Effect of Intra‐wound Vancomycin for Spinal Surgery: A Systematic Review and Meta‐analysis

Abstract

Intra‐site prophylactic vancomycin in spine surgery is an effective method of decreasing the incidence of postsurgical wound infection. However, there are differences in the prophylactic programs used for various spinal surgeries. Thus, this systematic review and meta‐analysis aimed to evaluate the effectiveness of using intra‐wound vancomycin during spinal surgery and to explore the effects of dose‐dependence and the method of administration in a subgroup analysis. A total of 628 citations or studies were searched in PubMed, Ovid, Web of Science, and Google Scholar that were published before August 2016 with the terms “local vancomycin”, “intra‐wound vancomycin”, “intraoperative vancomycin”, “intra‐site vancomycin”, “topical vancomycin”, “spine surgery”, and “spinal surgery”. Finally, 19 retrospective cohort studies and one prospective case study were eligible for inclusion in the systematic review and meta‐analysis. The odds of developing postsurgical wound infection without prophylactic local vancomycin use were 2.83‐fold higher than the odds of experiencing wound infection with the use of intra‐wound vancomycin (95% confidence interval, 2.03–3.95; P = 0.083; I ² = 32.2%). The subgroup analysis including the dosage and the method of administration, revealed different results compared to previous research. The value of I ² in the 1‐g group was 27.2%, which was much lower than in the 2‐g group (I ² = 57.6%). At the same time, the value of I ² was 0.0% (P = 0.792, OR = 2.70) when vancomycin powder was directly sprinkled into all layers of the wound. However, there is high heterogenicity (I ² = 60.0%, P = 0.007, OR = 2.83) when vancomycin powder is not exposed to the bone graft and instrumentation. There are differences found with the method of local application of vancomycin for reducing postoperative wounds and further studies are necessary, including investigations focusing on the dose‐dependent effects during spinal or the topical pharmacokinetic and other orthopaedic surgeries.

Introduction

Postoperative wound infection following various spinal surgeries is a serious complication1, 2. The incidence of post‐surgical wounds in spine surgery is high, and various researchers have reported different infection rates3, 4, 5. In addition, increased healthcare costs, prolonged lengths of stay in hospital, and reduced quality of life as a result of surgical site infections (SSI) are also major concerns6. Several methods for avoiding SSI, such as betadine irrigation, vacuum‐assisted closure, and intra‐wound vancomycin powder, have been used to reduce the rate of wound infection in spine surgery7, 8, 9. Use of local vancomycin has been popular because of its protective effects and lower cost. According to some reports, prophylactic administration of intra‐wound vancomycin powder before wound closure is an effective method for decreasing postoperative wound infection rates9, 10, 11, 12, 13; however, other studies have revealed a non‐significant effect of intra‐wound vancomycin use for decreasing the postsurgical wound infection rate14, 15, 16. In the present study, we combined data from existing studies to assess the effectiveness of the use of intra‐operative vancomycin powder for reducing postsurgical infections after various types of spinal surgery. We also performed a subgroup analysis of the dosage used and the method of administration to provide a reference for investigating dose‐dependent effects in the future.

Materials and Methods

Search Strategy

The systematic search strategy involved one website and three databases. All articles published before August 2016 in the PubMed, Ovid, and Web of Science databases and the Google Scholar website were searched using the terms “local vancomycin”, “intra‐wound vancomycin”, “intraoperative vancomycin”, “intra‐site vancomycin”, “topical vancomycin”, “spine surgery”, and “spinal surgery”. Studies were limited to articles published in English. In addition, the reference lists of all located articles were screened to determine whether any additional references could be included in this meta‐analysis. We also scanned the reference lists of systematic review and meta‐analysis articles concerning prophylactic vancomycin and spinal surgery. Three independent surgeons (Dan Pu, Jun Zhu and Lunli Xie) performed the searches and made decisions in terms of inclusion and exclusion criteria. When there was a question about a study's eligibility for inclusion and a consensus was not reached, the senior surgeon made the final decision after a group discussion.

