Keywords suprascapular neuropathy - rotator cuff tear - shoulder pain - suprascapular nerve
- supraspinatus - double hit phenomenon - double crush syndrome
Background
Several studies have shown that suprascapular neuropathy can occur in a significant
percentage of patients with large-to-massive rotator cuff tears,[1 ]
[2 ]
[3 ]
[4 ]
[5 ] but is a rare complication following primary rotator cuff repairs even when tears
are massive.[6 ]
[7 ] In a study of 26 patients with massive cuff tears who were evaluated with preoperative
electrodiagnostic studies, 7 (38%) had isolated suprascapular injury.[3 ] In six of these patients, the neuropathy improved after repair (the seventh was
irreparable). Costouros et al[3 ] speculated that improvement occurred because the repair untethered the suprascapular
nerve (SSN) from being compressed at the base of the scapular spine ([Fig. 1 ]). When improvement does not occur, or when postoperative neuropathy occurs but is
not detected preoperatively (as in this report), the differential diagnosis is broadened
to include (1) direct iatrogenic injury, (2) cervical radiculopathy, (3) brachial
plexitis, (4) compression by an extrinsic mass such as a hematoma or ganglion cyst
not detected preoperatively, and (5) indirect iatrogenic injury secondary to cuff-repair–related
traction of the SSN that was constricted in the suprascapular notch.[9 ] This latter possibility appears to have been an important factor in our unusual
case, where repair of the rotator cuff exacerbated a preexisting, but preoperatively
undetected, SSN dysfunction.
Fig. 1 Medial retraction of the suprascapular nerve (SSN) caused by a large-to-massive rotator
cuff tear. Illustration of how the SSN can experience traction by being pulled medially
at the base of the scapular spine following medial and inferior retraction of a large-to-massive
tear. In some cases, repairing these tears can un-tether the nerve by mobilizing it
laterally. Not only has this explanation been contested,[8 ] but it is also not applicable in our patient because his cuff tear was small and
significant suprascapular neuropathy occurred after the cuff repair. Reproduced from
Costouros et al[3 ] with permission of Elsevier Limited.
Case Presentation
A healthy 62-year-old (height: 1.83 m; weight: 98.4 kg; body mass index: 22 kg/m2 ) right-hand–dominant man, who was an accountant and avid golfer, was referred to
our clinic with a chief complaint of left shoulder pain. There was no history of shoulder
or neck trauma. Thirteen years ago, an arthroscopic acromioplasty was performed on
his left shoulder, and 6 months prior to this first clinic visit a revision arthroscopic
acromioplasty with cuff tendon debridement was performed for recurrent pain, but did
not help. His diagnosis had been stage-two subacromial impingement syndrome.[10 ] The pain was achy and occasionally sharp and was distributed over the anterior and
superior-posterior aspects of the shoulder. There was a mildly positive left-side
Spurling sign, but there was no winging of the scapula, and liftoff and glenohumeral
instability tests were negative. Active range of motion was 150-degree flexion and
135-degree abduction, and manual muscle tests showed normal strength. Pain was moderate
with impingement (Neer) and impingement reinforcement (Hawkins-Kennedy) maneuvers.
