Advertisement

A randomized controlled noninferiority trial of reduced vs routine opioid prescription after prolapse repair

Published:March 18, 2020DOI:https://doi.org/10.1016/j.ajog.2020.03.017

      Background

      Given the accelerating opioid crisis in the United States and evidence that patients use fewer opioid tablets than prescribed, surgeons may choose to decrease prescribed quantities. The effect this may have on patient satisfaction with pain control after hospital discharge is unknown.

      Objective

      The primary objective of this study was to compare patient satisfaction with postoperative pain control between patients receiving a routine or reduced quantity opioid prescription after prolapse repair. Secondary objectives included a comparison of opioid-related side-effects, the number of opioid tablets used, and the number of excess tablets prescribed between these groups.

      Study Design

      This was a single-center, unmasked, 2-arm, randomized controlled noninferiority trial of women who underwent a prolapse repair with a planned overnight hospitalization. Patients were assigned randomly to 1 of 2 study arms: routine (28 tablets of oxycodone 5 mg) or reduced (5 tablets) prescription of opioid tablets. Patients were eligible if they were at least 18 years of age and undergoing a prolapse repair with an anticipated overnight hospital stay. Exclusion criteria included a history of chronic pain, preoperative opioid use, intolerance to study medication, or a score of ≥30 on the Pain Catastrophizing Scale. In addition to their opioid prescription, all patients received multimodal pain medications at discharge. Patients were asked to complete 6 weeks of diaries to record pain and medication use. The primary outcome (patient satisfaction) was collected as part of a postoperative survey completed at patients’ routine postoperative visit 6 weeks after surgery. The sample size for noninferiority was calculated at 59 patients per group for a total of 118 patients.

      Results

      One hundred eighteen patients were assigned randomly; the primary outcome was available for 116. The majority of patients were white, postmenopausal, and nonsmokers; the mean age was 62±10.4 years. The most common surgery was a hysterectomy with native tissue repair (n=71; 60%). One hundred ten patients (93%) were satisfied with postoperative pain control. Statistical analysis constructed for noninferiority showed that the difference between the groups was <15% (93% vs 93%; P=.005). Subjects in the reduced arm reported requiring an additional opioid prescription more frequently than in the routine arm (15% vs 2%; P=.01). Patients in the routine arm used more opioid tablets than the reduced arm (median, 3 [interquartile range, 0–14] vs 1 [interquartile range, 0–3]), but overall opioid utilization was low. As such, patients in the routine arm had significantly more unused opioid tablets (median, 26 [interquartile range, 15–28] vs 4 [interquartile range, 2–5]).

      Conclusion

      Patient satisfaction with pain control was noninferior in patients who received a reduced quantity of opioid tablets after prolapse repair compared with those who received a routine prescription. A large quantity of excess opioid tablets was seen in both groups. Surgeons should consider prescribing 5–10 opioid tablets after prolapse repair surgery and consider applying these findings to postoperative prescribing after other gynecologic procedures.

      Key words

      The opioid abuse epidemic in the United States is a deadly issue. Overdose leads to >130 deaths per day; up to 40% of these deaths involve prescription opioids. Guidelines for postoperative prescribing based on expert consensus have been published for common surgical procedures,
      • Overton H.N.
      • Hanna M.N.
      • Bruhn W.E.
      • et al.
      Opioid-prescribing guidelines for common surgical procedures: an expert panel consensus.
      but there are no guidelines for postoperative prescribing after surgery for pelvic organ prolapse.
      In practice, surgeons often prescribe a standard amount of pain medication after surgery in anticipation of expected postoperative pain. Recent studies in the gynecology and urogynecology literature have reported that the quantity of prescribed opioid tablets exceeds the number of tablets actually used by patients, often by 2–3 times the necessary amount.
      • As-Sanie S.
      • Till S.R.
      • Mowers E.L.
      • et al.
      Opioid prescribing patterns, patient use, and postoperative pain after hysterectomy for benign indications.
      • Griffith K.C.
      • Clark N.V.
      • Zuckerman A.L.
      • Ferzandi T.R.
      • Wright K.N.
      Opioid prescription and patient use after gynecologic procedures: a survey of patients and providers.
      • Hota L.S.
      • Warda H.A.
      • Haviland M.J.
      • Searle F.M.
      • Hacker M.R.
      Opioid use following gynecologic and pelvic reconstructive surgery.
      • Swenson C.W.
      • Kelley A.S.
      • Fenner D.E.
      • Berger M.B.
      Outpatient narcotic use after minimally invasive urogynecologic surgery.
      Although new persistent opioid use after surgery fortunately is uncommon,
      • Brummett C.M.
      • Waljee J.F.
      • Goesling J.
      • et al.
      New persistent opioid use after minor and major surgical procedures in US adults.
      this opioid excess contributes to our current opioid crisis because excess tablets may be available for abuse or diversion. Researchers have attempted to find ways to better manage postoperative pain with minimization of opioid prescribing. One way of accomplishing this is through the use of perioperative multimodal pain therapy.
      • Reagan K.M.L.
      • O’Sullivan D.M.
      • Gannon R.
      • Steinberg A.C.
      Decreasing postoperative narcotics in reconstructive pelvic surgery: a randomized controlled trial.
      Despite recent data that opioid tablet reduction is both favorable and desirable, surgeons may be wary of prescribing fewer tablets for fear of impacting patient satisfaction. As such, the objective of this study is to compare satisfaction with postoperative pain control in women who receive either a routine or a reduced quantity opioid prescription after pelvic organ prolapse surgery. We hypothesize that patient satisfaction with postoperative pain control in patients who receive a reduced quantity of opioid tablets would be noninferior to the satisfaction of those who receive the routine quantity.

