Secondary Logo

Journal Logo

Pain Medicine: Research Report

Prolonged Opioid Use After Knee Arthroscopy in Military Veterans

Rozet, Irene MD*†; Nishio, Isuta MD, PhD*†; Robbertze, Reinette MBChB, FANZCA*; Rotter, Douglas BS, JD; Chansky, Howard MD; Hernandez, Adrian V. MD, PhD

Author Information
doi: 10.1213/ANE.0000000000000292
  • Free

The development of chronic (or persistent) postsurgical pain, also known as postoperative chronic pain syndrome, has been recognized in the last decade as a serious postoperative complication, leading to disability after elective surgery.1–3 Depending on the type of surgery, the incidence of chronic postoperative pain (CPP) may be as high as 60%.4 Multiple risk factors for the development of chronic pain have been identified, including the intensity of immediate postoperative pain.1,2

Posttraumatic Stress Disorder (PTSD) is an anxiety disorder that develops after severe psychological trauma.5 PTSD impairs the quality of life of patients and their families, particularly in young, working patients.6,7 The reported prevalence of PTSD in both civilian and military populations ranges from 1.8 % to nearly 50%, depending on the studied subgroup and methods used for diagnosis.8–13 The incidence of PTSD in the United States Army veterans of Operation Enduring Freedom and Operation Iraqi Freedom (OEF/OIF) may be as high as 40%.14,15

PTSD is associated with the development of chronic pain in up to 60% of individuals diagnosed with PTSD.16 Chronic pain leads to further physical and mental debilitation in patients with PTSD.9,16–18

We hypothesized that PTSD may contribute to CPP in young veterans after elective surgery. The purpose of this study was to (1) identify risk factors for CPP after minor orthopedic surgery in veterans, using prolonged postoperative opioid prescriptions (PPOP) longer than 3 months after surgery as a surrogate for CPP and (2) specifically characterize the extent to which PTSD was a risk factor for CPP.


Study Design

This was a retrospective cohort study approved by the IRB of the Department of Veterans Affairs, Puget Sound Health Care System (VA PSHCS), Seattle, WA, with the waiver of the informed consent to review medical records.

Subjects and Setting

Our study population included United States military veterans, women and men aged 18 to 50 years, who had undergone elective ambulatory knee arthroscopic surgery at the single 500-bed Veterans Affairs’ Tertiary Medical Center (Seattle Division, VA PSHCS), the largest referral medical center of the Veterans Integrated Service Network 20 (includes 8 medical centers serving 23% of the United States territory; Washington, Idaho, Oregon, and Alaska). We chose knee arthroscopy because it is the most common elective ambulatory surgery in the young subgroup of veterans in our institution.


We abstracted patterns of perioperative prescription of opioids from the pharmacy records and clinical notes. We categorized perioperative opioid use as (1) “opioid prescription before surgery” in cases where opioids were prescribed in a period of 3 months before surgery regardless of the duration, and (2) ”prolonged postoperative opioid prescription (PPOP)” in cases where opioids were prescribed uninterruptedly longer than 3 months after the surgery; PPOP was used as a surrogate for CPP.

Inclusion Criteria and Data Sources

We used the VA electronic Computerized Patient Record System to (1) create a list of eligible cases based on all knee arthroscopic procedures in the period between January 1, 2007, and October 1, 2010, using Current Procedural Terminology Codes 29870-29889 and (2) review medical records of subjects. Our exclusion criteria were (1) subjects older than 50 years on the day of surgery, (2) repeat knee surgery on the same leg, (3) previous knee arthroscopy on the contralateral leg within study period, (4) repair of an anterior or posterior cruciate ligament, and (5) any surgical procedure within 6 months before the knee arthroscopy.

During the study period, all anesthesia records in our institution were electronic. Recovery room records were initially charted manually and then scanned as PDF files in the Computerized Patient Record System. We assessed the use of opioids by reviewing clinical notes and pharmacy records.

The abstracted data included patient demographics (gender, age, race, and body mass index), ASA physical status class, comorbidities (obstructive sleep apnea, chronic obstructive pulmonary disease, asthma, hypertension, and heart disease), history of traumatic brain injury, site of deployment, and diagnosis of PTSD, which is made by either psychologist or mental health specialist, alcohol, tobacco and recreational drug use, chronic medications; perioperative variables included length and type of surgery and anesthesia, all anesthesia medications, and perioperative use of opioids (intraoperative and recovery room dose of opioids before discharge from the hospital).

