Can Topical Vancomycin Prevent Periprosthetic Joint Infection in Hip and Knee Arthroplasty? A Systematic Review

Abstract

Background

Periprosthetic joint infection (PJI) after hip and knee arthroplasty is a leading cause of revision surgery, inferior function, complications, and death. The administration of topical, intrawound vancomycin (vancomycin powder) has appeared promising in some studies, but others have found it ineffective in reducing infection risk; for that reason, a high-quality systematic review of the best-available evidence is needed.

Questions/purposes

In this systematic review, we asked: (1) Does topical vancomycin (vancomycin powder) reduce PJI risk in hip and knee arthroplasty? (2) Does topical vancomycin lead to an increased risk of complications after hip and knee arthroplasty?

Methods

A search of Embase, MEDLINE, and PubMed databases as of June 2020 was performed according to Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) guidelines. Studies comparing topical vancomycin in addition to standard infection prevention regimens (such as routine perioperative intravenous antibiotics) with standard regimens only in primary hip and knee arthroplasty were identified. Patients 18 years or older with a minimum follow-up of 3 months were included. No restrictions on maximal loss to follow-up or PJI definition were imposed. Studies were excluded if they included patients with a history of septic arthritis, used an antibiotic other than vancomycin or a different route of administration for the intervention, performed additional interventions that differed between groups, or omitted a control group. A total of 2408 studies were screened, resulting in nine eligible studies reviewing 3371 patients who received topical vancomycin (vancomycin powder) during a primary THA or TKA and 2884 patients who did not receive it. Groups were comparable with respect to duration of follow-up and loss to follow-up when reported. Study quality was assessed using the Newcastle-Ottawa scale, showing moderate-to-high quality for the included studies. The risks of PJI and overall complications in the topical vancomycin group were compared with those in the control group.

Results

One of nine studies found a lower risk of PJI after primary THA or TKA, while eight did not, with odds ratios that broadly bracketed the line of no difference (range of odds ratios across the nine studies 0.09 to 1.97). In the six studies where overall complications could be compared between topical vancomycin and control groups in primary THA or TKA, there was no difference in overall complication risks with vancomycin (range of ORs across the six studies 0.48 to 0.94); however, we caution that these studies were underpowered to detect differences in the types of uncommon complications associated with vancomycin use (such as allergy, ototoxicity, and nephrotoxicity).

Conclusion

In the absence of clear evidence of efficacy, and without a sufficiently large evidence base reporting on safety-related endpoints, topical vancomycin (vancomycin powder) should not be used in routine primary THA and TKA. Adequately powered, multicenter, prospective trials demonstrating clear reductions in infection risk and large registry-driven audits of safety-related endpoints are required before the widespread use of topical vancomycin can be recommended.

Level of Evidence

Level III, therapeutic study.

Introduction

Hip and knee arthroplasty are the third and second most commonly performed inpatient operations, respectively [2]. The lifetime risk of undergoing THA or TKA is more than 11% in the general population, with the risk estimated to increase by 50% by 2026 [4]. Unfortunately, periprosthetic joint infection (PJI) remains one of the most common and devastating complications of hip and knee arthroplasty. The PJI risk ranges from 0.2% to 1.6% for primary THA and from 1% to 2% for primary TKA [13, 16, 20, 28]. Patients with PJI often experience repeated hospitalizations and revision procedures, and they are at risk for serious complications, poorer function, and even death [1, 36]. An estimated USD 1.6 billion is spent on treating hip and knee PJI in the United States annually [17].

