Abstract
We examined the effect of varying multimodal pain management (MMPM) combinations on oral morphine milligram equivalents (OMME) and length of stay (LOS) after total knee arthroplasty (TKA). Five groups were compared based on the combination of multimodal analgesics ranging from no MMPM to full MMPM with acetaminophen, gabapentinoids, and celecoxib. After risk adjustment, MMPM was associated with decreased odds of LOS ≥2 days and OMME ≥75th percentile. MMPM protocols are effective at reducing LOS and postoperative narcotic requirements post-TKA. Patients appear to derive similar benefit from receiving all three medications, as well as various combinations of these drugs.
Keywords: Total knee arthroplasty, Multimodal, Pain management, Narcotic, Opioid
1. Introduction
Total Knee Arthroplasty (TKA) is a commonly performed procedure for relieving joint pain in patients with end-stage arthritis.1,2 While effective in restoring joint function, TKA is often followed by severe post-operative pain that must be adequately managed in order to enhance early recovery, postoperative rehabilitation, patient satisfaction, and overall outcomes.1, 2, 3 Therefore, pain management practices that include multimodal pain management (MMPM) protocols are increasingly being utilized to control acute postoperative pain following TKA.
The traditional standard to treat post-arthroplasty pain was for orthopedic surgeons to prescribe long-acting opioids.4 However, many opioid-naive individuals became at-risk or addicted to legal, physician prescribed medications following surgical procedures.4 Previously described risk factors for increased postoperative opioid use after total joint arthroplasty include younger patient age, tobacco use, greater symptoms of depression, private insurance, and TKA.5 Multimodal pain regimens gained popularity in the 2000s as a treatment method to deter opioid dependence through the synergistic usage of analgesics that work in various parts of the pain pathway to prevent pain transduction, transmission, and perception.4 Literature investigating methods of postoperative analgesia have cited MMPM protocols to be highly effective in diminishing postoperative pain scores and reducing the use of opioids for acute pain management.1,3,4 Studies have also cited MMPM to be associated with reduced hospital stays, reduced hospital readmissions and decreased inpatient rehabilitation.4,6 Decreased length of stay lowers the risk of infection, medication adverse effects, and improves early postoperative outcomes.7 Early mobilization, a key component of rapid recovery protocols that requires adequate postoperative pain control, is associated with reduced medical complications in patients and subsequent decrease in time to discharge.8
While there is no standard MMPM protocol that is currently being practiced across all healthcare institutions, the American Society of Anesthesiologists (ASA) nonetheless agrees that unless contraindicated, patients on MMPM therapy should be receiving an around-the-clock, individualized regimen of nonsteroidal anti-inflammatory drugs (NSAIDs)—including COX inhibitors, or acetaminophen.9 These agents can be combined to work synergistically to target different portions of the pain pathway, yielding to greater analgesic effects.1,3,10
1.1. Common analgesics used in MMPM
A 2019 study reported that the two most commonly used preoperative medications in multimodal analgesia regimens post-TKA were acetaminophen (74.5% of cases) and Celecoxib (54.1% of cases).11 Non-steroidal anti-inflammatory drugs are a staple addition to many MMPM protocols due to their ability to significantly reduce postoperative pain while retaining a generally favorable side effect profile as compared to opioid monotreatment.1
NSAIDs commonly used in MMPM therapy include nonselective COX inhibitors, preferential COX-2 inhibitors, and COX-2 selective NSAIDs.12,13 Preoperative NSAID administration has been shown to significantly decrease early postoperative pain and opioid consumption following total joint arthroplasty.12 Additionally, qualitative studies have shown that there is no significant difference in pain score or opioid use between selective and nonselective NSAIDs.12 While the drugs may appear to have similar efficacy in postoperative pain related outcomes, the different classes are associated with different side effect profiles that warrant consideration during prescription. Nonselective NSAIDs are classically associated with significant gastrointestinal and renal adverse events such as gastritis, bleeding, and acute renal failure.2 Compared to other nonselective NSAIDs, celecoxib is associated with less adverse events such as gastric ulcers and bleeding.1,3 Celecoxib spares gastroprotective COX-1 action and thrombotic action due to its selective COX-2 inhibitory effect.1,14,15 Despite the large body of data available on NSAID related adverse events, little has been reported within the context of total joint arthroplasty (TJA).12,13 Current guidelines on MMPM recommend continued monitoring of potential NSAID related adverse events.12,13
Additional analgesic agents that can be synergistically combined with NSAIDs include acetaminophen and gabapentinoids. Acetaminophen is considered a strong candidate for MMPM protocols as it is not only cost-effective, but also the most commonly used analgesic for postoperative patients.16, 17, 18 Although the exact mechanism of action is not well known, acetaminophen is generally a safe analgesic, with hepatotoxicity occurring only if doses exceed 4g/day.16 Gabapentinoids act on voltage gated calcium channels of the nervous system to modulate pain perception. The use of gabapentinoids in MMPM therapy has been widely established, however the efficacy and safety of the drug has not been well identified.19, 20, 21, 22
In October 2020, The Journal of Arthroplasty published usage guidelines for acetaminophen, gabapentinoids, opioids, and NSAIDs in total joint arthroplasty. The guidelines for each of the four analgesics were based on how the analgesic in question affects patient outcomes and opioid consumption after TJA. Although the four guidelines were written separately from one another, when analyzed in concert, the findings continue to corroborate the current literature on the importance of MMPM and recommend (1) the use of NSAIDs preoperatively and/or postoperatively; (2) use of pregabalin after discharge; and (3) use of acetaminophen during the perioperative period for an overall reduction in pain and opioid consumption.12,17,21 Additionally, guidelines call for providers to use the minimum amount of opioids necessary to reduce opioid consumption and risk of dependence.23,24
Based on the conclusions and recommendations published in the guidelines and recent literature, we believe further research into MMPM and how each of the four analgesics work synergistically to reduce opioid consumption and post-operative pain should be conducted. This study examines how different levels of adherence to a standard MMPM protocol containing a combination of acetaminophen, gabapentin/pregabalin, and celecoxib within a single institution affect oral morphine milligram equivalents consumed and hospital length of stay.
