Incidence and Predictive Factors for Additional Opioid Prescription after Endoscopic Skull Base Surgery

Abstract

Introduction  Postoperative pain management and opioid use following endoscopic skull base surgery (ESBS) is not well understood. A subset of patients requires additional opioid prescription (AOP) in the postoperative period. The objective of this study is to describe the incidence of AOP, as well as evaluate patient and surgical characteristics that may predict additional pain management requirements following ESBS.

Methods  A retrospective review of cases undergoing ESBS between November 2016 and August 2018 was performed. We reviewed patients' sociodemographic and clinical data, and Controlled Substance Utilization Review and Evaluation System (CURES) records. Stepwise multivariable logistic regressions were performed to evaluate the factors associated with AOP within 60 days following surgery.

Results  A total of 42 patients were identified. Indications for ESBS included intracranial mass (64.2%), sinonasal malignancy (23.8%), and skull base reconstruction (9.5%). AOP were recorded in nine patients (21.4%). There were no significant differences in operative factors, including approach, lesion location, or perioperative analgesia between the two cohorts. On multivariable logistic regression, we found that younger age (odds ratio [OR]: 0.891, 95% confidence interval [CI]: 0.79–1.00, p  = 0.050), comorbid depression (OR: 86.48, 95% CI: 1.40–5,379.07, p  = 0.034), and preoperative opioid use (OR: 104.45, 95% CI: 1.41–7,751.10, p  = 0.034) were associated with additional prescriptions postoperatively.

Conclusion  The requirement for extended postoperative opioid pain control is common after ESBS. Patient demographics including age and psychosocial factors, such as depression may predict the need for AOP after ESBS. These results suggest that patient-driven factors, rather than surgical characteristics, may determine the need for prolonged pain control requirements after ESBS.

Introduction

The epidemic of opioid-associated morbidity and mortality over the last two decades has been well documented. The Centers for Disease Control and Prevention (CDC) estimates that opioid overdose related deaths were six times higher in 2017 compared to 1999; prescription rates reached a high of 81.3 prescriptions per 100 Americans in 2012. ¹ Surgeons play a unique role in this public health emergency, as they are the second highest prescribers of narcotics, trailing only pain-medicine specialists among all practioners. ² A growing awareness that prescription diversion is a substantial source of illicit supply ^{3 4} has further placed prescription habits under scrutiny. Opioids remain the standard of care for postoperative pain management, as acute surgical treatment often requires more potent analgesia than standard nonsteroidal anti-inflammatory medications. ⁵ However, selected patients have extended opioid requirements beyond the immediate postsurgical phase. There is an urgent need for tools to better understand patient and procedural variables may predispose such patients to extended postoperative opioid therapy.

Indications and techniques for endoscopic skull base surgery (ESBS) have evolved considerably over the last 20 years and it is now recognized as a minimally-invasive approach to skull-base pathology. Minimally invasive approaches in other fields including neurologic spine, ⁶ transplant living donor surgery, ⁷ and pediatric urology, ⁸ have consistently demonstrated lower reported postoperative pain and less narcotic utilization when compared to open surgery. While a multidisciplinary ESBS collaboration has been accepted as an alternative to more invasive neurosurgical intervention, current literature in the has primarily focused on cerebrospinal fluid (CSF) leak complications, nasal care, debridement schedules, and quality of life (QOL) assessment. ^{9 10} Surveys into postprocedural practices have yielded insight into surgeon preferences for postoperative imaging, patient restrictions, and postoperative level of care. ^{11 12} There remains a significant gap in the understanding and description of pain management regimens, and little insight into factors predictive of greater postoperative opioid requirement.

Within this study, we aim to describe the incidence of and patient and surgical factors that influence persistent postoperative opioid analgesic requirements following ESBS. Persistent opioid use was defined as an additional filled prescription, beyond initial postoperative opioid prescription, within 60 days of surgery. To facilitate this analysis, we utilized a state-wide controlled substance database and examined previously described clinical and demographic risk factors. Our goal is to provide increased understanding on pain management following ESBS and identify predictors of nonstandard opioid use.

Methods

Patient Population and Inclusion Criteria

Adult patients (≥18 years old) who underwent ESBS were included in this retrospective study. Enrollment consisted of patients who underwent the procedure at the University of California San Diego between November 2016 and August 2018. Clinical data were collected up to 60 days after the procedure for each patient. The study was approved by the University of California, San Diego Institutional Review Board (IRB; 161005X).

