Prevention of Infection

Prevention of Infection

Yoav Rosenthal, MD


Although relatively uncommon, periprosthetic joint infection (PJI) is a serious complication of shoulder arthroplasty resulting in revision surgery, an extended hospital stay, prolonged use of antibiotics, and a negative impact on patients’ outcome and satisfaction.1,2,3 In recent years, the prevalence of shoulder arthroplasty has been steadily growing in the United States (from 13,837 shoulder arthroplasties in 1993 to 79,105 in 20144,5) and is expected to continue to grow.6 Therefore, despite the relatively consistent incidence of PJI at approximately 1% (ranging from 0.7%-1.8%) for primary cases and between 4% and 15% in revision cases,7 the total prevalence of PJI of the shoulder is expected to increase as well.7 This concerning increase in the prevalence of PJI of the shoulder has significant financial implications with rising healthcare expenditures. In North America, the most common management strategy for PJI of the shoulder involves a two-stage protocol.8,9 Between 2003 and 2012, the mean hospital cost of these staged procedures was $35,825, compared with $16,068 for primary shoulder arthroplasty.8,9 The potentially devastating effect on the patient’s quality of life along with the increased financial burden on the healthcare system compels the surgeon performing shoulder arthroplasty to take all the necessary measures to prevent infection in patients undergoing shoulder arthroplasty. Recognizing the organisms that cause infection in shoulder arthroplasty is essential to select the appropriate prophylactic regimen and to have the most significant preventive impact.

Common Pathogens in Shoulder Arthroplasty

Approximately 85% of PJIs involve a single bacterium.10 The most common pathogens are Cutibacterium acnes (formerly known as Propionibacterium acnes), Staphylococcus epidermidis, and Staphylococcus aureus. Other bacteria identified are coagulase-negative staphylococci, Corynebacterium, Pseudomonas spp., Peptostreptococcus spp., Finegoldia magna, Bacillus species, Enterobacter cloacae, Proteus mirabilis, Staphylococcus albus, diphtheroid, and Enterococcus.3,9,10,11,12 The remaining 15% of PJIs are polymicrobial, are not well understood, and may involve other types of bacteria.10 Staphylococci species are very commonly found in shoulder PJI (between 27% and 52%3,11,12); however, they are not unique to shoulder infection. Nonetheless, most of the prophylactic measures utilized—that is, antibiotics and skin solutions—target staphylococci species. C. acnes involves, as a single agent, 27% to 59% of primary PJIs of the shoulder and is more common in revision arthroplasty infections, where it involves up to 70% of isolated bacteria.3,11,12,13,14,15 Recently, C. acnes has been the focus of multiple studies designed to comprehend its role in shoulder infection and prevention.

Cutibacterium acnes

C. acnes is a slow-growing, non-spore-forming, gram-positive anaerobic bacillus. Unlike most of the aforementioned bacteria, which are considered part of the normal superficial skin flora, C. acnes colonizes within the acidic, anoxic environment of sebum-rich hair follicles but is also found among the other normal skin flora bacteria. Hair follicles and sebaceous glands are common around the head, neck, groin, back, and shoulder, especially the axilla. Therefore, C. acnes infection has been identified primarily about the spine and shoulder.13,14,16,17 C. acnes has also been cultured frequently from glenohumeral joint fluid and tissue specimens taken from patients undergoing primary shoulder arthroplasty. This has raised the possibility that C. acnes may play a role in the etiology of glenohumeral osteoarthritis.18 Several other studies yielded at least one positive C. acnes culture from specimen drawn in 20% to 33% of index primary shoulder arthroplasties.19,20 Maccioni and colleagues demonstrated a low rate of positive C. acnes culture by utilizing strict specimen collection techniques. Out of 32 arthroplasty cases, only 3 patients (9.3%) had positive cultures for C. acnes in only a third of the total specimens obtained from these patients. Overall, this low rate of infection (3.125% of the specimens), compared to the rates presented in the aforementioned studies, may reflect high rates of C. acnes contamination, rather than infection.21