Diagnosis of PTSD

We identified patients with PTSD based on either the patient’s “Problem List” or by reviewing clinical notes describing the patient’s mental health and psychological status. Diagnosis of PTSD within the Department of Veterans Affairs is made according to the Department of Veterans Affairs and Department of Defense Clinical Practice Guidelines (VA/DoD Evidence Based Practice; Clinical Practice Guideline: Management of Posttraumatic Stress Disorder. Guideline Summary 2010. Version 2 or We used the screening tools for PTSD as recommended by the guidelines implemented in the Veterans Affairs medical centers, consisting of one of the following screens: (1) Primary Care PTSD Screen (PC-PTSD); (2) PTSD Brief Screen; (3) Short Screening Scale for Diagnostic Statistical Manual of Mental Disorders (Fourth Edition) PTSD; (4) PTSD Checklist (PCL). The diagnosis of PTSD is made and recorded in the Problem List by either a primary care or mental health specialist using Diagnostic Statistical Manual of Mental Disorders (Fourth Edition) PTSD criteria.19 All patients with a diagnosis of PTSD had undergone mental status examination assessment of emotional and cognitive function to identify comorbid psychological or psychiatric conditions, as recommended by the guidelines.


The primary outcome measure was the PPOP, our surrogate measure for CPP.

Statistical Analysis

For power analysis and a sample size calculation, we used an a priori sample size calculation for multiple regression analysis ( based on an α level of 0.05, an effect size of 0.15, and a desired statistical power of 80%. The result was a minimum sample size between 67 (2 predictors) and 103 (7 predictors). We also calculated the sample size for an effect size of 0.02, yielding a sample size between 478 (2 predictors) and 721 (7 predictors). We recognized, a priori, that it would take several years to accrue adequate numbers, given that we perform approximately 40 knee arthroscopies per year. Therefore, we decided to review a 4-year period, which would provide us with approximately 160 cases.

Data are reported as mean ± standard deviation (SD), n (%), or median with interquartile range. Two-sample Student tests and univariate logistic regressions were used for analysis to compare the association of PTSD with preoperative smoking, alcohol use, and opioid use.

We used a classification tree to evaluate the association between patient characteristics and CPP. Classification tree analysis was performed using a recursive partitioning method with the rpart package of R software. SAS 9.2 (Cary, NC) and R software 3.0.1. ( was used for all statistical analyses. We also used multivariable logistic regression based on univariable analysis identifying associations with P < 0.2.

An α level of 0.05 was considered statistically significant. All P values are 2-sided.


One hundred forty-five patients (mean age 39 ± 8 years old, men = 87%, Caucasian = 76%, African American = 15%) were identified. Of 145 patients, only 3 patients were Vietnam War veterans, while the rest of the patients (98%) were veterans of OEF/OIF.

Prevalence of PTSD

In the whole cohort, 47 patients (32%, 95% confidence interval [CI], 25%–41%) had a diagnosis of PTSD. Patients with PTSD had a higher incidence of smoking, preoperative use of opioids, and CPP compared with patients without PTSD. (Table 1) While there was no difference in comorbidities, history of traumatic brain injury, length of surgery, and anesthesia management, patients with PTSD received significantly more opioids in the recovery room, than non-PTSD patients [14 (5–24) vs 8 (01.5–14.0), equivalent to morphine in milligram, median (interquartile range), P = 0.0006].

Table 1
Table 1:
Baseline Characteristics of Veterans With Posttraumatic Stress Disorder (PTSD) and Without PTSD (Non-PTSD)

Chronic Postoperative Pain

CPP measured as PPOP was identified in 30% (43 of 145) of patients, with a 53% incidence of PTSD (23 of 43 patients, 95% CI, 38%–69%) (Table 2).

Table 2
Table 2:
Comparative Data Between Patients With Chronic Postoperative Pain (CPP) and Without CPP (Non-CPP)

By univariable analysis, any preoperative opioid prescription (univariable P < 0.0001), and PTSD (univariable P = 0.0004) were found to be associated with CPP (Table 2).

In the subgroup of patients who were not using opioids chronically, longer than 3 months before surgery (n = 101), only 8% (8 of 101) developed CPP compared with 80% of chronic opioid users (35 of 44) (univariable P < 0.0001). In nonchronic opioid users, there was a strong univariable association between PTSD and CPP (P = 0.001; univariable odds ratio 10.3, 95% CI, 1.9–54.8).

