Discharge Readiness after Tricompartment Knee Arthroplasty: Adductor Canal versus Femoral Continuous Nerve Blocks—A Dual-center, Randomized Trial

TRICOMPARTMENT knee arthroplasty is among the most common surgical procedures, with over 700,000 procedures performed annually within the United States alone.¹  This surgery results in moderate-to-severe postoperative pain that often requires intravenous analgesics and impairs functional mobility, all of which can prolong hospitalization.²  Continuous femoral nerve blocks provide effective postoperative analgesia and are therefore widely used to provide analgesia after knee arthroplasty.³  However, this modality induces quadriceps weakness and is associated with an increased risk of falling,^4–6  leading clinicians to question its risk–benefit ratio and consider alternative analgesic approaches.^7–11 

A relatively new alternative is an adductor canal block in which local anesthetic is deposited within an aponeurotic tunnel in the middle third of the thigh containing multiple afferent sensory nerves—but only a single efferent motor nerve: a branch innervating the vastus medialis of the quadriceps muscle.^12,13  In both volunteers and surgical patients, a single-injection adductor canal block induces less quadriceps weakness and mobilization disability compared with a single-injection femoral nerve block.^14–16 

Because pain after knee arthroplasty usually outlasts the duration of a single-injection nerve block, a perineural catheter is often introduced to allow prolonged local anesthetic administration.^15,17–21  Two randomized comparisons of adductor canal and femoral continuous nerve blocks after knee arthroplasty demonstrated the quadriceps-sparing benefit of the adductor canal infusions, but both limited infusion duration to 24 or fewer hours and evaluated mobilization a maximum of 24 h.^20,22  Because discharge readiness is rarely achieved within the first 24 h after tricompartment knee arthroplasty, it remains unknown whether the use of continuous adductor canal blocks will have any appreciable effect on readiness for discharge.²³ 

We therefore conducted a dual-center, randomized, active-controlled, parallel-arm clinical trial to test the hypothesis that a continuous adductor canal block decreases the time to attain four specific discharge criteria compared with a continuous femoral nerve block after tricompartment knee arthroplasty: (1) adequate analgesia; (2) independence from intravenous opioids; (3) ability to independently stand; walk 3 m, return, and sit down; and (4) independently ambulate 30 m. Both treatment groups received intraoperative joint infiltration of ropivacaine, ketorolac, epinephrine, and tranexamic acid.

This study followed Good Clinical Practice and was conducted within the ethical guidelines outlined in the Declaration of Helsinki. The trial was prospectively registered at ClinicalTrials.gov (NCT01759277). The University of California San Diego Institutional Review Board (San Diego, California) approved all study procedures and provided oversight of the data and safety issues for the duration of the trial. Written, informed consent was obtained from all participating subjects.

Enrollment was conducted exclusively through the orthopedic clinic and was offered to adults (≥18 yr old) scheduled for primary, unilateral, tricompartment knee arthroplasty whose postoperative analgesic plan included a perineural local anesthetic infusion. Exclusion criteria were morbid obesity (body mass index >40 kg/m²), chronic high-dose opioid use (daily oxycodone equivalents >20 mg within 2 weeks before surgery and duration of use >4 weeks), history of opioid abuse, allergy to study medications, known renal insufficiency (creatinine >1.5 mg/dl), pregnancy, incarceration, any known neuromuscular deficit of the ipsilateral femoral nerve, obturator nerve, and quadriceps muscle (including diabetic neuropathy), and inability to ambulate 30 m preoperatively. The study was conducted at Thornton and Hillcrest hospitals, both of which are academic institutions in San Diego, California.

Both femoral and adductor canal sites were visualized with ultrasound using a 13–6 MHz 38-mm linear array transducer (M-Turbo; SonoSite, USA). Subjects were randomized to one of the two treatment groups—adductor canal versus femoral perineural catheter—only if both locations were considered acceptable for catheter insertion. Randomization lists were created by Investigational Drug Service personnel using a computer-generated randomization table in blocks of four, with a 1:1 ratio, stratified by both treatment center and surgeon. Treatment allocation was concealed using consecutively numbered, sealed, opaque envelopes that were opened only after confirmation by ultrasound that either insertion site would be acceptable.