C. acnes may gain access to the glenohumeral joint and arthroplasty implants by several routes. First, it may colonize the deep tissues to begin with, prior to any surgery, or can be introduced by prior surgery (as arthroscopy), injection, or by hematogenous spread. Second, it may enter the joint through surgeon’s manipulation of the subdermal layer during the surgical approach. Third, although C. acnes may be naturally present in deep tissue, it could also penetrate as a simple foreign contamination, originating from a contaminated surgical environment or a member of the surgical team.14,19

Multiple studies examined the efficacy of several measures taken throughout the phases of patient care, from preoperative preparations through peri- and intraoperative regimens and the postoperative course. Due to lack of high-quality evidence of some of these strategies, we will extrapolate some data from hip and knee arthroplasties, as well as nonarthroplasty shoulder surgery. In addition, sections of the recently published updated guidelines for the prevention of periprosthetic hip and knee joint infection by the American Academy of Orthopaedic Surgeons (AAOS)22 and the proceedings of the 2018 International Consensus Meeting on the Prevention of Periprosthetic Shoulder Infections will be presented.23


General Patient Characteristics

Matsen and colleagues retrospectively reviewed 342 shoulder revision arthroplasties for deep infections. By analyzing 101 cases of positive C. acnes cultures, several significant patient risk factors were identified: younger age, male sex, primary osteoarthritis as an indication for surgery, and a relatively lower American Society of Anesthesiologists score.24


Several studies examined obesity as a risk factor for shoulder PJI. Jiang and colleagues examined 4796 patients who underwent shoulder arthroplasty and found no significant difference in the short-term incidence of superficial or deep wound infection, wound dehiscence, or total wound complications among four different body mass index (BMI) groups of patients (18.5-25 kg/m2, 25-30 kg/m2, 30-35 kg/m2, and 35 kg/m2 and above). In fact, no significant difference was found for all complications examined in this study.25 Similarly, a meta-analysis performed by Klein and colleagues found no significant difference in the infection incidence in patients with BMI below and above 30 kg/m2.26

This may differ, however, for morbidly obese patients (BMI >40 kg/m2). Theodoulou and colleagues analyzed 10 shoulder arthroplasty studies and found a small increased odds ratio (OR = 1.94) for infection in morbidly obese patients, compared with nonobese patients.27

Statz and colleagues examined the outcome of primary reverse total shoulder arthroplasty (RTSA) in 41 morbidly obese patients and reported two cases of infection that requires revision surgery.28

Diabetes Mellitus

Diabetes mellitus (DM) is a highly prevalent disease with over 380 million affected worldwide in 2013. The increased postoperative risk of infection has already been established in other surgical fields;29 however, literature regarding the risk in shoulder arthroplasty is scarce. Richards and colleagues retrospectively investigated 1186 patients with DM who had undergone primary shoulder arthroplasty. The incidence of PJI among these patients was not significantly increased, compared to 3342 patients without DM (1.1% vs 1.0%, respectively) in this cohort.12

Cancienne and colleagues, however, queried a much larger national database, including 18,729 primary shoulder arthroplasty patients, of which 43% were previously diagnosed with DM (n = 8068). The corrected incidence of superficial wound complication among diabetic patients was 1.4%, compared with 0.9% among nondiabetic patients (OR = 1.22). The corrected incidence of deep infection requiring additional surgery among diabetic patients was 0.7%, compared with 0.4% among nondiabetic patients (OR = 1.47). Furthermore, they found that patients with a threshold HBA1c level greater than 8.0 mg/dL had a significantly higher risk of both wound complications and deep infection requiring surgical intervention.30 These findings emphasize the importance of examining the preoperative levels of HBA1c in patients with DM and support measures to optimize glycemic control preoperatively.


Smoking appears to be a significant risk factor for infection following shoulder arthroplasty. In a cohort including 1834 shoulder arthroplasties (814 smokers and 1020 nonsmokers), Hatta and colleagues identified an increased risk for PJI in smokers.31 Althoff and colleagues reviewed a database of 14,465 patients, which included 1513 smokers. Smokers had a significantly increased risk for wound complications, as well as superficial and deep surgical site infection.32 On the other hand, Morris and colleagues did not find an increased incidence of PJI among smokers undergoing RTSA.33 However, the latter study examined 301 patients, which included only 15 smokers.

