Secondary Logo

Journal Logo

Regional Anesthesia: Research Report

Neurostimulation in Ultrasound-Guided Infraclavicular Block: A Prospective Randomized Trial

Dingemans, Emmanuel MD*; Williams, Stephan R. MD, PhD*; Arcand, Geneviève MD, FRCPC*; Chouinard, Philippe MD, FRCPC*; Harris, Patrick MD, FRCSC; Ruel, Monique RN*; Girard, François MD, FRCPC*

Author Information
doi: 10.1213/01.ane.0000226101.63736.20

Infraclavicular brachial plexus block is an anatomically attractive approach for anesthesia of the arm and hand because at this level, the nerve cords are compactly distributed around the axillary artery. However, the clinical popularity of infraclavicular block has been limited by the lack of reliable surface landmarks and a reasonable fear of complications such as vascular puncture and pneumothorax (1,2). In recent years, ultrasound-guided (USG) brachial plexus blocks have gained popularity, allowing real-time visualization of the neurovascular bundle, the advancing needle, and local anesthetic deposition. USG has been shown to decrease procedural times and increase block quality in both adult (3,4) and pediatric patients (5) by providing more reliable landmarks for nerve localization and local anesthetic deposition than surface anatomy and neurostimulation alone.

Studies examining the quality of infraclavicular block using neurostimulation as the end point for local anesthetic injection have reported success rates for surgery in the range of 80%–90%, whether USG was used to locate the brachial plexus (1,6–8) or not. However, successful neurostimulator-guided infraclavicular block is dependant on obtaining a distal stimulation, which may be difficult in 5%–21% of patients (1,9), and occasional failures still occur despite adequate stimulation. In response to this problem, ultrasonic visualization of local anesthetic spread has been proposed as a new end point for injection. Published series using this end point have obtained complete blocks in 90%–95% of patients (10,11) and success rates for surgery approaching 100%. It may therefore be that neurostimulation is no longer required when performing USG infraclavicular block; the visualization of appropriate local anesthetic deposition guarantees successful block. To answer this question, we conducted a prospective and randomized trial comparing USG infraclavicular block with and without neurostimulation. We hypothesized that local anesthetic deposition guided by ultrasound alone would be more rapidly performed while offering a rate of successful block comparable to local anesthetic injection based on a neurostimulatory end point.


After IRB approval and written informed consent, 72 consecutive patients scheduled for hand, forearm, and distal arm surgery were recruited and randomized into one of two groups: Group U (USG alone) and Group S (USG with neurostimulation). Exclusion criteria were any condition precluding informed consent, significant coagulopathy, previous contralateral pulmonary resection, infection at the puncture site, known allergy to local anesthetics, or preexisting motor or sensory deficit in the operated limb.

Each patient received a titrated sedation (midazolam 0–2 mg IV and fentanyl 0–100 μg IV) during performance of the block. No other sedation was provided until evaluation of the block was completed. If required, light intraoperative sedation was provided with propofol to a maximum rate of 50 μg·kg−1·min−1 titrated to maintain verbal contact. No other IV anesthetics was administered unless general anesthesia was required.

After standard monitoring was applied, the neurovascular bundle was imaged in the lateral infraclavicular region with a 50-mm 7.5-MHz fixed-frequency linear ultrasonic scanning probe (Aloka, Tokyo, Japan). The probe was positioned in a parasagittal plane just medial to the coracoid process and adjusted to obtain a cross-sectional image of the axillary artery as it passes under the pectoralis minor. Under direct visualization, with the needle in line with the probe, a 22-gauge 50-mm Teflon-coated insulated needle (Pajunk, Geisingen, Germany) was advanced cauda-posteriorly to the neurovascular space. Block performance time was defined as the time between skin puncture and withdrawal of the needle after injection. In Group U, local anesthetic was injected in a U-shaped distribution posterior and to each side of the axillary artery with as few injections as possible, with the needle initially positioned just posterior to the axillary artery (Fig. 1). The cords of the brachial plexus were not always visualized and were not used to determine appropriate local anesthetic deposition in Group U. If adequate needle placement and anesthetic spread could not be obtained after 20 min, the technique was considered a failure and assigned a performance time of 20 min. In Group S, no specific cord was sought out. Rather, the needle was advanced superior to the axillary artery and then redirected just inferior to the axillary artery if appropriate neurostimulation was not obtained. The entire dose of local anesthetic was injected after obtaining a distal motor response (discrete hand or finger movement, as described by Borgeat et al. (1,9)) using a neurostimulator (Pajunk) with a 1-Hz 0.1-ms stimulation current between 0.3 mA and 0.6 mA. If such a stimulation was not obtained after 20 min, the technique was considered a failure with a performance time of 20 min. The anesthetic solution consisted of 0.5 mL/kg (to a maximum 40 mL) of 1:3 volumes of 0.5% bupivacaine and 2% lidocaine with 1:200 000 epinephrine. The site of injection and the cross-sectional spread of local anesthetic with respect to the axillary artery were noted in each case. Degree of patient discomfort related to the block procedure was reported on a numerical rating scale (0–10). All blocks were performed by the same senior anesthesiology resident (ED); previous experience with each of the two research techniques consisted of 10 blocks performed under supervision by a staff anesthesiologist.

