We prospectively enrolled 64 patients (Table ) undergoing unilateral THA for a diagnosis of osteoarthritis between November 2003 and April 2005. During this time, 452 THAs were performed in our unit in 440 patients for all diagnoses (Fig. ). Of these 440 patients, 143 were under the care of the surgeons contributing to this study, and of these 143 we assessed 115 for eligibility. Patients were identified at the preoperative assessment stage. We considered patients younger than 80 years weighing from 50 to 120 kg, who were able to provide informed consent for and cooperate with the study with a diagnosis of primary osteoarthritis. We excluded two patients owing to major psychological problems, previous drug dependency, allergies to any of the ingredients of the injection, renal insufficiency, abnormal liver enzymes, a history of stroke or major neurologic deficit, or uncontrolled angina and bifascicular blocks with prolonged QT intervals. Using randomization tables, patients were randomized to one of two groups: 32 patients received an intraoperative periarticular injection and 32 patients did not. Patients and assessors were blinded; owing to the study design, it was not possible to blind the surgeons. A power study, using an estimated 30% decrease in 24-hour PCA consumption as being clinically important, showed statistical significance (alpha = 0.05) was reached using 23 patients in each treatment arm (beta = 0.8); for a significance value of alpha = 0.01, 32 patients were required in each arm. Previous studies that have stated their power calculation have used a 20% decrease in narcotic requirement and other parameters [24
] or a 25% reduction in VAS scores [1
]. From our previous work, we believed a 30% reduction in PCA requirement would need to be achieved to be significant. No patients were lost to followup at the 6-week end point. All patients consented to inclusion in the study. We had prior approval of our local ethics committee (in accordance with the ethical standards of the responsible committee on human experimentation and with the Declaration of Helsinki of 1975 as revised in 2000).
Demographic data for patients participating in the study
A CONSORT flowchart describes patient selection.
The THAs were performed through a direct lateral approach. Operative anesthesia was either general or spinal. The anesthetic regime was standardized. No long-acting analgesics were used and spinal anesthesia was performed using spinal bupivacaine hydrochloride, ranging from 10 to 15 mg. Twenty-three patients in the injected group and 24 in the noninjected group had a general anesthetic. Nine patients in the injected group and eight in the noninjected group had a spinal anesthetic. With these small numbers, there was no difference (chi square test, p = 0.78) between general versus spinal anesthetic, which was selected at the anesthesiologist’s discretion.
The injection contained 400 mg ropivacaine (ropivacaine HCl; Astra Zeneca Canada Inc, Ontario, Canada), 30 mg ketorolac (Toradol®
; Sabex Inc, Boucherville, Canada), 0.6 mL epinephrine (1:1000; Abbott Laboratories, Abbott Park, IL), and 5 mg preservative-free morphine (Sabex Inc). These were mixed with sterile normal saline and made up to a combined volume of 100 mL in the operating room [5
]. During the operation, 20 mL of the mixture was injected into the posterior capsule after femoral component insertion and another 20 mL in the anterior capsule before hip reduction and capsular closure. The remaining 60 mL was placed in the fat and subcuticular tissues. Five patients in each group had blood samples taken at 30 minutes, 1 hour, and 4 hours postoperation to measure venous blood (protein-bound) ropivacaine levels.
All patients received PCA for 24 hours postsurgery (morphine: bolus 1.5 mg, lockout 6 minutes and maximum 15 mg/hour). All patients had a lower limb ultrasound to exclude a deep vein thrombosis at Day 5 postsurgery.
Postoperative PCA consumption was measured for the first 24 hours and the patient’s overall analgesic consumption until discharge was measured and converted to morphine equivalents to allow for comparison of the groups. Other analgesics used by the patients during the postoperative period included Percocet™ (acetaminophen 325–650 mg, oxycodone 5–10 mg; Endo Pharmaceuticals Inc, West Chester, PA), Tylenol®
(acetaminophen 325 mg, caffeine 15 mg, codeine 8 mg; McNeil Consumer Healthcare, Fort Washington, PA), and Tylenol®
3 (acetaminophen 300 mg, codeine 30 mg; McNeil Consumer Healthcare). The following conversion ratios were used for calculating morphine equivalents: morphine:oxycodone 1:2 and morphine:codeine 1:20 [13
]. The total dose of a drug was calculated for the period of interest and then converted to morphine equivalents for analysis. We measured ropivacaine levels in five patients 4 hours postoperation. The criteria for hospital discharge were the ability to safely walk up and down a flight of stairs and a safe environment for discharge.
Patients were assessed using VAS scores for pain, at rest and during activity, and for patient satisfaction, in the preoperative assessment clinic (2–3 weeks before surgery), on the day of surgery, in the postanesthetic care unit (PACU), at some time during the inpatient stay, and finally at 6 weeks’ followup. The VAS for pain and satisfaction ran from 0 mm (indicating no pain or completely dissatisfied) to 100 mm (indicating extreme pain or completely satisfied) in 10-mm increments. The activity performed relating to the VAS assessment of pain on activity was leg movements while supine within the limits of any regional anesthesia while in the PACU. At other times, the activity was the rehabilitation activity relevant to that stage of the patient’s recovery. Specific note was made of any signs of cardiac (chest pain, shortness of breath, or electrocardiogram changes) or central nervous system toxicity (visual and hearing disturbances, dysarthria, tingling, perioral numbness, dizziness, paresthesia, light-headedness, muscular twitching, or muscular rigidity), or wound complications. We observed wounds daily and noted any wound drainage, erythema, swelling, blisters, desquamation, dehiscence, protuberant suture material, or signs of infection. As part of our routine clinical care during the period of this study, Harris hip scores and WOMAC scores were collected prospectively for all patients (Table ). These data were available out to 2 years by the time of this study. All patients had a lower limb ultrasound to assess for the presence of deep vein thrombosis before discharge.
Preoperative and postoperative hip scores
We determined differences in postoperative PCA requirements, postoperative VAS scores for pain and satisfaction, the operating time, and the length of hospital stay between the injected and noninjected cohorts using a two-tailed unpaired t test. In all five of these analyses, the data passed the Kolmogorov-Smirnov test for normality (p < 0.05). We determined the difference in the incidence of complications between the injected and noninjected cohorts using Fisher’s exact test (nonparametric). SPSS® (Version 11.5; SPSS Inc, Chicago, IL) was used for the analyses.