Smoking may be a modifiable risk factor as it appears to be an independent risk factor for postoperative infection.31 Therefore, smoking cessation intervention should be considered in patients undergoing shoulder arthroplasty.

Immunosuppressant Therapy

Organ transplant patients with ongoing immunosuppressive therapy have an increased risk of PJI following both hip and knee arthroplasty.34 Data regarding the risk of shoulder PJI are limited. In a retrospective cohort of 30 primary shoulder arthroplasties in 25 solid organ transplant patients, compared with a nontransplant cohort of 120 patients, Hatta and colleagues did not identify a significantly increased risk of infection.35 Notably, Malcolm and colleagues examined perioperative complications of shoulder arthroplasty in organ transplant patients and noticed only a fivefold higher risk of genitourinary infection.36

Human Immunodeficiency Virus

According to the 2019 Joint United Nations Programme on HIV/AIDS data sheet, approximately 2.2 million people are living with human immunodeficiency virus (HIV) in North America and Western and Central Europe.37 Bala and colleagues investigated complications following total shoulder arthroplasty in patients with HIV infection. Retrospectively reviewing a database of 51 million patient records, they identified 2528 patients with HIV infection who underwent shoulder arthroplasty. The authors identified a significant increased risk of shoulder PJI both within 90 days postoperatively and within 2-year postoperatively.38

Asymptomatic Bacteriuria

Asymptomatic bacteriuria refers to the presence of true bacteriuria without any signs and symptoms of urinary tract infection. In a large meta-analysis, including 2043 patients in 11 studies undergoing hip and knee arthroplasty with preoperative asymptomatic bacteriuria, Gomez-Ochoa and colleagues found an increased proportion (twofold) of surgical site infection in patients with asymptomatic bacteriuria, compared with the comparison group. However, the same microorganism was identified in the both sites in only 12.7% of the patients with surgical site infection. As expected, the most common bacteria cultured in urine was Escherichia coli, whereas the most common bacteria causing surgical site infection was gram-positive cocci.39 This makes the causal relationship between asymptomatic bacteriuria and surgical site infection questionable. Therefore, routine urinary screening prior to elective total joint arthroplasty is not recommended, as antibiotic treatment of asymptomatic bacteriuria has not been shown to reduce the risk of PJI.40

Preoperative Intra-Articular Corticosteroid Injection

Garrigues and colleagues at the 2018 International Consensus Meeting on Orthopedic Infections identified four studies that directly investigated the effect of corticosteroids injections to the shoulder prior to shoulder arthroplasty. They found a significant increase in the risk of postoperative PJI following shoulder arthroplasty if performed within 3 months of an injection, compared with noninjection. No significant difference was observed for arthroplasties performed 3 to 12 months after an injection.23 These findings strongly suggest that shoulder arthroplasty should not be performed within 3 months following a corticosteroid injection to the shoulder.

In a recent large, retrospective cohort, Forsythe and colleagues investigated 12,060 patients who received a corticosteroid injection prior to arthroscopic rotator cuff repair. Patients receiving an injection within 1 month prior to surgery had a significantly increased risk of infection following surgery, whereas patients receiving an injection 1 to 3 months, 4 to 6 months, or 7 to 12 months prior to surgery were not at an increased risk of infection postoperatively.41 Although this study focused on arthroscopic rotator cuff repairs, the results provide further support for avoiding corticosteroid injections prior to planned shoulder arthroplasty.

