Infection prevention strategies in shoulder surgery: A survey of fellowship-trained shoulder surgeons

Abstract

Background

Infection is a major complication with modifiable risk factors. We aimed to report current infection control strategies used by fellowship-trained shoulder surgeons.

Methods

Members of the Shoulder and Elbow Society of Australia were invited to participate in an anonymous electronic REDCap survey. Data is presented as median(IQR) or n(%).

Results

Seventy surgeons with 18 (9–25) years of post-fellowship experience, who performed 250 (150–400) shoulder operations, and 50 (21–79) shoulder replacements annually responded. Staphylococcus aureus screening was performed by 11% pre-arthroscopy versus 44% pre-arthroplasty (26% and 50% gave decolonization regimens respectively). 21% prescribed anti-Cutibacterium acnes regimens pre-arthroscopy, versus 31% pre-arthroplasty. For prophylaxis, 89% used cefazolin and 10% used clindamycin pre-arthroscopy, versus 84% who used cefazolin and 60% vancomycin pre-arthroplasty; 43% did not routinely co-administer cefazolin when prescribing vancomycin. 99% administered antibiotics within one-hour of knife-to-skin. During arthroplasty: 40% enforced personnel limits, 31% enforced entry/exit rules, 50% used topical antibiotics, 27% used liquid-based disinfectants, and 95% changed gloves when handling implants. If cementing, 82% always used antibiotic-loaded cement. Postoperative antibiotics were given by 43% post-arthroscopy and 90% post-arthroplasty (96% and 95% gave first-generation cephalosporins respectively).

Conclusions

There is no consensus on infection prevention in shoulder surgery. Most surgeons are compliant with evidence-based guidelines, where they exist. The data we present are not suggestions of treatment, but rather the contemporary opinion of surgeons discussed in the context of the literature as it stands.

Introduction

Surgical site infection is a highly morbid, potentially lethal, and multidimensionally costly complication, whose treatment demands disproportionate consumption of healthcare resources. ¹ While infections are rare in arthroscopic shoulder surgery, periprosthetic joint infection is the third leading cause of revision in total shoulder arthroplasty.^2,3

A patient's risk of developing a surgical site infection is influenced by host factors and treatment decisions – much of which can be optimised by effective infection control strategies.

Conducting quality randomized trials on surgical infection control strategies is challenging given difficulty of achieving statistical power when studying ‘rare’ diseases, and based on the interventions given or withheld, may cause ethical controversy.^4,5 As a result, guidelines and recommendations often conflict, even between peak organisational bodies.^6–9

Despite the lack of consensus on what constitutes the optimal infection control regimen, we hypothesized that the majority of surgeons would abide by evidence-based guidelines, where they exist. The aim of the present study, therefore, was to provide an up-to-date report on how fellowship-trained shoulder surgeons from a modern healthcare system prevent infection, describing decision making around patient optimisation, to perioperative, and intraoperative practice.

Methods

Fellowship-trained shoulder surgeons from the Shoulder and Elbow Society of Australia (Australian Orthopaedic Association) were invited via email and a WhatsApp message to participate in an online questionnaire using REDCap (Research Electronic Data Capture tools hosted at the University of New South Wales). The electronic invitations containing the study aims and a hyperlink to the survey were sent on 10 May 2024, and the data was extracted on 2 July 2024. All responses were anonymous. Completion of the survey implied consent. This study was undertaken after ethical approval from the Human Research Advisory Panel of the University of New South Wales, which operates in accordance with the National Health and Medical Research Council's National Statement on Ethical Conduct in Human Research (Reference Number: iRECS4552). The survey was designed to take approximately 15 minutes to complete. Responses were summarized qualitatively, or where appropriate, with descriptive statistics.

