Shoulder surgery preparation: a systematic review and meta-analysis

Abstract

Background

Shoulder infections are an unfortunate and serious complication of surgery. The prevention strategy for infections is multimodal, with a strong reliance on surgical preparation solutions. There is great variability in the use of and effectiveness of surgical preparation solutions; this review aims to identify the most effective methods of shoulder surgical site preparation in the literature.

Methods

A systematic review and meta-analysis was conducted by 2 independent reviewers in accordance with the PRISMA (Preferred Reporting Items for Systematic Review and Meta-Analyses) framework. Articles were screened by title, then abstract, and finally by full text by each clinician independently. A third orthopedic clinician adjudicated any disagreement on the application of the inclusion criteria. Studies included were human clinical studies utilizing skin preparation methods for the shoulder in a simulated or real primary surgery setting. English language databases from 1980 until the first of January 2025 were accessed. The terms searched included “shoulder,” combined with each of the terms “skin,” “wound,” “antibiotic,” “decolonization,” “topical,” “eradication,” “preparation,” and “sterilization.” The results were pooled and then analyzed in subgroups according to the timing of skin preparation, the solution used, and the area of sampling.

Results

The search found 13,154 articles, of which 31 studies were included in the final study and captured a total of 2,115 patients who were cultured for organisms. The lowest rates of culture positivity at the time of surgery were associated with the use of benzoyl peroxide (27.1% culture positive) and prior-to-day-of-surgery preparation solution administration (30.5% culture positive). Patients who received preparation solutions with alcohol numbered 1,551, of which 577 (37.2%) had positive cultures. Patients who received preparation solutions without alcohol numbered 564, with 140 (24.8%) culture positive, P < .0001, chi-square = 28.285. Patients who had shoulder preparation only on the day of surgery numbered 2,067 with a total of 703 (34%) with positive cultures. The patients who did not have day-of-surgery preparation were numbered at 39, with 9 (23%) positive cultures, P = .027, chi-square = 4.89.

Conclusion

There is a wide range of available surgical preparation solutions to attempt to prevent day-of-surgery culture positivity. However, even the most robust regiments still have a substantially high culture positivity rate at the time of surgery. Further trials are warranted to unify protocols for the management of perioperative shoulder preparation.

There is a range of shoulder surgery procedures used to treat shoulder pathology, all of which carry the risk of a clinically significant infection.^16,25 As technology in bacterial sampling has improved, evidence suggests that the joints and deep tissue of native shoulders may exhibit bacterial activity when sampled in sterile environments.²⁵ Clinically significant infections are the most concerning facet of bacterial activity in the shoulder. Of the many possible bacterial origins, the most common cause of a deep surgical infection is the skin flora.²³ Surgical infections may take up to two years to develop clinical signs, and infection may be precipitated by dysfunction of the biological events that normally occur following surgical wound closure.⁹ The hip and knee literature have well established guidelines for diagnosing infection with the Musculoskeletal Infection Society criteria, which was modified by the International Consensus Meeting (ICM) on Musculoskeletal Infection in 2018.²⁰ A new category of infectious status is the “unexpected positive culture” result, where a shoulder arthroplasty is revised without signs or symptoms of infection, but the intraoperative cultures return with positive organism growth.³⁵

The etiology of shoulder surgical infection is difficult to study, as infection is a relatively uncommon occurrence after shoulder surgery.⁷ This may lead to the failure of even large-scale randomized studies to demonstrate a meaningful difference between treatment groups.⁵ Deep shoulder periprosthetic infection for shoulder arthroplasty has devastating consequences for the host, the surgeon, and the healthcare system. The conventional treatment involves weeks of intravenous antibiotic administration as well as a one-stage (one operation) or two-stage (2 operations) revision shoulder arthroplasty.^20,22 The subclinical nature of Cutibacterium acnes infections means that they may present with alternate sequalae such as rotator cuff repair failure.¹² The most commonly isolated pathogens in postoperative shoulder infections are C acnes (38.9%), Staphylococcus aureus (14.8%), Staphylococcus epidermidis (14.5%), and coagulase-negative Staphylococcus species (14%).⁹

The optimum timing and dosing of topical agents to reduce the burden of colonizing shoulder organisms is unclear.^21,23,29 Standard skin preparation can be conducted in 3 stages: at-home treatment, presurgical treatment, and sterile surgical preparation at the time of surgery.³ The existing literature suggests that alcohol in combination with chlorhexidine or iodine solutions is more effective at reducing bacterial load than solutions without alcohol.⁷

Clinically apparent shoulder infection is a substantial problem in orthopedic surgery, and the conventional prevention methods for this occurrence are fallible. This review is driven by the need for a contemporary assessment of the published skin preparation treatment options and their rates of success or failure in reducing organism colonization for the sterile surgical field. The focus of this systematic review is to establish the efficacy of surgical preparation solutions in eradicating commensal shoulder bacteria by the time of surgery. This includes preparations administered in the days prior to surgery, on the day of surgery, and at the time of surgery. We had 2 prospective hypotheses for this review; the first was that application of preparation solutions prior to the day of surgery as well as on the day of surgery would yield superior decolonization. The second was that the use of multimodal agents would result in reduced rates of bacterial colonization on the shoulder.

Materials and methods

Preferred Reporting Items for Systematic Review and Meta-Analyses

The authors performed this systematic review in accordance with the PRISMA (Preferred Reporting Items for Systematic Review and Meta-Analyses) framework (Fig. 1).

Figure 1.

PRISMA summary of our study review process. PRISMA, Preferred Reporting Items for Systematic Review and Meta-Analyses.

