The treatment of cysts at the spinoglenoid notch has varied from observation, needle aspiration, and open excision to arthroscopic decompression1-15. The results after both open excision1,2,4,9,10 and arthroscopic decompression1-9,12-15 have been satisfactory, and the latter avoids the morbidity of an open surgical procedure.
Several studies have shown a high prevalence of labral tears in connection with glenoid cysts1-15. These cysts are believed to develop as a ganglion when a labral or a capsular tear allows synovial fluid to be forced into the tissues, creating a one-way-valve effect1-15.
It is reasonable to assume that pain in this condition is caused both by cyst compression of the suprascapular nerve and by the labral tear. It is debatable whether these capsulolabral tears should be repaired as part of the procedure. Some authors have advocated labral débridement4,7,8 in combination with cyst decompression; others, repair of the labral lesion3,6,9,13-15; and a third group, use of both procedures, depending on the labral lesion1,2. One recent study described good results after labral repair without formal decompression in ten patients12.
On the basis of our previous experience with two patients, our protocol, since 1997, has been to treat all patients who had a labral tear and a cyst with débridement and labral repair alone, without decompression of the cyst. Our hypothesis was that labral repair alone would lead to cyst resolution and pain relief without the need of formal cyst decompression. The purpose of the present study was to assess clinically, and by magnetic resonance imaging, whether labral repair alone would lead to cyst resolution and pain relief.
Inclusion and Exclusion Criteria
Forty-nine consecutive patients with a ganglion cyst were treated with labral repair between January 1998 and January 2006. During the same period, 2529 shoulder arthroscopies were performed at our orthopaedic center. Most of the patients with a cyst were referred to us and had already had magnetic resonance imaging scans performed, but all other patients with a history and clinical symptoms suggesting a labral lesion and/or clinical signs of muscle atrophy had magnetic resonance imaging scans performed. The criteria for inclusion in the study were (1) a magnetic resonance imaging scan showing a spinoglenoid ganglion cyst, (2) pain in the posterosuperior aspect of the shoulder, (3) reduced or painful external rotation or abduction with or without atrophy of the infraspinatus and/or supraspinatus or teres minor muscles, and (4) an intraoperative finding of a labral lesion corresponding with the spinoglenoid cyst.
All such patients were included in the study, independent of whether they had undergone a subacromial decompression earlier. Patients with a history of another surgical procedure were to be excluded, but there were no such patients. We chose to include only patients with a cyst in the spinoglenoid notch. Thus, patients with a ganglion cyst at the anterior, anteroinferior, or posteroinferior aspect of the shoulder were excluded. The size of the cyst was not considered as an inclusion or exclusion criterion. There is no available information or evidence regarding what size of cyst affects the suprascapular nerve; thus, we chose to treat all patients with a cyst in a similar fashion. The indication for surgical repair was a labral tear with a spinoglenoid cyst, chronic posterior pain, and functional disability refractory to conservative care.
Seven patients were excluded from the study; two of them had only a small cyst at the border of the posteroinferior aspect of the glenoid rim. The labrum was avulsed, and these patients underwent labral fixation. Two patients had an anterosuperior cyst with a labral tear that was treated with fixation, and three patients had an anteroinferior avulsion of the labrum with cyst formation and atrophy of the teres minor and deltoid. The anteroinferior aspect of the labrum was stabilized. At the time of follow-up, all seven excluded patients had good or excellent functional results and the cyst had resolved. The remaining forty-two patients were included in the study. All patients agreed to undergo the procedure as described and gave their consent to participate in this study. Approval from our hospital review board was obtained.
Patients
All forty-two patients had a history of chronic pain in the posterior part of the shoulder that had been treated with medication and physical therapy for at least six months. No patient had atrophy of the supraspinatus muscle. Seven patients had clinical findings and/or a magnetic resonance imaging scan documenting atrophy of the infraspinatus muscle, one had atrophy of both the infraspinatus and the teres minor muscle, and one patient had isolated atrophy of the teres minor muscle. Those two patients both had a large multilocular cyst in the spinoglenoid notch, and no part of the cyst was extending to the posteroinferior aspect of the glenoid where it could affect the branch from the axillary nerve to the teres minor muscle. Four of the ten patients with pathological findings of the infraspinatus or teres minor muscle on magnetic resonance imaging had edema, no fatty changes, and moderate atrophy of the muscles, while six patients had minor to major fatty changes of the muscles in addition to the atrophy.
