Since venous competence linearly decreases with age, more than half of the adult population has some symptoms of venous incompetence, and the prevalence of chronic venous disease (CVD) has increases substantially (1,2). The discomfort associated with CVD significantly impacts patients’ health-related quality of life (QoL), which consequently affects health care costs. The surgical treatment of incompetent veins can reduce symptoms and improve the QoL of patients (3,4).

In the past, the standard method for surgical treatment of CVD, including varicose veins, was stripping surgery. Post-operative surgical site wound infection (SSWI) following stripping surgery is reported to vary from 0 to 7.5%; therefore, the use of prophylactic antibiotics was justified to reduce SSWI (5,6). Endovenous procedures for treating incompetent veins have gained popularity due to their proven effectiveness and safety. The minimally invasive nature of endovenous ablation does not necessitate routine prophylactic antibiotics for endovenous procedures (7). Recently a major guideline was revised, but no comment on the recommendation for prophylactic antibiotic use was made (8). Whether there is a need for prophylactic antibiotics in cases of combined endovenous ablation with concomitant phlebectomy needs to be clarified.

Therefore, we conducted a retrospective analysis to evaluate the efficacy of prophylactic antibiotics after combined endovenous surgery and phlebectomy.

Our primary aim was to compare the incidence of SSWI after the combined procedure between the prophylactic antibiotic use and non-use groups. The secondary outcomes were to analyze antibiotic requirement due to any cause three months postoperatively and determine the incidence of thrombophlebitis requiring antibiotic treatment. As our study end-points required suitable comparison for “outcomes measurement of two groups” as well as investigation of the efficacy for “treatment,” a randomized controlled trial (RCT) would be the best methodology to obtain the highest level of evidence. However, prior to conducting an RCT, we tried to estimate the efficacy of each treatment strategy by carrying out a single-center registry-based case-control study (9). By using the propensity score matching method, we tried to minimize the selection bias that may occur between groups. Therefore, this was a propensity score-matched case-control study that retrospectively analyzed prospectively designed case report form (CRF) data from August 2018 to January 2022 in a venous disease-specific center. The CRF consisted of 100 informational items related to CVD for each patient. The National Institute of Review Board and Ethics Committee approved the study protocol (number: 2022-0200-001), and all patients provided written informed consent. Because our treatment goal for CVD is QoL improvement through venous symptom relief regardless of the presence of varicose or reticular veins, indications for surgery were subjects with truncal saphenous vein reflux (>1 s on duplex ultrasound [DUS]) with persistent venous symptoms that could not be resolved despite a period of conservative treatment. Every symptom was reported by its presence or absence using Villalta–Prandoni’s five categories: heaviness (fatigue), calf cramping, pain, itching, and numbness (10). CVD-related complications may affect the primary outcome; therefore, it was defined as one or more of the following events: history of spontaneous bleeding, superficial vein thrombosis, lipodermatosclerosis, skin pigmentation, eczema, ulcer, routine pain killer medication, and deep vein thrombosis.

Study participants included only those who underwent the combined endovenous procedure with concomitant phlebectomy. A total of 587 patients (mean age of 55.8±14.2 and 296 (50.4%) women) classified as CEAP C0∼C6, Ep (primary), or As(superficial), and Pr (reflux) were eligible for the present study. In the prophylactic antibiotics group, a 675 mg single dose of amoxicillin-clavulanic acid was administered orally 2 hours before the procedure. After propensity score matching, 201 patients who received prophylactic antibiotics (PA group) and 201 who did not (nPA group) were selected, and the present study analyzed 402 study participants (Fig. 1).

Figure 1. STROBE flow of the study.

Patients were asked by an independent investigator to use a visual analog scale (VAS, 0∼10 points) to describe their perioperative pain. Obtained VAS pain scores were recorded blindly to reduce information bias. All patients were followed up at one day, one week, one month, and three months. The follow-up protocol included an interview to obtain information about phlebectomy wound status, changes in their QoL, clinical assessment, and a DUS examination. When the remanent varicosity or perforator was considered problematic or assumed to be associated with remanent CVD symptoms or signs during follow-up, additional phlebectomy was considered. SSWI was defined as reddish swelling at the site of an endovenous or phlebectomy procedure and pus formation or neutrophil dominant leukocytosis with chronic reactant protein increase confirmed by laboratory examination. If painful swelling was noted over the superficial vein where the two concomitant procedures were performed on ultrasound examination, and non-organized thrombus extraction, without pus, was possible through phlebotomy followed by subsequent symptom improvement, the condition was judged to be thrombophlebitis.

