Objective
To determine whether patient-controlled analgesia or scheduled intravenous analgesia provides superior pain relief and satisfaction with pain control after vaginal reconstructive surgery.
Study Design
Fifty-nine women scheduled for vaginal reconstructive surgery were enrolled in this randomized trial. Operative procedures and postoperative orders were standardized. Visual analog scales for pain and satisfaction with pain control were recorded during the hospital stay and 2 weeks after surgery.
Results
Patients receiving patient-controlled analgesia had less pain on postoperative day 1, 25 mm vs 39 mm, on visual analog scales ( P = .007). Although this group used twice as much hydromorphone (3.57 mg vs 1.48 mg, P < .001), there was no difference in side effects, length of hospital stay, or complications. For the sample overall, larger amounts of narcotic used correlated with higher pain scores (r = 0.364, P = .009) and worse satisfaction scores (r = −0.348, P = .012).
Conclusion
In patients undergoing vaginal surgery, patient-controlled analgesia offers superior pain relief on postoperative day 1 when compared with scheduled, nurse-administered hydromorphone.
Postoperative pain control is a major factor influencing surgical patients’ recovery and satisfaction. Uncontrolled pain can lead to decreased ambulation, delayed return to normal daily activities, and pulmonary dysfunction. As a result, extensive clinical and research efforts have attempted to select the ideal route of narcotic administration. Patient-controlled analgesia (PCA) is a delivery vehicle for systemic, intravenous (IV), patient-controlled medications. However, this route may not be optimal for all postoperative patients.
The surgical literature to date has compared several postoperative delivery methods and medications. In gynecologic oncology patients, PCA was found to be inferior to patient-controlled epidural analgesia (PCEA) following laparotomy, and in studies evaluating laparoscopic surgery, pain scores were significantly higher for patients receiving PCA compared with intramuscular injection of opioids. However, studies involving laparotomy and laparoscopy are not generalizable to the urogynecology patient population, but there has been little clinical research in this arena. Vaginal reconstructive surgery is unique because of its route, operative time, and dorsal lithotomy positioning.
Therefore, the purpose of this study was to identify differences in pain perception and satisfaction with pain control in women undergoing major vaginal reconstructive surgery receiving hydromorphone by either scheduled, IV nurse-administration, or PCA. As a secondary aim, we sought to compare differences in narcotic usage, postoperative complications, medication side effects, and length of hospital stay.
Materials and Methods
This was a randomized clinical trial conducted at TriHealth hospitals (Good Samaritan Hospital and Bethesda North Hospital) in Cincinnati, OH. It was registered through the US National Institutes of Health clinical trials website, ClinicalTrials.gov , before enrollment and was approved by the TriHealth Institutional Review Board (IRB). All participants gave written informed consent. Subjects were limited to patients of the Division of Urogynecology and Reconstructive Pelvic Surgery.
Subjects were eligible for enrollment if they were English speaking, planning to undergo vaginal reconstructive surgery that included an intraperitoneal vaginal vault suspension and posterior repair, were at least 18 years of age, and able to read the visual analog scales (VAS). Those undergoing concomitant procedures such as vaginal hysterectomy and/or suburethral sling were eligible for enrollment.
Subjects were excluded if they had any of the following: allergy to hydromorphone, currently taking chronic opioids (defined as daily use), renal failure or insufficiency, liver failure, planning to receive regional anesthesia, or undergoing concurrent abdominal or laparoscopic procedures.
Patients signed the informed consent statement in the preoperative area on the day of surgery. After enrollment, a sequentially numbered, opaque, sealed envelope was opened to reveal the patients’ randomization allocation. All procedures were performed by 2 fellowship trained urogynecologists (S.K., R.P.). All subjects were positioned in high dorsal lithotomy using adjustable stirrups. Surgical technique and suture material were identical for both surgeons. Local infiltration of 1% lidocaine with epinephrine (1:100,000) was used for both the anterior and posterior dissections. Subjects undergoing concomitant transobturator suburethral sling received the Monarc subfascial hammock system (American Medical Systems, Minnetonka, MN). Retropubic suburethral slings were placed using the Gynecare tension-free vaginal tape (Gynecare, Ethicon Inc, Somerville, NJ). All suburethral slings were performed using manufacturer recommended techniques.
Routine postoperative orders were standardized. All subjects received postoperative ketorolac, 30 mg every 8 hours for 3 doses. Those over the age of 65 years were administered a modified dose of 15 mg every 8 hours for 3 doses. At the completion of 3 doses, ketorolac was discontinued, and ibuprofen was initiated at a scheduled dose of 800 mg every 8 hours until discharge. To avoid the sedative side effect of promethazine, this medication was not given to any subjects during the postoperative period. Instead, ondansetron 4 mg IV was given initially for complaints of nausea. If nausea was refractory to treatment with ondansetron, metoclopramide 10 mg IV was used. Subjects who tolerated a liquid diet on postoperative day zero were advanced to a regular diet on the morning of postoperative day 1. Once tolerating a regular diet, all IV narcotics were discontinued, and oral narcotics were initiated. On day of discharge, all subjects underwent a voiding trail to verify return of bladder function. Subjects unable to void two-thirds of the volume instilled were discharged home with an indwelling transurethral catheter.