Search strategy

Two authors who are orthopedic clinicians conducted independent literature searches of institutionally accessed medical research databases. These were Medline, Embase, SCOPUS, and the Cochrane Register. The search range included all articles from January 1980 until the first of January 2025. The search utilized the term “shoulder,” combined with each of the terms “skin,” “wound,” “antibiotic,” “decolonization,” “topical,” “eradication,” “preparation,” and “sterilization.” The excluded studies were those that were animal and cadaveric studies, studies written in languages other than English, systematic reviews and meta-analyses, patients under the age of 18 years, and studies of reoperations. Included studies were those on human patients who underwent real or simulated shoulder surgery for the first time, with the use of at least one type of surgical preparation solution, with microbial assessment reported.

Data extraction

Two clinicians independently reviewed the articles, with a third clinician consulted for differences of opinion. Interrater reliability was assessed using Cohen's kappa. Each clinician collected the following itemized data: author name, year of publication, sample size, study type, follow-up duration, participant number and demographics, type of surgery, surgical preparation intervention type and duration, antibiotic administration, organism results, and clinical infection results.

Quality of evidence assessment

The quality of evidence was assessed and categorized into high, moderate, low, or very low quality evidence for each article using the Grading of Recommendations Assessment, Development and Evaluation system.

Statistical analysis

Data were subdivided according to the timing of exposure of each patient pool to a given preparation solution. Preparation application timing was grouped into: day prior to surgery, day of surgery, or intraoperative therapy. We then conducted an analysis of proportions based on the number of patients with positive cultures divided by the total number of patients. The subcategorized microbiology and demographic results were interpreted with the aid of forest plots. The chi-squared statistical test was used to measure the independence of observed association. The leave-one-out approach was used to indicate the relative robustness of results for sensitivity analysis of outcomes. We used a random-effects model and fixed-effects model according to heterogeneity (I2 > 50%) and homogeneity (I2 < 50%), respectively. Statistical significance was set at 5% (P = .05). Statistical analysis was performed using SPSS version 27 (IBM Corp., Armonk, NY, USA) and STATA version 15 (StataCorp LLC, College Station, TX, USA).

Results

Study selection

Our search strategy generated 13,154 articles, which was reduced to 132 abstracts, which were read and then reduced to 58 full-text articles. The 58 full-text articles were read completely and further reduced to 31 studies included in the final study (Fig. 1, Table I).

Table I.

Demographic table of all included studies.