2. Methods
2.1. Setting and subjects
The study was determined to be institutional review board exempt by the institutional clinical research committee. A single regional health system's administrative database was used to perform a retrospective observational cohort study of a consecutive series of 1648 elective primary unilateral TKAs. All procedures were performed between January 1, 2019 and June 30, 2020 by one of 11 board-certified orthopedic surgeons in a 350-bed acute care hospital.
2.2. TKA protocol
All patients were subject to the same perioperative protocols in a coordinated Joint Replacement Center and received preoperative education consisting of written materials, a preoperative education class for patients and their coaches, preoperative medical evaluation, and preoperative strengthening via a home exercise program or formal physical therapy. For patients deemed to benefit from additional support, in-home physical therapy was used for preoperative optimization, and one to two postoperative visits were provided to bridge patients to outpatient therapy. This institution has established a standard rapid recovery pathway for all TJA patients including a multimodal pain management regimen and short-acting opioids; patient-controlled analgesia and femoral nerve blocks were not used. The preoperative multimodal analgesia protocol recommends use of 925 mg of oral acetaminophen, 50 mg of pregabalin, and 200 mg of celecoxib. The final decision regarding which elements of the preoperative multimodal protocol to use were made by the surgical team on a patient specific basis. Anesthesia was either general or neuraxial determined by the anesthesiologist in consultation with the patient and surgeon. Patients also received intravenous or topical tranexamic acid and day of surgery ambulation when appropriate. Aspirin 325 mg bid was the primary pharmacologic deep vein thrombosis prophylaxis, with warfarin or apixaban used in select high-risk patients. Prior to discharge, all patients achieved adequate pain control using oral medication, had stable vital signs, were able to ambulate, and voided. Final discharge clearance included consensus from the surgical, medical, and therapy providers.
2.3. Study design and data analysis
Patients were grouped into 5 cohorts based on the combination of multimodal medications received during the preoperative period:
-
•
Group 1: No multimodal analgesia
-
•
Group 2: Acetaminophen only
-
•
Group 3: Acetaminophen with pregabalin or gabapentin
-
•
Group 4: Celecoxib with acetaminophen or pregabalin/gabapentin
-
•
Group 5: Full multimodal analgesia – celecoxib with acetaminophen and pregabalin/gabapentin
Patients receiving a combination of multimodal medications that did not fall into one of these categories were then excluded from the study. Patients discharged to skilled nursing facilities (SNF) were also excluded, to eliminate the confounding influence of the Medicare required 2-night stay for SNF eligibility on evaluation of hospital length of stay (LOS).
Differences in patient demographics and comorbidities were assessed across groups. Age, body mass index (BMI), gender, race, and the prevalence of the following comorbidities as coded in the patient's chart at the time of surgery were evaluated: diabetes mellitus, sleep apnea, chronic obstructive pulmonary disease (COPD), asthma, atrial fibrillation (AFIB), congestive heart failure (CHF), primary hypertension (HTN), coronary artery disease (CAD), end stage renal disease (ESRD) or chronic kidney disease (CKD), gastroesophageal reflux disease (GERD), and anemia. The international classification of disease 10th edition (ICD-10) codes used to define each comorbidity are presented in Appendix A. For each group the rate of use of each multimodal medication, median dose given, and interquartile range (IQR) of dose was evaluated. The prevalence of the comorbidities deemed most likely to influence prescribing patterns was assessed for each medication. Differences in anesthesia details, post-anesthesia care unit (PACU) recovery outcomes, hospital outcomes, and post-discharge outcomes were then compared across groups. The primary endpoints were hospital LOS ≥2 days and oral morphine milligram equivalents (OMME) ≥ 75th percentile of the full patient population. LOS days were calculated based on the number of midnights stayed in the hospital, and OMME were calculated using the conversion factors presented in Appendix B. All preoperative, PACU, and floor narcotics (intravenous and oral) were included in OMME calculations.
Univariate comparisons across groups were performed using one-way analysis of variance (ANOVA) and chi-squared tests for continuous and categorical variables, respectively. Fisher's Exact test was performed when the necessary assumptions for chi-squared testing were not met. For the primary endpoints, post hoc analysis of differences between individual groups was performed using Bonferroni adjustment. Patient demographics, comorbidities, and anesthesia details that were found to differ significantly (p < 0.05) between groups were entered into multivariate regression models for risk-adjusted evaluation of the primary endpoints. Multivariate logistic regression was performed to evaluate the risk of hospital LOS ≥2 days and OMME ≥75th percentile for the multimodal groups in comparison to patients not receiving multimodal analgesia, and for the groups receiving elements of the multimodal protocol in comparison to patients receiving all three multimodal analgesics preoperatively. Adjusted odds ratios and 95% confidence intervals were calculated for the regression analyses. Statistical significance was assessed at p < 0.05, and all statistical analysis was performed in SPSS version 26 (IBM, Armonk, NY).
3. Results
3.1. Study population
A total of 1648 patients undergoing TKA were retrospectively reviewed. Seventy-eight patients discharged to SNF and 24 patients receiving a multimodal analgesia combination not captured by the 5 study categories were excluded. One thousand five hundred forty six patients were included in the study. The average age of the patient population was 67.1 ± 8.6 years, average BMI was 31.7 ± 5.3 kg/m2, 57.9% of the population was female and 20.1% were of non-white race. Statistically significant differences in age, BMI, and race were observed across study groups (all p < 0.001), while gender distributions were similar between groups (Table 1). The most prevalent comorbidities in the population were primary HTN (61.4% of patients), GERD (34.9%), and sleep apnea (20.3%). Statistically significant differences in rates of diabetes mellitus (p = 0.012), CAD (p = 0.036), and GERD (p = 0.010) were observed between groups, while rates of sleep apnea, COPD, asthma, AFIB, CHF, primary HTN, ESRD or CKD and anemia were similar across the population (Table 1).
Table 1.
