Lumbar Plexus Block Reduces Pain and Blood Loss Associated with Total Hip Arthroplasty

After obtaining institutional review board approval and written informed patient consent, we conducted a randomized, controlled, double-blind trial of 60 consecutive patients undergoing elective total hip arthroplasty during general anesthesia. Exclusion criteria were contraindications to regional anesthesia, use of opioids during the preoperative period, and dementia preventing proper comprehension of the study.

Patients were randomly allocated to receive general anesthesia combined with an LP block (plexus group, n = 30) or general anesthesia alone (control group, n = 30). All patients were scheduled for surgery at 8:00 am. After premedication with 7.5 mg midazolam orally, general anesthesia was induced with 4–6 mg/kg sodium thiopental and 2 μg/kg fentanyl and maintained with 0.3–1% isoflurane and 60–70% nitrous oxide. Tracheal intubation was facilitated with use of 0.2 mg/kg mivacurium and lungs were mechanically ventilated (end-tidal carbon dioxide, 30–38 mmHg [4 to 5 kPa]). Patients were placed in the lateral decubitus position, operative extremity superior. After recovery from use of mivacurium (as assessed by train-of-four stimulation of the ulnar nerve or by the appearance of respiratory efforts interfering with mechanical ventilation), patients assigned to the plexus group were administered a single-injection posterior LP block using the approach described by Winnie et al.  8The LP was localized by inducing contractions of the quadriceps femoris with use of a nerve stimulator (DualStim; Life-Tech, Houston, TX), delivering 0.2–0.5 mA impulses of 50 μs at 2 Hz linked to a 23-gauge, 100-mm sterile needle (Pole Needle; Top Corporation, Tokyo, Japan). After aspiration to ensure absence of blood or cerebrospinal fluid, 0.4-ml/kg bupivacaine, 0.5%, with epinephrine 1/200,000 was injected. In the control group, lumbar skin was perforated with use of a needle, but no placebo was administered. In conformity with the blinded study design, the anesthesiologist responsible for the patient was temporarily absent during treatment allocation, leaving the patient in the care of an attending anesthesiologist. The attending anesthesiologist administered the block when assigned, then transferred the patient to the initial staff. The surgical procedure was standardized and was performed with patients positioned as previously mentioned.

Intraoperatively, increments of vecuronium (1 mg) or mivacurium (2 mg) were administered only if warranted to facilitate the surgical procedure or mechanical ventilation. Hemodynamic goals were to maintain mean arterial pressure (MAP) and heart rate within 70–130% of preinduction or baseline levels. MAP increases to more than 130% of baseline were treated by raising end-tidal isoflurane to 1 vol% and administering boluses of fentanyl, 1 μg/kg. A decrease in MAP to less than 70% of baseline was treated with ephedrine, 10 mg intravenously, followed by reduction of end-tidal isoflurane to 0.3%. In the postanesthesia care unit (PACU), a patient-controlled analgesia device was given to all patients and set to deliver intravenous morphine in 1-mg boluses, with a lockout at 6 min and a 4-h maximum of 40 mg. Patients in the PACU reporting pain greater than 3 on the VAS (0 = no pain, 10 = most severe pain) despite patient-controlled analgesia were administered boluses of intravenous “rescue” morphine as needed. In addition, all patients received propacetamol, 2 g intravenously every 8 h administered shortly before the end of surgery and continued for 24 h, and ibuprofen, 400 mg orally every 8 h started on the morning after surgery.

Before selection for the study, a majority of study participants had been enrolled in an autologous blood transfusion program (plexus group, 24 of 30 patients; control group, 19 of 30 patients, P = 0.25). Each patient predonated 2 to 3 units of blood during the weeks preceding surgery. In accordance with common practice at the Hôpitaux Universitaires de Genève, indications for perioperative autologous blood transfusions were not subject to specific guidelines.

Pain scores at rest and morphine consumption were assessed every 30 min in the PACU and then 6, 12, 24, and 48 h after randomization by observers who were blind to treatment allocation. Other recorded variables included intraoperative MAP, heart rate, and end-tidal isoflurane concentration; intraoperative opioid and muscle relaxant use; loss of blood, intraoperative (volume of blood measured in suction canisters and estimated from operative dressings [40 ml blood/dressing]) and postoperative (volume of blood recovered in suction drains before removal at 48 h); bilateral distribution of anesthesia, suggesting epidural spread (tested in the PACU with use of a cold stimulus applied to lower thoracic and lumbar dermatomes contralateral to the operated hip); evidence for block-related neurologic damage at 48 h; postoperative nausea or vomiting; and patient satisfaction with anesthetic and pain management (rated using VAS, 10 = very satisfied, 0 = very unsatisfied). All observations were made by blinded assessors.