      Why was this study conducted?

      Opioid tablet excess after surgery contributes to the deadly opioid epidemic. It is not known whether patient satisfaction may be impacted if surgeons reduce the quantity of opioid tablets prescribed after prolapse surgery.

      Key findings

      Patient satisfaction with pain control is not inferior when they compare 5 tablets vs 28 tablets of oxycodone after prolapse repair surgery. Excess opioid tablets prescribed was significantly >28 tablets.

      What does this add to what is known?

      This trial supports previous recommendations for decreasing opioid tablets that are prescribed after surgery. This new data regarding patient satisfaction may reassure surgeons who are concerned with the impact of decreasing postoperative opioid prescription quantities.

      Methods

      This was a 2-arm, unmasked, single-center randomized controlled noninferiority trial that compared a routine quantity (28 tablets of oxycodone 5 mg) with a reduced quantity (5 tablets of oxycodone 5 mg) of postdischarge opioid prescription. A noninferiority design was chosen because we wanted to show that reducing prescription quantities would have no impact on patient satisfaction, not that it would improve satisfaction. The study was approved after full review by the Cleveland Clinic Institutional Review Board and was registered on ClinicalTrials.gov (NCT03319277). Funding for this research was provided by an AAGL research grant. Eligible patients were women ≥18 years old who were undergoing pelvic organ prolapse procedures that included transvaginal vault suspension or minimally invasive sacrocolpopexy with or without hysterectomy with an anticipated overnight hospital stay. Patients were excluded if they had chronic pain, preoperative opioid use, or intolerance or allergy to study medications (oxycodone, acetaminophen, and ibuprofen). Patients also were excluded if the surgical plan included an anal sphincteroplasty, nonurogynecologic concurrent surgery that included a colorectal procedure, or if they were planning on discharge the same day as the surgery. Additionally, the Pain Catastrophizing Scale was administered to all eligible patients,
      • Sullivan M.
      • Biship S.
      • Pivik J.
      The pain catastrophizing scale: development and validation.
      and patients with a score of ≥30 were excluded because this 75th percentile score predicts a greater postoperative pain rating.
      • Jacobsen P.B.
      • Butler R.W.
      Relation of cognitive coping and catastrophizing to acute pain and analgesic use following breast cancer surgery.
      All patients were screened and invited to the study by the primary investigator.
      Patients were assigned randomly at their preoperative visit with the use of a sequentially numbered, opaque, sealed envelopes by either a research nurse or the primary investigator. These envelopes were created before the study entered the recruitment phase with the use of a random number sequence generator. Those women who were assigned randomly to the reduced arm received a paper prescription for 10 tablets of oxycodone 5 mg, which was an Institutional Review Board requirement for the study. This back-up prescription was dated so that it could not be filled until 2 days after anticipated hospital discharge. Inpatient pain regimens were not protocolized to the study, but the division’s standard postoperative order set included scheduled intravenous ketorolac and oral acetaminophen with “as needed” oral oxycodone for severe pain; intravenous hydromorphone was available for breakthrough pain. At the time of discharge, patients were prescribed their assigned quantity of opioid tablets (5 or 28 tablets of oxycodone 5 mg), and all patients received prescriptions for ibuprofen 600 mg and acetaminophen 500 mg. Patients in both arms were instructed on alternating, around-the-clock use of the nonopioid medications for 3 days with instructions to use these before oxycodone. They also received a standardized bowel regimen of polyethylene glycol and docusate.
      After discharge, subjects were asked to complete 42 days of diaries to record their pain, medication use that included oxycodone, ibuprofen, and acetaminophen, and the ability to perform specific daily activities. They were also asked to estimate their percentage returned to their baseline status (Appendix 1). In addition, at their routine postoperative visit, patients completed a final survey (Appendix 2) that assessed subjective outcomes, with the use of the validated Pain Treatment Satisfaction Scale, that included our primary outcome of satisfaction with pain control.
      • Evans C.J.
      • Trudeau E.
      • Mertzanis P.
      • et al.
      Development and validation of the Pain Treatment Satisfaction Scale (PTSS): a patient satisfaction questionnaire for use in patients with chronic or acute pain.
      Other outcomes that were assessed in this survey were patient-reported bother with constipation or nausea and dizziness, which was rated on a scale from 0–10 (0=no bother; 10=worst imaginable bother). Subjects were asked if they required a prescription for opioid tablets beyond their original discharge prescription and about their willingness to destroy excess opioid tablets present at home. The medical record was queried for relevant demographic and perioperative data that included a calculation of the opioid use while admitted after surgery as calculated by oral morphine equivalents (OME).
      Sample size was calculated based on the noninferiority nature of the research hypothesis: patient satisfaction with pain control in the reduced quantity group would be noninferior to the routine group. Satisfaction was a binary outcome defined a priori as a response of either “very satisfied” or “somewhat satisfied” to the question “Overall, how satisfied are you with your pain medication at home since your surgery?” (from the validated Pain Treatment Satisfaction Scale) that was asked at their postoperative visit. With a predicted 92% satisfaction in the control group
      • Crisp C.C.
      • Khan M.
      • Lambers D.L.
      • et al.
      The effect of intravenous acetaminophen on postoperative pain and narcotic consumption after vaginal reconstructive surgery: a double-blind randomized placebo-controlled trial.
      and a planned 15% inferiority margin, the sample size required was 57 patients per group to achieve a power of 90% at an alpha of .05. To account for a 3% no-show rate at the postoperative visit, 2 patients per arm were added for a total sample size of 118 patients. For the primary outcome, the Farrington-Manning score test for noninferiority was performed for satisfaction between the 2 groups with intention-to-treat analysis. A secondary per-protocol analysis was also planned for the primary outcome. Other comparisons were performed by Student t test, chi square test, or Wilcoxon rank sum for nonparametric data, as appropriate. Data are presented as number (%), mean±standard deviation, or median (interquartile range[IQR]) if the data were not distributed normally. Analyses were performed with JMP software (version 12.0) and SAS software (version 9.4; The SAS Institute, Cary, NC); P<.05 was considered statistically significant.

