Common Pathogens

Many of the bacteria responsible for infection in patients who undergo shoulder surgery are the same organisms that can cause an infection in any orthopedic patient. Typically, bacteria that are normally found as skin flora are most likely to invade the shoulder joint following any procedure that violates the skin barrier. Staphylococcus aureus, Staphylococcus epidermidis, and other coagulase-negative Staphylococcus species are the most common potential pathogens that colonize human skin. In several case series, S aureus has been found to account for as much as 60% of native joint infections. Similarly, when it comes to prosthetic joint infections S aureus and coagulase-negative Staphylococcus have been found to be responsible for as much as 23% and 43% of cases, respectively. These pathogens, however, are not unique to the shoulder and are usually the first targets of directed antibiotic therapy when an infection following orthopedic surgery is suspected. More unique to the shoulder is P acnes , which has been found to be responsible for as much as 51.3% of postoperative shoulder infections. A study by Patel and colleagues looked to characterize the colonization of bacteria at various sites common for orthopedic surgery, and found that although the burden of S aureus was increased at hip, knee, and shoulder compared with P acnes , the burden of P acnes was significantly greater for the shoulder region than for the lower extremity. This higher concentration of P acnes in the shoulder region has even led some to speculate that the bacterium is responsible for shoulder abnormality even before surgical treatment. Levy and colleagues collected shoulder aspirates and tissue samples from 55 consecutive patients before primary joint replacement, and found that 41.8% of patients were positive for P acnes colonization.

Surgical-Site Preparation

Preoperative surgical-site preparation with an antiseptic solution is often performed in the operating room just before incision. Various iodine and alcohol-based solutions are available that have been shown to have varying efficacy. Saltzman and colleagues performed a randomized prospective study of 3 commercially available solutions: ChloraPrep (2% chlorhexidine gluconate and 70% isopropyl alcohol; ChloraPrep, Leawood, KS, USA), DuraPrep (iodine povacrylex and 74% isopropyl alcohol; 3M, Minneapolis, MN, USA), and Betadine (Purdue Pharma, Stamford, CT, USA). The investigators found ChloraPrep solution to be significantly more effective than the other 2 solutions in eliminating coagulase-negative Staphylococcus. However, no solution was found to be superior in eliminating P acnes . Similarly, a randomized controlled trial that compared patients who used 2% chlorhexidine gluconate cloths at home before surgery with those who washed with standard soap and water found the positive culture rate for coagulase-negative Staphylococcus and overall positive culture rate to be lower in the patients who used the chlorhexidine cloths. A trend toward a lower positive culture rate was also observed for P acnes , but this was not a significant reduction. Given the tendency of P acnes to colonize the hair follicles and sebaceous glands deep to the skin surface, it is not surprising that these solutions are less effective in decreasing its numbers.

Given the ineffectiveness of surface cleaning solutions on P acnes , other methods of reducing infection are encouraged.

Shaving, Draping, Incision

Many techniques necessary for decreasing the rate of P acnes infection are implemented in the operating room at the time of surgery. Like the clinical guidelines used by orthopedic surgeons in other subspecialties, the use of proper hair clipping and draping of the surgical site are encouraged to reduce infection rates. In addition, the use of proper ventilation flows, reduced operating theater traffic, and sterile autoclaving of surgical instruments and equipment should always be implemented. Some investigators have advocated the use of separate surgical knives for skin and deeper tissues to avoid possible seeding of the deeper surgical planes. A study that used separate knives for the skin and deeper layers found that the rate of contamination of the skin blades with P acnes (among other flora) was more than twice that of the inside blades. As P acnes is often found in the deeper layers of the skin that are untouched by surface-cleaning solutions, the idea of using a separate knife blade for shoulder surgery certainly seems a cheap and efficient way to reduce deep-tissue contamination. Such a simple technique can easily be added to the normal sterile precautions used in orthopedic operating rooms worldwide.