Based on previous data, 13,15it was computed that a sample of 30 patients/group would detect a difference between groups in mean morphine consumption of more than 20 mg and a difference in VAS pain scores of more than 1.7 cm, with a power of 90% and a two-tailed significance level of 5% (β= 0.1, α= 0.05). In an analysis made after obtaining results and using other data, 16it was calculated that a sample of this size had 90% power to detect a difference of 130 ml in mean operative blood loss. Where relevant, data are presented as the mean ± SD. Comparisons between groups of continuous variables (MAP, end-tidal isoflurane concentration, morphine use, pain scores, satisfaction scores, blood loss) were assessed using the unpaired Student t  test, whereas comparisons of discontinuous data (proportions of patients) were evaluated using the chi-square test. P < 0.05 was considered to be significant. Some data were analyzed using StatMate and Prism software (GraphPad, San Diego, CA).

Preoperative characteristics were similar in the two groups (table 1). Two patients, one in each group, were excluded from analysis of postoperative data because of delirium, impeding accurate evaluation. One patient in the plexus group received a dose of local anesthetic incompatible with the study requirement and was excluded from intra- and postoperative analysis. Bilateral distribution of anesthesia, suggesting an epidural extension of local anesthetic, occurred in 3 of 28 (10.7%) patients undergoing the block.

Data related to anesthetic and perioperative management are summarized in table 2and figures 1 and 2. The number of patients requiring supplemental fentanyl and the mean end-tidal isoflurane concentration were significantly increased in the control group, whereas intraoperative MAP was lower in the plexus group during most of the procedure. These differences in MAP remained significant even when patients with epidural-type distribution were excluded from analysis (P < 0.01 for the difference between groups in MAP from the twentieth to the sixtieth min of surgery, and P < 0.02 for the difference between groups at the eightieth and ninetieth min). A reduction in blood loss was observed in the plexus group, during the surgical procedure (22% reduction in loss) and 48 h after (45% reduction). When patients who had epidural distribution were excluded from analysis, the intraoperative blood-sparing effect was not statistically significant (plexus group mean blood loss, 434 ± 196 ml vs.  controls, 538 ± 254 ml;P = 0.09), whereas the postoperative loss remained significantly lower (144 ± 97 ml vs.  310 ± 204 ml, P = 0.0006). Eight patients, four in each group, received a unit of autologous blood intraoperatively, and all remaining autologous units were transfused over the ensuing 48 h. No heterologous blood was administered.

Pain scores were significantly reduced in the plexus group until 6 h after randomization (VAS at 6 h, 1.4 ± 1.3 vs.  2.4 ± 1.4, P = 0.007). Differences in pain scores were particularly evident in the immediate postoperative period (fig. 3). Therefore, at arrival to the PACU, 18 of 28 patients in the plexus group reported no pain (VAS = 0) versus  3 of 29 in the untreated control group (P < 0.0001). Conversely, at this same time point, 17 of 29 control patients had severe pain (VAS > 5), versus  2 of 28 of treated patients (P < 0.0001). Morphine administration, including patient-controlled analgesia and rescue doses, was diminished to a significant degree in the treated group until 6 h after randomization (cumulative morphine at 6 h, 5.6 ± 4.7 vs.  12.6 ± 7.5 mg, P < 0.0001), with a trend toward reduced consumption at 12 h (fig. 4). Only 2 of 28 patients required rescue morphine in the plexus group, versus  22 of 29 among controls (P < 0.0001).

Postoperative nausea and vomiting occurred in a minority of patients, and the incidence was similar in both groups. No other side-effects or complications were noted. Patient satisfaction with anesthetic and pain management was not significantly different between the two groups.

Our results show that posterior LP block successfully reduces pain associated with total hip arthroplasty, decreasing intra- and early postoperative opioid requirements. Another interesting finding was a block-associated reduction in hemorrhage.

Peripheral nerve or plexus blocks, with or without perineural catheter placement, constitute an attractive option for anesthetic management of lower limb surgery, either substituting for or complementing other anesthetic techniques. Blocks are appreciated for the superlative and long-lasting analgesia they provide; moreover, when compared with other types of anesthesia, nerve blocks may enhance intraoperative hemodynamic stability, 4,17decrease risk of urinary retention, 3,18reduce risk of puncturing the dura mater, and diminish surgical blood loss. 16 

The LP is formed by the ventral rami of the first four lumbar roots, with a contribution from the twelfth thoracic root. These elements converge in the posterior part of the psoas muscle and produce seven to eight branches, all of which contribute to a varying degree to the sensory innervation of the hip region. 5The hip joint is innervated by branches of both the LP and the sacral plexus, including the femoral (L2–L4) and obturator nerves (L2–L4), which contain afferents from the anteromedial aspects of the joint, and the sciatic nerve (L4–S3), the nerve to the quadratus femoris (L4–S1) and the superior gluteal nerve (L4–S1), which cover the posterior aspect of the joint. 5,6Although the relative contributions of lumbar and sacral components to sensory innervation are not known, our results suggest that local anesthetic blockade of the LP is sufficient to provide effective analgesia early after total hip arthroplasty. Although this finding is in itself of interest, we predict that its clinical relevance will be further shown by the application of continuous catheter-based techniques, as indicated in a recent preliminary report. 19 

