This prospective study was conducted based on preoperative and perioperative characteristics for 3 operative types. From June 2008 to December 2009, 34 patients underwent laparoscopic hysterectomy (Group A), 25 patients underwent robotic hysterectomy (Group B), and 11 patients underwent vaginal hysterectomy (Group C) at our institution. All of these hysterectomies were performed for noncancerous indications, including adnexal mass, endometriosis, abnormal uterine bleeding, hyperplasia, dysplasia, leiomyomas, chronic pelvic pain, adenomyosis, and pelvic prolapse, among others.
To eliminate surgeon bias in our study, all surgeries were either performed by or supervised by the same surgeon. All laparoscopic and vaginal hysterectomy cases were performed by residents under the direct supervision of the first author, with more than 50% of cases performed by residents as first surgeons. In contrast, 40% of robotic cases were teaching cases. In robotic cases, resident participation was primarily in uterine manipulator placement, trocar placement, and vaginal cuff closure.
For our study, hysterectomy with or without bilateral salpingo-oophorectomy was considered to be “type 1.” Hysterectomies performed at the same time as additional surgeries were labeled as “type 2 or more.” According to this classification, 91% of laparoscopic cases were type 1 versus 84% of robotic hysterectomy cases. Of vaginal hysterectomy cases, 72.73% required 2 or more additional surgeries.
Each case was evaluated for its complexity based on preoperative diagnosis, prior pelvic or abdominal surgery, patient's body mass index (BMI), and uterine weight. Prior pelvic or abdominal surgery was categorized as none, 1 previous surgery, or 2 or more previous surgeries. Patient BMI was categorized as BMI less than 30 or BMI of 30 or greater. Uterine weight was categorized as either <250
g or ≥250
g. All cases were further categorized as being teaching or nonteaching cases.
Before the initiation of this prospective study, institutional review board approval was obtained from The University of Texas Medical Branch at Galveston for data collection on patients who consented for surgery. All the cases were performed by a single surgeon with 7 years of experience in vaginal and laparoscopic hysterectomy, while the robotic cases were the first 25 cases in his medical career. Residents were involved in all laparoscopic and vaginal hysterectomy cases, performing more than 50% of any given case with supervision. Robotic cases were considered to be teaching cases, and resident involvement included only vaginal cuff closing.
Patients with a uterine size not greater than 14 weeks requiring pelvic prolapse surgeries or Grade 2 Baden-Walker uterine descensus in addition to their hysterectomies were selected for the vaginal hysterectomy group. After the decision of minimally invasive hysterectomy was made, the choice between laparoscopy and robotic approach was left solely to the patient's discretion. All cases were performed under general endotracheal anesthesia. Antibiotics were given just before surgery. Patients were placed in the dorsal lithotomy position with Allen stirrups (Allen Medical Systems, Acton MA) for lower extremity positioning.
2.1. Robotic Hysterectomy
A Rumi System uterine manipulator with balloon tip (Cooper Surgical) was placed after appropriate preparation and draping. Patients were placed in a steep Trendelenburg position, with arms tucked at the sides and shoulder blocks placed to limit shift on the operating room table. Placement of an oral gastric tube at the beginning of the case was assured to prevent trauma to the stomach. All cases with previous abdominopelvic surgeries were started using a Veress needle to the left upper quadrant for insufflation of CO2 and followed by a 12-mm bladeless trocar introduced from the same location under direct visualization. In cases of no previous abdominopelvic surgeries, the Veress needle was placed in the umbilicus to obtain the necessary insufflation of CO2, and this was followed by introducing a 12-mm bladeless trocar for the camera at the umbilicus. Two additional 8-mm robotic trocars were placed on the patients' abdomens where additional retraction was necessary. The left upper quadrant 12-mm trocar was used on this group as well to allow the bedside surgical assistant to use a grasper or suction irrigation device. The robotic system was then docked at a 45° angle from the patient's left foot side. All cases were performed using monopolar EndoWrist scissors (Intuitive Surgical, Inc.) combined with a PK grasper. Hysterectomies were classified as American Association of Gynecologic Laparoscopists type IVE, defined as total laparoscopic removal of the uterus and cervix including vaginal cuff closure, and performed using the KOH Colpotomizer System (Cooper Surgical).
All pathologic specimens were removed using 1 of the 3 following methods: direct removal through the vaginal cuff opening, morcellation of the specimen using an endoscopic morcellator, or sectioning of the uterine specimen with the robot using an EndoWrist monopolar cautery instrument (Institutive Surgical, Inc.) to create portions small enough to be delivered vaginally. The vaginal cuff was closed robotically using 0 polyglactin (Vicryl, Ethicon Endo-Surgery, Inc., Cincinnati, OH) on a CT-1 needle in a single separate suture closure.
Sutures were instrument-tied using the robotic instruments. Patients were desufflated, and pedicles were checked at half desufflation for homeostasis. The robotic system was then undocked, and all trocars were removed under direct visualization. The 12-mm trocar sites received a single deep 0-polyglactin suture, and all skin incisions were closed with 4-0 polyglactin subcuticular sutures. Adhesive skin closures (Steri-Strips, 3M, St. Paul, MN) were placed as dressing.
The following times were recorded: docking time, defined as the time from first incision to placement of the robotic instruments into the patient, morcellation time, and total operative time, defined as time from when the patient was brought into the OR until she left the room. In addition, uterine weight, blood loss, conversions, and intraoperative and postoperative complications requiring intervention, as well as length of hospital stay, were monitored and recorded. Cases were stratified based on the level of complexity using BMI, uterine weight, presence of prior pelvic or abdominal surgeries, and preoperative diagnosis. We also reported outcomes for subgroups of patients with uterine weight of ≥250
g or BMI ≥ 30.
2.2. Laparoscopic Hysterectomy
Laparoscopic hysterectomy cases were done with the same trocar placement technique and locations applied in robotic cases, but, instead of an 8-mm robotic port, 25-mm bladeless trocars were used. Other than docking time, the same parameters were recorded as for robotic cases.
2.3. Vaginal Hysterectomy
The cervix was grasped with a tenaculum, and the cervicovaginal junction was circumferentially injected with a vasoconstrictive agent. A circumferential incision was made at the cervicovaginal junction, and the vagina was dissected off the cervix for several centimeters. Careful entry into the anterior and posterior cul-de-sacs was done using sharp dissection. Retraction of the bladder anteriorly off the cervix, visualization of the lower edge of the peritoneum, and careful sharp entry into the anterior cul-de-sac minimized the chance of cystotomy. After the peritoneum was entered anteriorly and posteriorly, the uterosacral and cardinal ligament complex was detached from the uterus bilaterally. Downward traction on the cervix brought the uterus and its remaining attachments closer to the operator, where they too could be seen, clamped, and transected under direct visualization.
A McCall's culdoplasty or simple peritoneal closure and intraperitoneal uterosacral ligament plication were performed after a vaginal hysterectomy to prevent vaginal vault prolapse. The vaginal cuff was closed by using 0-polyglactin CT-1 needle in a running suture closure.
2.4. Data Analysis
Patients' preoperative and operative characteristics were summarized using means, standard deviations and 95% confidence interval for the continuous variables, and proportions for discrete variables among the 3 surgery groups (laparoscopic, robotic, and vaginal). The 1-way analysis of variance (ANOVA) was used to compare the means of continuous variables among the surgery groups. The chi-square test was used to investigate the association between category variable and the surgery groups, and the normality of continuous variables was tested. The tests were assessed at the 0.05 level of significance. Statistical computations were carried out using statistical software, SAS release 9.2.