Orthopedics

Feature Article 

Preoperative Sedative Use and Other Risk Factors for Continued Narcotic Use After Total Knee Arthroplasty: A Comprehensive Analysis of a Mandatory Database

Zachary T. Wilt, MD; Timothy L. Tan, MD; Alexander J. Rondon, MD, MBA; William L. Wang, MD; Carol Foltz, PhD; Pedro K. Beredjiklian, MD; William V. Arnold, MD

Abstract

Opioids are used for pain control after total knee arthroplasty (TKA) and carry risk for abuse. Mandatory statewide databases have been created to monitor their use. The goal of this study was to identify patient risk factors for prolonged opioid use after TKA. The authors retrospectively reviewed a consecutive series of 676 primary TKA procedures performed between January 2017 and July 2017. Information on fulfillment of narcotic, sedative, benzodiazepine, and stimulant prescriptions was obtained from the Pennsylvania State Controlled Substance Monitoring website 6 months before and 1 year after the procedure. Bivariate and multivariate analyses were used to identify risk factors for the need for a second prescription and opioid use for longer than 6 months. Of this cohort, 30.3% used preoperative opioids, 60.5% filled a second opioid prescription, and 11.8% continued opioid use for longer than 6 months. Patients who had opioid use before the index procedure had more than 3-fold (odds ratio [OR], 3.29; P<.001) increased odds of filling a second opioid prescription and 8-fold (OR, 8.05; P<.001) increased odds of postoperative opioid use for longer than 6 months. Multivariate analysis was used to identify independent risk factors for requiring a second prescription, including discharge to a rehabilitation facility (OR, 2.77), bilateral procedures (OR, 1.88), preoperative narcotic use (OR, 1.70), and younger age (OR, 0.95). Independent risk factors for narcotic use for longer than 6 months included preoperative sedative (OR, 3.30) or narcotic use (OR, 1.49). This study identified several risk factors associated with prolonged narcotic use after TKA, including preoperative sedative use, and determined their relative weight. [Orthopedics. 2021;44(1):e50–e54.]

Abstract

Opioids are used for pain control after total knee arthroplasty (TKA) and carry risk for abuse. Mandatory statewide databases have been created to monitor their use. The goal of this study was to identify patient risk factors for prolonged opioid use after TKA. The authors retrospectively reviewed a consecutive series of 676 primary TKA procedures performed between January 2017 and July 2017. Information on fulfillment of narcotic, sedative, benzodiazepine, and stimulant prescriptions was obtained from the Pennsylvania State Controlled Substance Monitoring website 6 months before and 1 year after the procedure. Bivariate and multivariate analyses were used to identify risk factors for the need for a second prescription and opioid use for longer than 6 months. Of this cohort, 30.3% used preoperative opioids, 60.5% filled a second opioid prescription, and 11.8% continued opioid use for longer than 6 months. Patients who had opioid use before the index procedure had more than 3-fold (odds ratio [OR], 3.29; P<.001) increased odds of filling a second opioid prescription and 8-fold (OR, 8.05; P<.001) increased odds of postoperative opioid use for longer than 6 months. Multivariate analysis was used to identify independent risk factors for requiring a second prescription, including discharge to a rehabilitation facility (OR, 2.77), bilateral procedures (OR, 1.88), preoperative narcotic use (OR, 1.70), and younger age (OR, 0.95). Independent risk factors for narcotic use for longer than 6 months included preoperative sedative (OR, 3.30) or narcotic use (OR, 1.49). This study identified several risk factors associated with prolonged narcotic use after TKA, including preoperative sedative use, and determined their relative weight. [Orthopedics. 2021;44(1):e50–e54.]