      Results

      Recruitment began October 2017 and continued through February 2019. The results of recruitment, randomization, and follow up are demonstrated in Figure 1 in the CONSORT diagram. Six hundred sixteen patients were assessed for eligibility; 327 of them met inclusion criteria. Of those who met inclusion criteria, 36 patients declined to be in the study, and 173 were excluded. The most common reasons for exclusion were intolerance or allergy to study medications (n=63) or the presence of chronic pain (n=55). Common reasons for declining to participate were patient preference to receive no opioid tablets at all after surgery and a concern about postoperative pain control. Data are presented according to the planned intention-to-treat analysis.
      Figure thumbnail gr1
      Figure 1CONSORT diagram: enrollment, randomization, and follow up
      Flow chart of study enrollment, allocation, and patient follow-up.
      Davidson et al. Reducing postoperative opioids. Am J Obstet Gynecol 2020.
      Of the 118 randomly assigned patients, most were white, postmenopausal, and nonsmokers; the mean age was 62±10.4 years. Table 1 shows the demographic and perioperative data for all patients. The most common surgery in both arms of the study was a hysterectomy with native tissue prolapse repair (n=71; 60%). Seventy-five percent of the patients (n=88) underwent a posterior colporrhaphy during the surgery, which was not different between arms. Patients in the routine arm used more oral opioid tablets in their postoperative hospital stay than those in the reduced arm (15 vs 7.5 OME; P=.02); this equates clinically to 1 additional oxycodone 5-mg tablet. Otherwise the groups had no statistically significant demographic or perioperative differences.
      Table 1Demographic and perioperative data (n=118)
      VariableAll subjects (N=118)Reduced arm (n=59)Routine arm (n=59)P value
      Age, y
      Data are presented as mean±standard deviation, analysis by Student t test
      62±10.463±9.061±11.6.25
      Race, n (%)
      Analysis by chi square test
      .47
       White106 (91.4)56 (94.9)50 (87.7)
       Black5 (4.3)2 (3.4)3 (5.3)
       Hispanic1 (0.9)01 (1.8)
       Other4 (3.4)1 (1.7)3 (5.3)
       Missing202
      Current tobacco use, n (%)
      Analysis by chi square test
      2 (1.7)1 (1.7)1 (1.7).98
      Menopausal status, n (%)
      Analysis by chi square test
      .37
       Postmenopausal99 (83.9)51 (86.4)48 (81.4)
       Perimenopausal1 (0.01)1 (1.7)0
       Premenopausal18 (15.3)7 (11.9)11 (18.6)
      Pain Catastrophizing Scale, total
      Data are presented as median (range), analysis by Wilcoxon rank sum test because the data were not distributed normally
      6.5 (0–28)6 (0–28)7 (0–27).55
       Rumination subscale4 (0–12)4 (0–11)4 (0–12).76
       Magnification subscale1 (0–6)1 (0–6)1 (0–6)1.0
       Helplessness subscale2 (0–12)2 (0–12)3 (0–12).18
      Highest level of education, n (%)
      Analysis by chi square test
      .49
       Less than high school1 (0.9)01 (1.8)
       High school/equivalent33 (28.5)15 (25.4)18 (31.6)
       Some college40 (34.5)24 (40.7)16 (28.1)
       College graduate18 (15.5)9 (15.3)9 (15.8)
       Graduate/professional degree24 (20.7)11 (18.6)13 (22.8)
      Yearly household income, n (%)
      Analysis by chi square test
      .15
       <$30,00017 (15.7)8 (14.8)9 (16.7)
       $30–50,00013 (12.0)6 (11.1)7 (13.0)
       $50–75,00026 (24.1)18 (33.3)8 (14.8)
       >$75,000052 (48.2)22 (40.7)30 (55.6)
      Surgery type, n (%)
      Analysis by chi square test
      .79
       Vaginal hysterectomy and colpopexy71 (60.2)35 (59.3)36 (61.0)
       Vaginal colpopexy18 (15.3)9 (7.6)9 (15.3)
       Vaginal hysteropexy18 (15.3)8 (13.6)10 (17.0)
       Sacrocolpopexy
      All sacrocolpopexies were completed by laparoscopy, either conventional or with robotic-assistance.
      11 (9.3)7 (11.9)4 (6.8)
      Total oral morphine equivalents used postoperatively
      Data are presented as median (range), analysis by Wilcoxon rank sum test because the data were not distributed normally
      8.75 (0–112.5)7.5 (0–75)15 (0–112.5).02
      Davidson et al. Reducing postoperative opioids. Am J Obstet Gynecol 2020.
      a Data are presented as mean±standard deviation, analysis by Student t test
      b Analysis by chi square test
      c Data are presented as median (range), analysis by Wilcoxon rank sum test because the data were not distributed normally
      d All sacrocolpopexies were completed by laparoscopy, either conventional or with robotic-assistance.
