction in patients with chronic, unilateral diaphragmatic paralysis, diaphragm function was improved in eight of nine patients [31]. In a later nonrandomized study of 92 patients with symptomatic diaphragmatic paralysis, patients were assigned to phrenic nerve reconstruction, nonsurgical care, or diaphragm plication [30]. Average respiratory improvements in the phrenic reconstruction group included 13 percent improvement in forced expiratory volume in 1 second (FEV1) and 14 percent improvement in forced vital capacity (FVC), which were not significantly different compared with the plication group (17 and 16 percent, respectively). There were no significant differences in baseline parameters between these nonrandomized groups; however, patients undergoing plication were not evaluated using electrodiagnostic testing. The average postoperative FEV1 was 71 percent, and FVC was 73 percent in both the plication and phrenic nerve reconstruction groups. On electrodiagnostic testing, there were significant improvements with phrenic nerve reconstruction with a 69 percent improvement in conduction latency and a motor amplitude increase of 37 percent. Surgical morbidity associated with phrenic nerve reconstruction, particularly when performed in the cervical region, was lower (at 1 to 3 percent) compared with thoracic or laparoscopic approaches for plication.

In a 2021 retrospective review of 400 patients undergoing phrenic nerve reconstruction, postoperative FEV1 and FVC percent predicted values improved by 8 and 10 percent at one-year follow-up, by 22 and 18 percent at two years, supporting the notion of incremental recovery with longer follow-up and continued rehabilitation [97]. Furthermore, there was evidence of significant neuromuscular regeneration with an 82 percent increase in compound muscle action potential and 27 percent increase in resting diaphragm thickness.

In a retrospective review, 14 patients with an average of 21 months of ventilator dependency who had combined lesions of the cervical spinal cord and phrenic nerves were treated with simultaneous microsurgical nerve transfer and implantation of diaphragmatic pacemakers [79]. Recovery of diaphragm electromyographic activity was demonstrated in 13 of 14 patients. Two patients recovered voluntary control of diaphragmatic activity and regained the capacity for spontaneous respiration. An additional eight patients achieved sustainable periods of ventilator weaning (average 10 hours/day). In a separate series in which four patients sustaining traumatic cervical spine injuries underwent intercostal to phrenic nerve grafting and pacemaker implantation, time off the ventilator with pacing ranged from 2 to 24 hours at one-year or greater follow-up [95]. In a subsequent prospective analysis, a surgical treatment algorithm using pacemakers, phrenic nerve reconstruction, and/or diaphragm muscle replacement, as indicated, was evaluated in 10 patients with chronic ventilator dependency due to cervical tetraplegia [98]. Prior failed attempts at pacing alone occurred in seven of these patients. Following surgical treatment, 8 of 10 patients achieved partial or complete weaning. Though in a limited number of patients, this outcome supports a surgical option for ventilator-dependent cervical tetraplegics for reducing or eliminating mechanical ventilation.

In a study evaluating long-term outcomes, 180 patients (134 males and 46 females with an average age of 56 years; range: 10 to 79 years) with unilateral or bilateral diaphragmatic paralysis were treated with phrenic nerve reconstruction and followed for a median of 2.7 years [96]. No patients were ventilator dependent preoperatively, but more than 70 percent required positive pressure oxygen supplementation at night. Mean baseline percent predicted values for FEV1, FVC, vital capacity (VC), and total lung capacity were 61, 63, 67, and 75 percent, respectively. The corresponding percentage improvements in percent predicted values were 11, 6, 9, and 13 percent. Postoperative SF-36 physical functioning survey scores (a set of generic, coherent, and easily administered quality-of-life measures) were 65 percent, up from 39 percent preoperatively, with 89 percent of patients reporting overall improvement in breathing function. Nerve conduction latency improved by an average of 23 percent, and there was a corresponding 125 percent increase in diaphragm motor amplitude. All patients reporting improvement also demonstrated corrections on one or more objective measures (sniff test, electromyography, pulmonary function tests).

POSTOPERATIVE CARE AND FOLLOW-UP — Postoperative care after phrenic nerve reconstruction is minimal, though the postoperative course may be longer depending upon the extent and invasiveness of the procedure.

For patients undergoing neck surgery, only one night in the hospital is typically required, and normal activities can be resumed in roughly three weeks. Procedures that involve thoracic access (phrenic nerve reconstruction or plication) usually require placement of a chest tube and a two- to three-day hospital stay. The recovery for these procedures is usually on the order of four to six weeks. Minimally invasive plication surgery performed via video-assisted thoracoscopic surgery (VATS), robotically, or laparoscopically has reduced hospital stay and recovery time. In these cases, normal activities can be resumed in approximately six weeks.

Ventilator-dependent patients who are also being implanted with a diaphragm pacemaker will require intensive care unit care during an estimated two- to three-day hospital stay and will usually return home or to a facility on the ventilator. Pacing trials are most often initiated in two to three weeks following implantation. (See "Pacing the diaphragm: Patient selection, evaluation, implantation, and complications", section on 'Conditioning of the diaphragm'.)

