This prospective, single-center, randomized, observer-blinded, parallel group comparison study was conducted in accordance with the ethical principles of the Helsinki Declaration, and was approved (authorization number A2405) by the Regional Ethical Committee of Saitama City Hospital (Research Ethics Committee No. 11000176) in September 2012. The registry number of this RCT is UMIN-CTR ID: 000013364 R: 000015591. The patients were recruited from October 2012 to July 2013. Thirty-four subjects were randomly assigned to two groups: an SNB group (group S) or a local anesthesia group (group L). All subjects provided written informed consent to participate in this study. Anesthesiologists who did not perform the anesthesia enrolled the participants and assigned them to the groups. The allocation ratio was 1:1. In total, 34 numbered cards assigned as L1-L17 and S1-S17 in an opaque envelope were used for random allocation to groups. The envelope was concealed in a safe and was opened on the morning of the operation. The patients and evaluators were blinded to the group assignments throughout the study.
The inclusion criteria were as follows: Scheduled for unilateral TKA for degenerative arthritis, American Society of Anesthesiologists (ASA) physical status 1–3, weight 40–99 kg, and fully able to understand the study contents from oral and written descriptions. The exclusion criteria were: Scheduled for bilateral TKA, regular narcotic use, allergies to any study drug, neuromuscular disease, sensory disturbances of the legs, severe diabetes, heart failure (New York Heart Association classification greater than 2), renal impairment with an estimated glomerular filtration rate of < 60 mL min−1 1.73 m−2 or liver dysfunction (Child–Pugh classification greater than class B), and inability to be assessed via the Numeric Rating Scale (NRS). All patients received instruction in the use of the NRS, where the number 0 on the ruler represented the absence of pain and the number 10 represented the most severe pain they had ever experienced, and they were asked to indicate the number that best reflected their pain on that scale.
The same team of orthopedists performed all surgeries, and the same anesthesiologist performed the anesthesia for all cases. All patients underwent general anesthesia induced by 2 mg kg−1 propofol and 50 μg of fentanyl, followed by insertion of a laryngeal airway. General anesthesia was maintained with inhaled 1–2 % sevoflurane in oxygen and air. Additionally, remifentanil administration was started at incision from 0.1 μg kg−1 min−1, increased or decreased when the blood pressure or heart rate changed by approximately 10 %, and stopped at least prior to skin closure. No additional narcotics or analgesics including fentanyl were administered during the operation.
The operation was performed using the midvastus approach. None of the participants underwent patellar resurfacing. The resurfacing criterion in our facility is International Cartilage Repair Society Grade IV. The implant models used in this study were of the Triathlon (Stryker, Mahwah, NJ, USA) or Verilast (Smith & Nephew, Andover, MA, USA) type. The implant was fixed with cement (Simplex P Bone Cement with cefazolin) on the cut bone surfaces after confirmation of the joint gap, and balanced by trial and error. A vacuum drain was not used. Edoxaban tosilate hydrate (Lixiana®, Daiichi-Sankyo, Tokyo, Japan; 30 mg) was administered orally for 10 days from postoperative day 2.
Subjects were moved to a lateral recumbent position after insertion of the laryngeal airway. An ultrasound-guided (M-Turbo Fujifilm SonoSite, Tokyo, Japan) SNB approach was utilized in combination with a 0.5–1.0 mA nerve stimulus for 0.1 ms (Stimplex HNS12, B. Braun, Melsungen, Germany). By setting the echo-probe (C60x/5-2 MHz Transducer Fujifilm SonoSite, Tokyo, Japan) parallel to the line created by the posterior superior iliac spine and ischial tuberosity, we inserted a needle (Stimplex D Needle STD-2280, B. Braun, Melsungen, Germany) with an in-plane approach from the lateral aspect of the ultrasound probe and injected 20 mL of 0.375 % ropivacaine after the position of the needle tip was confirmed by a motor response from the ankle.
After SNB, the patient was moved into the supine position, and ultrasound-guided FNB was performed in combination with 0.5–1.0 mA nerve stimulation for 0.1 ms (Stimplex HNS12, B. Braun, Melsungen, Germany). We placed the echo-probe (HFL38x/13-6 MHz Transducer, Fujifilm SonoSite, Tokyo, Japan) parallel to the inguinal ligament region to identify the femoral nerve and then inserted a needle (Contiplex Touphy B. Braun, Melsungen, Germany) using an in-plane approach from the lateral aspect of the ultrasound probe. We then injected 20 mL of 0.375 % ropivacaine after the position was confirmed by contraction of the quadriceps muscle. After the ropivacaine bolus injection, a catheter tip (Aesculap, B. Braun, Melsungen, Germany) was placed 3 cm from the needle tip. A continuous FNB (5 mL h−1) of 0.2 % ropivacaine was commenced at the end of the operation.