Study Data Collection

The Controlled Substance Utilization Review and Evaluation System (CURES), a database of schedules II, III, and IV controlled substance prescriptions, dispensed in California, was reviewed for total amount and type of opioid prescribed to our study population. The controlled substances from 30 days prior to surgery to 60 days after the procedure were recorded. The two cohorts within the study population were defined as the patients requiring additional opioid prescription and the single prescription population. AOP was defined as an additional opioid prescription within the 60-day postoperative window.

Statistical Analysis

Descriptive analytics and univariate regression were used to evaluate clinical and sociodemographic variables significantly associated with additional opioid prescription. The primary outcome measure of the study was additional opioid prescription within 60 days of surgery. To determine predictive value of patient variables on AOP, logistic regression was used to assess continuous variables, including age, median income, distance to provider, perioperative acetaminophen, and steroid dosage. Binary variables, including sex, marital status, employment status, public insurance, and presence of medical comorbidities were analyzed using Chi-squared tests.

A multivariable logistic model was built using a stepwise regression with backward elimination at a p -value of 0.10. To control for patient sex, sociodemographic variables, and smoking status, these variables were included in a second multivariable logistic model. All α levels and error probabilities were set at the standard 0.05 level. Statistical analyses were performed using SPSS 22 and SAS.

Results

Forty-two patients were included in this retrospective cohort analysis. Demographically, 50.0% were male, 52.4% were White, 66.7% were married, 73.8% were employed, 52.4% had public insurance, and the mean age was 51.6 years. Further, 16.6% of patients had comorbid depression, 11.9% had comorbid anxiety, and 16.6% had history of prior opioid use ( Table 1 ).

Table 1. Demographic characteristics of the study population.

	Total ( n  = 42)	Additional opioid prescription ( n  = 9)	Single-use ( n  = 33)	p -Value
Demographics: number (percent) unless otherwise listed
Age, y (mean, SD)	51.62 (16.56)	42.22 (17.89)	54.18 (15.48)	0.030
Male	21 (50.0)	3 (33.3.)	18 (54.5)	0.454
Race/ethnicity				0.781
White	22 (52.4)	3 (33.3)	19 (57.6)
African American	6 (14.3)	3 (33.3)	3 (9.1)
Hispanic	11 (26.2)	3 (33.3)	8 (24.2)
Other	3 (7.1)	0 (0.0)	3 (9.1)
Marital status				0.338
Single	7 (16.7)	3 (33.3)	4 (12.1)
Married	28 (66.7)	5 (55.6)	23 (69.7)
Divorced	5 (11.9)	1 (11.1)	4 (12.1)
Widowed	2 (4.8)	0 (0.0)	2 (6.1)
Employment status				1.000
Employed	31 (73.8)	5 (55.6)	26 (78.8)
Unemployed	4 (9.5)	2 (22.2)	2 (6.1)
Retired	4 (9.5)	1 (11.1)	3 (9.1)
Disability	1 (2.4)	0 (0.0)	1 (3.0)
Student	2 (4.8)	1 (11.1)	1 (3.0)
Mean income (USD) (mean, SD)	70,324 (27,972)	65,840 (20,917)	71,546 (29,765)	0.585
> California median income ($71,805)	9 (21.4)	3 (33.3)	6 (18.2)	0.422
Distance to provider (miles) (mean, SD)	25.8 (27.0)	24.7 (21.0)	26.1 (28.7)	0.866
Insurance				0.267
Medicare	9 (21.4)	1 (11.1)	8 (24.2)
Medi-Cal	13 (31.0)	3 (33.3)	10 (30.3)
Private	20 (47.6)	5 (55.6)	15 (45.5)
Medical history
Depression	7 (16.7)	3 (33.3)	4 (12.1)	0.028
Anxiety	5 (11.9)	2 (22.2)	3 (9.1)	0.288
Migraine	1 (2.4)	0 (0.0)	1 (3.0)	1.000
Social history
Smoking history				0.514
Current smoker	4 (9.5)	0 (0.0)	4 (12.1)
Former smoker	11 (26.2)	4 (44.4)	7 (21.2)
Never	27 (64.3)	5 (55.6)	22 (66.7)
Current alcohol use	17 (40.5)	2 (22.2)	15 (45.5)	0.271
Current drug use	4 (9.5)	2 (22.2)	2 (6.1)	0.196
Preoperative opioid use	7 (16.7)	5 (55.6)	2 (6.1)	0.003
Initial postoperative prescription				0.919
Hydrocodone bitartrate–acetaminophen 5–325	16 (38.1)	4 (44.4)	12 (36.3)
Oxycodone HCl–acetaminophen 5–325	6 (14.2)	3 (33.3)	3 (9.1)
Oxycodone 5	9 (21.4)	0 (0.0)	9 (27.3)
Oxycodone 15	1 (2.4)	1 (11.1)	0 (0.0)
Tramadol 50	2 (4.8)	1 (11.1)	1 (3.0)
None	8 (19.0)	0 (0.0)	8 (24.2)
MME (mean, SD)	254.1 (331.2)	376.4 (627.6)	213.3 (110.7)	0.210

