Total knee arthroplasty (TKA) is generally associated with moderate to severe postoperative pain that often lasts for up to 48 hours and is particularly severe during mobilization.1,2 Several methods are available for pain management including opioids, administered IV or intrathecally, peripheral nerve block, and the recently introduced method of local infiltration analgesia (LIA). Although IV opioids are effective, they have the major disadvantage of having disabling side effects, and analgesia is sometimes inadequate, specifically during mobilization. Therefore, intrathecal morphine has been used as an alternative method for pain relief and provides satisfactory analgesia for at least 24 hours, a period of time when pain is often most severe.3–6 However, side effects of intrathecal morphine—including pruritus, urinary retention, and respiratory depression7—can be troublesome for the patient and require increased supervision and monitoring.
The LIA technique has increasingly become popular over the last 5 to 10 years, especially in the Scandinavian countries. A long-acting local anesthetic (ropivacaine), a nonsteroidal anti-inflammatory drug (NSAID) (ketorolac), and epinephrine are infiltrated intraoperatively in the LIA technique, and this solution is also injected via a catheter placed in the knee joint, postoperatively. A number of studies supporting the efficacy of LIA in TKA have been published recently,8–12 but only a few have compared this technique with other standard techniques for postoperative pain management after TKA.13–16 Lower pain scores were found in all these studies, and mobilization could be achieved earlier after LIA. However, no decrease in hospital length of stay (LOS) was shown.
In this prospective, randomized, double-blind study, we compared the LIA technique with intrathecal morphine during TKA. In an attempt to prolong the postoperative analgesia and improve mobilization, the intra-articular catheter was left in situ for 48 hours to allow an additional bolus injection on the first and second postoperative days. Our hypothesis was that LIA would provide better postoperative analgesia than intrathecal morphine and thereby reduce IV morphine consumption during the first 48 postoperative hours.
The regional ethics committee in Uppsala, Sweden (March 25, 2009, Dnr 2009/069), and the Swedish Medical Products Agency approved this study. It was also registered at the ClinicalTrials.gov (Code NCT992082) on October 7, 2009, and conducted in accordance with the Helsinki declaration and monitored by the Clinical Research Support Unit at Örebro University Hospital.
Fifty patients scheduled to undergo TKA because of osteoarthritis gave written, informed consent and were enrolled in this trial (Fig. 1). The inclusion criteria were age 40 to 85 years and ASA I–III. Exclusion criteria were allergy or intolerance to any of the study drugs, severe liver, heart or renal disease, inflammatory joint disease, chronic pain requiring opioid medication, bleeding disorder, and any other contraindication for spinal anesthesia. Patients fulfilling the above criteria had surgery between August 2009 and June 2010 at the Department of Orthopedic Surgery, Örebro University Hospital.
Randomization and Blinding
The hospital pharmacy randomized the patients into 2 groups: group M (morphine) and group L (local infiltration), 25 patients in each group, using computer-generated randomized numbers (Fig. 1). The surgeon was not blinded to group randomization and was not allowed to participate in postoperative patient care. All other persons involved directly or indirectly in the study—including the patients, the other investigators, the physiotherapist, and staff involved in postoperative patient care—were blinded to the study arm.
All patients received diazepam 10 mg orally 1 hour before planned surgery, and all operations were performed under spinal anesthesia using a 27-G pencil-point spinal needle at the L3/L4 or L2/L3 intervertebral space with the patient in the sitting position. In group M, morphine 0.1 mg (0.25 mL) was injected intrathecally, and in group L an equal volume of 0.9% saline, together with glucose-free bupivacaine 17.5 mg (3.5 mL) (Fig. 1). The study drug mixtures were prepared by the hospital pharmacy. If the spread of the sensory block (pinprick) was insufficient, the patient was excluded from the study, and general anesthesia was administered. All patients could receive propofol IV on demand or as continuous infusion during the operation. If the patient had pain during the operation, fentanyl IV was administered in bolus doses of 25 to 50 μg up to a maximum of 300 μg. If insufficient analgesia was achieved, the patient was excluded from the study, and general anesthesia was administered.