His initial American Society of Shoulder and Elbow Surgeons (ASES) score was 64 (best = 100)
and other measures of general health and shoulder function are listed in [Table 1 ].[11 ]
[12 ]
[13 ]
[14 ]
[15 ]
[16 ]
Table 1
The patient's shoulder function and general health
Six-year outcome data
First clinic visit
4 mo after RCR
SSN release
Months after SSN release
3
6
12
18
22
34
39
45
51
Active forward flexion
150°
70°
90°
105°
107°
105°
100°
120°
130°
130°
130°
10 cm VAS score on typical day (best equals 0)
2.3
–
–
3.9
2.9
3.3
4.4
1.7
1.2
1.5
2.7
ASES score (best equals 0)
64
–
–
51
59
54
45
73
–
85
77
WORC score (best equals 0 (100%))
–
1,118 (47%)
–
1,222 (42%)
926 (51%)
1,497 (29%)
672 (68%)
634 (70%)
564 (73%)
418 (80%)
Simple shoulder[a ] (No. of yes responses)
–
–
8/12
–
7/12
10/12
5/12
10/12
10/12
12/12
12/12
DASH score (best equals 0)
–
–
32.5
–
42.5
31.0
62.1
26.7
15.0
15.8
12.9
Short Forrm-36[b ]
Physical functioning
–
–
50
–
15
65
60
70
70
70
90
Role limitation due to physical health
–
–
0
–
0
0
0
0
25
25
100
Role limitation due to emotional problem
–
–
0
–
0
0
0
33.3
100
100
100
Energy/Fatigue
–
–
60
–
20
40
40
45
50
60
65
Emotional well-being
–
–
44
–
56
84
56
72
80
92
88
Social functioning
–
–
62.5
–
62.5
50
50
75
100
100
100
Pain
–
–
67.5
–
22.5
55
22.5
55
77.5
67.5
80
General health
–
–
45
–
30
85
75
65
70
62.5
75
Abbreviations: ASES, American Society of Shoulder and Elbow Surgeons; DASH, disabilities
of arm, shoulder, and hand; deg., degrees; RCR, rotator cuff repair; SSN, suprascapular
nerve; VAS, visual analog scale; WORC, Western Ontario rotator cuff.
a Number of yes responses/number of questions (“yes” responses correlate with better
function than “no” responses).
b All questions are scored from 0 to 100 representing the highest level of function
possible. Aggregate scores are compiled as a percentage of the total points, using
the RAND scoring table.
Images from a magnetic resonance (MR) scan obtained 7 months prior are shown in [Fig. 2A, B ]. An MR scan performed 6 weeks prior to his first visit showed a small tear of the
supraspinatus tendon with mild muscle atrophy ([Fig. 2C–E ]). Both scans showed mild fatty degeneration of the infraspinatus ([Fig. 2B, C ]). A subsequent neck MR scan showed disk bulges/herniations at C2–3, C3–4, C4–5,
and C5–6. A neurosurgeon then performed two-level disk fusions (C4–6), which relieved
the neck pain but only mildly reduced the shoulder pain. Three months later (December
2006), the senior author (J. G. S.) performed arthroscopic debridement of anterior
labral fraying and open repair of a small supraspinatus tear. The tear was retracted
15 mm and was easily mobilized without requiring tissue release beyond the glenoid
labrum. Two suture anchors were used to affect a quasi-double-row repair.[19 ] Active motion was not allowed for 8 weeks.
Fig. 2 MR images prior to rotator cuff repair. MR images 16 (A –C ) and 10 (D, E ) months prior to cuff repair; asterisk indicates the coracoid. (A, B ) Evidence of scarring in the supraspinatus (white arrows) 16 months preoperation.
(G = glenoid). (C ) Mild fatty degeneration (short arrow)[17 ]
[18 ] of the infraspinatus. Scarring in the supraspinatus (longer arrow). (D ) Scar tissue is seen in (A ) and (B ) (arrow), and mild fatty degeneration[17 ]
[18 ] of the supraspinatus is seen 10 months preoperation. (E ) Small supraspinatus tear (arrow).
At 10 weeks postsurgery, there was substantial left shoulder weakness with attempts
at active motion. Although the deltoid contracted well, there was (1) atrophy in the
infraspinatus muscle, and strength of 3/5 in external rotation and 3 −/5 in elevation
in the plane of the scapula, as well as (2) mildly positive left-side Spurling sign.