      The results of the postoperative visit survey are given in Table 2. Two patients in the reduced arm did not complete these surveys. As such, data on patient satisfaction were available for 116 patients, which met our sample size requirement. Of the survey respondents, 108 patients (93%) were satisfied with their postoperative pain control at home. Patient satisfaction in the reduced arm was noninferior based on the a priori 15% margin: (93% vs 93%; P=.005). In the per protocol analysis, there were 53 patients in the reduced arm and 52 patients in the routine arm; the absolute difference of satisfaction rates between the 2 groups was 0.11% (90% confidence interval, –9.5% to +9.7%), and the 1-sided probability value for the noninferiority test was .005. The difference in satisfaction between the groups remained <15%.
      Table 2Postoperative survey data (n=116)
      VariableAll subjects (n=116)Reduced arm (n=57)Routine arm (n=59)P value
      Satisfaction with pain control (yes), n (%)
      Analysis performed by Farrington-Manning score test for noninferiority with a 15% noninferiority index. Satisfaction in this binary analysis was defined a priori as a response of “Very satisfied” or “Somewhat satisfied”
      108 (93.1)53 (93.0)55 (93.2).05
      Analysis performed by Farrington-Manning score test for noninferiority with a 15% noninferiority index. Satisfaction in this binary analysis was defined a priori as a response of “Very satisfied” or “Somewhat satisfied”
      Satisfaction with pain control, n (%)
      Analysis performed by chi square test
      .88
       Very satisfied96 (82.8)47 (82.5)49 (83.1)
       Somewhat satisfied12 (10.3)6 (10.5)5 (10.2)
       Neutral2 (1.7)1 (1.8)1 (1.7)
       Somewhat unsatisfied1 (0.9)01(1.7)
       Very unsatisfied5 (4.3)3 (5.3)2 (3.4)
      Additional opioid prescription filled, n (%)
      Analysis performed by chi square test
      9 (8.3)8 (14.8)1 (1.9).01
      Bother with constipation and/or nausea
      Data are presented as mean±standard deviation, analysis performed by Student t test
      3.9±2.63.8±2.64.0±2.7.81
      Bother with dizziness
      Data are presented as mean±standard deviation, analysis performed by Student t test
      2.4±1.42.5±1.52.4±1.4.79
      Willing to destroy excess opioid tablets, n (%)
      Analysis performed by chi square test
      101 (95.3)50 (94.3)51 (96.2).65
      Opioid tablets used, n
      Data are presented as median (interquartile range), performed the Wilcoxon Rank Sum test because the data were not distributed normally.
      2 (0–5.75)1 (0–3)3 (0–14).03
      Unused opioid tablets, n
      Data are presented as median (interquartile range), performed the Wilcoxon Rank Sum test because the data were not distributed normally.
      5 (3–26)4 (2–5)26 (15.3–28)<.01
      Davidson et al. Reducing postoperative opioids. Am J Obstet Gynecol 2020.
      a Analysis performed by Farrington-Manning score test for noninferiority with a 15% noninferiority index. Satisfaction in this binary analysis was defined a priori as a response of “Very satisfied” or “Somewhat satisfied”
      b Analysis performed by chi square test
      c Data are presented as mean±standard deviation, analysis performed by Student t test
      d Data are presented as median (interquartile range), performed the Wilcoxon Rank Sum test because the data were not distributed normally.
      Patients in the routine arm used more opioid tablets, with a median number of 3 (IQR, 0–14) compared with 1 (IQR, 0–3) in the reduced arm (P=.03). Those patients in the routine arm had more excess opioid tablets, with a median of 25 (IQR, 15.3–28) leftover tablets of oxycodone per patient compared with 4 tablets (IQR, 2–5) in the reduced arm. Patients in the reduced arm were more likely to report filling an additional opioid prescription (15% vs 2%; P=.02). For opioid-related side-effects, overall bother with these symptoms was low (Table 2), with no difference in patient-reported bother of constipation and/or nausea or dizziness between groups. Most patients were willing to destroy remaining opioid tablets (n=101; 95%); this was not different between arms.
