Three U.S. Preventive Services Task Force Level 1 studies (prospective randomized surgical trials) show clear clinical advantages of laparoscopic myomectomy over abdominal myomectomy in women of reproductive age.1–3 Large studies of obstetric outcome after laparoscopic myomectomy report a risk of uterine rupture between 0.0% and 0.25%4,5; these numbers compare favorably with the published uterine rupture rate after a single term cesarean delivery.6 Despite this evidence, uterine myomectomy is still performed through an abdominal incision by most gynecologic surgeons.7 Robot-assisted laparoscopic myomectomy, pioneered by Arnold Advincula after the introduction of the da Vinci robotic surgical platform (da Vinci Surgical System),8 is poised to become an ethical alternative to abdominal myomectomy for surgeons who do not master advanced laparoscopic techniques. Advincula et al9 were also the first to compare short-term surgical outcomes and cost of robot-assisted laparoscopic myomectomy with those of open myomectomy. This retrospective cohort study compared robot-assisted laparoscopic myomectomy patients with open myomectomy patients of similar age, body mass index (BMI, calculated as weight (kg)/[height (m)]2), and myoma weight, and suggested that robot-assisted laparoscopic myomectomy has decreased operative blood loss, lower complication rates, and shorter length of hospital stay compared with open surgery.
Three studies have compared robot-assisted laparoscopic myomectomy and laparoscopic myomectomy. A small study by Nezhat et al10 matched 15 robot-assisted laparoscopic myomectomy patients with 35 patients treated with laparoscopic myomectomy. Robot-assisted laparoscopic myomectomy was associated with a longer mean surgical time compared with laparoscopic myomectomy (234 minutes compared with 203 minutes). There were no appreciable differences observed in blood loss, hospitalization time, or postoperative complications.10 A retrospective cohort study by Bedient et al11 compared 40 robot-assisted laparoscopic myomectomy and 41 laparoscopic myomectomy surgeries and, after adjusting for uterine size and leiomyoma size and number, did not find significant differences between the two groups for operating time (141 compared with 166 minutes), blood loss (100 mL compared with 250 mL), complications, hospital stay more than 2 days, readmissions, or symptom resolution.
Recently, a larger retrospective cohort study compared 393 open myomectomies (performed over the course of 14 years) with 93 laparoscopic myomectomies (performed over the course of 6 years) and 89 robot-assisted laparoscopic myomectomies (performed over the course of 2 years).12 The three patient groups (abdominal myomectomy, robot-assisted laparoscopic myomectomy, and laparoscopic myomectomy) were not homogeneous; the median BMI was significantly higher in the open group compared with both laparoscopic groups, and larger, heavier, and more numerous myomata were removed in the open and the robot-assisted laparoscopic myomectomy group compared with the laparoscopic myomectomy group. The authors comment openly regarding the selection bias in the patient population with respect to surgical type. Specifically, robot-assisted laparoscopic myomectomy was chosen as the modality of choice in those cases thought to be too challenging for laparoscopic myomectomy. Despite this, operative time was shortest for the open myomectomy group, but no difference in time was observed between robot-assisted laparoscopic myomectomy and laparoscopic myomectomy. Similarly, open myomectomy had higher blood loss and longer hospital stay compared with that in the laparoscopic groups, but there was no difference for such parameters between robot-assisted laparoscopic myomectomy and laparoscopic myomectomy.
None of the three studies offered insight into the level of expertise of the surgical teams performing robot-assisted laparoscopic myomectomy compared with laparoscopic myomectomy. Such information is essential in any study comparing the safety and efficacy of a novel technique with those of an established one. The objective of our study was to compare the surgical outcome of laparoscopic myomectomy and robot-assisted laparoscopic myomectomy performed by two high-volume surgical teams, and also to consider the effect of the robotic learning curve.
MATERIALS AND METHODS
This study was approved by the institutional review board of Brigham and Women's Hospital. This retrospective cohort study included all 289 women undergoing a laparoscopic myomectomy for symptomatic leiomyomas from February 2007 through September 2009 at the Division of Minimally Invasive Gynecologic Surgery and the Division of Reproductive Endocrinology and Infertility. Techniques were standardized; all laparoscopic myomectomies were performed by the Division of Minimally Invasive Gynecologic Surgery team and all robot-assisted laparoscopic myomectomies were performed by the Division of Reproductive Endocrinology and Infertility team. The Division of Minimally Invasive Gynecologic Surgery team included a Division of Minimally Invasive Gynecologic Surgery–trained specialist (J.I.E.) with 5 years of postresidency experience in high-volume academic surgical practices. The Division of Reproductive Endocrinology and Infertility team consisted of two Division of Reproductive Endocrinology and Infertility specialists with 11 (A.R.G.) and 5 (S.S.S.) years of postresidency experience in a high-volume academic surgical practice. The Division of Reproductive Endocrinology and Infertility specialists operated as cosurgeons in almost all robot-assisted laparoscopic myomectomy cases included in this study.