The differential diagnosis included iatrogenic injury, extrinsic compression (e.g.,
hematoma, cyst, etc.), brachial plexopathy, and/or concomitant cervical nerve-root
compression.[20 ]
[21 ] A nerve conduction study and electromyography (NCS/EMG) showed findings consistent
with SSN injury at the level of the suprascapular notch with absence of evidence of
re-innervation. MR scans obtained of the left brachial plexus and shoulder showed
humeral head elevation suggesting weakness of the superior rotator cuff ([Fig. 3 ]) but no evidence of (1) a space-occupying lesion along the SSN, (2) injury of the
brachial plexus and its nerve branches,[22 ]
[23 ]
[24 ] or (3) unusual morphology of the suprascapular notch or transverse scapular ligament
(retrospective examination showed that these scans were not sufficient to detect an
anomaly of the notch).
Fig. 3 MR images 4 months after rotator cuff repair. (A ) Elevation of the humeral head (white bracket) and signal changes, and a possible
defect in the supraspinatus insertion (arrow). (B ) Moderate supraspinatus fatty degeneration (arrow). Asterisk indicates the coracoid.
At 105 days after shoulder surgery, the patient saw his neurosurgeon for a 7-month
follow-up and the cervical fusion had healed. At 138 days after shoulder surgery,
there was increased burning and aching that radiated across the posterior and lateral
aspects of the shoulder and upper arm. Voluntary infraspinatus contractions were now
more obvious. Active shoulder flexion and abduction were < 80 degrees each. The changing
pain intensity and characteristics were presumed to reflect SSN recovery for which
Lyrica (pregabalin) 75 mg twice/day was prescribed. At 162 days after surgery, the
neuropathic pain was significantly decreased and usually did not require medication;
however, there was no improvement in strength in flexion or external rotation.
At 213 days after cuff repair, results of another NCS/EMG revealed findings that could
not be explained by an isolated suprascapular neuropathy: (1) denervation potentials
in deltoid, rhomboids, and C5 cervical paraspinal muscles, as well as in the supraspinatus,
suggesting C5 radiculopathy; (2) small amplitude fibrillation potentials in the abnormal
muscles, and complex repetitive discharges in the deltoid indicating chronic neuropathy;
and (3) evidence of re-innervation in the abnormal muscles. There was also no evidence
of conduction block along the suprascapular and axillary nerves. A diagnostic/therapeutic
suprascapular notch injection with local anesthetic and a corticosteroid was recommended.
However, the patient's pain (but not strength) improved, causing him to cancel the
injection. In view of evidence of early nerve recovery, and in the perspective that
intraoperative trauma had somehow occurred to the SSN, a consulting surgeon specializing
in peripheral neuropathies recommended to continue nonoperative management. This recommendation
reflected the general conclusions of Antoniou et al[25 ] who, in their analysis of the functional outcome of 23 patients, suggested that
traumatic lesions of the SSN can have equivalent responses to operative and nonoperative
treatment. (But, as discussed later, it was ultimately concluded that prompt scapular
notch decompression would have been best for our patient.)
By 322 days after surgery, the pain recurred with greater intensity. A suprascapular
notch injection with 0.25% bupivacaine was done using computed tomography (CT) guidance
and this yielded excellent, but temporary, improvement. An open surgical decompression
of the suprascapular notch was performed (nearly 1 year after the cuff repair, [Table 1 ]) by the same consulting surgeon. The nerve was grossly compressed beneath a hypertrophic
ligament in a narrowed suprascapular notch ([Fig. 4 ])[26 ] and it dramatically protruded upward after ligament transection. The first motor
branch to the supraspinatus was distal to the notch[28 ] and the supraspinatus motor branches were intact. There was no evidence of neuroma
or other gross neural, perineural, or vascular pathology, variation, or anomaly, and
no compression caused by the deep fascia proximal or distal to the suprascapular notch
or by the tissues forming the spinoglenoid notch.[26 ]
[29 ] Consequently, further surgical exploration[26 ]
[30 ] was not necessary. The transverse scapular ligament was excised and the notch was
enlarged.[25 ]
[31 ] During the following 6 months, the burning pain improved significantly but the weakness
persisted.
Fig. 4 Types of suprascapular notch morphologies. These original drawings are based on the
photographs in Rengachary et al.[27 ] Type V (6% incidence) was found in our patient at the time of notch decompression.