All patients received an AGC prosthesis (Biomet, Warsaw, IN) using a standard medial parapatellar approach. Surgery was performed using a femoral tourniquet to minimize blood loss and improve operative conditions. No drains were left in the knee joint after the operation. Cloxacillin 2 g was given IV preoperatively and continued until the intra-articular catheter was withdrawn after 48 hours. Dalteparin (5.000 IU) was administered subcutaneously for thromboprophylaxis once each evening for 10 days, starting on the evening before surgery. Ice packs were applied around the knee joint during the first 6 hours, which is a routine in our hospital. The whole lower limb had a compression bandage for the first 21 postoperative hours.
Local Infiltration Technique
In group L, 300 mg ropivacaine, 30 mg ketorolac, and 0.5 mg epinephrine (total volume 116 mL) were infiltrated by the surgeon into the soft tissues periarticularly during the operation. The injections were done systematically into all tissue that had been traumatized during surgery by injecting 40 to 50 mL in the posterior capsule and the collateral ligaments after the bone cuts had been made and before insertion of the prosthesis. After the prosthesis had been inserted, another 50 to 70 mL was injected in the capsule incision, in the quadriceps tendon, in the infrapatellar ligament, and around the posterior cruciate ligament. Finally, 50 mL ropivacaine (100 mg) without epinephrine or ketorolac was infiltrated into the subcutaneous tissue before skin closure. Thus, a total of 400 mg ropivacaine was administered in a volume of 166 mL. Patients in group M received no injection (Fig. 1).
Before wound closure, the surgeon placed an intra-articular catheter in all patients in both groups using a Tuohy 18-G needle, a multihole 20-G epidural catheter ,and a bacterial filter (B. Braun Medical, Melsungen, Germany). The needle was introduced percutaneously from the lateral side about 10 cm proximal to the skin incision through the vastus lateralis and into the knee joint. The catheter was inserted via the needle and passed along the medial femoral condyle, leaving the tip of the catheter in the posterior part of the knee joint. The needle was removed, the bacterial filter was connected, and the filter and the catheter were filled with 1 to 2 mL of ropivacaine for bacteriostasis as well as to ensure functional patency of the system.
On the first and second postoperative morning, after 21 and 45 hours, 200 mg ropivacaine, 30 mg ketorolac, and 0.1 mg epinephrine, total volume 22 mL, were injected intra-articularly via the catheter in group L, and a similar volume of saline was injected in group M (Fig. 1). These drugs were prepared by the hospital pharmacy to ensure blinding. The intra-articular catheter was removed after 45 hours, and the tip of the catheter was sent for culture.
The first attempt to mobilization was made on the first postoperative morning 1 hour after the intra-articular injection. The patients were encouraged to stand and to walk 6 to 8 steps. If the patients could not be mobilized, another attempt was made the following day, after the second intra-articular injection.
A visual analog scale (VAS; 0 mm = no pain, 100 mm = worst imaginable pain) was used for assessment of pain. At 48 hours, if pain at rest was VAS <40 mm during a 2-hour period, the patient-controlled analgesia (PCA) pump was discontinued, and tramadol was administered 100 mg orally up to 4 times daily as required to achieve VAS <40 mm.
Recordings and Measurements
PCA-morphine consumption was recorded during 0 to 24 hours and 24 to 48 hours postoperatively, and 0 to 48 hours was calculated.
Pain assessment (VAS) was made preoperatively and at 6, 12, 21, 22, 24, 45, and 46 hours postoperatively. Pain was assessed both at rest and on flexion of the knee by 60 degrees. Pain when walking was also recorded at 24 and 48 hours. After discharge home, all patients were asked to complete a questionnaire regarding postoperative pain on days 1, 3, and 14 and after 3 months.
The patients were also asked to give a verbal rating scale for satisfaction with the quality of analgesia (excellent = 4, good = 3, inadequate = 2, poor = 1) during the first and second postoperative days and after 7 days.