Blocks were performed preoperatively by regional anesthesia fellows under the guidance of attending anesthesiologists or by attending anesthesiologists. Subjects were positioned supine, given supplemental oxygen via nasal cannula, premedicated with intravenous fentanyl and midazolam, and prepped and draped in a sterile manner. The same perineural catheter kits were used for both treatment groups (FlexBlock; Teleflex Medical, Research Triangle Park, USA).

A transverse cross-sectional (short axis) view of the femoral nerve was obtained at the inguinal crease with the same ultrasound and transducer used for the initial scan before sterile preparation and draping. A local anesthetic skin wheal was raised lateral to the ultrasound transducer with a 25-gauge needle and lidocaine 1.5%. A 17-gauge Tuohy needle was inserted through the skin wheal in plane with the ultrasound probe toward the femoral nerve. The final needle position for the femoral group was posterior to the femoral nerve and lateral to the femoral artery.

The midpoint between the anterior superior iliac spine and the cephalad margin of the patella was measured with slight external rotation of the leg at the hip. Using ultrasound at this midpoint, the superficial femoral artery was identified in a short-axis view deep to the sartorius muscle, medial to the rectus femoris muscle and anterior to the adductor longis muscle, with the superficial femoral vein usually posterolateral to the artery (the position of the vein may vary relative to the artery at this level of the thigh). At this position, the ultrasound image of the saphenous nerve or nerve bundle is typically anterolateral to the superficial femoral artery within the adductor canal. The corresponding anatomical positions for each structure are saphenous nerve anterior to the artery, vein lateral to the artery, and sartorius medial to the artery.^24,25  A local anesthetic skin wheal was raised anterolateral to the ultrasound transducer with a 25-gauge needle and lidocaine 1.5%. A 17-gauge Tuohy needle was inserted in plane from the anterolateral side of the transducer, through the sartorius muscle with the final needle tip positioning between the artery and the saphenous nerve. If the saphenous nerve could not be well visualized, the needle tip was placed at 5 o’clock relative to the femoral artery within the adductor canal.²⁶ 

In both treatment groups, normal saline was injected via the needle for hydrodissection in the minimal amount necessary to open a space for catheter insertion. A 19-gauge perineural catheter was subsequently inserted 3 to 5 cm past the needle tip. The needle was withdrawn over the catheter, and the catheter hub affixed to the upper lateral thigh with sterile occlusive dressings and an anchoring device. Thirty milliliters of lidocaine 2% was injected via the catheter in divided doses after negative aspiration.

Catheter insertion success was defined as a change in cutaneous sensation to touch with an alcohol pad in the saphenous nerve distribution over the medial leg within 30 min after injection. Subjects with successful catheter placement per protocol and nerve block onset were retained in the study. Subjects with a failed catheter insertion or misplaced catheter indicated by a lack of sensory changes had their catheter replaced or were withdrawn from the study. A ropivacaine 0.2% infusion was initiated via the perineural catheter with a basal rate of 6 ml/h, a 4-ml bolus, and a lock-out of 30 min using a portable, programmable, electronic infusion pump (ambIT PreSet; Summit Medical, USA).

For surgical anesthesia, subjects received either a single- injection spinal with bupivacaine 0.5% (2 to 3 ml) or a general anesthetic with inhaled volatile anesthetic in nitrous oxide and oxygen. Intravenous fentanyl, hydromorphone, and/or morphine were administered intraoperatively, as needed. Implants were fixed with methyl methacrylate bone cement via a parapatellar approach (a tourniquet was used for all cases). After joint closure, the surgeon infiltrated the entire joint using 30 ml ropivacaine (0.5%), ketorolac (30 mg), epinephrine (5 μg/ml), and tranexamic acid (2 g).