Figure 1
Figure 1:
Figure 1.

The degree of sensory and motor block for each terminal nerve (musculocutaneous, median, radial, and ulnar) was recorded by the same investigator who performed the block every 5 min for 30 min or until complete block of all territories was achieved. Sensory block was self-evaluated by the patients, who compared the cold sensation elicited by ice in the central sensory region of each nerve (12) with the same stimulus delivered to the contralateral side. Patients quantified sensory block in a given territory using any score between 0 and 1, with 0 representing sensation equal to the contralateral side and 1 no cold sensation (for example, if cold sensation was half that of the contralateral side, patients reported a score of 0.5). Motor block was evaluated using forearm flexion, wrist extension, thumb and second digit pinch, and finger abduction (for the musculocutaneous, radial, median, and ulnar nerves, respectively) and scored as follows: no loss of force = score of 0; reduced force compared with the contralateral arm = score of 0.5; and incapacity to overcome gravity = score of 1.

Surgical anesthesia was defined as surgery without patient discomfort or the need for supplementation of the block. If a sensory region involved in the surgery was not completely anesthetized, the block was supplemented in that territory using a neurostimulator-guided mid-humeral approach or locally by the surgeon. If the patient still experienced pain despite supplementation, general anesthesia was induced by the attending anesthesiologist using his preferred technique.

A postblock chest radiograph was obtained if a patient complained of respiratory distress. All patients were hospitalized or observed in a phase 2 recovery unit after their surgery and then discharged with oral analgesics and a detailed care sheet, including contact information and data collection instructions on which they inscribed the timing of the first postoperative analgesic dose. They were contacted 1 wk later by the principal investigator and questioned about: (a) time of first analgesic; (b) persistent paresthesia or motor weakness in the anesthetized limb; and (c) postoperative respiratory difficulty. The duration of postblock analgesia was defined as the interval between block completion and the first postoperative analgesic.

Data are expressed as mean ± 1 sd or proportions with 95% confidence intervals, as applicable. Student’s t-test or the Fisher’s exact test for 2 × 2 contingency tables was used for statistical comparisons. A P value < 0.05 was considered significant. Survival curves were analyzed using statistical software (SPSS, Chicago, IL). The median time to complete block was considered significantly different if the 95% confidence interval of the ratio of the means did not encompass 1. Based on our previous study, assuming a performance time of 5.0 min in the group with neurostimulation, a standard deviation of 2.4 min for both groups, an α of 0.05, and a β of 0.2, it was calculated that a sample size of 36 patients per group would be required to show a difference of 1.6 min in performance time between the two techniques.