Inclusion and Exclusion Criteria

Inclusion criteria were: (i) retrospective and prospective studies; (ii) human clinical studies with local application of vancomycin powder before wound closure for various spine pathologies and surgeries (i.e. degenerative, traumatic, tumor, and deformity; instrumented, non‐instrumented, and decompression); (iii) studies including the incidence of postsurgical wound infection, the sides effects, the dosage, and follow‐up time; and (iv) for subgroup analysis of dosage, the studies revealing both the defined administration of methods and the incidence of postoperative wound infection.

Exclusion criteria were: (i) letters, case reports, and replies; (ii) studies lacking control groups; and (iii) ill‐defined dosage or unclear data regarding the dosage used. However, some studies may be suitable for one of the subgroups, including the dosage of usage and the method of administration.

Data Extraction and Study Assessment

Data collected included the year of publication, authors, type of study, the rate of infections in the treatment and control groups, dosage used, number of patients, follow‐up duration, micro‐floras, surgery location, type of wound infection, method of administration, and level of evidence (LOE). If the LOE was not included in the paper, two independent surgeons performed an assessment according to the Oxford Centre for Evidence‐Based Medicine for the included papers17.

The quality of included studies was, respectively, evaluated by applying the JADAD and Newcastle–Ottwa Scale (NOS) for retrospective cohort studies and prospective randomized controlled trials18, 19. The literature was regarded as of low quality when the score of the JADAD and the NOS was less than 2 and 5 points, respectively. Final score ≥6 and >3 was, respectively, regarded as high quality based on the NOS and the JADAD. The senior surgeon consulted with the independent surgeons when they had different opinions on the LOE and the quality of studies.

Data Analysis and Statistical Methods

The number of infections and the dosage of medication used in the treatment and control groups for each included study were documented and analyzed with Stata (Version 14.0, Texas, USA) using a random or fixed model to estimate the effects. Subgroup analyses were also performed for some factors (e.g. dose and method of administration). A funnel plot was used to assess publication bias. When I ² > 60%, we also performed a sensitivity analysis for the included articles.

Results

Search Result

Overall, 215 citations were revealed from Google Scholar, and 413 studies were identified from the PubMed, Ovid, and Web of Science databases. Of the former, 151 citations were excluded because they were not related to our topic, and 41 citations were excluded after review of their abstracts. Of the latter 413 studies, 262 were excluded because they were not relevant to our study, and 90 were excluded after being reviewed. After omitting the repeated studies, 20 studies met our inclusion criteria, consisting of 19 retrospective studies and 1 prospective study (Table 1; Fig. 1).

Table 1.

The overall findings of the included studies using local vancomycin in spinal surgery

