Introduction

The number of total joint arthroplasty (TJA) procedures, including total hip arthroplasty (THA) and total knee arthroplasty (TKA), have consistently risen in recent years, and the need for this procedure is expected to continue to increase [1]. New healthcare policies, with lower reimbursement rates and fewer available resources, have focused on improving surgical efficiency and outcomes [2]. Henc.e, reduction in the time spent during wound closure may be an objective in efforts to reduce overall operative time and cost while increasing surgical productivity. In addition, shorter operative time has also been shown to decrease the rate of infection [3, 4].

Traditionally, wound closure after TJA has been performed with several layers of continuous and interrupted sutures using a variety of both absorbable and non-absorbable sutures and skin with non-absorbable suture, stapes or 2-octyl cyanoacrylate (OCA) [5]. However, wound closure using these techniques may increase the time required [3]. Additionally, biomechanical studies have demonstrated that barbed sutures are stronger than interrupted braided sutures [6]. Despite the multiple potential benefits, Patel et al. reported that barbed sutures were associated with higher rates of complications than other closure methods such as staples (13.0 versus 3.9 %; p = 0.017) [4].

Although multiple studies over the last decade have evaluated different possible alternatives for wound closure [7, 8], no study has assessed all available level I trials (defined as prospective randomized trials) [9] evaluating the effectiveness, risk of complications, and cost-saving capacity of barbed sutures in TKA and THA. Hence, the purpose of our study is to analyze the highest evidence-based (level I) studies in order to compare (1) rates of minor and major complication, (2) differences in operative time, and (3) cost reduction with the use of barbed sutures in TKA and THA.

Methods

Following the Preferred Reporting Items for Systematic Review and Meta-Analyses (PRISMA) guidelines, a query of PubMed, EMBASE and Ovid databases of relevant reports through June 2015 was performed [10] using the search strings "barbed AND suture," "wound AND barbed," "barbed AND total," and "self-locking AND arthroplasty," which yielded 594 studies. We then excluded studies that were (1) performed in animals, (2) non-English, or (3) single case reports. These exclusions left a total of 20 studies, which were carefully assessed for relevance. We then cross-referenced these studies and found two additional studies that were included, for a total of 22. After careful review, we found that 18 did not provide sufficient information, were focused on other types of procedures, were review studies, or had a lower level of evidence. Hence a total of four level 1 studies were included in our final review (see Fig. 1).

Fig. 1
figure 1

Study flow chart of patients in this review

Full size image

In all studies, we specifically focused on determining patient demographic characteristics, the specific type of procedure (TKA or THA), size of the incision, time savings, rate of closure, minor and major complications, and the overall cost savings. The complications were defined as minor (prominent suture, superficial infection, stich abscess, erythema, or other) or major (deep infections, pulmonary embolism, wound dehiscence, dehiscence of extensor mechanism) according to the definition for each individual study. In addition, we recorded any other possible difficulty described by any of the studies, such as needle sticks or suture breakage during closure. All data was inputted into an electronic spreadsheet, and descriptive statistics were then performed for all of the previously compiled information. Utilizing a random effects model, we also calculated the odds ratio for developing a minor or major complication, which was performed with the aid of a statistical software (MedCalc Software bvba, Ostend, Belgium).

This study was performed with no external funding.

Results

A total of 588 patients were randomized to either barbed suture closure (n = 290 TJAs, 268 TKAs, and 22 THAs) or a matched conventional suture cohort (n = 298 TJAs, 279 TKAs, and 19 THAs). In the barbed cohort, 40 % (115) were men, with a mean age of 64 years (range of means, 59.2 to 68.1 years) and a mean body mass index (BMI) of 31.8 kg/m2 (range of means, 30.1 to 33.7 kg/m2). In the randomized comparison cohort, 38 % (112) were men, with a mean age of 65 years (range of means, 63.0 to 68.1 years) and a mean BMI of 31.4 kg/m2 (range of means, 30.1 to 33.0 kg/m2) (Table 1).

Table 1 Demographic characteristics and type of suture
Full size table

The mean length of incisions within the barbed cohort was 16.2 cm. Specifically, it was 16.4 cm for the knee and 16.2 cm for the hip. In the comparison cohort, the mean length of incisions was 15.9 cm (mean of 16.3 cm for TKA and 14.9 cm for THA), which was not significantly different from that of the barbed cohort (p > 0.05). In terms of time savings during wound closure, the barbed suture was 6.3 minutes faster than the conventional closure (12 vs. 18.3 minutes, respectively). This was true for knee arthroplasties, with time savings of 4.3 minutes (10.1 vs. 14.4 min), and for hip arthroplasty, with time savings of 5.4 minutes (9.6 vs. 15.0 minutes) (Table 2).

