BACKGROUND

Abnormal glenoid version positioning has been recognized as a cause of glenoid component failure caused by the rocking horse phenomenon. In contrast, the importance of the glenoid inclination has not been investigated.

Aims of the study: To assess the results of one-stage revision of infected reverse total shoulder arthroplasty as a new treatment for this major surgical complication.

Materials: 11 consecutive patients with infected reverse shoulder prosthesis (DePuy International Ltd, Leeds, UK) were treated with a direct-exchange reverse total shoulder arthroplasty between March 2005 and June 2007. These patients were retrospectively followed.

Methods: Via a superior extended deltoideopectoral, clavicular osteotomy approach, according to Redfern-Wallace, a removal of the prosthesis is performed. Multiple operative specimens (4 to 8) are taken and cultured. A thorough synovectomy is performed before implanting a new reverse total shoulder arthroplasty (7 Delta^® 3.2 tsp and 4 DeltaXtend^® tsp). A gentamycine (Duracol^®) impregnated membrane is interpositioned between baseplate and glenosphere. Immediate postoperative passive and active mobilisation is permitted if no pain is present.

Results: Clinical symptoms are seldom severe pain (3) or severe limitation of function (3). A fistula is mainly present (8) without alterating the function. No prosthetic loosening was present at the humeral and glenoid site confirming the absence of radiological infectious signs. All but one patient are considered free of infection at mean follow up of 24 months (12m – 36m) and without antibiotic treatment for minimum 6 months. In only 1 patient the infection persisted necessitating a two-stage revision, unfortunately without definitive cure. Only three early complications are seen (< 2 months: posterior dislocation, postoperative haematoma and clavicle fracture). The mean postoperative Constant-Murley score was 52 (14–81) at latest follow up.

Peroperative samples identified Propionbacterium species (5), Coagulase-negative staphylococci (4), MRSA (1) and with E.Coli (1) infection. Monobacterial infection was seen in 6 shoulders, multibacterial in 2 shoulders and in 2 shoulders cultures were negative.

Discussion: This cohort of patients has a different syndrome than the known infected anatomical prosthesis. This pathology seems to be rarely associated with severe pain (3) or limitation of function (3). Draining fistulas without alterating the function of the shoulder [CS: mean 43 (3–63)] are more frequently present making a preoperative diagnosis easier and more certain.

Recurrence rate of infection is comparable to the classical two-stage revision. Preoperative stiff and painful shoulders seems to have a bad prognosis despite definite cure of the infection. Supple shoulders (mainly associated with a fistula) can be treated with a good functional result.

Conclusion: One-stage revision arthroplasty is an attractive alternative treatment for infected reverse total shoulder arthroplasty.

We retrospectively reviewed 11 consecutive patients with an infected reverse shoulder prosthesis. Patients were assessed clinically and radiologically, and standard laboratory tests were carried out. Peroperative samples showed Propionbacterium acnes in seven, coagulase-negative Staphylococcus in five, methicillin-resistant staphylococcus aureus in one and Escherichia coli in one. Two multibacterial and nine monobacterial infections were seen. Post-operatively, patients were treated with intravenous cefazolin for at least three days and in all antibiotic therapy was given for at least three months. Severe pain (3 of 11) or severe limitation of function (3 of 11) are not necessarily seen. A fistula was present in eight, but function was not affected. All but one patient were considered free of infection after one-stage revision at a median follow-up of 24 months, and without antibiotic treatment for a minimum of six months. One patient had a persistent infection despite a second staged revision, but is now free of infection with a spacer. Complications included posterior dislocation in one, haematoma in one and a clavicular fracture in one. At the most recent follow-up the median post-operative Constant-Murley score was 55, 6% adjusted for age, gender and dominance.

A one-stage revision arthroplasty reduces the cost and duration of treatment. It is reliable in eradicating infection and good functional outcomes can be achieved.