Bilateral phrenic neuropathy responsive to intravenous immunoglobulin treatment

Abstract

The aetiology of phrenic neuropathy is often unknown, but immune mechanisms may play a role. In a typical case of bilateral phrenic neuropathy with paradoxical breathing (video), an inflammatory pathogenesis was suggested by prolonged distal latency of phrenic nerve compound muscle action potentials in nerve conduction studies and a clear-cut albumin-cytologic dissociation. This encouraged us to treat the patient with a standard dose of intravenous immunoglobulin. After obtaining a strong improvement at spirometry, we repeated the second cycle of intravenous immunoglobulin and observed normalization of symptoms within few weeks and no relapse after 3 years. This case suggests that lumbar puncture should be performed in the acute phase of phrenic neuropathies to detect potential responders to immunomodulatory treatment.

Introduction

The diaphragm is innervated bilaterally by the phrenic nerve and is the main respiratory muscle. Diaphragm dysfunction ¹ can arise from different aetiologies (metabolic or inflammatory disorders, after cervical trauma or cardiac surgery, mediastinal masses, myopathies, and neuropathies). While unilateral diaphragm paralysis is usually asymptomatic, a bilateral palsy often leads to orthopnoea and dyspnoea, especially during exercise or when lying supine, and a typical sign known as ‘paradoxical breathing’. This sign can be better appreciated in supine position and corresponds to the paradoxical inward motion of the abdomen during inspiration due to the contraction of accessory inspiratory muscles while the abdominal wall moves inwards (Supplemental Video).

Despite accurate investigations, phrenic paralysis ² often lacks an identifiable cause and is even difficult to be classified. ³ Moreover, the disease is frequently overlooked or referred late (after several months) to the pulmonologist and/or the neurologist. Although an immune-mediated attack against the phrenic nerve has been often postulated, ³ neurologists do not systematically perform a lumbar puncture to exclude inflammatory or infectious diseases, maybe also because of the delayed referral.

Routine diagnostic tests include pulmonary function tests, nerve conduction and needle electromyogram (EMG) studies and radiology imaging (chest X-ray, fluoroscopy and diaphragm ultrasound ⁴ ). In clinical practice, it is important to compare the pulmonary function tests obtained in upright and supine position, because in bilateral phrenic neuropathy, there is typically moderate-to-severe restrictive dysfunction (30–50% of the predicted value for total lung capacity) that becomes more severe when the patient is supine. This decline of function in supine position is due to the cephalad displacement of abdominal contents in concert with ineffective activity of the accessory inspiratory muscles and may account for a decrease in vital capacity up to 50%.

Case description

In September 2016, a healthy, physically active 45-year-old man developed sudden right shoulder pain followed within a few hours by orthopnoea, but no shoulder girdle or limb paresis nor reflex loss. Medical and family history was negative, in particular for neuromuscular diseases, including neuropathies. Symptoms appeared at rest, without any preceding trauma, surgical operation or identifiable infection. In the emergency ward, myocardial infarction was excluded by electrocardiogram and echocardiography. The patient required non-invasive ventilation with a portable bilevel positive airway pressure apparatus during sleep for 5 months, when he was finally referred to our centre (February 2017).

Because of paradoxical breathing (Supplemental Video), we immediately suspected a bilateral phrenic palsy that was confirmed by chest X-rays, computed tomography scan and pulmonary function tests. Forced vital capacity (FVC) was 3.1 L upright (62% of the predicted volume = 4.96 L), declining to 1.4 L supine (−55%). Common causes of neuropathy such as diabetes, kidney failure or vitamin B12/folate deficiency were excluded. Laboratory results showed instead mild acute hepatitis alanine aminotransferase (ALAT) 57 U/L, normal <50; Gamma-glutamyl transferase (GGT) 86 U/L, normal <76 U/L) of unknown origin that recovered spontaneously within 1 week. Autoimmune screening included antinuclear antibodies, extractable nuclear antigen antibodies and complement (C3, C4) dosing that resulted all within the normal range. Infectious screenings for human immunodeficiency virus (HIV), hepatitis B virus, hepatitis C virus, hepatitis E virus (HEV), Lyme disease, Mycoplasma Pneumoniae, HSV I and II, Varicella-Zoster, cytomegalovirus, Bartonella henselae, Treponema Pallidum were negative. Serum antigangliosides IgM and IgG antibodies (GD1a, GD1b, GM1, GM2 and GQ1b) were negative. The cerebrospinal fluid (CSF) analysis, performed 5 months after symptoms appearance, showed albumin-cytologic dissociation (protein 96 mg/dL, white blood cell 1/µL).

Recording of the phrenic nerve compound muscle action potentials (CMAPs) was performed during inspiration, as described elsewhere, ^5,6 using surface electrodes. ² Stimulation was carried out in the supraclavicular fossa, just above the clavicle, placing the active electrode 5 cm above the xiphoid process and the reference electrode 16 cm from the previous on the chest margin ipsilateral to the stimulated phrenic nerve. Stimulus intensity was bilaterally 50 mA and pulse duration of 0.2 ms. The phrenic nerve CMAP amplitudes resulted bilaterally in the normal range (i.e. 0.7 mA left, 0.5 mA right; normal values >0.46 mA), whereas CMAP latencies were bilaterally prolonged (i.e. 9.5 ms left, 11.6 ms right; normal values <8.0 ms) (Online Supplemental Figure 1). ⁵ Remaining nerve conduction studies (tibial, peroneal, sural, median and ulnar nerve; F-waves from tibial and ulnar nerve) were bilaterally normal. Needle EMG of the diaphragm was not performed.

