An excuse to misuse? Evaluating the use of prophylactic antibiotics in elective skin graft surgery: a systematic review and meta-analysis

Abstract

Introduction: 

Plastic surgeons may opt to administer prophylactic perioperative antibiotics in elective skin graft surgeries to curtail the risk of surgical site infections (SSI). However, the use of antibiotics brings the risk of adverse events or antibiotic resistance and evidence comparing the risk versus benefit is still lacking. Thus, there are no specific guidelines regarding the use of antibiotics prophylaxis in skin graft surgery and the potential for misuse of antibiotics is high.

Methodology: 

A systematic literature search was conducted to retrieve all articles comparing the use versus nonuse of “prophylactic antibiotics” in “skin graft surgery” on March 29, 2020. All comparative study designs were considered for inclusion. The primary outcome was occurrence of SSI. A meta-analysis using the Mantel-Haenszel method with random effect model was used to pool the effect estimates from included studies.

Result: 

A total of 1403 articles were identified, of which, 6 were included in the meta-analysis. The 6 articles included were all comparative studies (5 RCTs and 1 cohort study) and reported the outcomes of 391 patients and 404 grafts. A total of 202 patients received prophylactic antibiotics and 202 patients did not receive antibiotics. All 6 studies were included in the meta-analysis. The overall pooled effect estimates comparing treatment groups showed no significant difference for outcomes in terms of SSI (odds ratio: 0.43; 95% confidence interval: 0.18, 1.02; I² 0%; P=0.06).

Conclusion: 

Incidence of SSI in elective skin graft surgery is relatively low. Our data shows there is limited utility in the use of prophylactic antibiotics to prevent clinically significant SSI in patients undergoing elective skin graft surgery. However, there are still only a small number of studies conducted on the topic to date, and more work is required before definitive conclusions can be used.

Skin grafting is an essential skill in the plastic surgeon’s armamentarium and involves the surgical removal skin from a donor site, typically either an autograft (from the patient’s own body) or an allograft (another individuals’ body), and transferring it to a recipient site with either defective or deficient skin¹. Skin grafts have many indications, from trauma and burns to oncology, and function to restore the skin’s barrier functions or improve the cosmesis of the recipient site. Consequently, skin grafting can substantially improve a patient’s quality of life.

Surgical site infections (SSIs) are a dreaded outcome of any surgery, especially skin graft surgery. Indeed, SSIs are the second most common cause of skin-graft failure², and may lead to systemic infections or require additional revision surgery³. To prevent SSIs, many surgeons performing skin graft surgery opt to prescribe antibiotics in the perioperative period³. Use of prophylactic antibiotics may prevent infections at the surgical sites as well as at distal sites (eg, endocarditis)⁴. Antibiotics can either be administered routes topically at the donor and surgical sites, or systematically as orally with either intravenous injections intraoperatively and/or orally or in the perioperative or postoperative period⁵.

Currently, we are in the midst of an antibiotic crisis⁶; new strains of drug-resistant bacteria are being encountered with increasing frequency within wounds. Continued use of prophylactic antibiotics only further accelerates this crisis to a future where antibiotic therapy is ineffective for lab-proven wound infections. Furthermore, there are well-documented complications surrounding prolonged use of antibiotics³ which range from systemic side-effects like nephrotoxicity⁷ to anaphylaxis⁸. While skin grafts are used for emergent reconstruction of traumatic injuries such as degloving and burns, most commonly they are used in the elective setting (eg, for lesion removal), involving surgeries on clean skin, with low risk of contamination^6,9. This raises the question as to whether the perceived benefits of prophylactic antibiotics in elective skin graft surgery outweigh the risks.

Although recent research argues for more prudent use of antibiotics, guidance to plastic surgeons regarding elective operations has been lacking. Specifically, there remains no significant literature assessing the benefit of antibiotic prophylaxis in relation to skin graft surgery. The choice to give antibiotics for a given surgery is often left to the discretion of the operating surgeon. Recent polls suggest that most surgeons prescribe prophylactic antibiotics¹⁰ despite the lack of a supportive evidence base. Given the numerous indications for skin graft surgery, the potential for misuse of antibiotics is very high. We therefore conducted a systematic review and meta-analysis seeking to analyze and summarize the current literature published to date to assess how the benefits of antibiotic prophylaxis weigh in relation to skin graft surgery in order to provide clinical guidance as to the treatment of patients undergoing skin graft surgery.