Topical delivery of antibiotic powder is a simple intervention that some studies have suggested may reduce the risk of orthopaedic infection and PJI [3, 6, 7, 25]. Direct delivery of antibiotics to a target area allows for high local drug concentrations while perhaps limiting the likelihood of systemic side effects [12], although claims about the frequency of side effects must be considered in light of the fact that most studies on the topic are underpowerd. The potential drawbacks of topical antibiotics include wound-healing complications, reduced osteoblast activity, third-body wear, and contribution to antibiotic resistance [8]. Although systemic absorption is lower when antibiotic powder is used locally than it is with intravenous routes of antibiotic administration, complications such as allergy, ototoxicity, and nephrotoxicity remain possible [8, 9]. Proliferation of antibiotic-resistant organisms is also of concern.

In THA and TKA, topical administration of vancomycin powder for the primary prevention of PJI has been studied in several recent observational studies [3, 5, 10, 14, 15, 21, 24, 33, 35] but conclusions have been hampered because of the low frequency of PJI and high number of patients required to detect relevant differences. Some studies have found no benefit in infection prevention [5, 10, 14, 15, 35], and some studies have even pointed to complications of vancomycin powder [5, 10]. Meta-analyses found a decreased incidence of PJI after topical vancomycin administration in THA and TKA [11, 25, 32]. However, these reviews aggregated studies with low levels of evidence, which should not have been pooled using meta-analytic techniques because of high risk of bias, which may lead to arriving at inaccurate and overstated conclusions, and several large studies have since recently been published [3, 15, 33, 35]. This suggests that another systematic review may be helpful to aggregate evidence and see whether recommendations can be made on this important topic.

Therefore, in this systematic review, we asked: (1) Does topical vancomycin (vancomycin powder) reduce PJI risk in hip and knee arthroplasty? (2) Does topical vancomycin lead to an increased risk of complications after hip and knee arthroplasty?

Materials and Methods

Search Strategy and Criteria

This systematic review was performed according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines [18]. We searched MEDLINE, Embase, and PubMed databases from their inceptions to June 1, 2020 to identify studies comparing powdered vancomycin administration with standard care in THA and TKA for the primary prevention of PJI. The specific search terms were ((vancomycin OR antibiotic OR “Vancomycin”[MeSH]) AND (topical OR local OR powder)) and (THA OR THR OR “Arthroplasty, Replacement, Hip”[MeSH] OR TKA OR TKR OR “Arthroplasty, Replacement, Knee”[MeSH] OR ((hip OR knee) adj2 (replace* OR arthroplast* OR prosthe*))).

Inclusion and Exclusion

Studies were screened against the following inclusion criteria: studies investigating the impact of topical vancomycin powder in primary THA or TKA on PJI, patients ≥ 18 years old, and a minimum follow-up of 3 months. No restriction based on maximum loss to follow-up was set. Articles were excluded if they performed arthroplasty in the context of current or previous PJI or septic arthritis; performed other infection prevention interventions in addition to topical vancomycin powder that differed between groups; used an antibiotic other than vancomycin or a delivery route other than topical powder, excluding routine perioperative intravenous antibiotics; and failed to include a control group without vancomycin powder administration. If studies included revision arthroplasty or unicondylar knee arthroplasty in addition to primary THA or TKA, they were excluded if PJI rates for only primary THA and TKA could not be extracted. Level I to III evidence was eligible for review, including randomized and nonrandomized prospective or retrospective comparative studies. This study excluded review articles, conference abstracts, gray literature, and expert opinion.

Study Selection

Search results from each database were compiled using Covidence systematic review software (Veritas Health Innovation) and duplicates were removed. Two reviewers (MTW, SS) independently screened titles and abstracts. Potentially eligible studies were identified, and the full-text articles of these studies were retrieved for review. These studies were independently reviewed and included in the review if they satisfied the inclusion and exclusion criteria outlined above. The senior author (NMD) resolved disagreements where applicable. Reference lists of the full-text articles and related citations were searched for relevant articles that were not identified in the original search.