Year and authors	Study design	LOE	Quality score	Size	Location	Dosage (g)	Type of wound infection	Micro‐floras	Follow‐up duration	Methods
2016, Gaviola et al. 20	Retrospective cohort study	3	NOS: 6	116/210	Multilevel spinal fusion	2	Superficial and deep SSI	Gram‐positive bacteria, Gram‐negative bacteria; MSSA; MRSA	3 months	A
2016, Lee et al. 21	Retrospective case study	3	NOS: 8	275/296	Posterior lumbar surgery	1	Deep SSI	Staphylococcal infection; MRSA	8 months	A
2016, Dennis et al. 22	Retrospective comparative cohort study	3	NOS: 6	117/272	Instrumented spinal surgery	1	Superficial and deep SSI	Pseudomonas aeruginosa; MRSA	3 months	A
2015, Martin et al. 15	Retrospective cohort study	3	NOS: 6	115/174	Posterior cervical fusion surgery	2	Deep SSI	Gram‐positive cocci; Gram‐negative rods; multiple organisms	NR	B
2015, Heller et al. 23	Retrospective historical cohort design	3	NOS: 7	342/341	Instrumented spinal surgery	0.5–2	Acute staphylococcal; Deep staphylococcal; Deep MRSA infection	MRSA	3 months	A
2015, Schroeder et al. 24	Retrospectively reviewed study	3	NOS: 7	1224/2253	Degenerative spinal surgery	1–1.5	Deep infection	P. acnes, MRSA, E. coli	12 months	A
2015, Devin et al. 25	Retrospectively reviewed study	2	NOS: 7	966/1190	Posterior spinal surgery	1	Superficial and deep infection	Nothing	NR	B
2015, Ross et al. 26	Retrospective, descriptive, case‐control study	3	NOS: 7	70/140	Lumbar fusion	1	Deep wound infection	Nothing	12 months	A
2014, Martin et al. 16	Retrospective cohort study	2	NOS: 7	156/150	Posterior spinal deformity surgical procedure	2	Deep wound infection	Coagulase‐negative rods MSRA	NR	B
2014, Hill et al. 27	Retrospectively reviewed study	3	NOS: 6	150/150	Spinal surgery	1	Superficial and deep infection	Staphylococcus aureus; Enterococcus; MRSA	1 month	A
2014, Theologis et al. 28	Retrospective cohort study	3	NOS: 6	151/64	Thoracolumbar adult deformity procedure	2	Wound infection	Nothing	18 months	A
2013, Strom et al. 29	Retrospective cohort study	2	NOS: 7	79/92	Posterior cervical fusion	2	Wound infection	Coagulase‐negative staphylococcal, gram‐negative rods, MSSA, MSRA	12 months	B
2013, Pahys et al. 30	Retrospectively reviewed study	3	NOS: 7	483/195	Posterior cervical spine surgery	0.5	Superficial and deep infection	Nothing	3 months	A
2013, Strom et al. 31	Retrospective cohort study	3	NOS: 6	156/97	Lumbar laminectomy and fusion	1	Wound infection	MSSA, MSRA, coagulase‐negative staphylococcal, gram‐negative rods	12 months	B
2013, Caroom et al. 32	Retrospective comparative study	2	NOS: 7	911/821	Posterior cervical fusion	1	Wound infection	Methicillin‐resistant Coagulase‐negative Staphylococcus species, MRSA	6 months	A
2013, Kim et al. 33	Retrospective comparative cohort study	3	NOS: 5	34/40	Instrumented spinal fusion surgery	1	Superficial and deep infection	Nothing	NR	B
2011, Sweet et al. 9	Retrospective cohort study	3	NOS: 7	911/821	Thoracolumbar fusion	2	Deep wound infection	Coagulase‐negative staph	2.5 years	B
2011, O'Neill et al. 34	Retrospective case study	3	NOS: 5	56/54	Posterior spinal fusion for traumatic injury	1	Superficial and deep infection	Methicillin‐resistant Staphylococcus aureus for deep infection	0.5 month	B
2011, Rahman et al. 35	Retrospective case study	3	NOS: 5	586/334	Adult spinal deformity surgery	1–2	Wound infection	NR	NR	B
2015, Tubaki et al. 13	Prospective randomized controlled trial	2	JADAD: 3	433/474	Spinal surgery	1	Superficial and deep infection	Klebsiella, Staph. aureus	3 months	B

A, vancomycin was directly sprinkled into all layers of the wound; B, vancomycin was directly sprinkled into the muscle, fascia, and subcutaneous layers, but the bone graft and instrumentation were not exposed to it; LOE, level of evidence; MRSA, methicillin‐resistant Staphylococcus aureus; MSSA, methicillin‐susceptible Staphylococcus aureus; NOS, Newcastle‐Ottawa Scale; NR, not reported; SSI, surgical site infection.

Figure 1.

Flow chart of study selection in this systematic review and meta‐analysis.

