Can Multimodal Pain Management in TKA Eliminate Patient-controlled Analgesia and Femoral Nerve Blocks?

Abstract

Background

TKA pain management protocols vary widely with no current consensus on a standardized pain management regimen. Multimodal TKA pain management protocols aim to address pain control, facilitate functional recovery, and maintain patient satisfaction.

Questions/purposes

(1) Did changes to our pain management protocol, specifically adding liposomal bupivacaine, eliminating patient-controlled analgesia (PCA), and discontinuing femoral nerve blocks (FNBs), affect narcotic consumption after TKA? (2) Did these changes to our pain management protocols affect patient-reported pain scores? (3) Does the use of an immediate postoperative PCA affect rapid rehabilitation and functional recovery? (4) How did changes to our pain management regimen affect discharge disposition and pain-related Hospital Consumer Assessment of Healthcare Providers and Systems (HCAHPS) scores?

Methods

We retrospectively analyzed an institutional arthroplasty database between September 2013 and September 2015 containing 1808 patients who underwent primary TKA. Departmental pain management protocols were compared in 6-month periods as the protocol changed. All patients received a multimodal pain management protocol including preoperative oral medications, spinal or general anesthesia, a short-acting intraoperative pericapsular injection, and continued postoperative oral narcotics for breakthrough pain. From September 2013 to April 2014, all patients received an intraoperative FNB and a PCA for the first 24 hours postoperatively (Cohort 1). From May 2014 to October 2014, a periarticular injection of liposomal bupivacaine was added to the protocol and FNBs were discontinued (Cohort 2). After April 2015, PCA was eliminated (Cohort 3). No other major changes were made to the TKA pain management pathways. Narcotic use, pain scores on 8-hour intervals, physical therapy milestones, and discharge disposition were compared.

Results

Total narcotic use was the least in Cohort 3 (Cohort 3: 66 ± 54 morphine milligram equivalents versus Cohort 2: 82 ± 72 versus Cohort 1: 96 ± 62; p < 0.001). There was an increase in pain score immediately after surgery in Cohort 3 (4.0 ± 3.5 versus 1.2 ± 2.2 versus 1.2 ± 2.5, post hoc analysis of Cohort 2 versus 3: mean difference 2.6, 95% confidence interval [CI] 2.2-3.0; p < 0.001); however, it was not different for the remainder of the hospital stay. Patients who did not receive PCA reached functional milestones for both gait and stairs faster by postoperative day 1 (47% [328 of 698] versus 30% [158 of 527] versus 16% [93 of 583], p < 0.001; Cohort 3 versus 2: odds ratio 2.1, 95% CI 1.6-2.6; p < 0.001). Discharge to home occurred more frequently (84% [583 of 698] versus 78% [410 of 527] versus 72% [421 of 583], p = 0.010) in Cohort 3. There were no differences in pain-related HCAHPS scores across all cohorts.

Conclusions

Discontinuing PCAs and FNBs from our multimodal TKA pain management protocols and adding liposomal bupivacaine resulted in fewer narcotics consumed with no difference in pain control and faster functional recovery while maintaining high HCAHPS scores relating to pain.

Level of Evidence:

Level III, therapeutic study.

Introduction

Pain control in TKA is a crucial component to TKA care. Improving TKA pain management strategies has been examined with renewed interest in a value-based healthcare environment, because postoperative pain control has been associated with early ambulation and functional recovery as well as patient satisfaction [3, 4, 9]. Inadequate pain control leads to prolonged hospitalization, inferior functional outcomes, and patient dissatisfaction [4, 8, 26].

A multimodal approach to pain management effectively addresses postoperative pain in TKA. This approach involves interventions during all phases of care throughout the perioperative period. Multimodal protocols have resulted in improved pain control and increased patient satisfaction [4, 17]. However, these protocols vary widely between institutions with no current consensus on a standardized regimen. Preemptive analgesia [9], peripheral anesthesia techniques such as femoral nerve block (FNB) or adductor canal blocks [2], intraoperative periarticular injections (PAIs) [12] with mixed drug regimens or liposomal bupivacaine [3], patient-controlled analgesia (PCA), and oral narcotics [16, 18] have all been part of reported regimens.