Patient recruitment into the Division of Minimally Invasive Gynecologic Surgery and Division of Reproductive Endocrinology and Infertility teams was independent. Both teams had established referral patterns and functioned as separate regional referral centers at the time of this study.
Brigham and Women's Hospital is a teaching hospital of Harvard Medical School; therefore, Division of Minimally Invasive Gynecologic Surgery and Division of Reproductive Endocrinology and Infertility fellows or gynecology residents (or both) were involved in most of these procedures at different levels as deemed necessary. Data regarding the extent of trainee involvement were not collected and are not part of our study. The 289 consecutive surgeries were stratified into two groups according to the surgical technique used. Group 1 included 115 women who underwent laparoscopic myomectomy, and group 2 included the first 174 women who underwent robot-assisted laparoscopic myomectomy.
Demographic data obtained were age, race (Causasian, African American, Hispanic, Asian, other), gravidity, parity, cigarette smoking status (current compared with not), surgical history, and BMI as calculated from height and weight measurements. Tumor characteristics collected for each patient included: number of myomata removed; maximum diameter of the largest myoma removed; and total morcellated myoma weight. Surgical outcome measures included: actual operative time in minutes as per the operative record (incision time to complete wound closing time); estimated blood loss in milliliters as per best estimate of the surgical, anesthesia, and nursing teams; perioperative complications (reoperation, blood transfusion, postoperative antibiotics, deep venous thrombosis, pulmonary embolus, and ileus); and length of hospital stay.
The surgical technique for all laparoscopic myomectomy was as follows. A 10-mm umbilical camera port and two parallel 5-mm operative ports on the left side were generally used. A third operative port was used on the right side if required. The uterus was infiltrated with dilute vasopressin, 20 units in 60 mL of saline solution, taking care to use no more than 10 units each time.13 We generally preferred to create a horizontal incision into the uterus using the Harmonic scalpel. The myoma was then dissected out of the uterus with generous traction with a tenaculum and countertraction with an atraumatic grasper as well as the Harmonic scalpel, as needed. In case of inadvertent entry into the uterine cavity, the endometrial defect was closed with a running 2-0 polyglactin 910 suture, taking care to avert suture entry into the uterine cavity. Before implementation of barbed suture, we used 2-0 polydioxanone suture on a CT-1 needle in continuous fashion using intracorporeal knot-tying to secure each end of the suture. The hysterotomy was closed in as many layers as necessary to eliminate dead space in the myometrial defect. After converting to barbed suture in March 2008, the hysterotomy was closed in layers using a 14×14-cm 0 polydioxanone suture on a 36-mm half-circle tapered needle. If the hysterotomy was longer than 8 cm, then the 24×24-cm suture is preferred. Tacking the first needle to the opposite anterior abdominal wall helped to prevent suture tangling. The deeper layers were closed using the first needle, and the second needle was used to close the more superficial layers and the serosa if possible. The needles were cut, and a Lapra-Ty clip was applied if the suture was used beyond the barbed portion of the suture, although additional anchoring of barbed sutures is not generally needed or recommended. Sometimes three or four layers were needed to close a deep myometrial defect; 2-0 glycolide and e-caprolactone (Monoderm) also were used in some cases for the serosa, either continuously or using a baseball stitch. Barbed sutures were used in 67.9% of the cases in the laparoscopic myomectomy group. The hysterotomy site was generally covered with an adhesion barrier (Gynecare Interceed).