Management of postoperative pain for patients undergoing total knee arthroplasty (TKA) can be a challenge. Narcotics are frequently prescribed in the perioperative and postoperative periods and may lead to continued use. However, prolonged opioid use has been shown to be associated with negative outcomes after orthopedic surgery, including worse functional outcomes, delayed return to work, and increased rates of revision, periprosthetic joint infection, and morbidity and mortality.1–3

Because of the addiction potential and adverse effects of opioids, state prescription drug monitoring programs (PDMPs) have been developed to combat the excessive distribution of narcotic pain medications by providing an online resource to track filled prescriptions and prevent patients from obtaining prescriptions from multiple providers. Pharmacies in 49 states are required to report all distributed narcotics, and practitioners in many states are required to query the database before prescribing opioids. These databases have proven immensely accurate (97.1%), with excellent sensitivity and specificity (96.4% and 97.1%, respectively).4 Although most of the literature on continued opioid use after total joint arthroplasty is based on patient reporting or insurance registries, which have inherent limitations, this study examined these outcomes with objective data obtained from these databases.

The goal of this study was to identify independent risk factors for patients who require a second opioid prescription as well as those who have prolonged opioid use (>6 months) through a mandatory statewide PDMP.5

Materials and Methods

After institutional review board approval was obtained, a retrospective review was performed of 837 patients who underwent primary TKA between January 2017 and July 2017. Patients underwent procedures at an academic hospital or ambulatory surgery center. The study period of January 2017 to July 2017 was chosen because the PDMP for Pennsylvania and adjacent states required reporting of opioid prescriptions during this time and provided a reasonable number of TKA procedures for analysis of prescription patterns. This time frame also provided a reasonable number of TKA procedures for analysis of prescription patterns. An electronic chart query of the electronic medical record was performed to obtain demographic information, such as sex, race, age, body mass index, and Elixhauser Comorbidity Index,6 as well as details on the procedure, including length of stay, operative time, insurance status, discharge location, and complications during hospitalization. Patients who did not fill a prescription and patients who had another surgical procedure during the study period were excluded. An additional manual chart review was performed to identify and exclude patients who had undergone other procedures within a year of primary TKA (6 months before and 6 months after the procedure). A cohort of 676 patients remained after the exclusion criteria were applied. These patients underwent 568 unilateral and 108 bilateral TKA procedures.

Information on controlled substances was obtained through the Pennsylvania PDMP website.5 The Pennsylvania PDMP search engine also obtained prescription information from the following states: Connecticut, Delaware, District of Columbia, Illinois, Maine, Maryland, Massachusetts, Minnesota, New Jersey, New York, Ohio, Oklahoma, South Carolina, Texas, Virginia, and West Virginia.7 All prescriptions that were filled 6 months before and 9 months after the surgical procedure were recorded, and these included opioid pain medications, benzodiazepines, stimulants, and nonbenzodiazepine sedatives and hypnotics. In addition, the date and number of prescriptions, medication dosage, and quantity were recorded. Patients were grouped according to type of controlled substance into exposed and unexposed cohorts, depending on their exposure in the 6 months before surgery.

Outcome Variables

Primary outcomes were (1) fulfillment of a second opioid prescription and (2) prolonged opioid use, which was defined as fulfillment of a prescription more than 6 months after the indicated procedure. A period of 6 months was chosen as a reasonable time frame because it was unlikely that continued use of pain medication was related to the surgical procedure. The number of sedative, benzodiazepine, and stimulant prescriptions that were filled during this time frame was recorded as well.

Statistical Analysis

Change in “use status” from the preoperative period to the postoperative period was evaluated with the McNemar test (nonsignificant results indicate consistency over time). Pearson's chi-square tests were used to test postoperative group differences. Preliminary bivariate logistic regression analyses were conducted to determine whether various preoperative and patient variables were significantly related to any of the key binary outcomes (filling a second prescription and filling a prescription 6 months after surgery). The following variables were included as predictors of the primary outcomes: bilateral TKA, length of stay, age, sex, race, body mass index, Elixhauser Comorbidity Index,8 operative time, government insurance, discharge destination, preoperative benzodiazepine use, preoperative narcotic use, preoperative hypnotic or sedative use, and preoperative stimulant use. Variables that were related to any of the outcomes in the bivariate analyses (P<.25, as recommended by Bursac et al,9 because lower P values, such as .05, can exclude important predictors) were considered for inclusion in the final multivariate logistic regression predictive model. An alpha of 0.05 was used to evaluate significance. Area under the curve analysis was used to evaluate the strength of the predictive model.

