Association Between Opioid Use Disorder and Hysterectomy Outcomes

This is a retrospective cohort study using the 2016 to 2021 National Inpatient Sample (NIS) and the Nationwide Ambulatory Surgery Sample (NASS) databases comparing patients who underwent hysterectomy with opioid use disorder (OUD) versus those without OUD. The NIS and NASS databases are publicly available, nationwide databases created by the Healthcare Cost and Utilization Project (HCUP). The National Inpatient Sample is the largest publicly available inpatient hospital database. It contains data for 7 million hospital stays annually unweighted and estimates over 33 million hospitalizations nationally. The Nationwide Ambulatory Surgery Sample is a related database of major ambulatory, outpatient surgery encounters within hospital-owned facilities which includes data on clinical diagnostic and surgical procedure codes, patient disposition and costs. Both databases capture information during the evaluated encounter and do not include information after discharge. These datasets are analyzed separately in this study, as these represent distinct patient populations. The NIS and NASS databases are universally de-identified; the study was determined to be exempt by the Duke University Health System Institutional Review Board (Proo00106911). All patients 18 years and older who underwent hysterectomy between 2016 and 2021 were included. CPT and ICD-10 codes were used to identify patients undergoing all routes of hysterectomy including open, minimally invasive, and vaginal ( Appendix 1 , available online at http://links.lww.com/xxx). Both benign and malignant indications were included. CPT and ICD-10 codes were then utilized to identify patients with opioid use disorder within the cohort. Patients were separated into two groups, those with and without associated diagnosis of opioid use. Patients who were admitted from the emergency room and patients who did not have data available for age, race, and primary expected payer were excluded. The primary outcome was composite postoperative complications prior to discharge. Complications included acute renal failure (ARF), sepsis, deep vein thrombosis (DVT), pulmonary thromboembolism (PE), stroke, pulmonary complications, nausea and vomiting, blood transfusion, postoperative bleeding, mechanical ventilation, and in-hospital mortality. Secondary outcomes were length of stay and hospital charges. Hospital charges were adjusted to 2024 US dollars using the Consumer Price Index. 12 , 13 Demographic characteristics include age, primary expected payer, and median household income based on zip code. Baseline clinical characteristics include smoking status, alcohol use disorder, asthma, congenital heart disease, coronary artery disease, chronic kidney disease, chronic obstructive pulmonary disease (COPD), dyslipidemia, heart disease (ischemic), heart failure, HIV, hypertension, hypothyroidism, nonalcoholic fatty liver disease (NAFLD), obstructive sleep apnea, type 2 diabetes, chronic pain. Indications for surgery included uterine cancer, ovarian cancer, cervix cancer, leiomyoma, abnormal uterine bleeding, benign ovarian neoplasm, endometriosis, pelvic organ prolapse, dysplasia, dysmenorrhea and premenstrual syndrome, dyspareunia. Hospital characteristics including bed size, location and teaching status were included. Patient characteristics and outcomes were summarized as means (standard error of the mean, SEM), medians (IQR), and minimum and maximum for continuous variables and counts (percentages) for categorical variables by patients with and without OUD. A modified Poisson regression model was performed with robust standard errors to assess the association between OUD and composite complications for both inpatient and outpatient cohorts. Results from unadjusted and adjusted models were reported as rate ratios (RR) and 95% confidence intervals (CI). A negative binomial model was performed to assess the association between OUD and length of stay for the inpatient cohort. Results from unadjusted and adjusted models were reported as incidence risk ratios (IRR) and 95% CIs. A generalized linear model with a gamma distribution and log link was performed to assess the difference in hospital charges (adjusted to 2024 US dollars) for both cohorts. IRRs and 95% CIs were reported for both unadjusted and adjusted models. Covariates in all adjusted models included patient age, patient race, geographic region, malignancy, route of surgery, insurance type, year of surgery, and a composite comorbidity including coronary artery disease, chronic kidney disease, COPD, hypertension, smoking, diabetes, heart failure, and hypothyroidism. Patient race was included as a proxy for social and structural factors that might confound the relationship between OUD and surgical complications. All models incorporated discharged weights, hospital clusters, and hospital stratification provided by NIS and NASS. All analyses were performed in SAS v9.4 (SAS Institute Inc, Cary, NC) and R 4.1.0 (R Core Team, 2024). A p-value < 0.05 was considered statistically significant.

