Management of glenohumeral arthritis in young active patients is controversial and very challenging1-6. Several methods had been proposed for treatment of symptomatic arthritis, but most have not been supported by a convincing evidence of success7-22. Burkhead and Hutton15 proposed biological resurfacing of the glenoid with the interposition of soft tissue as a means of improving the outcome of hemiarthroplasty in young, active patients. They used articular capsule; autogenous fascia lata graft; and, recently, Achilles tendon allograft. Their midterm follow-up demonstrated a failure rate of 20%, and a longer follow-up showed a similar failure rate16. They concluded that biological resurfacing provided a bearing surface that improved the outcome compared with that achieved with hemiarthroplasty alone.
The purpose of this study was to evaluate our results with soft-tissue resurfacing of the glenoid in the management of young patients with glenohumeral arthritis as we were concerned that this type of procedure was leading to poor outcomes.
The study was approved by the institutional review board of the Massachusetts General Hospital. Between 2000 and 2006, thirteen (3%) of 492 shoulder arthroplasties performed by the senior author (J.J.P.W.) were in very active young patients who underwent soft-tissue resurfacing of the glenoid. These patients satisfied treatment with soft-tissue resurfacing of the glenoid not only because they were less than fifty years of age, but also because total shoulder replacement was deemed to be not optimal for them because of their high activity level, which was defined as manual labor or sports activities resulting in substantial loading. Such sports activities included volleyball, baseball, basketball, football, hockey, and heavy lifting. Golf and basic swimming were not considered to be high-level sports activities. All patients had had a complete preoperative evaluation, had complete operative records, and had been followed clinically for a minimum of two years or until revision surgery was done because the procedure had failed. The duration of follow-up averaged forty-eight months (range, six to 102 months). All patients with less than two years of follow-up had required revision surgery. There were nine men and four women, with an average age of thirty-four years (range, eighteen to forty-nine years). The dominant extremity was involved in ten patients. Three patients had not had previous surgery. The remaining ten patients had undergone a total of fifteen previous shoulder operations (Table I).
The indication for soft-tissue glenoid resurfacing was the presence of arthritis in an active patient who was less than fifty years of age, who expected to return to sports or work activity, and in whom conservative treatment including injections, physical therapy, and modification of aggravating activities had failed. Exclusion criteria included major glenoid osseous deficiency, advanced rheumatoid arthritis, a prior shoulder replacement, and ongoing infection. Patients who were included in this study had preoperative evidence of glenohumeral joint-space narrowing and sclerosis on radiographs and/or computed tomography with or without posterior glenoid erosion and an intraoperative finding of full-thickness wear of the glenoid articular cartilage. Soft-tissue resurfacing was combined with use of a stemless resurfacing implant or a hemiarthroplasty, depending on humeral head and glenoid congruity. Soft-tissue resurfacing of the glenoid was performed with use of an Achilles tendon allograft in eleven patients, the anterior shoulder joint capsule in one, and a fascia lata autograft in one.
Surgical Technique
The surgery was routinely performed with the patient positioned in a beach-chair device (T-Max Beach Chair; TENET Medical Engineering, Calgary, Alberta, Canada) and with use of an articulated, hydraulic arm-holder (SPIDER Limb Positioner; TENET Medical Engineering) to support and position the arm during the surgery. Regional interscalene anesthesia was used along with light general anesthesia. An extended deltopectoral incision was made, and a layer-by-layer tissue-plane dissection was performed, with release of adhesions from prior surgery when necessary. In all cases, the biceps tendon was dissected out of the bicipital groove and tenodesed into the superior border of the pectoralis major tendon. The lesser tuberosity was then osteotomized unless this was not possible because the subscapularis anatomy was distorted from prior surgery. In all cases, the axillary nerve was identified, mobilized, and protected with a retractor in order to permit safe release of adhesions about the subscapularis muscle and to have free and complete access to the glenoid surface.