Our search strategy identified 2408 studies (Fig. 1). Of these, 701 duplicates were removed, yielding 1707 unique articles. Through title and abstract screening, 41 studies were identified for full-text review. Most studies were excluded during full-text review for not performing topical vancomycin administration, performing additional infection prevention measures that differed between experimental and control groups, not reporting on PJI, or not including a control group. Eight studies from the original search met the inclusion criteria [5, 10, 14, 15, 21, 24, 33, 35]. One additional eligible study was identified from related citations but was not indexed by any of the databases searched [3]. This was added to the previously identified studies for a total of nine included studies.

Fig. 1.

This flowchart depicts the search strategy and screening. We systematically reviewed nine studies.

Assessment of Study Quality

There were eight retrospective, comparative studies [3, 5, 10, 14, 21, 24, 33, 35] and one prospective, comparative study [15]. Study sizes ranged from 115 to 1733 patients, who were recruited between 2010 and 2017. The quality of the included studies was analyzed by two reviewers (MTW, SSS) using the validated Newcastle-Ottawa scale for nonrandomized studies in meta-analyses [29]. Disagreements were resolved by mutual agreement. Scores on the Newcastle-Ottawa scale can range from 0 to 9, where a higher score indicates higher quality and a lower risk of bias [29]. In general, these studies scored well to very well on the Newcastle-Ottawa scale, suggesting moderate-to-low risk of bias and moderate-to-high study quality. One study received 5 points, one study received 6 points, four studies received 8 points, and three studies received 9 points (Table 1).

Table 1.

Newcastle-Ottawa Scale quality assessment of included studies

Author	Selection				Comparability	Outcome
	Representativeness of cohort	Controls selection	Ascertainment of exposure	Outcome not present at start of study	Comparability of cohorts	Assessment of outcome	Length of follow-up	Adequacy of follow-up
Crawford et al. [3]	*	*	*	*	*
Dial et al. [5]	*	*	*	*	**	*	*	*
Hanada et al. [10]	*	*	*	*	*	*	*	*
Khatri et al. [14]	*	*	*	*	*		*
Koutalos et al. [15]	*	*	*	*	*	*	*	*
Otte et al. [21]	*	*	*	*	*	*	*	*
Patel et al. [24]	*	*	*	*	**	*	*
Xu et al. [33]	*	*	*	*	**	*	*	*
Yavuz et al. [35]	*	*	*	*	**	*	*	*

Each star equals 1 point.

Data Collection and Abstraction

A predefined spreadsheet was populated from each of the included full-text articles by two reviewers (MTW, EMD). Abstracted data included study design, study population, demographic information, enrollment period, age, ratio of men to women, BMI, presence of diabetes, smoking status, and follow-up periods (Table 2).

Table 2.

Characteristics of included studies

Author	Year	Date range	Study design	Minimum follow-up in months	% of patients (n)		Mean age in years
					Vancomycin	Control	Vancomycin	Control
Crawford et al. [3]	2018	2011-2015	Retrospective controlled	NR	57 (987 of 1733)	43 (746 of 1733)	65	63
Dial et al. [5]	2018	2013-2016	Retrospective controlled	3	52 (137 of 265)	48 (128 of 265)	62	61
Hanada et al. [10]	2019	2010-2017	Retrospective controlled	12	51 (90 of 176)	49 (86 of 176)	75	74
Khatri et al. [14]	2017	2014-2016	Retrospective controlled	6	44 (51 of 115)	56 (64 of 115)	NR	NR
Koutalos et al. [15]	2020	2016-2017	Prospective controlled	24	50 (153 of 303)	50 (150 of 303)	67	68
Otte et al. [21]	2017	2012-2014	Retrospective controlled	3	51 (682 of 1346)	49 (664 of 1346)	NR	NR
Patel et al. [24]	2018	2016-2017	Retrospective controlled	3	76 (348 of 460)	24 (112 of 460)	64	65
Xu et al. [33]	2020	2015-2017	Retrospective controlled	18	51 (437 of 855)	49 (418 of 855)	67	67
Yavuz et al. [35]	2019	2012-2016	Retrospective controlled	24	49 (486 of 1002)	51 (516 of 1002)	66	63
Total					54 (3371 of 6255)	46 (2884 of 6255)