Abbreviations: MME, morphine milligram equivalents; SD, standard deviation; USD, U.S. dollars.

For the initial postoperative prescription, 16 patients (38.1%) received hydrocodone bitartrate–acetaminophen 5–325. Six patients (14.2%) received oxycodone HCl–Acetaminophen 5–325, nine patients (21.4%) received oxycodone 5 mg, one patient (2.4%) received oxycodone 15 mg, two patients (4.8%) received tramadol 50 mg, and eight patients (19.0%) did not receive any opioid analgesia. The average morphine milligram equivalent (MME) per initial prescription was 254.1 mg ( Table 1 ).

Nine patients (21.4%) required additional opioid prescriptions following endoscopic skull base surgery. AOP patients were significantly younger (42.2 vs. 54.2 years, p  = 0.050), more likely to have comorbid depression ( χ ²  = 6.36, p  = 0.012), and more likely to have history of opioid use prior to surgery ( χ ²  = 12.47, p  < 0.001). Of the patients with a previous diagnosis of depression, 75% of AOP patients were taking antidepressants at the time of study (three of four), while 33% of single-use patients were taking antidepressants (one of three). Of the patients with history of opioid use prior to surgery, AOP cohort regimens were: three patients taking hydrocodone bitartrate–acetaminophen 5–325, two patients taking oxycodone HCl–acetaminophen 5–325; both patients in the single-use cohort with pretreatment opioid use were taking hydrocodone bitartrate–acetaminophen 5–325. AOP patients' first prescription had an average of 376.4 MMEs, compared to 213.3 MMEs for single-prescription patients. However, this difference did not reach statistical significance ( p  = 0.210).

Eighteen (42.9%) patients had lesions in the anterior skull base, 23 (54.8%) in the skull base compartment, and one (2.4%) in the posterior skull base. Indications for ESBS included intracranial mass (64.2%), sinonasal malignancy (23.8%), and skull base reconstruction (9.5%). Perioperative characteristics, including American Society of Anesthesiologists (ASA) score, intraoperative steroid use and intraoperative acetaminophen use, and surgical approach did not vary significantly between the two cohorts ( Table 2 ).

Table 2. Perioperative characteristics of the study population.

	Total ( n  = 42)	Additional opioid prescription ( n  = 9)	Single-use ( n  = 33)	p -Value
Medications
Perioperative steroids (mean, SD)	7.57 (3.26)	8.22 (2.11)	7.39 (3.51)	0.500
Perioperative acetaminophen (mean, SD)	2.48 (0.63)	1.11 (0.33)	1.15 (0.56)	0.858
Surgical characteristics
ASA classification				0.989
1	1 (2.4)	1 (11.1)	0 (0.0)
2	22 (52.4)	4 (44.4)	18 (54.5)
3	17 (40.5)	4 (44.4)	13 (39.4)
4	2 (4.8)	0 (0.0)	2 (6.1)
Approach				0.477
Transnasal- transphenoidal	19 (45.2)	3 (33.3)	16 (48.4)
Extended endoscopic	23 (54.8)	6 (66.7)	17 (51.6)
Lesion location				0.783
Anterior skull base	18 (42.8)	4 (44.4)	14 (42.4)
Middle skull base	23 (54.8)	5 (55.6)	18 (54.5)
Posterior skull base	1 (2.4)	0 (0.0)	1 (3.1)
Indication				0.435
Sinonasal malignancy	8 (19.0)	2 (22.2)	8 (24.2)
Intracranial mass	27 (64.2)	7 (77.8)	20 (60.6)
Skull base reconstruction	4 (9.5)	0	4 (12.1)
Other	1 (2.4)	1 (11.1)	0 (0.0)

Abbreviations: ASA, American Society of Anesthesiologists; SD, standard deviation.

