1 Introduction

Narcotic induced overdose is the leading cause of accidental death in the United States, resulting in approximately 47,000 deaths in 2017.^1,2 Taken on its own, prescription opioid abuse contributes to nearly 40% of the aforementioned overdose deaths.³ The postoperative period for patients undergoing surgery has been shown to contribute to prescription opioid misuse and patterns of chronic abuse.^4,5 Specifically, spine surgery has been defined as a “high risk” episode for sustained postoperative opioid use, where historically pain control was regarded as the “5th vital sign”.^6–10 In anterior cervical discectomy and fusion (ACDF), one of the most commonly performed spine surgeries, postoperative opioid prescription is the standard of care.⁵ Though it is important to adequately address pain during the postoperative period, prescriptions have been found to commonly exceed the dose and duration necessary for a patient's return to function after surgery.^11,12

The profound morbidity and mortality associated with opioid misuse after ACDF necessitates a more granular understanding of underlying patient predictors for prolonged opioid use. In particular, elucidating preoperative risk factors allows physicians to address those that are modifiable before surgery to help decrease perioperative morbidity. To date, prolonged narcotic use after ACDF has been strongly associated with prior history of opioid use, substance abuse, and mental health conditions including depression and anxiety.^7,13–20 However, the role that preoperative symptom severity and functional impairment play in postoperative opioid use patterns has yet to be elucidated.

While motor weakness has been associated with reduced preoperative health-related quality of life (HRQOL) outcomes, the effect of weakness on postoperative opioid use remains unknown.²¹ In the present work, symptom severity is explored through motor function weakness on physical exam arising from neurologic (central or foraminal) compression. Nerve root ischemia and inflammation associated with neurologic compression may mediate neuropathic pain thereby increasing narcotic requirement.^22–24 Therefore, the goal of this study is to determine if preoperative weakness is an isolated risk factor for postoperative prolonged opioid use or abuse.

2 Material and methods

After Institutional Review Board approval (IRB #19D.508), all patients over age 18 who underwent one-to four-level primary or revision ACDF for cervical spondylosis associated with radiculopathy and/or myelopathy at an academic medical center between 2016 and 2019 were retrospectively identified. Waiver was granted for patient informed consent as a minimal risk research study. Additional inclusion criteria were complete patient demographic profiles, surgical characteristics, pre- and postoperative motor strength testing, and postoperative opioid use records. Patients were excluded if their descriptive data or opioid utilization could not be assessed or if their surgical indications included trauma, infection, or malignancy.

3 Results

3.2

3.2 Comparison of opioid use by preoperative weakness

Postoperative opioid use varied significantly based on patient preoperative weakness. Preoperatively weak patients were significantly more likely to be prolonged narcotic users, receive very high opioid dose on a single prescription, and exhibit inappropriate use patterns through prescriptions with overlapping dates (Table 2). Specifically, 3.9% of weak patients vs 0.05% of non-weak patients received prescriptions of ≥ 200 MME for postoperative pain reduction (p = 0.031) (Fig. 1). A Kaplan Meier Curve (Fig. 2) demonstrates that at any postoperative time point, opioid discontinuation was less probable in the preoperatively weak group. Furthermore, greater than 40% of the group with preoperative weakness continued to take opioid medications three years after surgery compared to less than 10% of the control group (Fig. 2). Despite over 80% of preoperatively weak patients achieving motor recovery, at least 20% of those recovered patients continued taking opioids three years after surgery.

Table 2 Comparison of postoperative use by preoperative weakness status.