These techniques may have certain limitations as well. FNBs have been associated with increased risk of falls and delayed rehabilitation [10, 23]. PAIs containing bupivacaine may have an average half-life < 3 hours [12]. Liposomal bupivacaine may provide sustained release for up to 48 to 72 hours, but some studies question its cost-effectiveness compared with more traditional alternatives [1, 22]. Several small, randomized clinical trials have been conducted with mixed results [6, 21, 24]. Furthermore, two recently published meta-analyses regarding liposomal bupivacaine resulted in conflicting conclusions [15, 25]. Although narcotics have been the mainstay of postoperative pain control and PCA is a useful modality to provide immediate pain control, its use can lead to excessive dosing, rehabilitation delays as well as complications and resource utilization.

We earlier reported successful use of liposomal bupivacaine in lieu of FNBs [27]. As part of a quality improvement effort, our institution has made stepwise adjustments to refine our pain management protocols, which in the past utilized all of these aforementioned modalities in some combination. Furthermore, the efficacy and value of liposomal bupivacaine as part of multimodal regimens remain unclear. We evaluated these systematic adjustments made to our pain management protocols to answer the following questions: (1) Did changes to our pain management protocols, specifically eliminating PCA and FNB, affect narcotic consumption after TKA? (2) Did changes to our pain management protocols affect patient-reported pain scores? (3) Does the use of immediate postoperative PCA affect rapid rehabilitation and functional recovery? (4) How did changes to our pain management regimen influence discharge disposition and pain-related Hospital Consumer Assessment of Healthcare Providers and Systems (HCAHPS) scores?

Patients and Methods

Institutional review board approval was obtained to conduct this retrospective study at a single institution, which consisted of 1808 patients undergoing unilateral, primary TKA by 30 orthopaedic surgeons between September 2013 and September 2015. Three patient cohorts were compiled by date, from September 2013 to April 2014 (Cohort 1: 583 patients), from May to October 2014 (Cohort 2: 527 patients), and from April 2015 to September 2015 (Cohort 3: 698 patients). Patients were excluded from this study if they received a simultaneous or second staged primary TKA, possessed inadequate records within the electronic medical record, or deviated from the standard protocol. Adoption of liposomal bupivacaine began in May 2014 and became routinely used in all patients undergoing total joint arthroplasty at our institution. Also during that time, FNBs were completely discontinued as part of our pain management protocol. In February 2015, routine PCA use was eliminated from the standard protocol. No other variables changed in the TKA care pathway during these time intervals.

Each patient included in this study received the pain management protocol. In Cohort 1, a preinduction, ultrasound-guided, single-shot FNB consisting of a slow fractionated injection of 20 cc 0.25% bupivacaine was used as an adjunct to the analgesic regimen. In Cohort 2, patients did not receive FNBs and were intraoperatively given a liposomal bupivacaine injection. In summary, Cohort 1 consisted of a single-shot FNB with PCA administration for the first 24 hours postoperatively, Cohort 2 consisted of an intraarticular injection of liposomal bupivacaine and PCA administration for the first 24 hours postoperatively, and Cohort 3 consisted of an intraarticular injection of liposomal bupivacaine with no PCA administration (Table 1).

Table 1.

Descriptive statistics between the two cohorts

All three cohorts during the postoperative period were controlled using oral and IV narcotics by patient request for breakthrough pain. The described differences in the pain management protocol were made to the department’s TKA Standard Clinical Pathway during the study period.

Of the 1808 patients included in our study, 583 were in Cohort 1 (FNB + PCA, no liposomal bupivacaine), 527 in Cohort 2 (liposomal bupivacaine + PCA, no FNB), and 698 in Cohort 3 (liposomal bupivacaine, no PCA, no FNB). There were no significant differences among the descriptive statistics of the three populations (Table 1).

Various multimodal analgesia protocols were applied to all patients in this study (Fig. 1). Before entering the operating room, patients in all three cohorts received three preemptive oral analgesics (200 mg celecoxib, 1000 mg acetaminophen, and 50 mg pregabalin). Intraoperative analgesia was chosen at the discretion of the anesthesiologist and preferentially consisted of spinal anesthesia with general anesthesia as a backup option. Just before closure, an analgesic mixture was injected into the overlying soft tissue and muscle as well as the periincisional subcutaneous layer, consisting of 40 cc 0.25% Marcaine (Hospira, Lake Forest, IL, USA), 5 cc of 1 mg/mL Duramorph (West-Ward, Eatontown, NJ, USA), and 1 cc of 30 mg/mL Toradol (Regency, Shirley, NY, USA) for patients in all three cohorts.

Fig. 1.

The multimodal analgesia regimen of each cohort is displayed. PRN = as needed.