The median duration of symptoms prior to surgery was thirty-six months (range, two to 264 months). The median age at the time of the operation was forty-three years (range, twenty-three to sixty-eight years); there were thirty-five men and seven women. Nineteen patients had a history of shoulder trauma before the onset of symptoms and eleven patients were engaged in repetitive overhead activity, but we found no apparent explanation for the labral tear in twelve patients. In ten of the nineteen patients with a history of trauma, the injury was a fall on an outstretched arm, a motor vehicle accident, or direct trauma to the shoulder, and nine patients had experienced a forceful pull of the arm. For thirteen patients, the injury was related to sports. Twenty (48%) of forty-two patients had preoperative symptoms of supraspinatus tendinosis, and three patients had had a subacromial decompression performed previously without pain resolution. Twenty-one patients described night pain as one of their main complaints. Twelve patients were on sick leave because of shoulder problems, with a median leave of four months (range, three months to two years). Three patients had filed a Workers' Compensation claim.
The preoperative magnetic resonance image showed a labral tear at the posterosuperior quadrant in thirty-three (79%) of forty-two patients but revealed no labral tear in nine patients (21%). As a labral tear is the most common cause of cyst formation, it is our practice to proceed with arthroscopy in patients with symptomatic cysts, despite the absence of a labral tear on the imaging. The average cyst diameter was 2.4 cm (range, 0.5 to 7.0 cm). One patient with a small cyst measuring only 0.5 cm in diameter had infraspinatus atrophy on the preoperative magnetic resonance imaging scan. She had had relapsing pain for several years. Ten patients had a cyst diameter ranging from 1 to 1.5 cm, and one of them had atrophy of the infraspinatus muscle. Thirteen patients had a cyst diameter ranging from 2 to 2.5 cm, eight patients had a cyst diameter of 3 cm, and ten patients had a cyst diameter of 4 to 7 cm. The ten patients with muscle atrophy had an average cyst diameter of 3.3 cm (range, 0.5 cm to 7.0 cm). In two patients, magnetic resonance imaging was performed twice preoperatively because of a delay in treatment. In both patients the cyst had increased in size between the first and second image, and in one of them the cyst had excavated the posterior aspect of the glenoid neck. There was one other patient who demonstrated such excavation.
Surgical Technique
Arthroscopic surgery was performed by three experienced surgeons (G.U., E.G., and C.P.S.) with the patient under general anesthesia and in a lateral decubitus position with 4 to 5 kg of lateral traction applied. Two or three portals were used. The shoulder joint was examined, and the structures were palpated. Synovitis, cartilage damage, and/or partial tears of the rotator cuff tendons were registered. Starting anterior to the biceps, the labrum was examined and tested circumferentially. The posterosuperior aspect of the labrum was tested thoroughly. In a few shoulders, there was only a fissure of the posterior portion of the labrum. Our experience has been that although the labrum in these shoulders appears to be attached, we often find a thin synovial fold masking the tear. This is easily penetrated with the probe revealing an abnormal attachment of the labrum, much like the lesion of the posteroinferior aspect of the labrum described by Kim et al.16. All patients had a labral tear. When this was identified, a soft-tissue shaver was used to débride the underlying cartilage down to bleeding bone to promote healing. In a few shoulders, there was visible extrusion of gelatinous liquid during this débridement, but no attempt was made to evacuate the cyst directly.
The majority (twenty-eight) of the forty-two tears were treated with resorbable tacks. In twenty-eight patients, the operation was done with use of a lateral percutaneous pinning technique in which resorbable tacks (Suretac; Smith and Nephew, Andover, Massachusetts) were placed according to the size and position of the labral tear. In fourteen patients, resorbable suture anchors (Bio-FASTak; Arthrex, Naples, Florida) were placed in a similar fashion with spinal needle guidance if the tear was posterior to the biceps tendon. If the tear also involved the anterosuperior aspect of the labrum, a tack or anchor was placed through the anteromedial portal. A suture lasso (Arthrex) placed through the anterolateral portal was used to penetrate the superior portion of the labrum. If the labral tear was at the posterior corner or further down the posterior aspect of the glenoid, the arthroscope was switched to the anterior portal and the posterior portal became the working portal. One, two, or three tacks or anchors were used to stabilize the labrum.