On multivariable analysis, including only factors meeting the significance cut-off, we found that younger age (odds ratio [OR] for increasing age: 0.902, 95% confidence intervals [CI]: 0.83–0.99, p  = 0.025), comorbid depression (OR: 54.55, 95% CI: 1.84–1,619.42, p  = 0.010), and preoperative opioid use (OR: 98.39, 95% CI: 2.95–3,285.14, p  = 0.021) were predictors for additional opiate requirements after ESBS ( Table 3 ). After controlling for clinical and demographic factors, including age, sex, ethnicity, marital status, smoking status, alcohol use, and anxiety, the factors younger age (OR for increasing age: 0.891, 95% CI: 0.79–1.00, p  = 0.050), comorbid depression (OR: 86.48, 95% CI: 1.40–5,379.07, p  = 0.034), and preoperative opioid use (OR: 104.45, 95% CI: 1.41–7,751.10, p  = 0.034) remained significantly associated with AOP ( Table 4 ).

Table 3. Multivariable logistic regression of clinical variables and postoperative opioid refill.

	OR	95% CI	p value
Age	0.891	0.79–1.00	0.050
Preoperative opioid use	98.39	(2.95–3,285.14	0.021
Depression	54.55	(1.84–1,619.42)	0.010

Abbreviations: CI, confidence interval; OR, odds ratio Preop, preoperative.

Table 4. Multivariable logistic regression of clinical variables and postoperative opioid refill controlling for demographic factors.

	OR	95% CI	p -Value
Age	0.891	0.79–1.00	0.050
Female sex	0.699	0.036–13.41	0.812
Married	1.114	0.052–23.90	0.945
Employed	3.435	0.12–95.09	0.466
Tobacco use	0.451	0.029–7.12	0.852
Alcohol use	34.60	0.002–275.93	0.299
Depression	86.48	1.40–5,379.07	0.034
Anxiety	0.675	0.006–74.00	0.870
Preoperative opioid use	104.45	1.41–7,751.10	0.034

Abbreviations: CI, confidence interval; OR, odds ratio.

Discussion

The increasing incidence in opioid-related deaths over the past decade has generated a substantial interest in strategies to mitigate misuse. ¹³ Perioperative analgesia has come under particular scrutiny, as many patients are first exposed to opioids in this setting. Although multiple organizations have proposed recommendations for management of acute postoperative pain, such as protocols including epidural anesthesia and regional blocks, unique anatomical constraints have made it difficult to adapt these guidelines to sinonasal surgery. ¹³ Within ESBS, persistent postoperative pain is a commonly reported morbidity, with up to 17% patients complaining of pain 15 months after surgery. ¹⁴ Improved understanding of risk factors for additional opioid requirements and preoperatively identifying high-risk patients could help minimize complications and assist surgeons in tailoring postoperative analgesic regimens to minimize over-prescription and diversion.

We found approximately 20% of patients met our criteria for persistent opioid use following ESBS. This percentage is similar to that previously reported in endoscopic sinus surgery ¹⁵ and neurosurgical procedures. ¹⁶ Within our cohort, rates of AOP did not vary significantly between anatomic location of the lesion and approach. This is also reflected in work by Brummett et al demonstrating that rates of persistent opioid use are independent of the extent of surgery. ⁵ This suggests that surgical factors may be less critical in informing patient-reported pain levels. Rather, demographic or clinical characteristics could play a larger role in the experience and tolerance of postprocedural pain. By analyzing clinical data associated with AOP, we sought to better describe these risk factors within ESBS.

We found that younger patients were more likely to have additional opioid requirements. Prior studies have demonstrated that younger patients report greater subjective postoperative pain. ^{17 18} Similarly, Clarke et al identified a relationship between younger age and an increased risk of prolonged opiate use following major elective procedures in the U.K. ¹⁹ While it is increasingly clear that younger patients experience more pain after surgery, the precise etiology of this phenomenon is less apparent. It has been suggested that nociceptive receptor responses and morphine equivalent requirements decrease with age, ²⁰ providing a physiological basis for these findings. Additionally, weaker social support and decreased access to alternative pain management could contribute to increase opioid dependence in younger patients ²¹ and should also be considered when determining postoperative opioid requirements.

