Our study cohort of 32,874 veterans (Fig. 2) had 11,747 patients (36%) with lower acute pain scores (≤4/10) and 21,127 patients (64%) with higher acute pain scores (>4/10). Chronic significant postoperative opioid use (mean dose, >30 mg/d) was present in 21% of the overall cohort (n = 6856). Opioid prescriptions at discharge exceeded 30 days in more than half of all patients and exceeded 30 mg/d in approximately 3 of 5 patients. Overall, comparing postoperative versus preoperative categories, a majority of patients (54%) had no significant change, a modest percentage (28% [n = 9170]) decreased dose, whereas a smaller percentage (18% [n = 6034]) escalated dose. Comparing baseline characteristics in the 2 exposure groups (Appendix 7, Supplemental Digital Content 7, http://links.lww.com/AAP/A264), the following differences were notable. Patients with lower acute pain were older, had surgery performed in earlier years (2010–2011 vs 2014–2015), had significantly lower preoperative pain, had less preoperative opioids (both dose and duration of use) and more tramadol use, were less likely to visit pain clinics, and had significantly lower rates of back pain, substance abuse, tobacco or alcohol abuse, and lower rates of psychiatric illness. Percentage of patients with chronic significant postoperative opioid use was 11% versus 26%, respectively, with lower versus higher acute pain. Opioid prescriptions at discharge exceeded 30 days and 30 mg/d in well under half of the patients with lower pain scores (Appendix 3, Supplemental Digital Content 3, http://links.lww.com/AAP/A260).
In the propensity score model (Appendix 4, Supplemental Digital Content 4, http://links.lww.com/AAP/A261), multivariable logistic regression showed that lower acute pain was more likely with later preoperative initiation of opioids, use of tramadol, Southeast or Central US regions (vs the Pacific), Hispanic race, age older than 60 years, obesity, TKA in facilities in the lower (vs intermediate) volume tercile or by surgeons in the lower or higher (vs intermediate) volume tercile, receipt of PCA, and nerve blocks. Conversely, lower acute pain was less likely with higher preoperative opioid doses (>15 mg/d over the year before admission), greater preoperative pain, TKA in later years (2012–2013 or 2014–2015 vs 2010–2011) or in the Midwest (vs the Pacific), black race, chronic pulmonary disease, tobacco or substance abuse, psychiatric illnesses, back pain, or receipt of general anesthesia. Model parameters were good (area under receiver operating characteristic curve ~71% and satisfactory goodness of fit, Appendix 5, Supplemental Digital Content 5, http://links.lww.com/AAP/A262), so we used the model to compute propensity scores for each patient. Matching patients with a 1:1 greedy match algorithm, we identified 10,463 pairs of patients for inclusion in the propensity-matched group (Fig. 2). Reexamining baseline differences in this propensity-matched cohort (Table 1), in nearly two-thirds of the study cohort (20,926 of 32,874 patients), we saw that all clinically significant differences were removed. In other words, propensity matching effectively allowed the comparison of postoperative outcomes to be reflective of differences in acute pain scores (Fig. 2).
In the large propensity-matched cohort (Fig. 2 and Table 2), we found that the percentage of patients with chronic significant postoperative opioid use (mean dose, >30 mg/d),18 was 12% in those with lower acute pain scores versus 16% in those with higher scores, P < 0.001. Discharge prescriptions at more than 30 days and greater than 30 mg/d were filled by 50% and 58% of patients with higher pain scores (vs 42% and 47%, respectively, by patients with lower pain scores, P < 0.001). In terms of change in dose categories, mean postoperative opioid use increased in 17% of patients with lower acute pain versus in 21% of patients with higher acute pain, P < 0.001. Conversely, a larger percentage of patients with lower acute pain scores, 29%, decreased postoperative (vs preoperative) dose when compared with patients with higher acute pain (26%), P < 0.001. Taken together, the overall difference in change in chronic opioid use in absolute terms was 7% (4% difference in dose escalation plus a 3% difference in dose decrease). In terms of the numbers needed to treat, this meant that for every 15 patients reporting lower versus higher acute pain there was 1 more patient with a desirable change in chronic postoperative opioid use. Lengths of stay were longer than 3 days in 51% of patients with higher scores and 47% of patients with lower scores, P < 0.001. Sensitivity analysis (Appendix 6, Supplemental Digital Content 6, http://links.lww.com/AAP/A263) showed that acute pain scores reduced the odds of an increase in postoperative dose category by 15% (odds ratio, 0.85; 95% confidence interval [CI] 0.80–0.91), independent of all other factors including discharge prescriptions. A larger daily dose (>30 mg/d) in the discharge prescription was more important (odds ratio, 1.68; 95% CI, 1.55–1.81) than supply for more days (>30 days; odds ratio, 1.25; 95% CI, 1.16–1.35).