A subacromial inspection was then performed, and if there were obvious degenerative changes of the supraspinatus tendon, an acromioplasty was performed. Such changes were found in six patients (14%), and they underwent an acromioplasty. As a group, these patients were older (median age, fifty years) and had a longer history (average, eighty-five months) of shoulder pain than the rest of the study group. All patients had an intra-articular injection of 20 mL of 0.5% Marcaine (bupivacaine) at the end of the procedure.
Postoperative Management
A sling was used for three weeks, and all patients were sent to a physical therapist for passive shoulder motion exercises within the first postoperative week. Active motion was allowed after three weeks except that active use of the biceps was not allowed before six weeks postoperatively. Strengthening exercises of the shoulder were not started until the patients had a normal or nearly normal range of motion, usually at six to twelve weeks postoperatively.
Functional Assessment
By protocol, the patients were prospectively evaluated at six weeks, three months, six months, twelve months, two years, five years, and at a final follow-up at seven years or more. An independent observer (O.S.) performed a clinical examination and calculated a Rowe score17 at two years and five years postoperatively and at the time of the final follow-up. This score is based on a total of 100 points. Five categories are evaluated: pain, stability, function, motion, and strength, depending on the aim of the surgical procedure. For instance, to evaluate the result after recurrent instability, pain would be given 15 points; stability, function, and motion, 25 points each; and strength, 10 points. We consider pain as the major complaint of patients with a cyst; thus, we chose the maximal number of points (25) for pain and 15 points for instability. The total scores were used to grade the overall results into the following categories: excellent (100 to 85 points); good (84 to 70 points); fair (69 to 50 points), and poor (=49 points). The patients also answered a questionnaire describing the subjective result of the treatment as excellent, good, fair, or poor. All results, except for those of the four patients who received additional surgery, represent the findings at the latest follow-up evaluation for each patient; the median duration of follow-up was forty-three months (range, fourteen months to nine years). Ten patients were followed between fourteen months and two years; nineteen patients, between two and five years; and thirteen patients, between five and nine years of follow-up.
Anatomical Assessment
Postoperative magnetic resonance imaging was planned by protocol for all patients, but no exact time was set for this study. Magnetic resonance imaging was performed at a median of fifteen months (range, three months to three years) postoperatively. As some of the studies were performed quite early in the postoperative period, we believed there was a need for a second study at a later stage. Thus, all patients would have magnetic resonance imaging done at least one year postoperatively. In addition, this would give us serial images with the possibility to detect eventual cyst recurrences in the patients with longer follow-up. The second imaging study was performed in connection with the latest clinical follow-up evaluation, at a median of forty-three months (range, fourteen months to nine years).
Follow-up magnetic resonance imaging was performed with either a 1.0-T scanner (Magnetom Expert; Siemens, Erlangen, Germany) or a 1.5-T scanner (Achieva; Philips, Eindhoven, The Netherlands). A dedicated shoulder coil was used. All patients underwent a standard imaging protocol with a coronal oblique STIR (short T1 inversion recovery) sequence, an axial turbo-spin-echo proton density-weighted fat-saturation sequence, and a sagittal oblique turbo-spin-echo T2-weighted sequence. All images were evaluated by one experienced radiologist (M.S.) who was blinded to the results of the primary imaging when reading the images from the first postoperative study. However, he was not blinded to these results when he read the images from the second study. The qualitative and the quantitative assessments of the infraspinatus and teres minor muscles were performed according to the method described by Fuchs et al.18. The qualitative assessment of fatty degeneration was done according to the method of Goutallier et al.19, with stage 0 indicating no fatty infiltration; stage 1, some fatty streaks; stage 2, less fat than muscle; stage 3, as much fat as muscle; and stage 4, more fat than muscle. The quantitative evaluation was done by measuring the cross-sectional areas of the rotator-cuff muscles on a picture archiving and communication system (PACS) console (Siemens), with the manufacturer's standard software.
