Although lumbar plexus blocks (LPBs) provide reliable analgesia for lower limb surgery,1 because of reports of total spinal anaesthesia,2,3 retroperitoneal, psoas and renal sub-capsular haematoma, they have been not been popular.4–8 Increasingly, anaesthetists rely on ultrasound guidance to perform LPB.9–13 Because of its depth, the lumbar plexus can be difficult to visualise,12 prompting most experts to recommend combining ultrasound and neurostimulation to elicit a quadriceps-evoked motor response prior to local anaesthetic injection.11,12 However the combined use of ultrasound and neurostimulation does not always improve block success. Studies of infraclavicular,14,15 axillary16 and femoral blocks17 have concluded that compared with ultrasound alone, combined ultrasound–neurostimulation unnecessarily lengthens the performance time without increasing the success rate.
An alternative ultrasound technique for LPB consists of simply injecting local anaesthetic inside the posteromedial quadrant of the psoas muscle (without elicitation of quadriceps-evoked motor response).18,19 Compared with the conventional ultrasound–neurostimulation method, the potentially slower onset of the simplified ultrasound technique can be compensated for by a quicker performance time. In this randomised trial, we sought to formally compare ultrasound–neurostimulation and ultrasound-guided LPBs. Our primary outcome was the total anaesthesia time (defined as the sum of performance and onset times). We hypothesised that it would be equivalent for the two groups. Consequently, we designed the current study as an equivalence trial.
Ethics committee approval (11 April 2016 for the Ramathibodi Hospital; 270 Rama VI Road; Ratchathewi; Bangkok, Thailand; protocol number: ID03-59-02; and 29 April 2016 for the Maharaj Nakorn Chiang Mai Hospital; 110 Intavaroros Rd, Amphoe Muang, Chiang Mai, Thailand; protocol number: ANE-2529-03724) was secured prior to patient recruitment. The current trial was registered at wwwclinicaltrials in the (study ID: TCTR20160427003) on 25 April 2016.
After obtaining written informed consent, we enrolled patients undergoing total hip or knee arthroplasty. Inclusion criteria were age between 18 and 75 years, American Society of Anesthesiologists physical status I to III and BMI between 18 and 35 kg m−2. Exclusion criteria included inability to consent to the study, coagulopathy, sepsis, hepatic or renal failure, allergy to local anaesthetic, pre-existing neuropathy (due to diabetes, vascular disease or any other cause), previous surgery in the lumbar spine and chronic opioid intake at home.
In the induction room, an 18-gauge intravenous catheter was placed in the upper limb and intravenous pre-medication (up to 2 mg of midazolam and 50 μg of fentanyl) were administered if required. Supplemental oxygen (nasal cannulae at 4 l min−1) was administered and an ECG, non-invasive blood pressure (BP) monitor and pulse oximeter were attached throughout the procedure. Using a computer-generated sequence of random numbers and a sealed envelope technique, patients were randomly allocated to receive an ultrasound–neurostimulation or ultrasound LPB. Patients were randomised in blocks of 10 to ensure equal distribution between the ultrasound–neurostimulation and ultrasound groups. Based on projected surgical volume, the Bangkok and Chiang Mai centres were assigned seven and four blocks of 10 patients, respectively. All LPBs were supervised by one of four co-authors (VA, TC, PL and WT) and performed by consultants or supervised trainees.
For both techniques, patients were placed in the lateral decubitus position with the surgical limb uppermost. After skin disinfection and draping, the 2 to 6-mHz curved array ultrasound transducer (Sonosite M-Turbo; SonoSite Inc, Bothell, Washington, USA or GE Logiq e, GE Healthcare, Chicago, Illinois, USA) was applied in a sterile fashion between the iliac crest and the costal margin to obtain a ‘shamrock’ view.13 The vertebral body, transverse process and psoas muscle were identified. The puncture site was located 4-cm lateral to the midline (Fig. 1a). A skin wheal was raised with 3 ml of lidocaine 1%. Using an in-plane technique and a posterior-to-anterior direction, the 100-mm, 22-gauge, short-beveled block needle (Uniplex Nanoline, Geisingen, Germany), which was connected to a neurostimulator (set at 1.5 mA, 0.1 ms and 2 Hz), was advanced until its tip was anchored in the erector spinae muscles.
