Background
Postoperative opioid prescribing has historically lacked information crucial to balancing the pain control needs of the individual patient with our professional responsibility to judiciously prescribe these high-risk medications.
Objective
This study aimed to evaluate pain control, satisfaction with pain control, and opioid use among patients undergoing isolated midurethral sling randomized to 1 of 2 different opioid-prescribing regimens.
Study Design
Patients who underwent isolated midurethral sling placement from June 1, 2020, to November 22, 2021, were offered enrollment into this prospective, randomized, open-label, noninferiority clinical trial. Participants were randomized to receive either a standard prescription of ten 5-mg oxycodone tablets provided preoperatively (standard) or an opioid prescription provided only during patient request postoperatively (restricted). Preoperatively, all participants completed baseline demographic and pain surveys, including the 9-Question Central Sensitization Index, Pain Catastrophizing Scale, and Likert pain score (scale 0–10). The participants completed daily surveys for 1 week after surgery to determine the average daily pain score, number of opioids used, other forms of pain management, satisfaction with pain control, perception of the number of opioids prescribed, and need to return to care for pain management. The online Prescription Drug Monitoring Program was used to determine opioid filling in the postoperative period. The primary outcome was average postoperative day 1 pain score, and an a priori determined margin of noninferiority was set at 2 points.
Results
Overall, 82 patients underwent isolated midurethral sling placement and met the inclusion criteria: 40 were randomized to the standard arm, and 42 were randomized to the restricted group. Concerning the primary outcome of average postoperative day 1 pain score, the restricted arm (mean pain score, 3.9±2.4) was noninferior to the standard arm (mean pain score, 3.7±2.7; difference in means, 0.23; 95% confidence interval, −∞ to 1.34). Of note, 23 participants (57.5%) in the standard arm vs 8 participants (19.0%) in the restricted arm filled an opioid prescription ( P <.001). Moreover, 18 of 82 participants (22.0%) used opioids during the 7-day postoperative period, with 10 (25.0%) in the standard arm and 8 (19.0%) in the restricted arm using opioids ( P =.52). Of participants using opioids, the average number of tablets used was 3.4±2.3, and only 3 participants used ≥5 tablets. On a scale of 1=“prescribed far more opioids than needed” to 5=“prescribed far less opioids than needed,” the means were 1.9±1.0 in the standard arm and 2.7±1.0 in the restricted arm ( P <.001).
Conclusion
Restricted opioid prescription was noninferior to standard opioid prescription in the setting of pain control and satisfaction with pain control after isolated midurethral placement. Participants in the restricted arm filled fewer opioid prescriptions than participants in the standard arm. On average, only 3.4 tablets were used by those that filled prescriptions in both groups. Restrictive opioid-prescribing practices may reduce unused opioids in the community while achieving similar pain control.
Why was this study conducted?
The opioid epidemic continues to be a crisis in the United States. Understanding factors that influence opioid use and dispersion after midurethral sling will assist surgeons in opioid stewardship efforts.
Key findings
Despite markedly less opioid dispersion, participants in the restricted arm reported similar pain and satisfaction with pain control to those routinely given postoperative opioids.
What does this add to what is known?
This study provided quantification on the number of opioids used after midurethral sling and informed providers of the role that our prescribing practices play in opioid use, pain control, and excess opioid dispersion.
Introduction
The opioid epidemic continues to loom over the United States, with the number of opioid-related deaths tripling between 2007 and 2015. Overprescribing is a leading contributor to the opioid crisis. More than 70% to 80% of prescribed postoperative opioids remain unused, and patients are faced with unclear disposal options. , These unused opioids are at risk of diversion, and 74% of individuals with opioid use disorder reported obtaining their pills directly from a provider or indirectly through another individual with an opioid prescription. , These gaps in our current opioid prescribing present an opportunity to mitigate avoidable harm by reducing unnecessary opioid use and misuse through better tailored opioid prescription practices.
Despite the potential dangers of opioids to both the individual and society, there is a paucity of data available to guide surgeon decision-making concerning the number of opioids prescribed to their patients during discharge. Postoperative opioid prescribing has few standards, wide variations, and inconsistent use. , , Within obstetrics and gynecology, literature establishing appropriate postoperative opioid-prescribing benchmarks for hysterectomy and pelvic organ prolapse reconstruction has been growing; however, there is currently scant information describing opioid use after many common genitourinary operations, including an isolated midurethral sling (MUS). , , ,
Since their introduction in the 1990s, MUS has provided a safe and effective treatment of stress urinary incontinence for millions of women. These procedures are done almost exclusively with same-day discharge and are minimally invasive, but postoperative opioid-prescribing guidelines for MUS are nonexistent. Within our local practices, we observed varied postdischarge opioid-prescribing practices, ranging from routine provision of opioid prescriptions at a preoperative visit to counseling patients that opioids would not likely be needed for pain control. Thus, this study was designed to provide a more objective rationale for opioid-prescribing practices after isolated MUS. We hypothesized that postoperative pain control on postoperative day (POD) 1 would be noninferior between patients routinely prescribed opioids and those not routinely prescribed opioids after this procedure. Moreover, we sought to evaluate differences in satisfaction with pain control and opioid use between these arms and to identify clinical or demographic factors associated with higher opioid use.