Statistical Analysis
Measurements are expressed as the median with the range. Preoperative and postoperative differences were evaluated with the Wilcoxon signed-rank test. Categorical variables are expressed as frequencies. Results from the last follow-up evaluation before additional surgery were used if patients had a later second operation. The six patients who had an additional acromioplasty at the time of the primary operation were included, but an evaluation was also done with these patients excluded. Age, gender, duration of follow-up, and preoperative subacromial pain were evaluated in a multivariate regression model. The independent t test was used both to investigate whether patients with preoperative muscle atrophy had different cyst sizes than the patients without atrophy and to test whether the mean duration of preoperative symptoms was different in the patients with persistent muscle atrophy on magnetic resonance imaging compared with those without atrophy.
Intraoperative Arthroscopic Findings
Forty-one patients had a type-II SLAP (superior labral anterior-to-posterior) lesion, and one had a type-IV SLAP lesion according to the classification system of Snyder et al.20. Twenty-four of the patients with a type-II SLAP lesion had a subtype-B (posterior) lesion according to the system described by Morgan et al.21. Five patients had cartilage damage and osteoarthritis in an early stage. A partial tear of the supraspinatus tendon was observed in three patients.
Functional Results
The median Rowe score improved from 61.5 points (range, 54.3 to 68.5 points) preoperatively to 98.0 points (range, 70.0 to 100 points) at the time of follow-up (p < 0.001). The results were the same when the six patients who received an acromioplasty were excluded, and they were the same whether we used the scores before or after the reoperation in the four patients who had additional surgery. The final score did not appear to be related to age, gender, follow-up time, or preoperative subacromial pain. The median time for sick leave postoperatively was eleven weeks (range, two to forty weeks), and all patients except the three with a Workers' Compensation claim returned to regular work after their postoperative sick leave.
Anatomical Results (Cyst Resolution)
Magnetic resonance imaging revealed total resolution of the spinoglenoid cyst in thirty-seven patients (88%). In five patients, there was some remaining cyst structure, but the size was clearly smaller than it had been preoperatively (Table I). One additional patient initially had a 40-mm multilocular cyst located in the spinoglenoid notch with edema and moderate atrophy of the infraspinatus muscle. At the first postoperative imaging at three years, he had a 10 × 15-mm cyst at the posteroinferior aspect of the glenoid (not in the spinoglenoid notch) and no atrophy of the infraspinatus. The image made at the nine-year follow-up evaluation revealed a larger cyst at the same posteroinferior location, but again no cyst in the spinoglenoid notch. There was no atrophy of the infraspinatus, but there was moderate atrophy with stage-2 fatty infiltration of the teres minor muscle, and a posterior labral lesion was seen. One of the other five patients with a persistent cystic structure at the time of the final follow-up had a 6.7-mm-diameter cyst at the posterior aspect of the glenoid rim that appeared after the first postoperative image, the others had the cyst both at the first and the second follow-up (Table I).
The cyst was absent on the final magnetic resonance images in eight of the ten patients with preoperative atrophy of the infraspinatus and/or teres minor muscle. The four patients with a primary image demonstrating edema, moderate atrophy, and stage-0 fatty infiltration had no atrophy and stage-0 fatty infiltration at the time of follow-up. The six patients with stage-1 to stage-4 fatty infiltration preoperatively had unaltered stages of fatty infiltration at the time of follow-up (Figs. 1-A, 1-B, and 1-C and Table II).
With the limited numbers studied, cysts were not significantly larger (p = 0.08) in the patients with preoperative atrophy compared with those without atrophy (mean difference, 9.2; 95% confidence interval, —1.3 to 19.6), and there was no identifiable difference between the patients with or without persistent atrophy at the time of the last follow-up with respect to the mean duration of pain preoperatively.
