Predictors of new persistent opioid use after benign hysterectomy in the United States.

This study shows that younger age, non-Northeastern location, preoperative psychiatric and pain diagnoses, psychiatric medication use, surgical complications, and >1 perioperative opioid prescription are associated with higher odds of NPOU. On the other hand, more recent surgery year, minimally invasive hysterectomy, and a first perioperative prescription after the procedure were associated with lower NPOU likelihood. Our NPOU estimate after hysterectomy (4.6%) is comparable to what has been reported by previous gynecologic (2.5–6.8%) 6 , 8 , 9 , 14 , 17 and non-gynecologic (2.8–6.5%) studies. 6 , 9 While Bicket, Brummett, and Young defined NPOU as an opioid prescription fulfillment between 90 and 180 days, 8 , 9 , 11 our definition required an initial prescription from 15–90 days and another prescription from 91–365 days, postoperatively. Thus, our definition may more successfully exclude confounding new prescribing episodes unrelated to the surgery of interest. Additionally, a systematic review showed that having at least one opioid fill from 90–365 days postoperatively significantly increases definition sensitivity. 18 Our results suggest that baseline characteristics represent predictors of NPOU in women undergoing hysterectomy. Our finding of an association between NPOU and younger age is consistent with other studies on hysterectomy and non-gynecologic procedures. 17 , 19 Likewise, Swenson et al. identified preoperative depression and anxiety and minimally invasive hysterectomy approaches as significant predictors of NPOU. 14 While the mechanistic aspects of these associations are poorly understood, they may be partially explained by differential levels of postsurgical pain. For instance, abdominal hysterectomy is associated with significant postsurgical pain compared with laparoscopic and vaginal approaches whereas older age may decrease peripheral nociceptive function, 20 , 21 possibly dampening pain perception and consequently affecting opioid use. A preoperative pain disorder may alternatively sensitize peripheral nociceptors and central nervous system neurons, leading to persistent postsurgical pain. 20 , 22 Psychological factors, such as preoperative anxiety and depression, were shown to predict persistent postsurgical pain, 23 which is concerning given that adults with mental health disorders are more likely to continue opioid use after initial prescription regardless of surgical history and across all pain levels. 24 Several studies and systematic reviews explored opioid prescription practices in hysterectomy patients and showed that women are being overprescribed opioids for postoperative pain. 10 , 25 For instance, As-Sanie et al. found that patients sought to use only about half of prescribed pills, with a median excess of 110 MMEs. 10 Such practices span across several gynecologic avenues, including urogynecologic procedures. 26 To prescribe appropriate, individualized amounts of opioids, Wong et al. proposed a Postoperative Opioid Calculator for Hysterectomy (POOCH) to predict low and high opioid users based on 9 preoperative components. 25 A major and potentially practice-changing finding in our study is that a 1 st perioperative prescription given preoperatively, as opposed to postoperatively, dramatically increases NPOU odds by >60%. If opioids are necessary, physicians may therefore consider not dispensing them until after the procedure. Based on our findings, opioids should be carefully prescribed for all patients, especially younger patients with psychiatric or pain disorders and those operated abdominally. Several modalities were established to optimize pain control and minimize opioid use. Enhanced Recovery After Surgery (ERAS) protocols involve pre-, intra-, and postoperative models of care that facilitate postsurgical recovery without compromising patient outcomes. 27 Outpatient and ERAS hysterectomy protocols integrating multimodal analgesic approaches were deemed safe and associated with lower opioid requirements and significant cost savings. 27 , 28 Ultrasound-guided transmuscular quadratus lumborum block was identified as a novel regional anesthesia technique providing effective postoperative analgesia while reducing opioid consumption after laparoscopic hysterectomy. 29 , 30 Transversus abdominis plane block has shown similar results postoperatively for abdominal hysterectomy. 31 Recent evidence suggests that perioperative insomnia is associated with increased postoperative pain and can therefore be an interventional target. 32 Collectively, these data should be utilized in shared-decision making about the need for perioperative opioids and counseling hysterectomy patients about the risk of persistent use postoperatively. Our findings present unique opportunities to investigate an enormous public health priority. Prospective research should focus on understanding the biological basis of transitioning from acute to chronic pain, possibly aiding in creating evidence-based measures for postoperative pain management, as suggested by the NIH Helping to End Addiction (HEAL) initiative. 