Background

Cardiovascular disease is the leading cause of deaths world-wide according to the World Health Organization, causing 17.9 million deaths annually [1]. Coronary artery disease (CAD) is the most prominent cause of morbidity and mortality [2]. Due to advances in both surgical techniques and percutaneous coronary intervention, an increasing percentage of patients with CAD are eligible for myocardial revascularization [3]. Coronary artery bypass grafting (CABG) remains the standard of care for patients with three-vessel or left main CAD and improves myocardial ischemia, cardiac function, and cardiac-related mortality rates [2, 4, 5]. CABG requires autologous conduits; one of the most frequently used is the long saphenous vein (LSV) [6, 7].

The reported incidences of wound related complications after LSV harvest ranges from 2 to 20% [8, 9]. Commonly reported leg wound complications related to saphenous vein harvesting include surgical site infection, pain, swelling, and impaired mobility [6]. Such complications are a major cause of patient morbidity, resulting in prolonged hospital stay and delayed recovery with the need for additional interventions such as debridement, antibiotic therapy, and dressing changes [6, 10, 11].

Ultrasonography (US) for preoperative vein mapping has been used to assess and determine the anatomical course of the LSV, thereby facilitating rapid and accurate location of the vein during surgery [12, 13]. Furthermore, preoperative vein mapping has been suggested to be of value in patients whose veins are not evident by physical examination due to edema, multiple varicosities, or complex saphenous vein anatomy [14]. Thus, unnecessary incisions can be avoided reducing complications at incision sites and wound healing disturbances [12].

A previous Best Bet study from 2013 supports the use of preoperative ultrasound assessment of the saphenous vein [15, 16]. However, only one of the included studies was comparing ultrasound mapping with conventional vein harvest [16]. Therefore, the aim of this present systematic review was to determine whether preoperative ultrasound mapping of the LSV reduces leg wound complications compared to conventional vein harvest.

Main text

Search strategy

The data search was performed on May 10th, 2022. This systematic review was conducted according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines [17]. No ethics approval was required for this review.

Publications were identified through a search of online databases: Cochrane Central Register of Controlled Trials, PubMed, and EMBASE. The search was not limited by language or publication status, and no date restriction was implemented. Following search terms were used to search the PubMed and adapted appropriately for EMBASE and Cochrane: (((“Coronary Artery Bypass”[Mesh]) OR (coronary bypass)) OR (cabg)) AND ((((“Ultrasonography”[Mesh]) OR (ultrasound)) OR (ultrasonics)) OR (mapping))) AND (((“Veins”[Mesh]) OR (saphenous)) OR (vein)). The bibliographies of all included studies were hand-searched for any relevant references that may have been missed by the literature searches.

Eligibility criteria

Citations retrieved from all literature sources were collated into an EndNoteTM 20 library. Eligible studies were included if they met the following inclusion criteria: participants were adults (18 years or older); participants were undergoing elective isolated CABG surgery with LSV as conduit; studies evaluating ultrasound mapping of the LSV prior CABG surgery compared to conventional harvesting; study designs included were randomized controlled trials (RCTs) and prospective and retrospective cohort studies, respectively.

Reviews and case reports were excluded.

Primary outcome measure

Occurrence of leg wound complications following harvest of the saphenous vein.

Secondary outcome measures

Additional leg incision, leg incision length, and harvest time.

Assessment of study eligibility

In accordance with the inclusion and exclusion criteria, titles and abstracts were screened on Covidence software (www.covidence.org) independently by two authors (AIE-A, RR). Full-texts articles were assessed for eligibility using a modified Cochrane Data Extraction Template. The review authors were not blinded to the authors’ names and affiliations. A consensus of the two reviewers was needed to include the studies in the literature review (Fig. 1).

Fig. 1
figure 1

Summary of the process used to identify and select studies for the review

Full size image

Data extraction

The following data were extracted from the eligible studies: demographic information, study design, intervention details, and all reported outcomes. Data were only extracted if stated in the text or tables of the article, not in graphic representations.