We examined the association between acute postoperative pain and chronic opioid use in a large population at significant risk, preoperatively opioid-using veterans undergoing TKA. Prior studies have looked at opioid-naive and nonnaive patient populations undergoing surgery and its direct correlation with chronic postoperative opioid use and found much higher incidences of persistent opioid use in those patients taking opioids prior to surgery.11,14 Given these prior findings, we focused on preoperative opioid users because the study outcome of chronic postoperative opioid use is more common in this population. Hence, the effects of variations in acute pain can be better examined. Comparing rates of chronic use beyond 3 months after discharge, at mean doses exceeding 30 mg/d (risk of overdose death in veterans),18 in patients with lower versus higher acute postoperative pain scores, we found that patients with lower acute pain had significantly less chronic opioid use when compared with similar patients who had higher acute pain scores. Furthermore, postoperative dose categories were more likely to have decreased and less likely to have increased (vs preoperative categories) in patients with lower acute pain scores. The discharge prescription was also significantly smaller with lower acute pain scores, but the association between acute pain and chronic opioid use was independent of the duration and dose of the discharge prescription.
Particularly important is our finding that patients with lower acute pain scores (≤4/10) had a rate of 23% versus 28% in patients with higher acute pain scores (>4/10), for the receipt of a chronic postoperative opioid dose of greater than 15 mg/d in OMEs beyond 3 months after discharge. Amounts less than this are associated with a low risk of adverse events.18 This 5% difference in absolute terms translates to a number needed to treat of 20 patients. In other words, for every 20 patients who has an acute pain score of greater than 4 (rather than ≤4) on the 11-point scale, 1 more patient will go on chronic opioid use at a dose of greater than 15 mg/d. Given the volume of TKAs, the population-level implications of this difference are significant.
These findings are consistent with previous reports11–14,25 and to the best of our knowledge are the first to specifically identify intensity of acute pain as a potentially modifiable risk factor in ongoing efforts to reduce chronic opioid use.7 This is notable for several reasons. We studied patients undergoing TKA: the leading elective inpatient operative procedure.9,13,14 We examined a hypothesis that cannot be directly tested in randomized trials, using rigorous propensity score–based methods to minimize baseline differences before comparing chronic opioid use in large groups of patients with differing acute pain scores. Thus, observed associations are less likely to be an epiphenomenon. Lastly, the favorable exposure—lower acute postoperative pain scores (on average ≤4/10)—is not common. Lower scores were present in only 36% of all patients and declined recently (ie, a smaller proportion reported lower acute pain scores in 2014–2015 vs 2010–2011).