Patient demographics and surgical procedures were comparable between groups (Table 1). All but one patient in Group S demonstrated adequate stimulation with discrete finger or hand movements within the 20 min allotted for each block. This patient was classified as a block failure with a procedure time of 20 min and received an alternate technique of brachial plexus block. Block quality was quantified by measuring surgical anesthesia rates, supplementation rates by territory, and the proportion of complete blocks. Figure 2 shows the progression of sensory block by territory over 30 min. Sensory block was significantly better in Group U at 30 min for all territories. Progression of motor block (not shown) paralleled that of sensory block. Figure 3 shows the proportion of patients for whom complete anesthesia of all territories was achieved. The proportion of complete blocks at 30 min was significantly larger in Group U (86%) than in Group S (57%; P = 0.007). The time at which 50% of patients were completely blocked was 15 min in Group U and 20 min in Group S (not significant). Surgical anesthesia rates and supplementation rates by territory are shown in Table 2 (no patient required general anesthesia). In Group U, desired anesthetic spread was achieved with a single injection posterior to the artery in 29 patients (81%), two injections in 6 patients (17%), and three injections in 1 patient (3%). The rate of complete block in Group U was not related to the number of injections (single injection [86%] versus multiple injection [86%]; P = 0.71). The rate of complete block in Group U was significantly better than that in Group S, even when patients in Group U who required more than one injection to achieve desired anesthetic spread were excluded from the comparison (P = 0.01). In Group S, the pattern of local anesthetic deposition was noted to correspond to the position of the needle tip just before injection. Local anesthetic spread posterior to the axillary artery, observed in 63% of patients in Group S, was associated with a relatively large proportion of complete block (79%); anterior deposition, seen in 37% of patients in Group S, was associated with a small proportion of complete block (22%; P = 0.007 versus posterior deposition).

Table 1
Table 1:
Demographic Data and Sedative Administration
Figure 2
Figure 2:
Figure 2.
Figure 3
Figure 3:
Figure 3.
Table 2
Table 2:
Supplementation and Surgical Block Rates at 30 min

Infraclavicular block was more rapidly performed in Group U than in Group S (3.1 ± 1.6 min and 5.2 ± 4.7 min, respectively; P = 0.006). Block-related discomfort was not significantly different between Groups U and S, with pain scores of 2.8 versus 2.5, respectively (P = 0.34); patients in Group U requested slightly less midazolam (mean dose, 0.6 mg versus 1 mg in Group S; P = 0.007). Rates of intraoperative propofol infusion for anxiolysis were not significantly different between groups (Group U, 11%; Group S, 26%; P = 0.1). Two patients in Group S received 100 μg of fentanyl during surgery for back pain; their data were included in all analyses.

At the 1-wk follow-up, patients in both groups requiring postoperative analgesia reported similar time intervals from the block to the first analgesic (Group U, 7 ± 3 h; Group S, 8 ± 5 h; P = 0.17). Complications related to the block included three axillary artery punctures (one in Group S and two in Group U), one patient with shoulder pain lasting 3 days (Group S), and one patient with paresthesia in the median territory lasting 7 days (Group U).


This prospective randomized trial demonstrates that, for infraclavicular brachial plexus block, ultrasonic visualization of local anesthetic spread around the axillary artery as an end point for local anesthetic injection is preferable to a single distal neurostimulation. Procedure times were shorter, and the proportion of complete blocks larger, in Group U.

The success rates obtained in this study are within the range reported in previous series, with values ranging from 40% to 100% for single-injection techniques and 53% to 100% for multiple-injection techniques (7–11,13–17). Differences in reported success rates are the result of variations in technique, operator experience, exclusion of patient subgroups, bias, and different definitions of a successful block. Operator experience in this study was controlled, relatively small, and equal in both groups. Our evaluation of motor and sensory block was not blinded, and bias is a theoretical possibility in the assessment of the number of patients with complete blocks. Success in both groups would have been 100% if defined as surgery without the need for systemic narcotic or hypnotic supplementation or general anesthesia, which is a possibly more objective but less stringent measure of block success.

Increased procedure times in Group S reflect the extra technical effort required to obtain adequate neurostimulation. In addition, despite USG, adequate neurostimulation is occasionally difficult to obtain. This difficulty has been reported by other authors (9,18,19) and may be a consequence of the depth of the nerves at the cord level, where they are not compressed by external pressure and, consequently, may roll freely over the advancing needle. In addition, despite obtaining a neurostimulatory response considered adequate, the proportion of complete blocks was less in Group S than in Group U. When appropriate neurostimulation was achieved, the stimulating needle was kept immobile during local anesthetic injection, as is recommended (20). Interestingly, the proportion of complete blocks in Group S seemed related to the pattern of local anesthetic deposition, with spread anterior to the axillary artery associated with lower rates of complete block than posterior spread. This secondary finding supports the hypothesis that the three-dimensional spread of local anesthetic inside the patient is highly related to the two-dimensional spread observed using USG and builds on those of Porter et al. (19), who described three cases in which failed infraclavicular block was associated with antero-lateral spread and proximal stimulation, whereas successful infraclavicular block was seen with postero-medio-lateral spread, despite the inability to elicit distal neurostimulation. It may therefore be that a stimulation end point indicating proximity to the posterior cord is more appropriate for single-shot neurostimulator-guided infraclavicular block (21). Multiple neurostimulator-guided injections have been shown to improve infraclavicular block quality (14,15,22). However, increasing the number of neurostimulations elicited will further lengthen procedure times, strengthening this advantage of USG alone. A combined end point using both local anesthetic spread visualization and adequate neurostimulation has been proposed. Such an end point could be compared with local anesthetic spread visualization alone in a future study.