A total of 1,695,075 patients underwent outpatient hysterectomy which corresponded to a national estimate of 2,292,869 patients ( Figure 1A ). All subsequent results are presented as weighted national estimates. In the overall cohort, 21,632 (0.9%) patients had prevalent OUD. The mean age of patients was 48.5 years (OUD 46.4, No OUD 48.5). The majority were in the 18–44 years (42.2%; OUD 49.1%, No OUD 42.2%) age groups. The largest subset of patients was in the second lowest median household income national quartile for patient ZIP code group (27.1%; OUD 30.0%, No OUD 27.1%), and 68.6% underwent surgery in urban teaching hospitals (OUD 67.6%, No OUD 68.6%). Most patients had private insurance as their primary expected payer (68.8%; OUD 61.9%, No OUD 68.8%). Most patients underwent a minimally invasive hysterectomy (98.6%, compared to 1.4% abdominal hysterectomies), and most patients received a hysterectomy for benign etiologies (90.7%, compared to 9.3% for malignant indications). A higher percentage of OUD patients had alcohol use disorder (0.3% versus 0.1%), asthma (12.8% versus 9.7%), coronary artery disease (1.6% versus 1.5%), chronic kidney disease (1.4% versus 1.2%), COPD (3.3% versus 1.5%), hypertension (27.4% versus 26.9%), obstructive sleep apnea (6.0% versus 4.4%), and tobacco use (20.4% versus 9.9%) ( Table 1 ). The mean total charges were $50,276 (SEM $374.4), with lower charges for patients with OUD compared to patients without OUD (OUD $48,027, No OUD $50,297). While most patients did not have a composite complication (96.6%), a larger percentage of patients with OUD experienced a complication compared to patients without OUD (3.8% versus 3.4%, respectively). Specifically, a larger percentage of patients with OUD had nausea and vomiting (1.1% vs 0.8%) and DVT (0.1% vs 0%) compared to those without OUD ( Table 2 ). Patients with OUD had similar rates of routine disposition compared to patients without OUD (99.4% versus 99.6%, respectively) ( Table 2 ). In both adjusted and unadjusted regression models, OUD was associated with an increased risk of composite complications (Unadjusted RR 1.15, 95% CI 1.02, 1.30 p=0.026; Adjusted RR 1.13, 95% CI 1.00, 1.28, p=0.048). There was no association between OUD and increased charges (Unadjusted IRR 0.96, 95% CI 0.91, 1.02, p=0.215; Adjusted IRR 0.96, 95% CI 0.91, 1.01, p=0.099) ( Table 3 ). A total of 188,178 patients underwent inpatient hysterectomy corresponding to a weighted national estimate of 934,789 patients ( Figure 1B ). In this cohort, 12,085 (1.3%, weighted) patients had prevalent OUD. The mean age of patients was 50.6 years (OUD 48.8, No OUD 50.6). The largest subset of patients was in the 18–44 years (35.3%; OUD 41.1%, No OUD 35.3%). The largest subset of patients (29.0%; OUD 31.3%, No OUD 28.9%) was in the lowest median household income national quartile for patient ZIP code group, and most hospitals were urban teaching (74.7%; OUD 76.5%, No OUD 75.6%). Most patients (56.1%; OUD 40.5%, no OUD 56.4%) had private insurance as their primary expected payer, followed by Medicaid (18.6%; OUD 23.6%, No OUD 18.5%). Most patients underwent abdominal hysterectomy (73.4%; OUD 73.0%, No OUD 73.4%). Most patients received a hysterectomy for benign etiologies (81.8%), compared to malignant indications (18.2%). A higher percentage of OUD patients had alcohol use disorder (0.8% versus 0.2%), asthma (13.6% versus 8.6%), coronary artery disease (3.8% versus 2.8%), chronic kidney disease (3.8 versus 2.6%), COPD (7.1% versus 2.6%), hypertension (40.6% versus 34.4%), obstructive sleep apnea (7.4% versus 4.7%), tobacco use (26.1% versus 9.5%), Type 2 Diabetes (14.9% versus 12.4%), and chronic pain (12.7% versus 7.2%) compared to patients without OUD ( Table 1 ). The mean length of stay for patients was 3.2 days, with longer length of stay for patients with OUD (OUD 4.3, No OUD 3.2). The mean total charges were $72,058 (SEM 513.8), with patients with OUD demonstrating higher charges compared to patients without OUD (OUD $91,258.3, No OUD $71,809.1). While most patients did not have a composite complication (88.2%), a larger percentage of patients with OUD experienced a complication compared to patients without OUD (14.7% versus 