Methods

This systematic review was conducted in line with recommendations specified in the Cochrane Handbook for Intervention Reviews V.5.1.0¹¹. This review is AMSTAR compliant and is reported in line with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement¹². Our protocol was published a priori¹³ and our systematic review registered was registered on Research Registry (www.researchregistry.com) (UIN: reviewregistry656).

Aim

To understand the benefits and drawbacks of antibiotic prophylaxis in skin graft surgery.

Inclusion criteria

The following inclusion criteria were used to enable a targeted analysis:

Study type

All comparative study designs [randomized controlled trials (RCTs), cohort studies, case-controlled studies] reporting original/primary data on one or more of the outcomes of interest. Unpublished data and reports were also considered for inclusion if the methodology and data were accessible. Duplicate articles, cost-effectiveness studies, studies not reporting on primary data (review articles, editorials, discussions, commentaries, letters) were not included.

Participants

Patients of all ages, sex, and ethnicities were included.

Interventions

The interventions of interest included all skin graft operations involving removal of lesions (either traumatic, congenital, malignant, benign), on any location in the body (head and neck, upper limb, lower limb, trunk). Reconstructions could be delayed or immediate, and involve either full-thickness or split thickness skin grafts. Studies reporting outcomes of cadaveric skin grafts were excluded. Skin grafts used for soft tissue coverage following amputation, ballistic or blast trauma, were excluded.

Comparators

Only studies comparing outcomes between patients receiving and not receiving prophylactic antibiotics were included. Prophylactic antibiotics were defined as antibiotics were used in the perioperative period to decrease the chance of having a postoperative SSI, and not to treat an ongoing infection. Studies not clarifying whether not reporting on the indication for skin graft surgery and not stating whether antibiotics were used prophylactically (vs. therapeutically, i.e., to treat superimposed infections either before or after skin graft surgery), were excluded.

Outcomes

The primary outcome was incidence of a SSI. Studies not reporting on SSIs in patients receiving skin graft surgery were excluded. Secondary outcomes, if reported, included: length of hospital stay (LOS), total wound healing time, whether revision surgery was required, cost of medical care, and graft take and cosmetic appearance. Results were deemed as significant where P<0.05. The outcomes were defined as follows:

• SSI—an infection that occurs at the grafted (not donor) site of the surgical procedure, from any point after surgery until the skin graft has fully healed.
• LOS—the length of time the patient resides in hospital after the procedure.
• Total wound healing time—time taken for surgical site to be completely healed.
• Revision surgery required—any surgery performed on the same recipient surgical site due to a suboptimal outcome from primary surgery.
• Graft take—the level of incorporation of the graft with recipient host tissue (eg, either measured as a percentage, or score on a visual analogue scale).

Search methods for identification of studies

The following electronic databases were searched from 1979 to 2020: PubMed, MEDLINE®, EMBASE, SCOPUS, CINAHL, PsychINFO, SciELO, The Cochrane Library, including the Cochrane Central Register of Controlled Trials (CENTRAL), Database of Abstracts of Reviews of Effect (DARE), the Cochrane Methodology Register, Health Technology Assessment Database, the NHS Economic Evaluation Databases and Cochrane Groups, ClinicalTrials.gov, Current Controlled Trials Database, the World Health Organization (WHO) International Clinical Trials Registry Platform, UpToDate.com, NHS Evidence and the York Centre for Reviews and Dissemination. The gray literature was also searched.

Search strategy and key terms

The search strategy was designed to identify articles focused on “use of antibiotics for skin graft surgery.” A search was conducted using appropriate keywords in English combined with Boolean logical operators as follows: ([antibiotics] OR [antibiotic]) AND ([skin graft] OR [graft survival] OR [surgical flap] OR [graft rejection] OR [skin) adapted to the appropriate syntax of each database. An example of the search strategy used on MEDLINE is shown in Table 1.

Table 1 - Example search strategy and search string given in the OVID database format.