Patient demographics and comorbidities.
| Patient Characteristic | All Patients N = 1546 | Group 1 No Multimodal N = 357 | Group 2 Acetaminophen Only N = 151 | Group 3 Acetaminophen with Pregabalin/Gabapentin Only N = 326 | Group 4 Celecoxib with Acetaminophen or Pregabalin/Gabapentin N = 57 | Group 5 Full Multimodal – Celecoxib, Acetaminophen, and Pregabalin/Gabapentin N = 655 | P-Value |
|---|---|---|---|---|---|---|---|
| Demographics | |||||||
| Age – avg. yrs. ± SD | 67.1 ± 8.6 | 67.3 ± 9.4 | 70.7 ± 7.3 | 66.2 ± 8.4 | 69.3 ± 8.4 | 66.4 ± 8.3 | <0.001 |
| BMI – avg. kg/m2 | 31.7 ± 5.3 | 31.2 ± 4.8 | 30.8 ± 4.9 | 32.3 ± 5.5 | 32.1 ± 5.5 | 31.9 ± 5.5 | <0.001 |
| Female – n (%) | 895 (57.9) | 211 (59.1) | 91 (60.3) | 204 (62.6) | 29 (50.9) | 360 (55.0) | 0.138 |
| Non-White Race –n (%) | 310 (20.1) | 49 (13.7) | 18 (11.9) | 82 (25.2) | 11 (19.3) | 150 (22.9) | <0.001 |
| Comorbidities – n (%) | |||||||
| Diabetes Mellitus | 276 (17.9) | 56 (15.7) | 19 (12.6) | 73 (22.4) | 16 (28.1) | 112 (17.1) | 0.012 |
| Sleep Apnea | 314 (20.3) | 71 (19.9) | 32 (21.2) | 77 (23.6) | 16 (28.1) | 118 (18.0) | 0.160 |
| COPD | 68 (4.4) | 15 (4.2) | 8 (5.3) | 13 (4.0) | 4 (7.0) | 28 (4.3) | 0.843 |
| Asthma | 157 (10.2) | 27 (7.6) | 14 (9.3) | 46 (14.1) | 6 (10.5) | 64 (9.8) | 0.076 |
| AFIB | 104 (6.7) | 24 (6.7) | 18 (11.9) | 18 (5.5) | 6 (10.5) | 38 (5.8) | 0.051 |
| CHF | 24 (1.6) | 8 (2.2) | 1 (0.7) | 5 (1.5) | 2 (3.5) | 8 (1.5) | 0.435 |
| Primary HTN | 950 (61.4) | 204 (57.1) | 103 (68.2) | 209 (64.1) | 35 (61.4) | 399 (60.9) | 0.149 |
| CAD | 173 (11.2) | 44 (12.3) | 24 (15.9) | 33 (10.1) | 11 (19.3) | 61 (9.3) | 0.036 |
| ESRD or CKD | 101 (6.5) | 25 (7.0) | 12 (7.9) | 26 (8.0) | 7 (12.3) | 31 (4.7) | 0.081 |
| GERD | 540 (34.9) | 118 (33.1) | 70 (46.4) | 122 (37.4) | 22 (38.6) | 208 (31.8) | 0.010 |
| Anemia | 30 (1.9) | 5 (1.4) | 3 (2.0) | 7 (2.1) | 2 (3.5) | 13 (2.0) | 0.850 |
P-Values < 0.05 in bold.
BMI- body mass index.
COPD – chronic obstructive pulmonary disease.
AFIB – atrial fibrillation.
CHF – congestive heart failure.
HTN - hypertension.
CAD – coronary artery disease.
ESRD – end stage renal disease.
CKD – chronic kidney disease.
GERD – gastroesophageal reflux disease.
3.2. Multimodal medication use
The distribution of preoperative multimodal medication use across groups and associated dose given is presented in Table 2. Across all groups standardized dosing was largely adhered to, with median doses of acetaminophen (975 mg), pregabalin (50 mg), gabapentin (300 mg), and celecoxib (200 mg) all having an interquartile range of 0 mg, demonstrating limited variability. In the groups including anticonvulsant medication, pregabalin was more commonly used over gabapentin (86.5% of patients in group 3, 33.3% in group 4, 98.0% in group 5). In group 4 (celecoxib with acetaminophen or pregabalin/gabapentin), acetaminophen was used in conjunction with celecoxib in 64.9% of cases and an anticonvulsant with celecoxib in 35.1% of cases. The prevalence of comorbidities deemed most likely to influence the decision to prescribe each multimodal medication is presented in Table 3. The only statistically significant trends observed were reduced likelihood of using acetaminophen in patients with CHF (p = 0.047) and reduced likelihood of using celecoxib in patients with GERD (p = 0.045). A potentially clinically significant trend toward reduced utilization of celecoxib in patients with ESRD/CKD—a 10% lower rate of use than in patients without this comorbidity—was observed, although the result was not statistically significant (p = 0.079).
Table 2.
Preoperative multimodal medication usage and dosing.
| Preoperative Medication | Group 2 Acetaminophen Only N = 151 | Group 3 Acetaminophen with Pregabalin/Gabapentin Only N = 326 | Group 4 Celecoxib with Acetaminophen or Pregabalin/Gabapentin N = 57 | Group 5 Full Multimodal – Celecoxib, Acetaminophen, and Pregabalin/Gabapentin N = 655 |
|---|---|---|---|---|
| Acetaminophen – n (%) | 151 (100.0) | 326 (100.0) | 37 (64.9) | 655 (100.0) |
| Acetaminophen mg – median (IQR) | 975 (975–975) | 975 (975–975) | 975 (975–975) | 975 (975–975) |
| Pregabalin – n (%) | 0 (0.0) | 282 (86.5) | 19 (33.3) | 642 (98.0) |
| Pregabablin mg - median (IQR) | N/A | 50 (50–50) | 50 (50–50) | 50 (50–50) |
| Gabapentin – n (%) | 0 (0.0) | 44 (13.5) | 1 (1.8) | 14 (2.0) |
| Gabapentin mg - median (IQR) | N/A | 300 (300–300) | 300 (N/A) | 300 (300–300) |
| Celecoxib – n (%) | 0 (0.0) | 0 (0.0) | 57 (100.0) | 655 (100.0) |
| Celecoxib mg- median (IQR) | N/A | N/A | 200 (200–200) | 200 (200–200) |
IQR – interquartile range.