Survey contents

The questions in this survey were developed to compare shoulder surgeons’ infection control practice to evidence-based guidelines and peer-reviewed data, where available, and to poll surgeons on practices of interest based on expert opinion from senior members of the Shoulder and Elbow Society of Australia, in consultation with specialist input from Infectious Diseases and Antimicrobial Stewardship. Respondents were asked about their years of post-fellowship experience as consultant/attending surgeons, their annual operating volumes, how they optimised patients preoperatively (smoking, glycemic control, nutrition, microbial decolonization regimens), how patients were prepared in the perioperative period (prophylactic antibiotics, clipping hairs, prepping and draping), intraoperative (topical antibiotics, glove change, antimicrobial wash/lavage, wound closure) and postoperative variables (postoperative antibiotics, dressing change). A full copy of the survey can be found in the supplement (Supplement File 1).

Results

Respondent characteristics

Seventy shoulder surgeons responded to the survey. In this cohort, the median (IQR) experience as orthopaedic surgeons was 18 (9–25) years. This group performed a median (IQR) 250 (150–400) shoulder operations and 50 (21–79) shoulder arthroplasty procedures annually.

Among surgeons who performed arthroplasty, 68 declared the number of years of post-fellowship experience they had: 19 (28%) had ≤10 years, 29 (43%) had 11–20 years, 15 (22%) had 21–30 years, and 5 (7%) had ≥30 years of experience respectively. There was a relatively even distribution of case volume between groups, ranging from a median of 45 to 60 arthroplasty cases performed annually. Sixty-eight surgeons declared their annual arthroplasty case volumes: 45 (66%) performed ≤50 cases annually, 16 (24%) performed 51–100 cases annually, and 7 (10%) performed >100 cases annually – the experience between each group ranged from a median of 18 (>100 cases annually) to 15 (51–100 cases annually). A breakdown of arthroplasty-specific responses based on years of experience (Supplement File 2) and annual arthroplasty volume (Supplement File 3) are available in the appendix.

Of the respondents who trained fellows, 14/48 (29%) formally assessed new fellows’ ability to scrub and don personal protective equipment using the aseptic no-touch technique.

Patient optimisation

Smoking

While 50/70 (71%) respondents routinely assessed patient smoking status preoperatively, 29/69 (42%) would change their management for patients who were current smokers, of whom: 23% would not proceed with surgery at all, 8% would not proceed with arthroplasty, 58% would delay surgery and offer counselling/smoking cessation programs, and 12% offered counselling with no hard-line approach to delaying/refusing surgery.

Diabetes and nutrition

59/70 (84%) respondents routinely assessed patient diabetic status preoperatively, of whom 63% defined uncontrolled diabetes as an HbA1C of 7%-8% (whereas 16% of respondents defined uncontrolled diabetes as HbA1C < 7%, and 12% defined it as HbA1C > 8%). For patients who met the respondents’ respective definitions for uncontrolled diabetes, 72% would postpone surgery until the patient achieved glycemic control, 17% would proceed with heightened monitoring, and 11% would proceed with surgery without specific intervention. 91% of respondents would refer patients with uncontrolled diabetes for help with glycemic control: 47% would involve the patient's general practitioner, 49% would refer to a specialist (i.e., endocrinologist), and 11% would involve an allied health worker (i.e., dietician).

Of the 9/67 (13%) of respondents routinely prescribed/recommended nutritional supplementation before surgery, 56% prescribed Vitamin D, 33% prescribed Vitamin C, 11% prescribed a multivitamin pill and 11% prescribed Vitamin K2.

Microbial screening & decolonization

In preparation for arthroscopic surgery, 8/70 (11%) respondents performed methicillin-resistant Staphylococcus aureus (MRSA) screening, and 18/70 (26%) prescribed a Staphylococcus aureus decolonization preoperatively. Specifically, 15/70 (21%) prescribed a whole-body antimicrobial wash, 15/70 (21%) prescribed a Cutibacterium acnes-specific antimicrobial regimen (i.e., phisohex/benzoyl peroxide), and 4/70 (6%) prescribed an antimicrobial nasal ointment (i.e., mupirocin) to arthroscopy patients preoperatively.

