Materials and Methods

A retrospective analysis of 181 consecutive patients who underwent surgery for endometrial cancer in 2005 and 2006 was conducted following approval from our institutional review board. Patients were identified from institutional tumor registry data in addition to anesthesia and operating room records. Data were abstracted using electronic and manual chart review. Women of all ages with biopsy-confirmed or clinical suspicion of endometrial cancer who underwent staging procedures consisting of either open (abdominal) or minimally invasive (laparoscopic-assisted vaginal, total laparoscopic, or robotic-assisted laparoscopic using the da Vinci Surgical System [DVSS; Intuitive Surgical Inc, Mountain View, Sunnyvale, CA]) hysterectomy were included. Pelvic and aortic lymphadenectomy were at the discretion of the surgeon. Patients who received an epidural in addition to general anesthesia were excluded. Cases that were initiated as a minimally invasive procedure and subsequently converted to open were included in the minimally invasive group. Each patient received an American Society of Anesthesiologists (ASA) score, with a score of 1 representing a healthy adult vs a score of 5, a person not expected to survive >24 hours. The perioperative period was defined as the time from admission to the preoperative holding area until 24 hours after tracheal extubation. Tabulation of anesthesia time, intraoperative and postoperative analgesic and antiemetic use, intraoperative use of vasopressors and antihypertensives, intraoperative volume resuscitation and blood loss, and length of stay in the recovery room and hospital were performed. Anesthesia time was defined as time from induction of anesthesia until tracheal extubation. The recovery room stay was calculated as the time from admission until the patient achieved discharge criteria as per our institution protocol. Data on analgesic and antiemetic use were collected in 3 time frames: in the operating room, recovery room, and first 24 hours postoperatively. We restricted the analysis to 24 hours because most patients in the minimally invasive group were discharged within 24 hours. Pain and nausea scores were recorded using a verbal numeric scale, with 0 representing no pain or nausea, and 10 representing worst possible pain or nausea. Patients received education preoperatively on how to rate their pain and nausea based on these scores. Patients received analgesics in the recovery room as prescribed by the anesthesiologist using standardized order forms. Type and frequency of analgesics in the first 24 hours postoperatively were given at the discretion of the surgeon using opioids and nonsteroidal antiinflammatory drugs (NSAIDs). Nurses administered opioids to patients undergoing minimally invasive procedures with the dose determined according to pain scores. All patients undergoing open procedures received patient-controlled analgesia. Nurses recorded pain scores per institutional protocol at the time of scheduled vital signs or patient assessment, and these were recorded on nursing assessment forms and patient-controlled analgesia flow sheets. All narcotics were converted to morphine equivalents for the analysis. Patients in the minimally invasive group routinely received local anesthetic infiltration (0.25% bupivacaine) prior to insertion of abdominal trocars. The patients did not receive scheduled antiemetics postoperatively. Rescue antiemetics were given for the treatment of established symptoms. Data were collected on intraoperative volume of lactated Ringer’s solution, normal saline, colloids, packed red blood cells, and fresh frozen plasma used for resuscitation. We collected information about comorbidities, and preoperative use of narcotics, antidepressants, and anxiolytics.

Since all of the continuous variables, with the exception of age, were nonparametric, the Wilcoxon rank sum test was used to compare the minimally invasive to open surgery groups. Age in the 2 groups was compared using the pooled t test. The Pearson χ ² test was used to compare categorical data based on type of surgery. A multivariate regression model was also performed with 24-hour opioid consumption as the outcome, and surgery type, NSAID use, and interaction between surgery type and NSAID use as predictors. P < .05 was accepted as statistically significant.