New Persistent Opioid use After Orthopaedic Foot and Ankle Surgery

A Study of 348 Patients

Emily E. Hejna, MPH; Nasima Mehraban, MD; George B. Holmes, Jr, MD; Johnny L. Lin, MD; Simon Lee, MD; Kamran S. Hamid, MD, MPH; Daniel D. Bohl, MD, MPH


J Am Acad Orthop Surg. 2021;29(16):e820-e825. 

In This Article


Patient Selection and Database

After institutional review board approval, patients undergoing orthopaedic foot or ankle surgery by one of four surgeons at a large urban United States medical center between January 1, 2019 and June 30, 2019 were identified. For each of these patients, charts were reviewed, and the statewide prescription monitoring program (PMP) database was accessed. The PMP is a state-run system to which all dispensations of controlled substances within the state must be reported. The program's website gives prescribers and dispensers the ability to view a patient's controlled substance dispensation history before prescribing or dispensing additional substances.

Among initially identified patients, patients were excluded for (1) preoperative opioid use (defined as filling an opioid prescription between 6 months and 30 days preoperatively), (2) out-of-state residency, (3) revision surgery within the prior 6 months, and (4) diagnosis of Charcot neuropathy or other chronic neuropathic pain. All patients had 6 months of follow-up in the PMP at the time of review.

Baseline and Perioperative Characteristics

For included patients, charts were reviewed to extract demographic information including age, sex, body mass index (BMI), marital status, current smoker status, alcohol consumption, and recreational drug use. The use of nonopioid prescription medications was collected including antidepressants, antiepileptics, benzodiazepines, and muscle relaxants. Comorbidity data were collected and used to calculate the Charlson Comorbidity Index (CCI).[7] Conditions were characterized as acute (symptoms for fewer than 30 days) versus chronic (symptoms for greater than 30 days). Surgical data were recorded, including postoperative diagnosis and procedure performed, which were further categorized as predominantly ankle/hindfoot surgery (versus midfoot/forefoot surgery) and bone surgery (versus soft-tissue only surgery).

The statewide PMP was used to identify nonopioid controlled substance prescriptions in the 6-month preoperative window, including for antidepressants, antiepileptics, antipsychotics, benzodiazepines, muscle relaxants, and nonbenzodiazepine sedative-hypnotics.[8–10] For opioid prescriptions during the perioperative period, the total dispensed morphine milligram equivalents (MMEs) per day were calculated using the conversion factors provided by the Centers for Medicare and Medicaid Services (Supplemental Table 1, These conversions are calculated relative to the potency of morphine and allow for standardization of prescription strength across different types of opioids. A total perioperative MME was calculated for each patient:

Total perioperative MME was categorized as greater (versus less than or equal to) the median value of 160 MME.


The primary outcome measure for this study was new persistent opioid use, defined as the filling of an opioid prescription between 2 and 6 months postoperatively. The definition and rates of new persistent opioid use have varied widely in the literature, ranging from as early as 2 months and up to 1 year postoperatively.[11] These criteria seem to differ widely among different types of surgery as well, for example, orthopaedics versus gynecologic surgery.[12] We decided to take a more liberal approach to the definition of new persistent use at 2 and 6 months postoperatively given the smaller size of this cohort compared with other studies. However, a standardized definition of new persistent opioid use would be beneficial in comparing outcomes across studies. The filling of opioid prescriptions written by our or any other institutions or providers was counted toward new persistent use, as long as the dispensation occurred during that specified time period.

Statistical Analysis

Statistical analyses were conducted in Stata Version 16.6. T-tests (for continuous variables) and Pearson chi-squared tests (for categorical variables) were used to compare baseline characteristics. Persistent opioid use was tested for association with each baseline characteristic using Pearson chi-squared tests. The level of significance was set at P < 0.05.