Lumbar plexus blocks were developed by Winnie et al. , who described both an anterior 7and a posterior 8technique. The anterior, or 3-in-1, approach was based on the hypothesis that a large volume of local anesthetic placed in the femoral nerve sheath would spread proximally to the LP and block other branches, including the obturator and lateral femoral cutaneous nerves. Although the original account supported this postulate, subsequent reports failed to show reliable blockade of obturator or lateral femoral cutaneous nerves with use of this technique, and the concept of the 3-in-1 technique as a plexus block is contested. 11,12With the posterior LP approach, local anesthetic is deposited within or adjacent to the dorsal portion of the psoas muscle. Several methods of performing the posterior LP block have been described, 8–10all of which consistently produce anesthesia of LP branches important to lower limb surgery. In addition, some authors using posterior techniques have noted simultaneous anesthesia of lumbar and sacral plexi, 8,20whereas others have not. 11,14 

Few studies have addressed the clinical interest of peripheral nerve or plexus blocks for surgery of the hip. Using a posterior LP technique, Chayen et al.  9reported successful anesthesia in 52 of 57 hip procedures, but recommended combining lumbar and sciatic blocks in this indication. In a trial that compared general anesthesia alone or combined, either with subarachnoid block or with posterior LP block for femoral neck fracture surgery, White and Chappell 17noted greater cardiovascular stability in the group receiving the block; but provided no data regarding other outcomes, such as postoperative analgesia. In a study of hip replacement, Odoom et al.  21compared bupivacaine with or without epinephrine in a posterior LP block and noted lower peak bupivacaine concentrations and a prolongation of mean analgesia in the epinephrine group. Dalens et al. , 20reporting a pediatric population of which 9 of 50 children underwent hip arthrotomies, contrasted posterior LP block performed according to the techniques of Winnie et al.  8or Chayen et al. , 9noting lumbosacral anesthetic distribution in the former group and bilateral or epidural-like distribution in the latter, without providing information about pain or analgesia. Finally, Fournier et al.  13compared general anesthesia combined with 3-in-1 block versus  general anesthesia alone for total hip arthroplasty and observed a longer opioid-free interval after surgery in the block group, although intraoperative use of opioids and isoflurane was not influenced by the block. Contrasting these results and ours suggests that the posterior LP block, by providing more extensive anesthetic coverage of the region, may be better suited to surgical procedures of the hip than is the 3-in-1 approach.

The decreased blood loss we report, intraoperatively and postoperatively, confirm the findings of Twyman et al. , 16although only intraoperative loss was significantly reduced in that study. The clinical importance of this hemorrhage-sparing effect in terms of transfusion requirements was not directly addressed in the current study because most patients predonated blood for which transfusion policy was not guideline-directed, prohibiting a proper statistical comparison between groups. Diminished hemorrhage has been documented with various regional anesthetic techniques, including spinal and epidural anesthesia, 1and is thought to result from attenuated sympathetic tone in medium and small vessels, with concomitantly reduced arterial and venous pressure. Two distinct mechanisms may influence blood loss in patients undergoing peripheral nerve block: a direct effect on vasoconstrictive sympathetic fibers contained in peripheral nerves and an indirect effect mediated by antinociception and reduced systemic blood pressure. In this regard, we noted lower intraoperative MAP in the plexus group (fig. 2). Because this difference was observed in the presence of equivalent baseline MAP measurements and similar prevalence of treated arterial hypertension in the two groups, with lesser administration of isoflurane and fentanyl in the plexus group, we suggest that it is attributable to attenuated nociception and autonomic arousal in the block-treated patients. Finally, lower intraoperative blood loss in the plexus group may be related to epidural-like blockade appearing in some of these patients.

In 3 of 28 patients in the plexus group, bilateral distribution of anesthesia developed, evoking spread of local anesthetic to the epidural space. Hypotension or extension of block to the upper thoracic dermatomes did not develop in any of these patients. Epidural extension of anesthesia is a known side effect of posterior LP block, occurring in 3–10% of adult patients without adverse consequences. 9,11,14Notwithstanding, we believe that this result should be systematically checked after posterior LP block has been performed. Reported complications of posterior LP block, such as total spinal anesthesia, 22renal subcapsular hematoma, 23and psoas hematoma with lumbar plexopathy, 24were not observed in this trial.

The strengths of this study include the fact that it was designed to detect targeted differences between groups in the two major end points (pain, morphine use). A post hoc  analysis showed that the trial also was sufficiently powered to detect observed differences in an ancillary end point: blood loss. Among the limitations, we noted that, although care was taken to ensure the double-blind design, observations such as asymmetric thigh mobility, epidural blockade, or decreased opioid requirements might have introduced bias among patients and data collectors.

In summary, LP block performed using a posterior approach is an effective complement to anesthetic and analgesic management of total hip arthroplasty, reducing opioid administration and blood loss during the surgical procedure and in the early postoperative period. Future studies should address the possibility of extending the duration of this benefit in the postoperative period, namely via  continuous catheter-based techniques. Studies are also needed to confirm reduced bleeding in the setting of peripheral nerve blockade and to explore the clinical importance of this effect.