Results

Of the 676 patients who underwent primary TKA, 30.3% (205 of 676) used opioids preoperatively, 60.5% (409 of 676) filled a second opioid prescription, and 11.8% (80 of 676) continued to use opioids more than 6 months after surgery. Postoperatively, a mean of 2.9 opioid prescriptions were filled, equal to 2084.5 morphine milligram equivalents for 47.0 days. When stratified by preoperative opioid use, 52.7% (248 of 471) opioid-naïve patients required a second prescription and 4.7% (22 of 471) continued narcotic use for longer than 6 months, whereas 78.5% (161 of 205) patients who had preoperative opioid use required a second prescription and 28.3% (58 of 205) continued narcotic use for longer than 6 months (Figure 1).

Percentage of patients who filled a second postoperative narcotic prescription and had continued narcotic use for longer than 6 months among those who did and did not use narcotics preoperatively.

Figure 1:

Percentage of patients who filled a second postoperative narcotic prescription and had continued narcotic use for longer than 6 months among those who did and did not use narcotics preoperatively.

Bivariate analysis showed an association of independent variables (Table 1) with an increased risk of filling a second opioid prescription, in descending order, as follows: preoperative narcotic use (odds ratio [OR], 3.29; P<.001), preoperative sedative use (OR, 3.08; P=.014), disposition to a rehabilitation facility or skilled nursing facility (OR, 2.31; P=.005), simultaneous bilateral TKA procedures (OR, 2.16; P=.001), preoperative benzodiazepine use (OR, 1.92; P=.005), government insurance (OR, 0.71; P=.32), and younger age (OR, 0.95; P<.001). Predictors of narcotic use for longer than 6 months postoperatively, in descending order of influence, included the following: preoperative narcotic use (OR, 8.05; P<.001), preoperative sedative use (OR, 5.59; P<.001), preoperative benzodiazepine use (OR, 2.36; P=.002), disposition to a rehabilitation facility (OR, 2.48; P=.005), Elixhauser comorbidities (OR, 1.28; P=.008), and length of stay (OR, 1.19; P=.016).

Bivariate Predictors of Second Prescription and Prolonged Opioid Use

Table 1:

Bivariate Predictors of Second Prescription and Prolonged Opioid Use

After controlling for potential confounding variables in a multivariate analysis (Table 2), independent risk factors for requiring a second opioid prescription after TKA, in order of decreasing influence, included the following: discharge to a rehabilitation facility (OR, 2.77; P=.006), bilateral surgical procedures (OR, 1.88; P=.019), preoperative narcotic use (OR, 1.70; P<.001), and younger age (OR, 0.95; P<.001). Preoperative use of benzodiazepine and sedative use did not show statistically significant risk factors for a second opioid prescription. Multivariate analysis for opioid use for longer than 6 months showed that preoperative sedative use was a greater risk factor (OR, 3.30; P=.014) than preoperative narcotic use (OR, 1.49; P<.001). The models for filling a second prescription and continuing to use opioids for longer than 6 months showed excellent area under the curve scores of 0.73 and 0.82, respectively.

Multivariate Predictors of Second Prescription and Prolonged Opioid Use

Table 2:

Multivariate Predictors of Second Prescription and Prolonged Opioid Use

For repeated measurements, the McNemar test found that opioid (P<.001) and sedative (P=.04) use status changed significantly after surgery. After TKA, 3.0% of patients who had not used sedatives previously were using these medications postoperatively. In contrast, 1.2% of all patients who had a history of sedative use discontinued use postoperatively. Notably, benzodiazepine (P=.80) and stimulant (P=.40) use did not change significantly after surgery.

