1 Introduction

Prescription opioid consumption in the United States is a well-known public health problem. From 1999 to 2010, prescription opioid misuse and abuse roughly quadrupled, and the number of opioid-related deaths quintupled.¹ This startling increase coincided with a broadening of opioid prescribing indications from acute to chronic pain.² Notably, orthopaedic surgeons have played a major role in opioid prescribing, ranking as the fourth highest opioid prescribers behind primary care physicians, internists and dentists.³

Several studies have focused on preoperative opioid use and its relationship to patient-reported outcomes.^4,5,6,7,8 A meta-analysis found opioid use prior to total knee and hip arthroplasty was associated with worse scores on patient-reported outcome measures.⁶ Preoperative opioid use was also associated with prolonged hospital stays and an increased likelihood of discharge to rehabilitation facilities.⁵ However, there is limited data surrounding preoperative opioid use and its association with outcomes after shoulder surgery.

In patients with primary glenohumeral osteoarthritis undergoing shoulder arthroplasty, lower pre and postoperative outcome scores have been associated with preoperative opioid use, but a need for more expansive investigations was emphasized.⁷ Alarmingly, preoperative opioid use in total shoulder arthroplasty (TSA) patients was associated with a five times greater incidence of opioid use one year after surgery.⁴

That National Institute of Health Patient-Reported Outcomes Measurement Information Systems (PROMIS) is a validated and reliable tool providing clinicians with important feedback regarding physical and psycho-social patient reported outcomes (PROs).⁸ The Maryland Orthopedic Registry (MOR) is a prospective research registry that includes PROMIS computer adaptive testing to comprehensively assess patients undergoing a variety of orthopedic surgical procedures.⁹ Previously, Nadarajah et al. used the MOR to study associations between preoperative opioid use and preoperative PROs in patients undergoing shoulder surgery and found preoperative opioid use to be independently predictive of worse preoperative function, pain, and social satisfaction.⁸ We aimed to expand these results by performing a follow-up study.

Thus, the purpose of this study was to investigate the relationship between preoperative opioid use and patient-reported outcomes at two-year follow-up in patients undergoing elective shoulder surgery. We hypothesized that patients using opioids preoperatively would have worse patient-reported outcomes.

2 Materials and methods

A prospective institutional review board-approved orthopedic registry, the MOR, was queried for patients that underwent elective shoulder surgery at a single urban institution from June 2015 to June 2017. All data was gathered using the Research Electronic Data Capture (REDCap™) data collection system. Inclusion criterion was patients undergoing elective shoulder surgery. Exclusion criteria consisted of: inability to speak English, younger than 12 years old, incarceration or ward of the state, inability to provide an email, or incomplete baseline survey information.⁹ Patients that opted to enroll in the registry completed electronic surveys one week perioperatively, and two years postoperatively. Self-reported parameters included demographic characteristics, social history and surgical history. Relevant past medical history, such as smoking status, American Society of Anesthesiologists (ASA) score, comorbidities, and preoperative opioid use, was obtained from the medical record.⁹ Of note, during medical record review, any patient with an opioid listed as an active medication within six weeks of surgery was defined as an ‘opioid user’. Common Procedural Terminology (CPT) codes were obtained from operative notes to identify procedure types and were listed by most common primary codes.

Enrolled patients also completed a number of preoperative questionnaires assessing physical function, psychosocial function, pain, activity and expectations. These included six PROMIS questionnaires (Physical Function (PF), Pain Interference (PI), Social Satisfaction (SS), Anxiety, Depression, and Fatigue), the American Shoulder and Elbow Surgeons (ASES) Standardized Shoulder Form,¹⁰ and two numeric pain scales (NPS) assessing shoulder and whole body pain. Of note, in certain domains (i.e., PROMIS PF and SS) higher scores are interpreted as “better” and in other domains (i.e., PROMIS PI, Fatigue, Anxiety, and Depression) higher scores are interpreted as “worse”. Activity measurements were collected with the Marx Activity Rating Scale.¹¹ Patient surgical satisfaction was assessed using the Surgical Satisfaction Questionnaire (SSQ-8), which allows the patient to rate their satisfaction in eight separate domains measured on a five-point Likert scale (1 = very satisfied; 5 = very unsatisfied).¹² Preoperative patient expectations and whether those expectations were met was recorded through the Musculoskeletal Outcomes Data Evaluation and Management System (MODEMS).