Dose amount for patients receiving medication only.
Table 3.
Prevalence of comorbidities by multimodal medication prescribed.
| Preoperative Medication | No Diabetes |
Diabetes |
P-Value |
No ESRD/CKD |
ESRD/CKD |
P-Value | No Primary HTN |
Primary HTN |
P-Value |
No CHF |
CHF |
P-Value |
No GERD |
GERD |
P-Value |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| N = 1270 | N = 276 | N = 1445 | N = 101 | N = 596 | N = 950 | N = 1522 | N = 24 | N = 1006 | N = 540 | ||||||
| Acetaminophen – n (%) | 955 (75.2) | 214 (77.5) | 0.412 | 1094 (75.7) | 75 (74.3) | 0.743 | 435 (73.0) | 734 (77.3) | 0.057 | 1155 (75.9) | 14 (58.3) | 0.047 | 758 (75.3) | 411 (76.1) | 0.739 |
| Pregabalin – n (%) | 767 (60.4) | 176 (63.8) | 0.298 | 888 (61.5) | 55 (54.5) | 0.163 | 357 (59.9) | 586 (61.7) | 0.484 | 930 (61.1) | 13 (54.2) | 0.489 | 618 (61.4) | 325 (60.2) | 0.632 |
| Gabapentin – n (%) | 44 (3.5) | 15 (5.4) | 0.122 | 56 (3.9) | 3 (3.0) | 1.000a | 24 (4.0) | 35 (3.7) | 0.732 | 57 (3.7) | 2 (8.3) | 0.232* | 43 (4.3) | 16 (3.0) | 0.199 |
| Celecoxib – n (%) | 584 (46.0) | 128 (46.4) | 0.906 | 674 (46.6) | 38 (37.6) | 0.079 | 278 (46.6) | 434 (45.7) | 0.712 | 702 (46.1) | 10 (41.7) | 0.664 | 482 (47.9) | 230 (42.6) | 0.045 |
P Values < 0.05 in bold.
Fisher's Exact Test.
3.3. Anesthesia details and patient outcomes
Spinal anesthesia was used in 62.6% of surgeries and 41.4% of patients had an ASA score of ≥3; differences in both variables were significant between groups (both p < 0.001). Examination of postoperative recovery outcomes in the PACU found no difference in rates of postoperative nausea or vomiting (PONV, 2.9% overall), last pain numeric rating scale (NRS) ≥ 7 (5.6%), or average minutes in PACU (155.3 ± 54.3). A statistically significant difference in average last pain NRS was observed across groups (overall average = 3.2 ± 2.3, p < 0.001). The overall average LOS (ALOS) was 1.08 ± 0.68 days, and average inpatient OMME consumed was 82.94 ± 53.42; the 75th percentile of narcotic consumption was 105 OMME. Statistically significant differences were found across groups for all hospital outcomes—ALOS days, ALOS hours, % LOS ≥2 days, inpatient OMME, and % OMME ≥75th percentile (all p < 0.001). Post hoc Bonferonni adjusted comparisons between groups were performed for the primary study outcomes. Patients in groups 4 and 5 had significantly lower rates of LOS ≥2 days when compared to group 1 (group 1 = 21.8%, group 4 = 5.3%, group 5 = 9.2%, p < 0.05), but statistically similar rates to each other, and to groups 2 (12.6%) and 3 (9.2%). Groups 2 and 3 did not demonstrate significantly different rates of LOS ≥2 days in comparison to group 1. For the OMME ≥75th percentile endpoint, groups 3 and 5 had significantly lower rates of high narcotic use when compared to group 1 (group 1 = 37.3%, group 3 = 23.3%, group 5 = 19.8%), but similar rates to each other, and to groups 2 (25.8%) and 4 (26.3%). Groups 2 and 4 did not demonstrate significantly different rates of OMME ≥75th percentile in comparison to group 1. No statistically significant differences in rates of 30-day readmission, 90-day readmission, or 30-day return to OR were observed across groups (Table 4).
Table 4.
Anesthesia details and hospital outcomes.
| Patient Characteristic | All Patients |
Group 1 No Multimodal |
Group 2 Acetaminophen Only |
Group 3 Acetaminophen with Pregabalin/Gabapentin Only |
Group 4 Celecoxib with Acetaminophen or Pregabalin/Gabapentin |
Group 5 Full Multimodal – Celecoxib, Acetaminophen, and Pregabalin/Gabapentin |
P-Value |
|---|---|---|---|---|---|---|---|
| N = 1546 | N = 357 | N = 151 | N = 326 | N = 57 | N = 655 | ||
| Anesthesia Details – n (%) | |||||||
| Spinal Anesthesia | 968 (62.6) | 216 (60.5) | 82 (54.3) | 176 (54.0) | 41 (71.9) | 453 (69.2) | <0.001 |
| ASA ≥3 | 640 (41.4) | 145 (40.6) | 79 (52.3) | 141 (43.3) | 33 (57.9) | 242 (36.9) | <0.001 |
| PACU Recovery Outcomes | |||||||
| PONV – n (%) | 45 (2.9) | 10 (2.8) | 5 (3.3) | 8 (2.5) | 2 (3.5) | 20 (3.1) | 0.977 |
| Last Pain NRS – avg. ± SD | 3.2 ± 2.3 | 3.5 ± 2.4 | 3.1 ± 2.2 | 3.5 ± 2.3 | 2.7 ± 2.7 | 2.9 ± 2.2 | <0.001 |
| Last Pain NRS ≥7 – n (%) | 86 (5.6) | 27 (7.6) | 7 (4.6) | 21 (6.4) | 3 (5.3) | 28 (4.3) | 0.238 |
| Minutes in PACU – avg. ± SD | 155.3 ± 54.3 | 153.3 ± 54.3 | 148.5 ± 50.0 | 155.8 ± 70.3 | 156.5 ± 58.3 | 157.7 ± 52.6 | 0.440 |
| Hospital Outcomes | |||||||
| LOS (Days) – avg. ± SD | 1.08 ± 0.68 | 1.23 ± 0.82 | 1.15 ± 0.76 | 1.05 ± 0.56 | 1.02 ± 0.30 | 1.00 ± 0.64 | <0.001 |
| LOS (Hours) – avg. ± SD | 31.97 ± 15.69 | 35.54 ± 19.52 | 33.31 ± 18.05 | 31.30 ± 12.53 | 30.05 ± 7.63 | 30.29 ± 14.42 | <0.001 |
| LOS ≥2 Days – n (%) | 190 (12.3) | 78 (21.8)a | 19 (12.6)a,b | 30 (9.2)a,b | 3 (5.3)b | 60 (9.2)b | <0.001 |
| Inpatient OMME – avg. ± SD | 82.94 ± 53.42 | 97.47 ± 60.82 | 80.33 ± 45.40 | 79.95 ± 46.55 | 89.16 ± 62.32 | 76.56 ± 51.79 | <0.001 |
| OMME ≥75th percentile – n (%) | 393 (25.4) | 133 (37.3)a | 39 (25.8)a,b | 76 (23.3)b | 15 (26.3)a,b | 130 (19.8)b | <0.001 |
| Post-Discharge Outcomes | |||||||
| 30-Day Readmission – n (%) | 29 (1.9) | 4 (1.1) | 7 (4.6) | 5 (1.5) | 0 (0.0) | 13 (2.0) | 0.069 |
| 90-Day Readmission – n (%) | 31 (2.0) | 5 (1.4) | 7 (4.6) | 5 (1.5) | 0 (0.0) | 14 (2.1) | 0.108 |
| 30-Day Return to OR – n (%) | 12 (0.8) | 3 (0.8) | 3 (2.0) | 1 (0.3) | 0 (0.0) | 5 (0.8) | 0.370 |
P-Values <0.05 in bold.