We diagnosed an isolated inflammatory bilateral phrenic neuropathy based on albumin-cytologic dissociation in the CSF and prolonged distal latencies detected in the phrenic nerve conduction studies, a possible sign of demyelination, as previously reported. ⁷ In accordance with three similar cases where a positive response was reported, ^{8 –10} we started a cycle of intravenous immunoglobulin (IVIg) 0.4 g/kg/day for 5 days.

We observed a dramatic and prompt (within 7–10 days) improvement of the respiratory pattern. Orthopnoea disappeared, allowing the patient to stop nocturnal ventilation support, and FVC values increased to 3.5 L upright (predicted = 4.96 L), 2.0 L supine. We, therefore, repeated the second cycle of IVIg, observing further improvement (video) confirmed by respiratory tests: FVC was 4.16 (+34%) sitting, 2.86 L (+104%) supine and respiratory symptoms had disappeared (video). The patient had no relapse after more than 36 months of follow-up.

Discussion

Our initial differential diagnosis included several acute and chronic neuropathies.

A genetic neuropathy seems very unlikely because of negative family history and of the acute onset. Common causes of metabolic neuropathies (diabetes, kidney failure, vitamin deficiency) were excluded by routine blood tests. Viral agents previously associated with phrenic nerve damage, such as HEV ¹¹ and HIV, ⁸ were excluded. Despite we were unable to identify an infectious agent, we cannot exclude that the immune reaction against the phrenic nerve was initially triggered by an unknown virus/bacterium. This is especially true if we consider that all neurological investigations were performed 5 months after disease onset due to late referral. On the other hand, routine screening for autoimmune diseases failed to identify an underlying systemic disease.

Bilateral phrenic neuropathy ³ can occur isolated or in the context of a more generalized inflammatory neuropathy, both acute, such as neuralgic amyotrophy (NA) ^{11 –13} or Guillain–Barré syndrome (GBS), ¹⁴ and chronic, such as chronic inflammatory demyelinating polyneuropathy (CIDP) ^7,15 or multifocal motor neuropathy (MMN). ¹⁶

However, our case does not fully satisfy diagnostic criteria of these diseases. Disease duration (>5 months) and normal upper and lower limb nerve conduction studies are not compatible with GBS. ¹⁷ A typical form of NA ¹⁸ cannot be diagnosed because of the lack of shoulder girdle paresis or amyotrophy and absence of clinical damage in other typically affected nerves (long thoracic, anterior interosseous, suprascapular and axillary).

MMN seems also unlikely considering the acute and painful onset, the symmetrical involvement, the normality of upper and lower limb nerve conduction studies (without conduction blocks) and the absence of IgM GM1 antibodies. ¹⁹ Finally, CIDP standard diagnostic criteria ²⁰ are not satisfied based on normal upper and lower limb nerve conduction studies.

Theoretically, our case may potentially represent a focal variant of NA ¹⁸ or CIDP ²¹ with mononeuropathic (phrenic nerve) involvement. However, focal CIDP accounts of less than 1% in all CIDP cases and our case differs from previously reported CIDP cases ^7,15 with secondary phrenic nerve involvement because the onset of respiratory problems was acute and not slowly progressive, and our patient had no relapses after 3 years of follow-up.

We, therefore, concluded for bilateral phrenic neuropathy of inflammatory origin based on three main observations, that is, albumin-cytologic dissociation in the CSF, prolonged distal latencies in phrenic nerve conduction studies (possibly compatible with demyelination) and prompt response to immunomodulatory treatment with IVIg.

The main limitations in our case are that we have no certainty on the cause of phrenic neuropathy, and our findings support only indirectly a demyelinating process because neither nerve biopsy nor high-resolution nerve ultrasound was performed. Moreover, needle EMG was not performed to avoid iatrogenic pneumothorax in a patient already requiring ventilatory support, but it would have been helpful in the diagnostic process. The finding of normal amplitude of phrenic nerve CMAPs is not incompatible with the diagnosis of phrenic neuropathy, as discussed by other authors, ²² especially if the pathogenesis is demyelinating ⁷ and not axonal, as postulated in our case.

In conclusion, isolated phrenic neuropathy is often considered idiopathic because it does not satisfy the diagnostic criteria of any other inflammatory neuropathies. Although it is known that the majority of phrenic nerve palsy spontaneously improve during their natural history, ¹ our case underlies the importance of considering an immune-mediated aetiology. Based on our literature review, lumbar puncture is not routinely performed in phrenic neuropathy. However, our case highlights the importance of this examination in the acute phase of phrenic neuropathy to identify potential responders to immunomodulatory treatment. It is also possible that the presence or absence of pain during the attack is associated with different treatment responses. ^18,23

We encourage larger collaborative studies on patients suffering from phrenic neuropathy to identify those who may benefit from an immunomodulatory treatment.