Discussion

Because opioids are frequently prescribed after total joint arthroplasty and because the complications associated with opioid overuse are so great, preoperative risk evaluation is needed to identify patients who are at risk. To combat the over-prescribing of opioid pain medications, PDMPs were created to allow for easy access to a patient's prescription fulfillment of such medications with high accuracy. This study allowed for the identification of preoperative risk factors for continued postoperative opioid use. Most notably, the authors found that patients who had preoperative sedative use showed even higher odds of prolonged opioid use than those with preoperative opioid use. The effect of preoperative sedative use and its effects on postoperative opioid use have not been studied extensively, but the association of substance use disorders with abuse of prescription opioids has been established.10 A cross-sectional study by Saunders et al11 found that 32% of patients who had prolonged opioid therapy also used sedatives and that risk factors for concurrent sedative use were female sex, younger age, depression, higher daily opioid doses, and use of opioids for more than 1 pain condition.

A strength of this study was the use of a mandatory pharmacy reporting PDMP database. Other studies have been performed with less granular insurance registries. An insurance database cohort study of 73,959 patients who had undergone TKA found that preoperative opioid use was the strongest predictor of prolonged opioid use, and younger age and female sex also increased the rate of opioid use.7 In another large insurance database study, Politzer et al8 found that 34.8% of patients who used opioids preoperatively had prolonged use postoperatively compared with 5% of opioid-naïve patients. Bedard et al7 also reported that preoperative opioid use increased from 30.1% in 2007 to 39.3% in 2014. Cancienne et al12 queried an insurance database and found an increased risk of emergency department visits, readmission, infection, stiffness, and aseptic revision among 31,733 patients who were prescribed opioids preoperatively. In a large study of patient factors associated with greater opioid use, Namba et al13 found that younger age, preoperative nonsteroidal anti-inflammatory drug use, anxiety, substance abuse, diabetes, preoperative opioid use, back pain, congestive heart failure, depression, fibromyalgia, hypertension, chronic pain, Black race, and chronic lung disease all were associated with increased postoperative opioid use. Increasing daily use of preoperative narcotics also appears to play a role in prolonged use. The use of more than 60 morphine milligram equivalents daily is reported to result in an 80% likelihood of postoperative use.14 Interestingly, in the current study, the authors showed that preoperative sedative use was an even greater independent risk factor than preoperative narcotic use for postoperative opioid use for longer than 6 months. This finding suggests the need to consider sedative use as part of informed consent because of the high risk of continued opioid use and the resulting complications.

In contrast to the studies mentioned earlier, a strength of the current study was the use of a mandatory multistate database of filled prescriptions from pharmacies vs individual reporting or insurance or institutional databases, as in much of the current literature.1,7,8,12,13,15 Institutional or insurance databases may not report opioid use accurately because patients may underreport use or may not fill prescriptions.

Few studies have used online substance abuse databases to investigate narcotic abuse after arthroplasty. Dwyer et al15 compared opioid prescription patterns between patients undergoing total hip arthroplasty (THA) and TKA and found that those undergoing TKA filled a greater number of opioid prescriptions and used a larger total morphine milligram equivalent dose. In this prescription database of 186 THA and 197 TKA procedures, the medical comorbidities also differed. Patients in the TKA group were older, had higher body mass index, and were more likely to have at least 1 medical comorbidity, including hypertension, diabetes, and cardiac issues. Of note, no difference was seen in rates of obesity, chronic lung disease, pain comorbidity, anxiety, or depression between the THA and TKA groups. This database also found that the presence of a medical comorbidity, anxiety, or depression correlated with the need for opioid refills only in patients undergoing TKA. This study showed that although THA and TKA may represent similar orthopedic pathology, patient demographics and risk factors for prolonged opioid use after surgery may not be the same and should be studied separately.

The current study is not without limitations. First, the authors could not ensure that all filled opioid prescriptions were recorded within the database. For example, a patient may have filled a prescription in an adjacent state that is not included in the Pennsylvania PDMP. Second, despite mandatory reporting and good accuracy, as described by Hozack et al,4 prescriptions may not be recorded accurately or errors in data entry may occur. Third, it is possible that patients may not obtain narcotic medications legally or through pharmacy distribution. Fourth, these data included only filled prescriptions, not consumed narcotics. Finally, it is possible that the study could be underpowered at 676 patients to identify risk factors with a low incidence in the population.