Study	Comments	Country of origin	Level of evidence, study design	Total number male	Total number female	Total	Culture or NGS or both	Patients with positive cultures	Type of skin prep	Prep specifics	Prep timing (day of surgery, 1 d prior, 2 d prior)	Perioperative IV antibiotics given	Clinically significant infections (number of patients)
Saltzman 200923	Shoulder surgery	USA	1	N/A		50	Culture (swabs, aerobic + anaerobic)	3	Chlorhexidine	ChloraPrep (2% chlorhexidine gluconate and 70% isopropyl alcohol)	Day of surgery	Y	0
						50		9	DuraPrep	DuraPrep (0.7% iodophor and 74% isopropyl alcohol)	Day of surgery	Y	0
						50		15	Povidone-iodine	Povidone-iodine 0.75% scrub, 1.0% paint	Day of surgery	Y	0
Dizay 20179	Arthroscopic shoulder surgery	USA	II, prospective cohort design; treatment study	43	22	65	Culture (swab)	8	BPO	Benzoyl peroxide 5% and clindamycin 1.2% gel	Night prior to surgery	Y	0
Namdari 201721	Shoulder arthroscopy	USA	1, RCT	74	0	74	Culture (swabs)	16	Oral	Oral doxycycline 100 mg daily for 7 d	7 d course starting 7 d before surgery	Y	0
Stull 202024	Shoulder arthroscopy	USA	1, RCT	70	0	70	Culture (swabs)	12	Hydrogen	3% hydrogen peroxide soaked gauzes prior to prep	At surgery	Y	0
				70	0	70		24	Povidone-iodine	7.5% povidone-iodine	At surgery	Y	0
Moor 202126	Open shoulder surgery deltopectoral approach	Switzerland	1, RCT	34	21	55	Culture (swabs)	15	Povidone-iodine	povidone-iodine solution	At surgery	Y	0
				17	39	56		14	Povidone-iodine	povidone-iodine solution - twice used	At surgery - second application after cut	Y	0
Hong 202321	Shoulder arthroscopy	Taiwan	1, RCT	12	20	32	Culture and PCR	13	Preop soap and water	Soap	For 3 d prior to surgery	Y	0
				17	12	29		14	Preop chlorhexidine 2%	Chlorhexidine 2%	For 3 d prior to surgery only on shoulder	Y	0
				11	18	29		5	Preop chlorhexidine 2%	Chlorhexidine 2%	For 3 d prior to surgery whole body	Y	0
Symonds 202234	Shoulder arthroplasty	Australia	II, RCT; treatment study	22	11	33	Culture (swabs)	5	BPO	BPO 5%	BD for two days prior to surgery with 5 applications total	Y	Not specified
				15	16	31		5	BPO + clindamycin	BPO 5% and clindamycin 1%	BD for two days prior to surgery with 5 applications total	Y	Not specified
				24	11	35		9	pHisoHex	pHisoHex (1% triclosan, sodium benzoate 5 mg/ml, benzyl alcohol 5 mg/ml)	BD for two days prior to surgery with 5 applications total	Y	Not specified
Chalmers 20193	Primary arthroplasty with hydrogen peroxide (n = 90) prep vs. without (n = 93), in addition to standard skin prep	USA	II, prospective cohort	16	15	31	Culture (swabs, aerobic + anaerobic)	14	70% ethyl alcohol + 2x ChloraPrep applicators	Step 1. 70% ethyl alcohol Step 2. 2x ChloraPrep applicators	Day of surgery	Y	Not specified
				13	17	30		11	70% ethyl alcohol +3% hydrogen peroxide + 2x ChloraPrep applicators	Step 1. 70% ethyl alcohol Step 2. 3% hydrogen peroxide Step 3. 2x ChloraPrep applicators	Day of surgery	Y	Not specified
Cotter 20215	Volunteers-no surgery	USA	Level I, randomized controlled trial	20	0	20	Cultures + PCR + Sanger sequencing	20	Benzoyl peroxide + Chlorhexidine gluconate + isopropyl alcohol	5% benzoyl peroxide gel to axilla, anterior shoulder, and lateral shoulder after showering in the morning and evening for a total of 5 topical treatments spanning 5 d +2% chlorhexidine gluconate +70% isopropyl alcohol	Day of surgery	NA	NA
				20	0	20		20	Blue light therapy + chlorhexidine gluconate + isopropyl alcohol	415 nm of blue light at 40 mW/cm2-centered over deltopectoral interval at 5 cm from the clean, dry skin, administered for 23 min to reach 55.2 J/cm2. Omnilux Blue +2% chlorhexidine gluconate +70% isopropyl alcohol	Day of surgery	NA	NA
				20	0	20		16	Benzoyl peroxide + blue light therapy + chlorhexidine gluconate + isopropyl alcohol	5% benzoyl peroxide gel to axilla, anterior shoulder, and lateral shoulder after showering in the morning and evening for a total of 5 topical treatments spanning 5 d 415 nm of blue light at 40 mW/cm2 - centered over deltopectoral interval at 5 cm from the clean, dry skin, administered for 23 min to reach 55.2 J/cm2.Omnilux Blue +2% chlorhexidine gluconate +70% isopropyl alcohol	Day of swab (no surgery)	NA	NA
Crutcher 20246	Volunteers - no surgery 18 males (36 shoulders) - one side intervention vs. one side control	USA	Level I, randomized controlled trial	18	0	18	Culture before and 60 min after prep	12	Chlorhexidine gluconate Isopropyl alcohol	2% chlorhexidine gluconate and 70% isopropyl alcohol	Day of swab (no surgery)	N/A	N/A
				18	0	18		12	Chlorhexidine gluconate Isopropyl alcohol	2% chlorhexidine gluconate and 70% isopropyl alcohol	Day of swab (no surgery)	N/A	N/A
Dizay 20179	Shoulder arthroscopy patients given topic benzoyl	USA	Level II, prospective cohort	65	43	108	Cultures (deep)	2	BPO/C gel every night before surgery Chlorhexidine 4% scrub + chlorhexidine gluconate + isopropyl alcohol	Clindamycin phosphate/Benzoyl peroxide 1.2%/5% gel every night, depending when they presented to PAC, which ranged from 1 to 10 d preop (for average 2.3 d/applications preop) 4% chlorhexidine scrub 2% chlorhexidine gluconate 70% isopropyl alcohol	Day of surgery	Y	0
Grewal 202116	Peroxide group	USA	I, RCT	16	14	30	Culture	6	Hibiclens wash night before, and morning of surgery Chlorhexidine gluconate cloth ChloraPrep Hi-Lite Orange Hydrogen peroxide (after electrocautery of dermal layer)	Hibiclens = 4% chlorhexidine gluconate solution 2 × 2% chlorhexidine gluconate cloths 2× BD ChloraPrep Hi-Lite Orange (2% chlorhexidine gluconate with 70% isopropyl alcohol) 3% hydrogen peroxide after electrocautery of dermal layer	1 d prior and day of surgery	Y
	Control group			13	17	30		5	Hibiclens wash night before, and morning of surgery Chlorhexidine gluconate cloth ChloraPrep Hi-Lite Orange	Hibiclens = 4% Chlorhexidine gluconate solution 2 × 2% chlorhexidine gluconate cloths 2× BD ChloraPrep Hi-Lite Orange (2% chlorhexidine gluconate with 70% isopropyl alcohol)	1 d prior and day of surgery	Y
Hancock 201819	RCT in healthy volunteers - intervention group	NZ	I, RCT	22	0	22	Culture	4	Benzoyl peroxide Chlorhexidine/alcohol	1 × 5% benzoyl peroxide 50 mg/g 1 × 2% chlorhexidine/alcohol (Soluprep 2% w/v chlorhexidine gluconate and 70% v/v isopropyl alcohol)	Day of testing (no surgery, volunteers)	N	N/A
				22	0	22		3	Chlorhexidine/alcohol	2 × 2% chlorhexidine/alcohol (Soluprep 2% w/v chlorhexidine gluconate and 70% v/v isopropyl alcohol)	Day of testing (no surgery, volunteers)	N	N/A