In comparison, 31/70 (44%) respondents performed MRSA screening, and 35/70 (50%) prescribed a Staphylococcus aureus decolonization regimen before arthroplasty. Specifically, 27/70 (39%) prescribed chlorhexidine gluconate, 22/70 (31%) prescribed a Cutibacterium acnes-specific antimicrobial regimen, 8/70 (11%) prescribed an antimicrobial nasal ointment to prospective arthroplasty patients.

Corticosteroid injections

Respondents would wait until median (IQR) 12 (6–12) weeks since a patient's last intra-articular corticosteroid injection before performing an arthroscopic shoulder procedure, and 12 (12–12) weeks before performing shoulder arthroplasty.

Preoperative

Antibiotics

Antibiotics offered prior to shoulder arthroscopy and/or shoulder arthroplasty are shown in Table 1. Respondents administered a second dose of cefazolin a median (IQR) of 3 (2–6) hours after the initial dose. When administering vancomycin, 16/28 (57%) always co-administered cefazolin, 9/28 (32%) sometimes co-administered cefazolin, and 3/28 (11%) do not consider co-administering cefazolin.

Table 1.

Antibiotics used for peri-operative prophylaxis in shoulder arthroscopy and arthroplasty.

Antibiotic	Always for shoulder arthroscopy	Sometimes for shoulder arthroscopy	Always for shoulder arthroplasty	Sometimes for shoulder arthroplasty	Never
Cefazolin	62/70 (89%)	4/70 (6%)	59/70 (84%)	1/70 (1%)	0/70 (0%)
Vancomycin	0/70 (0%)	8/70 (11%)	11/70 (16%)	18/70 (26%)	7/70 (10%)
Clindamycin	3/70 (4%)	10/70 (14%)	4/70 (6%)	11/70 (16%)	10/70 (14%)
Gentamicin	1/70 (1%)	3/70 (4%)	2/70 (3%)	5/70 (7%)	18/70 (26%)

With respect to timing of surgical prophylaxis, 17/69 (25%) of respondents administered antibiotics within 15 min of knife-to-skin, 24/69 (35%) did so between 15–30 min, 24/69 (35%) between 30–45 min, 3/69 (4%) at 1 h, and 1/69 (1%) at 2 h.

Regarding the location of administration of surgical prophylaxis, 32/67 (48%) of respondents nominated that it did not matter so long as their stipulated timeframes were maintained, 11/67 (16%) nominated that it did not matter as long as prophylaxis was administered prior to surgery, and 21/67 (31%) insisted that prophylaxis be administered in the anesthetic bay.

Clipping hairs

Prior to arthroscopy, 62/70 (89%) of respondents clipped hairs over the surgical site - 46/70 (66%) doing so in the immediate preoperative period in the anesthetic bay or operating room, 12/70 (17%) doing so within a day of surgery in the preoperative ward, and none recommended clipping at home by the patient. About 32/70 (46%) respondents had the same hair clipping protocol for arthroscopy as they did for arthroplasty.

Prior to arthroplasty, 59/65 (91%) of respondents clipped hairs over the surgical site - 46/65 (71%) doing so in the immediate preoperative period in the anesthetic bay or operating room, 13/65 (20%) doing so within a day of surgery in the preoperative ward, and none recommended clipping at home by the patient.

Surgical scrub

Surgeon scrubbing preferences were similar irrespective of whether they were performing arthroscopy or arthroplasty (Figure 1).

Figure 1.

Surgical scrubbing preferences.

Prepping & draping

The tendency of surgeons to perform a ‘pre-prep’ (performing a scrub of the planned surgical site using antimicrobial solutions, such as those available in dispensers in the scrubbing bay, prior to formal skin prep/before the sterile field is set up) increased when performing arthroplasty and decreased for arthroscopy (Figure 2).

Figure 2.

Proportion of respondents who performed a non-sterile scrub before applying sterile skin preparation.