In the ultrasound group, the needle was subsequently advanced until its tip was positioned in the posteromedial quadrant of the psoas muscle (Fig. 1b). Neurostimulation was used solely to document the occurrence of quadriceps-evoked motor response when the needle tip was positioned in the posteromedial psoas. Its position was not modified and was maintained despite the presence (or absence) of a quadriceps contraction. In the ultrasound–neurostimulation group, the operator searched for a quadriceps-evoked motor response at a stimulatory threshold between 0.2 and 0.8 mA.12 In both groups, after achieving the desired technical endpoint, the operator injected a bolus of 30 ml of lidocaine 1% and levobupivacaine 0.25% (obtained by mixing equal parts of lidocaine 2% and levobupivacaine 0.5%) with epinephrine 5 μg ml−1 and 5 mg of preservative-free dexamethasone. The injection was carried out in incremental fashion with repeated (negative) aspiration after each 5-ml aliquot.
For both groups, the imaging time was defined as the temporal interval between contact of the ultrasound probe with the patient and the acquisition of a ‘shamrock’ sonogram.13 The needling time (defined as the temporal interval between the start of the skin wheal and the end of local anaesthetic injection through the block needle) was also recorded. The sum of imaging and needling times made up the performance time. The number of needle passes was also recorded. The initial needle insertion counted as the first pass. Any subsequent needle advance that was preceded by a retraction of at least 10-mm counted as an additional pass.20 The incidence of vascular puncture and paraesthesia were also recorded. All the preceding observations (imaging and needling times, number of passes and the occurrence of paraesthesia and vascular breach) were assessed by the co-author supervising the block.
After local anaesthetic injection through the block needle, measurements of LPB were carried out by a blinded observer every 5 min for 30 min. Sensory block was assessed in the anterior, lateral and medial aspects of the mid-thigh. For each territory, blockade was evaluated using a 3-point scale: 0 = no block, 1 = analgesia (patient can feel touch, not cold) and 2 = anaesthesia (patient cannot feel touch). Motor block was assessed using knee extension and hip adduction. Knee extension was graded according to a 3-point scale: 0 = no block, 1 = paresis (decreased ability to extend the leg) and 2 = paralysis (inability to extend the leg). Hip adduction was evaluated by comparing it with baseline strength. Before the LPB, a BP cuff, inflated at 40 mmHg, was inserted between the knees of the patient: the latter was then instructed to squeeze the cuff as hard as possible and to sustain the effort. We recorded this baseline pressure measurement. After the LPB, the same manoeuvre was repeated every 5 min for 30 min. We defined hip adduction scores of 0, 1 and 2 points as decreases in strength of 0 to 20, 21 to 70 and 71 to 90%, respectively, compared with baseline measurement.
Overall an LPB was considered successful if, at 30 min, a minimal composite score of 8 points out of a maximum of 10 points was obtained. Onset time was defined as the time required to achieve a score of at least 8 points. The blinded observer also recorded the procedural pain (0 = no pain; 10 = worst imaginable pain) and the occurrence of local anaesthetic spread to the epidural space (defined as the presence of bilateral lower extremity sensory block at 30 min).
After the 30-min block assessment, patients were taken to the operating room for the start of surgery. The choice of intraoperative technique (general vs. neuraxial anaesthesia) was left to the managing anaesthetist. Postoperatively the patient received patient-controlled analgesia with intravenous morphine (bolus dose = 1 mg; lockout interval = 7 min). The blinded observer recorded the cumulative breakthrough opioid consumption at 24 h. One week after the surgery, the blinded investigator contacted all patients to inquire about complications such as persistent numbness, paraesthesia or motor deficit.
In a pilot study (n = 10), we observed a 90% success rate (minimal composite score of 8 points at 30 min), a mean (±SD) performance time of 5.3 (±2.2) min, a mean (±SD) onset time of 19.4 (±9.5) min and a mean (±SD) total anaesthesia time of 24.8 (10.2) min for the ultrasound–neurostimulation technique (unpublished data). We considered that a 30% difference in total time (7.4 min) had little clinical relevance, and we therefore set our equivalence margin at ± 7.4 min. For a statistical power of 0.9 and a type1 error of 0.025, we calculated that a sample size of 48 patients per group was required. As total time can only be calculated for successful blocks, a total of 110 patients were planned to account for the anticipated 10% failure rate.
We designed the current protocol as an equivalence instead of a superiority trial as our preliminary experience with ultrasound LPB suggested that its total anaesthesia time would be similar to the one provided by ultrasound–neurostimulation LPB. From a conceptual standpoint, we discarded the non-inferiority design as we aimed to show that ultrasound LPB was no better or no worse than its ultrasound–neurostimulation counterpart. In contrast, a non-inferiority trial would have been desirable had we only wanted to demonstrate that the new treatment (ultrasound LPB) was no worse than the active control (ultrasound–neurostimulation LPB).21 Statistical analysis was performed using SPSS version 21 statistical software (IBM, Armonk, New York, USA). For continuous data, normality was first assessed with Lilliefors test, and, if normal distribution was not rejected, we employed the Student t test. Data that did not have a normal distribution, together with ordinal data, were analysed with the Mann–Whitney U test. For categorical data, Pearson's χ2 test was used. When the expected count was less than five, Fisher's exact test was employed. All P values presented are two-sided, and values inferior to 0.05 were considered significant.