Heckmann 201920	Upper backs of 12 volunteers, 4 quadrants randomized to different preps and dermal punch biopsy taken	USA	II, prospective cohort	10	2	4	Culture	4	None (negative control)	No prep	Day of surgery	N	N/A
				10	2	1		1	Benzoyl peroxide	5% topical benzoyl peroxide	Day of surgery	N	N/A
				10	2	2		2	Topical clindamycin	1% topical clindamycin	Day of surgery	N	N/A
				10	2	2		2	Topical benzoyl peroxide + clindamycin	5% topical benzoyl peroxide + 1% topical clindamycin	Day of surgery	N	N/A
Horneff 202322	Topical 5-ALA skin solution activated by blue light photodynamic therapy vs. standard preop skin prep	USA	II, prospective cohort			25	Culture	12	ALA HCl skin solution activated by blue light photodynamic therapy Followed by “preprep” chlorhexidine gluconate and povidone iodine prep. Sterile prep with Hydrogen peroxide, chlorhexidine gluconate, and isopropyl alcohol	Aminolevulinic acid hydrochloride 20% sticks applied after showering on the day of surgery. Photodynamic therapy was applied for 1000s (16 min 40 s) to provide 10 J/cm2 light dose to activate the solution 2% chlorhexiidne gluconate clothes (Sage Products) 7.5% povidone-iodine solution scrub brushes (Betadine) 3% hydrogen peroxide gauzes ×5 2% chlorhexidine gluconate and 70% isopropyl alcohol ×2 26 mL stick applicators (ChloraPrep)	Day of surgery	Y	N/A (none reported - ?0)
						26		16	Control	2% chlorhexiidne gluconate clothes (Sage Products) 7.5% povidone-iodine solution scrub brushes (Betadine) 3% hydrogen peroxide gauzes ×5 2% chlorhexidine gluconate and 70% isopropyl alcohol ×2 26 mL stick applicators (ChloraPrep)	Day of surgery	Y	N/A (none reported - ?0)
Hsu 202013	Adult male shoulder arthroplasty patients randomized to preop CHG vs. preop BPO 100% +ive skin surface swabs - taken before sterile prep 61% +ive dermal edge swabs (CHG) vs. 46% (BPO) - taken after sterile prep	USA	I, RCT	25	0	25	Culture	14	4% chlorhexidine gluconate wash (night before + morning of) 2% chlorhexidine + 70% isopropyl alcohol (sterile)	4% chlorhexidine gluconate wash (night before + morning of) 2% chlorhexidine + 70% isopropyl alcohol (sterile)	1 d prior	Y	N/A
				24	0	24		11	10% benzoyl peroxide soap (night before + morning of) 2% chlorhexidine + 70% isopropyl alcohol (sterile)	10% benzoyl peroxide soap (night before + morning of) 2% chlorhexidine + 70% isopropyl alcohol (sterile)	1 d prior and day of surgery	Y	N/A
Hsu 202315	10 shoulders from 5 male volunteers	USA	V, basic science	10	0	10	Culture	20% at 3 min post-prep 50% at 30 min post-prep 88% at 60 min post-prep 88% at 120 min post-prep 100% at 240 min post-prep	Chloraprep	2% chlorhexidine gluconate and 70% isopropyl alcohol	Day of surgery	N	N/A
Hudek 201628	Intraoperative samples from shoulder surgery patients	USA	V, basic science	70	42	112	Culture	43	Alcohol	Skin rubbed with 6 consecutive sterile swabs drained in 100 mL of alcoholic disinfectant (Kodan Tinktur forte gefarbt 45.0 g 2-propanol, 10.0 g 1-propanol, 0.20 g biphenyl-2-ol; Schulke & Mayr GmbH, Norderstedt, Germany)	Day of surgery	N (Abx withheld until samples taken)	N/A
Jalali 202029	Pilot study in healthy volunteers - phase I = 10% PCA vs. IPA; phase II = 17% PCA vs. IPA	USA	II, prospective cohort			11	Culture (PCR and DNA sequencing if culture + ive)	Aerobic = 1 Anaerobic = 2	10% protocatechuic acid in 70% isopropyl alcohol		Day of surgery	N	N/A
						12		Aerobic = 1 Anaerobic = 2	17% protocatechuic acid in 70% isopropyl alcohol with propylene glycol and peppermint oil		Day of surgery	N	N/A
						23		Aerobic = 9 Anaerobic = 14	70% isopropyl alcohol		Day of surgery	N	N/A
Kim 202417	RCT - skin incision with scalpel vs. electrocautery (prep same between groups)	USA	I, RCT	48	16	64	Culture	10	Chlorhexidine gluconate showers Topical hydrogen peroxide, povidone iodine, isopropyl alcohol, and DuraPrep	2% chlorhexidine gluconate showers - night before and morning of surgery Topical 3% hydrogen peroxide, 7.5% povidone iodine, 70% isopropyl alcohol, and DuraPrep (iodine povacrylex and isopropyl alcohol)	1 d prior	Y	1
Koh 201616(p201)	30 consecutive primary shoulder arthroplasty patients swabbed for C acnes Swab 1 = 4% chlorhexidine gluconate wash (24 h preop) Swab 2 = 4% chlorhexidine gluconate scrub (pre-prep with sponge on the day of surgery) Swab 3 = 2x applications of Chloraprep (2% CHG in 70% IPA) Swabs 4-5 = deep cultures Swab 6 = control - superficial skin medial to incision was swabbed after layered wound closure but before final subcuticular closure	NZ	V, basic science (but it's a prospective clinical study!)	7	23	30	Culture Swab 1 = 4% chlorhexidine gluconate wash (24 h preop) Swab 2 = 4% chlorhexidine gluconate scrub (pre-prep with sponge on day of surgery) Swab 3 = 2x applications of Chloraprep (2% CHG in 70% IPA) Swabs 4-5 = deep cultures Swab 6 = control - superficial skin medial to incision was swabbed after layered wound closure but before final subcuticular closure	S1 (after decolonizing wash but before pre-prep) = 14 S2 (after pre-prep) = 12 S3 (after sterile prep) = 8 S4-5 (deep cultures) = 14	S1 = 4% chlorhexidine gluconate wash (24 h preop) S2 = 4% chlorhexidine gluconate scrub (pre-prep with sponge on day of surgery) S3 = 2x applications of Chloraprep (2% CHG in 70% IPA) S4-5 = deep cultures S6 = control - superficial skin medial to incision was swabbed after layered wound closure but before final subcuticular closure	S1 = 4% chlorhexidine gluconate wash (24 h preop) S2 = 4% chlorhexidine gluconate scrub (pre-prep with sponge on day of surgery) S3 = 2x applications of Chloraprep (2% CHG in 70% IPA) S4-5 = deep cultures S6 = control - superficial skin medial to incision was swabbed after layered wound closure but before final subcuticular closure	1 d prior	Y	N/A
Maclean 201928	Open shoulder surgery patients Swab 1 = following skin incision Swab 2 = 2-5 min after aqueous CHG to dermis Swab 3 = 60 min postapplication	AUS	II or IV?	22	28	50	Culture	21	Pre-prep = 200 mL of Chloraprep (2% CHG in 7% IPA) Sterile prep with second application of Chloraprep (2% CHG in 7% IPA) Ioban	Pre-Prep = 200 mL of ChloraPrep (2% CHG in 7% IPA) Sterile prep with second application of Chloraprep (2% CHG in 7% IPA) Ioban	Day of surgery	Y	N/A