Liposomal Bupivacaine Administration

All surgeons included in this study underwent standardized teaching protocols concerning the liposomal bupivacaine administration technique, which is crucial for efficacy. The liposomal bupivacaine injection (Exparel®; Pacira Pharmaceuticals Inc, Parsippany, NJ, USA) is prepared intraoperatively by the surgical technician before placing the final components in the knee. A homogeneous solution is created with 20 cc liposomal bupivacaine in 40 cc 0.9% normal saline solution. A total volume of 60 cc is separated into 3- to 20-cc aliquots and infiltrated to the knee in a three-layer fashion with equal amounts injected into each of the (1) posterior capsule; (2) overlying periosteum, muscle, and fascia; and (3) the subcutaneous fat and subcuticular layer using a 21-gauge needle. The posterior capsule is injected just before implantation of the final components, and the rest of the allocation is administered to the other two layers just before closure of the arthrotomy. Careful allocation of liposomal bupivacaine throughout the wound is essential for effectiveness, because liposomal bupivacaine does not widely diffuse to areas where it is not injected [13].

All patients received the standardized TKA clinical pathway and rehabilitation. Metrics recorded included (1) patient-reported pain scores (an 11-point Numeric Pain Rating Scale) collected within regular 2- to 8-hour intervals by nursing staff; (2) all narcotics administered postoperatively; and (3) physical therapy (PT) milestones, specifically when a patient functionally achieves stairclimbing and ambulates > 100 feet. A knee immobilizer was not used postoperatively for the patients who received a FNB (Cohort 1). Patients were discharged when they were medically stable, when pain was under adequate control, and when PT milestones were achieved. On discharge, each patient submitted an HCAHPS score providing patient-centered feedback. Two major HCAHPS domains were analyzed: (1) answer of “always” in overall pain management; and (2) answer of “always” for how well the pain was controlled. Patient discharge disposition (home versus rehabilitation facility) was also collected.

Patient information was abstracted from the electronic medical record (Epic Systems Corporation, Verona, WI, USA), including procedure type, date of surgery, age, gender, body mass index, and American Society of Anesthesiologists score. Pain scores were collected and recorded in 8-hour intervals, starting from the time of postanesthesia care unit admission and continuing through postoperative floor admission to the time of discharge. The pain score closest to the 8-hour mark was recorded for that time interval. Narcotic use was aggregated per postoperative day (POD) and converted to morphine equivalent dosages [14]. PT milestones (gait distance and stairclimbing) were recorded as the best effort for each POD.

Statistical Analysis

Descriptive statistics were used to compare the three cohorts. Analysis of variance with a post hoc Tukey test was utilized to detect statistical differences among the means of each variable of the three groups, chi-square analyses were used for categorical data, and a nonparametric Kruskal-Wallis test was utilized to detect differences in median for narcotic use and ordinal data. Statistical significance was set at p ≤ 0.05.

Results

Total narcotic use was the least in Cohort 3 (Cohort 3: 66 ± 54 morphine milligram equivalents [mme] versus Cohort 2: 82 ± 72 versus Cohort 1: 96 ± 62). Post hoc analysis pairwise comparison of Cohort 3 versus Cohort 2 demonstrated an increase in POD 1 narcotic use (23 ± 17 versus 12 ± 16 mme; mean difference [MD] 11; 95% confidence interval [CI] 9-13; p < 0.001); however, it was decreased for every POD thereafter (Fig. 2).

Fig. 2.

The average morphine equivalent narcotic consumption for each cohort was compiled per postoperative day.

There was an increase in pain score immediately after surgery in Cohort 3 (4.0 ± 3.5 versus 1.2 ± 2.2 versus 1.2 ± 2.5, post hoc analysis of Cohort 2 versus Cohort 3: MD 2.6, 95% CI 2.2-3.0; p < 0.001), which is greater than the minimal clinically important difference of approximately 2 points on a 10-point scale [20]. However, there were no statistical or minimal clinically important differences in pain score trends after 8 hours postoperatively among all three cohorts (Fig. 3).

Fig. 3.

The average pain scores are reported for each cohort throughout the hospital length of stay on an 8-hour basis.

Patients who did not receive a PCA or FNB experienced accelerated functional milestones for both gait and stairs. Most notably, there was a higher proportion of patients reaching the functional milestones by POD 1 in Cohort 3 (47% [328 of 698] versus 30% [158 of 527] versus 16% [93 of 583], p < 0.001; Cohort 3 versus Cohort 2: odds ratio [OR] 2.1, 95% CI 1.6-2.6). Overall, there was no difference in Cohort 2 and Cohort 3 patients ultimately reaching these milestones (Cohort 3: 90% [628 of 698], Cohort 2: 93% [490 of 527], OR 1.5 95% CI 0.97-2.2; p = 0.063), whereas Cohort 1 demonstrated a lower proportion compared with the other two cohorts (73% [426 of 583]) (Fig. 4).