The 110 patients, 55 in each group, were recruited over a period of 8 months (12 May 2016 to 10 January 2017) (Fig. 2). As planned, 70 and 40 patients were enrolled in Bangkok and Chiang Mai, respectively. There were no inter-group differences in terms of personal data (Table 1). Except for one patient (ultrasound–neurostimulation group), all patients underwent neuraxial anaesthesia for intraoperative management.
Analysis of our primary outcome revealed that compared with ultrasound alone, the ultrasound–neurostimulation technique resulted in a shorter mean (±SD) total anaesthesia time [15.3 (±6.5) vs. 20.1 (±9.0) min; mean difference, −4.8; 95% confidence interval, −8.1 to −1.9; P = 0.005].
Analysis of secondary outcomes also reveals a difference in mean (±SD) onset times between ultrasound–neurostimulation and ultrasound techniques [10.2 (±5.6) vs. 15.5 (±9.0) min, respectively; P = 0.004] (Table 2). From 10 to 20 min, more patients in the ultrasound–neurostimulation group reached a minimal composite score of 8 points (all P ≤ 0.007). However, no differences were found thereafter (Fig. 3). No inter-group differences were observed in terms of success rates (89 to 96%), imaging and needling performance, number of needle passes, procedural pain and opioid consumption at 24 h (Table 2).
In the ultrasound–neurostimulation group, quadriceps-evoked motor response was obtained at a stimulatory threshold between 0.2 and 0.8 mA in 100% of patients. In the ultrasound group, quadriceps contraction occurred at the final position of the needle tip (posteromedial quadrant of the psoas muscle) in only 58% of cases.
No inter-group differences were recorded in terms of adverse events (vascular breach, paraesthesia, local anaesthetic spread to the epidural space) (Table 2). Follow-up at 1 week revealed no sensory or motor deficit.
In this randomised, observer-blinded trial, we compared ultrasound–neurostimulation and ultrasound techniques for LPB. Our results show that the ultrasound–neurostimulation method provides a 5-min shorter total anaesthesia time. Although statistically significant, this difference falls within our a priori accepted 30%-equivalence margin (±7.4 min). Thus, within the limits set by our study, we conclude that neurostimulation provides small benefit for ultrasound-guided LPB. During the inception phase of the trial, based on the 24.8-min total time demonstrated by our pilot study, we selected a 30% margin to ensure clinical relevance. We reasoned that margins of 20% (5 min) or 25% (i.e. 6.3 min) would have little clinical significance.
In the current trial, the mean onset (10.2 min) and total anaesthesia times (15.3 min) for ultrasound–neurostimulation are significantly shorter than the ones previously recorded for the pilot study (onset = 19.4 min; total time = 24.8 min). As the latter employed similar operators, an identical local anaesthetic mix and resulted in a comparable performance time (5.3 min), we attribute the differences to its small sample size (n = 10).
Our assessment of hip adduction deserves special mention. Hip adduction results from the combined action of five different muscles: the pectineus, gracilis, adductor longus, adductor brevis and adductor magnus muscles. Although the first four muscles are supplied by the femoral nerve and/or the obturator nerve, the adductor magnus muscle receives dual innervation by the obturator and the sciatic nerve. Jochum et al.22 have estimated that the sciatic nerve is responsible for 11.3 (±7.0)% of hip adduction. Thus, even with a complete LPB, patients may still be able to partially adduct the hip. Therefore, to properly assess decreased hip adduction related to LPB, we elected to evaluate paresis by comparing it with baseline strength.
Our ultrasound technique requires discussion. Although a methodological argument could be made for further refining the position of the needle tip (aiming for a stimulatory threshold ≤ 0.8 mA) if quadriceps contraction was incidentally encountered in the ultrasound group, we intentionally refrained from doing so. According to our protocol, neurostimulation was used solely for documentation purposes in this group: we wanted to elucidate the proportion of patients in whom a quadriceps-evoked motor response would occur if the needle tip were simply positioned in the posteromedial quadrant of the psoas muscle.