Author	% of males (n)		Mean BMI in kg/m2		% with diabetes (n)		% who smoke (n)
	Vancomycin	Control	Vancomycin	Control	Vancomycin	Control	Vancomycin	Control
Crawford et al. [3]	49 (484 of 987)	48 (358 of 746)	31	31	13 (133 of 987)	16 (119 of 746)	NR	NR
Dial et al. [5]	47 (65 of 137)	50 (64 of 128)	30	30	12 (16 of 137)	15 (19 of 128)	16 (22 of 137)	15 (19 of 128)
Hanada et al. [10]	26 (23 of 90)	23 (20 of 86)	NR	NR	NR	NR	NR	NR
Khatri et al. [14]	63 (32 of 51)	69 (44 of 64)	NR	NR	33 (17 of 51)	36 (23 of 64)	12 (6 of 51)	8 (5 of 64)
Koutalos et al. [15]	27 (41 of 153)	25 (37 of 150)	NR	NR	NR	NR	NR	NR
Otte et al. [21]	NR	NR	NR	NR	NR	NR	NR	NR
Patel et al. [24]	40 (138 of 348)	43 (48 of 112)	31	31	10 (34 of 348)	13 (14 of 112)	11 (38 of 348)	13 (14 of 112)
Xu et al. [33]	28 (121 of 437)	31 (129 of 418)	25	25	10 (42 of 437)	14 (58 of 418)	22 (97 of 437)	21 (87 of 418)
Yavuz et al. [35]	30 (148 of 486)	30 (154 of 516)	29	29	21 (102 of 486)	25 (129 of 516)	NR	NR

There were no randomized trials; NR = not reported.

Primary and Secondary Study Outcomes

The primary outcome of interest was the overall rate of PJI in THA and TKA. Secondary outcomes studied were PJI proportions in the THA and TKA subgroups, as well as the overall complication rates and rate of wound-related complications. One study investigated primary TKA and unicompartmental knee arthroplasty, although the results of unicompartmental knee arthroplasty were not included in our analysis [10]. Similarly, a study investigated primary and aseptic revision arthroplasty [21], but only primary arthroplasty data were extracted in our systematic review.

Outcomes Assessment

In the studies in our review, PJI was most commonly determined using the International Consensus Meeting on PJI definition [22], followed by the Musculoskeletal Infection Society 2011 definition (Table 3) [23]. Two studies [3, 14] determined the presence of PJI based on unspecified synovial aspirate analysis and the need for debridement in the opinion of the treating team. Wound-related complication rates were variably reported and included PJI, superficial infection, dehiscence, delayed healing, and seroma or hematoma formation requiring surgical drainage. Overall possible complications included wound complications, acute kidney injury, anaphylactic reaction, myocardial infarction, cerebrovascular accident, deep vein thrombosis or pulmonary embolism, component loosening, dislocation, periprosthetic fracture, and death.

Table 3.

Study populations, vancomycin powder administration protocols, and PJI definitions

Author	Patient population	Vancomycin administration protocol	PJI definition
Crawford et al. [3]	Primary THA	1 g intraarticular	“Based on a hip fluid cell count, culture, and need for radical debridement”
Dial et al. [5]	Primary THA	1 g intraarticular and extraarticular	International Consensus Meeting on PJI
Hanada et al. [10]	Primary TKA and unicompartmental knee arthroplasty	1 g intraarticular	International Consensus Meeting on PJI
Khatri et al. [14]	Primary TKA	1 g intraarticular	“Evaluated by wound exploration and debridement”
Koutalos et al. [15]	Primary THA and TKA	2 g intraarticular	Musculoskeletal Infection Society 2011
Otte et al. [21]	Primary and aseptic revision THA or TKA	1 g intraarticular	Musculoskeletal Infection Society
Patel et al. [24]	Primary THA or TKA	1 g throughout wound including intraarticular	Musculoskeletal Infection Society
Xu et al. [33]	Primary TKA	0.5 g intraarticular	International Consensus Meeting on PJI
Yavuz et al. [35]	Primary TKA	2 g intraarticular	International Consensus Meeting on PJI