Matsen 202029	Primary shoulder arthroplasty patients Swab 1 = preoperative clinic Swab 2 = skin at surgery after home CHG shower but prior to sterile prep Swab 3 = dermis after incision of prepped skin	USA	III	44	22	66	Culture Swab 1 = preoperative clinic Swab 2 = skin at surgery after home CHG shower but prior to sterile prep Swab 3 = dermis after incision of prepped skin	Dermal cultures (swab 3) Total = 18 Cutibacterium = 16 Coagulase-negative staphylococcus = 2	4% CHG solution (decolonizing wash given at preop clinic) 2% CHG in 70% IPA (sterile prep)	4% CHG solution (decolonizing wash given at preop clinic) 2% CHG in 70% IPA (sterile prep)	1 d prior	Y	N/A
Moor 202126	Open shoulder surgery with vs. without subcutaneous povidone-iodine solution	Switzerland	I, RCT	17	21	38	Culture (deep)	Total = 15 C acnes = 10	Beta-septic Ioban	Beta-septic Ioban	Day of surgery	Y	0
				31	39	70		Total = 14 C acnes = 7	Beta-septic Ioban Betadine (povidone-iodine) wash to subcutaneous layer	Beta-septic Ioban Betadine (povidone-iodine) wash to subcutaneous layer	Day of surgery	Y	0
Murray 201125	Shoulder surgery with vs. without preoperative 2% CHG cloth shower vs. normal soap and water shower before surgery	USA	I, RCT	25	25	50	Cultures (skin only, posterior shoulder + axilla)	All organisms equals; 47 Coag-negative staph = 35 C acnes = 29	Shower with soap and water morning of surgery Chloraprep (2% CHG and 70% IPA) Ioban	Shower with soap and water morning of surgery ChloraPrep (2% CHG and 70% IPA) Ioban	Day of surgery (normal soap and water not counted as prep)	Y	0
				36	14	50		All organisms = 33 Coag negative staph = 15 C acnes = 15	2x chlorhexidine cloth 1 h after showering with soap and water evening before surgery ChloraPrep (2% CHG and 70% IPA) Ioban	2x Chlorhexidine cloth 1 h after showering with soap and water evening before surgery ChloraPrep (2% CHG and 70% IPA) Ioban	1 d prior	Y	0
Phadnis 201628	Open shoulder surgery patients (prospective) Swab 1 (pre-prep) = skin before prep but after IV abx Swab 2 = skin after Chloraprep (70% alcoholic chlorhexidine) Swab 3 = dermis Dermal biopsy	AUS	Basic science study	30	20	50	Culture	All organisms were C acnes Preprep = 21 Postprep = 6 Dermal swab = 24 Dermal biopsy = 17	Chloraprep (2% CHG and 70% IPA)	ChloraPrep (2% CHG and 70% IPA)	Day of surgery	Y	1
Polishchuk 201232	Healthy volunteers - skin cultures over shoulders and knees pre- vs. post-prep with ethyl chloride spray Shoulder data extracted only	USA	Basic science study	N/A	N/A	30	Culture	Before skin prep = 17 After skin prep (before ethyl chloride spray) = 2 After ethyl chloride spray = 1	Ethyl chloride spray (Gebauer's Ethyl Chloride Medium Jet Stream Spray) 70% isopropyl alcohol	Ethyl chloride spray (Gebauer's Ethyl Chloride Medium Jet Stream Spray) 70% isopropyl alcohol	Day of surgery	N	N/A
Sabetta 201533	First-time arthroscopic shoulder surgery patients got 5% BPO cream 48 h before surgery. Swabs taken after sterile prep	USA	IV, case series	23	27	50	Culture	(ie pooled all cultures/400 cultures from 50 patients and reported in terms of cultures not patients) After BPO but before skin prep (50 samples from each site) Anterior deltoid = 8/50 Contralateral anterior deltoid (control) = 16/50 Axilla = 4/50 Contralateral axilla = 14/50 After skin prep After skin Prep = 25/400 Control (cotton swab exposed to air) = 16/400	BPO	5% benzoyl peroxide cream 48 h before surgery Chloraprep (3x applications of 2% CHG and 70% IPA)	2 d prior	Y	0
Scheer 201834	Healthy volunteers - randomized to preoperative topical treatment at home with 5% BPO vs. 4% CHS	Sweden	II			20	Culture 1. Before any treatment 2. Before sterile prep 3. 2 min after sterile prep 4. 120 min after sterile prep	Sample 1 = 19 Sample 2 = 13 Sample 3 = 1 Sample 4 = 4	5% BPO applied to a 5 cm strip of skin on the morning and evening for 2 d before the trial. The fifth and final application was the morning of the trial day			N	N/A
						20		Sample 1 = 19 Sample 2 = 16 Sample 3 = 7 Sample 4 = 10	4% chlorhexidine soap was applied the day before the trial with 2 showers (min 2 h between) using 2 sponges each, and on the morning of the trial with 2 more sponges			N	N/A
Scheer 202135	Open elective shoulder surgery randomized to BPO or soap	Sweden	II, RCT (lower quality)	40	15	55	Culture Swab C = preoperative area (contralateral) Swab 1 = preoperative area Swab 2 = immediately before incision Swab 3 = after skin closure	Swab C = 55 Swab 1 = 50 Swab 2 = 12 Swab 3 = 34	5% BPO applied to a 5 cm strip of skin on the morning and evening for 2 d before the trial. The fifth and final application was the morning of the trial day 0.5% chlorhexidine solution in 70% ethanol (sterile prep)	5% BPO applied to a 5 cm strip of skin on the morning and evening for 2 d before the trial. The fifth and final application was the morning of the trial day 0.5% chlorhexidine solution in 70% ethanol (sterile prep)		Y (only after swab 3 was taken)	2
				23	22	45		Swab C = 42 Swab 1 = 26 Swab 2 = 2 Swab 3 = 21	Regular soap 0.5% chlorhexidine solution in 70% ethanol (sterile prep)	Regular soap 0.5% chlorhexidine solution in 70% ethanol (sterile prep)		Y (only after swab 3 was taken)	3
Unterfrauner 202236	Open shoulder surgery patients randomized to 7-d application of BPO cream vs. no cream	Switzerland	I, RCT	12	18	30	Culture Sample 1 = skin swab before therapy (1 week preop) Sample 2 = skin swab after therapy (day of surgery, before prep) Sample 3 = subcutaneous tissue sample Sample 4 = capsular tissue sample	C acnes Sample 1 = 18 Sample 2 + 3+4 = 8/18 Sample 2 = 8/18 Sample 3 = 1/18 Sample 4 = 0 S epidermidis Sample 1 = 19 Sample 2 + 3+4 = 8/19 Sample 2 = 8/19 Sample 3 = 1/19 Sample 4 = 0 S aureus Sample 1 = 3 Sample 2 + 3+4 = 1/3 Sample 2 = 1/3 Sample 3 = 0 Sample 4 = 0	50 mg/g benzoyl peroxide and 20 mg/g miconazole nitrate (ACNE Crème plus Widmer) for 7 nights preceding surgery Betaseptic (povidone-iodine, isopropyl alcohol, and ethanol)	50 mg/g benzoyl peroxide and 20 mg/g miconazole nitrate (ACNE Crème plus Widmer) for 7 nights preceding surgery Betaseptic (povidone-iodine, isopropyl alcohol, and ethanol)	7 d prior	Y	N/A
				15	15	30		C acnes Sample 1 = 19 Sample 2 + 3+4 = 16/19 Sample 2 = 13/19 Sample 3 = 2/19 Sample 4 = 4/19 S epidermidis Sample 1 = 21 Sample 2 + 3+4 = 11/21 Sample 2 = 11/21 Sample 3 = 0 Sample 4 = 0 S aureus Sample 1 = 6 Sample 2 + 3+4 = 1/6 Sample 2 = 1/6 Sample 3 = 0 Sample 4 = 0	Betaseptic (povidone-iodine, isopropyl alcohol, and ethanol)	Betaseptic (povidone-iodine, isopropyl alcohol, and ethanol)	Day of surgery	Y	N/A