The surgical technique for robot-assisted laparoscopic myomectomy performed has been detailed previously.14 The primary 12-mm trocar (camera port) was placed in the midline, 10 cm above the top of the uterus as palpated at the time of the pelvic examination under anesthesia. Another 12-mm trocar was placed in the right lower quadrant to be used by the assistant during the case. Robotic arms one and two were set 8–10 cm to the right and to the left of the primary port, respectively. When needed, a third robotic arm was set 8–10 cm to the left of the port for robotic arm two. A dilute concentration of vasopressin (20 international units in 20 mL of sterile saline for the robot-assisted laparoscopic myomectomy group and 20 international units in 40 mL of sterile saline for the laparoscopic myomectomy group) was injected into the myometrium overlying the largest myoma. Using dedicated robotic Harmonic shears, a transverse myometrial incision was made over the largest myoma. The operator at the da Vinci console enucleated the myoma using a robotic tenaculum or robotic bipolar forceps (or both) in addition to the robotic Harmonic shears. The bedside assistant provided traction on the myoma using a laparoscopic tenaculum. After removal of each myoma, myometrial closure was performed in an identical fashion as in an open myomectomy. Namely, incisions were closed in one to five layers (depending on size and depth). The endometrium (where needed) was reapproximated with running 3-0 polyglecaprone 25; the deep myometrial layer was repaired with interrupted figure-of-eight sutures of 0 polyglactin and the outer myometrial layer was repaired with running sutures of 0 polyglactin. Finally, the uterine serosa was closed with 2-0 or 3-0 polyglecaprone 25 in a running in-to-out baseball stitch burying the entire length of the suture. Barbed sutures were used in 5.0% of the cases in the robot-assisted laparoscopic myomectomy group, with materials and techniques similar to those described for laparoscopic myomectomy. An adhesion barrier was applied.
A hybrid robot-assisted laparoscopic myomectomy procedure was developed by our team as a way to overcome current limitations of the da Vinci surgical system related to multiquadrant surgery in the case of very large myomata. In this hybrid robot-assisted laparoscopic myomectomy procedure, the enucleation of the largest myoma was performed with the same technique described for laparoscopic myomectomy, followed by swift docking of the da Vinci patient-side cart and uterine reconstruction using the robot. During the time of this study, we exclusively used central docking of the patient-side robotic cart.
All statistical analyses were conducted using SAS 9.2. Age and race were included in the models a priori. Other covariates, including BMI (continuous), obesity (dichotomous), gravidity, parity, number of leiomyomas removed, weight of leiomyomas removed, largest uterine dimension, largest leiomyoma, previous myomectomy, and other procedures (ie, hysteroscopic procedures or other laparoscopic procedures such as ovarian cystectomies), were tested individually as potential confounders and included in the final model if the effect size changed by more than 10% as compared with the base model for any given outcome.15 When both BMI and obesity changed the effect estimate of a given outcome by more than 10%, obesity was included in the final model, because the relation between outcomes and body size is likely best-characterized by dichotomizing into obese or not rather than assuming a linear association with the outcome across the entire distribution of BMI. Specifically, the model for postoperative hospital admission (yes or no) was adjusted for age, race, obesity, weight of leiomyomas removed, and largest leiomyoma. The model for hospital stay more than 1 day (yes or no) was adjusted for age, race, weight of leiomyomas removed, largest leiomyoma, and number of leiomyomas removed. The model for any postoperative complications (yes or no) was adjusted for age, race, and largest leiomyoma removed. The model for operative time (continuous) was adjusted for age and race. Finally, the model for estimated blood loss (continuous) was adjusted for age, race, obesity, previous myomectomy, gravidity, largest uterine dimension, weight of leiomyomas, largest leiomyoma, and number of leiomyomas. For all outcomes except hospital stay more than 1 day, indicator variables for missing covariates were included in the models to include the maximum portion of the study population in any regression analysis. The adjusted model for hospital stay more than 1 day did not converge when indicator variables for missing covariates were included. Therefore, only patients with nonmissing values for these factors were included in that analysis.
Odds ratios (ORs) and 95% confidence intervals (CIs) were calculated from logistic regression models. Continuous outcomes were analyzed using linear regression estimating marginal means (least squares means). Because of distributions skewed to the right, the logged outcomes of operative time and estimated blood loss were analyzed. Therefore, adjusted means are the geometric means. Wald two-sided P<.05 was considered statistically significant.
The patient populations were similar in the laparoscopic myomectomy and robot-assisted laparoscopic myomectomy groups in terms of age, BMI, smoking status, and history of a myomectomy; however, racial distributions were different in each group (Table 1). Furthermore, tumor burden was comparable in each group, as defined by the number of myomata removed (2 [1–21] compared with 3 [1–16], median [range]), weight of morcellated myomata (201 [1–1,473] compared with 159 [8–780g], median [range]), and dimension of the largest myoma (7.5 [2.2–16.5] compared with 7.3 [3.1–13.8 cm], median [range]) in the laparoscopic myomectomy and robot-assisted laparoscopic myomectomy groups, respectively (Fig. 1).