In Group U, a U-shaped distribution of local anesthetic around the posterior, medial, and lateral aspects of the artery reliably produced good quality brachial plexus anesthesia without direct visualization of the nerve cords. Previous studies have promoted circumferential spread of local anesthetic around the axillary artery (doughnut sign) by injection on either side of the artery (11) or near each of the three cords (10). These techniques involve two needle punctures to deposit the local anesthetic. In contrast, the technique used in this study resulted in postero-medio-lateral spread in 81% of cases with a single injection at the posterior margin of the artery (5–7 o’clock position, when 12 o’clock is defined as the most anterior point of the axillary artery). In the remaining cases, the needle tip was repositioned along the posterior or posterolateral border of the artery, while generally respecting the original puncture line, to achieve adequate spread. Interestingly, block quality in the subgroup of patients in Group U, in which desired local anesthetic spread was achieved with a single injection, was equal to block quality in the multi-injection subgroup of Group U. Furthermore, block quality was similar to techniques using systematic multiple injections around the artery or next to each nerve cord (10,11), lending further support to the notion that visualization of adequate anesthetic spread, rather than the number of injections required to achieve it, is the key to successful USG blocks.

The need to specifically identify nerves rather than perineural structures in USG regional anesthesia remains controversial (10,23). Ultrasonic identification of nerves depends on multiple factors, including (a) quality and appropriate adjustment of ultrasound probe and machine; (b) depth and echogenicity of the patients’ nerves relative to surrounding tissues; and (c) operator experience. It is our experience, as well as that of others (3,18), that (a) the brachial plexus at the infraclavicular level is not reliably visualized in all patients; (b) neurostimulation is not always successful, even when the stimulating needle is placed immediately adjacent to a nerve (24); and (c) interpreting distal motor responses at the cord level can be difficult (25). Technical improvements in ultrasound technology, such as multifrequency probes and better data processing, are improving image quality and may further improve success rates. Switching to another site for USG brachial plexus block when visualization at the infraclavicular level is difficult may also be a good strategy. This study was performed with an older ultrasound platform and circumvented the problem of precise identification of neural structures in Group U by using the axillary artery as a vascular landmark defining correct anesthetic placement. With this approach, it was found that depositing local anesthetic around the postero-medio-lateral margins of the vessel resulted in a large proportion of complete brachial plexus blocks. This anatomic correlation may be limited to the infraclavicular region. At this time, with older or lower performance ultrasound platforms, an USG technique that relies on an easily visualized vascular landmark is clinically useful. However, this type of approach could become less popular if ultrasound equipment improves in terms of resolution, penetration, and structure differentiation and if more anesthesiologists gain experience interpreting ultrasonic images.

In conclusion, USG infraclavicular block is more rapidly performed and gives a brachial plexus block of better quality when neurostimulation is not used as the end point for injection. Visualization of anesthetic spread posterio-medio-lateral to the axillary artery, most often achieved with a single injection posterior to the axillary artery, reliably produces complete brachial plexus block.