Among medical and psychiatric comorbidities included in the analysis, depression was associated with an increased risk of AOR. This finding is in accordance with previous literature describing risk factors of chronic opioid use in postoperative patients. ^{5 22 23} Goesling et al showed that patients with depression had a higher probability of opioid dependence independent of pain severity. ²⁴ Psychological distress is known to place patients at risk for persistent postoperative pain; depression, specifically, is associated with preoperative apprehension and over-estimation of pain intesity. ²³ Negative affect has also been shown to mitigate the effectiveness of opioid analgesia and higher levels of tolerance. ²⁵ Furthermore, it has been hypothesized that patients with depression continue to use opioids to self-medicate for psychiatric symptoms. ²⁴ Within our population, 33.3% of AOP patients had depression prior to surgery, almost three times higher than the 12.1% incidence within the single use cohort. Together, the increased incidence and predictive value (OR: 86.48, 95% CI: 1.40–5,379.07) of depression on AOP demonstrates substantial clinical value in recognizing psychological vulnerability in the perioperative period.

Within our multivariable model, the strongest and most consistent predictor of AOP was opiate use within 30 days prior to surgery. Presence of preoperative pain has been identified as a significant risk factor for both prolonged pain and increased analgesic requirements in orthopaedics, ^{26 27 28} general surgery, ²⁹ as well as otolaryngology. ¹⁵ We found that patients with preoperative opioid use were significantly more likely (OR: 104.45, 95% CI: 1.41–7,751.10) than opioid-naïve patients to require AOP. Although the OR is likely skewed upwards by the relatively small sample size, this finding confirms clinical suspicion that this population is more susceptible to persistent opioid use. In a study of elective procedures of the upper extremity, Waljee et al found that opioid refill rates increased from 4.5% in opioid-naïve patients to 23.9% in those with previous use. ²⁸ Similarly, a systematic review of predictors of postoperative pain found that preexisting pain and analgesic use was highly correlated with extended pain after surgery. ²⁰ Both psychological and biological processes can play a role in this finding. Analgesic use has been shown to heighten the experience of novel pain, as well as decrease the threshold of subjective “catastrophic” pain. ³⁰ Opioid utilization can also lead to increased response of peripheral nociceptive pathways, upregulation of prostaglandin transcription, and enhancement of dorsal horn pain fiber excitability. ³¹ Prolonged iatrogenic activation µ -opioid receptors in the spinal cord suppresses inhibitory GABAergic interneurons and can heighten pain transmission through the superficial dorsal horn pathways. ³² Together, these mechanisms can lead to sustained perception of pain beyond expected tissue recovery.

Overall, an improved understanding and recognition of patient-level risk factors for additional opioid requirement can empower surgeons to have more nuanced counseling and expectation setting with patients in the preoperative setting and help in tailoring postoperative analgesic management. This is one of the first studies to focus on these clinical variables within ESBS. Addressing potentially modifiable preoperative risk factors, such as opioid use may minimize the need for extended opioid therapy. Studies investigating opioid tapering prior to total joint arthroplasty have shown have shown improved postoperative disease-specific and QOL outcomes, ^{26 33} and patients undergoing ESBS may also be benefitted. Similarly, addressing depression or anxiety prior to ESBS, through medication or counseling, may have additional benefit for postoperative pain control.

Although this study provides a perspective on the predictive factors in opioid use after ESBS, there are important limitations that warrant further discussion. We defined AOP as filling any opioid prescription after 60 days, a binary outcome may fail to provide insight into quantity consumed. It was also difficult to verify the indications for the subsequent prescription, particularly in cases, where physicians outside our institute prescribed the medication. Additionally, the described cohorts are from a single surgeon at a single academic center and our results had the predictable wide confidence intervals reflective of statistical variance within a small population. Lastly, it is unclear if the distribution of surgical indications within our cohort is representative of an average academic center. There is little data describing the indications of ESBS due to disparate current procedural terminology (CPT) coding practices. ¹² The particular proportions of lesion location and approaches may have implications on the uncovered risk factors. However, given the concordance of our findings with research from other surgical subspecialties, it is probable that these patient-level factors are implicated in persistent opioid use following ESBS.

Conclusion

The current study describes the incidence of and clinical variables associated with additional opioid requirements after endoscopic skull base surgery. Within this growing field, quantifying the occurrence of AOP and identifying clinical risk factors is critical for the development of prescription recommendations. Understanding the role of age, comorbid depression, and preprocedural opioid use can guide practitioners in counseling and postoperative management. Research into interventional strategies with longer follow-up periods will provide further insight into the importance of the identified factors on ESBS outcomes.