The humeral head was replaced by an anatomical prosthesis (Zimmer, Warsaw, Indiana), which was designed to restore the three-dimensional geometry of the proximal part of the humerus as closely as possible to normal, in ten patients because they had flattening of the humeral head and glenoid erosion that required reaming to create a concave, uniform surface. A hemiarthroplasty was preferred in this group to allow better glenoid exposure. In the remaining three patients, in whom the humeral head was flattened but there was no substantial eccentric glenoid erosion, the humerus was resurfaced with a device that permitted reconstruction of the proximal part of the humerus without the use of a stem (Global CAP; DePuy, Warsaw, Indiana) (Table II). If substantial posterior erosion was present, the glenoid was reamed with a reamer from the shoulder arthroplasty instrumentation set (Anatomical Shoulder System, Zimmer) in order to make its surface more congruent with the humeral head. Occasionally, multiple drill-holes were placed in the glenoid surface if the bone was found to be very sclerotic.
Closure of the subscapularis was performed with use of transosseous sutures placed around the stem of the humeral component23.
Soft-Tissue Resurfacing of the Glenoid with Interposition Capsular Arthroplasty
This technique was similar to that described by Burkhead and Hutton15. In a patient without severe capsular scarring or disruption from prior surgery, the anterior aspect of the capsule was dissected free from the deep surface of the subscapularis and then fixed over the glenoid surface with insertion of anchors into the glenoid and sutures into the remaining circumferential labrum. All anchors were absorbable and were placed in the center and at each quadrant of the glenoid. This configuration was chosen to make sure that the graft was securely fixed onto the prepared glenoid surface.
Before the capsule was fixed over the glenoid, the osseous surface was abraded with a high-speed burr until bleeding bone was encountered.
Soft-Tissue Resurfacing of the Glenoid with Autogenous Fascia Lata Graft
A free autogenous fascia lata graft was obtained from the ipsilateral thigh. The typical size of the graft was 4 by 8 cm and it was folded over itself and sewn into a patch that would cover the glenoid surface. The defect in the fascia lata was closed with interrupted, absorbable sutures. The graft was placed with use of the method described above.
Soft-Tissue Resurfacing of the Glenoid with Achilles Tendon Allograft
Allograft tissue was chosen when the anterior aspect of the capsule was either too scarred and short or otherwise insufficient. Moreover, it was our impression that Achilles tendon allograft provided a thicker and more robust soft-tissue interposition and thus was preferable to autograft in most cases.
The same method of exposure was used as described above. The Achilles tendon allograft was made two layers thick and sewn into a shape that would cover the glenoid. It was then fixed with use of the methods described above. Multiple nonabsorbable, strong sutures (number-2 FiberWire; Arthrex, Naples, Florida) were placed around the perimeter of the glenoid in the remaining labrum, and four bioabsorbable anchors (PANALOK 3.5-mm anchors; Mitek, a Johnson and Johnson Company, Westwood, Massachusetts) were used in the glenoid as well. The graft was then secured with use of these sutures after preparation of the glenoid surface.
Postoperative Management
The arm was kept in a shoulder immobilizer for four weeks. Passive range-of-motion exercises were commenced immediately, but within the limits imposed by the soft-tissue repair of the subscapularis, as determined at the time of surgery. Active motion for activities of daily living as well as active-assisted range-of-motion exercises were permitted after four weeks. Strengthening exercises were delayed until twelve weeks after the surgery, and recreational sports or lifting activities were delayed for a minimum of sixteen weeks after the surgery.
Evaluation
The results were graded with use of a visual analog pain scale, the subjective shoulder value, and the Constant and Murley score24. The patients used the visual analog pain scale to indicate their subjective feeling of pain, with 0 indicating no pain and 10 indicating the worst pain that they had ever felt. The subjective shoulder value represents a patient's subjective assessment of shoulder function expressed as a percentage of an entirely normal shoulder, which would score 100%25. All patients had preoperative and postoperative radiographs (three views) and computed tomography scans, and the axillary radiographs as well as the coronal and sagittal computed tomography scan cuts were reviewed by one of us (M.O.) to determine the degree of glenohumeral joint-space narrowing and glenoid erosion.
The Wilcoxon signed-rank test was used for statistical analysis of the outcomes. A p value of 0.05 was set as the level of significance.
Source of Funding
There was no external funding source for this study.