33 To individualize opioid prescription, biomarker use and understanding the pharmacogenetic perspectives of opioids may serve as novel avenues for future work. 11 , 34 Lastly, as our study predates the ERAS era peak, studies evaluating prescription practices in ERAS-integrated settings will provide contemporary evidence on the long-term risks of postoperative opioid use in gynecologic populations. Limitations of this study include those inherent to retrospective designs. Administrative databases do not capture factors, such as prescriber preferences and non-narcotic over the counter analgesics, which might impact opioid use. Although patients not having prescribed opioids within one year prior to hysterectomy were termed opioid-naïve, it is difficult to assume that they have not been prescribed opioids prior to that. In such case, their NPOU after hysterectomy may in fact represent a recurrence of past use. It was not possible to guarantee that the prescriptions included were for perioperative indications; however, excluding patients with additional procedures after hysterectomy and using two temporally successive postoperative prescriptions to identify NPOU attempted to minimize inadvertent misclassification of unrelated opioid dispensing. Our cohort is comprised of privately insured women, which may not be representative of publicly insured and uninsured women or experiences outside of the US. Strengths of the paper include its large sample size and inclusive number of demographic, clinical, and opioid prescription characteristics, providing a more comprehensive picture of the possible drivers of NPOU. Lastly, we sought to use a conservative definition of NPOU after hysterectomy to avoid erroneous classification of opioid use and accurately differentiate new persistent opioid users from those with use unrelated to their postoperative course. In our study, NPOU occurred in 4.6% of women undergoing benign hysterectomy, an observation that echoes the national concern about problematic opioid use. NPOU is associated with many patient demographics and clinical factors, operative parameters, and perioperative opioid prescription characteristics. Preoperative planning and evidence-based understanding of patient risk profiles may significantly decrease NPOU while optimizing postoperative analgesia and recovery. Judicious perioperative opioid prescription and increased physician awareness of the detrimental effects of overprescription practices may partially ameliorate rates of chronic opioid use.

From 2011–2016, 114,260 women met our surgical criteria and did not have a recent opioid prescription dispensed. Of these, 93,906 women (82.2%) received a perioperative opioid. Patients with and without a perioperative prescription are described in Table 1 . The median total dose of the first perioperative prescription was 225 MME, with the majority (79.8%) receiving it within 14 postoperative days. Only 1.4% of women received >1 perioperative prescription, with a median prescription frequency of 1 in the total population. Of note, we observed a decreasing trend in the number of hysterectomies and persistent opioid users over the study period, which corresponds to a decline in patient enrollment in the database from 2011 to 2016 (53,012,885 to 27,895,445 records). The sociodemographic, operative, and prescription-related characteristics of persistent and non-persistent users are described in Table 2 . NPOU incidence in our cohort was 4.6%. Compared with non-users, women with NPOU were more likely to be <45 years of age (50.1% vs 44%; P <0.001), from the South (55.2% vs 48.1%; P <0.001) and have had their hysterectomy before 2013 (46.7% vs 40.5%; P <0.001). Persistent users were more likely to have psychiatric and pain disorders, mood disorder medication use, smoking history, and surgical complications compared with their non-user counterparts. While the median total dose of the first perioperative prescription (225 MME) and days’ supply (5 days) were the same between the two groups, persistent users were more likely to have received their first prescription before hysterectomy (2.5% vs 1.4%; P 1 prescription compared with non-users (28.8% vs 19.8%; P <0.001). Our regression model identified patient’s age, geographic region, surgery year, psychiatric comorbidities and medications, pain disorders, hysterectomy approach, surgical complications, and certain perioperative prescription characteristics as independent predictors of NPOU in opioid-naïve women undergoing hysterectomy ( Table 3 ; Figure 3 ). Younger women (18–34 years) had 97% higher odds of developing NPOU (95% CI 1.69–2.30; P <0.001) whereas those from the South had a 2-fold odd increase (95% CI 1.79–2.27; P <0.001) compared with older and Northeastern women, respectively. For more recent years beginning with 2013, there was a significant consistent lower NPOU likelihood (e.g., 