The classification tree analysis depicting risk factors for CPP appears in Figure 1. At the top of the tree there is the most important predictor of CPP: preoperative opioid use (multivariable odds ratio 65.3, 95% CI, 14.5–293.0). The branches at the left half of the figure depict the subgroup of patients who were not taking opioids before surgery. In this subgroup, the next most important predictors of chronic pain were age >27.5 years and PTSD. The branches on the right half of the figure depict the subgroup of patients who were taking chronic opioids before surgery. In this subgroup, the next most important predictors of chronic pain were a diagnosis of arterial hypertension and nonsmoking status; PTSD was not an additional risk factor for CPP in veterans who were taking opioids preoperatively.

Figure 1
Figure 1:
Classification tree for chronic postoperative pain. The tree displays risk factors associated with chronic postoperative pain, from those of highest importance (top of the tree) to least important (bottom of the tree) in this study. Predictors with P < 0.2 in the univariable analysis (Table 2) were included in the classification tree analysis. The numbers at the bottom represent the cumulative fraction of subjects in each classified subgroup as read from left to right. The smallest number of patients at each branch (subgroup) was 5; “preoperative opioids” = preoperative prescription of opioids within 3 months before surgery; “osa” = obstructive sleep apnea.


This is the first study to evaluate the risk factors for CPP in veterans and address the contribution of PTSD. Our data demonstrate that the primary predictor of CPP, measured using the surrogate PPOP, is preoperative opioid use. Other risk factors may include age, hypertension, nonsmoking, and obstructive sleep apnea. PTSD likely is a risk factor for CPP among veterans who were not taking chronic opioids before surgery.

Data of our cohort of OEF/OIF veterans younger than 51 years undergoing minor elective surgery (knee arthroscopy) demonstrate that (1) preoperative prescription of opioids is the strongest predictor of CPP when measured as a prolonged (longer than 3 months after surgery) opioid prescription; (2) the prevalence of PTSD is at least 32% (95% CI, 25%–41%) and is associated with smoking, an increased preoperative and postoperative opioid prescription rate and an increased demand for opioids in the immediate postoperative period, and (3) veterans suffering PTSD who are not using opioids chronically are at higher risk for prolonged, uninterrupted opioid prescription after surgery, compared with those veterans without a diagnosis of PTSD and CPP.

Development of chronic postsurgical pain, CPP syndrome, has been associated with multiple factors, including comorbid inflammatory diseases,1,2 preexisting chronic pain before surgery, and hyperalgesia with chronic use of opioids,20 a high intensity of surgical pain, and acute postoperative pain in the recovery room.1–3 In addition, acute postoperative pain may depend on demographic characteristics such as genetics, length of surgery,1–3 and type of anesthesia.20 Regional and neuraxial analgesia use along with multimodal nonopioid analgesia has been shown to have an antihyperalgesic effect.20 Luckily, in our cohort, there were no differences in demographics and anesthesia management when comparing patients with CPP with those without. Our data showing a strong connection between preoperative use of opioids and development of CPP in a highly selective young cohort of patients undergoing minor elective procedures, on one hand, strengthen the previously reported relationship between preexisting chronic pain and postsurgical pain; however, it poses a question regarding efficacy and/or timing of the surgical procedure. Undoubtedly, development of hyperalgesia and opioid tolerance contributes to development of CPP syndrome. Whether earlier surgical intervention rather than initiation of opioid analgesia before surgery might minimize the risks for postoperative pain is unclear, and it would require a large prospective investigation to answer this question.

Our finding of an association between chronic use of opioids and PTSD is not surprising and is in full agreement with previous data.14 The largest national-level survey involving 291,205 OEF/OIF veterans within the Veteran Affairs health care system demonstrated that veterans with PTSD have the highest opioid prescription rate when compared with a subgroup without mental disorders and with a subgroup with mental disorders other than PTSD.14 The high opioid prescription rate was strongly associated with the higher risk for poor outcome, including opioid overdose and self-inflicted injuries.14 This may, however, reflect a nationwide problem with the shift to liberal prescription of opioids in the last 2 decades.21

To the best of our knowledge, neither pain perceptionr nor opioid requirement in the acute perioperative and postoperative periods has been described in patients suffering PTSD. These, however, are very important questions to clarify in the PTSD population that is more susceptible to developing chronic pain. There are conflicting data regarding acute perception of pain and causative interrelations between PTSD and chronic pain. On one hand, the phenomenon of decreased pain perception under stress, so called stress-induced analgesia, has been recognized and proven to play a significant role in PTSD.22,23 This contradicts the widely accepted concept of comorbidly enhancing interplay between psychological stress and pain, the concept primarily based on the phenomenon of stress-induced hyperalgesia.”24 However, opioid consumption further complicates PTSD/pain interactions. Although acute analgesia with a prescription of opioids (morphine) in injured combat soldiers has been suggested to have a possible protective effect against later development of PTSD,15 the chronic use of opioids could lead to opioid-induced hyperalgesia,25 resulting in decreased pain tolerance and increased opioid demand with chronic, and particularly, acute pain.