1) Antiseptic preparation

In this study cohort, all patients were instructed to shower for at least 10 minutes with soap on the morning of the operation. There were no restrictions on the continuous use of previously prescribed drugs. In the PA group, an oral amoxicillin-clavulanic acid tablet was administered at least two hours before the surgery. The operation was performed on an outpatient basis, and after the patient entered the procedure room, the lower body was disinfected using warm betadine. After that, the skin was dried for five minutes to increase the iodine activity. After all the procedures were performed, the surgical wound was covered with only an adhesive bandage, even in the cases of suture site.

2) Combined endovenous procedure with concomi-tant phlebectomy

The selection of the primary endovenous treatment modality type, either cyanoacrylate glue ablation (CA), endovenous laser ablation (EVLA), radiofrequency ablation (RFA), or ultrasonography guided foam sclerotherapy (UGFS), depended entirely upon the patient’s own decision after procedure-related information was provided. All procedures were conducted under local anesthesia without any level of sedation. The combined procedure applied in the present study population consisted of two steps. The first step was truncal vein obliteration from the saphenofemoral junction to the below the knee level of the great saphenous vein (GSV) or from the saphenopopliteal junction (SPJ) to the mid-level of the calf in the small saphenous vein (SSV) using an endovenous technique. The second step was concomitant phlebectomy for the varicosity of remanent tributaries of the GSV or pathologic perforating vein of the distal calf SSV to complete the venous reflux treatment.

We used CA glue (VenaSeal^TM Closure System, Medtronic, USA) in compliance with its instructions for use. However, we did not inject VenaSeal^TM into the below-the-knee (BTK) area of the GSV to prevent unintended glue spillage into the tributary or perforating veins. If additional sclerotherapy was required after a concomitant phlebectomy, another puncture was made in the target tributary vein area for UGFS. Foam sclerosant (0.75%) was prepared using sodium tetradecyl sulfate (STS) (Fibrovein; STD Pharmaceuticals, Hereford, UK), and 1.5 cc was injected into each target area. Our safe limit was 3 cc for each limb.

In the case of EVLA, we used a 1,470 nm diode laser and tried to ensure at least a 4 mm thick tumescent infiltration on the ultrasonic cross-section. The laser fiber was steadily pulled back at a speed of 1 mm/s, and at least 45 J/cm of energy was delivered. No heat ablation was used in the BTK area to prevent sensory nerve injury. Concomitant phlebectomy and additional UGFS were performed in the BTK area as in the CA procedure. RFA procedure was similar to EVLA in that it required tumescent anesthesia with a saphenous nerve protection strategy. When the patient required UGFS for endovenous truncal vein obliteration, we applied 1.5% of foam STS 4∼6 cc for each truncal saphenous vein. In these cases, the phlebectomy was conducted for high ligation of the saphenofemoral or saphenopopliteal junction.

Active ambulation was encouraged immediately after each procedure. All daily activities should have been possible. A compressive elastic stocking (20∼30 mmHg) was worn for at least 24 hours regardless of the primary procedure type. Routine postoperative antibiotics were not administered.

3) Propensity score matching and statistical analysis

All data were analyzed using SPSS 20.0 (IBM, New York, NY, USA) and Excel 16.0 (Microsoft, New York, NY, USA). We applied propensity score matching analysis to make the two groups comparable with respect to the preoperative demographics, clinical status, and surgical type and burden. The PA group was considered the treatment group, and the nPA group was considered the control group. The propensity score for each eligible case was estimated using logistic regression based on covariates including sex, age, CEAP clinical class, type of endovenous procedure, and the number of phlebectomies. The probability scores for each participants obtained after logistic regression analysis were stratified. Subsequently, a one-to-one match between the PA and nPA groups was obtained using nearest-neigh-bor matching in Excel. A total of 402 samples were selected at a 1:1 ratio. The percentage of patients who showed post-operative SSWI was analyzed between the PA and nPA groups using Fisher’s exact test to confirm the significance. All P-value calculations were two-sided, and the alpha level for all analyses was set to p<0.05.

4) Manuscript preparation

This paper was written based on the STROBE statement checklist of items that should be included in reports of observational studies published in 2007 (11). Based on reporting standards for CVD and endovenous ablation, we prepared this manuscript according to the recently revised American Venous Forum (AVF) CEAP task force report, the Society of vascular surgery (SVS) document for adverse events after medical devices, and the joint statement of the AVF and the Society of Interventional Radiology (12,13).

1) Unadjusted data analysis

A total of 587 patients were included after the eligibility test. The distribution of symptomatic CEAP clinical classes was similar in both groups (p=0.11), and class 2 was the most prevalent in the study group (166, 64.3% vs. 199, 60. 5%, p=ns). In the univariate analysis, demographic findings of sex, age, GSV size, and calf circumference were comparable in both groups. However, the number of recurrences, CVD-related complications, concomitant phlebectomies and types of endovenous procedures were statistically different between the PA and nPA groups (Table 1).