Each subscript letter denotes a subset of group categories whose column proportions do not differ significantly from each other at the .05 level after bonferroni adjusted post hoc analysis (ie. a does not significantly differ from a; a,b does not significantly differ from a or b).
ASA – American Society of Anesthesiologists classification.
PACU – Post anesthesia care unit.
PONV – Postoperative nausea and vomiting.
NRS – Numeric rating scale (0–10).
LOS – Length of stay.
OMME – Oral morphine milligram equivalents.
OR – Operating room.
After controlling for age, BMI, non-white race, diabetes mellitus, CAD, GERD, ASA ≥3, and spinal anesthesia, each of the multimodal groups demonstrated decreased odds of hospital LOS ≥2 days in comparison to patients not receiving any multimodal analgesia (group 2 OR: 0.464, p = 0.009; group 3 OR: 0.360, p < 0.001; group 4 OR: 0.164, p = 0.004; group 5 OR: 0.398, p < 0.001). Similarly, after risk adjustment all multimodal groups demonstrated decreased odds of OMME ≥75th percentile in comparison to patients not receiving multimodal analgesia, with the exception of group 4 (group 2 OR: 0.558, p = 0.010; group 3 OR: 0.375, p < 0.001; group 4 OR: 0.524, p = 0.053; group 5 OR: 0.351, p < 0.001) (Table 5). In only patients receiving multimodal pain management (groups 2–5), no significant differences in odds of hospital LOS ≥2 days were observed when comparing each group to patients receiving the full multimodal protocol (group 5) after risk adjustment. Within this population, no differences in odds of OMME ≥75th percentile were observed, with the exception of group 2, who was at significantly increased risk of high narcotic consumption when compared to group 5 (OR: 1.602, p = 0.038) (Table 6).
Table 5.
Multivariate logistic regression analysis of hospital outcomes in comparison to patients not receiving multimodal analgesia.
| Multimodal Group | Odds Ratio | OR 95% CI | P-Value |
|---|---|---|---|
| Outcome: Hospital LOS ≥2 Days | |||
| Group 2 Acetaminophen Only | 0.464 | 0.262–0.822 | 0.009 |
| Group 3 Acetaminophen with Pregabalin/Gabapentin Only | 0.360 | 0.223–0.580 | <0.001 |
| Group 4 Celecoxib with Acetaminophen or Pregabalin/Gabapentin | 0.164 | 0.049–0.553 | 0.004 |
| Group 5 Full Multimodal – Celecoxib, Acetaminophen, and Pregabalin/Gabapentin | 0.398 | 0.269–0.590 | <0.001 |
| Outcome: Hospital OMME ≥ 75thPercentile | |||
| Group 2 Acetaminophen Only | 0.558 | 0.357–0.871 | 0.010 |
| Group 3 Acetaminophen with Pregabalin/Gabapentin Only | 0.375 | 0.262–0.535 | <0.001 |
| Group 4 Celecoxib with Acetaminophen or Pregabalin/Gabapentin | 0.524 | 0.272–1.009 | 0.053 |
| Group 5 Full Multimodal – Celecoxib, Acetaminophen, and Pregabalin/Gabapentin | 0.351 | 0.258–0.478 | <0.001 |
P-Value <0.05 in bold.
Adjusted Odds Ratio controlling for age, BMI, non-white race, diabetes mellitus, CAD, GERD, ASA ≥3, spinal anesthesia.
Table 6.
Multivariate logistic regression analysis of hospital outcomes in comparison to patients receiving full multimodal analgesia protocol.
| Multimodal Group | Odds Ratio | OR 95% CI | P-Value |
|---|---|---|---|
| Outcome: Hospital LOS ≥2 Days | |||
| Group 2 Acetaminophen Only | 1.138 | 0.534–2.044 | 0.664 |
| Group 3 Acetaminophen with Pregabalin/Gabapentin Only | 0.909 | 0.566–1.461 | 0.694 |
| Group 4 Celecoxib with Acetaminophen or Pregabalin/Gabapentin | 0.408 | 0.121–1.373 | 0.148 |
| Outcome: Hospital OMME ≥ 75thPercentile | |||
| Group 2 Acetaminophen Only | 1.602 | 1.025–2.504 | 0.038 |
| Group 3 Acetaminophen with Pregabalin/Gabapentin Only | 1.070 | 0.764–1.499 | 0.692 |
| Group 4 Celecoxib with Acetaminophen or Pregabalin/Gabapentin | 1.464 | 0.766–2.798 | 0.248 |
P-Value <0.05 in bold.