For arthroscopic shoulder procedures, 43/70 (61%) preferred to use chlorhexidine gluconate as sterile skin preparation, 20/70 (29%) preferred alcohol iodine, 7/70 (10%) preferred isopropyl alcohol, 6/70 (9%) preferred povidone iodine, and 1/70 (1%) preferred hydrogen peroxide. 40/70 (57%) of respondents had the same sterile skin prep protocol for arthroscopy as they did for arthroplasty. For arthroplasty, 42/70 (60%) preferred to use chlorhexidine gluconate as sterile skin preparation, 20/70 (29%) preferred alcohol iodine, 8/70 (11%) preferred isopropyl alcohol, and 5/70 (7%) preferred povidone iodine, and 3/70 (4%) preferred hydrogen peroxide.

Only 3/70 (4%) of surgeons used microbial sealants in preparation for arthroscopic shoulder surgery, and 5/60 (8%) did so prior to shoulder arthroplasty. While 24/70 (34%) of surgeons used antimicrobial incise drapes in arthroscopic cases (14% over the incision sites, 11% at the peripheries of the surgical field, and 9% in the axilla), 61/70 (87%) used them in arthroplasty (69% over the incision 47% at the peripheries of the surgical field, and 50% in the axilla).

30/70 (43%) routinely changed gloves after draping the surgical field before shoulder arthroscopy, versus 61/70 (87%) who would do so before arthroplasty.

Intraoperative

Tranexamic acid

During arthroscopy, 43/70 (61%) of surgeons routinely prescribed tranexamic acid – 39/70 (56%) for hemostasis, and 7/70 (10%) for infection prophylaxis. During arthroplasty, 59/70 (84%) routinely prescribed tranexamic acid – 59/70 (84%) for hemostasis, and 9/70 (13%) for infection prophylaxis.

Wound closure

When closing arthroscopic portals, 44/70 (63%) used non-absorbable sutures (29% also reinforced superficially with elastic/adhesive skin closures), 19/70 (27%) used absorbable sutures (none reinforced using elastic/adhesive skin closures), 2/70 (3%) used elastic/adhesive skin closures exclusively, and 12/70 (17%) used a wound dressing.

Arthroscopy-specific

Regarding ‘re-prepping’ the surgical field during arthroscopy, 30/70 (43%) would re-prep when transitioning from an arthroscopic to an open approach, and 14/70 (20%) would re-prep if the case spanned past a median (IQR) 2 (2–2.5) hours.

64/66 (97%) of surgeons utilized an automated pump to provide continuous pressure during arthroscopy. The remaining 3% utilized a gravity flow-based system by hanging a fluid bag.

For procedures that required suture anchors, 41/66 (62%) utilized a cannula/instrument to prevent the sutures from touching the skin at all, whereas the remaining 38% inserted the anchors and sutures directly through the skin.

Arthroplasty

Personnel and foot traffic restrictions

39/65 (60%) of surgeons did not enforce a limit on the number of personnel in the operating room while performing arthroplasty. Of surgeons who enforced a personnel limit, the median (IQR) number of persons allowed was 7 (6–8) inclusive of scrubbed and un-scrubbed personnel.

3/65 (5%) of surgeons did not allow any entry/exit from the operating room during the case, 17/65 (26%) limited movements, 35/65 (54%) encouraged but did not enforce movement restrictions, and 10/65 (15%) did not have any entry/exit rules at all during arthroplasty cases.

Further intraoperative disinfection

Half the respondents added intra-wound topical antibiotics (i.e., vancomycin powder) intraoperatively, 2/70 (3%) after arthrotomy, 1/70 (1%) coated the prosthesis, 22/70 (31%) before closing the joint capsule after implantation, and 16/70 (23%) applied it to the subcutaneous layer before wound closure.

65/70 (93%) of surgeons performed further disinfection using a liquid-based solution (i.e., iodine swab/bath) disinfection during arthroplasty, most commonly before wound closure (37%), before implanting the prosthesis (31%) and after the initial skin incision (27%) (Figure 3).

Figure 3.

Timing of intraoperative liquid-based disinfection.