All subjects received oral acetaminophen (975 mg every 6 h), celecoxib (200 mg every 12 h), and sustained release oxycodone (oxycontin, 10 mg every 12 h). For breakthrough pain, subjects depressed the infusion pump bolus button (4 ml, 30-min lock-out). When necessary, rescue opioid and route of administration were titrated to pain severity using a numeric rating scale (NRS) of 0 to 10: mild pain (NRS <4): oral oxycodone (5 mg); moderate pain (NRS 4 to 7): oral oxycodone (10 mg); and severe pain (NRS >7): intravenous morphine (2 to 4 mg) or hydromorphone (0.5 mg). Within the postanesthesia care unit, a once-only 10 ml lidocaine (2%) bolus was given via the perineural catheter for moderate or severe pain.

The basal infusion rate was initiated at 6 ml/h and titrated to subject comfort (increased 2 ml/h for NRS >4) a maximum of twice per day up to a maximum of 12 ml/h. However, muscle strength took precedence to allow for ambulation and minimize the risk of falling. If the physical therapist determined that subject standing and/or ambulation was inhibited by quadriceps weakness, the basal infusion rate was titrated down (decrease 2 ml/h) a maximum of twice per day to a minimum of 2 ml/h. Ropivacaine perineural infusions were continued until the morning of postoperative day (POD) 3.

Failure to meet four criteria determine the majority of hospitalization days at our hospitals: (1) adequate analgesia (defined as NRS <4); (2) independence from intravenous opioids for at least 12 h; (3) ability to independently stand and sit down (evaluated with the Timed Up and Go test)^27,28 ; and (4) unassisted ambulation of at least 30 m (evaluated with the 6-min walk test).²⁹  For both the Timed Up and Go test and general ambulation, a four-legged walker was used by all subjects. The primary endpoint of this study was the time from surgical stop until all four of these criteria were fulfilled without a reversion to unfulfilled status. These criteria were assessed at the end of each 8-h nursing shift: 08:00, 16:00, and midnight. Pain scores were recorded every 4 h and when subjects requested analgesics. Subjects participated in physical therapy sessions twice daily, beginning as early as the afternoon of surgery if they reached the orthopedic wards by 14:00 the day of surgery. Neither study participants nor investigators were masked to treatment group assignment.

Secondary endpoints included each of the four individual discharge criteria of the primary endpoint, supplemental oral opioid consumption, attaining a standing position without assistance, passive knee flexion and extension (measured with a goniometer), catheter site leakage, and the incidence of catheter dislodgement. The time for catheter placement began with the insertion of the Tuohy needle and ended upon final needle withdrawal. Infusion pump memory was interrogated daily and provided the basal infusion rate, self-administered bolus dose attempts and delivery, infused volume, and infusion duration. Subjects were discharged home after meeting all four of the composite primary endpoint criteria, but not before POD 3. Perineural catheters were removed by medical personnel before hospital discharge.

One week after surgery (±2 days), subjects were called and verbally completed the Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC) questionnaire. The WOMAC is an instrument specifically designed to evaluate clinically important, patient-relevant changes in health-related quality-of-life following treatment interventions in patients with osteoarthritis of the knee.⁷  The WOMAC evaluates three dimensions of health-related quality-of-life: pain, stiffness, and physical functional disability with 5, 2, and 17 questions, respectively. An ordinal Likert scale from 0 to 4 is used for each question, with lower scores indicating lower levels of symptoms or physical disability.³⁰  Each subscale is summated to a maximum score of 20, 8, and 68, respectively. The individual dimensions are always analyzed separately, and investigators have often added a “global” score, which is calculated by summating the scores for the three subscales.³¹  The questionnaire may be self-administered or administered via a telephone call and takes 5 to 10 min to complete.³²  Because it is a proprietary instrument, the questionnaire itself may not be published and is therefore not included in an appendix. Since its inception 2 decades ago, the WOMAC has been translated into 60 languages and used in several hundred published clinical trials.³³  It has been rigorously examined, demonstrating excellent construct validity, responsiveness, and test–retest reliability in patients after total knee replacement,^30,34–36  and it is therefore recommended in the Osteoarthritis International Research Society’s guidelines for clinical trials.^33,37 

All analyses were two-tailed tests comparing the randomized (independent) groups and conducted to assess superiority.