But how can lower acute pain (over 3 days in the hospital) have an effect on chronic opioid use? First, discharge prescriptions are a key pathway. We found that more patients with higher acute pain scores had longer and larger discharge prescriptions (Table 2). It is intuitive that surgeons might be guided by proximal pain scores and opioid use.26,27 However, the link between postoperative prescriptions and chronic use has been recognized only recently.19–21,26 Thus, there may be an opportunity to reduce variations in the dose and duration of discharge prescriptions after surgery—by lowering acute pain.27 Second, the association between acute pain and chronic pain may explain increased chronic opioid use.28 Complex mechanistic pathways, operating in both the peripheral and central nervous systems, have been described in the transition from acute to chronic pain.28 We also found that association (between acute pain and chronic opioid use) to be independent of discharge prescriptions (Appendix 6, Supplemental Digital Content 6, http://links.lww.com/AAP/A263). Hence, lowering acute pain scores may reduce long-term opioid use independent of discharge prescriptions.28 Third, scores reported in the hospital after surgery may represent pain at operative and nonoperative sites.29 Over a third of the propensity-matched patients received continuous nerve blocks (Table 1), but this may not address pain outside the specific areas “blocked” by local anesthetics. Back pain was prevalent in 25% of our cohort (based on ICD diagnosis codes or preoperative treatment or imaging codes) and, when managed with opioids, may predispose to long-term use with increased rates of addiction and loss of efficacy secondary to the emergence of opioid tolerance and opioid-induced hyperalgesia.29 The presence of chronic pain at another site may reconcile apparent differences between our findings and reports showing that nerve blocks did not lower chronic opioid use.30
Our study has several strengths in addition to those described previously. We also found that late initiation of opioids before surgery (in the last 3 preoperative months rather than earlier), lower preoperative opioid doses (<15 mg/d), and use of tramadol increased the likelihood of lower acute pain scores (Appendix 4, Supplemental Digital Content 4, http://links.lww.com/AAP/A261). Nerve blocks had modest benefit on long-term opioid-sparing effects, but consistent with known effects of local anesthetic infusions (vs single injections),30 we observed an increase in acute pain from day 0 to day 1 (Fig. 1), suggesting that nerve blocks when used were likely to be for short durations in our study. Our study also has several limitations. As with a majority of studies pertaining to the veteran population, this study cohort is dominated by a male majority (>90%), so generalizability to a more equal sex-distributed population is limited. Pain scores, as a single number between 0 and 10, do not reflect the biopsychosocial complexity of pain. Unmeasured confounding cannot be ruled out, and an independent factor, such as less pain catastrophizing, may lower both acute pain and chronic opioid use, creating the appearance of an association where none exists. Another limitation is the use of pain scores that simply represent general intensity of pain by patient report on an 11-point scale in the inpatient and outpatient setting without granular information with respect to activity, site, nature, character, and other attributes of pain. Finally, we cannot make causal inferences, and generalizability to non-VA settings may be limited.
IMPLICATIONS AND CONCLUSIONS
We conclude that in a population exposed to preoperative opioid use acute postoperative pain was associated with chronic opioid use after TKA, independent of discharge prescriptions. Scores reported by patients in the first few days after surgery can thus have prognostic value. Whether interventions to reduce acute pain will be effective in reducing chronic postoperative opioid overuse is unclear because confounding cannot be completely ruled out in any observational study. However, our finding provides reasonable support for aggressive efforts to reduce acute postoperative pain. Future prospective studies should look at comprehensive strategies to address both operative and nonoperative pain in ways that may mitigate the long-term use of opioids.
The following individuals contributed to the initial conception, and to certain aspects of data interpretation, in this study: Alan R. Ellis, PhD, assistant professor, Department of Social Work, North Carolina State University, Raleigh, NC; Atilio Barbeito, MD, MPH, and Rebecca Schroeder, MD, MMCi, Department of Anesthesiology, and Juliann Hobbs, MD, MPH, Department of Anesthesiology, Duke University Medical Center/Durham VAMC, Durham, NC; and Edward Mariano, MD, MAS (Clinical Research), Anesthesiology and Perioperative Care Service, VA Palo Alto Health Care System, and professor of Anesthesiology, Perioperative and Pain Medicine Stanford University School of Medicine, Palo Alto, CA.