Table 1 . Demographic and treatment characteristics between PA and nPA group in unadjusted data.

	nPA, n=258	PA, n=329	p
Age, years	56.2±13.8	55.6±14.6	NS
Male sex: no. (%)	120 (46.5%)	171 (52.0%)	NS
Calf circumference	36.2±3.3	36.7±3.2	NS
GSV size (mean)	4.84±2.3	5.2±2.4	NS
Recurrence	9 (3.5%)	24 (7.3%)	0.047
Calf circumference	36.40±3.2	36.37±3.0	NS
Diabetes	26 (10.1%)	17 (7.5%)	NS
PAOD	4 (1.6%)	5 (1.5%)	NS
Combined SSV treatment	117 (45.3%)	145 (44.1%)	NS
No. Of CVD related complication	75 (29.1%)	129 (39.2%)	0.01
Main Endovenous Procedure
VenaSealTM	13 (5%)	14 (14.0%)	0.001
RFA	110 (42.6%)	96 (29.2%)	0.001
EVLT	08 (41.9%)	142 (43.2%)	NS
UGFS	27 (10.5%)	45 (13.7%)	NS
No. Of phlebectomy	3.27±2.3	4.04±2.96	0.001
CEAP Clinical Class
0	18 (7.0%)	16 (4.9%)	0.113
1	5 (1.9%)	16 (4.9%)
2	166 (64.3%)	199 (60.5%)
3	35 (13.6%)	37 (11.2%)
4	34 (13.2%)	57 (17.3%)
5	0	1 (0.3%)
6	0	3 (0.9%)

GSV: great saphenous vein, SSV: small saphenous vein, PAOD: peripheral arterial occlusive disease, CVD: chronic venous disease, RFA: radio frequency ablation, EVLT: endo venous laser ablation, UGFS: ultrasonography guided foam sclerotherapy..

2) Adjusted data analysis

After the propensity score matching, demographic factors and the preoperative and operative condition were evenly adjusted between the two groups (Table 2). The mean number of concomitant phlebectomies was 3.54±2.3 in the nPA group and 3.68±2.40 in the PA group (p=ns), and their immediate postoperative VAS pain scores were not statistically different (3.55±2.3 vs. 3.69±2.3, p=0.65. non-parametric Mann–Whitney U test).

Table 2 . Demographic and treatment characteristics between PA and nPA group in propensity score matched adjusted data.

	nPA, n=201	PA, n=201	p
Age, years	56.4±14.1	56.6±14.6	NS
Male sex: no. (%)	96 (47.8%)	106 (52.7%)	NS
Calf circumference	36.2±3.3	36.6±3.2	NS
GSV size (mean)	4.87±2.5	5.2±2.2	NS
Recurrence	6 (3.0%)	11 (5.5%)	NS
Calf circumference	36.40±3.2	36.37±3.0	NS
Diabetes	22 (10.9%)	8 (7.0%)	NS
PAOD	3 (1.5%)	3 (1.5%)	NS
Combined SSV treatment	93 (46.3%)	79 (39.3%)	NS
No. Of CVD related complication*	64 (31.8%)	66 (32.8%)	NS
Main Endovenous Procedure
VenaSealTM	8 (4%)	10 (5.0%)	NS
RFA	66 (32.8%)	76 (37.8%)	NS
EVLT	108 (53.7%)	98 (48.8%)	NS
UGFS	19 (9.5%)	17 (8.5%)	NS
No. Of phlebectomy	3.54±2.3	3.68±2.40	NS
CEAP clinical class
0	13 (6.5%)	8 (4.0%)	NS
1	2 (1.0%)	8 (4.0%)
2	132 (65.7%)	132 (65.7%)
3	31 (15.4%)	19 (9.5%)
4	23 (11.4%)	33 (16.4%)
5	0 (0.0)	1 (0.5%)

*Bleeding, thrombosis, pigmentation, lipodermatosclerosis, static ulcer, painkiller medication, deep vein thrombosis, eczema..

GSV: Great Saphenous Vein, SSV: Small Saphenous Vein, PAOD: Peripheral Arterial Occlusive Disease, CVD: Chronic Venous Disease, RFA: RadioFrequency Ablation, EVLT: EndoVenous Laser Ablation, UGFS: Ultrasonography Guided Foam Sclerotherapy..

3) Primary and secondary outcomes measurement (Table 3)

SSWI was identified in five patients, and the incidence was similar in both groups (4, 2.0% vs. 1, 0.5%. p=0.177 by Fisher’s exact test). SSWIs were found in four phlebectomy sites and one EVLA lesion above the knee GSV. All patients required intravenous antibiotics and wound debridement. However, no patient needed hospital admission.

Table 3 . Primary and secondary outcomes of the present study.