Characteristics of Included Studies

Combined, the included studies investigated 3371 patients with topical vancomycin powder administration and 2884 patients treated without topical antibiotic administration (Table 2). The groups were comparable with respect to age, ratio of men to women, BMI, proportion of patients with diabetes, and smoking status. Only two studies reported loss to follow-up rates [15, 35], which had no difference in loss to follow-up between groups. Follow-up duration was comparable between groups. The minimum follow-up period ranged from 3 months to 2 years. Before wound closure, 0.5 to 2 g of vancomycin powder was administered (Table 3). Five studies reported on PJI after THA [3, 5, 15, 21, 24] while seven reported on PJI after TKA [10, 14, 15, 21, 24, 33, 35]. Six studies listed overall complication rates [3, 5, 15, 24, 33, 35] and seven studies described wound-related complications [3, 5, 14, 15, 24, 33, 35].

Statistical Analysis

Odds ratios were calculated for each study and outcome and evaluated using the Fisher exact test, regardless of whether they were calculated in the original studies, as many studies used the Pearson chi-square test, which may overestimate effect sizes for smaller sample sizes. No data pooling or meta-analysis was performed because most of these studies were retrospective. P values less than 0.05 were deemed statistically significant.

Results

Does Topical Vancomycin Reduce the Risk of PJI in Hip and Knee Arthroplasty?

Only one of nine studies [3] found a lower risk of PJI in THA or TKA with topical vancomycin than in controls (Table 4), while the remaining eight did not, with odds ratios that broadly bracketed the line of no difference. Odds ratios for the nine studies ranged between 0.09 and 1.97 (eight had p > 0.05 and one reported a p value of 0.04). In THA alone, one of five studies found a lower risk of PJI [3] (p = 0.04). Again, the remainder of the studies found no difference, and odds ratios ranged widely on either side of the line of no difference, between 0.11 and 2.86 for the five studies (Table 5, four of five reported p > 0.05). In TKA, none of the seven studies showed an effect of vancomycin on PJI (Table 6). Of these seven studies, odds ratios were between 0.07 and 1.01 (all p > 0.05).

Table 4.

Percentage of patients with PJI after topical vancomycin administration compared with controls

Study	Vancomycin with PJI	Control with PJI	Odds ratio (95% CI)	p value
Crawford et al. [3]	0.1 (1 of 987)	0.9 (7 of 746)	0.11 (0.01-0.87)	0.04
Dial et al. [5]	0.7 (1 of 137)	5.5 (7 of 128)	0.13 (0.02-1.05)	0.06
Hanada et al. [10]	5.6 (5 of 90)	8.1 (7 of 86)	0.66 (0.20-2.18)	0.50
Khatri et al. [14]	7.8 (4 of 51)	9.4 (6 of 64)	0.82 (0.22-3.09)	0.77
Koutalos et al. [15]	1.3 (2 of 153)	0.7 (1 of 150)	1.97 (0.18-22.00)	0.58
Otte et al. [21]	0.6 (4 of 682)	0.9 (6 of 664)	0.65 (0.18-2.30)	0.50
Patel et al. [24]	0.3 (1 of 348)	2.7 (3 of 112)	0.10 (0.01-1.02)	0.05
Xu et al. [33]	0.0 (0 of 437)	1.2 (5 of 418)	0.09 (0.00-1.56)	0.10
Yavuz et al. [35]	0.8 (4 of 486)	1.0 (5 of 516)	0.85 (0.23-3.18)	0.81

Data presented as % (n).