PCA subjects received a 0.2 mg demand dose of hydromorphone with a lock out interval of 8 minutes, a 4-hour maximum of 5 mg, and no continuous rate. Subjects in the PCA group that reported breakthrough pain were managed by giving a 1-time bolus of 0.2 mg and increasing the demand dose by 0.1 mg. In the nurse-administered group, hydromorphone was scheduled every 2 hours at a dose of 0.5 mg. Patients had the option to decline the nurse-administered dose based on their perceived pain level. Breakthrough pain in the nurse-administered group was managed by giving a single dose of 0.2 mg and increasing the interval dose by 0.2 mg; the time interval for delivery remained the same. All patients were able to receive more narcotic if needed for breakthrough pain. Narcotic choice, dosage, and intervals were based on pharmacologic data and determined in coordination with our team of clinical pharmacists. Although initial amount of narcotic provided was different between groups, because of the nature of PCA and interval administration, the amount and timing were chosen to optimize pain control while minimizing potential adverse events for each route of administration.
Although in the postanesthesia care unit, pain was managed by the anesthesia provider and assessed by nursing staff using a 0 to 10 verbal pain scale. A verbal pain score of 4 or less was required before leaving the postoperative recovery area. On patient arrival to their room, it is standard protocol at our institution for nurses to record a verbal pain score while documenting initial vital signs.
The primary outcomes of pain and satisfaction with pain control were assessed using 2 VAS. Patients completed both VAS forms before 8 am on the morning of postoperative day 1. The pain-specific VAS assessed the patients’ current pain level at rest on a 0-100 mm scale with 100 mm being severe pain. The satisfaction-specific VAS assessed satisfaction with pain control and was reverse scored with 100 mm being completely satisfied. Postoperatively, subjects were seen for routine evaluation at 2 weeks. At this visit, they were given the same pain-specific and satisfaction-specific VAS forms and asked to score them based on their recall of pain symptoms while hospitalized after surgery ( Figure 1 ).
A power calculation using nQuery Advisor, version 7.0 (Statistical Solutions, Saugus, MA) indicated a sample size of 27 patients in each arm would be needed to achieve 90% power to detect a mean difference of 2.68 cm on a 100 mm VAS for pain with a significance level of .05. This difference was selected as previous literature has shown that a pain score difference between 2 and 3 cm is significant to most patients. The randomization schedule was created using nQuery Advisor with a 1:1 ratio using mixed blocks. Descriptive statistics were calculated for demographic and dependent variables. A 2-sided Student t test was conducted on continuous variables. Pearson’s correlation coefficient was used to measure the correlation between VAS scores and narcotic use. All statistics were calculated using IBM SPSS Statistics v.19 (SPSS Inc, Chicago, IL).
Results
Fifty-nine subjects were enrolled from the 2 participating hospitals between December 2010 and October 2011. All potential candidates agreed to participate in the study. Five subjects were excluded from analysis; 2 underwent an intraoperative change in surgical procedure, 3 received incorrect postoperative orders ( Figure 2 ). The majority (88%) were white, with a mean age of 61 years (standard deviation [SD] 11.5), and a mean body mass index of 28 (SD 4.1). There were no differences between groups in any demographic variables, surgical data, or length of hospital stay ( Table 1 ).
| Demographic | PCA (n = 30) | Scheduled, nurse-administered (n = 24) | P value |
|---|---|---|---|
| Age, mean (SD) | 61.2 (11.9) | 61.5 (11.2) | .925 |
| Ethnicity | .645 | ||
| White, n (%) | 26 (86.7) | 22 (91.7) | |
| African American, n (%) | 3 (10.0) | 2 (8.3) | |
| Middle Eastern, n (%) | 1 (3.3) | 0 (0) | |
| BMI, mean (SD) | 27.8 (4.2) | 28.5 (4.0) | .592 |
| Procedures performed | |||
| Anterior repair, n (%) | 30 (100.0) | 24 (100.0) | n/a |
| Posterior repair, n (%) | 30 (100.0) | 23 (95.8) | .444 |
| Enterocele repair, n (%) | 27 (90.0) | 23 (95.8) | .395 |
| Vaginal vault suspension, n (%) | 30 (100.0) | 24 (100.0) | n/a |
| Total vaginal hysterectomy, n (%) | 26 (86.7) | 22 (91.7) | .448 |
| Oophorectomy, n (%) | 1 (3.3) | 1 (4.2) | .696 |
| Transobturator sling, n (%) | 7 (23.3) | 6 (25) | .568 |
| Retropubic sling, n (%) | 0 (0) | 1 (4.2) | .444 |
| Cystoscopy, n (%) | 30 (100.0) | 24 (100.0) | n/a |
| Length of surgery (min), mean (SD) | 156.7 (36.7) | 170.4 (34.3) | .166 |
| Length of stay (h), mean (SD) | 32.8 (11.8) | 28.0 (9.1) | .107 |
| Lidocaine used (mL), mean (SD) | 28.7 (8.8) | 33.19 (8.8) | .068 |
Nursing documentation of verbal pain scores on arrival to the floor was compared. The mean pain score, obtained with initial vital signs, was 3.8 (SD 2.45) and did not differ between groups (PCA 3.3 SD 2.6, scheduled nurse-administered 4.33 SD 2.18). However, this similarity of pain scores did not continue through the postoperative period. On postoperative day 1, PCA subjects reported a lower value on the pain-specific VAS, 25 mm, compared with the nurse-administered arm, 39 mm ( P = .007), suggesting that patients in the PCA arm had less pain. The majority of patients (44/54, 81%) were discharged home on postoperative day 1. Of the remaining hospitalized subjects (n = 10) there was no difference in pain or patient satisfaction with pain control when comparing groups on postoperative day 2. However, IV hydromorphone had been discontinued, and all patients were receiving oral pain medication only. At the 2-week postoperative visit, patients were evaluated again using pain-specific and satisfaction-specific VAS forms. At this time, there were no differences between groups for pain or satisfaction ( Table 2 ).