At 22 months after notch decompression, the pain and function had worsened. A second
postoperative MR scan showed increased fatty degeneration of the supraspinatus and
infraspinatus ([Fig. 5 ]). By contrast, substantial overall improvement occurred at 27 months—the pain had
almost completely subsided and strength significantly improved. In fact, he was able
to remodel his home (manual labor for several months) without significant discomfort.
Active forward flexion was 120 degrees, but early fatigue with overhead motion was
reported. At final follow-up at 4.5 years after notch decompression, he had reached
a plateau in motion and strength (4 + /5) and the pain was minimal/tolerable except
when golfing. An axillary lateral radiograph showed evidence of glenohumeral arthritis,
which likely contributed to this activity related pain.
Fig. 5 MR image obtained 34 months after rotator cuff repair and 22 months after suprascapular
notch decompression showed worsening of fatty degeneration of the supraspinatus (short
arrow) and infraspinatus (long arrow). The coracoid is indicated with an asterisk.
Conclusion
Among the various potential factors contributing to our patient's shoulder dysfunction,
what is clear is that the SSN was constricted in a narrowed suprascapular notch. Because
there are no data suggesting that a narrowed notch and hypertrophic transverse scapular
ligament result from indirect iatrogenic injury during routine open repair of a small
cuff tear,[32 ] we conclude that the SSN was compressed in the notch prior to the repair. What likely
occurred was that first the preexisting narrowed notch and hypertrophied superior
transverse ligament made the nerve prone to injury even after typically inconsequential
minor perturbation (e.g., nerve and other tissue motion during the rotator cuff repair).[28 ] Second, a vicious cycle ensued that produced focal swelling that interfered with
axonal transport and caused vascular compromise, ultimately leading to more pronounced
edema along the nerve with resulting distal muscle dysfunction. The anatomical bottleneck
then became even more constrictive pronouncing the pathology, which in our patient's
case resulted in a functionally (but not morphologically) complete lesion. While timely
decompression of such a lesion can frequently produce good results, the delay in our
patient's surgical decompression likely contributed to the prolonged delay in achieving
a good result.[25 ]
[33 ]
A significant amount of our patient's preoperative pain (i.e., before the cuff repair)
was likely sensory dysfunction[34 ]
[35 ] from notch constriction—consistent with descriptions of “tunnel syndrome of the
SSN” or “carpal tunnel syndrome of the shoulder.”[26 ]
[36 ] Suprascapular neuropathy has been postulated to occur from excess motion of the
scapula because the nerve is tethered to at least two, and often three, anatomic locations:
the suprascapular notch, the spinoglenoid notch, and Erb point in the neck.[27 ]
[37 ]
[38 ] However, observations by Rengachary et al[27 ]
[38 ] in fresh cadavers revealed no obvious movement of the SSN across the suprascapular
notch with extreme shoulder motion, reflecting the fact that the neurovascular pedicle
is fixed to the periosteum in the notch and to the supraspinous fossa. They speculated
that SSN injury more likely results from kinking against the ligament (the “sling
effect”) because of its close apposition to its sharp inferior border with shoulder
depression, retraction, and hyperabduction. This “sling effect” might become even
more pronounced when the SSN is also constricted within a stenotic notch like that
in our patient.
Although a peripheral nerve can tolerate approximately 10% increase in its resting
length before neurapraxia occurs,[39 ] altered conduction can occur when a nerve is stretched beyond 6% of its resting
length.[40 ] Scarring of the neurovascular pedicle associated with some cuff tears can reduce
the distance that the nerve can be stretched without injury.[9 ] Support for the hypothesis that this is what happened in our patient includes: (1)
his avid recreational activities (golf) causing repetitive trauma, potentially increasing
constriction in his stenotic suprascapular notch,[33 ]
[41 ]
[42 ] and (2) the scar-like tissue in the “intramuscular tendinous core” of the anterior
supraspinatus[43 ] ([Fig. 2 ]) can reduce the ability to safely mobilize the tendon during repair.[44 ]
[45 ]
[46 ] These factors may have contributed to SSN injury in our patient even though tendon
mobilization was within the safe range in normal cases.[28 ] The C5 axonal pathology is also potentially important because the SSN contains fibers
primarily from C5 and C6. If the C5 radiculopathy caused impairment of SSN axonal
transport, then it might have been more prone to neuropathy from peripheral constriction.