Table 5.

Percentage of THA patients with PJI after topical vancomycin administration compared with controls

Study	Vancomycin with PJI	Control with PJI	Odds ratio (95% CI)	p value
Crawford et al. [3]	0.1 (1 of 987)	0.9 (7 of 746)	0.11 (0.01-0.87)	0.04
Dial et al. [5]	0.7 (1 of 137)	5.5 (7 of 128)	0.13 (0.02-1.05)	0.06
Koutalos et al. [15]	1.5 (1 of 67)	0.0 (0 of 63)	2.86 (0.11-71.63)	0.52
Otte et al. [21]	1.1 (3 of 282)	1.6 (4 of 252)	0.67 (0.15-3.01)	0.60
Patel et al. [24]	0.5 (1 of 187)	1.8 (1 of 56)	0.30 (0.02-4.81)	0.39

Data presented as % (n).

Table 6.

Percentage of TKA patients with PJI after topical vancomycin administration compared with controls

Study	Vancomycin with PJI	Control with PJI	Odds ratio (95% CI)	p value
Hanada et al. [10]	5.6 (5 of 90)	8.1 (7 of 86)	0.66 (0.20-2.18)	0.50
Khatri et al. [14]	7.8 (4 of 51)	9.4 (6 of 64)	0.82 (0.22-3.09)	0.77
Koutalos et al. [15]	1.2 (1 of 86)	1.1 (1 of 87)	1.01 (0.06-16.44)	0.99
Otte et al. [21]	0.3 (1 of 400)	0.5 (2 of 392)	0.49 (0.04-5.41)	0.56
Patel et al. [24]	0.0 (0 of 161)	3.6 (2 of 56)	0.07 (0.00-1.43)	0.08
Xu et al. [33]	0.0 (0 of 437)	1.2 (5 of 418)	0.09 (0.00-1.56)	0.10
Yavuz et al. [35]	0.8 (4 of 486)	1.0 (1 of 516)	0.85 (0.23-3.18)	0.81

Data presented as % (n).

Does Topical Vancomycin Lead to Increased Complications in Hip and Knee Arthroplasty?

Of the six studies that reported complications including nonwound-related complications, there were no differences in the complication risk (Table 7). Odds ratios for these six studies were between 0.48 and 0.94 (all p > 0.05). One study [3] showed a reduction in the frequency of wound complications including PJI (Table 8; p = 0.03), but the other six of seven studies reporting found no difference. Odds ratios for the seven studies were between 0.31 and 0.93 (six of seven reported p > 0.05). Importantly, no study was sufficiently powered to detect differences in uncommon complications associated with vancomycin including allergy, ototoxicity, and nephrotoxicity, nor was there sufficient follow-up to detect medium- and long-term complications.

Table 7.

Percentage of patients with any complication after topical vancomycin administration compared with controls

Study	Vancomycin with any complication	Control with any complication	Odds ratio (95% CI)	p value
Crawford et al. [3]	0.8 (8 of 987)	1.2 (9 of 746)	0.67 (0.26-1.74)	0.41
Dial et al. [5]	8.8 (12 of 137)	13.3 (17 of 128)	0.63 (0.29-1.37)	0.24
Koutalos et al. [15]	2.0 (3 of 153)	4.0 (6 of 150)	0.48 (0.12-1.96)	0.31
Patel et al. [24]	3.2 (11 of 348)	5.4 (6 of 112)	0.58 (0.21-1.60)	0.29
Xu et al. [33]	18.5 (81 of 437)	21.8 (91 of 418)	0.82 (0.58-1.14)	0.24
Yavuz et al. [35]	1.6 (8 of 486)	1.7 (9 of 516)	0.94 (0.36-2.46)	0.90

Table 8.