Balance between the adductor canal and femoral randomized groups on postrandomization catheter insertion and perioperative characteristics was assessed using standard summary statistics. The effect of perineural catheter location (adductor canal vs. femoral) on time to reach all four discharge criteria was assessed using the two-tailed Wilcoxon rank sum test, with difference in medians estimated using the method of Hodges–Lehmann. In addition, Kaplan–Meier analysis on these time-to-event outcomes was conducted, and groups compared with the log-rank test. Finally, hazard ratios were estimated using a Cox proportional hazards regression model, and the proportional hazards assumption was tested by assessing the group-by-time interaction.

Each of the four individual discharge criteria of the primary endpoint, including (1) adequate analgesia, (2) independence from intravenous opioids, (3) independent ambulation 30 m or greater, and (4) the ability to independently stand, walk 3 m, return, and sit down was compared between the two randomized groups using the same statistical methods as outlined in the previous paragraph for the primary endpoint. The treatment effects on (5) time to attaining a standing position without assistance and (6) time to physical therapy unlimited by quadriceps weakness were assessed by log-rank test as well. The treatment effects on (7) passive knee flexion and (8) extension (after logarithm transformation) and (9) average and (10) worst pain score over time were assessed by repeated-measures general linear model with an autoregressive covariance structure. The treatment-by-time interaction was assessed for each outcome.

A Student t test or Wilcoxon rank sum test was used to compare adductor canal and femoral groups on continuous or ordinal outcomes, including (11) total supplemental oral opioid consumption, (12) total local anesthetic administered, (13) basal rate, and (14) hours from surgical stop to discharge, as appropriate. Pearson chi-square test or the Fisher exact test was used for (15) catheter site leakage, (16) catheter dislodgement, (17) hospital discharge rate, and additional categorical data, as appropriate. For each secondary outcome, we used an overall significance criterion of 0.003 (i.e., 0.05/17, a total of 17 secondary outcomes, Bonferroni correction) to control the type I error at 0.05 for this set of outcomes. Throughout we refer to them as “95% CIs” to indicate that the significance level was controlled at 5% for each hypothesis. Last, the chi-square test was used for comparisons of categorical data. SAS software version 9.3 for Windows (SAS Institute, USA) was used for all statistical analyses.

This superiority study was powered for the composite primary endpoint. Based on previously published data for tricompartment knee arthroplasty, the estimated distribution of time-to-discharge readiness for the adductor canal (femoral) catheter groups was 30 h: 60% (25%); 45 h: 20% (25%); 54 h: 10% (25%); and 69+h: 10% (25%).²  The given distributions are consistent with mean time-to-discharge readiness of 39 h versus 49 h (assuming the last category equals 70 h). To ensure 90% power at the 0.05 significance level for the Wilcoxon rank sum test to detect differences in the distribution of time-to-discharge category at least as large as specified in the two preceding sentences, 38 subjects were required in each treatment group. Sample size calculations were made using the POWER procedure (twosamplewilcoxon statement) in SAS statistical software. Therefore, we planned to enroll a total of 80 subjects having tricompartment knee arthroplasty who reached all four discharge criteria and therefore had an evaluable primary endpoint.

From January 2013 to September 2014, 84 subjects signed an informed consent form. We discovered that one individual had an exclusion criteria (body mass index >40 kg/m²) before randomization and the subject was excluded from further study. The remaining 83 subjects were randomized to either an adductor canal (n = 40) or femoral (n = 43) catheter, and all perineural catheters were inserted per protocol. However, one subject with a femoral catheter did not develop a sensory block within 30 min as required per protocol and was therefore withdrawn from study before perineural infusion initiation and further data collection. Two subjects began their infusions but withdrew from the study on PODs 1 (femoral) and 2 (adductor canal) before meeting all four discharge criteria, leaving 80 subjects with an evaluable primary endpoint (table 1). Of postrandomization catheter insertion and perioperative characteristics, only catheter insertion time differed to a statistically significant degree between treatment groups, with adductor canal catheters requiring 50% more time than their femoral counterparts (table 2).