Adjusted Odds Ratio controlling for age, BMI, non-white race, diabetes mellitus, CAD, GERD, ASA ≥3, spinal anesthesia.
4. Discussion
Pain after total knee arthroplasty can be adequately managed using multimodal methods. Our single institution study found acetaminophen, gabapentinoids, and celecoxib to be effective adjuncts in reducing length of stay and oral morphine milligram equivalents after TKA. While our results demonstrate the overall effectiveness of these medications in managing acute pain, we are unable to recommend an ideal MMPM combination for patients undergoing TKA.
4.1. Length of stay
Length of stay (LOS) comparisons were done at our institution between MMPM and non-MMPM groups, as well as within the MMPM subgroups. After controlling for potentially confounding factors, we observed a significant reduction in risk for prolonged length of stay ≥2 days in all groups receiving MMPM. We postulate that this was a result of multimodal therapy yielding better early postoperative pain control, as indicated by a trend towards reduced early pain scores in the MMPM group, consequently permitting earlier ambulation and discharge. Because opioid consumption and prolonged bed rest have been known to cause complications such as urinary retention and constipation, it is likely that decreasing opioid consumption while shortening bed rest in our multimodal groups led to decreased time to void and further contributed to the earlier discharge status for these patients.
Our findings corroborate the systematic review and meta-analysis by Guerra et al. finding that early mobilization within 24 h after total joint arthroplasty can reduce LOS without increasing adverse events, infection rates, urinary retention, or delaying recovery.25 Similar to our findings, Ellis et al. demonstrate that the ability to ambulate on postoperative day 0 is closely related to pain level, while Hebl et al. indicate that multimodal regimens can shorten hospital LOS and reduce postoperative urinary retention.26,27 While none of the four Journal of Arthroplasty Guidelines directly examined the relationship between LOS and MMPM, our single institution LOS data nonetheless support the overarching recommendations favoring MMPM over the use of opioids alone.
4.2. Oral morphine milligram equivalents (OMME)
Our study shows that post-TKA patients who received MMPM, had significantly improved odds of reducing OMME ≥75th percentile consumption. Apart from Group 4, OMME ≥75th percentile appears to trend downward as the number of adjunct analgesics used in an MMPM protocol increases. When comparing the various MMPM subgroups amongst each other, there appears to be no significant difference in OMME between Groups 3, 4, and 5. Interestingly, Group 2 showed increased odds of consuming OMME ≥75th percentile, indicating acetaminophen-only multimodal therapy is inferior to combined celecoxib, acetaminophen, and pregabalin multimodal therapy. Our findings are consistent with Memtsoudis et al., who demonstrated increased addition of adjunct analgesics (more than 2 modes) in MMPM are associated with stepwise positive effects and lead to decreased OMME consumption.28 However, after the inclusion of two adjunct analgesics, our data fails to show any significant reduction in opioid consumption compared to full multimodal therapy. Therefore, we recommend at least two adjunct analgesics in an MMPM protocol. In the instance that full multimodal therapy is contraindicated, our data suggests any tolerable multimodal combination compared to opioid monotherapy is beneficial in reducing opioid consumption. This finding is again in alignment with recent guidelines suggesting that each component of an MMPM protocol can reduce opioid consumption when used separately.12,17,21 In addition to using a pharmacologic MMPM protocol, patients should receive individualized counseling regarding postoperative opioid use and pain management, as preoperative education has been shown to reduce postoperative opioid use and severity of self-pain after total joint arthroplasty.29
An unexpected finding in our study was the non-significant reduction in OMME of Group 4 compared to the non-multimodal group. Celecoxib or Acetaminophen monotherapy is known to be effective at minimizing postoperative opioid consumption.12,13,17,18 Therefore, it seems counterintuitive that dual therapy including celecoxib would not significantly reduce postoperative opioid consumption when compared to opioid monotherapy. Our finding is likely due to the heterogeneous nature of the subgroup in the analysis as well as the group being underpowered to show a difference in OMME, with only 57 patients receiving this drug combination. This theory is supported by our finding that Group 4 was comparable to the other multimodal groups in reducing LOS compared to the non-multimodal group and was non-inferior to the full multimodal protocol in reducing both OMME and LOS.
This study does have multiple limitations. First, as a retrospective single institution study, our patient population may not be representative of the broader population of patients undergoing TKA. However, the high standardization of practice within our coordinated rapid recovery protocol allows for evaluation of variability in MMPM prescribing patterns in a relatively controlled environment. Second, this study is exposed to selection bias, as it is likely that multiple factors ranging from comorbidity burden to surgeon preference influenced the multimodal drugs selected. Our examination of comorbidities by medication prescribed suggests that comorbidities had a limited influence on prescribing patterns, but that they most commonly influenced the decision to withhold celecoxib in patients with ESRD/CKD or GERD. We attempted to control for bias from confounding variables using multivariate statistical techniques, but it is likely that other factors do influence our results. Third, at our institution, hydrocodone is only available on formulary as hydrocodone with acetaminophen. This presents the potential limitation that acetaminophen could be withheld in patients also being prescribed hydrocodone. A total of 525 patients in this study received hydrocodone with acetaminophen. However, only 11 patients received this medication in the preoperative or PACU settings, therefore suggesting that hydrocodone with acetaminophen was not commonly used as an alternative to the multimodal medications described. Finally, while we did use post-discharge outcomes of return to OR, and 30- and 90-day readmissions as quality control measures, we were unable to evaluate rates of specific complications or reasons for prolonged length of stay or high narcotic consumption. An opportunity exists to repeat this study in patients undergoing revision TKA, as prior studies have demonstrated that enhanced recovery protocols using multimodal pain management techniques are safe and effective for improving outcomes in this high-risk population.30
5. Conclusion
Multimodal pain management using acetaminophen, gabapentinoids, and celecoxib are effective at reducing length of stay and postoperative narcotic requirements after total knee arthroplasty. Patients appear to derive similar benefit from receiving all three medications, as well as various combinations of these drugs. Further research is required to tailor multimodal pain management approaches on a patient specific basis.