Pulse lavage/irrigation was performed by 65/70 (93%) of respondents, most commonly before implanting the prothesis (70%), with 50% opting to do so at each of the following timepoints: after implanting the prosthesis, prior to joint capsule closure, and prior to wound closure (Figure 4).

Figure 4.

Timing of intraoperative pulse lavage/irrigation.

Glove change

62/65 (95%) of surgeons changed gloves when handling the prosthesis.

Antibiotic-loaded cement

If cementing, 53/65 (82%) always used antibiotic loaded cement, 8/65 (12%) use it sometimes (i.e., for complex cases), and 4/65 (6%) never use it.

Surgical drains

21/65 (32%) of surgeons routinely inserted a drain after arthroplasty.

Postoperative

Antibiotics

Following arthroscopy, 30/70 (43%) of surgeons routinely administered postoperative antibiotics, of whom 96% prescribed a first-generation cephalosporin – 25/29 (86%) intravenously, and 75% did not continue to give antibiotics for more than 24 h postoperatively.

Following arthroplasty, 63/70 (90%) of surgeons routinely administered postoperative antibiotics, of whom 95% prescribed a first-generation cephalosporin – 58/60 (97%) intravenously, and 89% did not continue to give antibiotics for more than 24 h postoperatively.

First change of dressings

Assuming no strike-through, surgeons recommended dressings to be first changed a median (IQR) of 8 (2–12) days after arthroscopy, and 10 (7–14) days post-arthroplasty.

Discussion

The present study found that infection control strategies vary widely between experienced shoulder surgeons. Although the majority of surgeons routinely prescribed cefazolin as prophylaxis before arthroscopy (89%) and arthroplasty (84%), a major finding of this survey was that 43% of surgeons who prescribed vancomycin did not routinely co-administer cefazolin.

An analysis of 22,549 major surgical procedures (not limited to orthopaedic surgery) found that the risk of methicillin-sensitive Staphylococcus aureus infection was 2.79 times greater if a patient only had vancomycin than if they only had a beta-lactam for prophylaxis (p < 0.001). ¹⁰ A randomized controlled trial of 4239 arthroplasty patients (n = 30 shoulder) with known methicillin-resistant Staphylococcus aureus were randomized to receive either vancomycin and cefazolin, or cefazolin and normal saline placebo as surgical prophylaxis. Overall, there was no difference in the 90-day risk of surgical site infections between groups. Interestingly, in the knee arthroplasty subgroup, the risk of surgical site infections was significantly greater for patients who had vancomycin and cefazolin, than those who had cefazolin alone (relative risk = 1.52; 95% CI, 1.04–2.23). ¹¹

A retrospective review of the Mayo Clinic's arthroplasty registry, which included 7713 primary shoulder arthroplasty cases (including hemiarthroplasty, anatomic and reverse total shoulder arthroplasty) by Marigi et al. ¹² showed that the risk of all-cause PJI and Cutibacterium acnes PJI respectively, was 2.32 (p = 0.010) and 2.94 (p = 0.028) times greater if vancomycin was used as the sole agent for prophylaxis instead of cefazolin. While their initial study included all patients who had vancomycin-only prophylaxis, a recent follow-up study of the same cohort by the same authors refined their inclusion criteria to compare patients who got ‘complete vancomycin’ (≥30 min between start of vancomycin infusion and skin incision, to allow for adequate tissue concentrations to be achieved) to cefazolin, and found no statistical difference in the rate of infectious complications or reoperations between groups. ¹³ It is important to note however, that 6879 (96%) of patients in their study got cefazolin, whereas only 298 (4%) got complete vancomycin. Notably 163/461 (35%) of all vancomycin patients were excluded because they got ‘incomplete vancomycin’ (<30 min between infusion and incision). The indications for vancomycin were: cephalosporin allergy (95%), MRSA colonization (2%), and ‘other’ (3%) reasons, such as physician preference for high risk patients or previous unsuccessful prophylactic use. ¹³ In contrast to the judicious use of vancomycin in the aforementioned study, 11/70 (16%) of respondents to this survey always gave vancomycin for shoulder arthroplasty. Per the Australian Therapeutic Guidelines, while the primary indication for vancomycin monotherapy is cefazolin allergy, adding vancomycin to cefazolin is permissible in patients at who are colonized with MRSA, or at increased risk of MRSA colonization (e.g., revision arthroplasty). ¹⁴ Therefore, liberal use of narrow-spectrum vancomycin as surgical prophylaxis, especially without co-administration of broad-spectrum beta-lactam antibiotics in the absence of cefazolin allergy or waiting at least 30 minutes between infusion to incision, is hazardous.