Subjects assigned to an adductor canal catheter (n = 39) met the four discharge readiness criteria in a median of 55 h (interquartile range, 42 to 63 h) compared with 61 h (49 to 69 h) for subjects who received a femoral catheter (n = 41; P = 0.12; fig. 1). The median was an estimated 6 h less (95% CI, 13 h less, 1 h more) for adductor canal than femoral. The estimated hazard ratio of meeting all four discharge criteria was 1.26 (95% CI, 0.80 to 1.96) for adductor canal versus femoral (proportional hazards assumption was not violated, P = 0.08). In other words, subjects with an adductor canal catheter were 26% more likely to meet all four discharge criteria at any one time than subjects with a femoral catheter. In terms of discrete days after surgery, on POD 1, seven (18%) of the subjects with an adductor canal catheter met the four discharge readiness criteria versus only two subjects (5%) with a femoral catheter (P = 0.004). However, by POD 2, there was no statistically significant difference between the treatment groups (51 vs. 44%, P = 0.157; fig. 1). Therefore, use of a continuous adductor canal catheter possibly hastened overall discharge readiness for 5 of 39 subjects (13%), who would have been discharged the following day if they had instead received a femoral perineural infusion.

On POD 1, 72% of subjects in the adductor canal group were able to fulfill both the Timed Up and Go test (fig. 2) and ambulation (fig. 3) criteria compared with 27% in the femoral catheter group (P < 0.001; table 3). In contrast, there were minimal differences between treatment groups in both pain scores at rest (fig. 4) and supplemental opioid requirements (fig. 5).

The femoral catheter group reported superior analgesia during physical therapy sessions compared with the adductor canal catheter group (table 4), and this group demonstrated a higher mean/median passive knee flexion with a femoral catheter although this difference did not reach statistical significance (7 to 15 degrees; P = 0.15; table 3). There were no statistically significant differences detected between the two groups regarding passive knee extension (table 3), catheter site leakage during the first 2 days of infusion, day of actual hospital discharge (table 5), or health-related quality-of-life 1 week after surgery (fig. 6). However, a larger percentage of subjects with a femoral catheter had their basal infusion rate decreased by POD 3 (P < 0.001), whereas a larger percentage of subjects with an adductor canal catheter had their basal infusion rate increased during the same period of time (P < 0.001; table 5). This resulted in subjects with an adductor canal catheter consuming more local anesthetic relative to those with a femoral catheter (P = 0.004; table 5).

One adductor canal catheter broke external to the subject in the late evening of POD 1 for an unknown reason. For purposes of analysis, this subject was retained in her treatment group per the intention-to-treat principle.³⁸  There were seven subjects erroneously discharged a day early on POD 2 after meeting all discharge criteria: four (10%) and three (7%) with adductor canal and femoral catheters, respectively. There were five (6%) falls total, two subjects (5%) with adductor canal catheters on PODs 1 and 3 and three subjects (7%) with femoral catheters on PODs 1, 2, and 3. There were no injuries or complications as a result of these falls. No catheter-related infections or nerve injuries were identified.

This dual-center, randomized, controlled, parallel-arm clinical trial provides strong evidence that a continuous adductor canal block does not appreciably decrease the time to overall discharge readiness compared with a continuous femoral nerve block after tricompartment knee arthroplasty. Although an adductor canal infusion did decrease the time to achieve adequate mobilization, this had minimal effects on overall discharge readiness because both groups experienced similar analgesia and intravenous opioid requirements that—in most cases—exceeded the time required for adequate mobilization. Because the adductor canal infusions provided similar analgesia at rest compared with their femoral counterparts, these findings suggest that continuous adductor canal blocks may be preferable due to their greatly decreased inhibition of mobilization and ambulation—both important components of recovery after knee arthroplasty.³⁹  Conversely, femoral catheters provided superior dynamic analgesia during physical therapy (P = 0.01 and 0.02), and this group had a higher mean/median passive knee flexion although this difference did not reach statistical significance (P = 0.15).

Although previously published randomized studies have reported similar findings for some of the outcome measurements of the current trial, this is the first study—to our knowledge—to determine the relative effects of using an adductor canal versus femoral perineural infusion of multiple days on discharge readiness after knee arthroplasty using objective, prospectively determined discharge criteria.