Maximum knee extension and flexion were assessed preoperatively, on day 3, at discharge, and after 2 weeks and 3 months postoperatively. Ability to climb 8 stairs was recorded at 24 and 48 hours. Time to Up and Go (TUG) test17 was assessed preoperatively and postoperatively on days 3, 7, and 14, and after 3 months. The TUG test involves timing the patient when he or she rises from an armchair, walks 3 meters, turns, walks back, and sits down again. Values <20 seconds indicate that the patient is independently mobile. Oxford Knee Score was determined preoperatively and at 2 weeks and 3 months postoperatively. Oxford Knee Score is a validated 12-item knee questionnaire that scores patients from 12 (best possible) to 60 (worst possible).18 EuroQol (EQ-5D) questionnaire was collected preoperatively and postoperatively at 3 months. EQ-5D is a standardized instrument for use as a measure of health outcome.19 It provides a single index value from 0 to 1 for which 0 represents poor health and 1 represents perfect health.
Home readiness and hospital stay.
After the second injection via the catheter at 45 hours, the time to fulfillment of discharge criteria (home readiness) was recorded by a physician and the study physiotherapist, who were unaware of group randomization. The discharge criteria were mild pain (VAS <30 at rest) sufficiently controlled by oral analgesics, able to walk with elbow crutches, ability to climb 8 stairs, eat and drink normally, and no evidence of any surgical complication. Time to fulfillment of discharge criteria was defined as the time from the end of the operation until the patient fulfilled the discharge criteria, which was assessed 3 times a day. Hospital LOS was recorded (day 0 = the day of operation) as actual time to home discharge.
The incidence of nausea, vomiting, pruritus, and sedation were recorded on the first and second postoperative days. Sedation was recorded at 24 hours and 48 hours using a 4-grade scale (1 = fully awake, 2 = drowsy [light sedation], 3 = asleep, 4 = deeply asleep [heavily sedated]). As part of our routine in our hospital, respiratory rate and arterial oxygen saturation were recorded during the first 24 hours, and respiratory depression is defined as respiratory rate <10/min combined with Sao2 <90%. All complications and adverse events were registered intra- and postoperatively, as well as after discharge. Any hospital readmission during the 3-month follow-up period postoperatively was also recorded.
Sample-size calculations were done using morphine consumption for 48 hours postoperatively as the primary endpoint. In an earlier study on patients undergoing TKA during general anesthesia,10 the mean ± SD morphine consumption for 48 hours postoperatively was 91 ± 36 mg in the placebo group versus 24 ± 23 mg in the LIA group. In a pilot study of 5 patients receiving spinal anesthesia with morphine added to the local anesthetic, the PCA-morphine consumption was 45 ± 11 mg. Therefore, assuming a mean of 45 mg in group M and 24 mg in group LIA, with SD of 23 in both groups, we calculated that 23 patients would be required in each group to detect this difference with an α of 0.05 and β of 0.2. Considering the risk of dropouts and the uncertainty in SD, 25 patients were included in each of the 2 groups. Repeated-measurements analysis of variance (ANOVA) with Huynh–Feldt corrected P values were used for the analysis of the primary endpoint (morphine consumption during the first 48 postoperative hours), and post hoc test from the ANOVA for the first 24 postoperative hours was also performed. Mean difference between groups and time points and their interaction were tested. To summarize each patient's VAS pain scores for the first 48 postoperative hours, the median value was calculated for each patient. The difference between groups was then analyzed using Mann–Whitney U test. Hospital stay, time to fulfillment of discharge criteria, knee function scores, and patient satisfaction scores were also analyzed using the Mann–Whitney U test. The Bonferroni–Holm method was used to correct for multiple measures when P < 0.05 in the secondary endpoints.20 Dichotomous data were analyzed using the χ2 test or Fisher's exact test, as appropriate. P < 0.05 was considered to be statistically significant. Confidence interval around median was calculated with Hodges–Lehmann method using Confidence Interval Analysis (CIA) Software (Statistics with Confidence, 2nd ed., BMJ Books 2000). All other analyses were made using computer software SPSS version 15.0 for windows (SPSS Inc., Chicago, IL).
Two patients in group M were excluded after randomization, the first because of failure to induce spinal anesthesia and the second because of an intraoperative conversion to unicompartmental knee arthroplasty (Fig. 1). The patient characteristics of the study groups were similar (Table 1).