Primary outcomes	nPA, n=201	PA, n=201	p
SSWI	4 (2%)	1 (0.5%)	NS
Requirement of post-operative antibiotics, any cause of.	8 (4.0%)	22 (10.9%)	0.008
Thrombophlebitis	3 (1.5%)	2 (0.5%)	2NS

SSWI: surgical site wound infection..

Types of endovenous treatment (p=0.79 for VenaSeal, p=0.42 for RFA, p=0.52 for EVLA, and p=0.38 for UGFS. All Fisher’s exact test) and the number of phlebectomies (3.61±2.3 in no SSWI group vs. 4.40±5.0 in SSWI group) were not associated with postoperative SSWI incidence. In the univariate analysis, the size of the GSV was larger in SSWI patients than in no SSWI patient, (4.93±2.1 mm vs. 14.4±14.1 mm) but statistical significance could not be confirmed because Leven’s test did not assume equal variance. In the multivariate analysis of the logistic regression test, after including the covariates for sex, age, history of DM, body mass index, numbers of concomitant phlebectomies, GSV size, and prophylactic antibiotic use, the preoperative GSV size was a unique risk factor for postoperative SSWI (odds ratio 1.442, 95% confidence interval 1.12∼1.80, p=0.004).

The incidence of thrombophlebitis was not different in both groups (3, 1.5% vs. 2, 1.0%. p=ns). However, antibiotic requirement due to any cause three months postoperatively was significantly higher in the PA group than in the nPA group (8, 4.0% vs. 22, 10.9%, p=0.008). Subgroup analysis revealed that the reasons for postoperative additional antibiotic requirement were SSWI control (n=5), routine immediate postoperative medication (n=13), risks of aggravation to infective thrombophlebitis (n=5), perivascular swelling with skin erythema (n=4), and poor postoperative wound hygiene (n=3).

The present study focused on the incidence of SSWI in patients who underwent the combined endovenous procedure with concomitant phlebectomy. We found an acceptable and similar incidence of SSWI after the procedure in both nPA and PA groups. The risk for SSWI was similar across all types of endovenous procedures. We also identified a similar rate of postoperative throm-bophlebitis. Although the additional antibiotic requirement was higher in the PA group than in the nPA group, subgroup analysis showed that the reasons were not only infection control but also included empirical administration. Additionally, the finding that the size of the GSV might be a risk factor for post-operative SSWI is notable.

Outpatient clinic-based day-surgery processes for CVD treatment are gaining popularity. This trend is accelerating with the advancement of the medicotechnical environment, and the safety issues are receiving proportionate attention. Fortunately, most adverse events after the endovenous procedures are self-limiting and minor (14). However, the risk of postoperative SSWI has still remained a concern, and recent large-scale surveys revealed SSWI rates of 0.5∼3% (13,15). They demonstrated that prophylactic treatment with systemic antibiotics administered immediately preoperatively before open surgery reduces the risk of wound infection by between two-thirds and three-quarters. In general, endovenous treatment has a lower risk of postoperative SSWI. A pooled analysis revealed that the risk in the EVLA group is 68% lower [Relative risk (RR) 0.32, 95 % CI 0.10∼0.9] than that in the high ligation and stripping group (16). Therefore, guidelines recommend PA for surgical stripping and no use of routine PA for endovenous-only treatment. However, to reduce saphenous nerve damage amid technical success, a combined endovenous procedure with concomitant phlebectomy is being accepted as a good option. The authors believe that evidence for the guideline is needed in this changing environment.

Singh and colleagues reported a SSWI rate of 8.2% after high ligation and endovenous RFA procedure (17). They noted that a single dose of PA reduced the rate of all infections, especially in patients with diabetes and BMI. However, direct comparison of the SSWI incidence to our study has weak validity because high ligation requires wider skin incision and aggressive dissection of the inguinal lesion, which has many lymphatic pathways. Unfortunately, there has been no data to evaluate the efficacy of PA or nPA in a setting like our study.

The present study has several limitations. A retrospective observational study with a small sample size has has limitation in terms of data collection, analysis, and interpretation. However, the propensity score matching method might have reduced the selection bias. Second, we did not investigate the incidence and nature of the side effects of PA in the study population (18). Oral admini-stration of single-dose PA might be less effective. Third, the use of antibiotics after surgery was high in the PA group due to changes in prophylactic antibiotic strategies over time. Finally, the surgical environment and individual hygiene are heterogeneous in every center and patient, respectively. The effect of these differences should be considered and therefore we need an optimized and individualized protocol for the use of antibiotics.

Both nPA and PA strategies have very low and comparable incidence of SSWI after combined endovenous surgery and concomitant phlebectomy for CVD treatment. However, the size of the GSV might be a risk factor for post-operative SSWI. Optimized and individualized protocols for the use of antibiotics, specific for different conditions, are needed.