IV, Intravenous; NGS, next generation sequencing; BPO, benzoyl peroxide gel; RCT, randomised controlled trial; PCR, polymerase chain reaction; BD, bis in die (twice daily); PAC, pre-admission clinic; ALA HCI, aminolevulinic acid hydrochloride; CHG, chlorhexidine gluconate; IPA, isopropyl alcohol; PCA, phenolic compound protocatechuic acid; CHS, chlorhexidine solution.

Study characteristics

The demographics of each included study are presented in Table I. A total of 2,115 patients were found in the 31 eligible studies. Two hundred ten of these patients participated in simulation studies on healthy volunteers where no surgery was performed.

Culture methods

Thirty-one studies conducted culture by conventional laboratory analysis with either blood agar, chocolate agar, or thioglycolate broth for an incubation time of 2 to thirty days. Three studies utilized polymerase chain reaction and DNA analysis in addition to traditional culture methods.

Intravenous antibiotic administration methods

Nine of 31 studies did not utilize intravenous antibiotic administration. One study utilized antibiotic administration after the final swab had been taken, which in effect gave results without antibiotic influence. All 9 studies were conducted on patients who did not undergo surgery.

Microbial results

All patients

Of the 2,115 patients included in the study, there were 717 (33.7%) patients who had positive results for organisms on the day of testing (Fig. 2).

Figure 2.

Forest plot of culture positive results for all studies.

Timing of preparation

Prior to surgery administration

Of the studies that analyzed patients receiving topical preparations prior to the day of surgery, there were 829 patients who had a total of 253 (30.5%) positive cultures. The patients who did not have prior-to-day-of-surgery preparation were numbered at 1,286 patients who had a total of 464 (36.1%) positive cultures, P = .008, chi-square = 6.96 (Fig. 3).

Figure 3.

Forest plot of culture positive rates for patients treated with prior to surgery solutions.

Day of surgery administration

Of the studies that analyzed patients receiving topical preparation on the day of surgery, there were 2,067 patients who received skin preparation, with a total of 703 (34%) with positive cultures. The patients who did not have day-of-surgery preparation numbered 39, with 9 (23%) positive cultures, P = .027, chi-square = 4.89 (Fig. 4).

Figure 4.

Forest plot of culture positive rates for patients treated with day of surgery preparation solutions.

Prior to surgery and day of surgery administration

Of the studies that analyzed patients receiving day of surgery in addition to prior-to-day-of-surgery preparation, there were 793 patients with a total of 248 (31.3%) culture-positive results. 1,322 patients only had preparation prior to or on the day of surgery, with 469 (35.5%) culture positive, P = .048, chi-square = 3.907.