      For the daily diaries, scatter plots with smooth curve fitted by Loess were created for maximum daily pain scores and percentage return to normal; the scores were averaged over the week (Figures 2 and 3). Table 3 summarizes this information in median (IQR). To prevent type 1 error rate inflation, we tested only the differences between groups for weeks 1 and 6. Although no significant differences were found in these 2 time points, there were trends towards improvement in patient-reported outcomes within the reduced arm. Figure 2 shows that 2 arms had similar maximum daily pain scores score at week 1; the reduced arm had noticeably lower (better) scores from weeks 2–5; then the scores became similar again at the week 6. Figure 3 shows that the reduced arm had a slightly, but consistently, improved return to baseline scores over all 6 weeks.
      Figure thumbnail gr2
      Figure 2Weekly averaged worst pain rating: Loess plot
      Patients rated their return to their baseline status each day after surgery from zero (not at all themselves) to one hundred percent (completely returned to normal). The reduced arm showed improved return to their normal status.
      Davidson et al. Reducing postoperative opioids. Am J Obstet Gynecol 2020.
      Figure thumbnail gr3
      Figure 3Self-rated return to normal (%): Loess plot
      Patients rated their return to their baseline status each day after surgery from zero (not at all themselves) to one hundred percent (completely returned to normal). The reduced arm showed swifter return to patients' preoperative status.
      Davidson et al. Reducing postoperative opioids. Am J Obstet Gynecol 2020.
      Table 3Subject diaries: medication use, pain, and return to baseline status (n=107)
      VariableTotal (N=107)Reduced arm (n=56)Routine arm (n=51)P value
      nMedian (interquartile range)nMedian (interquartile range)
      Worst daily pain
       Week 1 average4.0 (2.3–5.7)563.6 (2.4–5.4)504.5 (2.3–5.9).36
      Performed with the Wilcoxon Rank Sum test because the data were not distributed normally.
       Week 2 average1.9 (1.00–3.3)561.6 (0.86–3.0)492.0 (1.00–3.9)
       Week 3 average1.00 (0.29–2.1)531.00 (0.14–1.7)461.3 (0.57–3.0)
       Week 4 average0.77 (0.00–2.0)510.43 (0.00–1.7)451.00 (0.00–2.1)
       Week 5 average0.29 (0.00–1.3)510.00 (0.00–1.3)420.34 (0.00–1.3)
       Week 6 average0.00 (0.00–0.75)510.00 (0.00–1.00)380.00 (0.00–0.71).78
      Performed with the Wilcoxon Rank Sum test because the data were not distributed normally.
      Self-rated return to normal
       Week 1 average48.6 (33.1–65.7)5653.5 (38.7–67.3)4945.0 (28.7–60.0).082
      Performed with the Wilcoxon Rank Sum test because the data were not distributed normally.
       Week 2 average72.4 (53.3–84.9)5676.4 (60.8–85.9)4760.0 (45.7–80.7)
       Week 3 average79.2 (66.0–89.9)5183.1 (72.9–91.9)4572.9 (57.1–86.0)
       Week 4 average85.4 (68.6–94.4)5087.9 (80.0–94.6)4478.7 (61.1–91.9)
       Week 5 average91.4 (78.6–98.3)5092.3 (83.6–99.3)4189.1 (76.4–96.3)
       Week 6 average93.8 (87.1–99.6)5095.3 (89.3–99.8)3793.0 (85.0–99.3).56
      Performed with the Wilcoxon Rank Sum test because the data were not distributed normally.
      Ibuprofen tablets (600 mg) used, n20.0 (9.7–38.0)5620.0 (9.9–33.0)5016.5 (8.0–44.9).96
      Performed with the Wilcoxon Rank Sum test because the data were not distributed normally.
      Tylenol tablets (500 mg) used26.5 (11.0–43.0)5625.5 (14.0–42.5)5027.5 (6.0–45.0).75
      Performed with the Wilcoxon Rank Sum test because the data were not distributed normally.
      Davidson et al. Reducing postoperative opioids. Am J Obstet Gynecol 2020.
      a Performed with the Wilcoxon Rank Sum test because the data were not distributed normally.
      Regarding adjunct pain medication use (ibuprofen and acetaminophen), subjects used a median of 20 tablets of ibuprofen 600 mg and 27 tablets of acetaminophen 500 mg (Table 3). There were no differences between arms.
      As a post hoc analysis, we assessed whether different factors could be used to predict postoperative opioid intake. Linear regression was used for the routine arm that investigated if age, type of surgery (hysterectomy or not), or inpatient opioid use impacted postdischarge opioid use within the routine arm. Only inpatient OME was predictive of increased opioid use. This analysis also was performed in the reduced arm, and OME was also predictive of increased use of opioids in this arm.
      There were 11 protocol deviations: Four patients were discharged same day; 2 patients had sphincteroplasties added to their case, which was an exclusion criteria; 3 patients had severe pain postoperatively and had their discharge opioid plan changed from oxycodone to hydromorphone (pain was unrelated to nerve entrapment or other complications); 1 patient could not tolerate oxycodone and was switched to tramadol, and 2 patients had acute kidney injury that prohibited their ability to receive ibuprofen per protocol.