Subjective Results
The five patients with a spinoglenoid cyst remaining postoperatively were pain-free and satisfied with the shoulder function. Five of the six patients with edema or stage-1 fatty infiltration with muscle atrophy preoperatively, and normal or stage-1 fatty infiltration on the postoperative magnetic resonance imaging scan rated the result as excellent, while the one patient described above with a cyst at the posteroinferior aspect of the glenoid and fatty infiltration of the teres minor described the result as good. Two of the three patients with fatty degeneration (stage 2, 3, or 4) of the infraspinatus muscle on both primary and follow-up magnetic resonance imaging rated the outcome in the shoulders as good, while one rated it as excellent. The patient with fatty infiltration (stage 4) of the teres minor muscle, both primarily and at the time of follow-up, rated the shoulder as excellent (Table II). All patients reported the absence of the chronic pain as the main benefit. Of the ten patients with atrophy, only two patients had subjective weakness in external rotation. These had both persistent fatty degeneration (stage 3 and 4) and atrophy of the infraspinatus.
All patients reported the absence of chronic posterior pain at the three-month follow-up examination, but some had other complaints, which are described below. Thirty-one (74%) of forty-two patients assessed the result as excellent; nine, as good; and two, as fair. Of the six patients who had an acromioplasty in addition to the labral repair, one rated the outcome as good and five rated the outcome as excellent. The categorical subjective ratings correlated to the Rowe scores (Spearman r = 0.67).
Complications and Reoperations
No patient had nerve damage or wound infection. Six patients had stiffness and pain, which were believed to represent an adhesive capsulitis, at the six or twelve-week follow-up evaluation. They received an intra-articular cortisone injection and had improvement with regard to pain and mobility. Four patients underwent a second operation. One patient had pain, and a persistent superior labral lesion was diagnosed and repaired. The patient with the type-IV SLAP lesion had persistent pain from the long head of the biceps and was treated with a biceps tenodesis. Two patients had persistent subacromial pain and underwent a subacromial decompression. One had relief of pain, and the other had no change. For these four patients, the Rowe score improved a median of 14 points after the second operation, but the classification according to both the Rowe score and the subjective evaluation of the patient was unaltered; two had a fair result, and two had a good result. In all four patients, the cyst had resolved and the chronic posterior pain was relieved.
The main finding of this study is that most spinoglenoid cysts resolve after labral fixation even without cyst decompression, with high patient satisfaction. Several authors have noted an association between spinoglenoid notch cysts and superior labral pathology, even when the labral disruption is not evident on imaging2-9. Our results are consistent with the assumption that ganglion cysts develop when a capsulolabral tear allows joint fluid to be forced into the surrounding tissues, creating a one-way valve, and that closing the valve by repairing the labrum is an effective method for cyst resolution and relief of symptoms. This is in agreement with the findings in the smaller study by Youm et al.12. Five patients had a cystic structure remaining in the spinoglenoid notch at the time of follow-up. The size was clearly reduced after surgery, and this may explain the absence of pain, but we cannot exclude that other factors may explain the clinical improvement.
We found in this study that when edema and muscle atrophy are seen on imaging studies preoperatively, they are likely to normalize, but fatty infiltration, regardless of stage, remains unaltered. Despite this finding, patients reported pain relief and satisfaction; however, those with fatty infiltration and atrophy of the infraspinatus had subjective external rotation weakness.
In agreement with Tung et al.13, we found that about half of the patients had symptoms of cuff tendinosis. In a few patients, subacromial degenerative erosions of the rotator cuff may represent a coexisting, separate condition. Six of the patients had an acromioplasty performed at the primary operation, which may, in addition to the labral stabilization, have contributed to the improvement in the subjective evaluation but not to cyst resolution.
The labral cyst size seems to be of importance as to whether the patient has development of muscle atrophy. Tung et al.13 reported that cysts of =3.1 cm in diameter were associated with signs of muscle denervation. In our study, there was a tendency for the cysts to be larger in patients with atrophy; however, this finding was not significant (p = 0.08), and the confidence interval for the difference in size was wide. Thus, we cannot exclude that the failure to show a difference between the groups was due to inadequate sampling. The two patients with small cysts of 0.5 to 1.5 cm in diameter, who had atrophy of the infraspinatus muscle, suggest that cyst size is not the only determining factor. This is in agreement with Fritz et al.22 and Ogino et al.23 who found that cysts with a diameter of 1.0 cm caused nerve compression and muscle atrophy.