Materials and Methods
This was a randomized, open-label, noninferiority clinical trial to evaluate pain control, satisfaction with pain control, and opioid use in patients after isolated MUS. Approval was obtained from the Pennsylvania State University College of Medicine Institutional Review Board (STUDY13951), and the study was registered at ClinicalTrials.gov (identifier: NCT04277975 ). As an investigator-initiated trial, funding was received from an internal research grant.
Eligible participants were women undergoing isolated MUS by a Female Pelvic Medicine and Reconstructive Surgery staff physician at a Penn State Health hospital from June 1, 2020, to November 22, 2021. Patients who were pregnant, nursing, cognitively impaired, currently using a daily opioid or with opioid use disorder, allergic to the study opioid (oxycodone) or acetaminophen or nonsteroidal antiinflammatory drug, or unable to speak or read English were excluded. Patients with chronic pain disorders were included in this study. The participants gave informed consent and were enrolled by a member of the study team at their preoperative appointments or in the preoperative area the morning of their scheduled surgery. Preoperatively, baseline demographic and medical history information were obtained ( Table 1 ), and the participants completed validated surveys, including the 9-Question Central Sensitization Index (CSI-9), 13-Item Pain Catastrophizing Scale (PCS), and Likert pain score (scale 0–10). The total CSI-9 score was computed by summing responses to all 9 items and ranged from 0 to 36. The total PCS score was computed by summing responses to all 13 items and ranged from 0 to 52. A total PCS score of 30 was considered clinically relevant catastrophizing.
| Variable | Overall (N=82) | Standard (n=40) | Restricted (n=42) | P values |
|---|---|---|---|---|
| Demographics | ||||
| Age (y) | 53.1±11.9 | 55.0±11.9 | 51.2±11.8 | .15 |
| Parity | 2.5±1.2 | 2.2±1.0 | 2.7±1.2 | .03 |
| BMI (kg/m 2 ) | 31.5±7.0 | 31.6±6.0 | 31.5±7.9 | .93 |
| Hispanic ethnicity | 10 (12.3) | 6 (15.0) | 4 (9.8) | .52 |
| White race | 74 (91.4) | 36 (90.0) | 38 (92.7) | .71 |
| Married | 55 (67.1) | 25 (62.5) | 30 (71.4) | .48 |
| Commercial insurance | 59 (72.0) | 25 (62.5) | 34 (81.0) | .09 |
| Past medical history | ||||
| Current tobacco use | 10 (12.3) | 4 (10.0) | 6 (14.6) | .74 |
| Current alcohol use | 34 (42.5) | 18 (46.2) | 16 (39.0) | .65 |
| Depression | 30 (36.6) | 17 (42.5) | 13 (31.0) | .36 |
| Anxiety | 33 (40.2) | 12 (30.0) | 21 (50.0) | .08 |
| Panic attacks | 19 (23.2) | 8 (20.0) | 11 (26.2) | .60 |
| FM | 8 (9.8) | 5 (12.5) | 3 (7.1) | .48 |
| Restless leg syndrome | 4 (4.9) | 2 (5.0) | 2 (4.8) | 1.00 |
| Chronic fatigue syndrome | 1 (1.2) | 1 (2.5) | 0 (0.0) | .49 |
| TMJ disorder | 4 (4.9) | 2 (5.0) | 2 (4.8) | 1.00 |
| Migraine or tension headaches | 27 (32.9) | 10 (25.0) | 17 (40.5) | .16 |
| CPP | 2 (2.4) | 1 (2.5) | 1 (2.4) | 1.00 |
| IBS | 14 (17.1) | 7 (17.5) | 7 (16.7) | 1.00 |
| IC | 3 (3.7) | 2 (5.0) | 1 (2.4) | .61 |
| Chronic pain disorder | 39 (47.6) | 18 (45) | 21 (50) | .67 |
| Pain and psychometric scores | ||||
| Preoperative pain score | 0.8±1.5 | 1.0±1.8 | 0.6±1.1 | .29 |
| PCS total score | 7.4±8.4 | 7.3±8.3 | 7.5±8.5 | .89 |
| PCS rumination | 2.9±3.4 | 2.9±3.4 | 3.0±3.4 | .87 |
| PCS magnification | 1.5±1.9 | 1.7±2.2 | 1.4±1.6 | .50 |
| PCS helplessness | 3.0±3.9 | 2.8±3.6 | 3.2±4.2 | .63 |
| CSI-9 total score | 13.8±6.2 | 12.5±6.8 | 15.0±5.3 | .06 |
The participants were randomized to receive either a standard prescription of ten 5-mg oxycodone tablets provided preoperatively (standard) or an opioid prescription of ten 5 mg tablets provided only during patient request postoperatively (restricted). The randomization allocation sequence was generated by the statistician using permuted blocks of random sizes 2, 4, and 6 with a 1:1 allocation ratio to standard or restricted regimen. The randomization sequence was uploaded directly into Research Electronic Data Capture (REDCap; Vanderbilt University, Nashville, TN). Randomization was performed by the study team surgeon using the REDCap randomization module on the day of surgery. Providers, outcome observers, or study participants were not masked concerning the allocation after inclusion. All participants received detailed, standardized instruction on nonopioid pain control methods postoperatively, including the use of nonsteroidal antiinflammatories, acetaminophen, and heat and ice.