The issue of the frequency of intra-articular corticosteroid injections preoperatively and the risk of PJI has not been reported for shoulder arthroplasty. However, Chambers and colleagues demonstrated and increased risk of PJI (in patients receiving multiple injections within the 12 months preceding hip replacement).42

History of Axillary Lymph Node Dissection

Padegimas and colleagues reviewed 32 shoulder arthroplasty cases in female patients with a history of cancer, who had previously undergone axillary lymph node dissection (ALND). In this retrospective cohort, two patients developed incisional cellulitis and were treated successfully with oral antibiotics. There were no cases of PJI documented. Based on these data, the authors’ conclusion was that previously performed ALND is not a contraindication for shoulder arthroplasty performed through a deltopectoral incision.43

Properly addressing preoperative patient risk factor may play an important role in the reduction of the risk of PJI following shoulder arthroplasty. Published studies suggest that smoking cessation, optimization of glycemic control (>8.0 mg/dL), and perhaps weight reduction (to BMI <40 kg/m2) may diminish the risk of infection. Furthermore, avoiding arthroplasty within 3 months of an intra-articular steroid injection is recommended. Asymptomatic bacteriuria and a history of ALND do not appear to be a risk factor for infection following shoulder arthroplasty.


Preoperative Home Chlorhexidine Wash

Since a substantial percentage of PJIs is caused by normal skin flora, the important question to answer is what are the most effective prophylactic measures that can be
used preoperatively and intraoperatively? Murray and colleagues investigated the effect of preoperative home application of chlorhexidine versus standard soap-and-water shower before shoulder surgery (mostly shoulder arthroscopy). The overall positive culture rate from the posterior shoulder and axilla was significantly reduced for the chlorhexidine group. This reduction was most prominent in coagulase-negative Staphylococcus and Corynebacterium, but not in C. acnes.44 The latter finding was confirmed by Matsen and colleagues, as chlorhexidine was found efficient in reducing coagulase-Staphylococcus and several other bacteria but unsuccessful in reducing C. acnes.45

The complete home chlorhexidine protocol consists of showering with soap and water the evening before surgery, followed by wiping the axilla, shoulder, and ipsilateral chest and back, with a 2% chlorhexidine gluconate-impregnated cloth 1 hour after showering. In the morning of surgery, the patient is instructed to apply the 2% chlorhexidine gluconate-impregnated cloth again in the manner performed the previous night and within 2 hours of departing for the hospital. The authors concluded that it may be valuable to apply home chlorhexidine cloths, especially considering that they are inexpensive and safe.44

Nasal Decolonization

Between 17% and 28% of patients undergoing elective total joint arthroplasty have a positive preoperative nasal screening for S. aureus.46,47,48,49 The identified risk factors for nasal S. aureus colonization are DM, immunosuppression, and renal insufficiency.48 Preoperative Staphylococcus aureus nasal screening and decolonization programs have been proven to lower postoperative surgical site infection rates in cardiac surgery, as well as hip and knee replacement surgery.47,50,51 A single preoperative application of 10% povidone-iodine should be effective and sufficient for short-term suppression (up to 4-6 hours) of S. aureus during the perioperative period.46,52 Furthermore, a cost analysis performed by Stambough and colleagues revealed a significant economic gain for the health system with the use of a nasal screening and decolonization protocol as a result of reduced hospital costs.53

Preoperative Prophylactic Antibiotics

The evidence to support a specific prophylactic perioperative antibiotic regimen is of limited strength. Both AAOS guidelines and the proceedings from the 2018 International Consensus Meeting recommend cefazolin (2 g intravenous [IV] or 3 g if patient weight exceeds 120 kg) 30 to 60 minutes prior to incision as the first line of perioperative prophylaxis.22,23 For patients with history of methicillin-resistant S. aureus (MRSA) infection or colonization, vancomycin (15 mg/kg) within 2 hours prior to incision is recommended, and for patients with proven serious β-lactam allergy, clindamycin (15 mg/kg) within 2 hours prior to incision is the treatment of choice. Redosing should be given every 4 hours for the cefazolin regimen.23

Since the aforementioned regimens provide partial coverage of the most common organisms causing shoulder PJI, doxycycline was examined as an additional prophylactic option. In a randomized controlled study, Rao and colleagues did not observe a significant reduction of C. acnes cultures in the skin, dermis, or glenohumeral joint of shoulder arthroplasty patients receiving a combination of doxycycline and cefazolin, compared to cefazolin only.54

The routine addition of vancomycin to the standard prophylactic regimen is still a matter of debate.55 In addition, patients receiving prophylactic treatment with both cefazolin and vancomycin, have a markedly increased risk of developing acute kidney injury, compared with patients receiving cefazolin alone.56 Therefore, the addition of vancomycin should be reserved only for patients with a high risk of MRSA infection.