Footnotes
All work completed at Luminis Health Anne Arundel Medical Center.
Appendix A. Comorbidity Definitions based on ICD-10 Diagnosis Codes
| Comorbidity | ICD-10 Code(s) |
|---|---|
| Diabetes Mellitus | E08a, E09a,E10a, E11a, E13a |
| Sleep Apnea | G47.3 |
| COPD | J44a |
| Asthma | J45a |
| AFIB | I48a |
| CHF | I50a |
| CAD | I25a |
| ESRD or CKD | N18a, I12a, I13a |
| GERD | K21a |
| Primary HTN | I10a |
| Anemia | D50a - D53a & D55a - D59a |
Indicates all codes below that grouping level.
Appendix B. Oral Morphine Milligram Equivalents Conversion Factors
| Drug Name | Units | Oral MME Conversion Factor |
|---|---|---|
| Codeine (oral) | mg | 0.15 |
| Fentanyl (Injection) | mcg | 0.25 |
| Fentanyl (patch) | mcg | 2.4 |
| Hydrocodone (oral) | mg | 1 |
| Hydromorphone (injection) | mg | 20 |
| Hydromorphone (oral) | mg | 4 |
| Meperidine (injection) | mg | 0.4 |
| Meperidine (oral) | mg | 0.1 |
| Morphine (injection) | mg | 2.5 |
| Morphine (oral) | mg | 1 |
| Oxycodone (oral) | mg | 1.5 |
| Oxymorphone (oral) | mg | 3 |
| Tapentadol (oral) | mg | 0.4 |
References
- 1.Golladay G.J., Balch K.R., Dalury D.F., Satpathy J., Jiranek W.A. Oral multimodal analgesia for total joint arthroplasty. J Arthroplasty. 2017;32(9):S69–S73. doi: 10.1016/j.arth.2017.05.002. [DOI] [PubMed] [Google Scholar]
- 2.Li J.W., Ma Y.S., Xiao L.K. Postoperative pain management in total knee arthroplasty. Orthop Surg. 2019;11(5):755–761. doi: 10.1111/os.12535. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Pepper A.M., Mercuri J.J., Behery O.A., Vigdorchik J.M. Total hip and knee arthroplasty perioperative pain management. JBJS Reviews. 2018;6(12) doi: 10.2106/jbjs.rvw.18.00023. [DOI] [PubMed] [Google Scholar]
- 4.Trasolini N.A., McKnight B.M., Dorr L.D. The opioid crisis and the orthopedic surgeon. J Arthroplasty. 2018;33(11):3379–3382. doi: 10.1016/j.arth.2018.07.002. [DOI] [PubMed] [Google Scholar]
- 5.Schumacher C.S., Menendez M.E., Pagani N.R. Variation in perioperative opioid use after total joint arthroplasty. J Orthop. 2021;25:162–166. doi: 10.1016/j.jor.2021.05.003. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Schlosser M.J., Korwek K.M., Dunn R., Poland R.E. Reduced post-operative opioid use decreases length of stay and readmission rates in patients undergoing hip and knee joint arthroplasty. J Orthop. 2020;21:88–93. doi: 10.1016/j.jor.2020.03.003. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Baek H., Cho M., Kim S., Hwang H., Song M., Yoo S. Analysis of length of hospital stay using electronic health records: a statistical and data mining approach. PLoS One. 2018;13(4) doi: 10.1371/journal.pone.0195901. Published 2018 Apr 13. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Hunter A., Johnson L., Coustasse A. Reduction of intensive care unit length of stay: the case of early mobilization. Health Care Manag. 2014;33(2):128–135. doi: 10.1097/HCM.0000000000000006. [DOI] [PubMed] [Google Scholar]
- 9.American Society of Anesthesiologists Practice guidelines for acute pain management in the perioperative setting: an updated report by the American society of anesthesiologists task force on acute pain management. Anesthesiology. 2012;116(2):248–273. doi: 10.1097/aln.0b013e31823c1030. [DOI] [PubMed] [Google Scholar]
- 10.Elmallah R.K., Chughtai M., Khlopas A. Pain control in total knee arthroplasty. J Knee Surg. 2018;31(6):504–513. doi: 10.1055/s-0037-1604152. [DOI] [PubMed] [Google Scholar]
- 11.Hannon C.P., Keating T.C., Lange J.K., Ricciardi B.F., Waddell B.S., Della Valle C.J. Anesthesia and analgesia practices in total joint arthroplasty: a survey of the American association of hip and knee surgeons membership. J Arthroplasty. 2019;34(12):2872–2877. doi: 10.1016/j.arth.2019.06.055. e2. [DOI] [PubMed] [Google Scholar]
- 12.Fillingham Y.A., Hannon C.P., Roberts K.C. AAHKS anesthesia & analgesia clinical practice guideline workgroup, Hamilton WG, della valle CJ. Nonsteroidal anti-inflammatory drugs in total joint arthroplasty: the clinical practice guidelines of the American association of hip and knee surgeons, American society of regional anesthesia and pain medicine, American academy of orthopaedic surgeons, hip society, and knee society. J Arthroplasty. 2020 Oct;35(10):2704–2708. doi: 10.1016/j.arth.2020.05.043. Epub 2020 May 29. PMID: 32571593. [DOI] [PubMed] [Google Scholar]
- 13.Fillingham Y.A., Hannon C.P., Roberts K.C. The efficacy and safety of nonsteroidal anti-inflammatory drugs in total joint arthroplasty: systematic review and direct meta-analysis. J Arthroplasty. 2020 Oct;35(10):2739–2758. doi: 10.1016/j.arth.2020.05.035. Epub 2020 May 28. PMID: 32690428. [DOI] [PubMed] [Google Scholar]
- 14.Shin S. Safety of celecoxib versus traditional nonsteroidal anti-inflammatory drugs in older patients with arthritis. J Pain Res. 2018;11:3211–3219. doi: 10.2147/JPR.S186000. Published 2018 Dec 14. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Nissen S.E., Yeomans N.D., Solomon D.H. Cardiovascular safety of celecoxib, naproxen, or ibuprofen for arthritis. N Engl J Med. 2016;375(26):2519–2529. doi: 10.1056/NEJMoa1611593. [DOI] [PubMed] [Google Scholar]