In the ultrasound group, quadriceps contraction occurred in fewer than 60% of patients despite careful positioning of the needle tip in the posteromedial psoas. This finding could be partly explained by the anatomy of the lumbar plexus itself; cadaveric and clinical studies have revealed that in some cases, the lumbar plexus is actually adjacent to the lumbar transverse processes and the quadratus lumborum muscle, dorsal to the psoas muscle (and not inside it).23–25 More importantly, the relationship between needle–nerve proximity and evoked motor response may be more complex than previously thought.26 During the physical advancement of needles towards neural structures, three events can occur: needle–nerve contact, mechanical paraesthesia and evoked motor response. Although undeniable overlap exists amongst them, they also depend on additional factors that remain poorly understood.26 For instance, in 2006, Perlas et al.27 performed an ultrasound-guided axillary block in 102 patients and intentionally made contact with one of the terminal nerves with the needle tip. After ultrasound confirmation of needle–nerve contact, a neurostimulator was turned on and the current incrementally increased until a maximum of 2.0 mA (0.1 ms) or until the elicitation of an evoked motor response. These authors observed an evoked motor response in only 74.5% of cases.27
In our two centres, the routine use of LPB for total knee arthroplasty may appear surprising. In recent years in North America and Europe, there has been considerable effort to develop motor sparing nerve blocks (adductor canal blocks) for knee replacement surgery.28 In addition to protecting against falls, motor sparing strategies allow early ambulation and better cooperation with postoperative physiotherapy. However, in our practice, assisted ambulation only begins on postoperative day 2. This does not allow the benefits provided by motor sparing nerve blocks. Instead, combined femoral and obturator nerve blocks were our preferred analgesic routine for total knee replacement.29,30 Recently, in an effort to improve efficiency, we have gradually moved to LPBs. The latter offer similar analgesic benefits to their femoral-obturator counterparts but require only one procedure instead of two distinct blocks. Furthermore, the same patient position (lateral decubitus position with the operative leg uppermost) can be used for both the LPB and the subsequent spinal block.
Our protocol contains some limitations. First, we did not assess postoperative pain scores. This technical trial intended solely to validate the simplified ultrasound technique by comparing it with the conventional ultrasound–neurostimulation method. Future studies are required to investigate differences in postoperative pain control conferred by continuous ultrasound–neurostimulation and ultrasound LPBs. Second, to visualise the psoas muscle we used a ‘shamrock’ view.31 We recognise that alternative sonographic windows have been described.10–12 However, the latter should not influence our findings; irrespective of the ultrasound view, the two target endpoints (ultrasound and ultrasound–neurostimulation) for needle placement remain the same. Third, our results inherently depend on the local anaesthetic volume. In the current study, we employed 30 ml because a recent dose-finding trial has reported that the minimum effective volume (of ropivacaine 0.5%) required for a successful LPB in 95% of patients could be as high as 36 ml.32 However, we concede that a smaller injectate would have altered our findings and potentially favoured the ultrasound–neurostimulation technique. Finally, our results depend on the contemporary state of ultrasound technology. In the future, if ultrasound permits reliable identification of the lumbar plexus in all patients, the position of the needle tip could be refined beyond simple placement inside the posteromedial quadrant of the psoas muscle.
In conclusion, although the ultrasound–neurostimulation technique results in a statistically shorter total anaesthesia time compared with ultrasound alone, this difference falls within our accepted equivalence margin (±7.4 min). Thus, adjunctive neurostimulation seems to provide small benefit for ultrasound-guided LPB. Future trials are required to investigate the postoperative analgesia conferred by continuous ultrasound–neurostimulation and ultrasound LPBs.
Acknowledgements relating to this article
Assistance with the study: none.
Financial support and sponsorship for this study: none.
Conflicts of interest: none.
