Subacromial corticosteroid injections are a well-recognised management for chronic shoulder pain and are routinely used in general practice and musculoskeletal clinics. Mycobacterium tuberculosis (TB) of a joint is a rare presentation in the United Kingdom. International literature exists for cases of reactivated latent tuberculosis following intra-articular corticosteroid injections in a knee; however there are no reports of a primary presentation of undiagnosed TB in a joint following therapeutic corticosteroid injections.

A seventy-four year old lady presented with a one-year history of a painful shoulder, which clinically manifested as a rotator cuff tear with impingement syndrome. Following three subacromial depo-medrone injections, the patient developed a painless “cold” lump which was investigated as a suspicious, possibly metastatic lesion. This lump slowly developed a sinus and a subsequent MRI scan identified a large intra-articular abscess formation. The sinus then progressed to a large intra-articular 5×8 cm cavity with exposed bone (picture available). The patient had no diagnosis of TB but had pathogen exposure as a child via her parents.

The patient underwent three weeks of multiple débridement and intravenous amoxicillin/flucloxacillin to treat Staphylococcus aureus grown on an initial culture. Despite best efforts the wound further dehisced with a very painful and immobile shoulder. Given the poor response to penicillin and ongoing wound breakdown there was a suspicion of TB. After a further fortnight, Mycobacterium tuberculosis was eventually cultured and quadruple antimicrobial therapy commenced. Ongoing débridement of the rotator cuff and bone was required alongside two months of unremitting closed vacuum dressing. The wound remained persistently open and excision of the humeral head was necessary, followed by secondary wound closure. There were no extra-articular manifestations of TB in this patient. At present the shoulder is de-functioned, the wound healed and shoulder pain free.

This unique case study highlights that intra-articular corticosteroid can precipitate the first presentation of Mycobacterium tuberculosis septic arthritis. The evolution of symptoms mimic many other shoulder complaints making confident diagnosis a challenge. The infective bone and joint destruction did not respond to the management described in the current literature. There remains a further management issues as to whether arthroplasty surgery can be offered to post-TB infected shoulder joints.

Taking a TB exposure history is indicated prior to local immunosuppressant injection, particularly in the older age group of western populations and ethnicities with known risk factors.

The aim of the New Zealand National Shoulder Arthroplasty Register is to evaluate the provision of shoulder arthroplasty across the entire country by both recording accurate technical information and measuring the clinical outcomes of all shoulder replacements performed in New Zealand.

An initial form is completed at the time of surgery which includes details of the patient, surgical indications, the surgical procedure, the implant and the operating surgeon. Six months following surgery, all registered patients are asked to complete the Oxford Shoulder Score and comment on post-operative complications.

Data from 686 consecutive primary and 44 revision shoulder arthroplasties were prospectively collected from January 2000 until December 2003. 82 surgeons performed shoulder arthroplasty during the study period but only 9 performed on average more than 5 per year. Their results at 6 months were statistically superior to those provided by other surgeons. Amongst all diagnoses, osteoarthritis scored significantly better than the rest and for this condition total shoulder replacement scored higher than hemiarthroplasty. 15 different prostheses were used, many of them too infrequently. There was no difference in outcome amongst the 5 most commonly used prostheses. The number of complications reported by patients and the revision rate within the study period was low. No benefit was observed in the use of laminar flow theatre to prevent infection.

The combination of technical data about the joint implanted and the individual patient assessment has made the New Zealand Shoulder Arthroplasty Register unique in its own kind. Our findings are in general agreement with the current literature and supports the idea that shoulder arthroplasty is better provided by surgeons with a higher yearly case-load.