Database: Ovid MEDLINE(R) ALL <1946 to March 29, 2020>

Search strategy:

1. antibiotic*.mp. [mp=title, abstract, original title, name of substance word, subject heading word, keyword heading word, protocol supplementary concept word, rare disease supplementary concept word, unique identifier, synonyms] (333609)

2. exp Anti-Bacterial Agents/ (661400)

3. skin graft*.mp. [mp=title, abstract, original title, name of substance word, subject heading word, keyword heading word, protocol supplementary concept word, rare disease supplementary concept word, unique identifier, synonyms] (16845)

4. skin transplantation/ (34524)

5. 1 or 2 (811431)

6. 3 or 4 (41819)

7. 5 and 6 (1516)

8. perioperative.mp. [mp=title, abstract, original title, name of substance word, subject heading word, keyword heading word, protocol supplementary concept word, rare disease supplementary concept word, unique identifier, synonyms] (86422)

9. 7 and 8 (28)

10. outcome*.mp. [mp=title, abstract, original title, name of substance word, subject heading word, keyword heading word, protocol supplementary concept word, rare disease supplementary concept word, unique identifier, synonyms] (2054474)

11. exp treatment outcome/ (878748)

12. 10 or 11 (2075484)

13. 7 and 12 (356)

14. skin graft/ (0)

15. exp skin graft/ (0)

16. exp antibiotic agent/ (0)

17. 1 or 16 (333609)

18. 1 or 16 (333609)

19. 3 or 15 (16845)

20. 18 and 19 (582)

21. exp treatment outcome/ (878748)

22. 20 and 21 (128)

Identification and selection of studies

• The articles identified from the electronic and manual searches were recorded into a preformatted Microsoft Excel 2017 (Microsoft, Redmond, WA) spreadsheet. Duplicates were excluded at this point.

Two researchers, acting independently, then screened articles for inclusion in 2 stages:

• Stage (1) titles and abstracts were screened.
• Stage (2) the full-text of articles selected in stage 1 were screened for inclusion.

If there was uncertainty at with titles and abstracts at stage 1, the full texts were retrieved for scrutiny at stage 2. Articles that met inclusion criteria at stage 2 proceeded to data extraction. Reasons for article exclusion at every stage were recorded.

Data extraction, collection and management

Data extraction was performed by 2 independently acting researchers. Discrepancies over the inclusion of any particular study were resolved by consensus or arbitration by a senior author. Data was input into a preformatted Microsoft Excel 2017 (Microsoft) database under standardized extraction fields to facilitate easy and consistent data entry.

For each article the following data was extracted:

• Article demographic details—number of authors, title, year published, journal level of evidence (1–5), conflicts of interest, funding.
• Patient demographic details—number of patients in total/in each group, mean follow-up (weeks), loss to follow-up (%), relevant comorbidities (vascular compromise, diabetes or other immunocompromising conditions).
• Indication for skin graft surgery—traumatic tissue loss, congenital deformity, lesion removal (malignant or benign).
• Surgical site—head and neck, upper limb, lower limb, trunk.
• Type of skin graft—split-thickness, full-thickness.
• Use of prophylactic antibiotics—“yes” or “no,” route [per oral (PO), intravenous (IV), topical (TOP)], regimen (preoperative, intraoperative, perioperative, postoperative), duration (in days), class of antibiotics used (Fluoroquinoloes, penicillins, macrolides, etc.).
• SSI—“yes” or “no,” treatment (conservative, medical, surgical).
• LOS—(days).
• Total wound healing time—(days).
• Revision surgery required—“yes” or “no.”
• Cost of medical care—(dollars).
• Graft take—as measured by each study (eg, % of incorporation/failure, “yes” or “no”).

Data analysis

Characteristics of included studies is presented as counts and percentages. Continuous data are expressed as weighted means differences with 95% confidence intervals (CIs). Categorical variables are expressed as odds ratios (ORs) with 95% CIs. A meta-analysis using a random effects model was conducted on Review Manager version 5.1.7 (The Cochrane Collaboration, Oxford, UK) to compare the odds of rates of SSI with and without use of prophylactic antibiotic. For all statistical comparisons, significance was set to P< 0.05. A κ value was also calculated to determine the agreement between the two independent reviewers.