Table 2 Length of incision, time for closure, and closure rate
Full size table

There were no significant differences in rates of minor or major complication between the two cohorts (Table 3). In terms of minor complications, our random effects model demonstrated that the odds of developing a minor complication were nearly the same in both cohorts (odds ratio [OR] 1.04, 95 % confidence intervals [CI] 0.31–3.54, p = 0.95; Fig. 2). Specifically, six cases of prominent sutures, seven superficial infections, five stich abscesses, and two cases of skin erythema were observed within the barbed suture cohort. Similarly, the conventional closure cohort had two cases of prominent suture, four stich abscesses, three cases of skin erythema, and four other unspecified minor complications. In terms of major complications, there was no significant difference between cohorts (OR 2.94, 95 % CI 0.44–19.74, p = 0.27; Fig. 3). The specific major complications in the barbed cohort were three deep infections and a pulmonary embolism; in the comparison cohort, only one patient developed a deep infection.

Table 3 Complications
Full size table
Fig. 2
figure 2

Forest plot of level I studies comparing complications in patients with the use of directional barbed suture closure compared with traditional sutures. An odds ratio greater than 1 describes greater odds for minor complications with barbed sutures

Full size image
Fig. 3
figure 3

Forest plot of level I studies comparing complications in patients with the use of directional barbed suture closure compared with traditional sutures. An odds ratio greater than 1 describes greater odds for major complications with barbed sutures. Note that the studies by Ting et al. and Sah had no major complications in either cohort

Full size image

After accounting for the time savings, all studies reported cost savings with the use of barbed sutures. After all data were compiled for knee and hip arthroplasty, the total mean savings was US $298 (please see Table 4). Specifically, of the studies that provided specific data for TKAs, the savings ranged from US $58 to $365, and the study providing specific data for THAs reported mean savings of US $615. In terms of surgical difficulties, 12 suture breaks were reported and one surgeon had a needle stick in the barbed closure cohort, while there were three suture breaks and five needle sticks in the conventional cohort.

Table 4 Cost savings
Full size table

Discussion

In light of new healthcare policies, cost and time savings while lowering or maintaining current complication rates is crucial. Hence, we attempted to analyze level I trials that have assessed the use of barbed sutures in the arthroplasty world for the highest possible evidence. We found no significant increase in complication rates, while all studies demonstrated the capacity for high cost and time savings. Similar studies have evaluated the effectiveness of these sutures in other spheres of the surgical field. Among these, a prospective porcine gastrointestinal trial demonstrated a 25 % decrease in surgical time while obtaining similar closure profile, adhesion formation, and histology scoring [14]. Other studies in the obstetric-gynecologic and plastic surgery literature have also evaluated these sutures, finding either no difference or improved outcomes and faster surgical time [15, 16].

A recent meta-analysis by Meena et al. published in the European Journal of Orthopedic Surgical Traumatology evaluated the use of barbed sutures in the world of arthroplasty [17]. The authors included four level III retrospective studies [7, 8, 18, 19] and the level I study by Gililland et al. [20] Their primary outcome measure was superficial infection, and secondary outcome measures were deep infection, wound dehiscence, total operative time, closure time, and arthrofibrosis. As they included only TKA in their analysis, and did not account for cost or safety measures, we were unable to draw a meaningful conclusion. Hence, we believe that our study is necessary to better assess these sutures and to determine whether they can aid the adult reconstruction surgeon.

There were several limitations in our study. As with every review, we are limited to data available in the current literature, therefore, the limitations of all evaluated trials are also our limitations. Additionally, non-level I studies were excluded, which may have excluded many high-quality case series; however, our inclusion criteria were chosen in order to include only studies with the best evidence. There are also different types of barbed sutures available, which may have affected homogeneity between studies. We classified complications based on the authors’ subjective determination, and this can induce some bias. Also, various protocols for grading randomized controlled trials (level I) could be utilized, and there is no widespread agreement on the validity of this approach. Therefore, rejecting or accepting one versus another is controversial. Thus we have included all level I trials assessing the use of barbed sutures in TKA and THA.

