Proximal Median Neuropathy Following Anterior Shoulder Dislocation: The Use of Magnetic Resonance Neurography

Abstract

Proximal median nerve injury is an uncommon consequence of anterior shoulder dislocation, especially occurring in isolation of other upper limb peripheral nerve injury. We report the case of an 82-year-old woman with a median nerve injury as detected by clinical and neurophysiological examination following a fall and anterior shoulder dislocation. Magnetic resonance neurography confirmed the diagnosis, but also detected asymptomatic brachial plexus and ulnar nerve involvement. Management was non-operative and there has been some improvement over several months. Our case expands the differential diagnosis for proximal median neuropathy and discusses the utility of neurography in cases of neural injury.

Introduction

Dislocation of the shoulder is one of the commonest major joint dislocations, with a lifetime prevalence of approximately 1.7% in the general population.¹ Accompanying injuries to the structure of the joint itself are common and varied, including fracture, avulsion injury, or rotator cuff tear.¹ Associated neural injury occurs in 5 to 55% of shoulder dislocations,² and when neural injury does occur, the axillary nerve is most commonly affected.³ Typically, more than one peripheral nerve is affected on clinical and neurophysiological examination⁴; a clinically isolated median neuropathy is rarely reported. Magnetic resonance neurography (MRN) is an increasingly utilized modality in peripheral nerve lesions which improves diagnostic accuracy and influences management decisions.^5,6 We report the case of a proximal median neuropathy following shoulder dislocation with asymptomatic brachial plexus and ulnar nerve injury detected on MRN.

Case Description

An 82-year-old woman presented following a fall that resulted in right shoulder anteroinferior dislocation. X-ray revealed a 3 mm bony fragment, consistent with a glenoid rim fracture (“bony Bankart lesion”). Examination in the emergency department demonstrated no focal neurological disturbance in the right arm. The shoulder was reduced in the emergency department and she was discharged home for conservative management.

Three weeks after the fall, she was reviewed for complaints of severe shooting pain in her right arm radiating to her hand, with associated hand weakness and numbness in the tips of her first 3 fingers, which had begun approximately 3 days following the injury. Examination revealed weakness in right abductor pollicis brevis, flexor pollicis longus, and flexor digitorum profundus supplying digit 2, scored as Medical Research Council (MRC) grade 1 to 2. Her power was normal elsewhere. Reflexes were present and symmetrical. She had reduced sensation to pinprick and light touch in the median nerve distribution. Sensory examination was otherwise normal. In light of the involvement of both forearm and hand muscles, a proximal median neuropathy was suspected.

Nerve conduction study and electromyography were performed 6 weeks post-injury (Table 1). The right median (digit 2) sensory nerve action potential was absent. The right median compound motor action potentials recorded from abductor pollicis brevis and flexor pollicis longus were markedly reduced in amplitude, and the F waves were absent. The ulnar sensory and motor studies were normal. These findings were consistent with a severe axonal median neuropathy proximal to the anterior interosseous nerve on the right. Needle electromyography revealed neurogenic recruitment with large polyphasic motor units and active denervation changes in abductor pollicis brevis, flexor pollicis longus, and flexor digitorum profundus (digits II/III), consistent with neuronal injury of at least 6 weeks’ duration.

Table 1.

Nerve Conduction Studies and Needle Electromyography.

Nerve studies	Initial study						Follow up—6 months later
Sensory nerves	Rec. site		Amp (µV)	Lat (ms)	Vel (m/s)			Rec. site	Amp (µV)	Lat (ms)	Vel (m/s)
R. Median Digit II	Wrist		NR	–	–		Wrist		NR	–	–
R. Ulnar Digit V	Wrist		7.2	1.77	56.5		Wrist		9.4	1.95	56.4
Motor nerves	Stim. site		Amp (mV)	Lat (ms)	Vel (m/s)		Stim. site		Amp (mV)	Lat (ms)	Vel (m/s)
R. Median—APB	Wrist		0.3	5.7	–		Wrist		1.1	5.8	–
	Elbow		0.3	11.7	33.4		Elbow		0.8	10.8	39.6
R. Ulnar—ADM	Wrist		7.8	2.4	–		Wrist		10.9	2.5	–
	Below Elbow		7.5	6.1	60.3		Below Elbow		9.9	6.7	52.4
	Above Elbow		7.1	6.8	59.1		Above Elbow		9.2	8.3	52.9
R. Median—FPL	Elbow		0.5	3.8	–		Elbow		3.5	4.2	–
L. Median—FPL	Elbow		5.6	2.9	–		Elbow		7.2	3.1	–
F Waves	Nerve		Min F lat (ms)	Max F lat (ms)	% F		Nerve		Min F lat (ms)	Max F lat (ms)	% F
ADM	R Ulnar		26.82	27.76	100		R Ulnar		27.85	28.75	100
ADM	L Ulnar		27.03	28.33	100		L Ulnar		28.40	29.15	100
Needle electromyography—initial visit only
Muscle tested	Fib	PSW	Fasc	Amp		Dur		Poly	Recruitment pattern
R. FPL	3+	N	N	Inc		Inc		2+	Single unit
R. APB	3+	N	N	Inc		Inc		3+	Single unit
L. FPL	N	N	N	N		N		N	Normal
R. FDP (II/III)	2+	2+	N	Inc		Inc		1+	Reduced