1. Desroches J. The infraclavicular brachial plexus block by the coracoid approach is clinically effective: an observational study of 150 patients. Can J Anaesth 2003;50:253–7.
2. Klaastad O, Lilleas FG, Rotnes JS, et al. Magnetic resonance imaging demonstrates lack of precision in needle placement by the infraclavicular brachial plexus block described by Raj et al. Anesth Analg 1999;88:593–8.
3. Chan VW, Perlas A, Rawson R, Odukoya O. Ultrasound-guided supraclavicular brachial plexus block. Anesth Analg 2003;97:1514–7.
4. Williams SR, Chouinard P, Arcand G, et al. Ultrasound guidance speeds execution and improves the quality of supraclavicular block. Anesth Analg 2003;97:1518–23.
5. Marhofer P, Sitzwohl C, Greher M, Kapral S. Ultrasound guidance for infraclavicular brachial plexus anaesthesia in children. Anaesthesia 2004;59:642–6.
6. Deleuze A, Gentili ME, Marret E, et al. A comparison of a single-stimulation lateral infraclavicular plexus block with a triple-stimulation axillary block. Reg Anesth Pain Med 2003;28:89–94.
7. Jandard C, Gentili ME, Girard F, et al. Infraclavicular block with lateral approach and nerve stimulation: extent of anesthesia and adverse effects. Reg Anesth Pain Med 2002;27:37–42.
8. Arcand G, Williams S, Chouinard P, et al. Ultrasound-guided infraclavicular versus supraclavicular block. Anesth Analg 2005;101:886–90.
9. Borgeat A, Ekatodramis G, Dumont C. An evaluation of the infraclavicular block via a modified approach of the Raj technique. Anesth Analg 2001;93:436–41.
10. Sandhu NS, Capan LM. Ultrasound-guided infraclavicular brachial plexus block. Br J Anaesth 2002;89:254–9.
11. Ootaki C, Hayashi H, Amano M. Ultrasound-guided infraclavicular brachial plexus block: an alternative technique to anatomical landmark-guided approaches. Reg Anesth Pain Med 2000;25:600–4.
12. Miller RD. Miller’s anesthesia. Philadelphia: Elsevier, 2005:1687.
13. Koscielniak-Nielsen ZJ, Rotboll Nielsen P, Risby Mortensen C. A comparison of coracoid and axillary approaches to the brachial plexus. Acta Anaesthesiol Scand 2000;44:274–9.
14. Gaertner E, Estebe JP, Zamfir A, et al. Infraclavicular plexus block: multiple injection versus single injection. Reg Anesth Pain Med 2002;27:590–4.
15. Rodriguez J, Barcena M, Taboada-Muniz M, et al. A comparison of single versus multiple injections on the extent of anesthesia with coracoid infraclavicular brachial plexus block. Anesth Analg 2004;99:1225–30.
16. Minville V, N′Guyen L, Chassery C, et al. A modified coracoid approach to infraclavicular brachial plexus blocks using a double-stimulation technique in 300 patients. Anesth Analg 2005;100:263–5.
17. Kapral S, Jandrasits O, Schabernig C, et al. Lateral infraclavicular plexus block vs. axillary block for hand and forearm surgery. Acta Anaesthesiol Scand 1999;43:1047–52.
18. Perlas A, Chan VW, Simons M. Brachial plexus examination and localization using ultrasound and electrical stimulation: a volunteer study. Anesthesiology 2003;99:429–35.
19. Porter JM, McCartney CJ, Chan VW. Needle placement and injection posterior to the axillary artery may predict successful infraclavicular brachial plexus block: a report of three cases. Can J Anaesth 2005;52:69–73.
20. De Andres J, Sala-Blanch X. Peripheral nerve stimulation in the practice of brachial plexus anesthesia: a review. Reg Anesth Pain Med 2001;26:478–83.
21. Bloc S, Garnier T, Komly B, et al. [Efficiency of secondary posterior trunk single stimulation, low volume infraclavicular plexus block for upper limb surgery]. Ann Fr Anesth Reanim 2005;24:1329–33.
22. Rodriguez J, Barcena M, Lagunilla J, Alvarez J. Increased success rate with infraclavicular brachial plexus block using a dual-injection technique. J Clin Anesth 2004;16:251–6.
23. Greher M, Retzl G, Niel P, et al. Ultrasonographic assessment of topographic anatomy in volunteers suggests a modification of the infraclavicular vertical brachial plexus block. Br J Anaesth 2002;88:632–6.
24. Urmey WF, Stanton J. Inability to consistently elicit a motor response following sensory paresthesia during interscalene block administration. Anesthesiology 2002;96:552–4.
25. Borene SC, Edwards JN, Boezaart AP. At the cords, the pinkie towards: interpreting infraclavicular motor responses to neurostimulation. Reg Anesth Pain Med 2004;29:125–9.
© 2007 International Anethesia Research Society