Aside from preexisting PTSD, surgical stress and emotional stress (such as fear of surgery) might influence pain perception and potentially cause either stress-induced analgesia or stress-induced hyperalgesia.24 To minimize heterogeneity of surgical stress and pain, we intentionally chose knee arthroscopy, a low pain intensity surgery, and excluded lengthy and more painful repair of knee ligaments. Considering that opioid requirements in the operating room are objectively guided by vital signs, whereas in the recovery room, they are guided by the subjective visual analog scale score and pain behavior, these data may reflect less tolerance to acute postoperative pain and/or exaggerated pain behavior in our PTSD patients. A large prospective study would require, however, proving this assumption.

Our study is limited by its retrospective nature and small sample size. Retrospectively gathered data may affect the accuracy of estimation of the overall prevalence of PTSD and, in some cases, of opioid prescription. Although we analyzed opioid prescription data as a surrogate for opioid consumption, this may rather reflect providers’ practice in response to patients’ reactions and behaviors. However, because we studied a veteran population seeking medical help within the Veteran Affairs health care system only, our findings should be carefully extrapolated to the entire veteran community. A small number of events along with the small sample size are another drawback of our study. Particularly in our cohort, where the frequent association between chronic opioid use and PTSD was not surprising, and in the context of an expected strong association between preoperative and postoperative opioid use, the preoperative use of opioids affected the statistical significance of PTSD. Our study design, however, has some advantages because it allowed us to investigate a highly selective subgroup of patients subjected to consistent surgical and anesthesia strategies because the study was over 4-year period within 1 institution only.

In conclusion, our data indicate that young veterans consuming opioids have a higher risk of CPP and PPOP after minor elective surgery, and veterans suffering PTSD who are not using opioids before surgery have a higher risk of developing CPP. Further prospective research is warranted to support our data and to elucidate optimal surgical and anesthesia strategies to minimize perioperative pain, stress, and optimal opioid use in the perioperative setting.


Name: Irene Rozet, MD.

Contribution: This author helped design and conduct of the study, data collection, data analysis, and manuscript preparation.

Attestation: Irene Rozet approved the final manuscript and is the archival author.

Name: Isuta Nishio, MD, PhD.

Contribution: This author helped design and conduct of the study, data collection, data analysis, and manuscript preparation.

Attestation: Isuta Nishio approved the final manuscript.

Name: Reinette Robbertze, MBChB, FANZCA.

Contribution: This author helped conduct of the study, data collection, and manuscript preparation.

Attestation: Reinette Robbertze approved the final manuscript.

Name: Douglas Rotter, BS, JD.

Contribution: This author helped in data collection and manuscript preparation.

Attestation: Douglas Rotter approved the final manuscript.

Name: Howard Chansky, MD.

Contribution: This author helped manuscript preparation.

Attestation: Howard Chansky approved the final manuscript.

Name: Adrian V. Hernandez, MD, PhD.

Contribution: This author helped data analysis and manuscript preparation.

Attestation: Adrian V. Hernandez approved the final manuscript.

This manuscript was handled by: Steven L. Shafer, MD.


