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


The role of the healthcare system in the opioid epidemic has come under scrutiny, and surgeons have sought to find balance between the risk of opioid dependence and sufficient management of acute postoperative pain.[13–15] Pain management has emerged as an innovative aspect of patient care, and the development of prescribing guidelines and alternative analgesic therapies has created a culture of awareness in the prevention and treatment of opioid dependence.[16,17] Orthopaedic surgical patients are at a particularly high risk of long-term or harmful opioid consumption.[5] New persistent use among previously opioid-naive postoperative patients represents a poor outcome and has been reported to occur at rates between 1.4% and 8.2% after various general and orthopaedic surgical procedures.[8,9,13,14,18–20] However, the rates and risk factors for new persistent opioid use after orthopaedic foot and ankle procedures are less thoroughly understood. Previous studies, including two large retrospective analyses by Gossett et al and Finney et al, similarly found high rates of new persistent opioid use among patients undergoing foot and ankle procedures.[13,14] However, these studies surveyed nationwide insurance claims databases from 2009 to 2016 and 2010 to 2015, respectively. As such, they were unable to quantify variables not included in administrative databases such as alcohol consumption or recreational drug use—variables found to be critically important in the present analysis. Moreover, the opioid landscape is changing rapidly. With most of its procedures occurring nearly a decade later than those in the prior studies, the present study is more up to date and better captures recent shifts in policy and priority.

In the present study, the rate of new persistent opioid use after orthopaedic foot and ankle procedures was 8.9%. The strongest risk factor for new persistent opioid use was recreational drug use, which presented a threefold increase in risk for new persistent use. Other risk factors included greater perioperative opioid prescription, age ≥40 years, and consumption of ≥6 alcoholic beverages per week, each of which presented about a twofold increase in risk.

At 8.9%, the rate of new persistent opioid use after the foot and ankle procedures in the present study is high compared with the limited results reported for other orthopaedic procedures. For example, Inacio et al[15] retrospectively examined 9,525 patients to determine chronic opioid use rates before and after total hip arthroplasty. They concluded that the rate of new persistent opioid use was 2.1% (190 patients of 8,932 patients who were opioid naive before surgery). Similarly, Sun et al[9] characterized the risk of persistent opioid use in opioid-naive patients after 1 of 11 procedures from various surgical subspecialties. Their results indicated that the rate of new persistent opioid use ranged from 0.1% (for cesarean delivery) to 1.4% (for total knee arthroplasty). Of note, the results of the present study are in tandem with one study on the subject from the foot and ankle literature: Gossett et al[14] used a nationwide insurance claims database to identify opioid-naive patients who underwent surgical treatment of an unstable ankle fracture. The rate of persistent opioid use among opioid-naive patients was 8.8%. Thus, orthopaedic foot and ankle surgeons should be aware of the relatively high risk of opioid dependence after their procedures.

Recent evidence from other surgical specialties supports the idea that greater perioperative opioid prescription increases the risk for postoperative dependence. For example, Deyo et al[21] examined 536,767 postoperative patients who filled an opioid prescription as tracked in their state's prescription monitoring program. They concluded that prescriptions of less than 120 MMEs were associated with a 2.0% rate of persistent use, whereas prescriptions of greater than 120 MMEs were associated with a greater than 3.5% rate of persistent use. Moreover, the proportion of long-term use increased consistently with the increase of MME across the spectrum, up to 37.5% among patients receiving 2,400 to 3,199 MMEs, and 46.2% among patients receiving 3,200 to 3,999 MMEs. Similarly, Finney et al[13] conducted a retrospective review of 36,562 patients who underwent surgical treatment of hallux valgus and found that opioid-naive patients prescribed ≥337.5 MMEs perioperatively were 1.25 times more likely to become persistent users. Relatedly, a retrospective review of 1,353,902 postoperative patients was conducted by Shah et al[22] They used an insurance claims database and identified opioid-naive patients with no history of cancer or other substance abuse. The results of their study suggested that a perioperative dose as small as 90 MME could cause opioid-naive patients to become persistent users.[22] In general, studies have concluded that perioperative opioid dose is the single most-modifiable risk factor for becoming opioid dependent after surgery. The present study supports this theory and suggests that orthopaedic foot and ankle surgeons have the potential to limit postoperative dependence by minimizing perioperative prescription dose.