Previous studies examining the adductor canal block support its efficacy of perioperative analgesia for total knee arthroplasty relative to placebo.^19,21  A single-injection adductor canal block has also been shown to exhibit improved quadriceps strength and noninferior analgesia when compared with a single-injection femoral nerve block for total knee arthroplasty.¹⁵ 

Two previous randomized studies examined continuous adductor canal blockade versus continuous femoral nerve blockade for total knee arthroplasty. Jaeger et al. compared quadriceps muscle strength using a hand-held dynamometer the day after surgery, whereas Shah and Jain²⁰  compared mobilization ability using the Timed Up and Go, 10-m walk, and 30-s chair stand/sit tests.²²  Unlike the current study, both previous investigations included a large bolus of long-acting local anesthetic (30 ml of 0.5% or 0.75% ropivacaine) during postoperative catheter insertion. Furthermore, they provided perineural local anesthetic for 24 or fewer hours, leaving a relatively short period of time between the initial block resolution and perineural catheter removal. In addition, the primary endpoints for both investigations were measured at a maximum of 24 h postoperatively. Our study, in contrast, extended previous work by providing a 3-day perineural infusion and measuring the primary endpoint for 72 h or until discharge readiness was attained, whichever occurred later.

Among the four required discharge criteria, we did not detect a statistically significant difference in two: time until adequate analgesia and independence from intravenous opioids. However, there were remarkable differences for the remaining two criteria: time until able to independently ambulate 30 m and independently stand, walk 3 m, return, and then sit down. These findings are consistent with data from one of the two previously published studies comparing adductor canal and femoral continuous nerve blocks, by Shah and Jain,²⁰  and considerably extends the observation period. The second related study, by Jaeger et al., found improved quadriceps strength for subjects with adductor canal catheters (retaining 52 vs. 18% of baseline; P = 0.004), a difference that did not translate into improved ambulation or mobilization—in contrast with our current results.²²  Differences may be attributable to lower power in the previous study due to a smaller sample size (48 vs. 80 subjects), a large (30 ml) initial bolus of 0.5% ropivacaine combined with outcomes measured only 24 h later, or a fixed basal infusion (8 ml/h) without patient-controlled bolus doses.

One notable difference previously unreported in similar studies is that the subjects with adductor canal catheters reported greater degrees of pain during their physical therapy sessions compared with the group with femoral catheters (although a study involving single-injection blocks did detect a similar difference at 24 h).⁴⁰  This association may be due to subjects with adductor canal catheters ambulating further, inducing a greater degree of dynamic pain; or, it might be because femoral catheters provided superior analgesia, regardless of ambulation distance. The fact that subjects with a femoral catheter had a greater mean/median passive knee flexion (7 to 15 degrees; P = 0.15) suggests the latter. Although the causative relationship among these variables requires further study, the clinical association remains: the desire for dynamic analgesia (favoring femoral catheters) and improved mobility (favoring adductor canal catheters) appears to be at odds; and, practitioners will need to determine the relative importance of each before choosing a catheter insertion site.

We gave 30 ml lidocaine 2% for the initial bolus and ropivacaine 0.2% at 6 ml/h for the initial basal infusion. The basal infusion rate was adjusted in 2 ml/h increments up to twice per day according to pain needs and quadriceps strength. It is noteworthy that by the second POD, nearly a third of subjects with adductor canal catheters had been increased to 8 ml/h versus only 5% with femoral catheters. Conversely, nearly 50% of subjects with a femoral catheter had been decreased to 4 ml/h due to quadriceps weakness compared with not a single subject with an adductor canal catheter.

These results not only demonstrate the quadriceps sparing of adductor canal relative to femoral infusions but also suggest that adductor canal basal infusions of ropivacaine 0.2% should be initially set at a higher rate than their femoral counterparts. Consequently, the average infusion rate for the femoral catheters was less than the average infusion rate for the adductor canal block, 5 versus 8 ml/h. The optimal local anesthetic and concentration, basal infusion rate, bolus volume, and infusion regimen (basal-only, basal/bolus combination, and repeated bolus doses) remain unknown and require further study.