Primary Endpoint: Morphine Consumption
Mean morphine consumption was less in group L than in group M for the first 48 postoperative hours: 26 ± 15 vs 54 ± 29, i.e., a mean difference for each 24-hour period of 14.2 (95% confidence interval [CI] 7.6 to 20.9) mg from the ANOVA (Table 2) with between- subjects effects (group) P < 0.001, within-subject effects (time) P = 0.001, and interaction effects (time × group) P = 0.335. Because of a slight significance in Shapiro–Wilk test for skewness, a sensitivity analysis with square-root transformation (resulting in nonsignificant Shapiro–Wilk test) was performed, resulting in the same overall conclusions. Post hoc test from the same ANOVA during the first 24 hours showed a mean difference of 15.7 (95% CI 7.9 to 23.6) mg. There was a protocol violation in 6 patients, 3 in each group, who received 1 dose of oral tramadol postoperatively in the ward.
The median VAS pain score at rest and on flexion for the first 48 postoperative hours was determined for each subject. The medians at rest were lower in group L than in group M: 5 (0–33) versus 20 (3–48) mm (P < 0.001) (Fig. 2). Even on flexion, the median (range) VAS pain score was lower in group L than in group M: 30 (0–60) versus 59 (22–93) mm (P < 0.001) for the first 48 postoperative hours (Fig. 3). When walking, the median (range) VAS pain score was lower in group L at 24 hours, 19 (0–49) versus 58 (40–92) (P < 0.001), and at 48 hours, 10 (10–50) versus 39 (4–74) (P = 0.001).
Patients' satisfaction was greater in group L than in group M on postoperative day 1 (P = 0.001). No difference was found in patient satisfaction on days 2 and 7 (Table 3).
Knee extension, knee flexion, and the TUG test did not show any differences between groups postoperatively (Table 3). However, a significantly larger proportion of patients in group L were able to climb stairs at 24 hours: 50% (11 of 22) versus 4% (1 of 23), i.e., a difference of 46% (95% CI 23.5 to 68.5); and at 48 hours: 70% (16 of 23) versus 22% (5 of 23), i.e., a difference of 48% (95% CI 23 to 73). Oxford Knee Score and EQ-5D did not reveal any differences between the groups at any time postoperatively (Table 3).
Home Readiness and Hospital Stay
Median (range) time to fulfillment of discharge criteria was shorter in group L than in group M, 51 (24–166) hours versus 72 (51–170) hours. The difference was 23 (95% CI 18 to 42) hours (P = 0.001). The hospital LOS was shorter in group L than in group M, median (range) 3 (2–17) versus 4 (2–14) days (P = 0.029) (Fig. 4). One patient in group M was admitted for 14 days because of a fall in the ward and sustained a hip contusion, which delayed discharge. One patient in group L remained in the hospital for 17 days because of persistent urinary retention requiring repeated catheterization.
We found no differences in the incidence of nausea, vomiting, pruritus, or sedation between groups (Table 4). We found only sedation grade 1 (fully awake) or 2 (drowsy), and no patient had sedation grade 3 or 4. Therefore we analyzed the incidence of sedation, as shown in Table 4. We found 13 episodes of respiratory rate <10/min in the LIA group in comparison with 8 in the morphine group, but none of those with Sao2 <90%. There were 3 episodes of Sao2 <90% in the LIA group in comparison with 2 in the morphine group, but none of those with a respiratory rate <10/min. None of the patients was treated with naloxone. There were 7 positive cultures from the catheter tips, all with solitary coagulase-negative Staphylococcus, 3 in group L and 4 in group M, but no antibiotics were given, and no clinical signs of infection were found during the follow-up period. One patient in group M was readmitted after discharge because of a swollen knee and mild fever, 37.8°C. The maximum C reactive protein was 38 mg/L, and wound cultures were negative. Oral antibiotics were administered, and wound healing was complete and satisfactory. No evidence of deep infection was found at admission or during the 3-month follow-up. No other complications were reported.