1. Hojer Kolsen AP, Geisler A, Petersen PL, et al. Postoperative pain treatment after hip arthroplasty: a systematic review. Pain
2. Pousman RM, Mansoor Z, Sciard D. Total spinal anesthetic after continuous posterior lumbar plexus block. Anesthesiology
3. Dogan Z, Bakan M, Idin K, et al. Total spinal block after lumbar plexus block: a case report. Rev Bras Anestesiol
4. Aveline C, Bonnet F. Delayed retroperitoneal hematoma after failed lumbar plexus block. Br J Anaesth
5. Dauri M, Faria S, Celidonio L, et al. Retroperitoneal hematoma in a patient with continuous psoas compartment block and enoxaparin administration after total knee replacement. Br J Anaesth
6. Klein SM, D’Ercole F, Greengrass RA, et al. Enoxaparin associated with psoas hematoma and lumbar plexopathy after lumbar plexus block. Anesthesiology
7. Aida S, Takahashi H, Shimoji K. Renal subcapsular hematoma after lumbar plexus block. Anesthesiology
8. Weller RS, Gerancher JC, Crews JL, et al. Extensive retroperitoneal hematoma without neurologic deficit in two patients who underwent lumbar plexus block and were later anticoagulated. Anesthesiology
9. Tognù A, Cauli V, de Simone N, et al. In-plane ultrasound-guided lumbar plexus block using catheter-over needle technique in a 14-month-old baby. Reg Anesth Pain Med
10. Doi K, Sakura S, Hara K. A modified posterior approach to lumbar plexus using a transverse ultrasound image and an approach from the lateral border of the transducer. Anaesth Intensive Care
11. Karmakar MK, Ho AM, Li X, et al. Ultrasound-guided lumbar plexus block through the acoustic window of the lumbar ultrasound trident. Br J Anaesth
12. Karmakar MK, Li JW, Kwok WH, et al. Ultrasound-guided lumbar plexus block using a transverse scan through the lumbar intertransverse space. Reg Anesth Pain Med
13. Lin JA, Lu HT, Chen TL. Ultrasound standard for lumbar plexus block. Br J Anaesth
14. Dingemans E, Williams SR, Arcand G, et al. Neurostimulation in ultrasound-guided infraclavicular block: a prospective randomized trial. Anesth Analg
15. Gurkan Y, Tekin M, Acar S, et al. Is nerve stimulation needed during an ultrasound-guided lateral sagittal infraclavicular block? Acta Anaesthesiol Scand
16. Chan VW, Perlas A, McCartney CJ, et al. Ultrasound guidance improves success rate of axillary brachial plexus block. Can J Anesth
17. Sites BD, Beach ML, Chinn CD, et al. A comparison of sensory and motor loss after a femoral nerve block conducted with ultrasound versus ultrasound and nerve stimulation. Reg Anesth Pain Med
18. Kirchmair L, Entner T, Kapral S, et al. Ultrasound guidance for the psoas compartment block: an imaging study. Anesth Analg
19. Madison SJ, Ilfeld BM, Loland VJ, et al. Posterior lumbar plexus perineural catheter insertion by ultrasound guidance alone. Acta Anaesthesiol Scand
20. Casati A, Danelli G, Baciarello M, et al. A prospective, randomized comparison between ultrasound and nerve stimulation guidance for multiple injection axillary brachial plexus block. Anesthesiology
21. Lesaffre E. Superiority, equivalence, and noninferiority trials. Bull NYU Hosp Jt Dis
22. Jochum D, Iohom G, Choquet A, et al. Adding a selective obturator nerve block to the parasacral sciatic nerve block: an evaluation. Anesth Analg
23. Kirchmair L, Lirk P, Colvin J, et al. Lumbar plexus and psoas major muscle: not always as expected. Reg Anesth Pain Med
24. Mannion S, Barrett J, Kelly D, et al. A description of the spread injectate after psoas compartment block using magnetic resonance imaging. Reg Anesth Pain Med
25. Capdevila X, Macaire P, Dadure C, et al. Continuous psoas compartment block for postoperative analgesia after total hip arthroplasty: new landmarks, technical guidelines, and clinical evaluation. Anesth Analg
26. Tran DQH, Techasuk W, Finlayson RJ. Elicitation of paresthesia, peripheral nerve stimulation and intraneural injection: new insight provided by ultrasonography. Adv Anesth
27. Perlas A, Niazi A, McCartney C, et al. The sensitivity of motor response to nerve stimulation and paresthesia for nerve localization as evaluated by ultrasound. Reg Anesth Pain Med
28. Bendtsen TF, Moriggl B, Chan V, et al. The optimal analgesic block for total knee arthroplasty. Reg Anesth Pain Med
29. McNamee DA, Parks L, Milligan KR. Postoperative analgesia following total knee replacement: an evaluation of the addition of an obturator nerve block to combined femoral and sciatic nerve block. Acta Anaesthesiol Scand
30. Macalou D, Trueck S, Meuret P, et al. Postoperative analgesia after total knee replacement: the effect of an obturator nerve block added to the femoral 3-in-1 nerve block. Anesth Analg
31. Stried JMC, Sauter AR, Ullensvang K, et al. Ultrasound-guided lumbar plexus block in volunteers; a randomized controlled trial. Br J Anaesth
© 2018 European Society of Anaesthesiology
32. Sauter AR, Ullensvang K, Niemi G, et al. The shamrock lumbar plexus block. A dose finding study. Eur J Anaesthesiol