We found no significant improvement, with the small number of patients analyzed, between the preoperative and follow-up visual analog pain scale scores, subjective shoulder values, or Constant and Murley scores (Table III). The mean visual analog pain score decreased from 8 (range, 6 to 10) preoperatively to 6 (range, 0 to 8) at the time of follow-up (p = 0.19). The mean subjective shoulder value improved from 21% (range, 10% to 30%) to 33% (range, 20% to 70%) of a normal shoulder (p = 0.12). The average age and sex-adjusted Constant and Murley score improved from 24% (range, 12% to 32%) to 43% (range, 31% to 73%) (p = 0.17).
Eleven of the thirteen patients had persistent severe pain. These patients had an average glenohumeral joint space of 4 mm (range, 3 to 6 mm) demonstrated on standard axillary radiographs immediately postoperatively. However, radiographs made during their last follow-up visit (between six and 102 months postoperatively) showed findings consistent with graft failure, as evidenced by a complete loss of the joint space (Figs. 1-A, 1-B, and 1-C). In four of the thirteen patients, the pain was associated with stiffness, and arthroscopic capsular release was performed at a mean of six months (range, five to eight months) after the index procedure. At the time of arthroscopic débridement, there was no observable graft remaining over the glenoid. While all four patients had improvement in the range of motion, they had no reduction in the pain after this arthroscopic procedure, and all eventually underwent a second revision, to a total shoulder arthroplasty.
Ten patients underwent a revision total shoulder arthroplasty at a mean of sixteen months (range, six to thirty-three months) after the index glenohumeral soft-tissue interposition arthroplasty. Of the three patients who did not undergo revision surgery, one was an eighteen-year-old woman in whom severe arthritis had developed secondary to thermal capsulorrhaphy. She had had several unsuccessful arthroscopic débridements before she presented to us. She then underwent biological resurfacing of the glenoid with use of an Achilles tendon allograft and a humeral hemiarthroplasty. The pain resolved completely, although she had moderate limitation of shoulder flexion and internal and external rotation. At the time of the last follow-up, three years postoperatively, she was very satisfied with the outcome of the surgery and was able to perform most of her sports activities with no major limitations.
The postoperative course of the remaining two patients who did not undergo revision total shoulder arthroplasty was complicated by infection. In one of them, the infection was discovered within four weeks after the surgery and was treated with irrigation and débridement and a six-week course of intravenous antibiotics. The infection resolved, but the patient had persistent pain and a limited range of motion. The second patient had persistent infection and pain and ultimately underwent a resection arthroplasty, with a poor result.
Management of severe glenohumeral arthritis in young patients has been controversial. Because this group of patients can expect to return to a high level of activity, the loads exerted on the glenohumeral joint may be high and loosening of a glenoid component can be a major concern.
Interposition soft-tissue arthroplasty has been proposed as a treatment option for young active patients with glenohumeral arthritis26-31. The experience of Burkhead and colleagues14-16 has raised expectations regarding soft-tissue resurfacing of the glenoid as an alternative to placement of a glenoid component. However, the results were almost uniformly poor in the patients in the present study, with an overall failure rate of 92%. Complete loss of the graft with replacement by scar tissue was observed in the ten patients who underwent revision to a total shoulder arthroplasty.
In the shoulder joint, humeral rotation on the glenoid is accompanied by some degree of translation, and thus compression and shear forces can be high32. This is probably the most logical mechanical explanation for the failure of biological graft resurfacing of the glenoid. There is currently no evidence of a biological mechanism of graft failure. Since the surgical technique used in our patients was essentially the same as the technique described by Krishnan et al.16, we believe that patient age was the main factor leading to the substantial difference between their outcomes and ours. The average age of our patients was thirty-four years, whereas the average age of their patients was fifty-two years. Thus, Krishnan et al. may have treated a much less active population of patients.
Our study had two main limitations. First, it was retrospective. Second, because we used strict inclusion criteria, the number of patients who underwent the procedure was relatively small, limiting our ability to analyze the results statistically.
We conclude that, despite the conceptual attractiveness of soft-tissue resurfacing of the glenoid, its liberal application in young patients should be pursued with caution. Our results demonstrate no evidence that the graft material used in this study was a durable interposition surface, and our patients did not experience reliable pain relief or improved function. 