Good pain relief after TKA is important as it aids physiotherapy and promotes mobilization, which is central for a satisfactory outcome. Several recent studies have confirmed the efficacy of the LIA technique during TKA, and today the method is commonly used in Scandinavia as an alternative to regional blocks for postoperative pain management. Therefore, the main aim of our present study was to elucidate whether this method is equally effective as intrathecal morphine after TKA. To the best of our knowledge, this is the first study comparing the LIA technique with intrathecal morphine during TKA. We used intrathecal morphine as a comparator for several reasons. It is a simple method used routinely for pain management after major orthopedic surgery of the lower extremity. Many studies have been published using intrathecal morphine, and the method is well established in clinical practice. In a review on postoperative analgesia after TKA, Fischer et al. recommended the use of spinal injection of local anesthetic plus spinal morphine as an alternative to general anesthesia combined with femoral nerve block (FNB).21 Although intrathecal morphine may be associated with some side effects, the incidence of serious side effects such as respiratory depression is low. Respiratory depression is a cause for major concern and therefore the need for close monitoring may demand increased hospital resources. One drawback with intrathecal morphine is the relatively short duration of action, which limits analgesia to 24 hours and at best up to 48 hours.22,23 Results from our present study show several advantages of LIA over intrathecal morphine. First, we found significantly lower total rescue morphine consumption for 0 to 48 hours in the LIA group, which was our primary endpoint. In contrast, we found that the effect of intrathecal morphine was shorter than 24 hours, which resulted in significantly more morphine consumption during the first 24 postoperative hours in group M. Although morphine consumption is a surrogate endpoint, nevertheless our study demonstrates that the LIA technique is effective because we used PCA in both groups and patients were clearly instructed to self-administer morphine as required to achieve VAS <30 mm. In addition to the local infiltrations during the operation, we injected the drugs intra-articularly after 21 and 45 hours in the LIA group, which prolonged the analgesic duration and reduced morphine consumption, which is an advantage when using this method. In contrast, the inability of prolonging analgesia is a major disadvantage when using intrathecal morphine and severely limits its potential during TKA. Furthermore, the injection of drugs intra-articularly after 45 hours prolonged analgesia during mobilization, and this beneficial effect may also have resulted in earlier home readiness and discharge in comparison with our previous study in which the catheter was removed after 21 hours.10 Second, we found significantly lower pain scores in the LIA group on movement as well as when walking for 0 to 48 hours, which is important because better pain relief allows patients to be mobilized easily and aids physiotherapy. The latter is particularly important after knee surgery because it may promote quicker discharge home and earlier rehabilitation. Indeed, we did find a shorter time to home readiness as well as to home discharge in the LIA group. Because personnel evaluating home readiness were blinded to the study arm and we used previously described objective criteria for its assessment,10 we believe that our results are valid on this point. Therefore, less pain resulted in earlier mobilization, which in turn led to earlier home readiness in the LIA group. Finally, a subjective measurement of patient satisfaction with postoperative analgesia was also greater in the LIA group. When combined, our findings would confirm that LIA is effective and results in good pain relief, earlier mobilization, and quicker home readiness and discharge.
It could be argued that intrathecal morphine is not the “gold standard” for pain relief after TKA, and therefore an alternative comparator group could have been epidural analgesia (EDA) or FNB. In view of the highlighted risks of EDA in older patients,24 we were reluctant to use this method for pain management after orthopedic surgery. Furthermore, previous studies using EDA for TKA did not demonstrate better analgesia than did LIA.13–15 Could FNB then have been a better alternative to spinal morphine as a comparator? One study found the LIA technique to be better than FNB in terms of earlier mobilization, reduced pain intensity, and lower morphine requirements.16 Although the study by Carli et al. showed greater morphine consumption in the LIA group than in the FNB group, this study could be questioned in view of the fact that the FNB group also received a modified LIA technique, which may be a confounding factor.25 Therefore, these results are difficult to interpret.
In our study, the incidence of side effects was similar between the groups. Although urinary retention, pruritus, and nausea and vomiting are frequent side effects of intrathecal opioids,4,26,27 we were unable to find any significant differences between the groups. This could be because of the small number of patients we studied and because the difference between the 2 groups in rescue morphine consumption was small. Thus, it is possible that side effects of morphine are dose dependent and that when larger doses are administered, the incidence of side effects increases and then becomes more clinically important and significant.