Study Characteristics

This systematic review and meta‐analysis consists of 7 retrospective cohort studies, 4 retrospectively reviewed studies, 4 retrospective case studies, 3 retrospective comparative studies, 1 retrospective historical cohort study, and 1 prospective randomized controlled trial9, 13, 15, 16, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35. Most included studies were published within the last 4 years, ranging from 2011 to 2015 (Table 1). The sample sizes of the included studies ranged from 70 to 1224 in the treatment group. In the treatment group, patients received the standard intravenous antibiotic during the perioperative period plus intrawound vancomycin powder before closure of the wound during surgery. However, in the control group, the patients had the standard perioperative intravenous antibiotic alone. All the studies reported on spine surgeries, including for degenerative diseases, trauma, and deformities, but not tumors. The included studies revealed the rate of wound infection, side effects, and the results of the culture from the infected wound. Of the 20 studies, 18 showed that the use of intra‐wound vancomycin powder might prevent SSI after spinal surgeries. However, 3 studies showed little evidence that it prevented SSI. The level of evidence (LOE) of the included studies was 2 and 3, and most LOE were 3. The period of follow‐up ranged from 1 month to 2.5 years.

Risk of Bias Assessment

The included studies consisted of 1 prospective randomized trial and 19 retrospective cohort studies. The results of the risk of bias evaluation reported in Table 1 were evaluated by applying the JADAD for prospective randomized controlled trial (1) and the Newcastle–Ottwa scale (NOS) for retrospective cohort studies (19)18, 19. Most studies had low to moderate risk of bias when assessed by applying the Newcastle–Ottwa scale, with the score ranging from 5 to 7. Lee et al. report on posterior lumbar surgeries with application of intraoperative local 1 g vancomycin in 275 patients (NOS was 8)21. The quality score of the prospective study reported by Tubaki et al. was 3 evaluated by JADAD13. Higher scores in studies imply less risk of bias and better study quality. Bias in this systematic review and meta‐analysis derived primarily from adequate follow‐up duration. Few studies reported the result of the follow‐up time in the control group.

Outcomes for Meta‐analysis

Effects of Local Application of Vancomycin Powder

There was a lower incidence of postoperative wound infection in the study group compared to the control group, which was revealed by the pooled effects for the included studies. Our review indicated that local application of vancomycin powder shows benefits for reducing postsurgical wound infection. The pooled odds ratio (OR) was 2.83 (95% CI: 2.03–3.95), I ² = 32.2%, P = 0.083 (Fig. 2). The I ² value suggests that low‐to‐mild heterogenicity might exist for this meta‐analysis.

Figure 2.

The Forest graph of the total included studies. This figure reveals the rate of postoperative wound infection in spine surgery for total included studies. The pooled odds ratio (OR) was 2.83 (95% CI: 2.03–3.95), I ² = 32.2%, P = 0.083.

Subgroup Analyses

Dosage Used

Most studies use 1 g of vancomycin powder13, 21, 22, 25, 26, 27, 31, 32, 33, 34, whereas some studies use 2 g9, 15, 16, 20, 28, 29. In contrast, Pahys et al. use 500 mg to explore the anti‐infection effects of intraoperative vancomycin in posterior cervical spinal surgery30. In the subgroup analysis of dosage, we included the studies that used 2 and 1 g, but did not include the subgroup using 500 mg; the values would have been null because only one such study met our inclusion criteria. Low‐to‐mild heterogenicity was found in the 1‐g subgroup (I ² = 27.2%, P = 0.186); however, in the 2‐g subgroup, the heterogenicity (I ² = 57.6%, P = 0.051) was obvious (Fig. 3).

Figure 3.

The subgroup analysis of the dosage of vancomycin powder used. The value of I ² in the 1‐g group is 27.2%, which is much lower than for the 2‐g group (I ² = 57.6%). The total I ² = 35.5%, P = 0.079. CI, confidence interval; OR, odds ratio.

