UW Shoulder and Elbow Academy: Delayed onset of total shoulder pain and stiffness

A 60 year old patient presented after a right total shoulder arthroplasty 20 years earlier. The shoulder had become progressively stiff and painful over the last several years. On examination his shoulder was stiff and painful. There were no other clinical or laboratory suggestions of inflammation or infection. His SST was only 1/12. His radiographs showed some endosteal bone resorption near the distal end of his stem and some suggestion of thinning of the glenoid polyethlene.

At revision surgery the humeral component was somewhat loose, but the glenoid was secure, showing only minimal evidence of surface wear. There were no signs of joint fluid or inflammation within the joint. However, there was a thick leather-like humeral membrane, part of which is shown below.

The pathology on this membrane is shown below.

We performed a single stage exchange to a hemiarthroplasty, redraping and regowning after component removal and debridement and harvesting of five deep specimens for culture.

The new humeral stem was inserted using impaction allograft.

Because of our suspicion of infection - even though it had been 20 years since his index procedure, we started him on immediate Ceftriaxone and Vancomycin via a PICC line.

At 6 days after surgery, all of his cultures were positive for Propionibacterium

This case demonstrates the ability of Propionibacterium prosthetic shoulder infections to appear years after the index procedure without clinical or surgical evidence of infection. In male patients with delayed onset of pain and stiffness, we usually perform a single stage exchange to a hemiarthroplasty with a plan for IV antibiotics until the cultures are finalized at 3 weeks to make sure there is no gap in antibiotic coverage. If the cultures are positive, IV antibiotics are continued for 6 weeks followed by a 6 month course of oral antibiotics.

It is important to contrast this case with the following one:
Detritic synovitis can mimic a Propionibacterium periprosthetic infection

These authors illustrate that the clinical findings of detritic synovitis (the macrophage reaction to polyethylene, cement or metal debris) complicating a total shoulder arthroplasty can strongly resemble those of a ‘stealth’ periprosthetic shoulder infection with a low-virulence organism such as Propionibacterium, including a clinical presentation long after the index procedure. At present, the important differentiation between these two etiologies can only be ascertained by awaiting the results of cultures obtained at the time of revision surgery. The surgical and antibiotic treatment decisions must be made before the culture results become available.

Here is an informative case. A 76-year-old right hand dominant man presented with right shoulder pain and decreased range of motion. He had a history of bilateral total shoulder arthroplasties, his left 15 years prior and his right 14 years prior to his visit with us. Following his index surgeries he initially did well with full painless range of motion and was able to return to full activities. Eleven years after his right arthroplasty he experienced the insidious onset of worsening shoulder pain and stiffness with no known injury. He also noted painful catching and locking in his shoulder joint with certain shoulder movements. His symptoms were unresponsive to non-operative treatment, including exercises, anti-inflammatory medications and a corticosteroid injection. His left shoulder had some stiffness but was otherwise asymptomatic. The CBC, sedimentation rate and C-reactive protein were all normal.

Physical examination demonstrated a well-healed surgical scar with no erythema, drainage or evidence of infection. Both active and passive ranges of motion were decreased. There was palpable crepitus on range of motion. Rotator cuff strength was intact, as was neurologic function of the affected extremity. Radiographs demonstrated a thinned glenoid component with surrounding osteolysis, appearing grossly loose. The humeral component was well positioned with surrounding osteolysis of the medial and lateral proximal humeral bone. There were no radiolucencies around the distal stem and the prosthesis did not appear grossly loose.

The patient was advised to have a revision shoulder arthroplasty to manage his symptoms and loose glenoid component. Because of the high index of suspicion of an infection, the plan included a one-stage revision to hemiarthroplasty followed by a course of intravenous antibiotic therapy until culture results were finalized. At the time of revision surgery, perioperative antibiotics were held until tissue cultures were obtained. There was abundant scar tissue surrounding the shoulder. A synovial fluid aspiration prior to capsulotomy showed grossly cloudy fluid with a negative gram stain, with no polymorphonuclear cells or organisms seen.

There was diffuse membranous tissue around both the humeral and glenoid components. There was osteolysis of the proximal humerus, but the humeral component was securely fixed ; it was removed without complication. The glenoid component was grossly loose and easily removed. There was significant wear of the glenoid polyethylene and osteolysis of the underlying glenoid bone. The rotator cuff was intact.

A total of 8 samples for culture were taken from various locations within the glenohumeral joint, including the glenoid membrane, collar membrane between the modular humeral head and stem, humeral canal membrane, bursa, glenoid explant, and stem explant. Due to preoperative and intraoperative concerns of infection, including cloudy fluid, abundant membrane, glenoid loosening and osteolysis, the patient was treated with a one-stage revision consisting of removal of the loose glenoid and single stage exchange of the humeral component using Vancomycin soaked cancellous allograft to secure the stem by impaction grafting. The remaining glenoid bone was smoothed, no bone graft was added, and no glenoid component was reimplanted. Cultures were grown on four types of media: blood agar, chocolate agar, Brucella agar and brain-heart infusion broth as previously published. Postoperatively the patient was placed on IV Ceftriaxone 2g daily and Vancomycin 1g daily via PICC line. The Vancomycin was discontinued after 2 days when the cultures failed to grow MecA CoNS. All cultures were negative at 21 days at which time all antibiotics were discontinued. Permanent pathology of the deep tissues identified fibrotic tissue with chronic inflammation, the absence of neutrophils, and a foreign body giant cell reaction consistent with detritic synovitis

After surgery, he was placed on the standard post arthroplasty rehabilitation program focusing on range of motion in the first six weeks, followed by progressive anterior deltoid strengthening. At his six-month follow up visit, the patient was recovering well with no complaints of pain. His Simple Shoulder Test had improved from 5 out of 12 prior to his revision to 10 of 12, and radiographs showed a well-fixed humeral component.

Comment: In this case, it seems unlikely that the cultures in this were falsely negative – multiple tissue and explant specimens were obtained before antibiotic administration, the specimens were cultured on multiple media and observed for 21 days. Thus, we must consider the possibility that the detritic synovitis from polyethylene debris produced osteolysis and periprosthetic membrane formation similar to that of a Propionibacterium infection. Until better methods become available for differentiating the two conditions, our practice is to continue to treat shoulders with osteolysis and glenoid component loosening as if they were infected until cultures prove this not to be the case.

=====
The reader may also be interested in these posts:

Healing through joint replacement

Supporting progress in shoulder surgery