3 Results

Of the 353 patients enrolled, 257 (72.8%) completed two-year follow up questionnaires and preoperative opioid use was reported in 77 (30.0%) of those 257 patients. Among the primary CPT codes, open treatment of clavicular fracture (23515) and arthroscopic shoulder capsulorrhaphy demonstrated significantly higher and lower rates of preoperative opioid use, respectively (Table 1).

Preoperative opioid use was associated with several patient characteristics including increased BMI (p = 0.008), greater number of comorbidities (p = 0.004), greater number of total surgeries (p < 0.001), greater number of orthopedic surgeries (p = 0.003), lower income (p = 0.001), smoking (p = 0.001), actively living with caretaker (0.026), history of back pain (p = 0.009), and higher ASA score (p = 0.005) (Table 2).

Preoperative opioid use was associated with significantly worse scores on all patient reported outcome measures two years after surgery (Table 3). However, both groups improved significantly from baseline on all patients reported outcome measures except PROMIS Depression, MARS and NPS Whole Body Pain. Non-users improved in PROMIS Depression, neither group improved in MARS and both groups worsened in NPS Whole Body Pain (Table 4). The amount of change was not significantly different between opioid users and non-users except that non-users had significantly greater improvement in PROMIS PF (p = 0.041) and significantly greater increase in NPS Whole Body Pain (p = 0.028) (Table 4).

Multivariate analysis was used to assess if preoperative opioid use was an independent predictor of two-year and change in patient-reported outcomes scores. After controlling for confounding variables, preoperative opioid use was shown to be an independent predictor of worse two-year scores in PROMIS Physical Function, PROMIS Pain Interference, PROMIS Social Satisfaction, ASES Standardized Shoulder Form and Numeric Pain Score Operative Site (Table 5). Preoperative opioid use was not found to be an independent predictor of improvement in any of the outcome measures.

4 Discussion

With increasing emphasis on the deleterious effects of opioids in today's healthcare, understanding the effects of preoperative opioid use on surgical outcomes is essential. Specifically, we sought to better understand the relationship between preoperative opioid use and patient-reported outcomes two years after shoulder surgery. The results of this study supported our hypothesis that preoperative opioid use is associated with worse patient-reported outcomes two years after elective shoulder surgery. After controlling for confounders, preoperative opioid use was predictive of worse two-year scores in PROMIS Physical Function, PROMIS Pain Interference, PROMIS Social Satisfaction, ASES Standardized Shoulder Form and Numeric Pain Score Operative Site. These findings advocate against opioid use in the preoperative period for patients undergoing elective shoulder surgery.

This retrospective study examined a cohort of patients with a similar prevalence (30%) of preoperative opioid use compared to past literature.¹³ Preoperative opioid use has been shown to be a marker of worse general health status,¹⁴ which was consistent with our analysis of preoperative opioid use by two-year patient reported outcomes. Beyond the scope of our outcome parameters, preoperative opioid use has also been linked to higher readmission rates, longer hospital stays, higher incidence of revision surgery and greater discharge to rehabilitation facilities.^15,5,16,17

Past literature has associated preoperative opioid use with worse overall health profiles in a variety of orthopedic patients, including those undergoing shoulder surgery.^14,8 Similarly, we found preoperative opioid use to be associated with increased BMI, greater comorbidities, greater number of orthopedic and total surgeries, greater ASA score and history of back pain. Preoperative opioid use has also been shown to be one of the strongest predictors of postoperative opioid use.^18,19 Concerningly, recent literature has shown that patients undergoing shoulder arthroplasty who used opioids before surgery had a seven times greater incidence of opioid use four years after surgery.⁴ Although that study may not be generalizable to all shoulder surgeries, it highlights the potential progression of preoperative opioid use to chronic opioid use.

In previous literature concerning hip and knee surgery, preoperative opioid use was associated with worse patient outcomes.⁶ Similarly, preoperative opioid use was associated with worse outcomes in patients undergoing anatomic shoulder arthroplasty and reverse shoulder arthroplasty, respectively.⁷ Williams et al. also found that preoperative opioid use hindered patients from reaching peak functionality in multiple outcome measures (ASES, active range of motion (AROM), Simple Shoulder Test (SST)).²⁰ Our findings concur with these studies—preoperative opioid use was associated with worse scores in all PRO measures at two-year follow-up and multivariate analysis supports these findings in shoulder specific functional and pain outcomes. .