Method of Administration

In the subgroup analysis of the method of administration, the results for vancomycin sprinkled directly into all layers of the wound showed no heterogenicity (I ² = 0.0%, P = 0.792)21, 22, 23, 24, 26, 27, 28, 30, 32. However, moderate heterogenicity (I ² = 60.0%, P = 0.007) existed in the subgroup in which the vancomycin powder was sprinkled directly into the muscle, fascia, and subcutaneous tissue but was not exposed to the bone graft or instrumentation (Fig. 4)9, 13, 15, 16, 25, 29, 31, 33, 34, 35.

Figure 4.

The subgroup analysis of the method of vancomycin powder administration. Method of administration: A, vancomycin was directly sprinkled into all layers of the wound (I ² = 0.0%, P = 0.792, OR = 2.70). Method of administration B, vancomycin was directly sprinkled into the muscle, fascia, and subcutaneous layers, but the bone graft and instrumentation were not exposed to it (I ² = 60.0%, P = 0.007, OR = 2.83).

Complications

The 20 studies reported few complications in performing this method, such as liver or renal failure, the red man syndrome, pseudarthrosis, and other local or systematic side effects. However, the potential risks were also mentioned by the researchers.

Publication Bias

The funnel plot showed that the small sample sizes might have been the dominant factor for the small bias observed in this meta‐analysis (Fig. 5). In addition, few articles with negative results were published in the databases, which might become another factor.

Figure 5.

Funnel plot diagram of the total included studies. This funnel plot diagram indicates that the bias was primarily attributed to many studies having small sample sizes, and negative results of the articles without publishing in the databases may become another factor.

Discussion

Wound infection following spinal surgery can cause many problems1, 2. Many methods to decrease the incidence of postsurgical wound infection following spinal surgeries have been used, including vacuum‐assisted closure, incisional wound vacuum‐assisted closure, prophylactic flap coverage, and betadine irrigation7, 8, 36, 37. Since Sweet and O'Neill first reported it, the prophylactic intraoperative use of vancomycin powder in the spinal wound site has been supported by the majority of researchers9, 35. The primary viewpoint is that this prophylactic way can reduce the rates of postoperative wound infection, decrease healthcare costs and resources used, and improve the quality of life during hospitalization. In addition, few side effects of intra‐wound vancomycin use have been revealed during spinal surgeries. Notably, the topical use of vancomycin powder appears to greatly decrease the incidence of postsurgical wounds following spinal surgeries in high‐risk populations with diabetes mellitus or of advanced age, and during instrumented surgery. However, other studies have shown no significant benefit of vancomycin for preventing or reducing postsurgical wound infections, and side effects may become problematic in the future with this prophylactic approach. Our systematic review and meta‐analysis revealed that this prophylactic intrawound administration can reduce the incidence of postsurgical wound infection, and the included studies did not report side effects. This study respectively reported different heterogenicity in the subgroups based on dosage and method of administration, even if the results of total analyses revealed low‐to‐moderate heterogenicity.

Clinical Studies for Intra‐wound Vancomycin in Various Spinal Surgeries

Most studies support the conclusion that intra‐wound vancomycin powder used in spine surgery can reduced the postoperative incidence of wound infection without obvious side effects. Armaghani et al. demonstrate local levels well above the MIC for common pathogens and serum levels below the toxicity threshold (25 μg/mL)38. However, three articles demonstrate that there is no significant difference (Table 2).

Table 2.