Fig. 4.

A-B The day on completion of each physical therapy milestone was recorded, specifically for gait requirements of ambulating at least 100 feet (A, top) and climbing stairs (B, bottom).

There was an overall trend of increased discharge to home with Cohort 3 demonstrating the highest proportion of home discharge (84% [583 of 698] versus 78% [410 of 527] versus 72% [421 of 583], p = 0.010). The proportion of patient-reported HCAHPS relating to pain being always “well controlled” was not different among all three cohorts (74% [517 of 698] versus 70% [370 of 527] versus 69% [402 of 583]; p = 0.214). Similarly, an “always satisfied” score of “overall pain management” was not different among all three cohorts (79% [551 of 698] versus 77% [406 of 527] versus 77% [449 of 583]; p = 0.463).

Discussion

Achieving adequate pain control after TKA is challenging, and many different versions of multimodal regimens have been found to be effective [4-6]. As focus has shifted to quality and outcome-based reimbursement paradigms, pain management plays a vital role in functional outcomes, patient satisfaction, and cost-effectiveness [7, 17, 26]. Our study demonstrates that stepwise adjustments made to a multimodal pain regimen resulted in fewer narcotics consumed, earlier functional recovery, improvement of quality outcomes, maintenance of patient-reported pain control, and pain-related HCAHPS survey responses.

This study has several limitations. First are the limitations inherent to a retrospective study. Missing or incorrect data may result in retrieval error. The state of the contralateral knee is unable to be adequately standardized for analysis with retrospective data. Furthermore, preoperative narcotic use was not analyzed. Lastly, deviations from the protocol may not have been recorded and not appropriately excluded from our study. However, the large series of patients should help minimize this error. Additionally, there was no statistically significant difference in data exclusion among the three cohorts. Second, historical cohorts were used for analysis. Certainly time bias exists; however, this design allows for accrual of large series of patients during which a standardized protocol was implemented. Additionally, the study interval is short (approximately 6 months per cohort) and unlikely represents a severe limitation. The only two notable changes to our TKA pathway are represented by each cohort. Third, although this study was conducted at a single institution, 30 different orthopaedic surgeons contributed to the database. This is also unlikely to represent a considerable limitation, because each surgeon undergoes scrutiny for adherence to protocols as well as standardized training for proper liposomal bupivacaine administration and technique. Invariably, there may still be variation in how liposomal bupivacaine was administered, and its efficacy is highly technique-dependent. Fourth, patient-reported pain scores are subjective and there is likely wide variability of patient characteristics that confound how each patient responds to pain. Ideally, preoperative pain scores or narcotic use would be helpful to normalize this subjectivity.

It is also important to note that Cohort 3 contained more women in the patient cohort compared with the other two cohorts. Although this poses a difference in patient populations, we do not believe this demographic affects our results or conclusions. A subgroup analysis between sexes was conducted for Cohort 3 to investigate differences in pain score, narcotic use, and functional milestone achievement. Regarding pain score, the only statistically significant difference between sexes was at the initial pain score recording with females reporting a higher pain score versus males (4.2 ± 3.5 versus 3.1 ± 3.3, p < 0.001). This is below the minimal clinically important difference, however. Regarding narcotic use, males utilized more narcotics on the first postoperative day (31.2 ± 25.3 mme versus 27.1 ± 18.2, p = 0.017); otherwise use was not different on subsequent days. For PT milestone achievement as well as discharge disposition, there was no significant difference in outcome between sexes. Despite the differences outlined by the subgroup analysis, we feel that sex discrepancy noted in Cohort 3 minimally contributes to the results as presented, because the majority of endpoints is not different between sexes and for the two highlighted endpoints that were statistically different, the contributory effect is unclear.

Finally, time bias likely exists in the improvement of the discharge disposition metric reported. In today’s value-based environment, it is important to note that institutional economic motivation to discharge home versus a postacute rehabilitation facility may confound the results observed here. Although this may indeed have an effect, we feel that the degree is small, because disposition is primarily affected by achieving PT milestones. Therefore, the ability to achieve early functional recovery is the critical factor in improving these quality metrics, regardless of external pressure.