Bilevel positive airway pressure (BiPAP) may also assist in surgical recovery, regardless of whether the patient required such respiratory support preoperatively.

Diaphragm physical therapy — In addition to its role as the primary muscle of inspiration, the diaphragm plays a major role in stabilization of the trunk and spine. Diaphragmatic dysfunction resulting from disease or trauma leads to both respiratory and postural impairments. The physical therapist's role is well established in the management of diaphragm dysfunction based on a broader understanding of the multiple roles of the diaphragm and a targeted approach to rehabilitation [62]. Specific components of diaphragm retraining include thoracic spine and rib cage mobilization techniques, breathing neuromotor retraining, reducing postural demands on the diaphragm, endurance training, and airway clearance [63-67].

Physical therapy for diaphragm retraining can be extremely effective in postsurgical recovery and typically is provided for approximately three months, though the duration is variable at the discretion of the respiratory therapist. Diaphragm physical therapy enhances diaphragmatic activity and expedites diaphragmatic muscle strengthening following successful phrenic nerve reconstruction.

SOCIETY GUIDELINE LINKS — Links to society and government-sponsored guidelines from selected countries and regions around the world are provided separately. (See "Society guideline links: Thoracic trauma".)

SUMMARY AND RECOMMENDATIONS

●Phrenic nerve injury – Injury to one phrenic nerve leads to paralysis of the ipsilateral diaphragm, often leading to respiratory symptoms (eg, dyspnea, orthopnea) that may improve with time. If both phrenic nerves are injured, both diaphragms are affected, which often results in oxygen or ventilator dependency. Diaphragmatic paralysis may lead to sleep-disordered breathing. (See 'Clinical evaluation' above.)

●Pathophysiology – A variety of pathological processes can affect the central or peripheral pathways that control diaphragm muscle function. Common etiologies are injury due to iatrogenic and traumatic mechanisms and chronic nerve compression. Any surgical procedure that involves structures in the vicinity of the neuromuscular pathways in the neck or mediastinum (figure 2 and figure 3) has the potential to cause injury. The most common include coronary artery bypass grafting and mediastinal procedures, such as thymectomy. Traumatic phrenic nerve injury is most often from traction or whiplash injuries impacting the neck and shoulder. (See 'Etiologies' above.)

●Timing of surgery – Surgical intervention should be pursued in a timely fashion according to the mechanism of injury or underlying disorder. When phrenic nerve injury is a known transection or sacrifice of the nerve during surgery in proximity to its course, it would be optimal to repair the phrenic nerve in the same setting. A few institutions have this capability, but if this option is not available, early referral (days to weeks) for phrenic nerve repair or diaphragm plication during the index surgery (or later) may be appropriate depending upon the age and medical comorbidities of the patient. In other settings, a waiting period of six to eight months identifies those whose symptoms are not improving with optimal medical management (eg, diaphragm physical therapy) and who may benefit from phrenic nerve repair. (See 'Surgical referral' above and 'Timing' above and 'Candidates' above.)

●Patient evaluation – Clinical and diagnostic evaluation by the treating surgeon is necessary to determine feasibility of surgical correction, the treatment approach, and the timing of surgery. Diagnostic studies (table 3) are obtained (or repeated as necessary) depending on the clinical presentation. Identification of intact voluntary motor units (VMUs) in the diaphragm indicates the potential for functional diaphragmatic recovery after phrenic nerve reconstruction. (See 'Clinical evaluation' above and 'Electrodiagnostic testing' above.)

●Phrenic nerve reconstruction

•Candidates – Persistently symptomatic patients with phrenic nerve injury and intact VMUs are candidates for phrenic nerve reconstruction. (See 'Candidates' above.)

•Procedures – Phrenic nerve reconstruction may involve neurolysis, neurotization, or nerve interposition depending on the extent of the injury. If phrenic nerve reconstruction is available directly or through referral, then diaphragm plication is reserved for those without intact VMUs, or as a salvage procedure after the failure of other treatments. (See 'Procedures' above and 'Adjunctive procedures' above.)

•Rehabilitation – Following nerve reconstruction, diaphragm physical therapy is important for diaphragmatic muscle strengthening. If the patient cannot participate in postoperative diaphragm physical therapy, adjunctive placement of a diaphragm pacer can aid with recovery. (See 'Diaphragm physical therapy' above.)

•Outcomes – Outcomes for immediate reconstruction of anticipated nerve resection or known iatrogenic injury are excellent. For chronic diaphragmatic paralysis due to both identifiable and idiopathic etiologies, reported outcomes are also excellent, with most patients experiencing improvement in respiratory parameters and improved quality of life. (See 'Outcomes' above.)

ACKNOWLEDGMENT — The UpToDate editorial staff acknowledges Thomas L Bauer, MD, MBA, who contributed to earlier versions of this topic review.