Operative outcomes are illustrated in Table 2 and Figure 2. The robot-assisted laparoscopic myomectomy group had significantly longer operative times, with an adjusted mean difference of 77 minutes (195.1 compared with 118.3; P<.001). Furthermore, robot-assisted laparoscopic myomectomy patients had higher odds of being admitted to the hospital (OR 2.07, 95% CI 1.14–3.74) and of having a hospital stay more than 1 day (OR 5.73, 95% CI 1.58–20.81) as compared with laparoscopic myomectomy patients. The adjusted geometric mean for estimated blood loss among robot-assisted laparoscopic myomectomy patients was significantly higher than laparoscopic myomectomy patients (110.0 compared with 85.9 mL, P=.04).
The odds of postoperative complications were lower in the robot-assisted laparoscopic myomectomy group but not statistically different from the laparoscopic myomectomy group (OR 0.63, 95% CI 0.27–1.47). Only one reoperation occurred in each laparoscopic myomectomy and robot-assisted laparoscopic myomectomy group for laparoscopic management of small bowel obstruction and hemoperitoneum, respectively. No deep venous thrombosis or pulmonary emboli occurred in either group. Standard mechanical deep venous thrombosis prophylaxis (automatic intermittent pneumatic calf compression devices) was used in all patients. No pharmacologic deep venous thrombosis prophylaxis was used. The most common complications were administering of postoperative antibiotics for infections such as incisional cellulitis in both laparoscopic myomectomy (10, 9.1%) and robot-assisted laparoscopic myomectomy (6, 3.4%), and administering of blood transfusions in laparoscopic myomectomy (1, 0.9%) and robot-assisted laparoscopic myomectomy (10, 5.7%). Throughout the study period there were no conversions to an open procedure and no hysterectomies performed in either group.
When the data were reanalyzed including only robot-assisted laparoscopic myomectomy cases from 51 to 174 (accounting for a 50-case learning curve for robotic gynecologic surgery derived from published literature),16 these findings remained unchanged. We observed negligible change in these associations when the study population was stratified by race and by obesity. However, because of small sample sizes, formal tests for heterogeneity could not be performed.
Our study reports the experience of two high-volume surgical teams performing laparoscopic myomectomy and robot-assisted laparoscopic myomectomy and found that robot-assisted laparoscopic myomectomy was associated with longer operative times, increased estimated blood loss, and increased odds of hospital admission; however, perioperative complications were similar. One of the advantages of a high-volume surgical center is the ability to evaluate rapidly the advancement of a stable team on operative outcome and account for a learning curve.
Although robotic platforms can be described as highly enabling laparoscopic tools, they are complex machines that require the acquisition of a new skill set.8 Therefore, when considering surgical outcomes the inclusion of data pertaining to the learning curve of the teams for all surgical techniques being compared is paramount. Lenihan et al16 found that the total operative times for benign robot-assisted hysterectomies sequentially stabilized at approximately 95 minutes after 50 cases utilizing the da Vinci Surgical System. We used this conservative estimate of a learning curve for our subanalysis. When we compared the outcomes of laparoscopic myomectomy with those or robot-assisted laparoscopic myomectomy cases beyond the first 50 cases, our findings remained unchanged.
The difference in hospital admission rates noted in this study were reflective of the different practice patterns in each surgical team. When we introduced robot-assisted laparoscopic myomectomy at our hospital, we routinely admitted patients because we did not have any short-term or long-term perioperative complication outcome data. As was shown in this study, the complication rates are comparable with laparoscopic myomectomy group and, overall, very low. With these data, we now discharge patients on the same day of robot-assisted laparoscopic myomectomy or observe them for less than 24 hours when indicated.