Cigarette smoking increases the risk of infection by compromising the host's structural and immunological defenses. ¹⁵ While 71% of respondents routinely assessed smoking status preoperatively, only 42% of them modified management based on smoking status. An analysis of 14,465 total shoulder arthroplasty patients from the American College of Surgeons National Surgical Quality Improvement Program (ACS NSQIP) database found that smoking increased the risk of surgical site infections by 1.9 times (p = 0.042), ¹⁶ whereas an analysis of 18,594 arthroscopic rotator cuff repairs from the same database found that smoking increased the risk of sepsis by 4.7 times (p = 0.021), but not the risk of surgical site infections. ¹⁷ Of surgeons who modified management based on smoking status, 8% would not proceed with arthroplasty specifically, whereas 23% would not proceed with any shoulder surgery. This suggests that surgeons may have been more concerned about the impact of smoking on healing rather than infection in the context of arthroscopic shoulder surgery.^18,19

Persistent hyperglycemia causes immune dysfunction and creates an microenvironment favourable to bacterial proliferation.^20,21 63% of surgeons defined uncontrolled diabetes as a HbA1C of 7%-8%, in keeping with current guidelines, ²² while 16% had a more conservative definition of <7%, while 12% had a more liberal threshold of >8%. We found that 72% of surgeons would postpone surgery until glycemic control was achieved. An ACS NSQIP study of 99,970 non-arthroplasty shoulder surgery patients found that diabetic patients were 1.3 times more likely (p = 0.380) to get infections than non-diabetics. ²³ The evidence for infection risk in diabetic shoulder arthroplasty patients is mixed. An analysis of 113,713 shoulder arthroplasty patients found that diabetics were 1.2 times as likely to develop surgical site infections (p = 0.005) and postoperative infections (p = 0.003) than non-diabetics, ²⁴ whereas another study that reviewed 8819 patients from the Kaiser Permanente Shoulder Arthroplasty Registry showed no association between HbA1c and revision risk, but did find that patients with poor glycemic control had a 1.5 times greater risk of 90-day readmission compared to patients without diabetes (p = 0.032). ²⁵ Ultimately, respondents favoured a more conservative approach toward the decision to operate in patients with poor glycemic control. 91% of surgeons referred patients with uncontrolled diabetes to other medical and allied health colleagues for assistance, in keeping with evidence showing efficacy and cost-effectiveness of multidisciplinary care in diabetes management. ²⁶

Only 13% of respondents routinely prescribed nutritional supplementation preoperatively, although 56% of those who did, prescribed Vitamin D. An analysis of 1674 total shoulder arthroplasty patients found that the rate of all-cause revision was 3.3 times higher in Vitamin D deficient patients were than non-deficient controls (p < 0.0001), ²⁷ while a systematic review showed that Vitamin D deficiency was associated with a greater prevalence of septic versus aseptic revisions (p = 0.016), ²⁸ supporting Vitamin D use in preoperative optimisation.

Staphylococcus aureus is the most commonly implicated organism in orthopaedic surgical site infections. ²⁹ An analysis of 1574 elective shoulder operations found that the risk of infection was 3.6 times greater in open cases than in arthroscopic cases (2.5% vs 0.7%, p < 0.005), and significantly higher in arthroscopic cases with implants than in non-implant arthroscopic cases (0.7% vs 0%, p < 0.05). ³⁰ This finding, in addition to the fact that the majority of surgical site infections of the shoulder are caused by skin commensals, may be why 62% of respondents utilized special instruments to prevent anchors/sutures from touching the skin at all.