11.8%, respectively). Specifically, a larger percentage of patients with OUD had acute renal failure (4.7% vs 3.3%), nausea and vomiting (4.6% vs 2.8%), and sepsis (1.7% vs 1.0%) compared to those without OUD. A larger percentage of patients without OUD required a blood transfusion (5% vs 4.3%) compared to those with OUD ( Table 2 ). Fewer patients with OUD had routine disposition compared to patients without OUD (86.9% versus 92.2%, respectively) ( Table 2 ). In both adjusted and unadjusted regression models, OUD was associated with an increased risk of composite complications (Unadjusted RR 1.24, 95% CI 1.12, 1.36, p<0.001; Adjusted RR 1.18, 95% CI 1.07, 1.31, p<0.001). OUD was also associated with a longer length of stay (Unadjusted IRR 1.38, 95% CI 1.29, 1.47, p<0.001; Adjusted IRR 1.32, 95% CI 1.24, 1.41, p<0.001) and increased charges (Unadjusted IRR 1.27, 95% CI 1.20, 1.34, p<0.001; Adjusted IRR 1.22, 95% CI 1.16, 1.29, p<0.001) ( Table 3 ).

Within this large cohort of patients who underwent hysterectomies from the NASS and NIS nationwide databases, patients with OUD had higher complication rates compared to patients without OUD. Furthermore, patients with OUD who underwent inpatient hysterectomy experienced longer lengths of stay and had increased hospital charges, compared to patients without OUD. These findings are consistent with findings in other medical and surgical fields. OUD has been associated with increased rates of complication and increased hospital charges after lumbar spine surgery, bariatric surgery, and urologic oncology surgery, among others. 6 , 14 , 15 A large retrospective cohort study found higher resource utilization and readmission rates were noted among trauma patients with OUD, as well as a trend toward higher prevalence of OUD year over year 11 . In obstetrics, a large multistate analysis found that OUD predicts higher inpatient mortality and increased readmission rates in the post-operative period in patients undergoing cesarean delivery 12 . Among the above studies, there is a common acknowledgement of the paucity of data available evaluating the role of optimization of OUD as it relates to improving these outcomes. OUD is a heterogenous and complex disorder with mental health, environmental, and societal factors that can affect a patient’s overall health and post-operative recovery 8 , 9 . This study found an increased risk of composite complications for both cohorts and found increased total charges and length of stay for the NIS inpatient cohort, even when adjusting for factors such as age, race, geography, and insurance type. However, it is notable that a larger proportion of patients with OUD had Medicaid in both cohorts compared to those without OUD (NASS 19.2% vs 12.2%; NIS 29.5% vs 18.5%). Furthermore, a larger proportion of patients with OUD were in the lower half of household income for a patient’s zip code (NASS 55.2% vs 50%; NIS 59.9% vs 54%). Interestingly, a larger percentage of patients without OUD in the inpatient cohort underwent surgery for leiomyomas and abnormal uterine bleeding compared to patients with OUD, which may clinically explain the larger percentage of patients who required a blood transfusion (5% vs 4.3%). Outcomes should be interpreted and interventions should be evaluated within the context of a broader substance use disorder and with consideration of socioeconomic issues, healthcare access, payer-related differences, and care fragmentation. There have been multiple programs for prehabilitation and peri-operative optimization for improving patients’ cardiopulmonary reserve, smoking cessation, nutrition, physical activity, and comorbidities prior to elective surgery that have demonstrated reduction in complications and decreased hospital charges for high risk surgical patients, such as the Perioperative Optimization of Senior Health (POSH) Initiative or the Michigan Surgical and Health Optimization Program (MSHOP). 