The patients in this study received a high dose of ropivacaine (400 mg) initially, followed by 2 bolus injections of 200 mg each over a 48-hour period. No patient had any clinical symptoms of systemic toxicity. In an earlier study using 400 mg ropivacaine injected periarticularly, followed by 200 mg intra-articularly after 21 hours, we could show that the individual maximum unbound plasma concentrations were far below toxic levels.10,28 Our data are similar to those of other authors9,12 and confirm that the risk of local anesthetic toxicity is small or absent when injected intra-articularly in these doses.
The safety of leaving in an intra-articular catheter for 1 or 2 days can be questioned. A number of studies using LIA and intra-articular catheters have not reported any infections related to the use of the wound catheter.10,12,16,29,30 However, 2 studies did report deep infections. DeWeese et al. reported 1 deep infection in 91 patients, and Rasmussen et al. found 1 in 136 TKA when a catheter was left in situ for 72 hours.31,32 This low incidence of deep infection after TKA can be expected even without intra-articular catheters.33 It is important to stress that all of our patients were given antibiotics until the catheter was removed; the catheters were inserted during the operation by an orthopedic surgeon under sterile conditions; and a bacterial filter was used during all intra-articular injections postoperatively.
No differences were found at 3 months when comparing the general health outcome EQ-5D or the disease-specific Oxford Knee Score, because earlier mobilization and shorter hospital stay do not seem to affect the long-term outcome in any significant way.
One possible limitation of this study could be that group L received NSAID in the mixture injected in the knee perioperatively, whereas group M did not. Therefore, a systemic effect of NSAID administered intra-articularly cannot be excluded.34 It may have been an advantage to inject a similar dose of ketorolac IV postoperatively in group M to confirm a beneficial local effect. However, some studies in the literature have reported significantly better pain relief when ketorolac was administered intra-articularly in comparison with IV injection.13,14,35
Could a higher dose of morphine injected intrathecally have a better analgesic effect? In one study, the authors found that 0.1 mg or 0.2 mg resulted in similar postoperative pain relief after hip arthroplasty.4 In addition, 0.1 mg morphine was also found to provide the best balance between efficacy and side effects in elderly patients. Therefore, we chose to use this dose and also found few side effects in the present study.
In conclusion, the LIA technique was found to be superior to intrathecal morphine in providing good pain relief and resulted in early mobilization and greater patient satisfaction after TKA. These advantages translated into earlier home readiness and quicker home discharge without increasing any adverse effects. However, there was no improvement in patient-assessed long-term outcomes when using the LIA technique. Further clinical trials are warranted to define the best composition of drugs involved in the LIA mixture and the role of the intra-articular catheter in prolonging postoperative analgesia.
Name: Per Essving, MD.
Contribution: Study design, enrollment of patients, surgery, data analysis, manuscript preparation.
Attestation: This author reviewed the original study data and data analysis, approved the final manuscript, and is the archival author.
Name: Kjell Axelsson, MD, PhD.
Contribution: Study design, data collection, data analysis, manuscript preparation.
Attestation: This author reviewed the original study data and data analysis, and approved the final manuscript.
Name: Elisabeth Åberg, BSc.
Contribution: Data collection.
Attestation: This author approved the final manuscript.
Name: Henrik Spännar, BSc.
Contribution: Data collection.
Attestation: This author approved the final manuscript.
Name: Anil Gupta, MD, PhD.
Contribution: Study design, manuscript preparation.
Attestation: This author approved the final manuscript.
Name: Anders Lundin, MD, PhD.
Contribution: Data analysis.
Attestation: This author approved the final manuscript.
This manuscript was handled by: Spencer S. Liu, MD.
We thank Anders Magnuson of the Department of Epidemiology and Statistics for his expert help on statistic issues.