Heterogeneity

Inter-study heterogeneity was explored for each variable using the χ² statistic. I² values were calculated to quantify the degree of heterogeneity across trials that could not be attributed to chance alone. Significant heterogeneity was considered present when I² >50%. Funnel plots were used to evaluate publication bias and thus assess data validity and heterogeneity.

Quality scoring

The Grading of Recommendation Assessment, Development and Evaluation (GRADE) system¹⁴ was used to assess the overall strength of evidence of included studies¹⁵. The GRADE system offers 4 levels of evidence: high; moderate; low; very low. RCTs are considered highest level of evidence. For RCTs the following will be assessed: (1) whether or not clinically relevant outcomes are reported; (2) whether results are comparable with protocols and subsequent publications where available. Key missing information across all study types such as follow-up times was documented and assessed.

Assessment of bias

Risk of bias was assessed using the Cochrane risk of bias tool¹⁶. All included articles be subjectively reviewed and assigned a value of “yes,” “no,” or “unclear” to the following questions: (i) Was the allocation sequence adequately generated? (ii) Was allocation adequately concealed? (iii) Was there blinding of participants, personnel, and outcome assessors? (iv) Were incomplete outcome data sufficiently assessed? and (v) Are reports in the study free of the suggestion of selective outcome reporting?

Results

Search results

The initial literature search yielded a total of 1403 records for review. Two articles were identified from the search but were Russian and Chinese. The authors were contacted but there was no response and these papers were therefore excluded. After duplicates were removed, review of abstracts and titles (stage 1) identified 85 articles for full-text screening (stage 2). A total of 6 articles were included in the final selection for data extraction, shown in Table 3. The main reasons for article exclusion at stage 2 were; “no skin grafting” (n=39, 49%), “incorrect use of antibiotics” (n=29, 37%), “incorrect study design” (n=5, 6%), “not primary data” (n=1, 1%) and “failure to report outcomes” following skin graft surgery (n=5, 6%). Record selection and reasons for exclusion of articles at stage 2 are shown in Figure 1.

Figure 1: PRISMA flow diagram showing article exclusion at each stage and reasons for exclusion at full-text stage. ABX indicates antibiotics.

Article demographics

The κ score indicated a high level of agreement between data collectors (κ=0.823). The included studies report on outcomes published between 2006 and 2020 in 5 different countries; Sweden, Australia, the United States, Argentina, and The Netherlands, The highest level of evidence of our included studies was level 1 (RCT, n=5). The majority of studies (83%), were level 1b (n=5, randomized control trials) and one was a retrospective cohort study. The articles were generally of moderate quality; GRADE scores ranged from high with strong recommendation for prophylactic antibiotics to low with strong recommendation for use of prophylactic antibiotics (Table 2). Assessment of bias using the Cochrane bias criteria, revealed overall low risk of bias among the randomized control trials and a moderate risk of bias in the retrospective cohort study (Table 3).

Table 2 - Article demographic information including GRADE assessment^17–22.