2016 compared with 2011, aOR 0.58; 95% CI 0.51–0.65; P <0.001). History of anxiety (aOR 1.20; 95% CI 1.09–1.33; P <0.001), insomnia (aOR 1.34; 95% CI 1.18–1.52; P <0.001), and alcohol (aOR 1.70; 95% CI 1.17–2.47; P 0.005) and other substance use (aOR 1.82; 95% CI 1.21–2.75; P 0.004) significantly increased NPOU odds. Women taking mood disorder medications (aOR 1.51; 95% CI 1.40–1.64; P <0.001) and smokers (aOR 1.65; 95% CI 1.45–1.89; P <0.001) were also significantly more likely to continue opioid use. Women undergoing minimally invasive hysterectomy (laparoscopic or vaginal) had lower odds of NPOU compared with women operated abdominally (aOR 0.89; 95% CI 0.71–0.88; P <0.001; aOR 0.82; 95% CI 0.72–0.93; P 0.002, respectively). Although statistically nonsignificant, women prescribed ≥300 MME in their 1 st dispensing had a 9% higher odds of persistent use compared with women prescribed 1 perioperative prescription (aOR 1.53; 95% CI 1.24–1.88; P <0.001) and a 1 st prescription before the procedure (aOR 1.61; 95% CI 1.50–1.72; P <0.001) significantly predicted NPOU incidence. In addition, each additional perioperative day covered by an opioid significantly increased NPOU likelihood by 2% (95% CI 1.02–1.03; P <0.001). Lastly, women developing a surgical complication had 84% higher odds of continuing opioid use postoperatively (95% CI 1.69–2.00; P <0.001).

We utilized data from the IBM MarketScan Commercial Claims and Encounters Database, a validated national commercial insurance repository gathering healthcare information for patients enrolled in private health plans. It includes >263 million individuals as well as inpatient and outpatient services and pharmaceutical claims across the US. 12 This database has been previously used to study opioid use in gynecologic settings. 13 The study was exempt by the Johns Hopkins University Institutional Review Board as the database contains only de-identified data. In this retrospective cohort study, we included opioid-naïve women aged 18–64 years undergoing benign hysterectomy from January 2011 to December 2016 and having at least one perioperative opioid prescription ( Figure 1 ). We used the Current Procedural Terminology (CPT) codes to identify hysterectomies ( Appendix 1 ) and the index date was the date of hysterectomy. To verify opioid use preoperatively and NPOU postoperatively, patients must have prescription coverage and be continuously enrolled in the database for at least 1 year both prior to and after the index date. We defined perioperative opioid prescription as at least one prescription between 30 days before and 14 days after hysterectomy. We excluded patients with one or more opioid prescriptions from 365-31 days preoperatively (non-naïve) to better understand the independent association of perioperative prescription on NPOU. We excluded patients who underwent additional procedures or received anesthesia within one year after the index date to avoid erroneous overestimation of NPOU due to unrelated opioid prescriptions. Lastly, we excluded patients with radical hysterectomy, gynecologic malignancy within one year before or after hysterectomy, a delivery within 12 weeks prior to hysterectomy, and those with >1 procedure type on the index date. As opposed to other studies, 14 we sought to examine NPOU in women with exclusively benign pathologies to better delineate patterns of opioid use without the unique, complex implications of a cancer diagnosis. Opioid use was determined using pharmacy dispensing codes in the MarketScan database. To identify opioid prescriptions, therapeutic classes (THERCLS) and subclasses of generic drug identifiers (GENERID) for opioids were used ( Appendix 1 ). METQTY is the number of pills dispensed per prescription, which alongside the opioid type and number of days’ supply, was identified using the National Drug Code (NDC) in prescription claims data. For each prescription, Morphine Milligram Equivalent (MME) was calculated by multiplying METQTY by the strength/pill by the MME conversion factor provided by the CDC. Invalid opioid pharmacy claims (days’ supply of ≤0 or >365 or pill quantity of ≤0) were excluded as were outliers (pill quantity ≥99 th percentile). 15 For women with multiple perioperative opioid prescriptions, only the first prescription was included in the analysis as a potential predictor of persistent opioid use postoperatively. A priori , our primary outcome was NPOU after hysterectomy ( Figure 2 ), which we defined as the fill of at least one eligible opioid between postoperative day 15–90 and postoperative day 91–365, in addition to at least one perioperative prescription between 30 days before to 14 days after hysterectomy. This definition captures patients filling additional prescriptions beyond the resolution of postsurgical pain based on commonly accepted time periods that align with the typical resolution of postsurgical pain. 16 To evaluate NPOU-associated factors, we included sociodemographic, preoperative clinical, operative, and opioid prescription-related