However, the shorter operative time of laparoscopic myomectomy over robot-assisted laparoscopic myomectomy is significant in our view. At a superficial analysis, our results appear to differ from those of Barakat et al.13 These authors report no significant difference in actual surgical times between the laparoscopic myomectomy and the robot-assisted laparoscopic myomectomy groups: 155 minutes (98–200) compared with 181 minutes (151–265) (median [interquartile range]). However, one should consider that the operative times for robot-assisted laparoscopic myomectomy reported in their study are very similar to those reported by us, whereas their operative times for laparoscopic myomectomy appear to be quite longer than those reported by us. The tumor load in the robot-assisted laparoscopic myomectomy group of Barakat et al is significantly higher than tumor load in their own laparoscopic myomectomy group, but very similar to that reported in our study for both the robot-assisted laparoscopic myomectomy and the laparoscopic myomectomy groups. Therefore, the discordant findings between our study and the study by Barakat et al can be explained by their relatively long operative times for laparoscopic myomectomy. Barakat et al describe 93 laparoscopic myomectomies performed by multiple teams over the course of 6 years (an average of 15.5 cases per department, per year), whereas our study examines a single Division of Minimally Invasive Gynecologic Surgery team performing 114 cases over the course of 2.7 years (an average of 42.8 cases per surgeon, per year). Hence, the work by Barakat et al is comparing a high-volume robotic team (with an average of 44.5 cases per years) with a low-volume conventional laparoscopy team.
An even more relevant consideration to explain the faster operative time for laparoscopic myomectomy in our study is that of the type of suture used. Recent reports have highlighted the role of barbed sutures in significantly decreasing operative times and perioperative blood loss in laparoscopic myomectomy.17,18 In both laparoscopic myomectomy and robot-assisted laparoscopic myomectomy groups in the study by Barakat et al surgeons have used conventional suture, whereas in our study barbed suture was used in 67.9% of the patients in the laparoscopic myomectomy group but in only 5.0% of the patients in the robot-assisted laparoscopic myomectomy group. Barbed sutures allow running closure in layers without the need for tying knots; this translates to a faster uterine closure time, which is reflected in decreased blood loss and decreased operative time, which are precisely the outcomes in which differences were noted in our study.
The increased blood loss we observed in the robot-assisted laparoscopic myomectomy group (110.0 compared with 85.9 mL, P=.04) is marginally statistically significant and cannot be viewed as clinically significant given a mere difference of 25.9 mL on an estimated (not measured) parameter. There was a higher blood transfusion rate in the robot-assisted laparoscopic myomectomy group (5.7% compared with 0.9%); however, because of the small number of transfusions overall, it could not be determined if this was a significant difference. Three of the patients who received blood transfusion in the robot-assisted laparoscopic myomectomy group had significant chronic or preoperative anemia for which transfusion of blood products was planned. This brings the rate of unexpected postoperative transfusion to 4%. Furthermore, it is difficult to fully account for differences in practice patterns when comparing these numbers. Difference in practice patterns also might explain the increased rate of postoperative antibiotic administration seen in the laparoscopic myomectomy group.
It should be noted that data on the significant effect of barbed suture on these important perioperative outcome parameters were not available at the time of designing our retrospective cohort study. The robot-assisted laparoscopic myomectomy group utilized conventional sutures because of availability of obstetric data with such suture material. However, the results of this study have inspired a rethinking of our approach to barbed suture in some of our robotic cases, specifically in cases with larger myoma or in cases involving patient who have completed childbearing. Future studies comparing perioperative and reproductive outcomes of robot-assisted laparoscopic myomectomy and laparoscopic myomectomy performed with barbed suture are needed.
A major limitation of this and all the available studies comparing robot-assisted laparoscopic myomectomy and laparoscopic myomectomy to date is that they are retrospective and therefore have selection bias. Logistic limitations as well as conceptual limitations made it impossible for our group to consider a prospective randomized trial. The intent of our study was that of comparing two surgical techniques performed by teams with extensive experience. The alternative would have been to have both practices perform robot-assisted laparoscopic myomectomy and laparoscopic myomectomy, which would negate the main goal of the study.
In conclusion, we have analyzed data from two surgical practices with a specific high volume of robot-assisted laparoscopic myomectomies and laparoscopic myomectomoes, and have even taken the learning curve of robot-assisted laparoscopic myomectomy into consideration. Our data suggest that when these two techniques are performed by experienced teams, short-term clinical outcomes and complication rates are equivalent, but operative times are longer and estimated blood loss are larger for robot-assisted laparoscopic myomectomy. A major technical caveat in the interpretation and applicability of our results to other practices is that almost all robot-assisted laparoscopic myomectomies were performed with standard suture material, whereas most laparoscopic myomectomies were performed with the novel barbed suture material. Finally, our findings do not support the existence of a clinically significant learning curve for robot-assisted laparoscopic myomectomy within an advanced laparoscopic practice. We believe that our findings bear great clinical significance at a time when robotic surgery is set to penetrate our practices in ways and at a rate that are difficult to predict.