Nerve conduction studies demonstrated a severe axonal median neuropathy, with minor improvements on follow up study 6 months later. Needle electromyography demonstrated neurogenic changes consistent with the timing of her injury. Key: abnormalities highlighted in bold type face; R = Right, L = Left, Rec. site = recording site, Stim. Site = stimulation site, Amp = amplitude, Lat = latency, Vel = velocity, µV = microvolts, mV = millivolts, ms = milliseconds, m/s = meters per second, NR = no response, - = not calculated, APB = abductor pollicis brevis, ADM = abductor digiti minimi, FPL = flexor pollicis longus, FDP = flexor digitorum profundus, Max = maximum, Min = minimum Fib = fibrillations, PSW = Positive Sharp Waves, Fasc = fasciculations, Amp = amplitude, Dur = duration, Poly = polyphasia, N =normal or none.

Magnetic resonance neurography (MRN) was performed 7 weeks after her injury using a brachial plexus protocol on a Siemans 3 T MRI without contrast (Table 2). There was thickening and hyperintensity of the inferior trunk of the brachial plexus, extending into the medial cord and further distally in to both the median and ulnar nerves. Nerve roots were not involved. There was long segment hyperintensity of the median nerve extending to the distal forearm (Figure 1, panels A-B), with similar changes affecting the ulnar nerve up to the distal upper arm (Figure 1, panel C). Both nerves were intact along their course with no external compression or neuroma. Median-innervated forearm muscles demonstrated hyperintensity consistent with subacute denervation (Figure 1, panel D); there was mild denervation change in the ulnar-innervated muscle flexor carpi ulnaris.

Table 2.

Magnetic Resonance Neurography Protocol.

Sequence	Repetition time (ms)	Echo time (ms)	Slice thickness (mm)	Matrix
3D T2 Space STIR (coronal)	3500	196	2	320 × 320
3D T2 Space STIR (coronal)	3500	196	7	320 × 320
Proton density Dixon (axial)	2810	60	3	320 × 290
Proton density fat saturated turbo spin echo (transverse)	4300	60	3	312 × 384

The sequences and protocols used to acquire magnetic resonance neurography images in the present study. Abnormalities in nerve signal are best seen with fat saturated studies with echo time greater than 50 milliseconds to eliminate artifactual high signal in nerves. Dixon and STIR sequences are also utilized to minimize artifact and demonstrate nerve and muscle edema. Key: STIR = short tau inversion recovery; ms = milliseconds, mm = millimeters.

Figure 1.

Magnetic resonance neurography (MRN) of the right brachial plexus and peripheral nerves in the right arm performed 7 weeks after the initial injury. (A) Coronal plane of 3D acquisition T2 space short tau inversion recovery (STIR) sequence shows right brachial plexus inferior trunk, medial cord and right median nerve asymmetric thickening and high signal change extending through the axilla; (B) Coronal plane of 3D acquisition T2 space STIR sequence shows high signal in the shoulder consistent with joint effusion (long arrow) and right median nerve thickening and high signal change at the proximal humerus (short arrow); Axial proton density (PD) Dixon fat saturated sequence demonstrates high signal intensity and thickening of the median nerve at the upper humerus (C) and of the ulnar nerve at the mid/distal humerus (D); (E) Axial PD-Turbo Spin Echo (PD-TSE) sequence shows a thickened median nerve in the distal forearm (long arrow) with denervation of pronator quadratus muscle (short arrows).

In the absence of a compressive lesion, she was managed non-operatively, including with physical therapy. Follow up review 6 months later demonstrated improvement in strength in right abductor pollicis brevis (MRC grade 3), flexor pollicis longus (MRC grade 3), and flexor digitorum profundus supplying digit 2 (MRC grade 4). Repeat nerve conduction studies demonstrated improvement in the right median motor studies, specifically abductor pollicis brevis (originally 0.3 mV; follow up 1.1 mV) and flexor pollicis longus (originally 0.5 mV; follow up 3.5 mV) (Table 1). The right median sensory responses and F waves remained absent and the ulnar nerves remained normal. In view of her progress, it is anticipated that she will continue to improve, and in light of her comorbidities no additional intervention is currently planned.

Discussion

We report the case of a proximal median neuropathy following anterior shoulder dislocation, diagnosed by clinical and neurophysiological examination, and confirmed by magnetic resonance neurography (MRN). Clinical examination showed weakness of median-innervated muscles in both the hand and forearm, and neurophysiology was consistent with a severe axonal proximal median neuropathy. Radiographic assessment with MRN confirmed proximal median nerve involvement but also showed asymptomatic brachial plexus and ulnar nerve involvement.