      Comment

      Principal findings

      The goal of this study was to compare reduced and routine quantity postoperative opioid prescription and assess for an effect on patient satisfaction with pain control. Patient satisfaction with postoperative pain control was noninferior in patients who received fewer opioid tablets at discharge after prolapse surgery, despite an increased frequency of needing an additional opioid prescription in the reduced arm. We found that patients who received a reduced prescription used fewer opioid tablets and had fewer excess tablets at their postoperative visit.

      Results

      This work affirms what has been seen in previous work in urogynecology and gynecology: over-prescribing is common after surgery. Additionally, our finding that prescribing more opioid tablets leads to increased use is consistent with previous work in urogynecology
      • Swenson C.W.
      • Kelley A.S.
      • Fenner D.E.
      • Berger M.B.
      Outpatient narcotic use after minimally invasive urogynecologic surgery.
      and in urology. Bates et al
      • Bates C.
      • Laciak R.
      • Southwick A.
      • Bishoff J.
      Overprescription of postoperative narcotics: a look at postoperative pain medication delivery, consumption and disposal in urological practice.
      found that two-thirds of patients reported excess opioid therapy with 40% of all prescribed opioid tablets left unused; in addition, most patients did not know how to safely dispose of unused tablets and kept the excess at home. In our study, we estimate that >1500 tablets of oxycodone 5 mg went unused in the routine arm compared with 240 tablets in the reduced arm.

      Clinical implications

      Given that most hospitals do not have an easy way for patients to dispose of unused medication, these excess tablets are not only wasteful but also dangerous in their potential for abuse or diversion. Nationwide reduction in postoperative opioid prescribing of only 10% has been predicted to lead to 300 million fewer tablets available for misuse, 300,000 fewer people per year becoming persistent opioid users, and a cost-savings of >800 million dollars per year in drug costs alone.
      • Compton P.
      • Mena G.
      • Sethi P.
      • Sigman S.
      United States for non-dependence: an analysis of the impact of opioid overprescribing in America.
      Based on our own findings, we estimate that, for the anticipated 300,000 women who will undergo prolapse surgery per year,
      Urogynecologic surgical mesh: update on the safety and effectiveness of transvaginal placement for pelvic organ prolapse.
      a change in prescribing from 28 to 5 tablets of oxycodone would result in over 6 million fewer excess opioid tablets prescribed. This is especially relevant to the urogynecology patient population, because female sex and increased age are both risk factors for persistent opioid use.
      • Compton P.
      • Mena G.
      • Sethi P.
      • Sigman S.
      United States for non-dependence: an analysis of the impact of opioid overprescribing in America.