Conclusion

This study identified preoperative patient risk factors associated with the need for a second prescription and use of narcotics for longer than 6 months postoperatively. Other studies have identified preoperative opioid use as a risk factor for prolonged use, but this study is the first to identify it with a mandatory reporting database. Interestingly, in the current study, multivariate analysis showed that preoperative sedative use is the strongest preoperative predictor of postoperative opioid use for longer than 6 months. This study provides predisposing factors to allow for patient risk stratification for postoperative opioid use and may be of benefit during preoperative patient counseling for TKA. The use of these mandatory statewide databases appears to be an effective way to help to identify at-risk patients.

References

  1. Bell KL, Shohat N, Goswami K, Tan TL, Kalbian I, Parvizi J. Preoperative opioids increase the risk of periprosthetic joint infection after total joint arthroplasty. J Arthroplasty. 2018;33(10):3246–3251. doi:10.1016/j.arth.2018.05.027 [CrossRef] PMID:30054211
  2. Weick J, Bawa H, Dirschl DR, Luu HH. Preoperative opioid use is associated with higher readmission and revision rates in total knee and total hip arthroplasty. J Bone Joint Surg Am. 2018;100(14):1171–1176. doi:10.2106/JBJS.17.01414 [CrossRef] PMID:30020122
  3. Menendez ME, Ring D, Bateman BT. Preoperative opioid misuse is associated with increased morbidity and mortality after elective orthopaedic surgery. Clin Orthop Relat Res. 2015;473(7):2402–2412. doi:10.1007/s11999-015-4173-5 [CrossRef] PMID:25694266
  4. Hozack BA, Rivlin M, Graham J, Lutsky KF, Beredjiklian PK. Validation of the prescription drug monitoring program web site. J Opioid Manag. 2019;15(6):495–498.
  5. Prescription Drug Monitoring Program (PDMP). health.pa.gov. Accessed November 2017. https://www.health.pa.gov/topics/programs/PDMP/Pages/PDMP.aspx
  6. Elixhauser A, Steiner C, Harris DR, Coffey RM. Comorbidity measures for use with administrative data. Med Care. 1998;36(1):8–27. doi:10.1097/00005650-199801000-00004 [CrossRef] PMID:9431328
  7. 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. 2017;32(8):2390–2394. doi:10.1016/j.arth.2017.03.014 [CrossRef] PMID:28413136
  8. Politzer CS, Kildow BJ, Goltz DE, Green Cl, Bolognesi MP. Trends in opioid utilization before and after total knee arthroplasty. J Arthroplasty. 2018;33(7)(suppl):S147–S153. doi:10.1016/j.arth.2017.10.060 [CrossRef] PMID:29198871
  9. Bursac Z, Gauss CH, Williams DK, Hosmer DW. Purposeful selection of variables in logistic regression. Source Code Biol Med. 2008;3(1):17. doi:10.1186/1751-0473-3-17 [CrossRef] PMID:19087314
  10. Chou R, Fanciullo GJ, Fine PG, et al. American Pain Society-American Academy of Pain Medicine Opioids Guidelines Panel. Clinical guidelines for the use of chronic opioid therapy in chronic non-cancer pain. J Pain. 2009;10(2):113–130. doi:10.1016/j.jpain.2008.10.008 [CrossRef] PMID:19187889
  11. Saunders KW, Von Korff M, Campbell CI, et al. Concurrent use of alcohol and sedatives among persons prescribed chronic opioid therapy: prevalence and risk factors. J Pain. 2012;13(3):266–275. doi:10.1016/j.jpain.2011.11.004 [CrossRef] PMID:22285611
  12. Cancienne JM, Patel KJ, Browne JA, Werner BC. Narcotic use and total knee arthroplasty. J Arthroplasty. 2018;33(1):113–118. doi:10.1016/j.arth.2017.08.006 [CrossRef] PMID:28887020
  13. Namba RS, Singh A, Paxton EW, Inacio MCS. Patient factors associated with prolonged postoperative opioid use after total knee arthroplasty. J Arthroplasty. 2018; 33(8):2449–2454. doi:10.1016/j.arth.2018.03.068 [CrossRef] PMID:29753617
  14. Goesling J, Moser SE, Zaidi B, et al. Trends and predictors of opioid use after total knee and total hip arthroplasty. Pain. 2016;157(6):1259–1265. doi:10.1097/j.pain.0000000000000516 [CrossRef] PMID:26871536
  15. Dwyer MK, Tumpowsky CM, Hiltz NL, Lee J, Healy WL, Bedair HS. Characterization of post-operative opioid use following total joint arthroplasty. J Arthroplasty. 2018;33(3):668–672. doi:10.1016/j.arth.2017.10.011 [CrossRef] PMID:29128235