References	Country	Study Design	LoE	GRADE	No. Patients/Group	Mean Age/Group	Main Indication Skin Graft/Group	Donor Site/Group	Type of Skin Graft/Group	Which ABX & How Administered	Graft Recipient Site	Length of Stay (d)
Kuijpers et al18	Belgium	RCT	1b	High strong for ABX	Total: 30 16—ABX 14—no ABX	Total: 68.95 69.3—ABX 68.6—no ABX	Reconstruction nonmelanoma skin cancer nose	Pre and retro auricular, upper arm/preauricular, glabella	FTG	Azithromycin 500 mg OD PO	Nose	Not stated
Saleh et al20	Sweden	Retrospective cohort study	4	Low strong for ABX	Total: 18 9—ABX 9—no ABX	Total: 76.7	1 actinic keratosis, 2 SCC, 15 BCC	Not stated	FTG	Flucloxacillin 1.5 g 3 d PO	Nose	Not stated
Rosengren et al21	Australia	RCT	1b	High weak against Abx	Total: 8 3—ABX 5—no ABX	Total: 69.4 69.4—ABX 68.4—no ABX	BCC—ABX BCC—no ABX	Not stated	STG	One dose Cefalexin 2 g PO	Below knee	Not stated
Rosengren et al22	Australia	RCT	1b	High strong for Abx	Total: 9 3—ABX 6—no Abx	Total: 66.3 66.6—ABX 65.9—no Abx	BCC—ABX BCC—no ABX	Not stated	Not stated	One dose Cefalexin 2 g PO	Nose and ear	Not stated
Alexander et al17	USA	RCT	1b	High strong for ABX	Total—249 127—ABX 122—no ABX	Total—10.6 10.5—ABX 10.8—no ABX	Burns	Not stated	STG or FTG	Cephalothin 15mg/kg in 50 ml 5% dextrose preoperatively, immediately before skin incision and 4 hours after skin incision	Not stated	ABX—12.38 (0.29) Non ABX—13.66 (0.46)
Ramos et al19	Argentina	RCT	1b	High weak for ABX	(presented by grafts, 77 patients) Total—90 44—ABX 46—Non ABX	41.7	Burns	Head and neck, upper limb, lower limb, chest, back, others	Not stated	Cephalothin, 13 (29) Vancomycin, 13 (29) Minocycline, 6 (14 Ciprofloxacin, 6 (14) Piperacillin/tazobactam, 5 (11) Amikacin, 5 (11) Trimethoprim/sulfamethoxazole, 2 (4) Imipenem, 2 (4) Ceftazidime, 1 (2)	Not stated	The mean length of stay in the burn unit was 34.8 d (median, 28.5; IQR, 11.2–43.7 d)

ABX indciates antibiotics; IQR, interquartile range; LoE, level of evidence; BCC, basal cell carcinomas; FTG, full thickness graft; RCT, randomized controlled trial; STG, split thickness graft; SMX, TMP sulfamethoxazole-trimethoprim.

Table 3 - Cochrane risk of bias^17–22.

References	Random Sequence Generation	Allocation Concealment	Blinding Participants and Personnel	Blinding of Outcome Assessment	Incomplete Outcome data	Selective Reporting	Overall Bias
Randomized trials
Kuijpers et al18	Low	High	High	Low	Low	Low	Moderate
Rosengren et al21	Low	Low	Low	Moderate	Low	Low	Low
Rosengren et al22	Low	Low	Low	Moderate	Low	Low	Low
Alexander et al17	Moderate	Moderate	Low	Low	Low	Low	Low
Ramos et al19	Moderate	Low	Low	Low	Low	Low	Low
Reference	Confounding	Selection	Information	Reporting	Overall risk of bias
Nonrandomized trials
Saleh et al20	High	High	Low	Low	Moderate

Patient demographics

The total number of patients across all six articles included studies was 391. Patients had a mean weighted age of 70.3 years (ranging 10.5 y to 76.7).

Interventions

Outcomes are reported on a total of 404 skin grafts. The main indications for skin graft surgery was burns (n=339, 84%). Skin grafting was also used in the setting of lesion removal and specifically removal of squamous cell carcinoma or basal cell carcinomas (n=65, 16%). The most common donor site was the ear (n=28, 7%) and the most common recipient site was the nose.

Comparators

A total of 202 (50%) grafts received prophylactic antibiotics and 202 (50%) did not. The most common prophylactic antibiotic and dosing regimen used was single dose 2 g cefalexin PO (n=2).

• Indication for skin graft surgery—traumatic, congenital, malignant (SSC, BSC, melanoma), benign.
• Surgical site—head and neck, upper limb, lower limb, trunk.
• Type of skin graft—split-thickness, full-thickness.
• Use of prophylactic antibiotics—“yes” or “no,” route (PO, IV, TOP), regimen (preoperative, intraoperative, perioperative, postoperative), duration (in days), class of antibiotics used (Fluoroquinoloes, penicillins, macrolides, etc.).