Types of preparation methods

Subgroup analysis was performed for the use of specific preparation solutions. Patients who received chlorhexidine without povidone-iodine solutions numbered 1,199, with 417 (34.8%) culture positive. Patients who received povidone-iodine solutions without chlorhexidine numbered 361, with 110 (30.5%) culture positive, this difference was not significant, P = .129, chi-square = 2.302. Patients who received preparation solutions with alcohol numbered 1,551, of which 577 (37.2%) had positive cultures. Patients who received preparation solutions without alcohol numbered 564, with 140 (24.8%) culture positive, P < .0001, chi-square = 28.285 (Fig. 5). Patients who received preparation solutions with BPO numbered 639, of which 173 (27.1%) were culture positive. Patients who received preparation solutions without BPO numbered 1,476, of which 544 (36.9%) were culture positive, P < .0001, chi-square = 19.05.

Figure 5.

Forest plot of culture positive rates for patients treated with preparation solutions without alcohol.

Adjuvant therapies

Antibiotic administration

Of the studies that analyzed patients receiving intravenous antibiotic administration at the time of surgery, there were 1,689 patients, with a total of 532 (31.5%) with positive cultures. The patients who did not have antibiotics administered at the time of sampling numbered 426, of whom 185 (43.5%) had positive cultures, P < .0001, chi-square = 21.6 (Fig. 6).

Figure 6.

Forest plot of culture positive rates for patients who had intravenous antibiotic administration.

Blue light therapy

Two studies examined the effect of blue light therapy on bacterial colonization rates.

Cotter et al⁴ ran a randomized controlled trial on 60 male participants, subjecting them to BPO topical gel (5 treatments) or 23 minutes of exposure to 40 mW/cm² Omnilux Blue (GlobalMed Technologies, Glen Ellen, CA, USA). A third group had both treatment modalities combined. The patients who had blue light therapy alone had inferior bacterial decolonization to the groups which had BPO, with 20 out of 20 patients demonstrating positive bacterial results after blue light therapy. Horneff et al¹⁴ combined 16 minutes and 40 seconds of 10 J/cm² blue light therapy with topical 5-aminolevulinic acid (Levulan Kerastick, DUSA Pharmaceuticals inc., Billerica, MA, USA) in 25 randomized presurgical patients. The rate of colonization for these 25 patients was 44%, compared to the control group of standard preparation colonization rate of 64%, which was not significantly different (P = .049).¹⁴

Quality of evidence

The overall quality of evidence was assessed by Grading of Recommendations Assessment, Development and Evaluation as being predominately high, providing a relative certainty in the veracity of the results of this review (Table II).

Table II.

Grading of Recommendations Assessment, Development and Evaluation (GRADE) level of quality assessment.

Study	Initial level of certainty	Raise or lower	Final level of certainty
Saltzman 200923	H	-	H
Dizay 20179	L	-	L
Namdari 201721	H		H
Chalmers 20193	L	-	L
Stull 202024	H	-	H
Moor 202126	H	-	H
Hong 202321	H		H
Symonds 202327	H	-	H
Chalmers 20193	L	-	L
Cotter 20215	H	-	H
Crutcher 20246	H	-	H
Grewal 202116	H	-	H
Hancock 201819	H	-	H
Heckmann 201920	L	Increase (+1)3	M
Horneff 202422	H	-	H
Hsu 202013	H	-	H
Hsu 202315	H	-	H
Hudek 201628	L		L
Jalali 202029	H	-	H
Kim 202417	H	-	H
Koh 201618	L	-	L
Maclean 201928	L	-	L
Matsen 202029	L	-	L
Moor 202126	H	-	H
Murray 201130	H	-	H
Phadnis 201631	L	Increase (+1)3	M
Polishchuk 201232	H	-	H
Sabetta 201533	L	Increase (+1)3	M
Scheer 201834	H	-	H
Scheer 202135	H	-	H
Unterfrauner 202236	H	-	H

Quality of evidence: H, high; M, moderate; L, low; VL, very low. GRADE rating – 1: large effect (+1), very large effect (+2), 2: dose-response (+1), 3: all plausible confounding and bias (+1), 4: risk of bias (−1), 5: inconsistency (−1), 6: indirectness (−1), 7: imprecision (−1), 8: publication bias (−1).

Discussion

The purpose of this review was to establish the most effective way to reduce bacterial culture rates in a shoulder surgery patient, given all available shoulder surgery preparation modalities. In terms of timing for antibiotic administration, our hypothesis that more administration episodes would be more effective for bacterial culture rate reduction was not proven to be correct, as the most effective way to reduce culture-positive rates was to use prior-to-day-of-surgery preparation methods. Similarly, our hypothesis was not proven to be correct in terms of combination preparation products, as using more types of preparation products and the use of alcohol solutions was not more beneficial than the use of BPO alone. The caveat to this is the administration of intravenous antibiotics, which significantly reduced culture-positive rates. This study adds to the current body of knowledge surrounding shoulder surgery preparation methods as it provides comparative results on the timing of skin preparation application. A systematic review and meta-analysis by Sewpaul et al³² found similar superiority of effectiveness of BPO over other topical solutions but did not report on intravenous antibiotic usage or differentiate results based on the time of application. Green et al¹⁰ also established in their systematic review that BPO is an effective agent for skin decolonization compared to alternative preparation solutions. They did quantify BPO effectiveness at different time points during the operative procedure but did not compare home preparation to day-of-surgery preparation effectiveness.

Our results demonstrate the effectiveness of multiple preparation and therapeutic methods for reducing the culture-positive rate on the day of surgery. Independently analyzed factors in the administration of preparation solutions lowered the rate of culture positivity at the time of surgery. These included prior to surgery preparation, with or without day-of-surgery preparation; preparation solutions without alcohol; preparation solutions with the use of BPO; and with the administration of intravenous (IV) antibiotics.

