18.2.1 Physical Examination

The most common presentation for an indolent PSI is a painful and stiff shoulder. The patient with a PSI will complain of a constant dull ache in the shoulder. The pain will often worsen with activity. Signs of infection such as draining sinus, erythema, induration, and effusion are often absent in these patients. However, trace patchy erythema on a male patient should raise suspicion for infection (► Fig. 18.3).

18.2.2 Imaging

High-quality radiographs are essential in the evaluation of periprosthetic infections. Aseptic humeral loosening is a rare finding ⁷ and, if seen on an X-ray of a patient with a painful prosthesis, is often indicative of a PSI. Additional radiographic signs include generalized bone resorption and periprosthetic radiolucent lines ⁷ (► Fig. 18.4).

Advanced imaging such as computed tomography and magnetic resonance imaging with metal artifact reduction protocols can be helpful in confirming loosening noted on X-rays and aid in eventual operative planning, especially in the evaluation of glenoid bone stock and rotator cuff muscle quality, but are not generally diagnostic for infection in isolation. Nuclear medicine testing is of little value in the setting of a painful arthroplasty, as conventional radiography has higher sensitivity for infection. ⁹

18.2.3 Laboratory Testing

Laboratory testing is a cornerstone of the traditional workup for a prosthetic joint infection. WBC, ESR, and CRP are the most commonly used tests to evaluate or diagnose an ongoing infection. However, in the setting of a P. acnes PSI, CRP was elevated in only 42% of cases, while ESR was elevated in only 16% of cases in a study by Piper et al. ¹⁰ Unfortunately, WBC, ESR, and CRP values are useful when elevated, but normal values do not exclude the possibility of PSI. When elevated preoperatively, postoperative trends of laboratory markers are useful in monitoring for resolution or recurrence of infection.

18.2.4 Joint Aspiration

Joint aspiration is a valuable tool in the evaluation of PSI. It is a simple procedure and can be performed efficiently and easily in the office. It can be immediately diagnostic of a joint infection. Color, clarity, and viscosity of any fluid obtained should be noted and aspirate sent for gram stain, cell count with differential, and cultures including aerobic, anaerobic, acid fast bacteria, and fungal species. It is our standard of care to hold cultures for 14 days to maximize the yield of the slow growing P. acnes organism. ¹¹

Nevertheless, the sensitivity of cultures from joint aspirates for P. acnes is low. Dilisio et al found that aspiration and culture identified only 11% (1/9) of patients for whom P. acnes was eventually isolated at the time of revision surgery. ¹² This may be due to the fact that the P. acnes eventually found at revision surgery are in the sessile form and adhered to the biofilm on the prosthetic surfaces rather than freely floating in the joint fluid. Tissue cultures tend to have a higher yield than joint fluid cultures. ¹³ Despite poor sensitivity, it is our standard of care to aspirate all painful shoulder arthroplasty joints, because a positive culture can be diagnostic of PSI.

Newer inflammatory markers as well as gene amplification techniques have promise but have not been widely adopted at this time. Synovial interleukin-6 (IL-6) has sensitivity of 87% with specificity of 90% and is significantly elevated in the presence of P. acnes infection. ¹⁴ Serum IL-6 testing does not share the same usefulness as synovial IL-6. Synovial alpha-defensin has sensitivity and specificity of 87 and 63%, respectively. ¹⁵ Polymerase chain reaction (PCR) has sensitivity of 79% and specificity of 86% in the diagnosis of orthopaedic infections and is recommended for use in patients previously on antibiotics and when suspecting low virulence organisms. ¹⁶ Advanced technologies may improve our diagnostic accuracy when they become more widely available and affordable, but are not currently routinely used by Dr. Kelly.

18.2.5 Biopsy

Synovial biopsy is most useful when multiple specimens are obtained properly and are positive among a preponderance of samples. Therefore, in patients with a failed total shoulder arthroplasty and concern for infection in whom workup has been negative with no identifiable reason for failure, an open or arthroscopic biopsy should be considered. Biopsy can provide a higher degree of certainty prior to undergoing extensive revision. Results of arthroscopic biopsy have been shown to be consistent with open biopsy, and both techniques allow thorough inspection of components for mechanical loosening and wear. ¹² Open biopsy allows for removal of larger volume of tissue untouched by arthroscopic fluid and mechanical shaver. However, arthroscopy has the potential advantage of decreased recovery time and preservation of subscapularis insertion.

Proper biopsy technique is important to avoid cross contamination and false-positive results. Clean unused instruments should be used to obtain three to five tissue specimens of capsule or periprosthetic tissue. Separate instruments should be used for harvest of each specimen and a no-touch technique where each specimen is placed immediately into individual sterile specimen container for transport. Histopathologic evaluation should be performed on one specimen to evaluate for acute or chronic inflammation. ¹⁷

18.2.6 Intraoperative Frozen Section

Intraoperative frozen sections can be very useful during revision surgery to help guide clinical decision making prior to reimplantation. Results are available intraoperatively and help ensure infections are appropriately treated prior to definitive revision. Mirra’s criterion (5 polymorphonuclear leukocytes [PMNs] per high power field) is associated with sensitivity of only 50% for P. acnes. Therefore, a new guideline has been proposed: 10 PMNs per five high power fields. This guideline is associated with a sensitivity of 72% and specificity of 100%. ¹⁸ While a frozen section cannot exclude infection, the presence of PMNs in tissue is a sign of an acute inflammatory response, which is most commonly caused by an infection. Therefore, while the surgeon cannot depend on the test in all circumstances, it is highly indicative of an infection when significant numbers of WBCs are found, and therefore frozen section analysis is commonly used in revision arthroplasty cases.

18.2.7 Diagnosis

Making the diagnosis of a PSI continues to be a challenge given that there is currently no consensus or gold standard definition of an infection. The diagnosis of a PSI requires the synthesis of many pieces of data. In order to synthesize the data and facilitate later intraoperative decision making, we categorize patients preoperatively as definitely infected, probably infected, potentially infected, and unlikely infected. Classification is based on all available data, including clinical examination, imaging, laboratory data, and culture of aspirate and periprosthetic tissue. More data points allow for a more comprehensive analysis and hopefully a more accurate classification. Classification allows us to stratify the patient’s pre-revision risk of having a true infection and their likelihood of future or recurrent infections, which thereby facilitates the choice of treatment.