Primary outcome: SSIs

The total number of SSI in the included studies was 25 (6%), the incidence of SSI in patients not receiving antibiotics was 8.9% (n=18) and 3.5% (n=7) in patients receiving prophylactic antibiotics (Table 4). Of the SSI reported, 18 (72%) occurred in patients who did not receive prophylactic antibiotics and 7 (28%) were in patients who received prophylactic antibiotics. The overall pooled effect estimates comparing treatment groups showed no significant difference for outcomes in terms of SSI (relative risk: 0.70, 95% CI: 0.27, 1.82) (Fig. 2).

Table 4 - Study outcomes^17–22.

References	SSI (Total, ABX, Non ABX)	Graft Take/Survival (Mean % ±SD)	Cosmetic Appearance	Bacterial Load	Length of Stay (d)	Adverse Effects
Kuijpers et al18	Total: 0 ABX: 0 Non ABX: 0	ABX: 86.6 Non ABX: 36.2	Not stated	Not stated	Not stated	Not stated
Saleh et al20	Total: 8 ABX: 3/9 Non ABX: 5/9	Not stated	Not stated	Significantly higher in non ABX	Not stated	Not stated
Rosengren et al21	Total: 1 ABX: 1 Non ABX: 0	Not stated	Not stated	Not stated	Not stated	ABX: 0 Non ABX: 1 (nausea)
Rosengren et al22	Total: 2 ABX: 0 Non ABX: 2	Not stated	Not stated	Not stated	Not stated	ABX: 1 (nausea) P=1 Non ABX: 0
Alexander et al17	Total: 8 ABX: 1 Non ABX: 7	ABX: 99.9 (0.1) Non ABX: 98.9 (0.4)	Not stated	Not stated	ABX—12.38 (0.29) Non ABX—13.66 (0.46) Non infected: 12.8 (0.27) Infected: 19.25 (1.3)	ABX: 1 Non ABX: 1 Both patchy, cutaneous erythema
Ramos et al19	Total: 11 ABX: 4 Non ABX: 7	ABX: 97 Non ABX: 87	Not stated	Not stated	The mean length of stay in the burn unit was 34.8 d (median, 28.5; IQR, 11.2–43.7 d)	Lower colonization levels in ABX

ABX indciates antibiotics; IQR, interquartile range; SSI, surgical site infections.

Figure 2: Forrest plot showing odds ratio for the incidence of surgical site infections between patients receiving prophylactic antibiotics and those not receiving antibiotics. ABX indicates antibiotics; CI, confidence interval.

Secondary outcome

LOS in hospital was measured in 2 studies and was between 12.38 and 34.8 days, Alexander et al¹⁷ found that length of stay was higher in infected (19.25 d) than noninfected patients (12.8 d). Graft take was measured in 3 studies. Kujpers et al¹⁸ found mean graft take at 10 days postoperative to be higher in patients receiving prophylactic antibiotics (86.6%) than patients not receiving prophylactic antibiotics (36.6). Alexander et al¹⁷ demonstrated graft take of 99.9% in patient taking prophylactic antibiotics compared with 98.9% in patients not taking prophylactic antibiotics. Ramos et al¹⁹ show graft take was 10% higher in patients taking prophylactic antibiotics (97%) than patients not taking prophylactic antibiotics (87%) Kujpers et al¹⁸ reported no SSI in patients receiving or not receiving prophylactic antibiotics, however, showed graft survival to be significantly higher in patients who received prophylactic antibiotics (P=0.002). Saleh at al²⁰ demonstrated bacterial loads of surgical wounds were significantly higher in patients who did not receive prophylactic antibiotics (P=0.02) and a significant positive correlation between bacterial load and postoperative complications including SSI (P<0.01). Ramos et al¹⁹ report lower levels of wound colonization in patients taking prophylactic antibiotics. Adverse effects were reported in 3 studies. Rosengren et al²¹ found 1 patient in the nonantibiotic group who reported nausea. Rosengren et al²² report 1 incidence of nausea and 1 of diarrhea in the antibiotic group, and 1 incident of nausea in the nonantibiotic group. Alexander et al¹⁷ showed 2 episodes of patients reporting patchy, cutaneous erythema, 1 in each group (intervention and control).