A stepwise decrease in narcotic use was observed for each adjustment in our pain management protocol. First, the addition of liposomal bupivacaine and discontinued use of FNB led to slightly less narcotics consumed. After discontinuing the PCA from our regimen, a further and more substantial reduction was observed. Alleviation of narcotic burden is a key to achieving successful outcomes after TKA, because patients will experience a reduction in side effects and the facilitation of early rehabilitation. Dasta et al. [8] reported similar experiences with the addition of liposomal bupivacaine to their patients undergoing TKA, noting reductions in opioid use and opioid-related adverse events. The present study echoes these benefits with the addition of liposomal bupivacaine to a multimodal regimen and further substantiates the ability to rely on nonnarcotic means of pain control.

None of the cohorts exhibited statistically or clinically different pain levels and trends beyond 8 hours postoperatively. Barrington et al. [3] demonstrated improved pain scores in patients with liposomal bupivacaine-treated total joint arthroplasty, specifically a higher percentage of visual analog scale scores that were 0 (48.8% versus 40.2%; p < 0.001). Conversely, Jain et al. [11] in a small series determined that the addition of liposomal bupivacaine did not affect pain or visual analog scores and narcotic use. A recent meta-analysis studying the efficacy of liposomal bupivacaine in several randomized trials concluded that there was no difference in pain scores or narcotic use between liposomal bupivacaine and standard bupivacaine [15]. The included studies in this analysis, however, were all small trials. The true effectiveness of liposomal bupivacaine and its role in TKA remain to be determined; however, it seems clear that some type of a modern, periarticular local analgesic is an effective addition to a multimodal pain regimen, whether it is a short-acting cocktail or a combination of short- and long-acting cocktails [6, 12, 22].

In this study, patients undergoing TKA who did not receive a FNB or postoperative PCA experienced accelerated functional milestones for both gait and stairs. We believe that pain control should be achieved with the minimum amount of narcotics to avoid opioid-related adverse effects (ORAEs). The combination of liposomal bupivacaine and the periincisional short-acting injection made possible the elimination of PCA and the reliance on narcotics. Second, the ability to avoid early motor blockade from FNB as well as any hindrance to mobility (PCA, IV lines) allowed for more rapid rehabilitation in the early postoperative setting. As a result, a higher proportion of patients reached functional milestones by POD 1. Although FNB is an effective mode of postoperative analgesia in TKA, the associated motor blockade delays early rehabilitation [2, 10, 23]. PCA use also has unwanted associated complications including ORAEs and delays in rehabilitation. Langford et al. [16] demonstrated that PCA use versus an alternative fentanyl transdermal delivery poses delays in early rehabilitation.

Each subsequent modification to our pain management protocol resulted in decreased narcotic burden, equivalent pain control, and faster functional recovery. This, in turn, led to improvement in discharge disposition. PCA, although an effective mode of immediate postoperative pain control, is associated with substantial cost as a result of equipment needs, increased monitoring, and complications from overdosing [18, 19]. During the first 8 hours postoperatively, Cohort 3 reported an average pain score of 4.0, which was increased compared to the other two cohorts. A pain score of above 4.0 is an indication of a need for treatment in the postoperative phase of care. Our next iterative improvement process in the multimodal approach is to add intravenous Tylenol to the protocol to address this pain in the immediate post op period in the absence of PCA.

HCAHPS scores are becoming a relevant metric in quality and reimbursement. Maher et al. [17] conducted a study on pain-related HCAHPS scores in patients undergoing total joint arthroplasty and concluded that scores were negatively associated with greater opioid requirement and longer length of stay. We demonstrated that “best-response” pain-related questions were unchanged despite making adjustments to our pain management protocols that reduced narcotic use. Patient-centered feedback is crucial in determining the value of care delivered, because quality-driven interventions cannot accept detriments to the quality of patient care. Our results demonstrated that the discontinuation of the PCA did not affect our pain-related HCAHPS.

Discontinuing PCAs, eliminating FNBs, and adding liposomal bupivacaine to our multimodal TKA pain management protocols resulted in decreases in narcotic consumption, similar pain control as determined by the minimal clinically important difference, and faster achievement of functional milestones. All of this was achieved without affecting our pain-related HCAHPS scores. Multimodal protocols combining modern intraarticular anesthetics such as liposomal bupivacaine and a short-acting pain cocktail can be helpful in eliminating postoperative narcotic burden and facilitating early rehabilitation. Dedicated studies investigating the cost-effectiveness among various multimodal agents are required to substantiate adoption of newer agents in today’s medicoeconomic environment. Ultimately, more research is needed to develop superior analgesic modalities and devise strategies to improve our clinical TKA pathways and increase the value of care for our patients.