Tuesday, March 29, 2016

Periprosthetic joint infection - challenges of definition, diagnosis and treatment


Outcomes in the treatment of periprosthetic joint infection after shoulder arthroplasty: a systematic review

These authors performed a systematic review of 30 qualifying articles. They found that Propionibacterium acnes was most commonly reported, representing 38.9% of infections, followed byStaphylococcus species.

Risk factors for shoulder PJI include previous surgery, increased age, male gender, increased body mass index, and diabetes mellitus. The average white blood cell count in 13 studies was 7472 cells/┬ÁL. Ten studies reported a mean erythrocyte sedimentation rate of 27.6 mm/h, whereas 14 studies reported a mean C-reactive protein level of 2.6 mg/dL. 

P. acnes in intraoperative culture specimens was an independent risk factor for failed treatment for shoulder PJI. 

No statistical difference was found in the success rates of 1-stage, 2-stage, or resection arthroplasty revision; each displayed a success rate >90%. However, single-stage revision produced the highest mean Constant score; implant retention resulted in the best range of motion.

Comment: The challenge here is the difficulty in (1) defining "periprosthetic joint infection" in a useful way and (2) in distinguishing the 'stealth' form of shoulder arthroplasty failure with positive cultures (pain, stiffness, loosening, but no clinically apparent evidence of infection) from the 'obvious' form of shoulder arthroplasty failure with positive cultures (draining sinus, erythema, fever, positive laboratory studies). It is recognized that a high percent of shoulder arthroplasty revisions for pain, stiffness, or implant loosening are culture positive for organisms such as Propionibacterium and coagulase negative staphylococcus, even though these cases may fail to meet the criteria established by the Musculoskeletal Infection Society for periprosthetic hip and knee infections.

Rather than using the term 'unexpected positive cultures (UPC)', we now know to expect positive cultures in a high percentage of samples harvested from the surgical revision of failed shoulder arthroplasty. While this review found that 40% of the positive cultures were for Propionibacterium, it seems like that the actual rate may have been even higher, because of the variability in the culture protocols among the articles reviewed.

Our current approach is to use a single-stage exchange and prolonged postoperative antibiotic therapy for 'stealth' type cases (positive culture without clinical evidence of infection), although as this review points out, Propionibacterium can be remarkably difficult to eradicate in some cases. We reserve two-stage exchange and prolonged postoperative antibiotic therapy for 'obvious' type cases (clinical evidence of infection and positive cultures) and for those failing a single stage revision.

We suggest that the evaluation and management of shoulders with failed shoulder arthroplasties will better informed if authors report the actual results of the cultures (including the number and sources of specimens submitted) and the  clinical findings, rather than applying an arbitrary definition of 'infection'.