Axillary Hair Removal

Preoperative removal of axillary hair has been proposed as a method to decrease the rate of infection in shoulder surgery, especially due to the colonization of C. acnes in hair follicles. However, there is no consensus among surgeons regarding this regimen.57 Saltzman and colleagues examined the efficacy of different skin solutions in 150 patients. They noticed that 25% of the patients had voluntarily shaved their axillary hair as preparation for surgery. No statistically significant difference in the rate of positive cultures was detected between those who had shaved their axilla and those who had not.58

Marecek and colleagues also compared shaving and no shaving regimens and found no significant difference in the burden of C. acnes between shaved and unshaved axillae. Surprisingly, there was a significantly greater total bacterial burden in the shaved group.59

Both studies support the conclusion that preoperative shaving of the axilla has no beneficial effect on decreasing the risk of infection.

Surgical Skin Preparation

Skin cleansing at the operation site with antiseptic solutions is routinely performed in the operating room before draping and skin incision. This skin preparation aims to reduce the microorganism load present on the skin. Multiple studies have evaluated the efficacy of various skin solutions in reducing bacterial load before and after the incision. However, evidence demonstrating statistically significant differences in surgical site infection when comparing different routines is either limited or low.60


It is important to use an effective skin preparation solution to eliminate as much skin flora as possible and prevent seeding of the surgical incision. Saltzman and colleagues studied the native skin flora about the shoulder and the efficacy of different skin solutions. Based on their study, the most isolated bacteria on the skin of the shoulder, prior to preparation, was coagulase-negative staphylococci, C. acnes, and Corynebacterium. This level 1 study compared a solution composed of 2% chlorhexidine gluconate and 70% isopropyl alcohol, a solution containing 0.7% iodophor and 74% isopropyl alcohol, and povidone-iodine scrub and paint (0.75% iodine scrub and 1% iodine paint). The 2% chlorhexidine gluconate and 70% isopropyl alcohol solution was found most effective in eliminating the overall surface bacteria, as manifested by the rate of positive cultures obtained after skin preparation. The 2% chlorhexidine gluconate and 70% isopropyl alcohol solution was as effective as the 0.7% iodophor and 74% isopropyl alcohol solution in eliminating coagulase-negative staphylococci, and both were more effective than the povidone-iodine solution. None of the preparation solutions showed superiority in eliminating C. acnes, since the total number of positive C. acnes cultures obtained after preparation was too small to allow clinical significance.58 Two additional studies failed to show effective eradication of C. acnes with chlorhexidine skin preparations.61,62

Benzoyl Peroxide

Benzoyl peroxide was found effective for the treatment of acne, due to its ability to penetrate the follicles of sebaceous glands in the dermis.63 Therefore, several studies were conducted to test its efficacy in eradicating C. acnes around the shoulder.

A double-blind randomized controlled study performed recently by Van Diek and colleagues examined 30 patients who were screened and tested positive for the presence of C. acnes on the skin of their shoulder. Then they applied benzoyl peroxide gel five times, and skin swabs were cultured. Applying benzoyl peroxide gel was found to effectively reduce the presence of C. acnes by 51.4%, compared with the control group.64

An earlier study performed by Duvall and colleagues, which examined 34 volunteers, demonstrated significant reduction in C. acnes burden on the shoulder after application of benzoyl peroxide 5%.65

Sabetta and colleagues treated patients with a benzoyl peroxide 5% gel to the entire shoulder and axillary region, starting two mornings before scheduled shoulder arthroscopy. They demonstrated a significant reduction in the rate of positive C. acnes cultures from the skin of the anterior deltoid and axilla of the benzoyl peroxide-treated shoulder compared with the nontreated group.66

Only gold members can continue reading. Log In or Register to continue

Jun 23, 2022 | Posted by in ORTHOPEDIC | Comments Off on Prevention of Infection
Premium Wordpress Themes by UFO Themes