The antibiotic effectiveness is commensurate with published guidelines; IV antibiotic prophylaxis is recommended in the perioperative period, with the 2018 ICM guidelines recommending IV cefazolin 30-60 minutes prior to surgery, and up to 24 hours of postoperative therapy. Allergic patients are recommended to have vancomycin 1 to 2 hours prior to surgery. Cephazolin is most likely to provide optimal antibiotic tissue concentrations for prophylaxis against common postsurgical infections, although doxycycline is significantly more effective in treating acne vulgaris (from C acnes) than cephalosporins or vancomycin, and resistance to doxycycline is rare. There is a low rate of decolonization with the use of doxycycline, and this may be due to doxycycline being bacteriostatic rather than bactericidal. Clindamycin as a stand-alone alternative antibiotic treatment is not recommended due to a demonstrated 3.5 times higher risk of prosthetic joint infection with this regimen.²⁰ Prophylactic antibiotic administration for shoulder arthroplasty is typically administered on induction in Australia, although there is published evidence contrary to our findings, that antibiotics administered at induction do not affect intraoperative culture results.¹⁹

Sagkrioti, Glass and Arealis performed a systematic review of the most effective home and surgical skin preparation methods to reduce the C acnes skin load in shoulder surgery from published sources between June 1, 2015, and May 31, 2020. From 19 included studies, they concluded that the most effective solution to reduce the C acnes load is benzoyl peroxide gel (BPO) 5% (w/v) applied 3 times preoperatively, with the last application being on the morning of surgery.²⁸ This is consistent with our findings demonstrating lower culture-positive rates with prior-to-day-of-surgery preparation solution application and the use of BPO. The number of decolonization doses is an important factor in organism eradication, BPO application has been shown to have a dose-dependent effect on C acnes eradication.¹⁰ C acnes is relatively protected from surgical site disinfection and preparation due to its location in the pilosebaceous follicles in the deep skin of the shoulder.^19,24 The relative success of peroxide-based agents in eradicating C acnes is believed to be due to the lipophilic structure of peroxide, which allows deeper penetration to the pilosebaceous glands than traditional skin preparation solutions.³

Our study demonstrated the superiority of prior-to-day-of-surgery preparations; however, there was still a culture-positive rate of 30.5%. The timing of skin preparation before surgery is important, as it has been demonstrated that there is a “rebound effect” of recolonization with C acnes 7 days after decolonization. Recolonization has been measured at 175% of the original bacterial load of C acnes.¹⁰ Following skin preparation, recolonization of the skin flora starts after 30-60 minutes, and regeneration of the physiological flora is complete after three days.^1,8 Skin preparation agents may also fail to eradicate organisms simply because they are not being comprehensively applied. A mean of 27.25% of shoulder and axilla skin remains unprepared after the use of a chlorhexidine preparation stick, and 15.37% of skin remained unprepared when using gauze and a chlorhexidine preparation.⁷

Topical preparations are aimed at diminishing the bacterial load prior to surgery and often have a multifactorial approach but consistently fail to eradicate C acnes on the day of surgery.^15,33 Any instrument that passes through the skin and into the joint may come into contact with C acnes and provide deeper contamination and “false positive” growth of deep C acnes.² Topical preparations are not without risk; benzoyl peroxide has been reported to have a weak allergenic effect and cause contact dermatitis at application sites. Hydrogen peroxide has been recorded as a cause of skin blanching or blistering, and ChloraPrep has been associated with pruritic rashes and anaphylaxis.³³

Postoperative antibiotic administration does not appear to reduce the rate of C acnes colonization and therefore infection risk.⁶ The use of electrocautery for the skin incision in shoulder arthroplasty has been posited as a method to reduce the C acnes burden during surgery. The suggestion is that the high temperature generated by electrocautery decreases the number of viable bacteria that can be cultured intraoperatively. The minimum temperature required to prevent the transmission of C acnes is 70 degrees Celsius.¹¹

Adjuvant techniques to avoid deep infection include changing gloves after the initial incision, keeping implants from touching the subdermal layer, and repeatedly applying skin preparation solutions to the wound after the subdermal layer has been opened.²³

A caveat to our results is that we have pooled study level I to level IV results to provide microbial culture data, and this combination of study types means the data may be skewed by the heterogeneity in study design. However, at its core this review requires articles that report on the accurate use of a surgical preparation method and subsequent culture. This process is not necessarily performed better in a level I compared to level IV study, although the direct comparisons within a level I study are more likely to have confounders eliminated.

Conclusion

This review demonstrates the broad range of tested therapeutic options to reduce the bacterial load the time of shoulder surgery. Due to the demands of patient safety and the variability in surgical practice, there is no absolute certainty in causation between any of the factors analyzed in this review. However, there is a clear correlation that prior to surgery, application of preparation solutions with BPO, without alcohol, and with the inclusion of day-of-surgery intravenous antibiotic administration results in lower culture rates at the time of surgery. Infection will remain a ubiquitous issue for shoulder surgery well into the future, and our methods for reducing bacterial load and assessing the effect of these methods must continue to develop.

Disclaimers

Funding: No direct funding was received for this study in the form of grants, equipment, or other items for this study. No outside source of funds was involved in data collection, data analysis, or the preparation of or editing of the manuscript. The corresponding author is supported by an Australian Government Research Training Program Scholarship. The corresponding author is the recipient of a grant from the Australian Orthopaedic Association Research Foundation (AOARF).

Conflicts of interest: Rajpal S Narulla is supported by an Australian Government Research and Training Program (RTP) Scholarship. All the other authors, their immediate families, and any research foundations with which they are affiliated have not received any financial payments or other benefits from any commercial entity related to the subject of this article.