1. Strassels SA, Chen C, Carr DB. Postoperative analgesia: economics, resource use, and patient satisfaction in an urban teaching hospital. Anesth Analg 2002;94:130–7
2. Wang H, Boctor B, Verner J. The effect of single-injection femoral nerve block on rehabilitation and length of hospital stay after total knee replacement. Reg Anesth Pain Med 2002;27:139–44
3. Cole PJ, Craske DA, Wheatley RG. Efficacy and respiratory effects of low-dose spinal morphine for postoperative analgesia following knee arthroplasty. Br J Anaesth 2000;85:233–7
4. Murphy PM, Stack D, Kinirons B, Laffey JG. Optimizing the dose of intrathecal morphine in older patients undergoing hip arthroplasty. Anesth Analg 2003;97:1709–15
5. Sites BD, Beach M, Biggs R, Rohan C, Wiley C, Rassias A, Gregory J, Fanciullo G. Intrathecal clonidine added to a bupivacaine–morphine spinal anesthetic improves postoperative analgesia for total knee arthroplasty. Anesth Analg 2003;96:1083–8
6. Tan PH, Chia YY, Lo Y, Liu K, Yang LC, Lee TH. Intrathecal bupivacaine with morphine or neostigmine for postoperative analgesia after total knee replacement surgery. Can J Anaesth 2001;48:551–6
7. Cousins MJ, Mather LE. Intrathecal and epidural administration of opioids. Anesthesiology 1984;61:276–310
8. Andersen LO, Husted H, Otte KS, Kristensen BB, Kehlet H. High-volume infiltration analgesia in total knee arthroplasty: a randomized, double-blind, placebo-controlled trial. Acta Anaesthesiol Scand 2008;52:1331–5
9. Busch CA, Shore BJ, Bhandari R, Ganapathy S, MacDonald SJ, Bourne RB, Rorabeck CH, McCalden RW. Efficacy of periarticular multimodal drug injection in total knee arthroplasty. A randomized trial. J Bone Joint Surg Am 2006;88:959–63
10. Essving P, Axelsson K, Kjellberg J, Wallgren O, Gupta A, Lundin A. Reduced morphine consumption and pain intensity with local infiltration analgesia (LIA) following total knee arthroplasty. Acta Orthop 2010;81:354–60
11. Kerr DR, Kohan L. Local infiltration analgesia: a technique for the control of acute postoperative pain following knee and hip surgery: a case study of 325 patients. Acta Orthop 2008;79:174–83
12. Vendittoli PA, Makinen P, Drolet P, Lavigne M, Fallaha M, Guertin MC, Varin F. A multimodal analgesia protocol for total knee arthroplasty. A randomized, controlled study. J Bone Joint Surg Am 2006;88:282–9
13. Andersen KV, Bak M, Christensen BV, Harazuk J, Pedersen NA, Soballe K. A randomized, controlled trial comparing local infiltration analgesia with epidural infusion for total knee arthroplasty. Acta Orthop 2010;81:606–10
14. Spreng UJ, Dahl V, Hjall A, Fagerland MW, Raeder J. High-volume local infiltration analgesia combined with intravenous or local ketorolac+morphine compared with epidural analgesia after total knee arthroplasty. Br J Anaesth 2010;105:675–82
15. Thorsell M, Holst P, Hyldahl HC, Weidenhielm L. Pain control after total knee arthroplasty: a prospective study comparing local infiltration anesthesia and epidural anesthesia. Orthopedics 2010;33:75–80
16. Toftdahl K, Nikolajsen L, Haraldsted V, Madsen F, Tonnesen EK, Soballe K. Comparison of peri- and intraarticular analgesia with femoral nerve block after total knee arthroplasty: a randomized clinical trial. Acta Orthop 2007;78:172–9
17. Podsiadlo D, Richardson S. The timed “Up & Go”: a test of basic functional mobility for frail elderly persons. J Am Geriatr Soc 1991;39:142–8
18. Jahromi I, Walton NP, Dobson PJ, Lewis PL, Campbell DG. Patient-perceived outcome measures following unicompartmental knee arthroplasty with mini-incision. Int Orthop 2004;28:286–9