Percentage of patients with wound complications after topical vancomycin administration compared with controls

Study	Vancomycin with wound complications	Control with wound complications	Odds ratio (95% CI)	p value
Crawford et al. [3]	0.5 (5 of 987)	1.6 (12 of 746)	0.31 (0.11-0.89)	0.03
Dial et al. [5]	5.8 (8 of 137)	7.0 (9 of 128)	0.82 (0.31-2.19)	0.69
Khatri et al. [14]	9.8 (5 of 51)	12.5 (8 of 64)	0.76 (0.23-2.48)	0.65
Koutalos et al. [15]	2.0 (3 of 153)	4.0 (6 of 150)	0.48 (0.12-1.96)	0.31
Patel et al. [24]	1.1 (4 of 348)	2.7 (3 of 112)	0.42 (0.09-1.92)	0.26
Xu et al. [33]	13.3 (58 of 437)	14.1 (59 of 418)	0.93 (0.63-1.38)	0.72
Yavuz et al. [35]	1.4 (7 of 486)	1.7 (9 of 516)	0.82 (0.30-2.23)	0.70

Discussion

PJI after hip and knee arthroplasty has severe consequences both for patients and for the healthcare systems in which they receive care. Multiple studies comparing the addition of topical vancomycin (vancomycin powder) with routine intravenous perioperative antibiotics have been performed but are underpowered to detect differences in the risk of PJI postoperatively, and some of those studies have reached very different conclusions regarding its efficacy [3, 5, 10, 24]. Therefore, we performed a systematic review to identify trends across these individual studies to better understand the effect of topical vancomycin on PJI and overall risk of complications. In our review, a preponderance of the evidence did not find that topical vancomycin reduced infection risk, and there was insufficient evidence on the question of safety. Considering those findings, we recommend that surgeons not use topical vancomycin (vancomycin powder) in routine THA and TKA.

Limitations

This review is limited by the low level of evidence in the included studies. The included studies are predominantly retrospective and most have small sample sizes. Higher-level evidence, including randomized controlled trials, is lacking. Therefore, we refrained from pooling data or performing a meta-analysis. Most studies compared sequential groups of patients without vancomycin administration followed by a series in which topical vancomycin was used, or they evaluated one surgeon using vancomycin compared with a different surgeon as a control who did not use it. However, we note that despite there being no randomized trials on this topic, study quality—as measured by the Newcastle-Ottawa scale—was generally good to excellent (Table 1), which tends to support the validity of the main findings here. Although any differences in protocol between groups other than vancomycin administration (for example, povidone-iodine irrigation, systemic gram-negative antibiotic coverage, or different venous thromboembolism prophylaxis) led to article exclusion, it is plausible that other factors not measured or accounted for may influence the results. Additionally, results are limited by the relatively short minimum follow-up period. Although most PJIs occur within 3 months [37], longer follow-up is required to understand the effects of topical vancomycin administration on delayed PJI and chronic complications such as component loosening or wear. Also concerning is the differing definition of PJI between studies, especially where definitions were based on surgeon opinion [3, 14] and may not have been reproducible. Despite this, results remain valid as seven of nine studies used evidence-based and validated criteria [22, 23], while the two other studies [3, 14] applied their criteria uniformly between groups. Finally, varied reporting of complications other than PJI limits the strength of the current findings with respect to wound complications or overall complications.

Does Topical Vancomycin Reduce the Risk of PJI in Hip and Knee Arthroplasty?

Odds ratios of PJI after vancomycin compared with controls broadly bracketed the line of no difference, suggesting that vancomycin powder may have, in fact, been associated with an increase or a decrease in the risk of PJI. Small sample sizes and infrequency of the event of interest (PJI) led to low precision in the estimate of the PJI risk. The only study in this review that favored vancomycin on the basis of a statistical finding was a retrospective study with a definition of PJI that was vague and subjective [3]. The PJI risk in the control group was comparable to previously published risks of PJI for THA and TKA [13, 16, 20, 28]. In addition, the odds ratios of most studies comparing topical vancomycin and standard care are concordant with prior estimates in THA and TKA [11, 19] and estimates from other procedures such as spine, elbow, and foot and ankle surgeries [6-8, 31, 34].