There is limited knowledge about the natural course of asymptomatic cysts. We do not know whether they will expand and eventually cause symptoms. There are reports of cysts that have spontaneously resolved1,2,9,22. The cysts actually expanded in the two patients in our study who had two preoperative magnetic resonance images performed a year apart. In agreement with Tirman et al.5, we observed that expanding cysts may cause excavation of the posterior portion of the glenoid neck. The possibility of cyst expansion supports an approach of repairing symptomatic labral tears with associated cysts even when the cyst is not believed to compress the nerve.
Tung et al. reported that six of forty-six shoulders with spinoglenoid notch cysts had infraspinatus atrophy and one had teres minor atrophy13. They postulated that variant innervations of the teres minor muscle by a branch of the suprascapular nerve explained the teres minor atrophy. The finding of isolated atrophy of the teres minor muscle in two patients and atrophy of both the infraspinatus and teres minor in one patient in the present study supports this view. Such variant innervation of the teres minor muscle has been described24. That atlas also noted that, in some cases, the two muscles form one anatomical unit.
Another possible explanation for isolated atrophy of the teres minor is compression of the posterior branch of the axillary nerve. The anatomic study by Ball et al.25 described the course of this branch. It runs posteriorly, adjacent to the inferior aspect of the glenoid rim for an average distance of 10 mm (range, 2 to 17 mm) before dividing into the superior-lateral brachial cutaneus nerve and the nerve to the teres minor muscle. The motor branch courses medially along the posterior aspect of the inferior part of the glenoid rim for an average distance of 18 mm (range, 11 to 25 mm). Thus, a cyst located at the posteroinferior part of the glenoid, or a large spinoglenoid notch cyst that expands downward sufficiently, may compress the nerve to the teres minor. This could explain how the atrophy of the teres minor muscle developed without affecting the innervation of the deltoid muscle. The three patients in our study with an affected teres minor muscle had a cyst located in the spinoglenoid notch, and the cyst did not expand to the posteroinferior part of the glenoid.
Labral cysts were infrequently described prior to the widespread use of magnetic resonance imaging for the evaluation of shoulder abnormalities. We agree with Moore et al.2 that shoulder pain is the dominant clinical problem associated with labral cysts both for patients with and for those without atrophy. The sensory branch of the suprascapular nerve innervates the posterior part of the capsule, and, when the nerve is compressed, it may elicit pain sensation in the posterior aspect of the shoulder, as entrapment of the nerve anywhere along its course is associated with deep, poorly localized shoulder pain22.
There were no serious complications in these series, but the observed complications have to be considered when labral repair is recommended to patients with shoulder pain and a spinoglenoid cyst. The four patients who required additional surgery experienced some improvement, which was sufficient to justify the reoperation.
The advantage of the present study is that both clinical and imaging findings were assessed prospectively by an independent observer in a relatively large number of patients. A disadvantage may be the shorter duration of follow-up of ten patients. As the radiologist was not blinded to the results from the first postoperative imaging study when reading the images from the second, we cannot exclude that this may have biased the findings. The fact that only one cyst recurred between the time of the first and second magnetic resonance imaging studies of the thirty-two patients who were followed from two to nine years postoperatively supports the assumption that if the cyst is resolved at short-term follow-up, this result is likely to persist at a longer-term follow-up. Another weakness of the study is the fact that a concomitant subacromial decompression was performed in six patients. The statistical evaluation showed similar differences in Rowe scores when these patients were excluded, and, in our opinion, this justifies their inclusion.
The present study was not randomized to compare the effectiveness of expectant treatment with that of traditional cyst decompression and labral repair. Other studies have not indicated that cysts of the size and location of those treated in this study will resolve spontaneously1,2,5,10. It therefore may be unethical not to treat these patients, and consequently a future multicenter randomized study probably should compare labral repair with and without cyst decompression.
In the present study, we found that most spinoglenoid cysts resolve, as verified by magnetic resonance imaging, after labral repair alone. Our results suggest that labral repair is a safe and effective procedure with high patient satisfaction. Functional and subjective results and imaging observations following labral repair alone are comparable with those in other reports of cyst decompression combined with labral repair3,6,9,13-15. This approach may provide a way to obviate the potential risk of endangering the suprascapular nerve during cyst decompression. We recommend that patients with labral tears and associated cysts be treated with labral repair only, to prevent cyst enlargement and to avoid the risk of suprascapular nerve compression. 