19. Fransen M, Edmonds J. Gait variables: appropriate objective outcome measures in rheumatoid arthritis. Rheumatology (Oxford) 1999;38:663–7
20. Holm S. Simple sequentially rejective multiple test procedure. Scand J Stat 1979;6:65–70
21. Fischer HB, Simanski CJ, Sharp C, Bonnet F, Camu F, Neugebauer EA, Rawal N, Joshi GP, Schug SA, Kehlet H. A procedure-specific systematic review and consensus recommendations for postoperative analgesia following total knee arthroplasty. Anaesthesia 2008;63:1105–23
22. Beaussier M, Weickmans H, Parc Y, Delpierre E, Camus Y, Funck-Brentano C, Schiffer E, Delva E, Lienhart A. Postoperative analgesia and recovery course after major colorectal surgery in elderly patients: a randomized comparison between intrathecal morphine and intravenous PCA morphine. Reg Anesth Pain Med 2006;31:531–8
23. Gehling MH, Luesebrink T, Kulka PJ, Tryba M. The effective duration of analgesia after intrathecal morphine in patients without additional opioid analgesia: a randomized double-blind multicentre study on orthopaedic patients. Eur J Anaesthesiol 2009;26:683–8
24. Moen V, Dahlgren N, Irestedt L. Severe neurological complications after central neuraxial blockades in Sweden 1990–1999. Anesthesiology 2004;101:950–9
25. Carli F, Clemente A, Asenjo JF, Kim DJ, Mistraletti G, Gomarasca M, Morabito A, Tanzer M. Analgesia and functional outcome after total knee arthroplasty: periarticular infiltration vs continuous femoral nerve block. Br J Anaesth 2010;105:185–95
26. Jacobson L, Chabal C, Brody MC. A dose–response study of intrathecal morphine: efficacy, duration, optimal dose, and side effects. Anesth Analg 1988;67:1082–8
27. McDonnell NJ, Paech MJ, Browning RM, Nathan EA. A randomised comparison of regular oral oxycodone and intrathecal morphine for post-caesarean analgesia. Int J Obstet Anesth 2010;19:16–23
28. Knudsen K, Beckman Suurkula M, Blomberg S, Sjovall J, Edvardsson N. Central nervous and cardiovascular effects of i.v. infusions of ropivacaine, bupivacaine and placebo in volunteers. Br J Anaesth 1997;78:507–14
29. Bianconi M, Ferraro L, Traina GC, Zanoli G, Antonelli T, Guberti A, Ricci R, Massari L. Pharmacokinetics and efficacy of ropivacaine continuous wound instillation after joint replacement surgery. Br J Anaesth 2003;91:830–5
30. Essving P, Axelsson K, Kjellberg J, Wallgren O, Gupta A, Lundin A. Reduced hospital stay, morphine consumption, and pain intensity with local infiltration analgesia after unicompartmental knee arthroplasty. Acta Orthop 2009;80:213–9
31. DeWeese FT, Akbari Z, Carline E. Pain control after knee arthroplasty: intraarticular versus epidural anesthesia. Clin Orthop Relat Res 2001:226–31
32. Rasmussen S, Kramhoft MU, Sperling KP, Pedersen JH. Increased flexion and reduced hospital stay with continuous intraarticular morphine and ropivacaine after primary total knee replacement: open intervention study of efficacy and safety in 154 patients. Acta Orthop Scand 2004;75:606–9
33. Jamsen E, Varonen M, Huhtala H, Lehto MU, Lumio J, Konttinen YT, Moilanen T. Incidence of prosthetic joint infections after primary knee arthroplasty. J Arthroplasty 2010;25:87–92
34. Stalman A, Tsai JA, Segerdahl M, Dungner E, Arner P, Fellander-Tsai L. Ketorolac but not morphine exerts inflammatory and metabolic effects in synovial membrane after knee arthroscopy: a double-blind randomized prospective study using the microdialysis technique. Reg Anesth Pain Med 2009;34:557–64
© 2011 International Anesthesia Research Society
35. Convery PN, Milligan KR, Quinn P, Scott K, Clarke RC. Low-dose intra-articular ketorolac for pain relief following arthroscopy of the knee joint. Anaesthesia 1998;53:1125–9