1. Mais V, Ajossa S, Guerriero S, Mascia M, Solla E, Melis GR. Laparoscopic versus abdominal myomectomy: a prospective, randomized trial to evaluate benefits in early outcome. Am J Obstet Gynecol 1996;174:654–8.
2. Seracchioli R, Rossi S, Govoni F, Rossi E, Venturoli S, Bulletti C, et al.. Fertility and obstetric outcome after laparoscopic myomectomy of large myomata: a randomized comparison with abdominal myomectomy. Hum Reprod 2000;15:2663–8.
3. Palomba S, Zupi E, Falbo A, Russo T, Marconi D, Tolino A, et al.. A multicenter randomized, controlled study comparing laparoscopic versus minilaparotomic myomectomy: reproductive outcomes. Fertil Steril 2007;88:933–41.
4. Sizzi O, Rossetti A, Malzoni M, Minelli L, La Grotta F, Soranna L, et al.. Italian multicenter study on complications of laparoscopic myomectomy. J Min Invas Gynecol 2007;14:453–62.
5. Seracchioli R, Manuzzi L, Vianello F, Gualerzi B, Savelli L, Paradisi R, et al.. Obstetric and delivery outcome of pregnancies achieved after laparoscopic myomectomy. Fertil Steril 2006;86:159–65.
6. Spong CY, Landon MB, Gilbert S, Rouse DJ, Leveno KJ, Varner MW, et al.. Risk of uterine rupture and adverse perinatal outcome at term after cesarean delivery. Obstet Gynecol 2007;110:801–7.
7. Liu G, Zolis L, Kung R, Melchior M, Singh S, Cook EF. The laparoscopic myomectomy: a survey of Canadian gynaecologists. J Obstet Gynaecol Can 2010;32:139–48.
8. Advincula AP, Song A, Burke W, Reynolds RK. Preliminary experience with robot-assisted laparoscopic myomectomy. J Am Assoc Gynecol Laparosc 2004;11:511–8.
9. Advincula AP, Xu X, Goudeau S 4th, Ransom SB. Robot-assisted laparoscopic myomectomy versus abdominal myomectomy: a comparison of short-term surgical outcomes and immediate costs. J Minim Invasive Gynecol 2007;14:698–705.
10. Nezhat C, Lavie O, Hsu S, Watson J, Barnett O, Lemyre M. Robotic-assisted laparoscopic myomectomy compared with standard laparoscopic myomectomy–a retrospective matched control study. Fertil Steril 2009;91:556–9.
11. Bedient CE, Magrina JF, Noble BN, Kho RM. Comparison of robotic and laparoscopic myomectomy. Am J Obstet Gynecol 2009;201:566.e1–5.
12. Barakat EE, Bedaiwy MA, Zimberg S, Nutter B, Nosseir M, Falcone T, et al.. Robotic-assisted, laparoscopic, and abdominal myomectomy: a comparison of surgical outcomes. Obstet Gynecol 2011;117(2 Pt 1):256–65.
13. Frishman G. Vasopressin: if some is good, is more better? Obstet Gynecol 2009;113:476–7.
14. Quaas AM, Einarsson JI, Srouji SS, Gargiulo AR. Robotic myomectomy: A review of indications and techniques. Rev Obstet Gynecol 2010;3:185–91.
15. Greenland S. Modeling and variable selection in epidemiologic analysis. Am J Public Health 1989;79:340–9.
16. Lenihan JP Jr, Kovanda C, Seshadri-Kreaden U. What is the learning curve for robotic assisted gynecologic surgery? J Minim Invasive Gynecol 2008;15:589–94.
17. Alessandri F, Remorgida V, Venturini PL, Ferrero L. Unidirectional barbed suture vs. continuous suture with intracorporeal knots in laparoscopic myomectomy: a randomized study. J Min Invasive Gynecol 2010;17:725–9.
18. Einarsson JI, Chavan NR, Suzuki Y, Jonsdottir G, Vellinga TT, Greenberg JA. Use of bidirectional barbed suture in laparoscopic myomectomy: Evaluation of perioperative outcomes, safety, and efficacy. J Minim Invasive Gynecol 2011;18:92–5.