- 16.Candiotti K.A., Bergese S.D., Viscusi E.R., Singla S.K., Royal M.A., Singla N.K. Safety of multiple-dose intravenous acetaminophen in adult inpatients. Pain Med. 2010;11(12):1841–1848. doi: 10.1111/j.1526-4637.2010.00991.x. [DOI] [PubMed] [Google Scholar]
- 17.Fillingham Y.A., Hannon C.P., Erens G.A. AAHKS anesthesia & analgesia clinical practice guideline workgroup, Hamilton WG, della valle CJ. Acetaminophen in total joint arthroplasty: the clinical practice guidelines of the American association of hip and knee surgeons, American society of regional anesthesia and pain medicine, American academy of orthopaedic surgeons, hip society, and knee society. J Arthroplasty. 2020 Oct;35(10):2697–2699. doi: 10.1016/j.arth.2020.05.030. Epub 2020 May 26. PMID: 32571591. [DOI] [PubMed] [Google Scholar]
- 18.Fillingham Y.A., Hannon C.P., Erens G.A. The efficacy and safety of acetaminophen in total joint arthroplasty: systematic review and direct meta-analysis. J Arthroplasty. 2020 Oct;35(10):2715–2729. doi: 10.1016/j.arth.2020.05.037. Epub 2020 May 28. PMID: 32563592. [DOI] [PubMed] [Google Scholar]
- 19.Ohnuma T., Raghunathan K., Moore S. Dose-dependent association of gabapentinoids with pulmonary complications after total hip and knee arthroplasties. J Bone Joint Surg. 2020;102(3):221–229. doi: 10.2106/jbjs.19.00889. [DOI] [PubMed] [Google Scholar]
- 20.Lee J.K., Chung K.-S., Choi C.H. The effect of a single dose of preemptive pregabalin administered with COX-2 inhibitor: a trial in total knee arthroplasty. J Arthroplasty. 2015;30(1):38–42. doi: 10.1016/j.arth.2014.04.004. [DOI] [PubMed] [Google Scholar]
- 21.Hannon C.P., Fillingham Y.A., Browne J.A., Schemitsch E.H. AAHKS anesthesia & analgesia clinical practice guideline workgroup, buvanendran A, Hamilton WG, della valle CJ. Gabapentinoids in total joint arthroplasty: the clinical practice guidelines of the American association of hip and knee surgeons, American society of regional anesthesia and pain medicine, American academy of orthopaedic surgeons, hip society, and knee society. J Arthroplasty. 2020 Oct;35(10):2700–2703. doi: 10.1016/j.arth.2020.05.031. Epub 2020 May 26. PMID: 32616442. [DOI] [PubMed] [Google Scholar]
- 22.Hannon C.P., Fillingham Y.A., Browne J.A. The efficacy and safety of gabapentinoids in total joint arthroplasty: systematic review and direct meta-analysis. J Arthroplasty. 2020 Oct;35(10):2730–2738. doi: 10.1016/j.arth.2020.05.033. Epub 2020 May 26. PMID: 32586656. [DOI] [PubMed] [Google Scholar]
- 23.Hannon C.P., Fillingham Y.A., Nam D. AAHKS anesthesia & analgesia clinical practice guideline workgroup. Opioids in total joint arthroplasty: the clinical practice guidelines of the American association of hip and knee surgeons, American society of regional anesthesia and pain medicine, American academy of orthopaedic surgeons, hip society, and knee society. J Arthroplasty. 2020 Oct;35(10):2709–2714. doi: 10.1016/j.arth.2020.05.034. Epub 2020 May 26. PMID: 32571594. [DOI] [PubMed] [Google Scholar]
- 24.Hannon C.P., Fillingham Y.A., Nam D. The efficacy and safety of opioids in total joint arthroplasty: systematic review and direct meta-analysis. J Arthroplasty. 2020 Oct;35(10):2759–2771. doi: 10.1016/j.arth.2020.05.032. Epub 2020 May 27. PMID: 32571589. [DOI] [PubMed] [Google Scholar]
- 25.Guerra M.L., Singh P.J., Taylor N.F. Early mobilization of patients who have had a hip or knee joint replacement reduces length of stay in hospital: a systematic review. Clin Rehabil. 2015;29(9):844–854. doi: 10.1177/0269215514558641. [DOI] [PubMed] [Google Scholar]
- 26.Ellis T.A., 2nd, Hammoud H., Dela Merced P. Multimodal clinical pathway with adductor canal block decreases hospital length of stay, improves pain control, and reduces opioid consumption in total knee arthroplasty patients: a retrospective review. J Arthroplasty. 2018;33(8):2440–2448. doi: 10.1016/j.arth.2018.03.053. [DOI] [PubMed] [Google Scholar]
- 27.Hebl J.R., Dilger J.A., Byer D.E. A pre-emptive multimodal pathway featuring peripheral nerve block improves perioperative outcomes after major orthopedic surgery. Reg Anesth Pain Med. 2008;33(6):510–517. [PubMed] [Google Scholar]
- 28.Memtsoudis S.G., Poeran J., Zubizarreta N. Association of multimodal pain management strategies with perioperative outcomes and resource utilization: a population-based study. Anesthesiology. 2018;128(5):891–902. doi: 10.1097/ALN.0000000000002132. [DOI] [PubMed] [Google Scholar]
- 29.Rucinski K., Cook J.L. Effects of preoperative opioid education on postoperative opioid use and pain management in orthopaedics: a systematic review. J Orthop. 2020;20:154–159. doi: 10.1016/j.jor.2020.01.020. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Joseph Z., Calvert N., Salmon M. Enhanced recovery principles applied to revision hip and knee arthroplasty leads to better patient outcomes. J Orthop. 2020;22:543–547. doi: 10.1016/j.jor.2020.11.004. [DOI] [PMC free article] [PubMed] [Google Scholar]