Category	Opioid Outcome Measure	No Preop Weakness	Preop Weakness	P Value
Dosing	Peak MME ≥ 200 (n = 481):			0.031
	No	402 (99.5%)	74 (96.1%)
	Yes	2 (0.50%)	3 (3.90%)
	Peak MME ≥ 90 (n = 481):			0.280
	No	300 (74.3%)	52 (67.5%)
	Yes	104 (25.7%)	25 (32.5%)
Prolonged Use	>1 Postop Opioid Prescriptions (n = 382):			0.005
	No	217 (67.8%)	30 (48.4%)
	Yes	103 (32.2%)	32 (51.6%)
	≥ 120 Days of Postop Opioid Use (n = 385):			<0.001
	No	299 (92.6%)	45 (72.6%)
	Yes	24 (7.43%)	17 (27.4%)
	>10 Postop Opioid Prescriptions (n = 385):			<0.001
	No	307 (95.0%)	47 (75.8%)
	Yes	16 (4.95%)	15 (24.2%)
Appropriateness	Appropriateness of Postop Opioid Use (n = 335):			0.001
	Sole Provider	213 (75.5%)	32 (60.4%)
	Multiple Provider + Appropriate Use	60 (21.3%)	12 (22.6%)
	Multiple Provider + Inappropriate Use	9 (3.19%)	9 (17.0%)

Close

Fig. 1 Postoperative peak MME by preoperative weakness status.

Export to PPT

Close

Fig. 2 Kaplan Meier survival curve to measure time to opioid discontinuation by preoperative weakness status.

Export to PPT

4 Discussion

The prescription opioid epidemic has profound medical and socioeconomic repercussions.^26–28 Orthopaedic surgeons are among the highest opioid prescribers with spinal surgery representing a “high risk” care episode.¹⁰ Currently there is an incomplete understanding of the various patient risk factors that predict high postoperative opioid requirement and extended usage patterns. To date, this is the first study to examine preoperative weakness as a predictor of postoperative narcotic usage. The results of our study show that preoperative weakness is associated with preoperative opioid tolerance as well as prolonged and inappropriate opioid use with a high postoperative dose requirement. After taking into account associated preoperative risk factors, preoperative weakness was shown to be an independent predictor of prolonged postoperative opioid use.

Debilitating neurologic compression resulting in motor weakness serves as a proxy for the underlying disease severity and often warrants urgent referral and intervention to improve the likelihood of motor recovery.^21,29 However, it remains unclear whether the severity of neurologic compression is associated with increased pain, therefore increasing presurgical and postsurgical opioid requirements. Cervical foraminal compression is associated with both neuropathic mediated pain and sensorimotor deficit.^22,24 Radicular pain arises from ischemia and inflammation triggering mechanical and nociceptive pain pathways mediated by sensory neuropeptides and interleukins.^22–24 Theoretically, more severe mechanical compression resulting in motor weakness would increase nerve root ischemia and inflammation and further the release of nociceptive mediators. However, it remains to be determined if the degree of radicular compression scales with neuropathic pain generation clinically.

Previous literature has shown a significant association between preoperative weakness and higher VAS Neck and VAS Arm pain scores.²¹ In the present work, VAS Neck was greater in patients with preoperative weakness, but this relationship was not significant. Therefore, the significantly greater opioid tolerance in preoperatively weak patients in the present study cannot be attributed to pain alone. One possible explanation is the difference in preoperative patient optimization between those presenting with symptoms of pain compared to those presenting with weakness. While patients presenting with pain would be encouraged to wean opioid usage before being considered surgical candidates, patients with weakness would likely undergo early surgical intervention for the best chance of improving postoperative function despite continued opioid use.²¹ As suggested by this study, as well as others, preoperative opioid use is a risk factor for prolonged opioid use postoperatively.^7,13–18,30

Preoperative weakness was observed to be a strong risk factor for postsurgical opioid use. Univariate analyses of preoperatively weak patients compared to non-weak controls demonstrated greater postoperative opioid dose, increased inappropriate prescriptions, and greater prolonged usage. These findings are likely confounded by characteristics associated with both weakness and opioid use, including preoperative opioid tolerance and potentially preoperative pain. However, logistic regressions demonstrated weakness to be an independent predictor of prolonged postoperative opioid use. Duration of preoperative symptoms alone was not significantly associated with postoperative opioid use. This suggests that although weak patients may require more urgent surgical intervention in the setting of neurologic deficits, earlier surgery does not impact postoperative opioid use. Furthermore, postoperative weakness was not related to postoperative opioid use metrics suggesting that the association between preoperative weakness and extended usage exists regardless of motor recovery. It is unclear what unidentified pathology or patient characteristics explain this unique relationship. Future research should investigate HRQOL outcomes specific to preoperatively weak patients that abuse narcotic prescriptions postoperatively and compare delta HRQOL with and without motor recovery.

