Association Between Spine Surgery and Availability of Opioid Medication

IMPORTANCE: Prolonged prescribing of opioids after spine surgery is often perceived as a negative outcome, but successful opioid reduction may occur despite continued prescribing. Improved characterization of opioid availability before and after surgery is necessary to identify these successes. OBJECTIVE: To evaluate the association between spine surgery and modification of opioid availability postoperatively by using consistent definitions to classify opioid availability before and after surgery. DESIGN, SETTING, AND PARTICIPANTS: This population-based cohort study included 2223 adults (age ≥18 years) who underwent spine surgery in Olmsted County, Minnesota, from January 1, 2005, through December 31, 2016. Data were analyzed from April 1, 2019, to December 1, 2019. EXPOSURES: Preoperative opioid availability based on prescription data in the 180 days before surgery in accordance with Consortium to Study Opioid Risks and Trends (CONSORT) definitions. MAIN OUTCOMES AND MEASURES: Successful modification of opioid availability, defined as an improvement in CONSORT status postoperatively (assessed from 181 to 365 days after surgery) compared with preoperative status, or continued absence of opioid availability for patients with no preoperative availability. Multivariable logistic regression was used to assess the association between preoperative opioid availability and successful modification by 1 year after undergoing surgery. RESULTS: Of 2223 patients included in the study, 1214 were male (54.6%), with a median age of 55 years (interquartile range, 43-68) years. Patients were classified as having no (778 [35.0%]), short-term (1118 [50.3%]), episodic (227 [10.2%]), or long-term (100 [4.5%]) preoperative opioid availability. Of the 2148 patients (96.6%) who were alive at 1 year, postoperative opioid availability was classified as no (1583 [73.7%]), short-term (398 [18.5%]), episodic (104 [4.8%]), and long-term (63 [2.9%]). A total of 1672 patients (77.8%) had successful modification of opioid availability, with success of 83.0% for those with no preoperative availability, 74.9% for those with short-term preoperative availability, 79.8% for those with episodic preoperative availability, and 64.4% for those with long-term preoperative opioid availability. In multivariable analysis, success was significantly associated with preoperative opioid availability (odds ratio [OR] for short term, 0.61 [95% CI, 0.48-0.77]; OR for episodic, 0.95 [95% CI, 0.64-1.40]; OR long term, 0.49 [95% CI, 0.30-0.82]; P < .001 overall vs no availability). CONCLUSIONS AND RELEVANCE: In this study, when following standardized CONSORT definitions, 4 of 5 adults undergoing spine surgery in a population-based cohort met the criteria for a successful pattern of postoperative opioid prescribing. Similar methods to objectively assess changes in opioid prescribing may be clinically useful in other perioperative settings.