Although single-injection femoral nerve blocks may not be associated with an increased risk of falls,^9,41 continuous peripheral nerve blocks involving the femoral nerve have been associated with an increased risk of falling.^4–6  It remains unknown to what degree each induced deficit—motor, sensory, and proprioception—contributes to increased risk.⁴²  The greatly reduced motor block induced by both single-injection and continuous adductor canal blocks relative to their femoral counterparts has raised hopes of decreasing the risk of falls.^14,21  Our current study, with only two (5%) adductor canal and three (7%) femoral block subjects falling is underpowered for this important outcome and should not be taken as supporting or refuting the potential benefits of adductor canal blocks on falls.

The anatomic location of catheter insertion used in this study mirrors that used by the majority of recent investigations involving the adductor canal block. The technique used was selected for its benefit shown by Lund et al.²⁶  and Jaeger et al.²¹  as well as the authors’ examination of the underlying neuroanatomy and clinical observations. This technique has demonstrated analgesic benefit and preserved quadriceps strength for patients after knee arthroplasty. It also minimizes interference with the surgical field, thus decreasing the potential for catheter dislodgement and possibly reducing the theoretical impact of bacterial colonization or potential localized infection affecting the new implant.⁴³  However, the optimal catheter insertion site has not yet been determined.^44–46  Therefore, our results may be applicable exclusively to catheters inserted midway on the line between the anterior superior iliac spine and the cephalad margin of the patella.

Similarly, the optimal perineural local anesthetic infusion regimen has yet to be elucidated. For continuous blocks involving the femoral nerve, dose appears to be the predominant determinant of infusion effects relative to local anesthetic concentration or basal rate; therefore, we do not believe that a change in ropivacaine concentration would produce differing results.^47–49  However, it remains unknown whether local anesthetic introduced into the adductor canal as regularly scheduled bolus doses—as opposed to a basal infusion and added patient-controlled bolus doses—would change the pharmacodynamics of the infusion.⁵⁰  In addition, both treatment groups received intraoperative joint infiltration of ropivacaine, ketorolac, epinephrine, and tranexamic acid. It thus remains unknown whether the results of our study would be different if this infiltration was not included or modified. Furthermore, practitioners were more experienced placing femoral compared with adductor canal catheters, at least during the first half of enrollment; thus, our finding that adductor canal catheters required 50% more time for insertion than their femoral counterparts may have been strongly influenced by this difference in familiarity.

Last, subjects and investigators were not masked to treatment group. Although it is unlikely that subjects had a predisposition toward one insertion site versus another, outcome assessors (nursing staff, physical therapists, and investigators) may have had preconceived bias toward one of the two treatments. In addition, caretaker bias may have been subconsciously transferred to patients, and therefore biased the results.

Compared with a continuous femoral nerve block, a continuous adductor canal block decreased the time to achieve adequate mobilization for discharge after tricompartment knee arthroplasty. This resulted in minimal overall earlier discharge readiness because both groups experienced similar analgesia and intravenous opioid requirements that—in most cases—exceeded the time to attain adequate mobilization. Adductor canal blocks may nonetheless be preferable to femoral nerve blocks because they better preserve quadriceps strength and shorten time until adequate mobilization is achieved while providing comparable analgesia at rest. Conversely, femoral catheters provide superior dynamic analgesia, and practitioners thus must decide the relative importance of this factor versus the relative benefits of adductor canal catheters on quadriceps strength and mobilization. These results may have been influenced by the unmasked design of this study.

The authors appreciate the invaluable assistance of the University California San Diego Department of Physical Therapy (San Diego, California), without which this study would not have been possible.

Funding for this project was provided by the University California Academic Senate (San Diego, California); Summit Medical (Sandy, Utah); and Teleflex Medical (Research Triangle Park, North Carolina). Teleflex Medical also provided the perineural catheters used in this investigation. Neither company had input into any aspect of study conceptualization, design, and implementation; data collection, analysis, and interpretation; or article preparation. The contents of this article are solely the responsibility of the authors and do not necessarily represent the official views of the funding entities.

The authors declare no competing interests.