Quality assessment

The possible risk of bias was assessed using either the Cochrane Risk of Bias Assessment [18] or the ROBINS-I tool [19] for the RCTs [20, 21] and non-RCTs [6, 12, 22], as appropriate (Table 1). Both risk of bias assessment tools include several domains whereby in which regard each study is evaluated independently. The two reviewers completed the risk of bias assessments independently and any disagreement was resolved by discussion.

Table 1 Quality assessment of included studies. The Cochrane risk of bias tool was used to assess the quality of the randomized controlled trials while the ROBINS-I tool was used to assess the non-randomized trials
Full size table

Statistical analysis

Due to the considerable differences regarding levels of evidence between the RCTs and cohort studies, the results from the RCTs were evaluated independently and the results from cohort studies were pooled in a meta-analysis. The meta-analysis was calculated using the Mantel-Haenszel estimator. The primary outcome was tested according to the null hypothesis that there is no difference in the risk of leg wound complications. The fixed or random effect was utilized depending of the heterogeneity of the studies included in the meta-analysis using I2, i.e., if heterogeneity was small (I2 ≤ 50%) the fixed-effect model was used while the random-effects model was assumed in case of high heterogeneity (I2 > 50%). The effect size of the meta-analysis was reported in log risk ratio and presented as exponent values.

The statistical analysis was performed using Stata (StataCorp. 2019. Stata Statistical Software: Release 16. College Station, TX: StataCorp LLC).

Results

Description of included studies

Of 4514 papers screened in this systematic review, 36 papers underwent full-text assessment with final inclusion of five studies [6, 12, 20,21,22] (Fig. 1). Of the five included studies, two were RCTs [20, 21] and three were cohort studies [6, 12, 22], respectively. Studies were mainly excluded as the outcomes of interest were not reported.

Quality of the included studies

An overview of the methodological assessment of the included studies is displayed in Table 1. The Oxford Centre for Evidence Based Medicine tool was used to classify the quality of included studies [23]. Only one of the included studies used concealed allocation [21]. One of the randomized studies [20] did not provide information regarding the randomization protocol. None of the studies used blinding.

Baseline characteristics

Patient baseline characteristics are reported in Table 2. Apart from two studies, there was no statistically significant difference in baseline characteristics between the two groups. In Temizkan et al. [12], the mean age of patients and the rate of hypertension were higher in the unmapped group (p = 0.038). In the randomized study by Zarrepur et al. [21] there was a significantly higher number of diabetic patients in the mapped group (US) (p = 0.024).

Table 2 Demographics and comorbidities of the study population
Full size table

Outcomes

All five studies reported occurrence of leg wound complications following harvest of the saphenous vein as an outcome. Details of the included studies is shown in Table 3. Outcome measures varied notably between the studies.

Table 3 Main results of the included studies
Full size table

Harvest wound complications

Due to significant differences in the study designs and quality, we chose to assess the RCTs individually separating them from the non-RCTs, thus pooling the results from the three non-RCTs in a meta-analysis. In the meta-analysis, we found a statistically significant reduction in the risk of leg wound complications in the ultrasound group with a risk ratio of 0.32 (0.19–0.55) compared to the group without preoperative ultrasound mapping (Fig. 2). As the I2 value was very low, the fixed-effect model was assumed.

Fig. 2
figure 2

Forest plot for leg wound complications. (Yes; number of patients with harvest wound complications, and no; number of patients with no harvest wound complications)

Full size image

In Zarepur et al., each leg wound complication type was reported separately [21]. No statistically significant difference was found between the two groups regarding postoperative infection (p = 1.000), nor incidences of edema, hematoma, or deep vein thrombosis.

Luckraz et al. assessed the presence of postoperative leg wound complications by an ASEPSIS score (additional treatment, serous discharge, erythema, purulent exudates, separation of deep tissues, isolation of bacteria, stay as patient prolonged over 14 days) [20, 24]. The ASEPSIS score ranges from 0 to 70 with higher scores indicating higher probability of infection. There was a statistically non-significant higher mean ASEPSIS score of 25 in the group without ultrasound mapping compared to a mean score of 10 in ultrasound group (p = 0.08).