Clinical studies for intra‐wound vancomycin in various spinal surgeries

Year and authors	Location and types of surgery	LOE	Dosage (g)	Study Group		Control Group		P‐value	Reducing wound infection
				Size	ROWI	Size	ROWI
2013, Strom et al. 29	Cervical fusion	2	2	79	2.5%	92	10.9%	0.0384	Support
2013, Caroom et al. 32	Cervical fusion	2	1	911	0%	821	15%	0.007	Support
2015, Martin et al. 15	Cervical fusion	3	2	115	5.2%	174	6.9%	0.563	No‐support
2013, Pahys et al. 30	Cervical fusion	3	0.5	483	0%	195	1.86%	0.048	Support
2011, O'Neill et al. 34	Spinal fusion for traumatic injury	3	1	56	0%	64	13%	0.02	Support
2014, Theologis et al. 28	Adult deformity procedure	2	2	151	2.6%	64	10.9%	0.01	Support
2014, Martin et al. 16	Posterior spinal deformity surgery	2	2	156	5.1%	150	5.3%	0.963	No‐support
2011, Rahman et al. 35	Adult spinal deformity surgery	3	1–2	586	0.7%	334	5%	<0.0001	Support
2016, Gaviola et al. 20	Multilevel spinal fusion	3	2	116	5.2%	210	11.0%	0.08	Support
2013, Strom et al. 31	Lumbar laminectomy and fusion	3	1	156	0%	97	11%	0.0000182	Support
2015, Ross et al. 26	Lumbar fusion	3	1	70	0%	140	0.5%	0.1585	Support
2016, Lee et al. 21	Lumbar surgery	3	1	275	5.5%	296	10.5%	0.028	Support
2015, Schroeder et al. 24	Degenerative spinal surgery	3	1–1.5	1224	0.41%	2253	1.33%	0.04	Support
2016, Dennis et al. 22	Instrumented spinal surgery	3	1	117	0.9%	272	6.3%	0.049	Support
2015, Heller et al. 23	Instrumented spinal surgery	2	0.5–2	342	1.1%	341	3.8%	0.092	Support
2015, Devin et al. 25	Posterior spinal surgery	2	1	966	2.2%	1190	5.1%	<0.001	Support
2014, Hill et al. 27	Spinal surgery	3	0.5–2	150	0%	150	4%	0.0297	Support
2013, Kim et al. 33	Instrumented spinal fusion surgery	3	1	34	0%	40	12.5%	<0.033	Support
2011, Sweet et al. 9	Thoracolumbar fusion	3	2	911	0.2%	821	2.6%	<0.0001	Support
2015, Tubaki et al. 13	Spinal surgery	2	1	433	1.68%	474	1.61%	N	No‐support

LOE, level of evidence; N, null value; ROWI, rate of wound infection.

Suh et al. performed a retrospective study and declared that the use of intra‐wound vancomycin powder was ineffective for preventing SSI after instrumented lumbar fusion surgery, based on cultures of the wound drain, erythrocyte sedimentation rate, and C‐reactive protein levels14. Ghobrial et al. demonstrated that the incidence of gram‐negative or polymicrobial spinal infections increased after using prophylactic intraoperative vancomycin during various spinal surgeries, with negative culture outcomes for wound infections derived from degenerative disease but positive cultures for those following spinal trauma surgery39. In addition, minimal effectiveness of intra‐wound vancomycin powder use during spinal tumor surgery was revealed by Liu and Okafor40, 41. Okafor also revealed an infection rate of 4.9% after evaluating the use of intra‐site vancomycin powder during spinal tumor surgeries, and Liu revealed an insufficient benefit of it for patients undergoing spinal tumor surgeries (8.0% vs. 14.8%, P = 0.442).

In our meta‐analysis, we reviewed the published literature in accordance with the location of surgery, including cervical, lumbar, and other locations. For spine tumor surgery, almost all previous studies suggested that the local application of vancomycin powder in spine surgery can reduce the postoperative wound infection rate without any obvious side effects. In fact, these studies emphasize the possibility of potential side effects, such as the inhibition of osteoblast viability or bone fusion in future. Thus, a suitable concentration or dosage for decreasing the postoperative wound infection without any potential side effects should be investigated in the future.