Subjects in group L underwent FNB similarly to those in group S. The local anesthetic mixture was prepared from 100 mL of 0.2 % ropivacaine by adding 0.5 mL (0.5 mg) of adrenaline. The local anesthetic mixture was administered three times in measured doses as follows:
20 mL of the local anesthetic mixture was administered intracutaneously to the surgical region at the start of the operation.
50 mL of the local anesthetic mixture was administered as infiltration anesthesia to sites posterior to the articular capsule near the incised part, namely the vastus intermedius, vastus lateralis, and lateral collateral ligament, before the injection of cement.
30 mL of the local anesthetic mixture was administered intraarticularly at the end of the operation.
A continuous femoral block of 0.2 % ropivacaine at 5 mL h−1 was commenced at the end of the operation, as in group S.
The participants were administered a 25-mg diclofenac suppository if their NRS score was > 3 upon exiting the operating room, and if at 3, 6, 12, and 24 h after that time or at any other time they requested an analgesic and had an NRS score higher than 3. Repeated administrations of 25-mg diclofenac suppositories were allowed after a 3-h interval. If this method was inadequate and the patient could not cope with the pain (NRS score > 3), 15 mg of pentazocine was injected intramuscularly. We did not apply a strict fast-track recovery program in this study.
The primary outcome was the change in NRS scores between groups L and S at five time points: upon exiting the operating room and 3, 6, 12, and 24 h later. Nurses in the operating room and on the ward who were blinded to the group allocations recorded the NRS scores.
Other outcome measures included NRS score changes within groups, the area under the curve of the NRS scores, total dose of diclofenac, change in knee flexion and extension, pain control satisfaction, nausea and vomiting, and hospital stay duration.
Physiotherapists measured knee flexion and extension using an angle meter every postoperative day. The Knee Society Score (The 2011 Knee Society Knee Scoring System)  was recorded before the operation and at discharge. Furthermore, the participants were asked about numbness or unusual sensations in their tongue and mouth, metallic tastes, dizziness, tinnitus, and agitation, and they were also checked for possible ropivacaine side effects by evaluating their level of consciousness and convulsions upon exiting the operating room and at 30 min, 1 h, and 3 h after that time. Nausea and vomiting were evaluated at the same time by inquiring about the presence or absence of symptoms. Nurses on the ward who were blinded to the group allocations checked these items. In addition, during the first 24 h after the operation, the nurses and physicians in charge observed the participants closely on the ward every 4 h. If symptoms of ropivacaine toxicity were detected, the continuous infusion of ropivacaine was discontinued, and symptomatic treatment was performed accordingly. In cases of unanticipated side effects, the details were recorded.
Analgesic administration times were recorded, and the total doses were calculated. On the third postoperative day, a survey was administered relating to pain control satisfaction, consisting of a five-point scale (1 = extremely severe pain; 2 = quite severe pain; 3 = neutral; 4 = generally content; 5 = extremely content). No changes to trial outcomes or methods occurred after the trial commenced.
Repeated measures analysis of variance (ANOVA) was used to compare the changes in NRS scores between groups S and L. The differences in NRS scores at each time point and the area under the curve of the NRS scores were analyzed using the Mann–Whitney U test. The changes in NRS scores within each group were analyzed by the Friedman test. The values at 3, 6, 12, and 24 h were compared to those at 0 h by the Wilcoxon signed-rank test with Bonferroni correction in both groups, after differences within groups were confirmed. The P value for the primary outcome measure was determined using a two-tailed test. Other ordinal variables and continuous variables were analyzed by the Mann–Whitney U test or Wilcoxon signed-rank test. Categorical variables were analyzed by a chi-squared test for independence.
Continuous variables are presented as the mean and standard deviation. Ordinal variables are presented as the median and interquartile range. Categorical variables are presented as percentages.
The sample size of the primary outcome measure was calculated by G*Power 3 (Erdfelder & Buchner, 2007). With an alpha error of 0.05, a beta error of 0.2, and an effect size, f, of 0.25, the required sample size for detecting a significant difference between the groups was calculated to be 28. We enrolled 34 cases (17 to each group), anticipating that some patients may be lost to follow-up. Statistical analyses were performed using SPSS (SPSS Statistics 19 for Windows; SPSS Inc, Chicago, IL). Significance was defined as P < 0.05.