Physical Examination

P acnes infections of the shoulder are difficult to assess on physical examination because of the indolent nature of the bacteria. Unlike an infection with Staphylococcus species, patients with P acnes infections do not tend to present with the classic signs of inflammation such as erythema, fever, micromotion tenderness, or swelling. The most common complaint of postoperative patients harboring a P acnes infection tends to be persistent shoulder pain. In a retrospective review of 10 patients diagnosed with a P acnes deep shoulder infection, 9 patients had only a chief complaint of pain for a mean duration of about 3 months. The remaining patient also presented with pain with the additional prototypical infection symptoms of fever, swelling, and wound drainage. In a patient population that presents preoperatively with shoulder pain, it is easy for the treating clinician to disregard infection in the differential in favor of normal postoperative pain or adhesive capsulitis as the offending cause. However, there should be earnest concern for infection when a patient continues to complain of a painful motion arc postoperatively.

In some instances, symptoms do not even occur until years after surgery. Zeller and colleagues studied P acnes infections in 50 patients who underwent arthroplasty for various joints including the hip, knee, and shoulder, and found that patients could be divided into two groups: those who developed symptoms within 2 years of their index operation and those who developed symptoms more than 2 years after their operation. The most common symptoms were pain and limited range of motion. Symptoms of infection were significantly more frequent in the group that had an asymptomatic interval of less than 2 years; however, positive intraoperative cultures were significantly greater in those patients with a longer asymptomatic interval. This finding led the investigators to suggest that different pathophysiologic mechanisms of infection can occur. In the former group a more acute P acnes infection is manifested, whereas in the latter group a colonization of the arthroplasty occurs first, which later leads to symptomatic infection. If a patient does develop the prototypical symptoms of infection, these will tend to manifest sooner in the postoperative period rather than later. Fever, local inflammation, and sinus tract formation was seen in 18 of 35 patients within 2 years of surgery, compared with only 1 of 15 patients who developed symptoms beyond 2 years.

The shoulder surgery patient who presents postoperatively with continued difficulty should always be approached in a similar manner, with the treating physician having concern for infection even years after surgery. Physical assessment should include a thorough examination of the patient’s surgical incisions to assess for any erythema, swelling, or drainage. In the setting of a P acnes infection, incisions are likely to appear normal. However, there is the possibility of a multiple organism infection, and treatment should be tailored to account for all organisms. In addition, the patient should have a complete examination of shoulder range of motion and pain level. Stiffness and continued pain are the most likely symptoms to present in the P acnes –infected patient. Dictating thorough physical examinations in the patient’s medical record are essential for comparison, and may reveal a decreasing range of motion arc or increased pain levels. Documenting a worsening examination allows the patient and physician to recall prior presentations and guide further evaluation if deemed necessary.

Imaging

Evaluation of the painful postoperative shoulder should always include plain radiographs (anterior-posterior view, scapular-Y view, and axillary view). Radiographs are important to rule out subluxation of the humeral head, heterotopic ossification, hardware failure, or joint arthrosis or other mechanical causes that contribute to joint pain and stiffness in the postoperative patient. With early pyogenic infection, radiographic findings are typically normal. By contrast, radiographs in subacute and delayed cases of pyogenic infections may show osteopenia, lucencies around a prosthetic component, and pseudosubluxation of a prosthetic humeral head. Most P acnes infections tend to fit into this latter category of indolence. A retrospective study of 52 patients with P acnes infections of various orthopedic implants divided patients into 3 categories: definite infection (based on local signs and at least 2 P acnes –positive isolates), probable infection (based on local signs and one positive P acnes isolate), and possible infection (based on no signs of infection but signs of prosthetic dysfunction or arthrosis and one positive P acnes isolate). Of note, all 15 patients found to have changes on plain radiographs were in the possible infection group (n = 17). None of the patients in the more clinically concerning infection groups were found to have radiographic changes.

The divergence between clinical and radiographic presentation in cases of P acnes creates a challenge for the treating physician. On one hand, a patient who appears to have an infection of the postoperative shoulder on physical examination may not have any infectious signs on radiography; conversely, a patient with signs of infection on radiography may not present with the most concerning physical examination. Pottinger and colleagues prospectively followed patients undergoing revision arthroplasty of the shoulder, and found that patients with evidence of humeral component osteolysis radiographically had a 10-fold increase in the prognosis of a P acnes infection ( P <.01), whereas those with evidence of humeral component loosening in the absence of osteolysis had a 3-fold increase ( P <.01). Similarly, evidence of glenoid loosening was also found to significantly correlate with a positive P acnes culture ( P <.01). For the treating orthopedic surgeon, these studies show that one should maintain a high index of suspicion for infection when either the physical examination or radiographic imaging is concerning. In patients without visible hardware (ie, synthetic polymer anchors) or no implants, plain radiographs are most often unremarkable. Advanced imaging techniques, such as magnetic resonance (MR) imaging or computed tomography, are often unnecessary unless a raised clinical suspicion on physical examination or plain radiographs requires further investigation. In the absence of cystic absorption around anchor material, no clear MR imaging findings have been observed in patients with P acnes infections in the native shoulder joint.