A systematic review of 19 studies reported that institution of Staphylococcus aureus screening and decolonization protocol in orthopaedic patients uniformly reduced surgical site infections, wound complications, and was cost effective. ²⁹ It is unsurprising therefore, that the proportion of respondents who performed Staphylococcus aureus screening (44% vs 11%) and who routinely prescribed Staphylococcus aureus decolonization washes preoperatively (50% vs 26%) was greater before arthroplasty than before arthroscopy given the risk of periprosthetic joint infection. However, we note that a greater number of respondents performed Staphylococcus aureus decolonization than screening, suggesting that they perform universal decolonization. Although this goes against current Australian guidelines, which recommends decolonization only if screening returns positive, there is evidence to show that universal decolonization is more cost-effective than targeted decolonization, although this is controversial due the increased risk of antimicrobial resistance.^29,31,32

Cutibacterium acnes is the leading cause of infection in the shoulder. ³³ It is a skin commensal with an affinity for pilosebaceous glands in the subdermal layer, evading most skin preparations, ³⁴ and there is mounting evidence showing that it forms part of the native shoulder microbiome. ³⁵ However, only 31% and 21% of respondents prescribed local Cutibacterium acnes antimicrobial regimens before arthroplasty and arthroscopy respectively. In this study, less respondents gave decolonization regimens for Cutibacterium acnes than for Staphylococcus aureus, despite the theoretically higher risk of Cutibacterium acnes infection. This is likely due to the virulence of Staphylococcus aureus, which is significantly more aggressive than Cutibacterium acnes, which typically causes indolent infections. ³⁶

Most respondents clipped hairs over the surgical site (89% pre-arthroscopy, 91% pre-arthroplasty), all of whom recommended doing so in the hospital on the day of surgery in keeping with current evidence. ³⁷ Chlorhexidine gluconate was the preferred skin preparation agent for both arthroscopy (61%) and arthroplasty (60%), while isopropyl alcohol was the second most common agent used (29% in both groups). Saltzman et al. ³⁸ randomized 150 shoulder surgery patients to receive either Chloraprep (2% chlorhexidine gluconate and 70% isopropyl alcohol), Duraprep (0.7% iodopor and 74% isopropyl alcohol), or povidone iodine scrub and paint (0.75% iodine scrub and 1% iodine paint), and found that positive culture rates were significantly lower when Chloraprep was used compared to povidone iodine (p < 0.0001) and Duraprep (p = 0.010) respectively. However, there was no difference in the rate of Cutibacterium acnes-positive cultures between groups. ³⁸

High foot traffic and personnel volume in the operating room disrupts airflow, reducing the clearance of airborne contaminants and increases the risk of contamination. ³⁹ However, only 40% of respondents enforced a personnel limit in the operating room during arthroplasty, while most surgeons either encouraged, but did not enforce movement restrictions (54%) or limited movement (26%) during the case.

Surgical glove contamination can occur through external contact, or through micro-perforations causing contamination from the inside. To combat this, gloves should be changed every hour, and specifically during high-risk periods: after draping and before handling implants. ⁴⁰ For arthroscopy, 43% of respondents routinely changed gloves after draping. For arthroplasty however, 87% changed gloves after draping, and 95% changed gloves when handling the prosthesis.

Half of all respondents routinely used intra-wound topical antibiotics during arthroplasty, mostly before closing the joint capsule after implantation of the prosthesis (31%), and into the subcutaneous layer before wound closure (23%). A retrospective analysis of 827 total shoulder arthroplasty showed that the incidence of periprosthetic joint infection was lower in patients who had intra-wound topical vancomycin than those who did not (0% vs 3.2%, p < 0.001). ⁴¹ A break-even analysis found that topical vancomycin for infection prophylaxis in shoulder arthroplasty was highly cost effective, further supporting its use. ⁴² If cementing, 82% of respondents always used antibiotic-loaded cement, and 12% used it sometimes. A review of 501 primary reverse total shoulder replacement patients showed that no infections developed where antibiotic-loaded cement was used, versus 3% where it was not used (p < 0.001). ⁴³ In contrast, a retrospective review of 7188 cemented total shoulder replacements (6409 total shoulder replacements for osteoarthritis, 20% got antibiotic cement; 779 reverse total shoulder replacements, 19% got antibiotic cement) from an American registry found no difference in infection risk between patients who got antibiotic cement versus those who got plain cement. ⁴⁴ Therefore, the utility of antibiotic cement in preventing infection in primary shoulder arthroplasty remains undetermined.