Dose‐dependent Effects of Intra‐wound Vancomycin during Spinal Surgeries

Most studies reported a dosage of 2 or 1 g of vancomycin used during posterior spinal surgeries, and one included study confirmed the use of 500 mg of vancomycin powder for anterior cervical surgeries9, 13, 15, 16, 20, 21, 22, 23, 27, 28, 29, 32, 33, 34, 35. No side effects were attributed to intra‐site vancomycin powder use in retrospective clinical cohort studies, even though animal and in vitro studies report inhibition of osteoblast or dural cell viability. However, no large, randomized, prospective study has provided sufficient evidence of these complications. “New infections” refers to secondary infections developed after using topical vancomycin powder during spinal surgeries, and these cannot be ignored, although the mechanism underlying such new infections is still unclear (Table 1). There have been changes in the microbial trends of wound infections following the use of local vancomycin powder during spinal surgeries, and Ghobrial also demonstrated that the gram‐negative or polymicrobial bacterial genus might become the primary infected floras42. Other authors have also examined dose‐dependent or concentration‐dependent effects of prophylactic vancomycin use in in vitro studies43.

Moreover, the basic or in vitro studies may provide no direct evidence of such side effects. Claudia Eder et al. revealed that when the dosage used was >3 mg/cm², the dose‐dependent injury to osteoblasts in basic research studies on the human vertebral bone healing demonstrated significant inhibition of the cell proliferation and migration potential of osteoblasts and a decline in pH44. Godschmidt also indicated in an in vitro study that a concentration‐dependent inhibiting action existed on dural cell growth and fibroblast growth when the dose of intra‐wound vancomycin powder was >4000 μg/mL43. To date, no clinical trials have reported the dose‐dependent inhibition of osteoblast or dural cell viability.

Thus, the dose‐dependent microbial trends and inhibition effects of osteoblast or dural cell viability should be paid greater attention by researchers. Performing a systematic review with the previous data to investigate the relation of dose and possibility of side effects could provide valuable information for future study.

Methods of Intra‐site Vancomycin Powder Administration in Clinical Studies of Spinal Surgeries

The intraoperative use of local vancomycin varies for different spinal surgeries. There are two primary methods that can be used without any considerable side effects, especially for bone graft fusions. Some authors emphasize that the vancomycin powder should be placed directly on the deep wound and subfascial muscle tissues, but the bone graft, instrument, and dura should not be exposed to it9, 13, 15, 16, 25, 29, 31, 33, 34, 35. Other studies have shown no side effects on bone graft fusion in follow‐up observation, even if the bone graft was exposed to vancomycin powder20, 21, 22, 23, 24, 26, 27, 28, 30, 32; most studies have emphasized spraying or sprinkling vancomycin powder into the wound bed or deep fascial tissue.

The observed effectiveness without substantial side effects, documented in previous studies, has generated a considerable benefit in the application of intrawound vancomycin powder as a strategy for decreasing the incidence of postsurgical wound infection. Our hypothesis, supported by the systematic review and meta‐analysis, is that vancomycin powder should be sprayed into all layer of wounds, even if the basic research has demonstrated that higher concentration of local vancomycin powder has the ability to inhibit osteoblast and dural cell viability; however, none of related side effects have been reported in clinical studies.

Local Pharmacokinetics of Intrawound Vancomycin Powder in Orthopedic Surgery

At present, the pharmacokinetics of intra‐wound vancomycin powder in spine surgery remains unclear. Many studies have investigated the systematic and local drug concentration. Gans et al. surveyed the postoperative systematic vancomycin levels in plasma drawn on postoperative day (POD) 1 and on POD4 in 101 consecutive patients who received 500 mg of intra‐site vancomycin powder, who weighed over 25 kg, and who underwent spinal surgery, but excluded pediatric patients who received the growing rod expansion and vertical expandable prosthetic titanium rib45. The results of that study showed that serum vancomycin concentration in most cases had been drawn on POD1, and half of patients remained relatively steady until POD4; no cases involved the topical level of intra‐wound vancomycin powder. Armaghani et al. performed a retrospective review of pediatric patients with deformities who underwent the appropriate posterior spinal surgery and received 1 g of local vancomycin powder38. He reported that the serum levels in the immediate postoperative period averaged 2.5 μg/mL, with levels from 1.9 to 1.1 μg/mL on POD1 and POD2. The local vancomycin levels from the drainage were 403 μg/mL on POD0, followed by 251 and 115 μg/mL on POD1 and POD2, respectively. Sweet et al. also explored the serum and local vancomycin levels, and showed that systemic levels were not detected in most patients with a minimum sensitivity of 0.6 μg/mL, and the intra‐wound concentration was up to a 1000‐fold higher than the minimal inhibition concentration (MIC) for MRSA9. However, the authors did not reveal the mechanism of the pharmacokinetics of local vancomycin.