      Research implications

      Future research in this field includes studying the effect of patient-guided prescribing that includes variable quantities that include no opioids prescribed at discharge. In the obstetrics literature, researchers have found success in individualized, patient-guided opioid prescribing with no impact on patient satisfaction and decreased opioid prescribed.
      • Osmundson S.S.
      • Raymond B.L.
      • Kook B.T.
      • et al.
      Individualized compared with standard postdischarge oxycodone prescribing after cesarean birth: a randomized controlled trial.
      Previous work in urogynecology has shown that inpatient use of opioid predicts postdischarge opioid use,
      • Ramaseshan A.S.
      • Tunitsky-Bitton E.
      • O’Sullivan D.M.
      • Reagan K.M.L.
      • Steinberg A.C.
      Predictive factors of postdischarge narcotic use after female pelvic reconstructive surgery.
      which was also seen in our work. Implementation of a prescription protocol based on inpatient use should be studied prospectively. The findings regarding potentially improved return to baseline and decreased pain scores in the reduced opioid group should also be studied prospectively; perhaps there is a psychologic effect of prescribing more opioid tablets, increasing anticipation of pain, or “nocebo” effect. In addition, hospital systems should study the best ways to encourage and support disposal of unused opioid medications. At our large hospital system during the study period, only 3 hospital sites had medication disposal areas. Discharge instructions routinely given to all our patients include safe disposal options that include finding drug take-back locations, flushing, and techniques to safely place in the garbage as recommended by the Food and Drug Administration.

      Strengths and limitations

      The primary strength of this study is in its randomized controlled design. Although it is not the first trial executed with a goal to minimize opioid use after postoperative discharge, we believe it is the first to assess patient satisfaction with pain control after discharge when quantities are reduced. Additionally, the variety of surgeries included in the study increase this work’s generalizability to a real urogynecology practice. Finally, these results may contribute to evidence-based guidelines for postoperative opioid prescribing for patients who undergo urogynecology surgery. The results may further be extrapolated to recommendations for general gynecology and surgery practices.
      This study does have limitations. One important limitation is that masking was not possible because the opioid quantity was the intervention. Because the primary outcome was subjective, this could lead to bias in either direction; patients with fewer tablets could conceivably expect less pain and those with more tablets could expect more pain (“nocebo effect”) or vice versa. We also did not include intraoperative opioid use within the inpatient opioid use calculations because of variability in practices by anesthesiologist and hospital location; this limits our ability to draw firm conclusions of the effect on total perioperative opioid use.
      In addition, there were multiple protocol deviations, most of which involved changing the postoperative opioid plan before discharge. However, these data can still guide postoperative prescribing with some modifications for oxycodone-intolerant patients or those who cannot use the multimodal nonopioid therapies. For example, surgeons could consider a similarly reduced number of tramadol tablets for oxycodone-intolerant patients. For patients with greater inpatient opioid requirements, which we did find to be predictive of postdischarge use, surgeons may increase the number of tablets prescribed.
      Another limitation to this work is that data regarding medication use were patient-reported and that not all patients completed all 42 days of the 6-week patient diaries, which limited the precision of our estimates of medication use and pain scores. However, based on clinical logic and the completed diaries, it is likely that the majority of pain medication was used in the first weeks of recovery.
      Finally, we did not include same-day discharge surgeries. However, our findings could still be applied to same-day discharge patients with consideration of prescribing a limited number of additional tablets considering the overnight opioid needs.

      Conclusions

      It is imperative that surgeons adapt to the current climate of the opioid epidemic. We must act as better stewards of these potentially dangerous prescriptions by acknowledging the addictive potential of these medications. The results of this study provide some guidance for postoperative opioid prescribing. Additionally, these data should reassure surgeons that minimizing opioid prescriptions will not impact patient satisfaction with pain control. We recommend that surgeons change their routine postoperative prescribing practice to <10 opioid tablets in addition to multimodal adjunctive therapy in an effort to balance a patient’s pain control with wasteful and dangerous excess prescribing.