Among the multiple ways of closing the same type of wound, all studies stratified their information within two cohorts, a cohort utilizing a barbed suture and a comparison cohort without a barbed suture (Fig. 4). To illustrate these, we chose to describe the study by Smith et al. [11]. They used two cohorts undergoing different closure techniques following TJA. In the study cohort, running Quill (SRS; Angiotech Pharmaceuticals, Inc. Vancouver, Canada) for fascia, fat, subcutaneous and subcuticular layer with #2, #1, #0, 2–0 Monoderm was utilized, which was compared to #1 Ethibond (Ethicon Inc., Somerville, N J, USA) for fascia by running distally and interrupted proximally, running 0-Vicryl and interrupted 2–0 Vicryl for deep and superficial subcutaneous layers, and subcuticular layer with running 3–0 Monocryl (Ethicon Inc., Somerville, NJ, USA). The authors showed that after controlling for patient BMI, length of incision, and number of physicians closing the wound, 9.72 minutes was saved by utilizing barbed sutures.

Fig. 4
figure 4

a, b Illustration of a fascial closure with a barbed and b conventional sutures

Full size image

Similarly, Gilliand et al. prospectively evaluated 411 TKAs randomized to barbed running suture (191 knees, Quill SRS; Angiotech Pharmaceuticals, Inc. Vancouver, Canada) or knotted interrupted suture (203 knees, Ethibond or Monocryl; Ethicon Inc., Somerville, NJ, USA; see Table 1) [20]. They found that closure time was significantly faster, by 4.6 minutes, with no difference in KSS (Knee Society Score) 6 weeks postoperatively, and with similar patient satisfaction and cosmesis.

Our meta-analysis demonstrated that barbed sutures did not significantly increase the odds of developing minor or major complications. However, other studies with lower levels of evidence (levels III and IV), which we did not include in this analysis, have demonstrated conflicting results. Of these, Patel el al. used absorbable barbed suture only for the subcuticular level, and showed a significantly higher minor and major complication rate compared to that with staples (p = 0.033) [4]. Wright et al. described three cases of extensor mechanism failure that occurred with the use of bidirectional barbed sutures, and recommended avoiding this type of closure in morbidly obese, diabetic or rheumatoid patients undergoing TJA [21]. Campbell et al. found higher rates of infection and overall complications, and hence recommended avoiding this type of sutures in superficial closure [19].

Smith et al. combined their original prospective cohort of 34 patients with a retrospective cohort of 100 patients [11]. Although the differences in complication rates were not significant, they nonetheless prompted the discontinuation of the use of barbed sutures for closure, as their complication rate increased from 5.5 to 8.2 % for minor and 0 to 2 % for major complications (p = 0.45 and 0.488, respectively).

Gilland et al. reported five needle sticks of surgical staff in the control group and only one stick in the barbed group, with no statistical difference between groups [20]. There was also no statistical difference in the rate of major or minor postoperative complications, including wound infection, and the authors found no correlation with cellulitis, dermal closure, ASA score, smoking, age, or BMI. Ting et al. showed a lower rate of wound-related complications that was not significant at three weeks or three months (p > 0.05) [12]. The Hollander wound score, patient-derived Likert, and VAS revealed no significant difference in wound outcomes or patient satisfaction. Sah reported a significantly lower number of suture handoffs between scrub technicians and surgeons with barbed sutures, with seven to nine sutures versus 14 to 16 passes with standard sutures, although there were no intra-operative clinical complications using either suture type [13]. The authors found no dehiscence or wound drainage in either group. There were three suture abscesses with traditional sutures and none with barbed sutures (p = 0.24). No antibiotics were used, and no surgical interventions were necessary in either group.

Although no study commented on the intra-operative difficulty of using these sutures, we have found in our experience with Quill sutures that they are difficult to use in tough scar tissue (as the needle is not cutting), and the tip bends easily when repeatedly held with needle driver, with the risk of tip breakage.

Conclusions

There is overall consensus in the literature regarding the time savings that can be extrapolated to total cost savings associated with the use of barbed sutures. However, the potential increase in complication rates described by some authors may dissuade surgeons from taking advantage of this apparent cost benefit. In our evaluation of all level I trials, we found no significant increase in major complication rates, and minor complication rates were nearly equivalent between cohorts. Although the specific closure technique utilized is surgeon-dependent, when faster surgical time is a goal, the adult reconstruction surgeon could attempt to utilize this closure technique. Although larger cohort studies must be performed before a final conclusion can be made regarding whether major complication rates differ, current evidence supports the use of these sutures (see Fig. 5 for a clinical intra-operative photograph in a case utilizing barbed sutures).

Fig. 5
figure 5

Clinical intra-operative photograph of procedure utilizing barbed sutures, demonstrating closure of the deep arthrotomy in a patient undergoing total knee arthroplasty

Full size image