Prior authors have sought to investigate the impact of topical vancomycin on PJI after THA and TKA with meta-analyses. Heckmann et al. [11] reviewed six studies and 3298 patients, Xu et al. [32] reviewed nine studies including 4607 patients, and Peng et al. [25] reviewed nine studies and 4512 patients. However, these authors combined studies that were predominantly retrospective and of poor quality, limiting their ability to draw accurate conclusions. In general, better orthopaedic (and general medical) journals do not permit meta-analyses of retrospective studies of treatments for this reason, and we caution readers against having too much confidence in them in this setting. In addition, these authors included studies in which patients underwent a combined intervention in which the surgical wound was irrigated with povidone-iodine solution in addition to topical vancomycin administration [27, 30], and one study investigated the effect of topical vancomycin in the setting of irrigation and debridement with liner exchange for acute PJI rather than for PJI prevention [27]. By pooling data from these studies, prior meta-analyses almost certainly overestimated the apparent benefit of vancomycin powder, and likely do not provide an accurate estimate of the true impact of topical vancomycin on PJI. Restricting our inclusion criteria in the current review to patients without a history of septic arthritis undergoing primary THA or TKA and limiting interventions to only topical vancomycin powder administration allowed this systematic review to better represent the existing evidence supporting topical vancomycin for the primary prevention of PJI in hip and knee arthroplasty.

Does Topical Vancomycin Lead to Increased Complications in Hip and Knee Arthroplasty?

No study in this systematic review detected differences in the overall complication risk, including medical or wound-related complications, although individual studies were underpowered to do so. Inconsistent reporting of medical complications and their infrequency precluded our ability to compare suspected vancomycin-related complications between groups with great statistical precision. Concerns about nephrotoxicity, ototoxicity, anaphylactic reactions, development of vancomycin-resistant organisms, and increased component wear have been raised [8, 9, 11, 26], although their frequency remains undocumented, perhaps because of small sample sizes. Topical administration leads to peak serum vancomycin concentrations that are one-third of therapeutic levels but local concentrations are 60 times the therapeutic floor, which may explain the apparent infrequency of systemic complications compared with intravenous vancomycin administration [13].

Reporting and classification of wound complications varied between studies and was inherently subjective for some complications. Only one study showed a reduction in the risk of wound complications, although this was likely driven by the reduction in PJI. Previous authors have noted increased sterile wound complications after topical vancomycin administration, such as seroma formation and delayed wound healing [5, 10]. Based on the reviewed studies, insufficient evidence exists to confirm the safety of topical vancomycin administration. Even multicenter studies are likely going to be too small to capture differences in frequency of uncommon but serious events that may be associated with use of vancomycin powder, and so we recommend using insurance databases and national health registries to try to determine whether patients who received topical vancomycin (vancomycin powder) in this context are more likely to experience complications like nephrotoxicity, ototoxicity, anaphylactic reactions, development of vancomycin-resistant organisms, and premature revisions perhaps related to increased component wear.

Conclusion

PJI is a devastating and difficult-to-treat complication after THA and TKA. Surgeons and researchers had hoped that topical vancomycin (vancomycin powder) would reduce infection risk without causing drug-related complications. Unfortunately, the available studies have low levels of evidence and methodologic flaws, and in aggregate, they were not sufficiently convincing on the topic of efficacy to allow us to recommend the routine use of topical vancomycin in THA and TKA, and until large, safety-related audits of healthcare systems or registries are performed to ascertain whether use of topical vancomycin is safe—small, single-center studies are unlikely to be large enough to detect an increase in the proportion of patients with uncommon or rare complications, should they be present—we recommend against its routine use in this setting. Multicenter studies on efficacy and large-database studies on safety are needed.