1. Lorenz KA, Sherbourne CD, Shugarman LR, et al. How reliable is pain as the fifth vital sign? J Am Board Fam Med
2. Terkawi AS, Mavridis D, Sessler DI, et al. Pain management modalities after total knee arthroplasty: a network meta-analysis of 170 randomized controlled trials. Anesthesiology
3. Karlsen AP, Wetterslev M, Hansen SE, Hansen MS, Mathiesen O, Dahl JB. Postoperative pain treatment after total knee arthroplasty: a systematic review. PLoS One
4. Sinatra R. Causes and consequences of inadequate management of acute pain. Pain Med
5. Lindberg MF, Miaskowski C, Rustøen T, et al. The impact of demographic, clinical, symptom and psychological characteristics on the trajectories of acute postoperative pain after total knee arthroplasty. Pain Med
6. Dunwoody CJ, Krenzischek DA, Pasero C, Rathmell JP, Polomano RC. Assessment, physiological monitoring, and consequences of inadequately treated acute pain. J Perianesth Nurs
7. Murthy VH. Ending the opioid epidemic—a call to action. N Engl J Med
8. Sun EC, Darnall BD, Baker LC, Mackey S. Incidence of and risk factors for chronic opioid use among opioid-naive patients in the postoperative period. JAMA Intern Med
9. McDermott KW, Sun R. Trends in hospital inpatient stays in the United States, 2005–2014. HCUP Stat Brief
10. Dowell D, Chou R. CDC guideline for prescribing opioids for chronic pain—United States, 2016. MMWR Recomm Rep
11. Brummett CM, Waljee JF, Goesling J, et al. New persistent opioid use after minor and major surgical procedures in us adults. JAMA Surg
12. Mudumbai SC, Oliva EM, Lewis ET, et al. Time-to-cessation of postoperative opioids: a population-level analysis of the Veterans Affairs Health Care System. Pain Med
13. Bedard NA, Pugely AJ, Westermann RW, Duchman KR, Glass NA, Callaghan JJ. Opioid use after total knee arthroplasty: trends and risk factors for prolonged use. J Arthroplasty
14. Goesling J, Moser SE, Zaidi B, et al. Trends and predictors of opioid use after total knee and total hip arthroplasty. Pain
15. Memtsoudis SG, Poeran J, Cozowicz C, Zubizarreta N, Ozbek U, Mazumdar M. The impact of peripheral nerve blocks on perioperative outcome in hip and knee arthroplasty—a population-based study. Pain
16. Noël PH, Copeland LA, Perrin RA, et al. VHA Corporate Data Warehouse height and weight data: opportunities and challenges for health services research. J Rehabil Res Dev
17. Aspinall SL, Sales MM, Good CB, et al. Pharmacy benefits management in the Veterans Health Administration revisited: a decade of advancements. 2004–2014. J Manag Care Spec Pharm
18. Bohnert AS, Logan JE, Ganoczy D, Dowell D. A detailed exploration into the association of prescribed opioid dosage and overdose deaths among patients with chronic pain. Med Care
19. Hill MV, Stucke RS, Billmeier SE, Kelly JL, Barth RJ Jr. Guideline for discharge opioid prescriptions after inpatient general surgical procedures. J Am Coll Surg
20. Scully RE, Schoenfeld AJ, Jiang W, et al. Defining optimal length of opioid pain medication prescription after common surgical procedures. JAMA Surg
21. Shah A, Hayes CJ, Martin BC. Characteristics of initial prescription episodes and likelihood of long-term opioid use—United States, 2006–2015. MMWR Morb Mortal Wkly Rep
22. Babalonis S, Lofwall MR, Nuzzo PA, Siegel AJ, Walsh SL. Abuse liability and reinforcing efficacy of oral tramadol in humans. Drug Alcohol Depend
23. McDonald DC, Carlson K, Izrael D. Geographic variation in opioid prescribing in the U.S. J Pain
24. Stürmer T, Wyss R, Glynn RJ, Brookhart MA. Propensity scores for confounder adjustment when assessing the effects of medical interventions using nonexperimental study designs. J Intern Med
25. Carroll I, Barelka P, Wang CK, et al. A pilot cohort study of the determinants of longitudinal opioid use after surgery. Anesth Analg
26. Hill MV, McMahon ML, Stucke RS, Barth RJ Jr. Wide variation and excessive dosage of opioid prescriptions for common general surgical procedures. Ann Surg
27. Voscopoulos C, Lema M. When does acute pain become chronic? Br J Anaesth
28. Deyo RA, Von Korff M, Duhrkoop D. Opioids for low back pain. BMJ
29. Sun EC, Bateman BT, Memtsoudis SG, Neuman MD, Mariano ER, Baker LC. Lack of association between the use of nerve blockade and the risk of postoperative chronic opioid use among patients undergoing total knee arthroplasty: evidence from the Marketscan database. Anesth Analg
30. Richman JM, Liu SS, Courpas G, et al. Does continuous peripheral nerve block provide superior pain control to opioids? A meta-analysis. Anesth Analg
Supplemental Digital Content
Copyright © 2018 by American Society of Regional Anesthesia and Pain Medicine.