Apart from preoperative weakness, the current study investigates known preoperative risk factors for elevated or extended opioid use after ACDF and proposes new correlates for future investigation. Preoperative opioid use has been consistently identified as the strongest predictor of prolonged postoperative opioid use.^{7,13–18,30,31} Evidence suggests that opioid tolerant patients are four to five times more likely that opioid naïve patients to become prolonged opioid users.^15,17,18 This odds ratio continues to increase with extended durations of preoperative opioid use.^7,17 Other emerging risk factors for prolonged postoperative opioid abuse after ACDF include depression, anxiety, substance abuse, tobacco use, younger age, myelopathy, and insurance status (Medicare/Medicaid).^7,13–20,31 Elevated opioid dose in the direct postoperative period has been correlated with similar factors including preoperative opioid and benzodiazepine use and increased number of levels fused.¹⁶ Knowledge of these risk factors is crucial for presurgical optimization by counseling patients on opioid discontinuation.

In the present study, preoperative opioid tolerance, depression, and current smoking status corroborate the existing literature as predictors of prolonged usage (>1 script). The connection between preoperative opioid tolerance and increased MME requirement is also supported. Younger age, which has previously been correlated with opioid chronicity, also predicted high peak MME.³⁰ The present study demonstrates novel correlations between high perioperative opioid dose and male sex, medical comorbidities (higher CCI), and prior cervical surgery. Sex variations could reflect metabolic differences or differing sociocultural expectations such as desire to return to work.³² Medical comorbidities, namely HIV and liver disease, are commonly associated with opioid use disorder but the association with postoperative opioid dose requires further validation.³³ Likewise, revision procedures are often longer and more complex than primary surgery leading to larger postoperative pain control requirements. A novel association unique to this study, preoperative VAS Neck is shown to correlate with high postoperative opioid doses and prolonged usage patterns. To the best of the authors’ knowledge, this is the first study linking specifically preoperative pain to opioid abuse after ACDF.¹⁸ This demonstrates that opioid misuse depends on preoperative pain status regardless of the extent of postsurgical pain improvement.

This study is not without limitations. First, the study is subject to biases inherent to retrospective study design and cohort generation. Although patients were selected in a systematic method, potentially relevant cases were identified based on availability of PDMP data and surgical characteristics. In particular, drug monitoring data was limited geographically to the state of Pennsylvania, however patients may have sought continued long-term pain management in adjacent states. To mitigate the impact of this limitation, patients with New Jersey residences were not included in any opioid use metrics outside of immediate postoperative opioid dosing. Additionally, narcotic data was reported by prescription and is subject to inaccuracy based on patient filling patterns. The primary variable of weakness (MMT≤3) is a heterogenous measure dependent on provider physical examination and patient effort. For this reason, plus and minus additions to MMT strength grades were not utilized. Lastly, duration of preoperative symptoms was based on patient reporting and therefore is subject to recall bias.

5 Conclusion

Overall, preoperative weakness was found to be an independent predictor of prolonged opioid use after ACDF. Preoperative opioid tolerance, neck pain, duration of symptoms, and motor recovery incompletely explain the relationship between preoperative weakness and prolonged opioid usage. The benefits of presurgical optimization with narcotic discontinuation and multimodal pain management in preoperatively weak patients should be thoroughly weighed against the relative urgency of surgical intervention.