Neural injury following shoulder dislocation occurs in 5 to 55% of patients and most commonly affects the axillary nerve due to its close association with the glenohumeral joint.^2,3 Typically, injury to the surrounding neural structures occurs as a result of traction and stretching of the nerve over the dislocated humeral head, and less frequently due to compression from fractured bone, hematoma, or resultant fibrous tissue as hematoma resolves.^4,7 In the present case, it is presumed that the injury was caused by traction, as the bony fragment was small in size and there was no evidence of current or previous hematoma. Further, the median and ulnar nerve involvement on MRN is likely to be caused by the same injury given the continuity of signal change beginning in the brachial plexus, and subacute muscle denervation changes in both median- and ulnar-innervated muscles.

Combinations of nerve lesions following shoulder dislocation are frequently observed. Visser et al⁴ in their series of 37 patients found an average of 1.8 nerves involved per patient as detected by EMG, most commonly affecting the axillary and musculocutaneous nerves in combination. It should therefore not be surprising in our case to see that more than one nerve has been injured, however it was unexpected to detect such extensive ulnar nerve involvement on MRN without associated clinical or neurophysiological abnormalities.

Shoulder dislocation leading to median and ulnar nerve injury is infrequently reported, possibly because of the relatively greater distance between the humerus and the course of these nerves.³ A clinically determined isolated median neuropathy, as in our case, is rare. This was shown to occur in just 1.9% of patients in 2 case series of 106 patients evaluated by clinical examination and neurophysiology, though not with imaging.^2,7 It is feasible that other asymptomatic nerve lesions would have been found in those cases if MRN was used. In fact, a review of MRN use in the evaluation of neuropathies found that additional diagnoses were made in 45% of patients.⁸

The use of MRN in the diagnosis of neuropathy could be considered when there is doubt as to the cause of the clinical presentation, or to determine the precise location and etiology of injury.^5,6 Typical findings of nerve injury include focal or diffuse enlargement of the nerve, hyperintensity on T2-weighted and fat saturated imaging, enlargement or disruption of fascicles, or discontinuity.⁹ In fat saturated imaging, echo time greater than 50 milliseconds is advised to eliminate the “magic angle phenomenon”; that is, artifactual high signal in nerves due to the orientation of the magnetic field. Further, advanced imaging including diffusion tensor imaging can provide information regarding nerve function.¹⁰ MRN also has the advantage of further confirming neural injury by detecting denervation changes within associated muscles, characterized by diffuse homogenously increased T2 signal intensity.¹⁰

The sensitivity of MRN in neural injury is very high at approximately 95%,^5,6 however specificity is considerably lower at approximately 30%.¹¹ Nevertheless, MRN has been shown to have a moderate or major impact on therapeutic decisions in 84% of patients with upper limb neuropathy.⁶ In the setting of neural injury following shoulder dislocation specifically, accurate diagnosis of the neural structures involved is important to guide operative management as this has been shown to result in favorable outcomes in the majority of patients.⁷ In the present case, non-operative management was favored owing to our patient’s age and comorbidities, and there is evidence that recovery can occur without intervention in up to 90% of patients within a period of 12 to 45 weeks.⁴

The mechanism that explains the findings on MRN is debated. Interestingly, the findings in our case of extensive signal change distal to the site of injury are similar to nerve ligation studies in rats, which resulted in axonal degeneration and endoneurial edema.¹² Our case also highlights that these structural changes are possible without clinically detectable functional disturbance.

Proximal median neuropathy as a clinical presentation is uncommon, encountered in a neurophysiology laboratory to a far lesser degree than median neuropathy at the wrist. In a retrospective series by Gross & Jones,¹³ proximal median neuropathy occurred in 0.2% of patients, as compared with 23% of patients diagnosed with median neuropathy at the wrist. Thus, patients with this condition may initially present for workup as a suspected case of carpal tunnel syndrome. To make this distinction, clinical and neurophysiological examination should include an assessment of proximal median-innervated muscles (pronator teres, flexor carpi radialis), and anterior interosseous-innervated muscles (pronator quadratus, flexor pollicis longus, the median half of flexor digitorum profundus). Proximal median neuropathy occurs most commonly in the context of anterior interosseous syndrome, pronator teres syndrome, entrapment at the ligament of Struthers, and direct injuries at the shoulder girdle such as following trauma.¹⁴ Our case is a reminder that anterior shoulder dislocation is also a potential cause.

The present case describes the uncommon clinical presentation of a proximal median neuropathy following anterior shoulder dislocation. The use of MRN in this case also detected asymptomatic ulnar nerve involvement, presumably caused by the same injury. In the setting of neural injury following anterior shoulder dislocation, magnetic resonance neurography is suggested to confirm the diagnosis and clarify the extent of nerve involvement.