Patient-level patterns in daily prescribed opioid dosage in single level lumbar fusion are associated with postoperative opioid dosage and adverse events: a retrospective analysis of claims data

Background: Patients undergoing lumbar spine surgery have high rates of preoperative opioid use, which is associated with inferior outcomes and higher risks for opioid dependency postoperatively.

Purpose: Determine whether there are identifiable subgroups of patients that follow distinct patterns in pre- and postoperative opioid dosing. Examine how preoperative patterns in opioid dosing relate to postoperative opioid patterns, opioid cessation, and the risk for adverse events.

Study design/setting: Retrospective analysis of an administrative claims database (Meritive^TM Marketscan® Research Databases 2007-2015).

Patient sample: The 9,768 patients undergoing primary single level lumbar fusion.

Outcome measures: Primary: daily morphine milligram equivalent (MME) opioid dosing calculated from prescriptions dispensed for 1 year before and after surgery; secondary: 90-day all-cause readmission and complications, 90-day acute postoperative pain, 90-day and 1-year reoperation, surgical costs, length of stay, and discharge disposition.

Methods: Distinct patient subgroups defined by patterns of daily MME pre- and postoperatively were identified via group-based trajectory modeling. Associations between these groups and outcomes were assessed with multivariable logistic regression with risk adjustment for patient and surgical factors.

Results: Among primary single level lumbar fusion patients, 59.5% filled an opioid prescription in the 3 months preceding surgery, whereas 40.5% were opioid naïve (Naïve). Five distinct subgroups of daily MME were identified among those filling opioids preoperatively: (1) Naïve to 3m (21.2% of patients): no opioids until 3 months preoperatively, escalating to 15 MME/day; (2) Low to 3m (11.4%): very low or as needed dose until 3 months preoperatively, escalating to 15 MME/day; (3) 6m Rise (6.9%): no opioids until 6 months preoperatively, escalating to >30 MME/day; (4) Medium (9.8%): increased linearly from 10 to 25 MME/day across the year before surgery; (5) High (10.0%): increased linearly from 60 to >80 MME/day across the year before surgery. These five preoperative opioid groups were related to postoperative opioids filled in a dose-response manner. The two preoperative patient groups with chronic Medium to High-dose opioid dosing were associated with increased adverse events, including all-cause readmission, reoperation, and pneumonia, whereas a low baseline group with a large, earlier preoperative rise in opioid dosing (6m Rise) had increased encounters for acute postoperative pain. Postoperatively, only 9.5% of patients did not fill an opioid prescription. Five distinct postoperative subgroups were identified based on their patterns in daily MME: Two groups ceased filling opioids within the year following surgery (33.6% of patients), and three groups declined in opioid dosage following surgery but plateaued at low (0-5 MME/day, 29.1%), medium (10-15 MME/day, 12.0%), or high (70-75 MME/day), 13.1%) doses by 1 year. Patients within the higher preoperative opioid groups were more likely to belong to the postoperative groups that were unable to cease filling opioids.

Conclusions: Identification of a patient's preoperative time trend in daily opioid use may provide significant prognostic value and help guide pain management and risk reduction efforts.

Level of evidence: III.