Bivariate Predictors of Second Prescription and Prolonged Opioid Use

PredictorSecond prescriptionProlonged opioid use (>6 mo)


OR95% CIPOR95% CIP
Bilateral total knee arthroplasty2.161.36–3.45.0010.930.49–1.79.829
Length of stay1.030.92–1.15.6331.191.03–1.36.016
Age0.950.93–0.97<.0010.980.96–1.01.245
Female sex1.290.94–1.76.1100.940.58–1.50.781
White0.920.62–1.36.6680.860.49–1.52.612
Body mass index1.020.99–1.05.2981.040.99–1.09.096
Elixhauser score0.980.86–1.12.7511.281.07–1.54.008
Disposition to rehabilitation2.311.28–4.14.0052.481.31–4.69.005
Government insurance0.710.52–0.97.0321.290.81–2.07.280
Operative time1.000.99–1.00.4341.011.00–1.02.080
In-hospital complications1.190.70–2.01.5170.710.30–1.70.444
Preoperative opioid use3.292.25–4.81<.0018.054.76–13.61<.001
Preoperative benzodiazepine use1.921.22–3.02.0052.361.38–4.03.002
Preoperative sedative use3.081.25–7.55.0145.592.66–11.74<.001
Preoperative stimulant use1.310.33–5.28.7040.930.12–7.54.946

Multivariate Predictors of Second Prescription and Prolonged Opioid Use

PredictorSecond prescriptionProlonged opioid use (>6 mo)


OR95% CIPOR95% CIP
Bilateral total knee arthroplasty1.881.11–3.20.0190.950.43–2.10.906
Length of stay0.880.76–1.02.0901.040.87–1.26.658
Age0.950.93–0.97<.0011.000.96–1.03.774
Female sex1.200.54–1.71.3120.840.47–1.48.538
White1.000.65–1.54.9880.950.48–1.87.881
Body mass index1.010.70–1.45.9661.590.87–2.90.130
Elixhauser score0.940.80–1.10.4511.040.83–1.31.719
Disposition to rehabilitation2.771.34–5.74.0062.110.87–5.14.101
Preoperative opioid use1.701.38–2.10<.0011.491.32–1.68<.001
Preoperative benzodiazepine use1.430.86–2.39.1741.261.32–1.68.514
Preoperative sedative use2.050.75–5.58.1623.301.27–8.55.014
Authors

The authors are from Rothman Orthopaedic Institute at Thomas Jefferson University, Philadelphia, Pennsylvania.

Drs Wilt, Tan, Rondon, Wang, and Foltz have no relevant financial relationships to disclose. Dr Beredjiklian holds stock in Cross Current Business Analytics, Dimension Orthotics LLC, Force Therapeutics, Matador Inc, OBERD, and Wright Medical Technology Inc. Dr Arnold has received research support from Zimmer and holds stock in Franklin Bioscience.

Correspondence should be addressed to: William V. Arnold, MD, Rothman Orthopaedic Institute at Thomas Jefferson University, 125 S 9th St, Ste 1000, Philadelphia, PA 19107 (balec. arnold@comcast.net).

Received: July 25, 2019
Accepted: November 25, 2019
Posted Online: November 03, 2020

10.3928/01477447-20201026-01

Sign up to receive

Journal E-contents