Discussion

We conducted a systematic review and meta-analysis to assess whether the benefits of antibiotic prophylaxis in skin graft surgery outweigh the risks. The ultimate aim is to provide clinical guidance as to the treatment of patients undergoing skin graft surgery. We included a total of 6 papers reporting on outcomes of 404 skin grafts. Overall, we found no evidence to support the use of prophylactic antibiotics in skin graft surgery.

It is perhaps not surprising that prophylactic antibiotics had little effect on graft take or success. Skin graft surgery is typically an elective surgery and is considered “clean” and with a low inherent risk for SSI. Use of prophylactic antibiotics in other “clean” surgical procedures has also not demonstrated to be of significant benefit²³. Despite knowing that prophylactic antibiotics may provide little benefit, many plastic surgeons continue to use antibiotics prophylactic, possibly out of a fear of graft failure and the need for potential revision surgery if a skin graft fails²⁴. Interestingly, Kujpers et al¹⁸ demonstrated a significantly higher rate of graft survival (but not SSI) in patients taking antibiotics postoperatively, suggesting prophylactic antibiotics may justified to prevent graft failure possibly by preventing subclinical infection of the skin graft. Further research is required to establish this property. While wound beds with reduced bacterial colonization exhibit improve graft survival, there are a number of alternatives to antibiotics which may be equally or more effective in eradicating bacteria at the sugical site and do not carry the same risk. Iodine containing products, like betadine ointment, have proven antiseptic qualities in wounds and are associated with minimal complications²⁵. Honey and silver have also displayed success in preventing skin wound infection²⁶. Indeed, silver impregnated dressings are widely used in plastic and dermatological surgery and present a viable alternative to antibiotics²⁷. Studies have even demonstrated silver released in a moist wound surface environment significantly increases formation of new epithelial tissue in skin grafts²⁸. These options provide appropriate alternatives to antibiotics as they mitigate chances of subclinical infections and enhance graft survival, while also having a more favorable side-effect profile. The side-effect profile from antibiotics is well-documented and includes symptoms ranging from nausea and vomiting to life threatening complications²⁹. In our data, only 2 patients in 2 separate studies experienced side effects from antibiotics taken prophylactically; one patient experienced nausea¹⁷, and a second patient presented with patchy, cutaneous erythema²¹. Interestingly, in this second study, a control patient (not receiving antibiotics) had a similar episode of cutaneous erythema²¹. Although adverse effects related to prophylactic antibiotics were only reported in 2 patients, it is possible that many patients experienced subclinical adverse effects that were never documented. Future work may consider a rigorous method for reporting possible subtherapeutic effects.

A secondary aim of this study was to also investigate whether antibiotic prophylaxis may be more important in certain surgeries or patient groups. Unfortunately, there were too few patients and too few studies with homogenous outcomes which precluded a mathematical synthesis of data. Furthermore, the nonantibiotics group in the study by Kujpers et al¹⁸ included 4 (13%) of the 5 smokers included in the study, which could have skewed the results. Smokers in the study had a mean graft survival of 2% compared with 75.9% for nonsmokers on the second follow-up visit. It is well known that nicotine impairs healing through vasoconstriction, leading to a reduction of blood flow, and a higher risk of ischemia due to increased platelet activity. Smoking may also alter the microbiome of the wound, and detrimentally impacts the proliferation and migration of blood cells and fibroblasts. All these processes are essential for normal wound healing^30,31. Another secondary aim was to investigate whether prophylactic antibiotics are more essential for surgeries on the lower versus upper body. Lower limb surgeries are seen as “high risk” and “dirty,” often being colonized by bacteria such as methicillin resistant Staphylococcus aureus. Antibiotic prophylaxis may therefore theoretically have a more important role maintaining optimum levels of tissue colonization before surgery^32,33. An RCT comparing the incidence of SSIs following skin lesion excision in the lower limb with compared to without prophylactic antibiotics showed a relative reduction of 65% in the group taking antibiotics³². The lower body and the lower limb especially have a greater variation in circulation compared to the upper body which is relatively well perfused. In an elderly population with a degree of venous insufficiency this may be a greater factor in determining success or failure of skin graft. Studies suggest that factors related to blood flow such as body max index may have a greater role in skin graft success that SSI³⁴. No studies compared the incidence of SSI between upper and lower body unfortunately, and this is an interesting question for future research.