After MUS, participants completed a daily diary for 1 week, that is, POD 0 to POD 7, to determine average daily pain score, opioid use and number used, other forms of pain management, satisfaction with pain control, perception of the number of opioids prescribed, and need to return to care for pain management ( Appendix ). Diaries were collected electronically via REDCap or on a paper log if there is no Internet access. Participants unable to complete their diary in REDCap received a structured phone call at POD 1 and POD 7 to collect paper diary information. Furthermore, all participants were asked to return their paper diary (if applicable) at a postoperative visit. The online Prescription Drug Monitoring Program (PDMP) was queried for all participants to determine if prescriptions for opioids were filled during the postoperative period.
The primary outcome for this noninferiority study was POD 1 pain based on a Likert scale ranging from 0 to 10, where higher scores indicate worse pain. The secondary outcomes included whether the participant filled an opioid prescription (indicated by the online PDMP), opioid use (yes or no), satisfaction with pain control (on a scale of 1=“much worse” to 3=“same” to 5=“much better” than expected), and how participants felt about the number of opioids prescribed (on a scale of 1=“prescribed far more” to 3=“prescribed the right number” to 5=“prescribed far less” opioids than needed). Opioid use was determined using the daily POD 0 to POD 7 diaries and defined as using at least one 5-mg oxycodone tablet. An average score per participant was calculated for the 5-point Likert scale questions regarding satisfaction and number of opioids prescribed, which were both assessed on the daily diaries.
A 2-point difference on an 11-point Likert pain scale has previously been described as a clinically relevant difference in the level of pain experienced among vaginal surgery treatment groups. Therefore, we believed using a 2-point margin of noninferiority was appropriate for sample size estimation of our 2-arm trial. Using data reported from this previous study, we estimated a pain scale standard deviation (SD) of 3 points. Furthermore, for sample size estimation, we assumed the true difference between the means of the 2 arms was zero. Given these assumptions, a sample size of 37 per arm was calculated to provide 81% statistical power to detect noninferiority between the 2 arms using a 1-sided test having a significance level of.025. Concerning the potential for a 10% loss of information for the primary outcome (eg, dropout), the target sample size for the study was 84 participants.
Analyses were performed using SAS software (version 9.4; SAS Institute, Cary, NC). A 1-sided, 2-sample t test was used to test the noninferiority hypothesis where the margin of noninferiority was 2 points on the Likert pain scale. The 2-sample t tests were used to compare continuous variables, including area under the curve (AUC) pain scores, average satisfaction scores, and average rating of the number of opioids prescribed. Chi-squared tests, or Fisher exact tests if expected counts were small, were used to compare categorical variables, including whether a prescription was filled and whether opioids were used. In addition, a linear mixed model was fit and contrasts constructed to compare the 2 arms at each POD concerning pain scores. The linear mixed model included a random subject effect with a first-order autoregressive covariance structure to account for the correlation of repeated measurements per subject. Residual diagnostics were used to assess parametric assumptions (eg, normality) for continuous variables, and no violation of these assumptions was noted.
Exploratory analyses were performed to investigate predictors of postoperative pain score and opioid use. A multiple binary logistic regression model was used to assess the association of baseline factors (≥1 chronic pain disorders, PCS score, CSI-9 score, postanesthesia care unit [PACU] opioid use, baseline pain score, and randomization assignment) with postoperative opioid use. A multiple linear regression model assessed the association of these same baseline factors with postoperative pain scores.
Results
A total of 83 participants were randomized on the day of surgery; 1 participant was later excluded for complications incurred during surgery and inability to place a sling. The final cohort for analysis included 82 participants: 40 in the standard arm and 42 in the restricted arm. Figure 1 shows a Consolidated Standards of Reporting Trials flow diagram of study participants. Table 1 depicts the demographic and clinical characteristics of the participants. The 2 arms were similar in baseline demographics. Most participants, 33 (82.5%) in the standard arm and 36 (85.7%) in the restricted arm, opted to receive their diaries electronically.