The strongest risk factor for postoperative opioid dependence in the present study was recreational drug use, which presented a threefold increase in risk for new persistent use. Although only 4.3% of the opioid-naive cohort reported using recreational drugs (n = 15), over one-quarter of those patients became new persistent users of prescription opioids in the postoperative period. This association is consistent with prior literature tying use of other recreational drugs to opioid use.[9,10,18] Therefore, the importance of a thorough substance use history, particularly among opioid-naive surgical patients, cannot be overstated.

It was interesting that older patients (aged ≥40 years) were at greater risk for new persistent use. This is in tandem with the literature.[9,21,22] Older individuals tend to have more chronic pain generators and the use of the opioids postoperatively may provide the extra benefit of helping to ease additional sources of pain distinct from that for which the operation is taking place.[23] In addition, we speculate that older patients have more accumulated experiences involving trauma, including loss of loved ones and loss of social roles, potentially contributing to anxiety and depression. These challenges, although obviously not limited to the older cohort, can cause older patients, as well as people of all ages, to become more susceptible to relying on opioids to help ease emotional and physical pain. It is also important to note that metabolism of narcotic medication decreases with age. Therefore, drug dependence in the elderly may result from lower doses compared with the general cohort.[24]

Finally, patients with a history of ≥6 alcoholic beverages per week were at risk for new persistent use. Alcohol abuse has long been associated with increased opioid abuse in patients who have chronic pain.[25,26] However, only recently have studies shown that alcohol abuse is also a risk factor for new persistent use after surgery.[9,18] It is important to recognize and properly manage any alcohol misuse in patients when prescribing opioids.

The results of this study should be considered within the context of its limitations. First, the preoperative and postoperative time periods selected for analysis were constrained by the scope of data available through the statewide PMP in this retrospective review. Our state's PMP provides data from only the most recent 12 months, giving us a short window to analyze 6 months of both preoperative and postoperative opioid use around the time of surgery. Within these constraints, and given the retrospective nature of the study, the sample size was based on the total number of patients among the four foot and ankle surgeons at our large urban medical center over a 6-month period. This sample size is small compared with those in some of the studies referenced, which were large administrative database studies, which have their own limitations compared with ours. Nevertheless, we recognize the limitations of our sample size, and with greater power, we may have detected additional factors predictive of new persistent opioid use. Second, the statewide PMP tracks only dispensations occurring within the state. Although we excluded out-of-state residents to minimize the numbers of patients who might have sought to fill a prescription out of state, out-of-state dispensations may still have occurred on a limited basis among our cohort. Third, the available data did not enable monitoring of actual narcotic consumption, rather only the filling of narcotic prescriptions. Hence, if a patient took someone else's narcotic medication or his or her own narcotic medication that he or she had left over from a prior prescription, this would have gone undocumented. Similarly, if narcotic medication was filled but not consumed, this could have resulted in the inappropriate categorization of a patient as having persistent use. Fourth, self-reported substance use may be underestimated in this cohort as cohort surveys of drug use typically demonstrate underreporting of true consumption rates.[27–30] Limited data in this area resulted in the grouping of all patients under a single category of recreational drug use rather than stratifying by substance type, which could be potentially problematic.[31] Fifth, patients who recreationally used opioids before surgery without a prescription (eg, heroin or prescription opioids under someone else's name) should have in theory been pushed out of the cohort for having preoperative use; however, no patient reported recreational use of narcotic pain medication in the present study. The possibility of unreported prior opioid use and/or dependency therefore represents a potential confounding factor that may have altered rates of new persistent opioid use within this patient cohort. Sixth, although we studied patients who were on medications for psychiatric disorders, not all patients with depression or other psychiatric disorders are on medication. Patients with psychiatric disorders not on medications would not have been captured. Finally, this limited analysis of drug use did not account for polysubstance users, where the complex interaction of opioids with multiple substances may carry more potential for harm. Additional research is needed to uncover both accurate rates of drug use and the scope of substances being consumed to fully understand substance use as a risk factor for new persistent opioid use among orthopaedic surgical patients. Ultimately, this requires a shift in medical culture in which patients feel comfortable disclosing full and honest accounts of their drug use and where physicians are appropriately educated and trained in a harm reductionist approach with respect to substances.[32,33]