Laboratory Evaluation

Peripheral blood tests are often an element of the infection workup. ESR and CRP are 2 of the most common inflammatory markers obtained along with a peripheral leukocyte count. These tests are nonspecific and are normally elevated even after an uncomplicated surgical procedure. After an uneventful operation, the ESR will slowly decline while the CRP level declines more rapidly, normalizing within about 2 weeks. In cases of acute purulent infections with virulent organisms, these inflammatory markers can be significantly elevated and can help reveal an infectious process. Unfortunately, they are typically within the range of normal values in the setting of a P acnes infection. In a retrospective study of 107 patients undergoing revision shoulder arthroplasty who were found to have unexpected positive cultures, 88% of patients had all blood test results (WBC, ESR, CRP) within normal reference ranges. Similarly, a prospective examination of 108 revision shoulder arthroplasty cases associated with positive cultures found no significant association with the WBC, ESR, or CRP values. Zeller and colleagues found a small correlation between positive P acnes cultures and elevated ESR/CRP values, but only in one-third of cases. Lastly, a comparison of definite prosthetic shoulder infections with cultures positive for P acnes versus those with cultures positive for other microorganisms showed a trend of the former group being less likely to have elevated ESR or CRP counts in contrast to the latter. In short, peripheral blood tests for inflammatory markers are often unremarkable in the setting of P acnes infection. Regardless of the observed low sensitivity and specificity of general inflammatory markers, such laboratory measures should still be obtained if a patient has an indolent or early infection from another microorganism that could possibly elevate their levels.

Culture

The culturing of P acnes is a difficult process in comparison with that of other microorganisms responsible for orthopedic infections. The growth cycle of P acnes is inherently slow. Many laboratories routinely discard culture samples after 3 to 5 days. Most bacterial species responsible for orthopedic infections, such as Staphylococcus species, are identified within this time frame, but the slow growth of P acnes requires further incubation. Studies have placed the average incubation time for a positive P acnes culture at between 9 and 15 days. In 108 revision shoulder arthroplasties performed for increased pain and stiffness, positive P acnes cultures taken at the time of surgery went from 45% within 1 week to 100% within 4 weeks. This finding indicates that more than half of potential P acnes infections can be missed within the typical hospital laboratory incubation time and that a large false-negative rate can occur. Some investigators argue that this increased incubation time is not required as long as proper handling of the culture samples is carried out to encourage P acnes growth. Shannon and colleagues performed multiple broth and plate culture of 14 subjects who had more than 1 positive shoulder culture for P acnes , and found that 52 of 53 broth and 27 of 28 plate cultures were positive within 7 days of sampling. The investigators argued that the difference in their results in comparison with those in most of the literature is that the collection and transportation of their samples were under anaerobic conditions, unlike in prior studies. Typically, hospital laboratories place initial culture specimens under both aerobic and anaerobic conditions, although the latter is sometimes not strictly devoid of oxygen when performed.

It is important for the treating orthopedic surgeon to be familiar with the practices of the microbiology laboratory to determine the most accurate culturing of a suspected shoulder infection. The need for extended incubation times can be impractical for many laboratories because of increased labor costs and the costs associated with increased recovery of nondiagnostic isolates. A 7-day incubation period seems to be sufficient time for P acnes growth if an institution’s microbiology laboratory is able to perform all the tasks of collection and preparation of culture samples under ideal anaerobic conditions; however, if exposure to aerobic conditions is used, the incubation time should last at least 2 weeks. Regardless of preparation protocols, it is best for multiple culture samples to be taken in the operating room at the time of surgery, as multiple cultures positive for P acnes at the time of surgery appear to correlate more convincingly with true infection than with a contaminant.