      Appendix

      Figure thumbnail fx1
      Appendix 1Daily diary (example)
      Davidson et al. Reducing postoperative opioids. Am J Obstet Gynecol 2020.
      Figure thumbnail fx2
      Appendix 1Daily diary (example)
      Davidson et al. Reducing postoperative opioids. Am J Obstet Gynecol 2020.
      Figure thumbnail fx3
      Appendix 2Postoperative survey
      Davidson et al. Reducing postoperative opioids. Am J Obstet Gynecol 2020.

      References

      1. The opioid epidemic: By the numbers.
        (Available at:) (Updated June 2016. Accessed June 2, 2019)
        • Overton H.N.
        • Hanna M.N.
        • Bruhn W.E.
        • et al.
        Opioid-prescribing guidelines for common surgical procedures: an expert panel consensus.
        J Am Coll Surg. 2018; 227: 411-418
        • As-Sanie S.
        • Till S.R.
        • Mowers E.L.
        • et al.
        Opioid prescribing patterns, patient use, and postoperative pain after hysterectomy for benign indications.
        Obstet Gynecol. 2017; 130: 1261-1268
        • Griffith K.C.
        • Clark N.V.
        • Zuckerman A.L.
        • Ferzandi T.R.
        • Wright K.N.
        Opioid prescription and patient use after gynecologic procedures: a survey of patients and providers.
        J Minim Invasive Gynecol. 2018; 25: 684-688
        • Hota L.S.
        • Warda H.A.
        • Haviland M.J.
        • Searle F.M.
        • Hacker M.R.
        Opioid use following gynecologic and pelvic reconstructive surgery.
        Int Urogynecol J. 2018; 29: 1441-1445
        • Swenson C.W.
        • Kelley A.S.
        • Fenner D.E.
        • Berger M.B.
        Outpatient narcotic use after minimally invasive urogynecologic surgery.
        Female Pelvic Med Reconstr Surg. 2016; 22: 377-381
        • Brummett C.M.
        • Waljee J.F.
        • Goesling J.
        • et al.
        New persistent opioid use after minor and major surgical procedures in US adults.
        JAMA Surg. 2017; 152e170504
        • Reagan K.M.L.
        • O’Sullivan D.M.
        • Gannon R.
        • Steinberg A.C.
        Decreasing postoperative narcotics in reconstructive pelvic surgery: a randomized controlled trial.
        Am J Obstet Gynecol. 2017; 217: 325.e1-325.e10
        • Sullivan M.
        • Biship S.
        • Pivik J.
        The pain catastrophizing scale: development and validation.
        Psychol Assess. 1995; 7: 524-532
        • Jacobsen P.B.
        • Butler R.W.
        Relation of cognitive coping and catastrophizing to acute pain and analgesic use following breast cancer surgery.
        J Behav Med. 1996; 19: 17-29
        • Evans C.J.
        • Trudeau E.
        • Mertzanis P.
        • et al.
        Development and validation of the Pain Treatment Satisfaction Scale (PTSS): a patient satisfaction questionnaire for use in patients with chronic or acute pain.
        Pain. 2004; 112: 254-266
        • Crisp C.C.
        • Khan M.
        • Lambers D.L.
        • et al.
        The effect of intravenous acetaminophen on postoperative pain and narcotic consumption after vaginal reconstructive surgery: a double-blind randomized placebo-controlled trial.
        Female Pelvic Med Reconstr Surg. 2017; 23: 80-85
        • Bates C.
        • Laciak R.
        • Southwick A.
        • Bishoff J.
        Overprescription of postoperative narcotics: a look at postoperative pain medication delivery, consumption and disposal in urological practice.
        J Urol. 2011; 185: 551-555
        • Compton P.
        • Mena G.
        • Sethi P.
        • Sigman S.
        United States for non-dependence: an analysis of the impact of opioid overprescribing in America.
        (Available at:) (Published 2017. Updated September 26, 2017. Accessed June 2, 2019)
      2. Urogynecologic surgical mesh: update on the safety and effectiveness of transvaginal placement for pelvic organ prolapse.
        (Available at:) (Published 2011. Accessed October 26, 2019)
        • Osmundson S.S.
        • Raymond B.L.
        • Kook B.T.
        • et al.
        Individualized compared with standard postdischarge oxycodone prescribing after cesarean birth: a randomized controlled trial.
        Obstet Gynecol. 2018; 132: 624-630
        • Ramaseshan A.S.
        • Tunitsky-Bitton E.
        • O’Sullivan D.M.
        • Reagan K.M.L.
        • Steinberg A.C.
        Predictive factors of postdischarge narcotic use after female pelvic reconstructive surgery.
        Female Pelvic Med Reconstr Surg. 2019; 25: e18-e22
        • Food and Drug Administration
        Safe disposal of medicines.
        (Available at:) (Accessed)

      Linked Article