Our study aimed to assess preoperative opioid use in in a mixed cohort of shoulder procedures. A significant difference in preoperative opioid use was observed between primary CPT groups and these differences were incorporated into our multivariable analysis. Our study also incorporated PROMIS instruments in addition to legacy assessments such as ASES, MARS and NPS. PROMIS has been shown to correlate well with other orthopedic-specific PRO measures,⁹ and our results support those findings. The correspondence of preoperative opioid use with worse patient reported outcomes, both for PROMIS and legacy assessments, supports PROMIS as a consistent and valid tool.

Past literature has indicated that improvement in patient-reported outcome measures from baseline to two-year follow up did not vary significantly between the opioid and non-opioid users.^6,21,7,20 However, our study did not definitively support these findings. Although patients using preoperative narcotics may experience worse two-year outcomes after shoulder surgery, their degree of improvement relative to preoperative scores indicates surgical treatment is still efficacious. This finding may suggest that although surgery is beneficial regardless of preoperative opioid use, preoperative opioid may act as a barrier to maximizing postoperative outcomes.

Demographically, preoperative opioid use was found to be associated with greater BMI, greater number of comorbidities, greater number of total surgeries, greater number of orthopedic surgeries, and lower expectations. Preoperative opioid use was also found to be significantly higher in females, low income patients (<70k), smokers, patients with history of back pain, patients with friends/family as caretakers, and patients with higher ASA scores. Previous studies found similar associations and encouraged an emphasis on the importance of integrating the biopsychosocial model of medicine in treatment plans for patients in this population.^8,22 Our study reiterates the importance of these factors, which when taken into consideration with our previously discussed findings, paint a complicated picture of opioids’ role in the health of our surgical patients.

This study is not without limitations. This is a retrospective cohort study and the limitations which typically apply to such a study design also apply here. Most notably, causality cannot be ascertained from this study. The use of a multivariable regression with step-wise elimination helped mitigate the presence of any confounders, such as procedure type, that may be present, but is not fail-safe for eliminating selection bias in our patient population. Opioid use was defined as a binary variable, which hinders our ability to query how the quantity of preoperative opioid use impacts patient-reported outcomes. Additionally, as preoperative opioid use was partially determined by patients self-reporting a narcotic as an active medication, it is possible that some patients may have not felt comfortable admitting opioid use. Finally, as previously mentioned, our patient population included a large number of procedures, which could introduce confounders and limit the study generalizability. However, we believe our results identify an important trend that may be common across multiple shoulder procedures. Our analysis of primary CPT codes followed by controlling for confounders through multivariate analysis attempts to mitigate these shortcomings. While this is the case, this study utilized a large and diverse patient cohort to study preoperative opioid use and postoperative outcomes after elective shoulder surgery.

5 Conclusion

In this study of patients undergoing elective shoulder surgery, preoperative opioid use was associated with worse scores in all PRO measures at two-year follow-up including all six PROMIS domains, ASES, MARS, NPS pain scores, MODEMS, and SSQ-8. Furthermore, preoperative opioid use was independently associated with worse two-year scores in the PROMIS PF, PROMIS PI, PROMIS Social Satisfaction, ASES, and Numeric Pain Score Operative Site. Our findings provide further evidence that orthopedic surgeons should consider alternate methods of preoperative pain management, when possible.

Funding

This work was supported by a grant from The James Lawrence Kernan Hospital Endowment Fund, Incorporated.

Disclaimer

None.

CRediT authorship contribution statement

Michael A. McCurdy: Conceptualization, Writing – original draft, Writing – review & editing. Cameran I. Burt: Conceptualization, Writing – original draft, Writing – review & editing. Matheus B. Schneider: Writing – original draft, Writing – review & editing, Methodology, Data curation, Formal analysis. Tina Zhang: Writing – original draft, Writing – review & editing, Methodology, Data curation, Formal analysis. Michael J. Foster: Writing – original draft, Writing – review & editing. Ali Aneizi: Writing – original draft, Writing – review & editing. Mohit N. Gilotra: Writing – original draft, Writing – review & editing, Supervision. S. Ashfaq Hasan: Writing – original draft, Writing – review & editing, Supervision. R. Frank Henn III: Writing – original draft, Writing – review & editing, Methodology, Data curation, Formal analysis, Project administration, Supervision, Funding acquisition, Validation.