To date, the pharmacokinetics of intra‐wound vancomycin powder in total hip or knee arthroplasty seems to be well established. Johnson et al. investigated the pharmacological mechanism of local vancomycin powder in total hip arthroplasty (THA) and total knee arthroplasty (TKA) at three standardized times post‐operation: 3, 12, and 24 h. They estimated that the half‐life for intra‐wound vancomycin was 7.2 h using a longitudinal mixed model with a random intercept46. The paper revealed that 1 g vancomycin powder was used for the surgical site, and an additional 1 g of vancomycin powder was mixed with 2 g tranexamic acid and 50 mL saline. In fact, they injected the mixture retrograde through the drainage tube. This half‐value period of intra‐site vancomycin powder may not be suitable for spinal surgery; however, the research method can be used for designing trials in the future. At present, the pharmacokinetics of vancomycin powder applied locally to spinal wounds remain uncertain, as only one trial of intra‐site vancomycin pharmacokinetics has been performed, by The Washington University School of Medicine47.

Our meta‐analysis reported differences compared to the viewpoint of previous studies. Moreover, there exist some questions and limitations relating to using local vancomycin in various spinal surgeries, such as the lack of a relatively long or standard follow‐up period (at least 90 days), according to the Centers for Disease Control and Prevention (CDC) for human clinic research, and a comprehensive study combining radiographic assessment with laboratory analysis for the prospective random control trial using vancomycin in various spine surgeries. Investigating intra‐wound vancomycin use for high‐risk populations, the dose‐dependent polymicrobial trends of spinal surgical site infections, and the intra‐site pharmacokinetics of vancomycin in the wound site may become directions for future research. Finally, the multi‐level fusion or the different surgical site that without any analysis may become a deserving study in an advanced analysis.

Conclusion

Some limitations of this study cannot be ignored. For instance, there were some inclusive studies with small sample sizes, which could have caused relative heterogeneity. That there are few unpublished negative studies on this topic may also have caused a bias. Moreover, the quality of the studies included in our research was uneven, but most studies were LOE 3. Some studies with a relatively low level of evidence may have reduced the power of our results in our meta‐analysis. The selection bias from the database could also have influenced our results. In addition, all of the included studies emphasized that future large, randomized, prospective studies need to be performed, but did not reveal the dose relationships whether appeared the side effects for trends of polymicrobial or decreased the viability of osteoblasts and dural. Based on our meta‐analysis and systematic review, especially of the subgroup analysis of dosage used, the 500 mg and 2 g dosages may become topics for future studies. The administration of intra‐wound vancomycin powder may also become a future topic in relation to the dose‐dependent effectiveness of decreasing the incidence of postsurgical wound infection. In addition, the intra‐wound vancomycin used for high‐risk populations (e.g. for the postoperative leakage of cerebrospinal fluid in spinal surgeries), the time‐dependent effects in other orthopedic surgeries, the dose‐dependent polymicrobial trends of spinal surgical site infections, and the intra‐site vancomycin pharmacokinetics in wound sites may become hot spots for future study. Notably, a consistent rating scale needs to be applied in future research for high risk populations, which may be useful in formulating the guidelines for intraoperative use of vancomycin powder.