Antibiotic stewardship is a hot topic and has featured publications across many surgical subspecialties. A recent systematic review and meta-analysis of urologic procedures reported that there is limited utility in postoperative prophylactic antibiotics, and prophylactic antibiotic treatment may be associated with unpleasant side effects, allergic reactions, and development of antibiotic resistance⁸. Studies in cardiothoracic surgery have also highlighted the importance of proper antibiotic stewardship. A 4-year cohort study comparing outcomes after Coronary Artery Bypass Grafting surgery showed that many patients were receiving prophylactic antibiotics for over 48 hours despite this showing no reduction of SSI³⁵. As a result of indiscriminate antibiotic prescribing, antibiotic resistance is increasing, leading to significant morbidity and mortality globally²¹. Increased use of antibiotic prophylaxis may accelerate the increase in resistant bacteria and subsequently more challenging SSIs in the future. Furthermore, antibiotics are not without potential adverse effects. A study of adverse effects of antibiotic prophylaxis in ambulatory cystoscopy showed that up around 10% of patients may experience systemic symptoms such as; diarrhea, malaise, abdominal pain and, hypersensitivity reactions^36,37. In our data, the groups of patients receiving antibiotics were not free of complications. One patient reported nausea and another patient reported a cutaneous erythema; however, it is possible that other adverse effects occurred but were not reported.

The main limitations of this manuscript include the limited number of studies in general, the paucity of comparative studies, and the inclusion of only a small number of patients in each of the studies included. In addition, data published on this topic to date has been of low quality and at high risk of bias. Furthermore, the incidence of SSI is low, with often no patients in 1 study group experiencing an SSI complication. In addition, data pertaining to donor and recipient sites, length of dosage and patient comorbidities was not always available. More and higher-quality comparative studies such as RCTs reporting on SSI, graft take, and esthetic outcomes must be conducted in this area in order to better understand the dynamics of wound healing and the role, if any, of antibiotic prophylaxis in order to help clinicians choose the best perioperative care for their patients. Furthermore, comparative studies using different antibiotics will allow clinicians to determine their efficacy and availability and will aid proper antibiotic stewardship in the future. Future work is also required to understand the vascular dynamics in patients with known risk factors such as smokers and the obese, or to understand whether site of surgery (eg, upper or lower limb) is at more risk of having an SSI. Despite these limitations, these data suggest that the benefits of prophylactic antibiotics in elective skin graft surgery do not outweigh the risks. It is surprising given this highly topical and controversial topic that only few comparative studies have addressed this question to date, and may reflect how plastic surgeons are guided by their own complication rates or by a fear of litigation or accusation of malpractice. Indeed, there is an interesting disparity of who (patient vs. surgeon) experiences regarding the consequences of prophylactic antibiotics or versus no antibiotics. Regardless of the reason, clearly more information is required to develop gold standards and ensure all patients are receiving the highest quality of data-supported care possible.

This review provides an up-to-date summary of the use of prophylactic antibiotics in skin graft surgery. We demonstrate statistically that antibiotics do not decrease the incidence of SSI. Furthermore, antibiotics are thought to increase graft survival through eradication of microscopic surgical site bacterial, however, we show no increased survival rates with antibiotics. Given their extensive side-effect profile, the use of prophylactic antibiotics is not justified by the evidence to date. With growing concerns over antibiotic resistance and higher quality antibacterial stewardship taking place more high-quality evidence will be published prompting an updated review of this topic in the future. We hope this data will help clinicians decide against prescribing antibiotic as a prophylactic measure and will encourage more sparing use of antibiotics amidst growing bacterial resistance. Additional articles reporting on SSI, graft survival, and esthetic outcome will further guide future practice.

Ethical approval

Not required for an audit.

Sources of funding

This work did not receive any funding.

Author contribution

M.R.B. conceived of the idea. M.R.B., V.S., and M.L.L. drafted the manuscript. R.A., K.E., and A.M.R. reviewed the final manuscript.

Conflicts of interest disclosure

The authors declare that they have no financial conflict of interest with regard to the content of this report.

Research registration unique identifying number (UIN)

reviewregistry656.

Guarantor

Mimi R. Borrelli.