This phenomenon is sometimes referred to as a “double crush syndrome” or the “double
hit phenomenon.”[4 ]
[47 ]
[48 ]
[49 ]
Because there were periods of improvement during his early postoperative course, a
prolonged period of recovery was allowed. This opinion was based on the presumption
that the SSN sustained intraoperative trauma and hence this opinion was consistent
with the general conclusions of Antoniou et al.[25 ] The SSN lesions in their 23 patients that were caused by trauma (including traction
and direction closed injuries) showed no statistical difference in the response to
operative (n = 12) and nonoperative (n = 11) treatment. But the p -value for this difference was not provided, leaving the reader with the impression
that, based on the subsample means provided, the sample sizes were insufficient to
discern clinically and statistically significant differences when all of these issues
are considered: (1) trauma mechanism, (2) depth of nerve injury, (3) lesion level,
(4) intraoperative electrophysiological evaluation result, and (5) microsurgical repair
method chosen, which can have a significant effect on results obtainable (e.g., external
neurolysis and/or nerve mobilization, vs. graft repair vs. end-to-end repair after
neuroma excision). However, Antoniou et al[25 ] did show that, in contrast to “traumatic lesions,” “compressive lesions” attributable
to suprascapular notch entrapment (which our patient had) had the best outcome with
surgical decompression. Therefore, prompt scapular notch decompression would have
been the best choice for our patient. This is supported by additional reports showing
that when transverse scapular ligament release is done relatively soon after the diagnosis
is made that milder muscle atrophy is more likely to subside along with shoulder pain.[50 ] But when ligament release was delayed and there was advanced atrophy, the pain improved
but not the atrophy.
Recent studies show that a large percentage of massive rotator cuff tears with significant
muscle atrophy can have significant clinical improvement after surgical repair without
SSN decompression.[51 ] This is one reason why current guidelines for the preoperative workup of most rotator
cuff tears do not include electrodiagnostic testing.[4 ]
[5 ]
[8 ]
[9 ]
[32 ] But when functional deficits that do not resolve, and there are signs of a possible
nerve lesion, easy and cost-effective electrodiagnostic studies as well as nerve imaging
should be included early in the differential diagnostic workup. SSN impairment can
be diagnosed quite well electrophysiologically by NCS/EMG, and well differentiated
from C5 injury. In conjunction with MR imaging of supraspinatus/infraspinatus muscle
wasting and high signal due to denervation edema, a concomitant nerve lesion can be
readily diagnosed. If this is the case, then exploration and macro-/microsurgical
decompression or nerve repair should follow promptly. In less clear cases, neuroimaging
in the form of MR neurography of the notch area (thin slice, 1 mm, in high resolution)
or neurosonography, or both, can be used to enhance the ability to define the lesion.[52 ]
[53 ] This includes the condition of the muscles (including wasting and fatty degeneration)
and the nerve can be traced from its takeoff from the upper trunk to the notch area
and beyond to reveal transection, neuroma, constriction, or edema.
In summary, our patient had an unusual postoperative course because of the exacerbation
of preexisting entrapment of the SSN in the suprascapular notch. Three years was required
to achieve a good result following suprascapular notch decompression, which was done
nearly 1 year after the diagnosis was clearly established. There should be heightened
awareness of this problem in patients who do not have satisfactory improvement in
shoulder pain from previous shoulder and neck surgery, and this should prompt more
immediate surgical decompression in some cases. The underlying C5 radiculopathy that
our patient had may have created a “double crush syndrome” that contributed to the
propensity for injury and prolonged recovery.