1. Kalkman CJ, Visser K, Moen J, Bonsel GJ, Grobbee DE, Moons KG. Preoperative prediction of severe postoperative pain. Pain. 2003;105:415–23
2. Janssen KJ, Kalkman CJ, Grobbee DE, Bonsel GJ, Moons KG, Vergouwe Y. The risk of severe postoperative pain: modification and validation of a clinical prediction rule. Anesth Analg. 2008;107:1330–9
3. Grosu I, de Kock M. New concepts in acute pain management: strategies to prevent chronic postsurgical pain, opioid-induced hyperalgesia, and outcome measures. Anesthesiol Clin. 2011;29:311–27
4. Perkins FM, Kehlet H. Chronic pain as an outcome of surgery. A review of predictive factors. Anesthesiology. 2000;93:1123–33
5. American Psychiatric Association. Diagnostic and Statistical Manual of Mental Disorders. 2000 Washington, DC American Psychiatric Association
6. Mansfield AJ, Kaufman JS, Marshall SW, Gaynes BN, Morrissey JP, Engel CC. Deployment and the use of mental health services among U.S. Army wives. N Engl J Med. 2010;362:101–9
7. Mansfield AJ, Kaufman JS, Engel CC, Gaynes BN. Deployment and mental health diagnoses among children of US Army personnel. Arch Pediatr Adolesc Med. 2011;165:999–1005
8. Gillespie CF, Bradley B, Mercer K, Smith AK, Conneely K, Gapen M, Weiss T, Schwartz AC, Cubells JF, Ressler KJ. Trauma exposure and stress-related disorders in inner city primary care patients. Gen Hosp Psychiatry. 2009;31:505–14
9. Phifer J, Skelton K, Weiss T, Schwartz AC, Wingo A, Gillespie CF, Sands LA, Sayyar S, Bradley B, Jovanovic T, Ressler KJ. Pain symptomatology and pain medication use in civilian PTSD. Pain. 2011;152:2233–40
10. Sundin J, Fear NT, Iversen A, Rona RJ, Wessely S. PTSD after deployment to Iraq: conflicting rates, conflicting claims. Psychol Med. 2010;40:367–82
11. Fear NT, Jones M, Murphy D, Hull L, Iversen AC, Coker B, Machell L, Sundin J, Woodhead C, Jones N, Greenberg N, Landau S, Dandeker C, Rona RJ, Hotopf M, Wessely S. What are the consequences of deployment to Iraq and Afghanistan on the mental health of the UK armed forces? A cohort study. Lancet. 2010;375:1783–97
12. Smith TC, Wingard DL, Ryan MA, Kritz-Silverstein D, Slymen DJ, Sallis JFMillennium Cohort Study Team. . PTSD prevalence, associated exposures, and functional health outcomes in a large, population-based military cohort. Public Health Rep. 2009;124:90–102
13. McNally RJ. Are we winning the war against posttraumatic stress disorder. Science. 2012;336:872–4
14. Seal KH, Shi Y, Cohen G, Cohen BE, Maguen S, Krebs EE, Neylan TC. Association of mental health disorders with prescription opioids and high-risk opioid use in US veterans of Iraq and Afghanistan. JAMA. 2012;307:940–7
15. Holbrook TL, Galarneau MR, Dye JL, Quinn K, Dougherty AL. Morphine use after combat injury in Iraq and post-traumatic stress disorder. N Engl J Med. 2010;362:110–7
16. Sharp TJ, Harvey AG. Chronic pain and posttraumatic stress disorder: mutual maintenance? Clin Psychol Rev. 2001;21:857–77
17. Asmundson GJ, Coons MJ, Taylor S, Katz J. PTSD and the experience of pain: research and clinical implications of shared vulnerability and mutual maintenance models. Can J Psychiatry. 2002;47:930–7
18. Beck JG, Clapp JD. A different kind of co-morbidity: Understanding posttraumatic stress disorder and chronic pain. Psychol Trauma. 2011;3:101–8
19. American. Psyciatric Association.Diagnostic and Statistical Manual of Mental Disorders. 20004th ed Washington, DC: American Psychiatric Association;
20. Richebé P, Beaulieu P. Perioperative pain management in the patient treated with opioids: continuing professional development. Can J Anaesth. 2009;56:969–81
21. Phillips DP, Barker GE, Eguchi MM. A steep increase in domestic fatal medication errors with use of alcohol and/or street drugs. Arch Intern Med. 2008;168:1561–6
22. Lewis JW, Cannon JT, Liebeskind JC. Involvement of central muscarinic cholinergic mechanisms in opioid stress analgesia. Brain Res. 1983;270:289–93
23. Pitman RK, van der Kolk BA, Orr SP, Greenberg MS. Naloxone-reversible analgesic response to combat-related stimuli in posttraumatic stress disorder. A pilot study. Arch Gen Psychiatry. 1990;47:541–4
24. Geuze E, Westenberg HG, Jochims A, de Kloet CS, Bohus M, Vermetten E, Schmahl C. Altered pain processing in veterans with posttraumatic stress disorder. Arch Gen Psychiatry. 2007;64:76–85
25. Angst MS, Clark JD. Opioid-induced hyperalgesia: a qualitative systematic review. Anesthesiology. 2006;104:570–87
© 2014 International Anesthesia Research Society