Leg incision

Only the study by Lopes et al. [22] reported the risk for bilateral leg incision. Six patients in the unmapped group (8.95%) required an additional incision, and this outcome did not occur in the mapped group (p = 0.0067). Furthermore, in the RCT by Luckraz et al. there was a significant reduction in the leg wound incision length in the mapped group (mean 16.8 vs 24.1 cm, p = 0.005) [20].

Harvest time

Aguirre et al. reported that the median harvest time was significantly shorter in the mapped group (25 min [19, 21,22,23,24,25,26,27,28,29,30] vs 40 min [30–50]; p = 0.001) [6]. Similarly, Luckraz et al. found a significantly shorter time for vein harvesting per length of vein graft for the mapped group (36 ± 13 min compared to 47 ± 17 min for the non-mapped group, p = 0.04) [20].

Discussion

This systematic review assessed the existing literature regarding the effects of preoperative ultrasound mapping of the LSV on harvest site infections in patients undergoing CABG. We screened 4514 studies with final inclusion of 5 studies; 2 RCTs and 3 non-RCTs. Furthermore, we assessed reporting regarding additional leg incisions, incision length, and harvesting times.

Our meta-analysis based on three non-RCTs indicated a significant reduction of postoperative leg wound complications following preoperative ultrasound mapping [6, 12, 22]. The two RCTs included were assessed independently, and neither showed effects of preoperative vein mapping on harvest site infections [20, 21]. One of the RCTs, Zarepur et al. reported wound complications separately, indeed a higher number of patients would be needed to find any statistically significant difference between the groups. Only two studies [6, 22] reported the mean BMI as a confounder, while the study by Luckraz et al. [20] reported a mean body surface area, which is an important factor correlated with leg wound healing and complications.

Furthermore, one study showed statistically significant reduction in incision lengths, one study showed reduced risk of additional incisions, and two studies showed significantly shorter harvesting times. However, these outcomes are highly dependent on the experience of the surgeon performing the procedure which was not specified in the studies. Moreover, these secondary outcomes were not reported consistently in all of the included studies.

Despite the above reservations regarding the findings of the present systematic review and meta-analysis, our findings indicate that preoperative vein mapping leads to a better preoperative qualitative assessment of the LSV thus possibly reducing harvest site infection, incision lengths, the need for additional incisions, and harvesting times as compared to blind leg incision.

Other aspects of significance not within the scope of this present study are the learning curve and interobserver variability of utilizing ultrasound mapping as well as the cost-benefit aspects. All of these factor should be considered and investigated into before we can encourage the routine use of preoperative ultrasound mapping of the LSV. None of the included studies mentioned the effects of preoperative ultrasound mapping of the LSV on long term graft patency or perioperative graft strategy. Furthermore, as endoscopic vein harvesting was not within the scope of this study, this technique is known to reduce the risk of leg wound complications [26,27,28,29,30]. However, endoscopic equipment increases the cost of surgery and has different impediments as compared to a conventional harvesting technique. Therefore, open vein harvesting is still a widely used technique and efforts to improve this approach are warranted.

The strength of this study is a large comprehensive literature search performed using multiple databases by two independent reviewers. Thereby, the potential risk of missing relevant studies was very low. The meta-analysis in this study is based on three observational studies only, which is a considerable limitation. Furthermore, the findings from the meta-analysis regarding reduced harvest site infections were not confirmed by the two RCT’s. All included studies were ranked as lower quality studies with poor methodical quality. Furthermore, the lack of blinding and concealment in the included studies might induce a type 1 error, as the investigators might possibly evaluate the wounds of the UL group more favourably. The lack of randomization could lead to both type 1 or type 2 errors, while we find it more plausible the UL sound assessment were used more frequently on indication leading to utilizing UL on patients with more complex anatomy, possibly leading to type 2 errors.

Conclusions

Postoperative wound complications are a major concern following coronary bypass surgery. This systematic review including a metanalysis showed that the existing evidence in this field is of low quality. However, the studies suggest that preoperative ultrasound mapping might be beneficial in terms of reducing wound complications and harvest times while preventing unnecessary leg incisions. Indeed, these findings encourage large-scale randomized prospective studies in this field.