Few cases of PVNS were accidentally discovered during surgery for joint instability or rotator cuff repair [4, 14–17]. In some patients, PVNS causes shoulder swelling, due to an excessive production of articular fluid or to a massive intraarticular hemorrhage. In all cases, the pathology is responsible for serious shoulder pain and a decrease in range of motion. These symptoms are even more evident if PVNS is associated with rotator cuff tear and/or glenohumeral osteoarthritis. We believe that PVNS may worsen rotator cuff tearing or cause itself a cuff tearing because of its proliferating nature, with a secondary degenerating effect of cuff tendons. It is also plausible that it may be due to the release of cytokines with tendon tearing effect.

Generally, total synovectomy is the standard treatment for PVNS; however, when the shoulder is involved, rotator cuff and articular surface status have to be considered because the final clinical outcome is strongly influenced by tendinous and cartilage conditions. Since the almost totality of the rotator cuff tear in cases of shoulder PVNS is classified as massive and because all cases of shoulder PVNS have different degrees of degenerative changes of articular surfaces, the recommended surgical options which should be associated with synovectomy are as follows: arthroscopic debridement of cuff tear [3, 5], partial cuff repair [3], shoulder hemiarthroplasty [6, 7], and total shoulder arthroplasty [8].

Recently, we examined the largest series of patients with shoulder PVNS, massive irreparable rotator cuff tear, and glenohumeral osteoarthritis treated with arthroscopic debridement and synovectomy [18]. Our purpose is to evaluate the result of the treatment in patients with shoulder PVNS and to compare these results to those obtained from patients of similar age and having massive irreparable cuff tear associated with hemorrhagic synovitis with or without glenohumeral arthropathy treated with debridement and synovectomy.

From 2003 to 2013, we prospectively treated nine patients (six females and three males, mean age 65.8 years, range 63–70 years; seven right shoulders; right side dominant in all cases) with PVNS of the shoulder (group I) and who responded to these inclusion criteria: age between 60 and 70 years and presence of a massive and irreparable rotator cuff tear with slight glenohumeral arthropathy [scored as Grade 1 or 2 according to Hamada method [19]], and arthroscopic histology confirming the diagnosis of PVNS characterized by pigment deposition, histiocytic cell infiltration, multinuclear giant cells, and foam cells in the hyperplastic synovial villi as the major histologic findings (Fig. 2a, b).

Age was considered as a primary factor because it was our intention to compare clinical results obtained after arthroscopic synovectomy and debridement to those registered in a control group (group II) composed of 20 consecutive patients (12 females and 8 males recruited in the same period, from 2003 to 2010), of comparable age (mean age: 66.7; range 62–70 years), and who had undergone arthroscopic debridement for massive and irreparable cuff tear associated with hemorrhagic synovitis (confirmed by mean of arthroscopic histology) and who had no (12 patients: 6 females and 6 males) or slight (8 patients: 6 females and 2 males) glenohumeral arthropathy (Grade 1 or 2 according to Hamada method). In group II, patients without glenohumeral arthropathy refused a latissimus dorsi transfer because they preferred to wait and to verify the clinical results of the arthroscopic treatment. Furthermore, patients of this group were not considered as candidates to receive a reverse shoulder prosthesis because of their young age and/or because the quality of the glenohumeral joint was interpreted as good or scarcely degenerated.

Shoulder arthroscopy was performed in the beach-chair position. Common posterior, lateral, anterolateral, and midglenoid portals were used. Synovectomy, tenotomy of the long head of biceps, and debridement of the cuff tear margins and of the degenerated articular surfaces plus regularization of the glenoid labrum were carried out. We did not perform acromioplasty to prevent coracoacromial ligament damage and shoulder cranialization. Patients affected by PVNS did not receive any method of intraarticular radiation therapy because the diagnosis was not known at the moment of the surgical procedure. After surgery, the shoulder was immobilized for 24–48 h in an internal rotation sling for comfort; then, passive and active exercises were permitted once the sling was removed.

Upon follow-up, none of the 9 patients with PVNS showed sign of a recrudescence of the pathology. The demographic data of patients relating to the two groups and those relating to their shoulder function preoperatively and at the follow-up are reported in Table 1 (Constant score) and Table 2 (subjective shoulder value). Results of each patient with PVNS are reported in Table 3.