factors as covariates. Sociodemographic factors included patient’s age based on categories from the Marketscan database (18–34, 35–44, 45–54, or 55–64), procedure year (2011–2016), geographic region, union status, wage type, and admission status. Clinical factors included preoperative psychiatric and pain diagnoses, mood disorder medication use, tobacco use, and Charlson Comorbidity Index whereas operative factors included hysterectomy approach, surgical indication, length of stay, and surgical complications. Patient demographic and clinical characteristics and the MarketScan, Current Procedure Terminology (CPT), International classification of diseases (ICD), and Clinical Classification system (CCS) diagnosis codes by which they were obtained are in Appendix 1 . We included 4 variables to characterize perioperative opioid use: (1) total MME of the 1 st prescription; (2) perioperative days’ supply (number of days covered by an opioid prescription); (3) having dispensed >1 prescription; and (4) 1 st prescription timing (pre- or postoperatively), with the former two as continuous variables and the latter two as binary variables. We analyzed the data using SAS version 9.4 (SAS Institute, Cary, NC). We conducted bivariate analyses for baseline (sociodemographic and preoperative clinical), operative, and prescription-related covariates for women with and without a perioperative prescription and NPOU using chi-squared and Mann-Whitney U tests for categorical and continuous covariates, respectively, and unadjusted odds ratios with 95% CIs for developing NPOU were calculated. We used multivariate logistic regression analysis to estimate adjusted odds ratios (aOR) and 95% CIs for developing NPOU and identify its independent predictors. Variables found to have statistical significance ( P 10 and variables with known associations with NPOU were included in the multivariate analysis. We used the median values of continuous variables as dichotomous cutpoints, except for the total MME of the 1 st perioperative prescription, wherein the 75 th percentile served that purpose and days’ supply, which was analyzed as a continuous variable. 11 We assigned the groups below the cutoff values as the referent. Statistical significance was set at P <0.01 with 2-sided tests for all analyses. We used a P value cutoff of 0.01 rather than the usual 0.05 to decrease the possibility of type 1 error due to multiple comparisons and large sample size.

In the United States (US), inappropriate prescription opioid use has reached epidemic proportions, with greater than 11.5 million Americans reporting prescription opioid misuse in 2016. 1 According to the Centers for Disease Control and Prevention (CDC), opioid overdoses claimed the lives of 450,000 people over two decades from 1999–2018, with greater than 50% of deaths involving a prescription opioid. 2 Inappropriate prescribing practices served as a primary driver in the initial years of the epidemic, and while more recent increases in overdoses have been due to the use of non-prescription opioids, prescribing still serves as the initial exposure for many persons who transition to illicit use. 3 Although the overall opioid prescription rate in the US has been declining since 2012, 2 opioids continue to be overprescribed in many surgical settings. 4 , 5 Concerningly, up to 10% of patients receiving postoperative opioids may become long-term users, including those naïve to opioids preoperatively. 6 Hysterectomy is the most common major gynecologic surgery in the US and can be associated with significant postsurgical pain, 7 increasing the risk of new persistent opioid use (NPOU). NPOU incidence estimates after hysterectomy have been reported to range from 1.4 to 6.7%. 6 , 8 , 9 NPOU risk may differ based on operative approach; for example, Clarke et al. 6 found persistent use estimates of 1.5% and 2.5% after minimally invasive and abdominal hysterectomy, respectively. Other studies examined predictors of perioperative opioid prescriptions. 8 , 10 While this adds to our understanding of opioid prescription patterns around hysterectomy time, independent predictors of persistent opioid use remain largely unexplored. Furthermore, studies that have indeed assessed NPOU after hysterectomy remain scarce and omit some clinical factors, including surgical indication, mental illness, and opioid prescription characteristics, 6 , 8 although some were shown to predispose to NPOU in other surgical subspecialties. 11 Equipping the gynecologist with the knowledge of risk factors for persistent opioid use could help inform perioperative prescription practices and assist national efforts to reduce opioid-related morbidity and mortality. Thus, we aim in this population-based study to identify the incidence and predictors of NPOU after benign hysterectomy, with emphasis on previously unmeasured clinical and prescription-related factors, including surgical complications, duration and frequency of perioperative opioid supply, and a wider range of mental disorders.