16 , 17 Patients with OUD may benefit from similar programs for optimizing their health prior to elective surgery. This study has many strengths. The opioid use epidemic is a known public health crisis, and this study evaluates the association between OUD on outcomes after hysterectomy, which is one of the most performed procedures on females in the United States. Second, the NIS and NASS databases are large, nationwide databases inclusive of all payer types, which makes this study generalizable due to the diverse and large sample size. By using both databases, we were able to capture patients who underwent surgery in both an inpatient and ambulatory surgery setting. Third, we were able to adjust for several other characteristics known to be associated with increased complication rates. There are also a number of limitations. First, this study does not capture complications that occur outside of the index surgical admission, such as vaginal cuff dehiscence, post-operative infection, or deep vein thrombosis which may occur days to weeks post-operatively. These outcome data are necessary to gain a full picture of post-operative complication risks. Second, a relatively small proportion of patients had OUD. However, the NIS OUD hysterectomy cohort still included 12,085 patients, and the NASS OUD hysterectomy cohort included 21,632 patients. These cohorts are still larger than previous studies on this issue. For example, only 1,096 of 376,467 patients who underwent major upper abdominal surgery had OUD and only 2,205 of 159,455 patients who underwent bariatric surgery had OUD. 6 , 18 Third, the database does not consider pre-operative and post-operative optimization measures to reduce morbidity. Fourth, while we captured both inpatient and ambulatory surgeries, this study has the limitations inherent to retrospective administrative databases, such as missing and/or inconsistent data and misclassification. This is especially relevant to this patient population due to the stigma, sensitive nature, and legal implications of OUD. Therefore, there is a risk of under-coding and inconsistent capture of OUD across hospitals, payer types, and care settings. There may also be additional unmeasured confounders. This study found that patients with OUD are at increased risk of complications, longer length of stay, and increased hospital charges after hysterectomy. Understanding the relationship between opioid use disorder and risks of post-operative complication could help surgeons identify who may benefit from additional pre-operative optimization and education to improve post-operative outcomes and decrease post-operative healthcare expenses.

Opioid use disorder (OUD) in the United States is a major public health concern and continues to increase. 1 In 2022 alone, 4 million female patients were in need of OUD treatment. 2 With this growth, physicians, including gynecologic surgeons, must understand the challenges and consequences of the disease on peri-operative management and outcomes. There is significant evidence that OUD is associated with increased complication risks, length of stay, readmission rates, and costs within the surgical literature, from cardiac to major abdominal surgery 3 – 5 . For example, in a National Inpatient Sample (NIS) study of 159,455 patients who underwent bariatric operations from 2016–2020, OUD was associated with an increased risk for complications, such as bleeding and postoperative nausea and vomiting. OUD was also associated with longer length of stay (LOS) and increased charges. 6 While opioids may increase complication risk due to side effects, including decreased bowel motility and urinary retention, OUD has also been associated with a number of other pre-operative risk factors, such as mental health disorders, socioeconomic status, and risk of fragmented healthcare. 5 , 7 – 9 Hysterectomy is one of the most performed surgeries for women in the United States, with approximately 500,000–600,000 surgeries performed every year. 10 , 11 Yet, there have been limited studies assessing if patients with OUD undergoing hysterectomy are at higher risk of complications than patients without OUD. Further evaluation of the association of opioid use disorder on outcomes after hysterectomy could inform both peri-operative optimization strategies and larger policy decisions. The aim of this study was to evaluate the association between opioid use disorder (OUD) and complications, length of stay, and costs for patients undergoing a hysterectomy.