The World Health Organisation and Centers for Disease Control and Prevention (CDC) advocate for intraoperative wound irrigation with dilute betadine.^7,45 Accordingly, 93% of respondents performed further intraoperative disinfection using liquid-based solutions during arthroplasty, most commonly after skin incision (27%), presumably to eliminate organisms in the deeper skin layers, before implanting the prosthesis (31%), and before wound closure (37%).

Interestingly, 1/3 of surgeons routinely used a drain following shoulder arthroplasty, despite a lack of evidence to support its use. Trofa et al. randomized patients undergoing anatomic and reverse total shoulder arthroplasty to receive a closed-suction drain (n = 50) or no drain (n = 50). There were no differences in transfusion rates, perioperative complications, hemoglobin or hematocrit levels, length of stay, or cost of stay between groups. ⁴⁶ A retrospective study of 103 reverse total shoulder replacement patients again found no difference in change in hemoglobin or change in hematocrit or complications irrespective of whether a patient received a closed-suction drain (n = 45) or not (n = 58). In this study however, patients who got drains for elective shoulder replacement had longer lengths of stay (2.6 vs 1.8 days, p < 0.01). ⁴⁷ Insertion of drains therefore, is not supported by the evidence.

Postoperative antibiotic prophylaxis is controversial. A systematic review of four randomized controlled trials with 4036 hip and knee arthroplasty patients showed no difference in infection rates between patients who got, versus those who did not get postoperative prophylaxis. ⁶ The CDC and Australian Therapeutic Guidelines do not recommend postoperative antibiotic prophylaxis, citing that there is insufficient evidence to show that a single dose of prophylaxis is as effective as 24 h of prophylaxis.^7,14 This reflects the move toward shorter antibiotic courses. ⁴⁸ However, at the time of writing, the American Association of Hip and Knee Surgeons does not agree with this recommendation. ⁸ Not only was there a high proportion of surgeons administering postoperative antibiotics (post-arthroplasty and 43% post-arthroscopy), but 15% and 11% of those who did continued to give antibiotics past 24 h in oral form following arthroscopy and arthroplasty respectively. One can argue that the risk of additional prophylaxis post-arthroscopy outweighs the benefits, given the significantly lower risk of infection compared to arthroplasty.

The limitation of the present study, as with all survey-based studies, was the potential for response bias, selection bias, and survey fatigue. This survey was not designed to determine the efficacy of the various infection control strategies. Rather, it was designed to provide an up-to-date report of the current practice in an experienced cohort of dedicated shoulder surgeons. We did not distinguish between whether surgeons were responding based on diagnostic arthroscopy or corrective arthroscopy, which may affect antibiotic prophylaxis recommendations. Nonetheless, questions were phrased to ascertain surgeons’ routine practice, of which corrective arthroscopy would account for the overwhelming majority of cases.

Conclusions

Infection control strategies were heterogenous among experienced shoulder surgeons. Most surgeons’ infection control strategies are compliant with evidence-based guidelines, where they exist. However, even guidelines from major bodies differ in their recommendations, such as the ongoing debate on whether or not to prescribe postoperative antibiotics. Nonetheless, this highlights the uncertainty and lack of evidence in some aspects of our practice in relation to infection control. It is important to note that 43% of surgeons did not routinely co-administer beta-lactam antibiotics when giving vancomycin as prophylaxis, which may increase breakthrough Staphylococcus aureus infection risk.